Abstract
Plenary Session I: Imaging Approaches for Biology and Medicine
Abstract ID: 001
The explosion of progress in the fields of cell biology, biochemistry, and molecular biology has offered unprecedented knowledge of the components involved in embryonic development. This dramatic progress poses the challenge of integrating this knowledge into an understanding of the developmental mechanics that pattern and construct the embryo. Advanced imaging techniques offer an important stepping stone between these disparate approaches, permitting questions about cellular and molecular signaling events to be posed in the most relevant setting of the intact embryo. Confocal laser scanning microscopy or two-photon laser scanning microscopy can be used to follow cells as they migrate in the intact embryo. To validate this in vivo data, we have been employing a novel form of sectioning microscopy, in collaboration with Resolution Sciences Corporation. The technique automatically sections through the tissue at 1-mm increments, capturing a fluorescent image of the block face after each section. Because the block face is not deformed by the sectioning, these images can be easily registered and computationally reconstructed to yield the full three-dimensional volume of the tissue at subcellular resolution. In vivo imaging of multiple labels should offer the ability to test the roles proposed from the time-lapse and fixed specimen imaging. GFP color variants offer the possibility of following multiple molecular species in the same cell, and fluorescence resonance energy transfer (FRET) allows molecular interactions to be assayed as it takes place. Multispectral approaches involve acquisition of the spectrum of the emitted light from each pixel, followed by decomposition into its component parts by simple mathematics. Fluorochromes as similar as GFP and fluorescein can be separated unambiguously, and even small amounts of FRET can be detected by our approach.
S.E. Fraser, Zeiss 6; Metaprobe 4.
Abstract ID: 002
At the end of the nineteenth century, Ramon y Cajal's application of the Golgi method revolutionized neurobiology by showing that the neuron is the brain's organizational unit. For most of the twentieth century, Cajal's wonderful drawings and analysis remained the last word on the cellular organization of many regions of the adult and developing brain. However, all this is changing due to the convergence of 2 new sets of technical developments. First are the methods of genome manipulation that have made it possible to insert genes from jellyfish and other aquatic creatures that encode fluorescent proteins stably into lines of transgenic mice that express these fluorescent proteins in the brain. Second is the development of a suite of confocal, multiphoton, low-intensity, and computational methods that allow imaging of fluorescent neurons in living animals at a resolution previously obtainable only in thin sections of fixed tissue. Together, these methods now make it possible to label different neurons in different colors so that synaptic circuitry can be untangled. Moreover, it is now possible to view neurons in living animals so that the same individual cells and synapses can be monitored over minutes or months as they change in response to experience, aging, or disease. Perhaps the most exciting uses of such transgenic animals is that they provide the first way to assay the cellular alterations that underlie behavioral changes such as memory formation. We have used these fluorescent mice to monitor dramatic remodeling of synaptic circuits that takes place in early postnatal life. This remodeling is likely central to the way young mammals use experience to mold their nervous systems to conform to the world they live in.
J.W. Lichtman, None.
Abstract ID: 003
In has taken many years to conceptualize addiction as a disease of the brain that has similarities to many other chronic and relapsing illnesses. Many of the major advances in understanding how drugs of abuse affect the human brain, as well as the delineation of brain circuits that are disrupted in addiction have come from molecular imaging. Advances in radiotracer design, chemistry, and quantitation have made it possible to image dopamine and other neurotransmitter systems and to assess their role in behavior, in disease, and in mediating the actions of drugs of abuse. We have used positron emission tomography (PET) imaging and carbon-11-labeled drugs, as well as radiotracers that selectively bind to the dopamine transporter, the dopamine D2 receptor, and monoamine oxidase to probe both the distribution and pharmacokinetics, as well as the effects of drugs of abuse on the human brain. Many of these studies have been coupled with PET measures of brain function using 18FDG, a fluorine-18-labeled glucose derivative, which measures glucose metabolism in all brain regions simultaneously. This multitracer approach has provided new insights on the neurochemical systems that are linked to the behavioral and toxic effects of abused substances and has suggested new approaches to treatment. (Supported by the U.S. Department of Energy, Office of Biological and Environmental Research and the National Institutes of Health (NIDA and NINDS)).
J.S. Fowler, None.
Abstract ID: 004
We are endeavoring a gene therapy approach for carcinoma of the breast based upon selective transduction of target cells. In our initial studies, we addressed this technical goal via a novel “double targeting” strategy whereby transgene expression was directed selectively to target cells employing a combination of transductional and transcriptional targeting approaches. This strategy was designed to mitigate the native hepatotropism of the employed adenoviral vector while also augmenting transduction of the target cells. In the present period, we strove to studies to further optimize the targeting capabilities of our approach towards realizing a favorable therapeutic index in the context of our planned gene therapy intervention. Of note, we have heretofore monitored induced transgene expression (luc via luminometery, cea via immunohistochemistry) but have not had the means to directly monitor viral particle trafficking. As particle and induced transgene profiles may be affected by our retargeting approaches, we recognize that this key parameter required an assay approach. On this basis, we have developed a novel method based on direct incorporation of the fluorescent protein egfp into the ad capsid protein pix. The derived particles retain their full vector capacities but embody the means to be monitored by fluorometric methods via direct fluorescent microscopy. This powerful new approach will thus allow us to study in vivo vector trafficking in our in vivo models. Such analysis will provide a key guide rationalizing our vector efforts towards the goal of breast cancer specific delivery.
D.T. Curiel, None.
Plenary Session II: New Targets and Novel Strategies
Abstract ID: 005
Cancer is a genetic malady, mostly resulting from acquired mutations and epigenetic changes that influence gene expression. Accordingly, a major focus in cancer research is identifying genetic markers that can be used for precise diagnosis or therapy. Over the last half century, investigators have used reductionism to discover such markers through the study of simple genetic changes like balanced chromosomal translocations. For example, fundamental insights into the nature of the bcr-abl gene translocation product resulted in the precise molecular classification of chronic myelogenous leukemia and recently led to the development of the abl-targeted tyrosine kinase inhibitor Gleevec for the treatment of this disease. Ninety percent of human cancers, however, are epithelial in origin and display marked aneuploidy, multiple gene amplifications and deletions, and genetic instability, making the resulting downstream effects difficult to study with traditional methods. Because this complexity probably explains the clinical diversity of histologically similar tumors, a comprehensive understanding of the genetic alterations present in all tumors is required. The initial sequencing of the human genome, coupled with technologic advances, now make it possible to embrace the genetic complexity of common human cancers in a global fashion. Tools are currently available, or are being developed, for the identification of all changes that take place in cancer at the DNA, RNA, and protein levels. We have used oligonucleotide microarrays to obtain global views of gene expression in multiple human cancer types and have developed analytic methods that permit the accurate, multiclass, molecular diagnosis of human cancer. We have also used information in this large database to develop molecular predictors of metastasis and treatment response and to gain insight into molecular mechanisms that are operative in tumor growth. These studies and others suggest that information-rich genomic approaches will prove important in the development of novel diagnostic and therapeutic strategies for human cancer.
S. Ramaswamy, None.
Abstract ID:
Blood vessels of tumors have multiple abnormalities that impair blood flow, influence delivery of chemotherapeutics, and provide potential targets for imaging agents and angiogenesis inhibitors. These abnormalities involve all components of the vessel wall: endothelial cells, pericytes (mural cells), and basement membrane. Defective endothelial barrier function is one of the best documented abnormalities of tumor vessels. Blood vessel leakiness is important because it facilitates angiogenesis and enables antibodies and other macromolecules to reach tumor cells from the bloodstream. Yet, vessel leakiness also contributes to high interstitial pressure that restricts drug delivery to tumor cells. Also, because of heterogeneity of the leakiness, antibodies and other proteins extravasate in a patchy distribution and do not have uniform access to tumor cells. By comparison, antibodies directed against integrins and other membrane proteins that are selectively expressed on tumor vessels can readily access their target. In vivo phage display is a powerful approach for identifying membrane proteins that are potential vascular targets. Libraries of filamentous bacteriophage, genetically engineered to express random peptides, can be injected into the bloodstream to identify cognate receptors on tumor vessels. Still, endothelial permeability determines whether endothelial cells, pericytes, and tumor cells are all potential targets of the injected phage. Microscopic imaging studies are revealing the cellular distribution and accessibility of molecular targets in tumors and how the abnormalities of tumor vessels can be exploited in cancer therapy.
D.M. Mcdonald, None.
Abstract ID: 007
Peptides are ubiquitous ligands in biology, and therefore hold great promise as starting points for the discovery of new therapeutics and imaging agents. However, peptides often suffer from pharmacologic liabilities that limit their effectiveness in animals. Our laboratory has developed chemistry-based technologies to “morph” peptides into more bioactive polyketide-like structures and to stabilize peptides in defined secondary structures that can penetrate cells and are resistant to proteolyic degradation. Progress on these fronts will be reviewed.
G.L. Verdine, Enanta Pharmaceuticals 1, 2, 4.
Abstract ID: 008
A common goal of molecular therapeutics is to specifically deliver therapy or imaging agents to target cells. In nature, multiple and sequential binding events are a common theme to establish and ensure specificity of interaction. There are numerous examples, such as virus-host, cell-cell, and DNA-DNA interactions. Specificity and avidity are obtained through multiple weak cooperative binding events. In this work, the objective is to develop agents that are reverse-engineered against the expressed epitope profiles of target cells. These agents contain multiple independent ligands directed against combinations of epitopes defined by a combined genomic-proteomic approach. Combinations of up to 7 different epitopes are theoretically possible, yielding trillions of distinct potential targets. We have investigated gene expression in a series of Pancreatic Cancer cell lines, pancreatic cancer biopsies and normal pancreas using a 4000 named gene array. From these data, a set of genes have been identified that are appropriate as targets. This is followed up with Affimetrix arrays and proteomic analyses using isothiocyanate labeling of cell surface epitopes. As proof-of-principle, multimeric ligands have been developed to contain GPCR ligands: oxytocin, MSH, CCK, and enkephalin. These are labeled with lanthanide (e.g., Eu) chelates and assayed using time-resolved fluorescecne with attomole sensitivity. Alternatively, ligands are labeled with fluorophores (e.g., Cy 3) and subcellular fate determined by multispectral microscopy. Proof-of-principle is established by determining the avidity and cooperativity of ligands directed against cells engineered to express defined combinations of GPCR.
R.J. Gillies, None.
Symposium I: Advances in Radiochemistry and Hyperpolarized Gases
Abstract ID: 009
The successful effort to sequence the human genome has ushered in an entirely new era of medicine. In order to use the information derived from this effort, we must now map the genetic information to specific phenotypes. The mouse and rat are the most amendable models for this task. This talk will focus on image-based methods to phenotype the rat and mouse using magnetic resonance and X-ray microscopy. The challenges in moving imaging methods from man to mouse are best viewed in the context of the relative volumes. The mouse at 25 g is nearly 4000 times smaller than the man at ˜100 kg. The problem in scaling clinical imaging technologies to the mouse is further exacerbated by the temporal scale. The R-R interval in the mouse is ˜100 msec, nearly 10 times shorter than that in the man. No single “trick” will allow one to image the mouse with the necessary temporal and spatial resolution. One must instead use an integrated approach. We have applied this approach in both functional and structural phenotyping using both magnetic resonance microscopy and X-ray microscopy. Technical approaches and the applications that can benefit will be included.
G. Johnson, MRPath, Inc. 4.
Abstract ID: 010
The base for the design of an imaging probe for a given application is first dictated from the chosen Imaging modality, which, in turn, is dependent upon the concentration and localization (vascular, extracellular matrix, cell membrane, intracellular) of the target molecule. MRI-Gd(III) agents are much less sensitive than radionuclear and optical imaging probes. Therefore, molecular imaging approaches based on MRI invariantly involve the need of accumulating a high number of contrast enhancing units at the site of interest. Our work has aimed at exploring different routes to the entrapment of Gd(III) complexes in various types of cells at amounts sufficiently large to allow MRI visualization. Namely, the obtained results can be summarized in terms of internalization via (i) phagocytosis, (ii) pinocytosis, and (iii) receptor-mediated endocytosis. Often, the amount of Gd(III) that can be entrapped in each cell is remarkable. Much work has been done in order to enhance the uptake of Gd(III) complexes into tumor cells. Monomeric and multimeric Gd(III) complexes targeting hyperexpressed receptors or up-regulated transporters of nutrients and pseudo-nutrients have been developed.
Abstract ID: 011
MRI is the only imaging modality that allows “real molecular imaging” since NMR is the only in vivo modality allowing information about the structure of molecules. Other modalities only give information on where the reporter group or atom is localized within the body. In SPECT or PET, the reporting unit is a radiant atom (In, F, etc.). In light imaging, it is a reporting unit. MRI in general is not considered sensitive enough to generate relevant information on the events at the molecular/cellular level. Hyperpolarization of C-13 has dramatically increased the potential. We have developed C-13 polarization techniques and imaging techniques that makes it possible to follow the specific molecules injected. The method allows to follow the fate of a large number of biologically/medically interesting molecules. It is concluded that it is finally possible to do real molecular imaging.
K. Golman, Amersham Health R & D AB, Sweden 5.
Abstract ID: 012
During the last decade, peptide radiopharmaceuticals have become an important class of tracers for the detection and localization of malignant neoplasms by peptide receptor imaging and for therapeutic intervention by peptide receptor radiotherapy. Various radiometallated peptides have entered clinical studies or found broad application. In contrast, radiohalogenated peptides could not benefit from this development. Especially with respect to the growing number of new peptidic structures with high receptor affinity and the increasing demand for tracers for quantification of the receptor status for therapy planning and control, the development of improved methods for production of suitable PET tracers (i.e., 18F-labeled peptides) and modification of their in vivo behavior is still an important objective in radiopharmaceutical research. Consequently, as part of our ongoing efforts in this field, new labeling strategies and methods to improve the in vivo behavior of peptide radiopharmaceuticals were investigated. Carbohydration of peptides was found to be a valuable tool to modify the pharmacokinetics, that is, excretion pathways, blood clearance, tumor uptake, unspecific binding, or renal uptake, leading to excellent in vivo characteristics and imaging quality of radiohalogenated peptides. Furthermore, the development of homo- and heterodimeric and multimeric peptides seem to be a promising strategy to improve targeting with respect to the heterogeneity of peptide receptor expression in neoplasms. New labeling methods were evaluated allowing chemoselective two-step radiohalogenation of unprotected peptides. The most promising approach recently developed will allow large-scale production of 18F-labeled peptides within less than 1 hr in up to 50% overall yields in a chemoselective manner. The combined approach, chemoselective conjugation of unprotected, carbohydrated peptides seems to have the potential for a redirection and reevaluation of the future of radiohalogenated peptides in nuclear medicine.
H.J. Wester, None.
Symposium II: Advances in Systems Architectures and Informatics
Abstract ID: 013
Selection of in vivo imaging modalities (i.e., X-ray, radionuclide emission, ultrasound, magnetic resonance, light absorption, fluorescence, bio-luminescence) can be logically approached by evaluating the following parameters relative to a biomedical imaging objective (e.g., fibrin on plaques vs. plaque calcium; brain inflammation vs. brain atrophy): spatial resolution, temporal resolution, signal to noise, sensitivity, accessible image field, effective tissue depth, object size, radiation dose, biologic safety, contrast sensitivity. Contrast sensitivity, of highest importance for modality selection in most cases, is defined as the signal to background for the desired biochemical or physiological parameter and is a complex function of target specific kinetics of the molecular probe to background concentration of the probe. The objective of this article is to present examples of trade-offs that are at the essence of some new instrumentation for imaging vulnerable atherosclerotic plaques, small animals with both fluorescent and radioactive probes simultaneously, and specialized organ specific imaging instruments (breast and prostate). Specifically, the method for evaluating the presence of a vulnerable plaque requires evaluation of the information supplied by competing modalities. High-resolution MRI may give somewhat specific information on tissue characteristics, but radionuclide imaging and molecular beacons have the potential to signal the presence of fibrin, though at a spatial resolution 10 times less than MRI, yet, a contrast recovery for the presence of fibrin which is specific and many times greater than the contrast sensitivity of MRI. Sensitivity, limited angle reconstruction, background, safety, and biokinetics of the probe have a bearing on whether fluorescence or radionuclide emission methods are optimal. A design for carotid plaque biologic imaging is shown in Figure 1.
High sensitivity Fresnel zone plate imaging of a vulnerable plaque.
T.F. Budinger, None.
Abstract ID: 014
A range of systems and technologies for applications in molecular imaging are being developed in our laboratory. These include a high-resolution microPET scanner, an integrated microPET/microCT scanner, a MRI compatible PET system, and a system for 3-D in vivo fluorescence imaging. The design and performance of these imaging systems will be discussed and several examples will be shown to highlight how these technologies might be utilized to open up new opportunities in the field of molecular imaging.
S.R. Cherry, Concorde Microsystems, Inc. 2.
Abstract ID: 015
Local hyperthermia has been suggested for many purposes, for example, tumor ablation, control of gene therapy, local drug delivery, heat-activated chemotherapy. Focused ultrasound (FUS) with a wavelength of about 1 mm is capable of noninvasively depositing energy in a target area deep within the body without harming neighboring tissues. However, FUS energy absorption and heat conduction depend on tissue composition and physiological processes like perfusion. Hence, continuous thermometry of the target area is necessary. MRI is an ideal tool for guiding FUS because of its unique temperature mapping capabilities, and its soft tissue contrast for target definition. Temperature MRI can be used for guidance of FUS based on the temperature dependence of T1, diffusion, or the proton resonance frequency (PRF) of water. A nonmagnetic focused ultrasound transducer was incorporated in a Philips Intera 1.5 T MRI. Automatic MR control of the FUS hyperthermia was developed employing: (1) fast PRF-based temperature mapping methods; (2) lipid suppression; (3) motion artefact suppression; (4) real-time data processing and visualization; (5) automatic feedback-controlled FUS power to guarantee a predefined temperature trajectory in the FUS focal point. The initial hardware can be further refined by using phased-array ultrasound transducers, allowing rapid steering of the focal spot and proper focusing of the beam even for inhomogeneous tissues. Much attention has to be paid to use MR compatible materials and filtering methods allowing simultaneous MRI and high power ultrasound transmission. Preliminary applications have concentrated on gene therapy, local drug deposition and ablation. The combination of a thermosensitive promoter (like the human heat shock protein, hsp70, promoter) and local hyperthermia was demonstrated to allow local control of transgene expression in a modified glioma tumor cell line. The method may offer a noninvasive way of controlling, in space and in time, the expression of therapeutic proteins in gene therapy.
C.T. Moonen, Philips Medical Systems 1, 2.
Abstract ID: 016
Several years ago, we showed that dipolar couplings between distant spins permit NMR spectra and MR images where the contrast comes from intermolecular multiple-quantum coherences—simultaneous flips of spins in widely separated molecules (typical separations are 10-100 μm) [1–3]. The contrast provides a direct measure of the spatial distribution of the magnetization density and the resonance frequency variations on a distance scale much smaller than the voxel size, and this has been shown to be useful in tumor detection and functional MRI [4]. Here we present a new approach that permits simultaneous acquisition of double-quantum, zero-quantum, and (conventional) single-quantum images. This enhances signal intensity, highlights subtle variation in image contrast, and allows clean distinctions between the various contrast mechanisms frequently present in biological tissue. Applications to phantoms and small animals will be presented. (This work is supported by the Center for Molecular and Biomolecular Imaging and by the NIBIB/NIH under grant EB2122.)
Symposium III: Genomics, Proteomics, and Reporter Gene Strategies
Abstract ID: 018
1Department of Genetics, Harvard Medical School, USA; 2Harvard Medical School, USA; 3Fred Hutchinson Cancer Research Center, USA.
Within a developing tissue, the intersections of gene expression domains define unique coordinate positions and gene expression codes, providing an important means for patterning cell types. As a paradigm for studying how such intersecting axes of gene expression relate to cell fate, we report the simultaneous use of Cre and Flp recombinases to engineer genetic manipulations in only those cells occupying a defined coordinate position in the developing mouse neural tube. Central to this approach is a dual recombinase responsive indicator mouse (designated R26FLAP) that we have developed in which expression of a marker, placental alkaline phosphatase (PLAP), is conditional on both Flp- and Cre-mediated excision events. Using rhombomere-specific Cre alleles and a dorsal neural tube-specific Flpe allele, which together subdivide the embryonic hindbrain along orthogonal gene expression axes, we show how the R26FLAP dual responsive indicator strain can be used to study specific populations within populations of cells, here the dorsal-most cell populations within individual rhombomeres. In compound animals harboring Cre, Flpe, and R26FLAP alleles, we show that virtually all cells occupying the intersection of the Cre and Flpe expression domains indeed switch-on expression of PLAP. By following the fate of marked cells, we have revealed unexpected features of the development of the critical organizing center, the hindbrain roof plate. This intersectional gene activation method, its derivative tools, and modifications to them should prove useful in studies of many other complex biological processes.
S.M. Dymecki, None.
Abstract ID: 019
The field of proteomics aims to characterize dynamics in protein function on a global scale. However, several classes of enzymes are regulated by posttranslational mechanisms, limiting the utility of conventional proteomics techniques for the characterization of these proteins. Our research group has initiated a program aimed at generating chemical probes that interrogate the state of enzyme active sites in whole proteomes, thereby facilitating the simultaneous activity-based profiling of many enzymes in samples of high complexity. Progress towards the generation and utilization of active site-directed chemical probes for the proteomic characterization of several enzyme classes will be described. These enzyme classes fall into two general categories: (1) enzymes for which active site-directed affinity agents have been well-defined, and (2) enzymes for which active site-directed affinity agents have been lacking. The application of activity-based protein profiling to the functional characterization of enzyme activities that vary in human cancer specimens will be highlighted, as will be the use of this strategy as a screen to discover potent and selective reversible enzyme inhibitors.
B.F. Cravatt, Activx Biosciences 2, 4.
Abstract ID: 020
We are developing technologies for genome-wide mapping of gene expression patterns in the brain. These technologies combine the mathematical concepts of biomedical imaging with the high-throughput tools of genomics. One method, voxelation, employs analysis of spatially registered cubes or voxels. Another method, gene expression tomography, employs analysis of sets of parallel slices rotated about multiple axes of rotation. Both approaches use reconstruction to provide multiple volumetric images of gene expression, reminiscent of the images obtained from biomedical imaging systems.
D.J. Smith, None.
Symposium IV: Endogenous Optical Signatures and Exogenous Probe
Abstract ID: 021
Low coherence interferometry and angular light scattering spectroscopy can both be used to obtain optical thickness images with resolution of 1 nm or less. As opposed to OCT, these techniques are based on spectroscopy. The concept of an all optical microscope capable of providing such 3-D images of living tissues, cells and organelles, and their dynamics, in their native state, without fixation or other processing, will be presented. Examples will be given.
M.S. Feld, None.
Abstract ID: 022
Intravital microscopy is a powerful method for visualizing cellular interactions in the native environment of living animals and is particularly useful for monitoring leukocyte trafficking in vivo. Leukocyte trafficking is tightly controlled by the adhesion molecules expressed on vascular endothelial cells. However, the endothelial cells are usually not visualized in conventional (epifluorescence) intravital microscopy. The aim of our work is to develop methods for imaging endothelial cell surface markers and studying leukocyte-endothelial interactions in vivo. We have constructed a video-rate confocal/two-photon fluorescence microscope that is capable of taking high-resolution images (optical sections) noninvasively through the intact skin of live mice at depths of greater than 200 μm into tissue. The optical sectioning capability allows us to image vascular endothelial cells labeled with fluorescent antibodies, while rejecting tissue autofluorescence background. Using antibodies against specific cell adhesion molecules, we are able to image the differential expression of inducible markers such as E-selectin before and after LPS-induced inflammation. Separate confocal and two photon detection channels allow in vivo double labeling experiments using more than one molecular probes.
C.P. Lin, None.
Abstract ID: 023
Near-infrared (NIR) fluorescence imaging is usually used in concert with molecularly targeted contrast agents that not only provide enhanced contrast but also, more importantly, reveal specific molecular events associated with cancer formation and progression. The specificity of a novel epidermal growth factor (EGF)-Cy5.5 fluorescent optical probe in the detection of EGF receptor was assessed using continuous-wave fluorescence imaging accomplished via an intensified CCD camera. Human mammary MDA-MB-468 (EGFr+) and MDA-MB-435 (EGFr–) cancer cells were incubated with Cy5.5, EGF-Cy5.5, or the anti-EGFr monoclonal antibody C225 or EGF followed by EGF-Cy5.5 and were examined under a fluorescence microscope. In vivo imaging was performed on mice with subcutaneous MDA-MB-468 and MDA-MB-435 tumors. Images were obtained every 6 sec for 20 min after intravenous injection of Cy5.5, indocyanine green (ICG), or EGF-Cy5.5 (1.0 nmol eq Cy5.5/mouse) and every 24 hr after injection for up to 192 hr. Additionally, mice with MDA-MB-468 tumors were injected iintravenously with C225 24 hr before injection of EGF-Cy55. EGF-Cy55, but not Cy55 or ICG, bound to MDA-MB-468 cells. Binding of EGF-Cy55 was blocked by C225 and by EGF. In contrast, binding of EGF-Cy5.5 to MDA-MB-435 cells was not observed. Monitoring of the time-sensitive fluorescence intensity in mice confirmed that ICG and Cy5.5 had no favorable binding to tumor regardless of EGFr expression level. In contrast, EGF-Cy5.5 accumulated only in MDA-MB-468 tumors. Moreover, tumor uptake of EGF-Cy5.5 was blocked by C225, demonstrating specificity of the targeted contrast agent. ICG and Cy5.5 fluorescence was completely absent from the tumor site, regardless of EGFr expression level, 24 hr after injection. Little EGF-Cy5.5 fluorescence was detected in MDA-MB-435 tumors 24 hr after injection. Our data suggest that EGF-Cy5.5 may be used as a specific NIR contrast agent for noninvasive imaging of EGFr expression and monitoring of responses to molecularly targeted therapy.
C. Li, None.
Abstract ID: 024
Diffuse Optical Spectroscopy and Imaging (DOS, DOI) are non-invasive diagnostic techniques that employ near-infrared (NIR) light to quantitatively characterize the optical properties of thick, multiple-scattering tissues. Although NIR was first applied to breast diaphanography more than 70 years ago, quantitative optical methods employing time- or frequency-domain photon migration technologies have only recently been used for optical mammography. This presentation reviews principles of photon migration (i.e., Diffuse Optics) and describes the development of broadband DOS for quantitatively measuring the bulk absorption and scattering properties of thick tissues. The spectral dependence of optical properties derived from broadband DOS can be used to determine tissue hemoglobin concentration (total, oxy, and deoxy forms), tissue hemoglobin oxygen saturation, blood volume fraction, water content, lipid content, and cellular/matrix density. Clinical study results will be shown, highlighting DOS sensitivity to breast tissue metabolic changes associated with aging, hormonal stimulation, tumor growth, and chemotherapy. These findings will be placed in the context of conventional breast imaging methods in order to assess the role of optics in breast cancer research and clinical management.
B.J. Tromberg, None.
Symposium V: Molecular Imaging of Targets in the Vasculature
Abstract ID: 025
Atherosclerosis is the leading cause of death in Western civilization. To date, the development of targets for molecular imaging has been hindered by a limited understanding—specifically in the blood vessel wall—of the molecular and genetic basis of this disease. Therefore, a custom vascular cell-specific microarray has been constructed with ink-jet technology and employed to identify genes differentially regulated in the atherosclerotic vessel. This cDNA microarray consists of approximately 8000 genes, including 5206 known genes and 1360 ESTs. These genes are primarily derived from cDNA libraries generated by suppression subtraction hybridization methods to enrich for genes differentially expressed by human smooth muscle cells treated with cytokines and growth factors implicated in atherosclerosis. In vitro and in vivo studies are being performed in parallel to investigate this candidate gene set through expression profiling. In vitro studies include stimulation of individual vascular cell types and cocultured vascular cell combinations with physiological and hormonal factors, as well as biophysical forces. Human atherosclerotic coronary tissue specimens of varying disease stage are being harvested from heart transplant patients and for used for chip analysis. Individual gene expression profiles are being evaluated in the context of each subject's clinical history, risk factor status, and vascular lesion classification, all of which is maintained in a comprehensive relational database. This approach is yielding new specific targets that are being evaluated with regard to cell specificity and appropriateness for molecular imaging utilization.
T. Quertermous, None.
Abstract ID: 026
The endothelial cells that line tumor blood vessels represent an attractive target for the delivery of therapeutic and imaging agents to tumors. However, few molecular markers have been identified that can distinguish normal endothelium from angiogenic endothelium associated with disease. In order to identify markers of tumor endothelium, we compared global gene expression patterns in endothelium isolated from either normal colonic mucosa, or colorectal cancer. These studies led to the identification of 5 novel cell surface tumor endothelial markers, called TEM1, TEM5, TEM7, TEM7R, and TEM8. To begin exploiting the products of these genes, we have begun to generate a panel of monoclonal and polyclonal antibodies to each of these cell surface TEMs. Importantly, cell surface TEM proteins appeared to be expressed predominantly in tumor endothelium by various techniques including immunohistochemical and immunofluorescence staining. These studies demonstrate that cell surface TEMs maintain preferential expression in tumor endothelium at the protein level, and thus represent attractive new targets for the delivery of novel therapeutic and imaging agents to tumors.
B. St. Croix, Genzyme 6.
Abstract ID: 027
To refine the search for vascular targets in solid tumors, we analyzed gene and protein expression patterns in human glioblastoma multiforme using contrast-enhanced magnetic resonance imaging (MRI) to sample highly vascularized angiogenic regions within the tumor. As an internal control, we compared the gene and protein expression profiles of samples from 10 human glioblastoma patients obtained from contrast-enhancing and nonenhancing regions within the same tumor using MRI and functional genomics and proteomic analysis. We have analyzed 12,000 genes from these samples and compared 2-D gels using DIGE (2-D fluorescence difference gel electrophoresis) and mass spectral analysis. We have found unique proteins that are up-regulated in the contrast-enhancing areas of several tumors and confirmed these proteins using immunohistochemical analysis. Several of these proteins encode extracellular matrix and two proteins (aFGF and laminin receptor) are associated with the tumor vasculature. This study demonstrates that MRI can serve as a powerful, noninvasive tool for characterizing vascular and angiogenic regions of tumors to guide genomic and proteomic analysis in the search for new vascular targets.
M. Bednarski, None.
Abstract ID: 028
Molecular medicine has discovered many targets leading to therapies that work in vitro but not in vivo where barriers restrict access. Limited progress has been made in achieving tissue-specific drug and gene delivery in vivo in part because of poor accessibility caused by significant barriers existing in vivo. Using our recently developed subfractionation techniques for purifying luminal endothelial cell plasma membrane and caveolae directly from normal organs, as well as tumors, we have produced in different tissues high-resolution proteomic maps of the endothelial cell surface directly in contact with the circulating blood. We have used 2-D gel electrophoresis, immunoblotting and mass spectrometry to develop a proteomic database of accessible vascular targets. Tumor-induced neovascular targets have also been identified. These newly discovered targets have been used to image the vasculature of whole organs and tumors selectively after intravenous injection of specific antibodies. Intravital microscopy reveals transendothelial transport and improved tissue penetration for those antibodies targeting caveolae. Electron microscopy shows explicit targeting of caveolae and subsequent rapid transcytosis for release to underlying tissue cells. Intravenously injected antibodies permit rapid visualization of single organs or solid tumors in vivo by SPECT. At 30 min after injection, up to 90% of the antibody can accumulate in a single tissue. Our novel profiling strategy has identified distinct molecular signatures for the endothelial cell surface in each organ and in solid tumors. Profiling accessible vascular and caveolae targets is an important logical step not only for achieving site-directed pharmacodelivery and selective molecular imaging of normal organs as well as tumors in vivo but also for overcoming the normally restrictive endothelial cell barrier to transport drugs and possibly even genes to their intended targets, the underlying tissue cells.
J.E. Schnitzer, None.
Symposium VI: Imaging And Therapy Combinations
Abstract ID: 029
Positron emission tomography (PET) allows a noninvasive assessment of molecular mechanisms in humans and animals in vivo. With the developments in tracer technology, a variety of endogenously expressed and exogenously introduced genes can be analyzed by PET. The intriguing advantage of molecular imaging is the kinetic analysis of a given molecular event in the same experimental subject over time. This allows a noninvasive characterization and “phenotyping” of animal models of human disease at various disease stages, under certain pathophysiological stimuli and after therapeutic intervention. Important endogenous enzymes and receptors in neuroscience include: cellular hexokinase ([18F]FDG) and thymidine kinase ([18F]FLT), aromatic aminoacid decarboxylase ([18F]FDOPA), acetylcholine esterase ([11C]MP4A), dopamine D2 receptor ([11C]RAC), benzodiazepine receptor ([11C]FMZ), and amino acid transporter ([11C]MET). The respective tracers are applied to elucidate pathophysiological mechanisms, to monitor disease progression, or to provide diagnostic or prognostic information in neurological diseases such as neurodegeneration, movement disorders, gliomas, epilepsy, or cerebral ischemia. For example, they are being used for early detection of Alzheimer's disease ([18F]FDG, [11C]MP4A), in the differentiation of Parkinson's disease from multiple systems atrophy ([18F]FDOPA, [11C]RAC, [18F]FDG), in the grading of gliomas and differentiation of radionecrosis from recurrent tumor ([18F]FDG, [18F]FLT, [11C]MET), in the exact localization of epileptogenic foci in partial epilepsy syndromes ([18F]FDG, [11C]FMZ), and in the assessment of neuronal integrity after stroke ([11C]FMZ, [18F]FDG). The overall attractive goal in neuroscience research is that with the advent of animal PET, these tracers can be applied in animal models for human disease in the further development of experimental therapeutics, such as gene and stem cell therapies. Most importantly, the same in vivo imaging parameters can be used preclinically and clinically enabling a direct comparison when translating experimental molecular imaging markers into clinical application. (Supported in part by MSWF 516–40000299, ZMMK-TV46 and DFG JA 981/1–2.)
A. Winkeler, None.
Abstract ID: 030
Advances in molecular biology and cellular biochemistry are providing new opportunities for diagnostic medical imaging to “see” beyond the anatomical manifestations of disease to the earliest biochemical signatures of disease. Today, a broad array of molecular contrast systems are emerging, which are designed to specifically recognize, bind, and amplify the signal from biomarkers uniquely reflective of underlying pathological processes. Although molecular imaging was once the province of nuclear medicine, today numerous, highly active research programs are found for all clinically relevant noninvasive imaging modalities. Moreover, these site-directed molecular imaging systems are becoming hybridized with therapeutic agents to provide local drug delivery, which can be confirmed and quantified noninvasively at the target of interest. We have developed a novel multimodal site-targeted contrast agent for sensitive and specific imaging of molecular epitopes and local therapy. This “platform” approach comprises a nanoparticle that is applicable to ultrasound, nuclear, CT, and magnetic resonance. This novel platform has been used to detect angiogenesis, fibrin, tissue factor and collagen III and to locally deliver a variety of therapeutic agents through a novel “contact-facilitated” mechanism. Molecular imaging platforms for diagnostic application are or soon will be in clinical trials followed shortly thereafter by image-based targeted pharmaceuticals. These advances in diagnostic and therapeutic contrast systems in conjunction with complementary advancements in medical imaging equipment promise to catapult current clinical practice to new standards of care within the next decade.
G.M. Lanza, Kereos, Inc. 4.
Abstract ID: 031
Several studies have demonstrated that the assessment of response to therapy in cancer is diagnostically challenging. Morphological criteria have been used for many years as surrogate endpoints identifying responders and nonresponders. However, the limitations of anatomic imaging are increasingly appreciated and new markers describing the functional response of tumor tissue to therapy are being investigated. PET, especially in combination with CT, allows for in vivo characterization of tumor biology with various radiochemicals. F-18 deoxyglucose as marker of glucose metabolism represents a clinically robust measurement of viable tumor cells, which has been show to be highly reproducible in the quantitative assessment of tumor metabolism. Several studies indicate that loss of metabolic activity is superior to morphological response by CT or MR in comparison with histology or clinical outcome. Furthermore, recent studies suggest that early assessment of metabolic response may predict clinical outcome with high diagnostic accuracy. Measurement of FDG uptake as soon as 14 days after initiation of neoadjuvant chemo-radiotherapy in esophageal cancer identified a subgroup of patients with good long-term outcome. Newer approaches include the assessment of proliferation rate by F-18 FLT or the identification of protein expression as specific therapeutic targets. In neuro-endocrine tumors somatostatin receptors can be visualized with high sensitivity and resolution using PET. In addition, biological targets such as integrins may serve as therapeutic targets inhibiting angiogenesis. Drugs binding to such proteins may be specifically monitored using designed radiochemicals. The combination of individual tumor biology, monitoring of therapy effect, and correlating functional and morphological information may provide all the necessary tools for individualized therapy in cancer.
M. Schwaiger, None.
Abstract ID: 032
The exciting progress we experience today in regard to targeted therapies based on advanced molecular and genetic understanding, as well as other advanced therapies challenge us to rapidly facilitate noninvasive imaging for biologic response assessment. The disconnect or the long development path between small animal molecular imaging and the commercial availability of highly specific MR agents for patient studies is creating the need to advance the capabilities of today's available and accessible MR technologies and contrast agents. New MR contrast agents have been very slow in approval in the US, and there are a number of agents available or soon available in other countries including the EU. Technical advances in MR include higher fields, advanced coils and parallel acquisition schemas. Therefore, a key focus for today's capabilities are improved image acquisition and advanced processing. This is facilitating the use of MR to monitor therapies; even while unspecific, the biologic response can be assessed. One of the biggest challenges is that we lack standardization and vendor independence both in the acquisition approach and the postprocessing. However, MRI is being established as a capable and important response assessor during therapies, not only for the traditional ones such as cytotoxic and radiation, but also for current biologic approaches such as cytostatic or genetic. MR is uniquely able to facilitate both high-resolution morphologic as well as functional/metabolic imaging. The key advances relate to improving sensitivity to probes and quantitative assessment, with immediate superposition of detailed anatomy and all with a noninvasive, nonionizing radiation exposure methodology.
M.V. Knopp, Berlex 1, 2; Bracco 1, 2; Amersham 2; Novartis 1, 2.
Plenary Session III: Emerging Medical Needs: Disease Detection and Outcome Measures
Abstract ID: 033
Tumors arise as result of the accumulation of mutations that lead to the activation of protooncogenes and the inactivation of tumor suppressor genes. By the time tumors are detected, many genetic alterations have occurred obscuring the sequence of events in the tumorigenic process. It is likely that not all the oncogenic mutations equally contribute to the tumor phenotype. It is therefore important to design model systems in which the importance of the various mutations can be assessed. My laboratory has made a significant investment in developing new mouse tumor models, using Cre/Lox-mediated switching of tumor suppressor genes and oncogenes. The methodology enables us to switch multiple oncogenes and/or tumor suppressor genes within cells in vivo at a defined time. This permits the induction of highly specific tumors within a narrow time window, and allows us to correlate specific genetic defects with tumor characteristics. The general picture that perspires is that mouse models show closer resemblance to the human tumors when an increasing number of mutations is shared. We utilize these models also for evaluating therapeutic intervention strategies. Therefore, we employ sensitive in vivo imaging. We have introduced a conditional luciferase transgene into mice which is activated in conjunction with the introduction of the oncogenic mutations by Cre recombinase expression. In this way, it is possible to follow cells that have been mutated through the light the tumors emit. Also, we can follow tumor growth and its response to therapy. We expect that this advanced modeling of tumorigenes in mice will help us to define the critical signaling pathways in a variety of tumors and permit us to test the activity of drugs acting on these specific pathways. We expect that this can accelerate the identification of the most suitable targets and the preclinical testing of promising drugs.
A.J. Berns, Xenogen 2, 4.
Abstract ID: 034
Development of anticancer drugs has long relied on structural imaging modalities to follow treatment response in terms of tumor volume. However, since current drug development focuses on agents aimed at specific molecular targets, which may not directly impact tumor volume as single agents, simple measurements of tumor volume are of limited utility. Development of these drugs is critically dependent on biomarkers to determine if candidate agents interact with the desired target and/or cause downstream effects associated with the mechanism of action. Molecular (and functional) imaging has the potential to provide such biomarkers, often in both clinically relevant animal models and humans. Thus, molecular imaging has great potential to impact the drug development process, especially for internal decision making, such as Go/No-Go decisions, Proof of Concept evaluations, and guidance with dose selection, and, ultimately, for purposes of drug registration.
J.L. Evelhoch, Pfizer 5.
Abstract ID: 035
Optimal clinical evaluation of novel targeted therapies will require reproducible and quantitative methods for measuring treatment-induced changes in presumed biological targets. In current practice, investigators must rely on the use of surrogate tissues or invasive tumor biopsies collected pre- and posttreatment. 17-AAG is an inhibitor of Hsp90 currently in Phase 1 clinical trial at our institution. This agent causes the degradation of HER2 by causing its ubiquitination and targeting to the proteasome. Our group has developed a method for the noninvasive quantification of HER2 protein status. HER2 expression in tumors was quantitated using 68Ga-labeled antibody fragments and positron emission tomography (PET). The rapid elimination of this radiotracer from the blood pool allowed for serial determinations of HER2 expression pre- and posttreatment with 17-AAG. Treatment with 17-AAG caused a reproducible decrease of over 50% in tracer uptake within tumors by PET imaging. Imaging results correlated with direct determinations of tracer uptake by gamma counter and with changes in HER2 protein expression by immunoblot. This technology has immediate clinical applications as a pharmacodynamic marker in the ongoing clinical trials of Hsp90 inhibitors. It also highlights the potential utility of molecular imaging technologies in the translation of novel targeted therapies from the laboratory to patients.
D.B. Solit, None.
Abstract ID: 036
The PTEN tumor suppressor gene is a target of somatic mutation in a number of malignancies including prostate and endometrial cancer, glioblastoma, and melanoma. Furthermore, germ line mutations in the PTEN gene lead to the development of Cowden disease, an inherited hamartoma syndrome associated with an elevated risk of breast and thyroid cancer. PTEN is a lipid phosphatase with specific activity against the D3 position of phosphatidylinositol-3,4,5-trisphosphate and phosphatidylinositol-3,4-bisphosphate both of which are direct products of the phosphoinositide-3 kinase (PI3K). Thus, in this capacity, PTEN antagonizes signaling through PI3K. PI3K, in turn, activates the downstream serine-threonine kinase Akt. While PTEN loss is one mechanism through which this pathway is activated, the PI3K pathway is activated by any number of upstream activate oncogenic growth factor tyrosine kinases including BCR-ABL, HER2/neu, c-kit, and PDGF-R. Robust therapeutic development impinging on this pathway has focused on the upstream activated tyrosine kinases, PI3K and Akt themselves, and the downstream effectors of Akt most notably mTOR. Multiple tyrosine kinase inhibitors including Imatinib and gefitinib alter PI3K activity. Thus, successful strategies for functional imaging of PI3K pathway activation and inactivation will be broadly applicable to a number of current and emerging therapeutics. Our group has recently deployed cell-based imaging of translocation of the FOXO3A transcription factor, a downstream target of Akt, in broad high-content genetic screens for both cDNA, siRNA, and chemical inhibitors of the PI3K pathway. In addition, prostate-specific mouse models, as well as prostate organ transplant models are being used to discover and validate appropriate biomarkers that may be either detected in serum or may be suitable for imaging. Ongoing imaging activities in this field relevant to drug discovery and human application will also be discussed.
W.R. Sellers, Novartis Pharmaceuticals 1, 2; Dharmacon, Inc. 1, 2.
Plenary Session IV: Markers of Cell Function and Trafficking
Abstract ID: 037
Stem cells serve as source of derived cells for cell-based therapies. While efforts are being made to differentiate stem cells in culture, the major test for the function of the cells are in grafts to animals. In assessing the “authenticity” of derived cells, that is, that they function appropriately in grafted sites, it is necessary to identify (all) the transplanted cells through the use of markers that allow for the identification of the transplanted cells and their progeny. Marker systems that allow for real-time analysis of function, real-time analysis of cells, and histologic identification are most desirable, as the number of animals required for study would be reduce. Also, effective marker systems will allow for an assessment of safety, that is, inappropriate differentiation, inappropriate cell migration, and tumor formation. The need for such marker systems will be emphasized through examples of grafted derivatives of human stem cells.
J. Gearhart, None.
Abstract ID: 038
Stem cells and progenitor cells have the potential to treat and possibly cure a variety of disorders, particularly those of the central nervous system and the myocardium. To develop successful clinical therapies, the fate of these cells after grafting must be monitored in a noninvasive manner. MR imaging offers a high spatial-temporal resolution and is therefore ideally suited for tracking magnetically labeled cells. Superparamagnetic iron oxide particles can be used as magnetic markers, but, in order to be taken up by nonphagocytic cells, they have to be modified in order to induce sufficient uptake. This can be achieved by coating the magnetic nanoparticles with either internalizing monoclonal antibodies or transfection agents, including dendrimers and poly-
J.W. Bulte, None.
Abstract ID: 039
T cells are critical to both the immunologically based graft-versus-host disease (GVHD) and graft-versus-tumor (GVT) reactions, which occur following hematopoietic cell transplantation. In an effort to visualize these complex biological processes, we utilized bioluminescent based imaging with the North American firefly luciferase gene, which was introduced into T cell populations. Following successful introduction into diverse populations of T cells, biological activity could be readily assessed following transfer into syngeneic and allogeneic animals. GVHD characterized by the massive proliferation of allogeneic T cells was readily imaged in the gut, liver, and skin. In contrast-specific T cell trafficking to tumor sites could also be readily assessed. The role of regulatory T cells in these complex biological processes revealed that with the timed introduction of specific T cell populations at defined timepoints in the process the harmful effects of GVHD could be minimized, whereas GVT retained. The use of bioluminescent imaging holds great promise in elucidating these complex biological processes to gain unique in vivo insights to direct therapeutic interventions.
R.S. Negrin, None.
Abstract ID: 040
Significant progress in stem cell research over the past decade is now gradually translating into a number of new and promising methods of therapy of various diseases, including cancer. It is quite obvious that continuous monitoring of the efficacy of many emerging stem-cell-based therapeutic approaches cannot be performed using invasive methods (e.g., multiple biopsies), except for blood sampling. This dictates the need to develop techniques for noninvasive monitoring of the fate of stem cells after their administration to patients. The aim of current molecular-genetic and cellular imaging research is to develop novel approaches for repetitive in vivo imaging of the fate of stem cells after their administration into an organism. However, several components must be developed first, including: (a) vectors for transduction of progenitor cells; (b) in vitro models for validation of tracer accumulation in transduced cells; (c) in vivo imaging of reporter gene expression in transduced tumor cells (as a model for validation of the reporter gene imageability in vivo); (d) in vivo imaging of adoptively transferred stem cells; (e) in vivo models of disease to study the therapeutic approaches using progenitor cells; and (f) synthesis and validation of novel radiotracers for reporter gene imaging. The lecture will describe recent advances in noninvasive multimodality imaging of spatial and temporal whole body dynamics of adoptively transferred antigen specific T cells for therapy of cancer and transplanted stem cells for bone marrow reconstitution, for organ regeneration, and for stromally targeted anticancer therapies. Representative PET/CT fusion images of bone marrow-derived stem cells engraftment in at 1 week postintravenous injection in a mouse model are shown in the figure below.
J.G. Gelovani, None.
Symposium VII: Applications of Imaging: Use of Clinically Approved Molecular Imaging Agents
Abstract ID: 041
The folate receptor (FR) is overexpressed in a large fraction of human tumors but only minimally distributed in normal tissues (Garin-Chesa, 1993; Ross, 1994). The FR therefore serves as an excellent tumor marker, as well as a functional tumor-specific receptor. Folic acid, a high-affinity ligand for the FR (Kd ˜ 0.1 nM for the type α receptor), retains its receptor-binding and endocytosis properties when covalently linked to a wide variety of molecules (Leamon, 2001), including conjugates of radionuclides and chemotherapeutic agents (Ladino, 1997). This implies that the nature of the attached “cargo” that can be targeted to tumors with folate is inconsequential. In a recent Phase I/II study (Siegel, 2003), a folate-targeted delivery system was applied towards the targeting of a diagnostic imaging agent (111In-DTPA-Folate) to ovarian cancer. This “proof-of-principle” study demonstrated the ability of folate to target drugs to folate-receptor-positive (FR+) cancer tissue. Studies are now underway in which FolateScan [a folate-targeting ligand (EC20) designed to chelate technetium-99m] is being tested in the imaging of ovarian and renal cell cancers. FolateScan is being developed as a diagnostic agent intended to provide accurate information on folate receptor status and aid in selecting patients that may benefit from folate-targeted therapies. Such therapeutic approaches targeting the folate receptor on cancer tissues are being investigated in parallel with the imaging agent. The advantages of this targeted drug delivery approach to both diagnosing and treating cancer and other diseases, as well as results of recent clinical trials, will be presented.
J.S. Engel, Endocyte 5.
Abstract ID: 042
Reimbursement of positron emission tomography (PET) for most major tumors has fostered the availability of high-resolution PET scanners throughout the US. The clinical benefits of improved staging are being felt in the management of lymphomas, lung cancer, colorectal cancer, esophagus, head and neck cancer, thyroid cancer, breast cancer, and other common tumors. Most of the studies performed to date depend on selective FDG uptake in tumors, the molecular basis for this being, an up-regulation of glycolysis rate, due to increased hexokinase and down regulation of glucose-6-phosphatase in human tumors. The reason for this up-regulation is still unclear, but may relate to adaptive responses to stress, which shield the cancer cell from death. Much biologic information is embedded in simple measures of FDG uptake: For example, SUVmax, a standardized uptake value for FDG in the primary tumor correlates with survival and progression free survival in thyroid cancer, esophageal cancer, lung cancer, and lymphoma. Other imaging probes for the characteristic phenotype of human cancer are also being developed and are likely to help improve clinical management, and biologic characterization of human tumors in vivo. New tracers with significant promise include F-18 misonidazole (hypoxia), F-18 dihydrotestosterone (androgen receptor), and Fluorothymidine. In the future, positron-emitting radiolabeled antibodies, such as Gallium-68 (Fab')2 herceptin for her2 expression imaging, will be used to select appropriate treatment and monitor response in breast cancer. In summary, PET imaging permits characterization of the characteristic phenotype of human cancers, which is likely to lead to improved differential diagnosis, staging, treatment selection, and monitoring treatment response for common human cancers.
S.M. Larson, GEMS 1; Amersham 2; DOE 1; NCI 1; CPS 2.
Abstract ID: 043
Detecting the extent of local or regional lymph node metastases is of clinical importance in virtually any type of primary tumor. Current cross-sectional imaging modalities like CT and MRI have a reported sensitivity of around 50% for determining which lymph nodes harbor metastases. This poor performance stems from reliance on size criteria for differentiating benign from malignant lymph nodes. Surgically invasive techniques for node dissection and pathologic analysis, although considered the gold standard, expose the patient to procedural complications and risks. Ultrasmall superparamagnetic iron oxide (USPIO) particles (Combidex®; Advanced Magnetics, Cambridge, MA; also known as Sinerem®; AMI-7227; AMI-227; BMS 180459; MION) are a class of nanoparticles specifically developed for MR lymphangiography targeting the SRA receptors on the macrophages. Given their macrophage uptake, they are specific for reticulo-endothelial system uptake of which lymph nodes are part of. The uptake of USPIO within nodal macrophages provides information on lymph node morphology and function allowing for improved metastases detection within the lymph nodes. After intravenous injection, the USPIO particles are transported into the interstitial space and subsequently via lymph vessels into the lymph nodes. Once within normally functioning nodes, the iron particles are taken up by the macrophages, reducing the signal intensity of normal lymph node tissue in which they accumulate, because of the T2* and susceptibility effects of iron oxide. In areas of lymph nodes replaced with malignant cells, there is absence of macrophage activity and hence lack of USPIO uptake. Thus, the post-USPIO MRI is able to identify metastatic areas in the lymph nodes independent of the lymph node size. Several studies have demonstrated enhanced sensitivity and specificity for lymph node evaluation after USPIO administration, for pelvic, head and neck, and chest malignancies.
M.G. Harisinghani, None.
Abstract ID: 044
Annexin V has been radiolabeled with 99m-Technetium for human use. Thus far, studies have been performed in several groups of patients: (1) Heart transplant recipients to detect the presence and severity of transplant rejection. The technique proved capable of identifying patients with moderate to severe rejection. The imaging findings were correlated with endocardial biopsy results. (2) Acute myocardial infarction. The technique demonstrated the striking presence of apoptosis in the territory of ischemia in 80% of patients. (3) Cancer patients, with lung cancer, lymphoma, and head and neck tumors. These patients had a remarkable increase in uptake after a single dose of chemotherapy or radiation. In addition to the human studies, laboratory investigations demonstrate striking uptake in animals with immune mediated inflammation. These studies support the use of radionuclide imaging to detect apoptosis in vivo.
H. Strauss, Theseus Imaging 2.
Abstract ID: 045
Homogenous distribution of cardioplegia delivered to the myocardium has been identified as an important predictor of postcardio-pulmonary bypass ventricular recovery and function. At present, a method to directly determine adequate distribution of cardioplegia in patients during cardiac surgery does not exist. The goal of this study was to evaluate the feasibility of visualizing cardioplegia delivery using a novel, noninvasive optical method. Such a system would permit instantaneous imaging of jeopardized myocardium and allow immediate, intraoperative corrective measures. Invisible near-infrared (NIR) light penetrates relatively deeply into myocardial tissue, and background autofluorescence in the NIR region is extremely low. We have developed a portable, intraoperative NIR fluorescence system for use in large animal cardiac surgery, which simultaneously displays color video and NIR fluorescence images of the surgical field. By introducing exogenous, nonisotopic NIR fluorophores, specific cardiac functions can be visualized in real time. In an open-chest, cardiopulmonary bypass porcine model, we demonstrate that the FDA-approved intravascular fluorophore indocyanine green (ICG, CardioGreen(tm)) permits real-time assessment of cardioplegia delivery. ICG was injected into an aortic root and/or transatrial coronary sinus catheter during delivery of cold crystalloid cardioplegic solution. Segmental distribution was immediately noted at the time of injection. In a subset of animals, simulated coronary occlusions resulted in imaging defects consistent with poor cardioplegic delivery and jeopardized myocardium. Videodensitometric analysis was performed on-line to quantify right and left ventricular (RV, LV) distribution. We report the development of a novel, noninvasive, intraoperative technique that can easily and safely delineate the presence or absence of regional cardioplegia (antegrade and/or retrograde) and offers the potential to quantitate the relative segmental distribution during cardiac surgical procedures.
J.V. Frangioni, None.
Abstract ID: 046
Sensitive, noninvasive in vivo imaging strategies are desirable for the early monitoring of treatment response in rheumatoid arthritis (RA). Such technologies would be valuable to initiate early appropriate therapy and dosing, judge clinical outcome, and allow more effective drug development. Furthermore, molecular reporters could serve as indicators for stage and severity of the disease, prognosis, and enable early diagnosis. Matrix degrading enzymes like certain cathepsins and matrix metalloproteinases are highly up-regulated in RA, leading to the destruction of inflamed joints. Therefore, these enzymes are suitable targets for in vivo imaging of early treatment response. Using a cathepsin B activateable, near-infrared fluorescence (NIRF) imaging probe, we examined whether methotrexate (MTX) treatment would alter cathepsin B activity. Our results show that 24 hr after intravenous injection of the reporter, inflamed paws of arthritic mice (n = 10 mice) showed sevenfold higher fluorescence intensity than paws of healthy mice (n = 6 mice). In MTX treated animals (35mg MTX/kg 48hrprior injection of the probe, n = 10 mice), a significantly (p <.001) lower fluorescence signal (inflamed paws: 50%, inflamed toes: 70%) was observed. Histologically, fluorescence of the cathepsin B probe and cathepsin B antibody staining are localized inboth major cell types in the hyperplastic synovia, synovial fibroblasts, and synovial macrophages. In conclusion, cathepsin B activateable NIRF imaging probes are suitable reporters for imaging of the early treatment response of antirheumatic drugs such as MTX. Moreover, the data show for the first time that MTX treatment resultsin a decrease of cathepsin B activity in the inflamed synovial tissue.
A. Wunder, None.
Symposium VIII: Low Light Imaging Instrumentation and Reporters
Abstract ID: 047
Fluorescence molecular tomography is a recently developed imaging modality that allows three-dimensional localization and quantification of gene expression products in vivo and in deep tissues. In this talk, current progress on a set of technologies that allow tomography of proteases, fluorescent proteins, and response to treatment based on fluorescent signatures are presented. The development of noncontact optical tomography and the optimization of tomographic schemes to yield submillimeter resolution in small animal imaging are also discussed.
V. Ntziachristos, None.
Abstract ID: 048
The opportunities for molecularly based fluorescence-enhanced optical imaging depend upon the ability to measure reemitted signals at the tissue interface and to tomographically reconstruct a three-dimensional interior map of the targeted or reporting fluorophore concentration or decay kinetics. At the Photon Migration Laboratories, our work has focused upon developing time-dependent measurements of fluorescent light generation and propagation within clinically relevant tissue volumes. Using frequency-domain approaches, we measure the phase-delay and amplitude attenuation of reemitted fluorescent light in response to excitation light modulated nominally at 100 MHz by employing a gain-modulated image intensified CCD camera as an area detector in order to eliminate the need for background or ambient light subtraction, as well as to provide enhanced time-resolved (subnano-second) information for tomographic reconstructions from expectedly diminutive fluorescent signals [1]. Using constrained nonlinear optimization techniques [2] and Bayesian approaches [3], we have reconstructed targets in clinically relevant phantom volumes. Illustrated in the figure is a typical result showing three-dimensional reconstruction of three distinct targets contrasted with micromolar concentrations of indocyanine green in an experimental phantom model of the human breast from point illumination of 785 nm light. Additional tomographic results from experimental data using area illumination and detection approaches are also shown indicating the potential of a handheld device for volumetric imaging from surface measurements. The ability to sense femtomolar quantities of dye in large phantoms [4] indicates the relevancy of NIR techniques for molecularly based diagnostic imaging of targeting and reporting fluorescent contrast agents.
Anterior and lateral view of the actual breast phantom (a,d), reconstructed breast phantom using coarse mesh (b,e) and reconstructed breast phantom using fine mesh (c,f). Three targets of 0.5 — 0.6 ml volume were used in the actual breast phantom of — 1087 ml volume.
E.M. Sevick-Muraca, None.
Abstract ID: 049
We previously developed optical imaging systems for noninvasive tomographic human brain structural and functional imaging, and codeveloped the original lucierfease imaging systems and approach in animals. We now set out to develop a flexible system for imaging the distribution and localization of targeted fluorescent agents in humans in room light surgical operating room conditions. A filtered, gated, intensified CCD system with a low-f custom lens was constructed and characterized. A tunable solid-state laser source allows operation with different contrast agents. The system was mounted on a standard OR mounting system, with the display suspended from the same ceiling mount (Heraeus, Germany; Figure A—Palomar™ Suspended System). We imaged the binding, kinetics, biodistribution, and excretion of multiple targeted agents, including our own SPX-Pr-045 directed at cells of prostate luminal origin, SPX-Lm-035, directed at the lymphatics, as well agents from other groups, for example, 2DG (Penn), Cytate (Wash U). We found that the system readily imaged tumors and other binding sites in small animals, in real time and in room light. Images were displayed with 1-sec update speed, including all processing/colorization of tumor fluorescence overlaid on a black and white background image (Figure B—Real-Time Palomar™ Screen Image). The system is now available for general research use and will be installed at 2 medical centers in 2004. An image using FDG-PET-analogue Cy-2DG is shown (Figure C—Cy7-2DG Tumor Imaging). We conclude that this system will perform in the OR environment and may have application in bone scans and postsurgical follow-up.
Real-Time Palomar™ Imager. (A) O.R. Ceiling Mount. (B) Real-Time Screen View. (C) Tumor imaged with Cy2DG.
D.A. Benaron, Spectros Corporation 1, 4, 5.
Abstract ID: 050
Optical coherence tomography (OCT) is an emerging tool for real-time in situ tissue imaging with micrometer-scale resolution and millimeter-scale imaging depth in living tissues. Real-time OCT systems have been integrated into clinical medical diagnostics and functional extensions such as polarization-sensitive, Doppler, and spectroscopic OCT have been demonstrated. However, OCT isinherently insensitive to incoherent scattering processes such fluorescence or spontaneous Raman scattering, which are critical to current forms of molecular imaging. We introduce a novel technique for contrast enhancement in OCT, which enables molecular specific imaging for the first time. A pump-probe technique is employed in which a pulsed pump laser is tuned to ground-state absorption in a molecule of interest. The location of the target molecule population is derived from the resulting transient absorption of OCT sample arm light acting as probe light. Preliminary results exhibiting contrast enhancement in cross-sectional OCT images using transient absorption in methylene blue are presented. An alternative approach that takes advantage of long lifetime dark state transitions in bacteriorhodopsin is also described. These novel techniques for interferometric detection of differential induced absorption have been demonstrated in tissue-simulating phantoms and appear to be sufficiently sensitive for molecular contrast imaging in biological tissues. The spectral selectivity of the technique offers flexibility to choose different contrast agents dependingonthe excitation source and transient spectra. The technique could readily be extended to exploit other processes such as two photon induced excited state absorption, single state saturable absorption, stimulated emission, and others. The spatial and spectral selectivity of this nonlinear optical approach has the potential for combining the advantages of OCT and molecular imaging. (Supported by NIH R24 EB000243.)
J.A. Izatt, None.
Abstract ID: 051
Newly developed instrumentation and optical probes allow measurements of dynamic processes within complicated biological systems, including live animals. For example, we have used two-photon imaging to quantify NAD(P)H from deep within living tissue, but use of intrinsic probes limits the information that can be obtained about physiological function. To broaden these optical approaches, new probes must be developed. We have developed in vivo detection tools for tumor-associated matrix metalloproteinase-7 (MMP-7 or matrilysin) activity using a novel polymer-based fluorogenic substrate, PB-M7VIS, which serves as a selective “Proteolytic Beacon” (PB). PB-M7VIS is built on a polyamido amino dendrimer core of 14.2 kDa, covalently coupled to a fluorescein (***Fl)-labeled peptide, ***Fl-M7, and to tetramethylrhodamine (TMR). PB-M7VIS is efficiently and selectively cleaved by MMP-7 with a kcat/Km of 2.2 × 105 as measured by the rate of increase in Fl fluorescence (up to 17-fold for complete cleavage of an optimized PB-M7VIS) with minimal change in the TMR fluorescence. With MMP-2 or MMP-3, the kcat/Km is ˜60- or ˜14-fold lower, respectively, than with MMP-7. Thus, in PB-M7VIS, Fl-M7 is a selective optical sensor of MMP-7 activity with TMR serving to quantify the uncleaved and cleaved reagent, each of which can be visualized as subcutaneous fluorescent phantoms in a mouse and optically discriminated based on the ratio of green/red (Fl/TMR) fluorescence. In vivo studies with phantoms prepared using a near-infrared (NIR) analogue of PB-M7VIS show enhanced sensitivity with greatly reduced background in the NIR. The in vivo specificity of PB-M7VIS was tested in a mouse xenograft model. Intravenous administration of PB-M7VIS gave significantly enhanced Fl fluorescence (520 ± 140 counts/10 sec/pixel) from MMP-7 positive tumors but not from control tumors, both originally derived from SW480 human colon cancer cells. Thus, PB-M7VIS functions as a “Proteolytic Beacon” for in vivo detection and imaging of MMP-7 activity.
D.W. Piston, None.
Abstract ID: 052
GFP-labeled in vivo tumor models and GFP-based transgenic systems have proven to be powerful research tools, revealing details of tumor growth, angiogenesis, and metastatic spread, for example. However, achieving useful images can be challenging if the label is buried in deep structures (because of scattering and absorption processes) or if bright autofluorescence is present. Spectral imaging generates a series of images at multiple wavelengths, associating an optical spectrum with every pixel. We investigated whether this approach, using liquid crystal tunable filters and mathematical unmixing algorithms, could usefully be applied to in vivo imaging. Visible-range LCTFs typically have a 10–30-nm-wide window that can be electronically tuned, with 1-nm precision, to lie anywhere in the range 420–720 nm. Such devices have been shown to be useful for multicolor FISH and for resolving multiple fluorescent reagents with overlapping emission spectra. In addition, they have permitted the elimination of interfering autofluorescence in microscope-based applications. Using LCTF-based imaging, we examined mouse and zebrafish model systems and found that spectral imaging could enhance the detection and quantitation of GFP-labeled tissues. Standard GFP imaging equipment was modified by replacing the usual color camera with a cooled monochrome camera and an LCTF positioned in front of a conventional macro lens. A series of images was taken every 10 nm from 500 to 650 nm. Using predefined or experiment-specific GFP and autofluorescence spectra, and appropriate algorithms, the original image could be unmixed to yield a GFP signal uncontaminated by autofluorescence. In one model system, this approach clearly revealed GFP signals emanating through skin from tumors located in the mouse lung. Similar results were also achieved imaging zebrafish embryos and adult fish expressing transgenic GFP fusion proteins in various locations. Thus, we have demonstrated the general utility of spectral imaging for in vivo fluorescence-based studies.
R. Levenson, CRI, Inc. 5.
Symposium IX: Targeting Pathways in Signaling Networks
Abstract ID: 053
Postnatal mammalian cardiac myocytes respond to mechanical stress, growth factor and hormonal action, and metabolic and sarcomeric abnormalities by enlarging, but these cells are unable to proliferate. Cardiac hypertrophy is a potent risk factor in humans for the development of congestive heart failure, but it is unclear whether hypertrophy is an adaptive or maladaptive response. My group has focused on the hypothesis that the relative activation of various intracellular signaling pathways determines whether compensated cardiac hypertrophy or overt cardiac dysfunction develops in response to provocative stimulation, such as pressure overload. We have employed a series of mouse genetic model systems to test this model, including Grb2 haploinsufficient mice, and transgenic mice with cardiac-specific expression of dominant negative p38 MAPK, 14-3-3, and Raf-1. Based on the phenotypic analysis of these and other mice, we conclude that compensated hypertrophy occurs as a result of isolated ERK MAPK or Akt activation, whereas “pathological” hypertrophy occurs in response to the simultaneous activation of multiple pathways, including p38 MAPK and JNK.
A.J. Muslin, None.
Abstract ID: 054
Green fluorescent protein (GFP) has now become an important fluorescent tag for imaging intracellular proteins that are otherwise technically difficult or impossible to observe in living cells with light microscope. This technical achievement offers enormous potential for biological and medical research in living cells. In the present symposium, first, we focused on the nuclear-cytoplasmic trafficking of corticosteroid receptors. There are two types of corticosteroid receptors, mineralocorticoid receptor (MR) and glucocorticoid receptor (GR), both of which are ligand-dependent transcription factors. Given the differential action of MR and GR in the central nervous system, it is important to elucidate how the trafficking of these receptors between cytoplasm and nucleus is regulated by various kinds of intra- and extracellular environments. We have shown dynamic changes in subcellular localization of GFP-MR and GFP-GR in response to ligand, cytoskeletal disruption, inhibition of hsp90. To examine the simultaneous trafficking of MR and GR within single living cells, different spectral variants of GFP, yellow fluorescent protein (YFP), and cyan fluorescent protein (CFP) were linked to MR and GR, respectively. We found that MR and GR exhibit differential nuclear translocation patterns depending on ligand concentrations or cell types. By using this double-label GFP technique, we also examined the possible involvement of transport factors, such as importin alpha and importin beta, for nuclear import of receptors. Furthermore, we demonstrated possible heterodimerization of MR and GR by using GFP-based fluorescence resonance energy transfer (FRET) using CFP and YFP in living cells. A successful FRET system yields information about the location and exact timing of the interaction that is not available from conventional detection systems such as coimmunoprecipitation and yeast two-hybrid assay. Our findings may provide new insights into the dynamic status of stress-related corticosteroid receptors in living neural cells.
M. Nishi, None.
Abstract ID: 056
Steroids as small lipophilic molecules can cross cell membrane and bind to the intracellular receptors with high affinity. They can distribute to almost every tissue including brain as they can cross blood-brain barrier. Ligand-binding specificity and mechanisms of action of the receptors have been extensively studied and fairly well understood. Estrogen and its antagonists have been radiolabeled and used to image estrogen receptor (ER) expressing tumors for in vivo diagnosis, which demonstrates the efficacy and safety of the radiolabeled ligands in human. The tissues that highly express ER are limited. Taking advantage of these properties, we designed a new in vivo reporter gene imaging system using ER. We used the ligand-binding domain (LBD) to avoid activation of the downstream genes. For further refinement, we also attempted to use a mutated ER with affinity to 4-hydroxytamoxifen but not estrogen to avoid the interference from endogenous estrogen. To show the applicability of the system for monitoring cell survival after ES cell transplantation, we established ES cell lines expressing ER-LBD either stably or inducibly. The cell lines expressing the LBD showed higher uptake of 4-hydroxytamoxifen than parent ES cells and the teratoma formed after subcutaneous transplantation of the cells kept the ability to express the LBD. For the application in gene therapy monitoring, we constructed plasmids in which ER-LBD coding gene was connected to therapeutic gene through IRES. Transiently transfected cells with the plasmid expressed the thrapeutic gene and the LBD simultaneously at a parallel level. Based on the advantage/problem of the ER system, future direction of in vivo reporter gene imaging will be discussed.
T. Furukawa, None.
Abstract ID: 057
Understanding of the tumor microenvironment is important to attempt to design new strategies for development of new chemotherapies for the treatment of cancer. Tumor hypoxia is one condition that may influence many factors, including efficacy of drug treatment, the potential for metastasis, and induction of an aggressive tumor phenotype. Currently, immunohistological staining of hypoxic regions of tumors is possible with a 2-nitroimidazole compound such as EF5. Noninvasive in vivo imaging of hypoxia is possible using a reporter gene driven by a hypoxia-inducible promoter. A series of promoters have been created using hypoxia-responsive elements (HREs) from genes up-regulated by hypoxia-inducible factor, HIF. Mammalian expression vectors were designed using multiple repeats of the HREs from carbonic anhydrase IX (CA9), phosphoglycerate kinase (PGK), or vascular endothelial growth factor (VEGF), a minimal CMV promoter and a firefly luciferase reporter. The induction of luciferase activity using these promoters can be optimized by selection of the cell line in which they are expressed. Quantitative RT-PCR was used to determine the up-regulation of CA9, PGK, and VEGF in several cell lines exposed to hypoxic conditions. Cell lines were transfected with the luciferase-HRE vectors, CA9, PGK, PGK/CA9, and VEGF. Cells were exposed to normoxia or hypoxia, lysed, and assayed for light production. Cell lines with the greatest increase of mRNA for a specific gene also had the greatest induction of light using the corresponding promoter after exposure to hypoxia. Using this information, cell line and vector combinations can be chosen for optimal induction and increased sensitivity of a reporter for in vivo imaging of tumor hypoxia.
B. Baggett, None.
Abstract ID: 058
Tissue hypoxia is one of the most important factors in tumor neo-angiogenesis. A transcriptional activator has been identified as the major mediator of cellular hypoxic responses, named hypoxia-inducible factor 1 (HIF-1). Under hypoxic conditions, HIF-1 binds to the hypoxia-responsive element (HRE) in the regulatory region of its target genes and initiates transcription. Recently, we proposed an approach for monitoring hypoxia-induced up-regulation of different endogenous genes in situ and in vivo by noninvasive imaging using a hypoxia-sensitive reporter system. Based on this system, we have developed a vector with a new reporter gene (TKnesGFPLuc) placed under control of HRE, and we have shown that the vector can be effectively used in identification of the inhibitors for HIF-1α-induced transcriptional activity. To demonstrate the response of the new fusion reporter to hypoxic conditions, we transduced C6 and RG2 cell lines with the vector. Hypoxia was induced by exposing cells to various concentrations of CoCl2 (1–500 μM) or 2% O2 during 24 hr. Expression of the TKnesGFPLuc reporter gene was equally activated either by oxygen deficiency or addition of CoCl2 (50–500 μM), and activation of the reporter gene was easily assessed by fluorescence microscopy, FACS, and bioluminescent analysis. To demonstrate the efficacy of this reporter system for the detection of HIF-1 signaling inhibitors, we used a recently developed compound 3-(5′-hydroxymethyl-2′-furyl)-1-benzylindazole (YC-1) that inhibits HIF-1 activity in vitro by blocking HIF-1α expression at the posttranscriptional level. When tumor cells carrying the HRE/TGL reporter system were treated with YC-1 during hypoxia the level of HIF-1-mediated transcriptional activation of the reporter system was significantly inhibited as compared to nontreated hypoxic cells. These results demonstrate that the current HRE/TGL reporter system can be used to study anti-HIF-1 drugs in vitro, and that PET and optical imaging could be used for in vivo assessment of novel drugs.
I.S. Serganova, None.
Symposium X: Novel Probes and Activation Strategies
Abstract ID: 059
The specific delivery of metal ions to physiologic sites such as tumors is a central problem in medical imaging. Practical approaches to the targeted delivery of directly labeled or tagged reagents include the development of labeled small molecules such as the technetium-99m complexes; the use of receptor-binding small molecules such as peptides, or target-binding macromolecules such as antibodies and their fragments; and the advent of pretargeting, which combines an often slow target localization process with a fast label capture step. The advantages of pretargeting are that it combines the highly developed target selection of antibodies with the rapid localization and clearance of small molecules in a wide-ranging format. The first successful reagents for the label capture step involved the unusually stable complex between the protein avidin (or streptavidin) and the small organic molecule biotin, which not only is specific but also can be practically irreversible over the course of several hours in living systems. The immunogenicity of this system and the endogenous presence of biotin limit its applications. Because antibodies can be humanized to reduce immunogenicity, alternative systems involving capture antibodies that recognize metal chelates or small synthetic haptens do not have these limitations, but lack the prolonged avidin-biotin binding. We address this by preparing antibodies with infinite affinity, which capture labels irreversibly. The methods involve engineering complementary reactive groups, one in the binding site of an antibody and another on the synthetic ligand the antibody recognizes. Recently, we have been able to prepare capture proteins that recognize and bind irreversibly to a set of small molecules known to be useful in vivo, which carry indium, yttrium, or any of the lanthanide elements. Because of the diverse physical properties of these elements, the probes may find application in a variety of imaging techniques from PET to MR to IR.
C.F. Meares, Lexrite Labs 2, 4.
Abstract ID: 060
We have demonstrated spatial and temporal control of therapeutic gene expression by linking the radio-inducible/chemo-inducible DNA sequences from the Egr-1 promoter (CArG sequences) to a cDNA encoding human tumor necrosis factor-alpha (TNF-α). This chimeric genetic construct was cloned into a replication deficient adenovirus forming the Ad.Egr-TNF vector. Preclinical studies demonstrated synergistic antitumor effects following combined treatment with Ad.Egr-TNF and ionizing radiation in various tumor model systems. We also demonstrated enhanced antitumor responses following combined treatment with Ad.Egr-TNF and chemotherapy. Taken together, these findings demonstrate that the combination of Ad.Egr-TNF and cytotoxic agents overcomes both radio-resistance and chemo-resistance. The mechanism responsible for the observed antitumor effects is likely destruction of the tumor microvasculature by high intratumoral levels of TNF-α, which are induced following exposure to ionizing radiation or chemotherapy. Importantly, the enhancement in the therapeutic ratio is achieved without an increase in local or systemic toxicity. To probe the physiologic effects of this gene therapy, novel images have been obtained from human prostate tumors (PC-3) implanted in the legs of athymic nude mice treated with this gene therapy. Preliminary spectroscopic electron paramagnetic resonance images demonstrate an initial increase in hypoxia 3 days following radiation and gene therapy. Tumor reoxygenation occurs 2 weeks later.
R.R. Weichselbaum, GenVec 1, 2, 4; MediGene 1, 2.
Abstract ID: 061
In our search on radioactive Cu-labeled radiopharmaceuticals, we found that some Cu complexes have an affinity to redox enzymes in electron transport chain (ETC) and show reductive retention in the living cells in redox-potential-dependent manner. From this findings, we surveyed various Cu-complexes to selectively visualize tissues of normal redox state, as well as abnormal redox state using Cu-radiopharmaceuticals in vivo. Hypoxia can be characterized as an abnormally reduced state, and it can be visualized using Cu-complexes with low redox potential. Among the surveyed Cu-complexes, we selected Cu-diacetyl-bis(N4-methylthiosemicarbazone) (Cu-ATSM) as a hypoxia imaging agent. Using Cu-ATSM, we could visulalize acute ischemic heart, hypoxic tumors and so on. Metabolic studies of Cu-ATSM clarified that dominant redox systems in nontumor cells and tumor cells were completely different; the former being mitochondrial ETC, whereas the latter being microsomal ETC. The role of microsomal ETC in tumor cells is still unclear, but it might be one of the keys for tolerance to hypoxia. Redox imaging agent will bring us new insight of energy metabolism in tumor cells and might become a key of tumor treatment.
Y. Fujibayashi, None.
Abstract ID: 062
Future medical technology may be envisioned to operate intelligently, linking sensed diagnostic information to therapeutic actuation via on-board controls and information processing capabilities. While a complete system of this nature may be a longer-term proposition, it may be maintained with confidence that the achievement of each step in this three-link chain will trigger momentous advances in medicine. One key enabling technology towards the advent of multifunctional, self-regulating devices is the broad field of nanotechnology. Drawing from past history and current development with liposomes and adenoviral delivery nanovectors, the vision that may be formed encompasses antigen-targeting, biological barrier-avoiding nanoparticulates that act as contrast agents with molecular marking capabilities, as well as the ability to deliver therapeutic action. More recent, and very promising technological entries in this grand challenge include dendritic polymers, metal nanoshells, C-60 constructs, gold colloids, quantum dots, silicon, iron oxide or solid lipid nanoparticles, and perfluorocarbonate emulsions, with still many other nanovectors appearing constantly on the horizon. Fully confident of the revolutionary beneficial potential of the single-platform diagnostics/therapeutics nanovector approach, the National Cancer Institute is developing a plan that will aid in its coming to full clinical fruition. In this talk, a brief review of established and newly proposed nanovector/molecular targeting technologies will be offered, the merits of its integration with therapeutic functionalities will be discussed, and a chart of key application areas and technological challenges will be presented.
M. Ferrari, iMEDD Inc 2, 4.
Abstract ID: 063
T.A. Beresten, None.
Abstract ID: 064
Pretargeting—specifically placing artificial receptors on target cells in vivo and capturing small probe molecules at these sites—can lead to highly specific images with little background. Pioneering work in this area has involved the strong avidin-biotin binding pair and weaker antibody-hapten binding pairs; our objective is to develop a humanized system with properties superior to both. We have solved the crystal structure of the rare earth-DOTA binding antibody 2D12.5 and have used it to identify preferred residues on the protein for engineering mutants capable of forming a permanent bond to a captured rare earth-DOTA ligand. We have designed and expressed a collection of chimeric antibody fragments incorporating site mutations and tested them for reactivity with complementary rare earth-DOTA molecules. The relative reactivity of each mutant depends on accessibility to the ligand molecule in the binding site. Details of the molecular engineering and irreversible reactivity of each antibody mutant will be described. The class of metallic elements that can bind to this antibody is rich in useful physical properties. For example, the nuclear emissions of yttrium, lutetium, and several of the other rare earths are suitable for radioimaging and/or radiotherapy. The luminescence properties of terbium and europium feature long excited-state lifetimes and sharp emission lines, and the paramagnetism of gadolinium and other lanthanides has led to contrast agents for magnetic resonance imaging. With its ability to bind all of the rare earths with high affinity, this system represents a molecular docking station with broad applications for imaging and therapy.
T.M. Corneillie, None.
Symposium XI: Multimodality Imaging: Combined Instruments and Multifunctional Probes
Abstract ID: 065
The information provided by different in vivo imaging modalities is in many situations complementary. X-ray computed tomography (CT) is a well-established imaging modality that provides detailed anatomical information especially for the skeletal system, while positron emission tomography (PET) provides information for both the temporal and spatial distribution of different radiolabeled molecular probes. By combined imaging of the same subject with these two modalities, the complementary information content brings a new dimension to the imaging results and elevates them to a higher level. We will present here details on the implementation, use, and pitfalls as well as some applications for combined X-ray micro-CT and micro-PET imaging, with two commercially available separate tomographic instruments. Specifically, we will discuss issues related to animal handling, positioning, and control as it refers to both routine screening scans and longitudinal studies. Furthermore, we will present data-independent solutions for image registration and image visualization with data fusion (Fig.) as well as usage of the X-ray CT image data for corrections of the PET images for photon scatter and attenuation. Additionally, we will also discuss the levels of necessary animal radiation exposure as well as methods to minimize it if necessary in repeated longitudinal studies. Finally, we will present examples of interesting applications for in vivo imaging of small animal models.
A.F. Chatziioannou, Concorde Microsystems 6.
Abstract ID: 066
Over the past 5 years, combined SPECT/CT imaging systems have emerged from the research laboratory and are now available for clinical use in humans and for experimental studies of small animals. These systems allow gamma-emitting radiopharmaceuticals to be imaged with SPECT in a way that can be correlated with anatomical information obtained with CT. While radionuclide and X-ray images can be fused in software using data from separate SPECT and CT systems, an integrated SPECT/CT system acquires both datasets during a single procedure to simplify and improve co-registration of the structural and functional images. The resulting data help the user to differentiate sites of normal radionuclide uptake from those that indicate disease. Current applications of SPECT/CT include in vivo imaging of radiolabeled antibodies or other receptor-binding agents for assessing malignant disease. In a clinical setting, this can improve diagnosis or prognosis, and can facilitate planning for radiation therapy or surgery. X-ray CT also can be used to generate a patient-specific map of attenuation coefficients and correction factors for scatter radiation and for partial volume errors. Therefore, SPECT/CT can improve the image quality and the quantitative accuracy of radionuclide data, for example, in myocardial perfusion studies or other functional assessments performed with SPECT. SPECT/CT also has the potential to improve the quantitative accuracy of internal radiation dosimetry of radiopharmaceuticals used therapeutically. Finally, while dual-modality imaging initially has been developed for clinical applications, it is now recognized that these systems have important applications for imaging small animals involved in experimental studies including basic investigations of mammalian biology and development of new pharmaceuticals for diagnosis or treatment of disease.
B. Hasegawa, GE Medical Systems 1, 6; Photon Imaging, Inc. 1; Radiation Monitoring Devices, Inc. 1.
Abstract ID: 067
In vivo molecular imaging is increasingly applied for noninvasive studies of biological processes, providing unique ability to follow and correlate dynamic processes. The imaging methods applied today cover a wide range of spatial and temporal resolution and offer multitude contrast mechanisms for specifically probing a particular process. However, each particular imaging modality is limited in its capacity to provide a full comprehensive description of the process. Multimodality contrast material can be used for cross validation, and for exploiting the particular advantages of each imaging method. Thus, dual MRI/optical detectable contrast agents, allow for reduced invasiveness, higher penetration, and improved in vivo deep tissue spatial resolution of MRI along with the superior spatial and spectral resolution of optical methods for thin samples. We applied albumin, triple labeled with biotin, GdDTPA, and either fluorescein or rhodamine, for evaluating the regulation of pathological and physiological angiogenesis. Tumors showed vascular permeability, interstitial convection, and lymphatic drain that were dependent on the expression of vascular endothelial growth factor (VEGF). Separation between vascular leak and lymphatic drain was achieved by elimination of the intravascular contrast material using avidin chase with fluorescently labeled avidin. In another study, biotinylated albumin-GdDTPA was applied for correlating vascular permeability with maturation (as manifested by BOLD contrast MRI changes in response to hypercapnia), in ovarian carcinoma tumor xenografts. Wide heterogeneity in maturation and permeability and dissociation between these two markers of neovasculature reflect independent regulation during the early stages of tumor development. A similar approach was applied for following angiogenesis and hyperpermeability during recovery from ischemic limb injury, in the normal ovary, in preservation of ovarian implants by xenotransplantation and in placental angiogenesis in the mouse. These studies demonstrate the strength of combining MRI and optical methods and the use of multimodality probes for molecular imaging as a tool in biological research.
M. Neeman, Varian Biosynergy 1.
Abstract ID: 068
I will describe an atlas-based approach to multimodality in vivo imaging studies in mice. We have developed a preliminary atlas consisting of a segmented volume in which major organs and structures are labeled. The atlas is constructed from co-registered X-ray CT and cryosection images of a nude mouse [1]. To align or register the structures in this atlas to those in an anatomical (MR or CT) scan of an individual mouse, we need to be able to accommodate differences in the size and shape of the anatomical structures defined in the atlas. We achieve this using a nonrigid deformation of the atlas co-ordinate system to that of the individual mouse. The inverse of this deformation will map the individual mouse back into atlas co-ordinates. Nonrigid registration methods based on linear elastic and viscous fluid models have been widely studied in computational neuroanatomy [2] and are applicable here. Co-ordinate deformations are computed to maximize a similarity measure between the deformed atlas and the individual image while constraining the deformation fields through the differential equations governing linear elastic media or viscous fluids. Potential applications range from automated segmentation and region of interest definition in high-throughput imaging studies, to statistical characterization of anatomical or functional differences between populations based on warps to a common atlas co-ordinate system. I will show examples of co-registration of the atlas to CT and PET imagery and demonstrate the use of the atlas for defining optical scatter and absorption properties in the mouse for forward calculations in fluorescence tomography.
Abstract ID: 069
. Ponomarev, None.
Abstract ID: 070
Bioluminescence (BLI) and PET imaging are two modalities currently available for noninvasive imaging in vivo, but each has strengths and limitations. To exploit the strengths of each and facilitate dual-modality imaging, a novel reporter construct was developed in which expression of firefly luciferase (F-luc) fused to a mutant herpes simplex virus thymidine kinase previously enhanced for imaging (mNLS-SR39TK) was driven by the ubiquitin promoter. A 12 amino acid linker between the two proteins was engineered to minimize steric constraints on activity; the construct also featured a modular design with restriction sites at the end of the linker to provide flexibility for future enhancements to the construct. Dual functionality of the fused enzymes was validated in cell culture. Transiently transfected HeLa cells showed full recapitulation of both F-luc and mNLS-SR39TK activity as measured by a cooled CCD camera (IVIS) and 3H-penciclovir uptake, respectively. HeLa cells transiently transfected with the fusion reporter showed F-luc activity at 133% of cells transfected with F-luc alone and more than 16-fold above control background. Similar results were obtained for mNLS-SR39TK activity. 3H-Penciclovir uptake in fusion construct-transfected cells was 112% of mNLS-SR39TK alone (ubiquitin promoter) and more than 100-fold above control background. Transfected cells in Matrigel were then implanted subcutaneously into mice to create pseudo-tumor xenografts. BLI showed equivalent high F-luc signals, while 3H-penciclovir biodistribution analysis showed the fusion construct at 330% of control compared to 312% of control for cells transfected with mNLS-SR39TK alone (mean of 2 animals). This fusion construct will be especially useful for concurrently obtaining both bioluminescence and quantitative PET data from living animals.
A.H. Kesarwala, None.
Symposium XII: Cell Migration: Stem Cells and Immunology
Abstract ID: 071
New preparations, fluorescent probes, and imaging techniques are providing the means to observe the behavior of cells in the tissue environment of lymphoid organs. In particular, when combined with two-photon laser microscopy, intravital imaging of surgically exposed lymph nodes provides a unique view of lymphocyte migration and antigen presentation as it occurs within the living animal. Real-time imaging of lymphoid cells within native tissue has begun to provide vital insight into how cells of the immune system physically communicate. The immune response is initiated by physical contact between antigen-bearing dendritic cells (DC) and antigen-specific naïve T cells. Using two-photon microscopy, we have resolved the real-time behavior of in situ-labeled DC interacting with naïve CD4+ T cells in lymph nodes. Endogenous DC were visualized in the lymph node following a novel subcutaneous labeling procedure. Dendritic cells migrated slowly while vigorously probing the T-cell area with dendritic extensions, an adaptation that increases the effective capture volume. Highly motile T cells encountered DC randomly, indicating that T cells are not attracted along chemotactic gradients. Early antigen recognition events occurred preferentially on DC dendrites, were transient (<30 min) frequently involving serial contacts with multiple DC, and preceded a robust immune response. We conclude that the immune response is triggered by random—rather than directed—encounters, and describe a stochastic mechanism of antigen recognition that is optimized by T-cell motility and by dynamic projection of DC dendrites.
M.D. Cahalan, None.
Abstract ID: 072
Hematopoietic or bone marrow stem cell transplant is used for a variety of conditions including hematologic disorders, metabolic storage diseases, immune deficiencies, and rescue following treatment for cancer. Although great advances have been made, the problems of inefficient stem cell engraftment and graft-versus-host disease (GVHD) remain major complications post-transplant. In animal models of bone marrow transplantation (BMT), analyses of tissue-specific events usually require dissociation of cells from the affected organs or sectioning of harvested tissues, resulting in loss of spatial and geographic information as well as limiting the areas being analyzed. We have developed a mouse system in which to track the migration and homing of cells following systemic infusion using eGFP transgenic cells and a simple application of a Leica MZFLIII fluorescence stereomicroscope outfitted with a Magnafire color CCD camera (Optronics). Whole body images of anesthetized mice taken at various time points after cell infusion revealed early migration of allogeneic cells to peripheral lymphoid organs with later infiltration of GVHD target organs. Localization of GFP Tg cells could be seen through the skin of shaved mice and internal organs were easily discernable. Following allogeneic or syngeneic GFP Tg cell infusion, representative mice were dissected to better-visualize deeper internal organs and tissues. Infusion of different cell subsets revealed different homing patterns and this method also provided a simple way to identify the critical time points for expansion of the transplanted cells in various organs. This simple application of the fluorescence stereomicroscope will be valuable for studies in which visualization of cellular migration, expansion, and cell-cell interactions will be more informative when analyzed by such an intravital method.
A. Panoskaltsis-Mortari, None.
Abstract ID: 073
Immune responses involve multiple cells acting contemporaneously and sequentially with one another and infectious agents. An important aspect of adaptive immunity involves initiation of antigen (Ag)-specific T-cell responses following recognition of peptide/MHC ligands presented by dendritic cells (DC) within secondary lymphoid tissues. To better understand these key events, we have used multiphoton microscopy to collect 4-D (XYZ and time) data on the interactions of multiple distinct fluorescent Ag-specific T cells of distinct lineage with each other and with fluorescent Ag-bearing DC in an intact lymph node. Using an ex vivo method, high-resolution imaging of these labeled cells in explanted lymph nodes is possible for several hours. The majority of T cells show prolonged adherence to one Ag-bearing DC, eventually resulting in activation. This prolonged contact is paralleled by the formation of an immunological synapse, visualized as the exclusion of CD43-GFP from the region of membrane contact. Detachment from the antigen-bearing DC follows, sometimes in conjunction with cell division. The limited movement of naïve cells observed in early confocal imaging studies seems most related to a region of low lymphocyte motility in the subcapsular area less than 100 μm from the surface of the organ. These data suggest that full T-cell activation largely follows from prolonged lymphocyte association with individual antigen-bearing DC rather than from summation of signals from many brief encounters with such presenting cells. They also indicate the importance of multilineage cell clusters in effective cell-mediated immune responses. Experiments are in progress to examine effector T-cell development based on cytokine gene activation and use FRET to analyze the interactions of membrane proteins or signaling molecules. These studies should provide a more accurate picture of the molecular, cellular, spatial, and temporal aspects of the cell zinteraction and signaling events involved in host immune responses.
R.N. Germain, None.
Abstract ID: 074
We investigated the potential role of local delivery of immunomodulatory molecules by gene therapy, including interleukin 4 (IL-4) to prevent or treat T1D in NOD mice and the IL-12/IL-23 receptor antagonist IL-12p40, to inflamed joints as a means of treating collagen-induced arthritis (CIA). Collagen type II (CII)-reactive CD4+ T-cell hybridomas and NOD bone marrow-derived dendritic cells (DCs) were transduced with retroviral vectors encoding either IL-12p40 or IL-4 and yellow fluorescent protein (YFP) or luciferase in a bicistronic internal ribosome entry site (IRES)-containing construct. Transduced T-cell hybridomas were FACS-sorted based on expression of YFP and adoptively transferred into immunized DBA/1 mice by intravenous injection one day before booster immunization with bovine CII (Day 20 following priming). NOD DCs were transduced to express IL-4 and an aliquot was assayed for transduction efficiency before adoptive transfer into adult prediabetic NOD mice or recent onset hyperglycemic NOD mice. Treatment with both IL-12p40 expressing T-cell hybridomas as well as IL-4 expressing DCs significantly decreased disease severity. Bioluminescence imaging, using transduction of luciferase-encoding retroviral constructs into T-cell hybridomas and DCs, showed that the hybridomas, as well as the DCs, migrated to and accumulated in the inflamed tissues. There was no significant change in the cytokine milieu of the draining lymph nodes nor in the systemic levels of the auto-antibody subtypes in treated mice, further supporting local delivery. We conclude that the beneficial clinical effects observed in this model were most likely based on the local actions of IL-4 and IL-12p40 in the inflamed tissues. The local delivery (and effects) of regulatory cytokines or receptor antagonists by T cells and DCs constitutes a novel and effective method of treating organ-specific autoimmune diseases and of minimizing the systemic adverse effects of immune-modulating therapy.
C. Fathman, None.
Abstract ID: 075
Tracking the fate of cells in vivo is crucial to elucidate cellular mechanisms and functions. Our goal was to develop a biocompatible superparamagnetic nanoparticle for highly efficient in vivo detection of systemically administered cells by MRI and to apply MR cell tracking to a relevant adoptive transfer mouse model. Based on principles of multivalency, we developed a novel Tat peptide derivatized superparamagnetic nanoparticle, CLIO-HD, and optimized its cellular uptake-efficiency in CD8+ T cells. Labeling with CLIO-HD did not affect T-cell physiology as assessed via in vitro cytotoxicity and cell killing assays, and a physiological flow model of leukocyte-endothelial interactions. CLIO-HD labeled ovalbumin-specific (OT-I) CD8+ T-cell recruitment in an antigen-specific murine tumor model (B16OVA vs. B16F0) was analyzed by MRI (8.5 T) and the results confirmed by immunohistochemistry, autoradiography, and biodistribution studies. CLIO-HD labeling efficiency was optimized 200-fold, and labeling was stable in cells for at least 5 days. This allowed the 3-D visualization of antigen-specific T-cell recruitment to, and departure from, target tumors “dynamically” and “quantitatively” in “live intact animals” with both “very high sensitivity (<3 cells/voxel)” and “high spatial resolution.” Using this approach, we demonstrated that the homing pattern of OT-I CD8+ effector T cells (CTLs) to tumors is “highly heterogeneous” and that “sequential adoptive transfers result in infiltration of CTLs into different tumoral compartments,” potentially resulting in a more efficient anti-tumor attack. The current study is the first to demonstrate that the recruitment of systemically injected cells can be imaged by MRI in live mice, and holds promise for the further investigation of cellular behavior in vivo and in the evaluation of novel cell-based therapies.
Time-course of CLIO-HD labeled ovalbumine-specific (OT-I) CD8+ T cell homing to ovalbumine-expressing (B16-OVA) and control (B16FO) tumor. Serial high-resolution MR imaging was performed following adoptive transfer into a mouse carrying both B16F0 and B16-OVA melanomas.
j, axial; k, saggital; l, coronal plane slices through the 3D reconstruction shown in i.
M.F. Kircher, None.
Abstract ID: 076
With the progress in stem cell biology, application of these cells for regeneration is now being investigated in experimental disease models. For this purpose, it is highly desirable to observe, noninvasively, their location and migration after implantation into host tissue. Here we describe the systematic investigation of the in vivo MRI detection sensitivity using an optimized lipofection method to label cells with iron-oxide nanoparticles, a highly efficient MR contrast agent. Cells of a GFP-expressing murine embryonic stem cell line were labeled with superparamagnetic iron-oxide nanoparticles (SINEREM; Guerbet, France) by lipofection. Electron microscopy reveals accumulation of SINEREM in lysosomes. These iron-oxide particles demarcate the treated cells within the host tissue with high T2* contrast in the MR images. The MR visibility of such labeled cells was assessed and quantified with cell suspensions in gel-phantoms. Cell migration was observed in male Wistar rats, subjected to focal cerebral ischemia. Two weeks later, stereotactic implantation of a murine stem cell suspension (3 × 104 cells) into the cortex and striatum of the healthy hemisphere was achieved. High-resolution in vivo MR imaging was performed on a 7-Tesla animal scanner by 3-D gradient-echo MR imaging at isotropic resolution of 80 pm. Our labeling procedure allowed the detection even of single cells in phantoms. In culture under conditions of unrestricted proliferation, cells were still detectable after three or more divisions. In animals, iron-oxide labeled embryonic stem cells were easily MR-detectable for at least 4 weeks after implantation. After implantation, they showed a directed migration from the cortical injection site along the corpus callosum towards the lesioned hemisphere where they integrated into the border zone of the infarcted area. This was confirmed by GFP fluorescence and GFP immunohistochemistry. We conclude that regeneration processes with stem cells can be studied noninvasively with high-resolution in vivo MR imaging using magnetically labeled cells.
E. Kuestermann, None.
Plenary Session V: Advances in Models Systems of Human Disease
Abstract ID: 078
Spontaneous mouse tumors have been used to study carcinogenesis. While valuable, spontaneous murine tumors rarely resemble human tumors. Even xenographs of human tumors do not retain the structure of primary human tumors. However, the introduction of specific oncogenes has resulted in new tumor types. These new tumor types have molecular and structural resemblance to human cancers. Thus, mouse models offer opportunities to study the structure and function of cancer in intact organisms. The functional modifications start with insertion or ablation of specific genes. In the mammary gland, the prostate, and other organs, each gene results in unique, or signature, tumor phenotypes. The tumor phenotypes are so characteristic that pathologists can identify molecular pathway activated from the histology. For example, mammary tumors from the wnt pathway are distinct from those from the erbB pathway. The prostatic tumors created by SV40 Tag are different from those from the PTEN/AKT pathway. Pathologists “validate” mouse models of human disease. Pathologists use “destructive” techniques to “preserve” the tissue and to create the histological “artifacts” they use for “validation.” However, the structural changes observed by the pathologist are reflected in other functional changes in the tumor in vivo. The imaging scientist has developed tools for assessing tumors in live animals. Unfortunately, little thought has been given to merging the two data sources. The challenge is to couple the functional insights of the imaging scientist with the structure to create nondestructive analysis of structure and function. The pathologist has powerful tools for the capture, analysis, and display of images. The imager has equally powerful tools but on a different resolution level. The challenge is to define and register the functional anatomy from the imaging modalities. As the two disciplines share their insights and technologies, our ultimate goal, a nondestructive pathology, will arise that truly merges function and structure.
R.D. Cardiff, None.
Abstract ID: 079
Neurodegenerative illnesses, including Alzheimer's disease (AD) and Parkinson's Disease (PD), are associated with characteristic clinical signs, genetic risk factors, dysfunction/death of specific subsets of neurons, and disease-defining pathological/biochemical abnormalities. Transgenic and knockout (KO) studies are valuable in understanding disease mechanisms, identifying potential therapeutic targets, and designing/testing novel treatments. In autosomal dominant familial AD, mutant genes encoding the amyloid precursor protein (APP) or presenilin 1 (PS1) influence the levels/length of Aβ, 4-kDa toxic peptides, which are generated by cleavages of APP by β-secretase 1 (BACE 1) and by γ-secretase activities. Aβ accumulates in the brain and deposits are at the extracellular core of neuritic plaques in the brains of individuals with AD. Mice overexpressing mutant APP/PS1 develop age-associated increases of Aβ, neuritic plaques within the hippocampus and cortex, and behavioral deficits in working memory. To illustrate how knockout strategies are used to identify therapeutic targets, we describe work in which we targeted genes encoding proteins critical for pro-amyloidogenic secretase activities. Studies of BACE1 –/– mice demonstrate that BACE1 is the neuronal β-secretase; without BACE1, Aβ is not produced. APPswe; PS1δE9; BACE1 –/– mice do not form Aβ deposits or plaques in brain, nor do they show memory deficits. In PD, mutations in α-synuclein (α-syn) are implicated in a subset of familial cases, and α-syn is also a principle component of Lewy bodies/neurites. A53Tα-syn mice develop motoric deficits, ubiquitinated α-syn inclusions in neurons, α-syn peptide fragments and peptide aggregates. Cleavages of α-syn appear to generate intracellular truncated toxic peptides. The hypothetical “synucleinases” responsible for these cleavages may represent targets for treatment of PD. The results of studies of these models have significant implications for the development of mechanism-based therapies designed to prevent accumulation of Aβ in brains of individuals with AD and possibly other toxic peptides in neurodegenerative diseases like PD.
D.L. Price, None.
Abstract ID: 080
Cancer is a leading cause of death for companion animals; greater than 50% of dogs that live to 10 years of age will develop at least one type of malignancy and the incidence is believed to be similar in cats. Common spontaneous tumors in cats and dogs include non-Hodgkin's lymphoma, mast cell tumors (MCTs), sarcomas (fibrosarcoma, osteosarcoma, hemangiosarcoma), carcinomas (mammary carcinoma, squamous cell carcinoma, transitional cell carcinoma), and malignant melanoma. The biologic behavior of these tumors often mimics the biologic behavior of similar tumors in humans: they arise over months to years and microscopic metastatic disease is usually present at the time of diagnosis. Recent evidence suggests that such spontaneous tumors express molecular targets similar to those found in human tumors. Examples include activating mutations in Kit, aberrant expression of Met, and mutations in p53. As such, spontaneous malignancies in the pet population serve as a relevant model to test the safety and efficacy of novel diagnostic approaches and antineoplastic therapeutics. A Phase I clinical trial investigating the utility of a novel kinase inhibitor (SU11654) was recently completed in dogs with cancer. SU11654 exhibits inhibitory activity against VEGFR, PDGFR, Kit, and Flt-3, resulting in both anti-tumor and anti-angiogenic activity. Objective responses were observed in 16 dogs, including MCTs, mammary carcinomas, sarcomas, and multiple myeloma, for a response rate of 28% (16/57). Stable disease for >10 weeks occurred in an additional 15 dogs for a resultant overall biological activity of 54% (31/57). The highest response rate was observed in MCTs, which, like their human counterpart, often possess Kit mutations. This study provides the first evidence that multi-targeted kinase inhibitors can exhibit broad activity against a variety of spontaneous malignancies. Given the biological similarities of canine and human cancers, it is likely that such agents will demonstrate comparable activity in people.
C. London, SUGEN/Pharmacia 1.
Abstract ID: 081
A.A. Lackner, None.
Plenary Session VI: Advancing Molecular Therapies Through Imaging
Abstract ID: 082
Better mechanistic understanding of disease through mapping of the human and mouse genomes enables rethinking of human infirmity. In the case of cancer, for example, we may begin to associate disease states with their underlying genetic defects rather than with the organ system involved. That will enable more selective, nontoxic therapies in patients who are genetically predisposed to respond to them. Because one of the major goals of molecular imaging research is to interrogate gene expression noninvasively, it can impact greatly on that process. Most of molecular imaging research is undertaken in small animals, which provide a conduit between in vitro studies and human clinical imaging. We are fortunate to be able to manipulate small animals genetically, and to have increasingly better models of human disease. The ability to study those animals noninvasively and quantitatively with new, high-resolution imaging devices provides the most relevant milieu in which to find and examine new therapies.
M.G. Pomper, None.
Abstract ID: 083
Recent technological advances in visible light imaging allowed us to use this technique to noninvasively detect and quantify tumor cells, and to follow responses to drug candidates at a molecular level in mouse models of human cancers. Human prostate tumor cells tagged with a firefly luciferase (PC-3M2AC6) were used in an orthotopic model, and in a model of experimental bone metastasis in athymic nude mice. Upon intravenous injection of
P. Lassota, None.
Abstract ID: 085
The successful gene therapy is highly dependent upon the degree of vector design development. Most of the human gene therapy protocols currently rely on ex vivo gene transfer manipulations, in which certain cells or tissues must be removed from the patient, transduced in vitro, selected for the expression of the transgene, and then reinfused into the patient. It would greatly benefit from the possibility of applying gene therapy approaches based on minimally invasive in vivo gene delivery, and may only require either an injection or oral administration of therapeutic agents. The in vivo transduction approach would allow for a broader application of gene transfer technology in therapy. Gene delivery systems based on nonviral vectors are relatively easy to manipulate, not infectious, and are not very toxic. However, the lack of specific targeting, the low transfection efficiency and transient expression make difficult the in vivo applications of nonviral gene deliver systems. Two nonviral delivery systems are developed in improving the in vivo performance of nonviral gene delivery for cancer gene therapy. Plasmid vectors containing promoters and HSV1-tk gene complexed with delivery reagent have been successfully transfected into Huh 7/SCID mice with tumor growth of 1 cm through tail vein injection. The other approach is to deliver HSV1-tk gene to the tumor growth on NG4TL4/FVB mice by intravenous administration of transgenic stem cells. The stem cells are transfected by HSV1-tk transgenic retroviruses ex vivo. Expression of HSV1-tk gene in the targeted tumor is well demonstrated by [F-18]FHBG microPET imaging. The preliminary results are encouraging in vector development for cancer gene therapy.
R. Liu, None.
Symposium XIII: Drug Discovery and Development
Abstract ID: 086
The vitamin folic acid enters cells either through a carrier protein, termed the reduced folate carrier, or via receptor-mediated endocytosis facilitated by the folate receptor (FR). Because folate-drug conjugates are not substrates of the former, they penetrate cells exclusively via FR-mediated endocytosis. Overexpression of FR on cancer cells and activated (but not resting) macrophages allows selective targeting of folate-drug conjugates to malignant tissues and sites of inflammation, respectively. Clinical data to date suggest that cancers of the ovary, cervix, endometrium, kidney, breast, brain, lung, and colon may all overexpress FR. Importantly, FR expression appears to be even further up-regulated in cancers that are resistant to standard chemotherapy and in higher grade and later stage neoplasms. Thus, precisely those cancers that are most difficult to treat by classical methods are most easily targeted by folate-linked imaging and therapeutic agents. Folate-mediated tumor targeting has been exploited to date for delivery of the following molecules and molecular complexes: (1) protein toxins, (2) chemotherapeutic agents, (3) genes, (4) oligonucleotides, (5) ribozymes, (6) radioimaging agents, (7) MRI contrast agents, (8) liposomes with entrapped drugs, (9) radiotherapeutic agents, (10) immunotherapeutic agents, and (11) enzyme constructs for prodrug therapy. Where live animal studies have been conducted, the data reveal significantly improved response rates with little or no measurable toxicity. Clinical studies of two folate-targeted imaging agents and one folate-targeted therapeutic agent will be briefly discussed. Applications of folate targeting to the diagnosis and treatment of several inflammatory diseases also look promising, and results of animal studies on models of rheumatoid arthritis, lupus, and atherosclerosis will be discussed. (Supported by grants from Endocyte Inc. and NIH grant CA89581.)
P.S. Low, Endocyte Inc. 1, 4.
Abstract ID: 087
Antigens on the surface of malignant cells constitute potential targets for antibody-based therapeutics. These antigens fall into several classes that include: receptors, oncogenes, differentiation antigens, and the carbohydrate portions of various glycoproteins and glycolipids. With rare exceptions, the treatment of cancer with antibodies alone is rarely curative. Because of this, various strategies have been proposed to increase the cytotoxic activity of antibodies and antibody-based molecules. One of these is the arming of antibodies with potent protein toxins. By combining the enzyme activity of the bacterial toxin, Pseudomonas exotoxin, with recombinant antibody fragments, we have produced a number of candidate therapeutic proteins termed recombinant immunotoxins. Here we report on several of these; with a particular emphasis on the activity of BL22, which is directed to CD22 on the surface of various B-cell leukemias and lymphomas. BL22 is composed of PE38 (from Pseudomonas exotoxin) fused to a disulfide stabilized-Fv from the RFB4 antibody. Antitumor activity is reported including many complete remissions in patients with Hairy Cell Leukemia.
D.J. Fitzgerald, Foreign (non-US) rights to a portion of the immunotox-in technology 6.
Abstract ID: 088
Their exists a significant opportunity to bridge the great divide between in vitro and in vivo studies in cancer research and anticancer drug development through the use of novel molecular imaging approaches. A number of novel imaging strategies have been developed that enable noninvasive monitoring of enzymes such as caspases, p53 activation as a measure of carcinogenic potential, diffusion imaging as a measure of cellular density and therefore therapeutic efficacy. Examples will be presented wherein the ability to image luciferase activity as a correlate for cytosolic caspase-3 activity was used to evaluate the therapeutic efficacy and of a novel anticancer agent. The ability to noninvasively image p53 transcription factor activity as a correlate for carcinogenic activity provides for a rapid and sensitive way to evaluate a number of lead drug candidates for potential DNA damaging activity. Lastly, the ability to quantitatively measure cellular density using diffusion-weighted MRI provides for a unique, rapid, and cost-effective way to optimize the dosing and schedule of an anticancer agent.
B.D. Ross, Molecular Therapeutics, Inc. 4.
Symposium XIV: Mechanisms of Oncogenesis and Disease Progression in Tumor Models
Abstract ID: 089
The use of molecular imaging for evaluation of the cellular biology and therapeutic response of orthotopic animal tumor models is essential for improving understanding of the underlying disease process and development of improved therapies. The myriad of new therapeutic cancer targets and targeted antineoplastic agents underscores the need for development of noninvasive surrogate markers for preclinical evaluation. The application of noninvasive imaging methods for quantitating the effects of experimental treatments will accelerate the process of drug development for translation into clinical trials. We have used MRI for quantitating a number of physiological parameters including blood flow (perfusion MRI), tissue cellularity (diffusion-weighted MRI), and tumor growth rates (T1/T2-weighted MRI). Blood flow measurements can be used to evaluate anti- as well as pro-angiogenic potential of an agent as well as the effects of changes in tumor biology (e.g., VEGF expression). Diffusion-weighted MRI has been used to quantitatively follow the cellularity of a tumor over time, thus facilitating drug dose and schedule optimization. Thus, MRI has been shown to be invaluable for assessing the biology and treatment response of orthotopically transplanted xenografts as well as genetically engineered animal models.
B.D. Ross, Molecular Therapeutics, Inc. 4.
Abstract ID: 090
Although distinct pathological stages of human breast cancer have been described, the molecular differences among these stages are largely unknown. Through the combined use of laser capture microdissection (LCM) and cDNA microarrays, we have generated in situ gene expression profiles of the premalignant, preinvasive, and invasive stages of human breast cancer. At the global gene expression level, our data reveal extensive similaritiesat the transcriptome level among the distinct stages of progression and suggest that gene expression alterations conferring the potential for invasive growth are already present in the preinvasive stages. In contrast to tumor stage, different tumor grades are associated with distinct gene expression signatures. In addition, a subset of genes associated with high tumor grade is quantitatively correlated with the transition from preinvasive to invasive growth.
D.C. Sgroi, None.
Abstract ID: 091
An important goal of molecular imaging will be to provide in vivo pharmacokinetics which can be used to enhance treatment efficacy and decrease toxicity. Pharmacokinetics has been studied extensively by both nuclear magnetic resonance (NMR) and radionuclide imaging. Radionuclide imaging is significantly more sensitive than magnetic resonance and can be used to study tracer doses of drugs. NMR spectroscopy and imaging require pharmacologic doses of drug, but have the advantages of studying stable isotopes which are necessary for multi-step complicated syntheses and have an indefinite shelf life. Most importantly, NMR spectroscopy can be used to identify the various metabolites present and thereby measure true concentrations of each metabolite, in addition to being able to sometimes resolve bound versus free chemical species. This session will review some of the uses of NMR spectroscopy in pharmacology, particularly as focused on oncology. Since NMR has relatively low inherent sensitivity, much of the research has focused on studying drugs that contain, either naturally or have been labeled with, a fluorine moiety. The fluoropyrimidines (5-fluorouracil and fluorodeoxyuridine) have been studied in some detail, although other drugs (both with and without a fluorine moiety) have been studied by more sophisticated NMR techniques. Within the last few years, NMR has also been used to monitor the effect of molecular genetic alterations in cancer cells by transfecting genes that activate different drug metabolism pathways. These NMR spectroscopic and imaging studies will be discussed to provide a better understanding of the strength and utility of this technique.
J.A. Koutcher, None.
Symposium XV: Late Breaking Session
Abstract ID: 092
Light passing through tissues is absorbed, scattered, reflected, and refracted; the problems of image reconstruction using laser CT (CTLM) are therefore much greater than in X-ray CT. These problems delayed the development of optical imaging, but have been solved with a level of accuracy permitting absorption and/or fluorescence imaging in human patients. Using appropriate molecules or fluorophores, CTLM provides CT and 3-D images of anatomic, physiologic, and metabolic processes, and can characterize the effects of interventional and pharmacologic therapies. CTLM is mechanically similar to X-ray CT but the X-ray tube has been replaced by a laser and there are two rings of detectors, one configured for absorbed light, from an external laser source, and the other for fluorescence, from a fluorophore within the tissue.
E.N. Milne, Imaging Diagnostic Systems Inc. 2.
Abstract ID: 093
A growing variety of cell surface markers is intimately associated with onset and progression of human disease. Growth factors and antibodies that recognize these markers can serve as targeting vehicles for imaging radionuclides and other imaging agents. Importantly, imaging of multiple markers might lead to better characterization of each patient, and therefore, to development of personalized treatment regiments and to a rational selection of patients for experimental therapeutics. Lack of efficient technology for “loading” imaging agents onto proteins prevents a widespread use of targeted imaging. Current loading strategies are based on random chemical modification of targeting proteins and share common problems such as: (1) partial inactivation of functional domains; (2) heterogeneity of final products; and (3) customized conjugation procedure for every targeting protein. An attractive alternative would be assembly of targeting imaging complexes that include a universal radiolabeled adapter protein that can be noncovalently “docked” to any targeting protein expressed with a corresponding “docking tag.” In addition to being nondestructive, this approach allows the use of multiple existing and newly discovered targeting proteins in a rapid and uniform fashion. We have recently developed adapter/docking tag system based on interactions between an 18–127aa fragment of human RNase I and a 1–15aa fragment of RNase I fused to a targeting protein. The docking system was used to assemble vascular endothelial growth factor (VEGF) driven 99mTc-labeled complexes for imaging of vasculature in tumor-bearing mice. We found that these complexes selectively and specifically image tumor neovasculature in mouse tumors as small as 2–3 mm in diameter. We propose the use of standardized adapter/docking tag system as an alternative to direct derivatization of any targeting proteins with chelators for imaging and, potentially, therapeutic radionuclides. Availability of multiple imaging proteins might have tremendous implications for the development and evaluation of novel anti-cancer and, specifically, anti-angiogenic therapies.
J.M. Backer, SibTech, Inc. 4, 5.
Abstract ID: 094
The in vivo fate, including engraftment and expansion, of transplanted bone marrow cells leading to hematopoietic reconstitution is still poorly understood. Previous preclinical studies on in vivo tracking of transplanted bone marrow cells have had to rely on invasive techniques for harvesting bone marrow from recipients. Despite new insights into fundamental processes such as chemotaxis, migration, and proliferation of transplanted progenitor cells, overall understanding of hematopoietic reconstitution dynamics remains incomplete. We hypothesized that a combined functional and structural multimodality imaging approach can noninvasively monitor the dynamics of hematopoietic progenitor cell engraftment and expansion in the bone, following both systemic as well as intraosseous administration. Bone marrow cells retrovirally transduced to express a fusion reporter gene, that allowed for functional imaging using both bioluminescence (BLI) as well as positron emission tomography (microPET), were used as donor cells. Following administration into congenic recipient animals, these marked cells were tracked repeatedly using multimodality imaging. BLI was used over a period of 28 days with daily monitoring for the first week followed by weekly imaging thereafter. Importantly, we integrated into this imaging sequence, a dual microPET and computed tomography (microCT) imaging of the transplanted cells and the mouse skeleton, respectively. Image reconstruction enabled us to determine the precise location and expansion of the transplanted cells that homed to the bone. The integration of microPET/CT in a sequence of BLI provides a structural reference for the transient optical signal origin. Following the repeated noninvasive imaging over 4 weeks, we validated the BLI signal to be predominantly of bone origin through invasive bone and organ sampling and subsequent visualization of luciferase activity using the same substrate-enzyme reaction in vitro as in vivo. We conclude that multimodality imaging of transplanted bone marrow cells is feasible with potential applications in the field of cell-based biotechnology in vivo.
D. Banerjee, None.
Abstract ID: 095
Overexpression of the peripheral benzodiazepine receptor (PBR) in cancerous tissue has been suggested as an effective target with which to direct contrast agents for the identification of cancer and CNS disease [1]. To this end, we have been focusing our efforts towards effectively delivering lanthanide chelate contrast agents to this receptor in order to image brain cancer. An agent that facilitates multiple scales of imaging from the same compound (both MR and fluorescence), such as the PK11195 conjugated complex demonstrated here, could be particularly useful when disease mapping is coupled to surgical resection, especially with tumors of an infiltrating nature such as primary brain cancer (glioblastoma). We recently reported the synthesis and spectroscopic characterization of a new receptor-mediated PBR targeted molecular imaging agent that produces visibly bright fluorescence [2]. This was accomplished by conjugating an analogue of PK11195 (targeting moiety) to a trifunctional lanthanide chelate (ln-qm-ctmc) (signaling moiety) [2,3]. The true utility of any multimodal agent is in the ability to generate more than one signature from a single matrix, thereby enhancing the s/n, improving the resolving power, and increasing the potential for multiple scales (both macro scale and micro scale) of imaging. After separately demonstrating both fluorescence and MRI signatures from the ln-qm-ctmc-pk11195 agent, we have used a “cocktail” of both eu3+ and gd3+ complexed agent and collected both fluorescence and MR signatures from a single group of C6 glioma cells. We have demonstrated that the cocktail approach is a viable method for multimodal imaging. By dosing C6 glioma cells with a cocktail of the eu3+-pk11195 complex and the gd3+-pk11195 complex, both fluorescence and MRI signatures can be observed from the same cells. Future and ongoing experiments center around demonstrating that enhanced brain tumor detection and diagnosis is possible with this new molecular imaging agent.
D.J. Bornhop, None.
Abstract ID: 096
Transplanted pancreatic islets have emerged as a promising therapy for patients with Type I diabetes. Given that a patient can be normoglycemic with as few as 2% of the native islets being functional, there is a clinical need to assess in vivo islet function to monitor and improve diabetes treatment. We propose manganese (Mn)-enhanced magnetic resonance (MR) microimaging of islet function as an approach to noninvasive in vivo scoring of islet activity. We use Mn as an activation-based contrast agent to observe islet function. Per the fuel hypothesis, glucose activation of pancreatic islet b cells results in calcium influx through Normalized Signal-to-Noise Ratio (SNR) for activated (20 mM glucose, 2.5 mM Mn) versus non-activated (5 mM glucose, 2.5 mM Mn) islets reveals an increase in SNR for islets activated in the presence of the T1 contrast agent manganese, indicating manganese uptake in activated islets (N=7). T1 weighted MR images were obtained on a 11.75 T Bruker Avance spectrometer, with TE = 7.2 ms, TR = 400 ms.
B. Gimi, None.
Abstract ID: 097
Oral cancer is a major health problem worldwide. Despite improvements in therapeutics, survival rates for oral cancer have not improved in the last 30 years. Thus, better early detection and staging tools are needed. Molecular characterization of cancer would have important clinical benefits: earlier cancer detection based on molecular characterization, the ability to predict the risk of cancer progression, real-time margin detection, the ability to rationally select molecular therapy and to monitor response to the therapy. We present a new class of molecular-specific contrast agents for vital optical imaging of intraepithelial lesions—gold nanoparticles conjugated with monoclonal antibodies specific for cancer biomarkers. The nanoparticles can resonantly scatter visible light that provides an opportunity for optical interrogation using reflectance spectroscopic and imaging modalities. We present data to image two molecular targets which are up-regulated at an early stage in many cancers, including cancer of the oral cavity: the epidermal growth factor receptor (EGFR), and matrix metallopreinases (MMPs). Using progressively more complex biological models of oral cancer, ranging from cancer cells in suspension to an in vivo animal model, we have demonstrated that these optical contrast agents have the potential to provide sensitive tool to image the molecular changes associated with oral cancer. Using cancer cells in suspension and in 3-D tissue culture, we demonstrated labeling of EGFR up to a depth of 500 μm using simple penetration enhancers. Similarly, we visualized MMP expression on cellular membranes and along collagen fibers in 3-D tissue cultures. Fresh tissue slices from normal and precancerous human oral mucosa show overexpression of EGFR in neoplasia detected optically using these contrast agents. Finally, we present data from an animal model of oral cancer, the DMBA-hamster cheek pouch model, which illustrates that contrast agents can be used in vivo to image EGFR overexpression in neoplasia.
A.M. Gillenwater, None.
Symposium XVI: Research and Industrial Collaborations
Abstract ID: 098
Philips is a major player in molecular imaging and diagnostics and has embarked on an ambitious and exciting program. The company recognizes the paradigm shift that molecular imaging will cause not only in its own area of competence—imaging, in which it has a leading position in most of the modalities—but also in the areas of contrast and therapeutic agents, as well as in related peripheral equipment such as molecular diagnostics. Philips is building new competencies to meet future diagnostic and therapy guiding requirements related to human clinical biochemistry. The company is optimizing several of its current imaging modalities, such as MRI, PET/CT, SPECT, and ultrasound, for molecular medicine applications. New combination platforms (SPECT-CT, MR-X-ray, CT-X-ray) are under development or already installed in academic centers. Preclinical imaging systems and tools are recently introduced and part of our product offer now. One of Philips' core molecular imaging activities is the creation of collaborative programs with academic programs focusing on the creation of new molecular markers, developing new multimodality imaging strategies ranging from small animals to whole body applications, and ultimately translating these new concepts in clinical applications. Clinical focus areas are cardiology, oncology, and neurodegenerative disease. These programs involve multidisciplinary interactions between Philips' corporate research laboratories, the research and development centers, and the required competencies of our academic partners. The program encompasses all relevant imaging modalities, as well as close interactions with contrast agent manufacturers and drug companies. Philips has developed collaborative agreements with several leading pharmaceutical companies to create imaging systems that optimize image content, as well as software that provides physicians with quantitative clinical information. “Collaboration is in our DNA” and Philips' extended user network of research partners in many imaging research area's is a testimony of Philips' commitment to help moving this new field forward.
P. Luyten, Philips Medical Systems 5.
Abstract ID: 099
Molecular imaging (MI) is clearly an emerging technique. Siemens is well aware that it will take some time until there is a considerable MI market besides nuclear imaging. However, encouraging advances achieved in clinical PET and animal research indicates that MI will continue to evolve into an indispensable diagnostic tool. MI will complement morphological imaging procedures and in vitro diagnostics, resulting in a substantially improved patient treatment. MI research is heavily intradisciplinary, as it requires an in-depth understanding of devices, IT, algorithms, probes, and molecular medicine. As a solution provider, we are going full force on all of these fronts. We are getting involved in small animal imaging (SAI) and in research on MI probes to further the field of clinical molecular imaging. By testing cutting-edge applications in animals, we hope to alleviate the bottleneck in translational research. We are leveraging this knowledge base to enable Siemens equipment and software for MI clinical trials. One luxury of being a leader in the field of imaging devices and medical IT is our broad experience. Our MI effort not only involves our device divisions, but our IT groups, our medical and biology experts, and our luminary collaboration partners as well. Siemens provides the resources to fully exploit the synergies between all of these areas. In addition, we have a corporate tradition of committing on basic research that will only pay off over the long term. Therefore, we are committed to this field in spite of the time horizon. The business opportunity we are putting into place today will not pay off this year or next, but in the future it will.
M. Naraghi, Siemens Medical Solutions 5.
Symposium XVII: Funding Sources
Abstract ID: 101
Funding for molecular imaging research projects is available from a variety of NIH institutes and centers. This presentation will provide an overview of funding opportunities available for individual and team projects, as well as training grant mechanisms. Differences in policies and procedures among institutes and centers will be discussed, as well as issues related to application assignment and review. Ample time for questions will be provided.
D.C. Sullivan, None.
Abstract ID: 102
An overview will be presented for NCI support for molecular imaging systems and methods and related informatics. Examples of research include system optimization for molecular imaging, related multimodality imaging systems, data extraction, and validation methods. Different funding sources and mechanisms will be described.
L.P. Clarke, None.
Abstract ID: 103
The newest institute of the National Institutes of Health (NIH), the National Institute of Biomedical Imaging and Bioengineering (NIBIB) has the mission to improve health by promoting fundamental discoveries, design and development, translation and assessment of technological capabilities in biomedical imaging and bioengineering. The NIBIB has programs and initiatives in a variety of areas of imaging and bioengineering including molecular imaging, image-guided interventions, imaging device development, and biosensors. The primary conduit of information, the NIBIB website, may be viewed at http://www.nibib.nih.gov. Grant applications that propose to test or develop a new technology are of particular interest to the NIBIB. The NIBIB has a variety of funding mechanisms for training and research, including Exploratory-Development (R21) Grants that fund innovative, high-impact research projects that (1) demonstrate the feasibility of new technologies that could have a major impact in a specific area; (2) involve high-risk experiments that could lead to a breakthrough in the field of biomedical imaging and/or bioengineering; or (3) generate pilot data to assess the feasibility of a novel avenue of investigation. Source: http://www.nibib.nih.gov.
J.W. Haller, None.
Abstract ID: 104
Over the past several years, the National Cancer Institute (NCI) has been keenly aware of the potential power of molecular imaging techniques. The Biomedical Imaging Program (BIP) http://cancer.gov/bip/default.htm of the Division of Cancer Treatment and Diagnosis (DCTD) of the NCI is responsible for the extramural grant portfolio and programs related to oncologic imaging. Imaging has been identified as an area of “Extraordinary Opportunity” in the past several “NCI Bypass Budgets” http://plan.cancer.gov/scipri/imaging.htm. Numerous NCI supported molecular imaging programs and initiatives were made possible due to imaging being identified as an “Extraordinary Opportunity for Investment.” To assure that the initially defined goals for molecular imaging are met and completed in future years, the NCI continues to support programs and initiatives that support investigators interested in molecular imaging. These include: (1) Support for In Vivo Cellular and Molecular Imaging Centers (ICMIC) http://www3.cancer.gov/bip/ICMICs.htm; (2) Support for the Development of Clinical Imaging Drugs and Enhancers (DCIDE) program http://cancer.gov/bip/dcide.htm which will facilitate and promote preclinical development and validation of important imaging agents and ligands; (3) NCI will, on a competitive basis, synthesize, test, and distribute probes that image the physiological and functional status of tumor tissue in the human body; (4) Establish a publicly available database of molecular imaging agents together with information on their chemical structure, properties, and development status. This presentation will provide an overview of the molecular imaging oriented initiatives and programs supported by the Biomedical Imaging Program (BIP) of the NCI. Cancer-specific funding opportunities for molecular imaging will be discussed. Ample time for questions will be provided.
J.M. Hoffman, None.
Poster Session 01: Chemistry: Novel Probes and Activation Strategies
Abstract ID: 105
The enhancement characteristics of a contrast agent are dependent on its pharmacokinetics within the body. In the case of macromolecular contrast agents, prolonged enhancement of the blood pool is seen after the first dose limiting opportunities for repeated injection in the same session. If the enhancement within the blood pool could be intentionally and effectively “switched off,” macromolecular contrast agents could be used both to define blood volume and vessel permeability, properties that could be useful in studying angiogenesis. In this current study, we attempted to develop a contrast agent system in which the blood pool enhancement might be controlled and intentionally “switched off.” The avidin-biotin chase system coupled with a blood-pool type, generation-6 polyamidoamine dendrimer-based macromolecular MRI contrast agent (PAMAM-G6) was employed to switch enhancement from the blood pool to the liver. A biotin-conjugated PAMAM-G6 (bt-G6) containing five biotin molecules and 205 Gd ions with a molecular weight of ˜240 kDa was synthesized. The bt-G6 was studied with dynamic MRI inmice followed by the injection with either avidinor PBS 4 min postinjection of bt-G6. Because the avidin chase causes rapid trapping of the contrast agent in the liver, the enhancement with bt-G6 in the blood pool was cleared within 2 min of the avidin injection. However, the enhancement with bt-G6 was not altered by saline injection. In conclusion, this system, macromolecular MRI contrast agents coupled with avidin chase, can be applied to all dendrimer-based MRI contrast agents to investigate blood volume and vascular permeability. Moreover, it permits the repeated injection of the contrast agent and the “avidin switch” during a single MR experiment.
H. Kobayashi, None.
Abstract ID: 106
Intercellular communications through gap junctions play essential roles in maintaining cell homeostasis and synchronizing biological functions. We are studying how cellular calcium activity is transmitted through gap junctions and how it affects signaling of coupled cells. We have developed several photo-caged and cell-permeable probes. These compounds can be loaded noninvasively into fully intact cells. Using localized photolysis, we can turn on cellular calcium activity with well-defined temporal and spatial characteristics. We then integrate intercellular transmissions by tracking both the calcium waves and the spread of fluorescent tracers simultaneously. Inositol 1,4,5-trisphosphate (IP3) is a second messenger inducing calcium release from internal stores. When the calcium activity transfers to neighboring cells, it is unclear of the relative roles of IP3 and Ca2+ signaling in the process. We have developed IP3 analogues of varied cellular metabolic stabilities. Local uncaging of these molecules causes calcium increases in one chosen cell among a group of coupled cells. Comparisons between these IP3 analogues showed that cellular calcium activity has profound effects on cell coupling, whereas IP3 by itself appears to be insufficient for propagating calcium waves. To investigate the mechanisms of gating of connexin channels, we have also developed caged and cell-permeable fluorophores. Remarkably, the uncaging cross sections of these molecules are nearly two orders of magnitude higher than those of previously described caged compounds. Upon photo activation, these caged compounds exhibit an optical enhancement of over 300 folds. The extraordinary photochemical properties of these dyes ensure that we can generate a robust fluorescence signal jump in cells with a minimum dose of UV light. In addition, the spectral properties of these fluorophores allow us to carry out simultaneous uncaging and multicolor imaging in live cells. The applications of these new probes in studying the spatiotemporal dynamics of cellular signaling in vivo will be discussed.
W. Li, None.
Abstract ID: 107
With the progress in the fields of more sophisticated optical signal-detection devices and fluorophores with greater tissue-penetration properties, molecular optical imaging for small animals is gaining great interest because of its merits of sensitivity, convenience, economy, and avoidance of radioactive hazards. The objective of this project is to develop an optical imaging agent for the investigation of apoptotic processes in small animals and to make the newly emerged small animal fluorescence imaging technology more accessible to scientific researchers. Annexin V was conjugated with a near-infrared fluorochrome, cyanine-5.5 (Cy5.5). The dye to protein ratio of Cy5.5 to annexin V (D/P) was characterized. The in vivo biodistribution and blood clearance, binding affinity, and target specificity were determined. In vivo imaging characteristics of the agent was evaluated in mice models using an IVIS small animal fluorescence imaging system. The D/P was optimized to 2 in terms of fluorescence quantum yield. The binding affinity, Kd, of the Cy 5.5-annexin V to phosphatidylserine (PS) was measured as 7.12 nM, comparable with the unmodified annexin V, Kd = 1.66 nM. Flow cytometry and confocal fluorescence microscopy showed the Cy5.5-annexin V targets mouse thymocytes in the same patterns as FITC-annexin V. Fluorescence images of mice with apoptotic MCA-29 tumor demonstrated that the CAV5.5 accumulated more in an irradiated tumor compared with the untreated tumor. We have demonstrated that the recombinant annexin V can be conjugated to cyanine 5.5 with its binding affinity to PS remained. In vivo fluorescence images demonstrate that the CAV5.5 is a promising agent for apoptosis detection in small animals.
D. Pan, None.
Abstract ID: 108
Whole genome sequencing of many organisms provides a large number of sequence information. To understand each gene function in greater detail, expression profiling is necessary to be made. Here we present a novel method to monitor gene expression in yeast cells by using 31P MRI. Yeast cells accumulate a large amount of inorganic polyphosphate (PolyP), which can be readily detected by 31P NMR/MRI. In Saccharomyces cerevisiae, a family of genes, PHM1–5, was shown to regulate PolyP accumulation [Ogawa N., et al, Mol. Biol. Cell, 
H. Tochio, None.
Abstract ID: 109
Near-infrared or infrared quantum dots are required for biomedical imaging applications in order to maximize photon yield. However, many combinations of emission wavelength and hydrodynamic diameter are not accessible using previously synthesized quantum dots. The major cancer staging procedure termed sentinel lymph node mapping is one such procedure that requires novel quantum dot engineering. We have developed novel “type-II” core/shell quantum dots that permit emission wavelength to be tuned, with flexibility, into the near infrared, and beyond, using well-characterized combinations of semiconductor materials. Our objective was to determine whether such quantum dots could be utilized for sentinel lymph node mapping in small and large animals. We present a simple colloidal synthesis for type-II QDs, which permitsfine tuning of emission wavelength and particle size using unique combinations of semiconductor materials. We also present a novel oligomeric phosphine organic coating that renders QDs stable in serum while maintaining quantum yield and photostability. Using 860 nm near-infrared quantum dots in conjunction with our previously developed intraoperative NIR fluorescence imaging system, we demonstrate that sentinel lymph node mapping can be performed in porcine models, and in real time. Sentinel lymph node imaging is both sensitive and specific, and provides visual guidance to the surgeon before, during, and after nodal resection. All sentinel lymph nodes were confirmed histologically, and were also shown to co-localize with the current gold standard, isosulfan blue. Sentinel node mapping in 35 kg pigs could be performed with as little as 400 pmol of quantum dots, with a fluence rate of only 5mW/cm2. The metastasis of many human tumors is via lymph channels. We have engineered novel type-II quantum dots for both near-infrared fluorescence emission and a particle size optimized for sentinel lymph node mapping. Properly sized, near-infrared fluorescent quantum dots make light-based sentinel lymph node mapping a potential clinical reality.
J.V. Frangioni, None.
Abstract ID: 110
FDG has proven successful as a PET imaging agent for detection and localization of many forms of cancer but is ineffective for the diagnosis and/or (re)staging of prostate carcinoma. Identification of alternative PET agents able to identify the molecular determinants of prostate cancer would be of great importance to the management of this disease. Prostate cancer is angiogenesis dependent and the cell adhesion receptor integrin αvβ3 is highly expressed in prostate cancer tumor cells and angiogenic prostate tumor blood vessels. The objective of this study is to demonstrate the feasibility of suitably radiolabeled αv-integrin antagonists for PET imaging of prostate cancer and its metastasis to the bone. 125I-labeled cyclic RGD peptide c(RGDyK) showed rapid and high uptake in subcutaneous PC-3 tumor models (3.98 ± 0.45%ID/g at 30 min pi). MicroPET imaging with 18F-RGD and 64Cu-DOTA-RGD allowed clear visualization of the αvβ3-expressing PC-3 tumor with high tumor/background ratios in accordance with direct tissue sampling technique. These two tracers also demonstrated the potential for PET imaging of prostate cancer bone metastasis. We obtained a bone xenograft model by intrafemoral injection of PC-3 cells to the right tibia. As seen from the figure below, FDG showed slightly higher uptake on the right thigh than on the normal left thigh. [18F]F− was unable to distinguish bone xenograft from normal bone tissue; both 64Cu-DOTA-RGD and 18F-RGD indicated significant uptake in the osteoblastic bone lesions, with virtually no uptake in the contralateral intact tibia. In this pilot study, radiolabeled RGD peptides were superior to FDG and [18F]F− for assessing prostate cancer bone metastasis. (Supported in part from DOD PC020544, ACS-IRG-58007–42, 5P20 CA86532, R24 CA86307).
X. Chen, None.
Abstract ID: 111
Bispecific antibody (bsAb) pretargeting presents an excellent opportunity to couple the molecular specificity of antibody targeting with the rapid circulatory elimination of a later-administered hapten, in order to obtain high target-to-nontarget ratios in multiple disease imaging modalities. A particularly attractive application is the use of bsAb pretargeting to deliver Ga-68 for positron emission tomography (PET). We sought to develop methods and compositions to enable the use of Ga-68-labeled species in this application, and to test their utility in a preclinical model of human cancer. We used Gallium-68 taken directly from a germanium-68/gallium-68 generator system to radiolabel the DOTA ring of the peptide DOTA-Phe-Lys(HSG)-
G.L. Griffiths, Immunomedics, Inc 5.
Abstract ID: 112
We describe a new approach to the design of cell specific probes, where a library of polyvalent nanoparticles with chemically modified surfaces was synthesized and screened for interactions with normal and apoptotic cells. Central to the development of these probes is the ability to synthesize libraries of nanoparticles with a variety of surfaces and to screen them for their ability to bind target cells identified with a dual wavelength FACS screening method. To illustrate this concept, we screened out of a larger library 10 surface modified nanoparticles modified with neutral, positive, and negative compounds against apoptotic cells. Normal and camptothecin-treated Jurkat T cells were incubated with a mixture of the FITC labeled CLIO nanoparticle and Cy5.5–annexin V, a standard protein marker for apoptosis. One nanoparticle with a high surface density of guanidinium groups (CL-45) was highly selective for apoptotic cells: In untreated or camptothecin-treated cell samples, normal cells and apoptotic cells (binding cells) bound annexin V and CL-45 similarly. For example, after camptothecin treatment, 45.4% of cells were labeled poorly by both reagents, and 50.4% of cells were highly labeled with both reagents. Hence, the binding of Cy5.5-annexin V and CL-45 correlated in 95.8% of the cells analyzed. Apoptotic cells had 11-fold higher CL-45-fluorescence intensity than normal cells. In common with some phage display selection methods, our method targets cellsindifferent functional states, normal and apoptotic, without the knowledge of the molecular target. Since all chemistry of probe design precedes screening, bioactivity cannot be lost due to the conjugation procedures. The principle demonstrated here provides a novel approach to the development of nanoparticle-based materials including quantum dots, colloidal gold, or crystalline forms of drugs for a wide range of applications like MR and optical imaging, drug delivery, and purification.
E.A. Schellenberger, None.
Abstract ID: 113
Many animal models of cancer progression rely on sensitive detection of metastases or of minimal residual disease after therapy. This detection usually involves invasive procedures. Recently, reporter gene constructs such as GFP and luciferase have been used to create cell lines whose metastases are detectable at earlier stages of disease. However, the visible photons produced by these reporters have limitations when imaging from deep tissues in larger organisms. We present a noninvasive method for imaging experimental tumors and metastases in vivo in real time using the sodium-iodide symporter (NIS) gene. A selectable mammalian expression vector containing the human NIS gene was used to transfect the human head and neck squamous cell carcinoma cell line FaDu. Individual clones were assayed for in vitro [125I]iodide-accumulating ability. One clone, NIS-FaDu-4, was identified which concentrated radioiodide to levels approximately 50-fold higher than the parent cell line. NIS-FaDu-4 cells were injected subcutaneously in the flanks of athymic mice. After tumor xenografts formed, mice were injected via tail vein with 500 microcuries of [99mTc]pertechnetate or with 1 millicurie of [125I]iodide and were then imaged by gamma camera scintigraphy, or serially sectioned and imaged by autoradiography to confirm and extend the gamma camera results. NIS-FaDu-4 tumors as small as 4 mm were easily visualized by gamma camera, whereas tumors in animals injected with the parent cell line were not. Similarly, 4-mm tumors were clearly visible in serial sections by autoradiography, whereas FaDu tumors were undetectable. Studies using spontaneously metastasizing NIS-expressing tumor cells are underway to determine the threshold of detection of metastases. Detection of experimental tumors and their metastases is crucial to studies of cancer progression and detection of minimal residual disease. The properties of NIS-expressing cancer cells make them amenable for detection of early metastases. (Supported by NIH CA91709)
M.M. Graham, None.
Abstract ID: 114
Noninvasive, tissue-specific imaging in transgenic animal models of disease offers the ability to study the growth and development of tumors and response to therapy. This study reports the development and imaging of a novel transgenic mouse line carrying the firefly luciferase (fl) transgene under the control of a prostate-specific promoter. Transgenic mice were generated in FVB background by standard techniques using the DNA fragment PSEBCVP2-G5-fl. The fl positive mice (F1 progeny, identified by PCR genotyping) were imaged at different intervals of time using a cooled charge coupled device (CCD) camera and
M. Iyer, None.
Abstract ID: 115
Molecular imaging strategies using noninvasive imaging techniques such as MRI, PET, SPECT, or optical imaging are currently being developed in mouse model systems where these strategies can be optimized. Endogenous receptor/labeled ligand reporter systems are typically used in these approaches. However, imaging strategies based on endogenous receptors can suffer from a number of potential drawbacks such as the competition of labeled ligand with endogenous ligand and the expression of the receptor at sites outside the target tissue. Our goal is to develop a nonmammalian receptor/ligand imaging system to overcome these potential difficulties. We are evaluating the Er-1mem/Er-1 (receptor/ligand pair). Er-1mem is a type II transmembrane cell surface receptor from the marine protozoan cilate Euplotes raikovi. Er-1mem has no mammalian homologues and a small (5 kDa) ligand, Er-1, which can be readily labeled for any imaging modality. We have placed the Er-1mem gene under the control of the CMV promoter for constitutive expression, and under the hypoxia response element promoter for conditional expression (i.e., hypoxia induced) in mammalian cells. We have shown that a green fluorescent Er-1mem fusion protein can be expressed in transiently transfected breast cancer cells in a manner that appears to be similar to that of a known plasma membrane protein claudin-7 fused to yellow fluorescent protein. We have evidence that Er-1mem can be expressed under hypoxia mimicking conditions (CoCl2), at the surface of breast cancer cells. A fusion gene of Er-1 and glutathione S-transferase has been prepared to facilitate the preparation of pure Er-1 (ligand). The Er-1mem/Er-1 system will be useful for wide-ranging applications in detecting gene expression, or gene expression driven by microenvironmental conditions such as hypoxia. In combination with the functional imaging capabilities of PET and MRI, the Er-1mem/Er-1 system will allow a molecular/functional imaging approach to understanding the impact of a gene of interest.
P.T. Winnard Jr, None.
Abstract ID: 116
Molecular imaging methods have previously been employed to image tissue-specific reporter gene expression by a two-step transcriptional amplification (TSTA) strategy (Iyer et al. PNAS, 98(25):14595–600; 2001). We have now developed a new bidirectional vector system based on the TSTA strategy that can amplify expression using a relatively weak promoter simultaneously for both a ‘target gene’ and a reporter gene. We used the synthetic renilla luciferase (hrl) and the firefly luciferase (fl) reporter genes to validate the system in cell cultures and in living mice. When mammalian cells were transiently co-transfected with the Gal4-responsive bidirectional reporter vector (fl-G5-hrl) and different doses of the activator plasmid encoding the Gal4-VP16 fusion protein, pSV40-Gal4-VP16, a high correlation (r2=.95) was observed between the expressions of both reporter genes. High correlation (r2 =.82) was also observed in vivo when the transiently transfected cells were implanted subcutaneously in mice and imaged. This work establishes a novel bidirectional vector approach utilizing the TSTA strategy for both target and reporter gene amplification. The approaches validated should have important implications for development of novel gene therapy vectors, as well as for transgenic models, allowing noninvasive imaging for indirect monitoring and amplification of target gene expression.
S. Ray, None.
Abstract ID: 117
Imaging in the visible light region is typically associated with high absorption, scattering, and autofluorescence. In this work, we describe a new water-soluble near-infrared fluorochrome (NIR820), which can be synthesized in large quantities using a three-step reaction. Furthermore, NIR820 can be purified by flash column chromatography and does not require HPLC purification, like many other NIR fluorochromes. Selective activation of NIR820 with succinimide ester yields a product which can be used to label proteins and other affinity ligands. In addition, the chemical stability of NIR820 allows it to be used directly during solid-phase peptide synthesis. A series of novel metalloproteinase sensitive NIR820 probes were synthesized, and we demonstrate the simultaneous imaging of activities using these probes in combination with other NIR fluorochromes. In vivo testing of these enzyme-sensitive probes is currently in progress.
W. Pham, None.
Abstract ID: 118
Bioluminescence resonance energy transfer (BRET) is an advanced, sensitive, cell-based assay to measure protein-protein interactions among hormone receptors and for mapping signal transduction pathways. Our objective is to optimize the BRET2 strategy for noninvasively imaging protein-protein interactions in living mice. 293T cells were transfected with CMV-GFP2-Rluc or MV-Rluc plasmids and 30 hr posttransfection bioluminescence was measured by adding 0.5μg/well deep blue coelenterazine (DBC). We used suitable filters to measure the donor Rluc-DBC (peak at 395 nm) and the acceptor GFP2 (peak at 510 nm) emission light separately in a Xenogen IVIS CCD camera setup. After implanting cells subcutaneously in nude mice with both transiently transfected and stable cells, the mice were imaged by delivering 25 μg DBC/mice/scan via tail vein and using the specified filters. For detection of BRET2 signal from specific protein-protein interactions, MyoD and ID proteins were cloned in combinations at N-and C-terminus with the acceptor and donor molecules. 293T cells expressing BRET2 give max of 98.16 × 104 p/sec/cm2/sr using a 500–570 nm filter, which is significantly greater than 74.47 × 104 p/sec/cm2/sr with a 360–460 nm filter, whereas, Rluc expression alone gives a max of 5.61 × 104 p/sec/cm2/sr and 37.42 × 104 p/sec/cm2/sr with the respective filters. Similarly, BRET2 signal is also detectable from nude mice with implanted cells expressing BRET2 in comparison to Rluc (see figure), as well as GFP2-MyoD and ID-Rluc together in comparison to ID-Rluc alone. The CCD camera can be used to monitor transient and stable BRET2 signal from the surface of nude mice. The ability to measure the BRET signal from living mice should have significant impact for studying fundamental biological processes dependent on protein-protein interactions and drugs targeted against protein-protein interactions.
Detection of BRET2 signal in a living mouse. Five million 293T cells expressing Rluc (L) and BRET2 (R) plasmid were implanted s.c. After an hour, DBC was injected and then the mouse was imaged to detect GFP emission for first 5 minutes and then the filter was changed to detect Rluc-DBC emission for next 5 minutes
A. De, None.
Abstract ID: 119
Liposomes are artificially constructed spheres of lipid bilayers frequently used for the delivery of drugs and nucleic acids into cells. An important development for liposome technology is the ability to monitor both their pharmcokinetic behavior and site of delivery in vivo, especially in long-term serial studies. This could potentially be achieved by the development of novel liposome membranes consisting of labeled components, such as contrast agents for magnetic resonance imaging (MRI). The aim of this research is to design and synthesize a series of gadolinium lipids for incorporation into the liposome membrane. Lipids will be analyzed by MRI, both individually and within the liposome membrane, allowing elucidation of a structure-activity relationship for these new compounds. Successful candidates will be used in experiments imaging liposome delivery in vivo. To date, four lipids have been synthesized. The structure of these imaging agents comprises a simple phospholipid or cholesterol-based lipid, bonded through the lipid head group to an octadentate ligand, namely, DOTA or DTPA, which permits coordination of the Gd metal ion (see figure). Preliminary T1 relaxation measurements for these compounds demonstrate them to be extremely effective contrast agents, reducing the T1 of the solvents to 1–3% of their original value. These target compounds are presently being evaluated in in vivo studies. Additionally, we are developing a novel solid-phase synthesis methodology for the functionalization of DOTA or DTPA and the development of new ligands with varying linkers. This methodology will allow rapid and facile functionalization of these polyamines and permit easier and large-scale purification of the final ligand.
M.H. Oliver, None.
Abstract ID: 120
Vectors containing wild type herpes simplex virus-1 thymidine kinase (HSVTK) or mutants are useful as PET reporters in molecular imaging and as therapeutics in suicide gene therapy. To better understand the cellular response of these reporter systems, quantitative in vitro enzyme kinetics were studied. Wild-type HSVTK was cloned into a bacterial expression vector and the enzyme was purified by affinity chromatography under nondenaturing conditions. In vitro kinetics studies were completed with the substrates 3H-thymidine, 14C-FIAU, and 3H-penciclovir. For purified wild-type HSVTK, the following parameters best fit the data: thymidine (Km = 0.30 uM, kcat = 0.50 1/sec), FIAU (Km = 0.89 uM, Ki = 0.407 uM, kcat = 2.42 1/sec), penciclovir (Km = 2.5 uM, Ki = 1.5 uM, kcat = 0.0536 1/sec). The rank order correlated well with cell culture uptake/washout studies previously reported from the laboratory. Interestingly, strong substrate inhibition was observed for both FIAU and penciclovir at concentrations lower than the Km. In vitro kinetic analyses such as these permit comparison of the efficacy of various substrates over a wide range of substrate concentrations from PET tracer levels to clinically relevant pharmaceutical concentrations independent of confounding factors present in whole cell assays.
S.T. Gammon, None.
Abstract ID: 121
A rapid and sensitive bioluminescent assay is being developed for imaging enzymatic activity both in vitro and in vivo, in live cells and whole organisms. Luciferin is catalyzed by luciferase and produces light in the presence of ATP and Mg2+/O2. Replacing the hydroxyl group on luciferin with an amine group (aminoluciferin) does not adversely affect its use as a substrate. This amine group also creates a site for adding amino acid residues to synthesize substrates for proteases. Thus, cleavage of these attached peptides by proteases reactivates the protected aminoluciferin. Aminoluciferin serves as a precursor for automated synthesizers and can be readily incorporated into synthetic peptides, RNA, DNA, and carbohydrates. Activateable luciferin analogs form the basis of a new molecular probe for in vivo imaging, as uniform expression of the luciferase enzyme can be engineered in cells and tissues. The initial proof of concept of this application has been demonstrated and can lead to development of novel imaging assays for pathogen detection, testing protease inhibitors, probing cell physiology, and assessing protease activity in oncogenesis. Activation of the luciferyl peptide by chymotrypsin was demonstrated in in vivo mouse models using a transgenic mouse (from the immediate early gene of human cytomegalovirus) to express a luciferase that had been modified for optimal expression in mammalian cells. Upon addition of the protease, chymotrypsin, aminoluciferin was cleaved from the luciferyl peptide. Aminoluciferin produced visible light and oxyaminoluciferin. To facilitate the synthesis of peptide-based molecular probes, we modified the aminoluciferin with Fmoc groups that allow its use in automated peptide synthesizers for rapid incorporation into peptide substrates for a variety of proteases, within a peptide substrate or as the terminal amino acid residue. We have synthesized substrates for HIV-1 and Prostate Specific Antigen (PSA) protease. The sensitivity and selectivity of these probes to the proteases is being evaluated.
R.R. Shinde, None.
Abstract ID: 122
We are investigating adenoviral protein scaffolds to develop novel ligands for molecular imaging applications. Our initial focus is on the recombinant “knob” domain of the adenovirus fiber protein, which interacts with its in vivo cellular receptor CAR. The knob will be engineered to interact with molecular targets (e.g., tumor-associated antigens) using in vitro evolution methods. This study reports our baseline results on the biodistribution of I-131-labeled knob administered intravenously in mice. First, NaI-131 was incubated with one Iodobead for 5 min in Tris pH 6.8. The reactive iodine species was separated, and the knob protein was added to it for 1 min. This method produced I-131-knob without SDS-PAGE detectable conformational damage, or loss of CAR-binding activity. In mice, there was expected uptake in CAR-binding tissues (liver, kidney, heart, and lung), but blood and whole body clearance were rapid. In mice that were injected with excess unlabeled knob (cold competitor), the level of I-131-knob (% id/g) in blood was increased (19.4 vs. 2.5) with a corresponding decrease in metabolic deiodination (6.5 vs. 33.5 in stomach). I-131-knob mutants, which are devoid of CAR-binding activity, showed high levels (% id/g) in blood at 6 hr postinjection (31.0), whereas wild-type I-131-knob in blood at 6 hr had decreased to background level (2.0). In summary, knob proteins with intact CAR-binding activity localize in tissues with CAR receptors, but their blood clearance and metabolism are rapid. Knob mutants (unable to interact with CAR) are retained longer in the blood and have slower metabolism. Blood and body clearance seem to be both CAR-dependent and saturable. These results suggest that engineered ligands based on knob scaffolds with inhibited CAR-binding activity would have slower breakdown and circulate long enough to specifically interact with their molecular targets in tumors or other tissues. (Work supported by U.S. DOE (OBER) under Contract #DE-AC02-98CH10886.)
S.C. Srivastava, None.
Abstract ID: 123
The aim of this study was to assess the feasibility of using a cholesterol-based contrast agent to label cell membranes. A novel cholesterol-based lipophilic MR contrast agent, Gd-DTPA-Chol, was synthesized. In phantom experiments, Gd-DTPA-Chol induced significant relaxation of solvent protons by both T1 and T2 mechanisms. Gd-DTPA-Chol (0.2 mmol/kg) was administered in 100% DMSO injection vehicle to a muscle of the hind leg of a mouse. DSMO was administered to the other hind leg (control). T1-weighted MRI (4.7 T) was performed at t=0, 6, 24 and 48 hr. The figure shows a representative slice at 24 hr. The arrow indicates the side in which the Gd-DTPA-Chol was administered. The DMSO vehicle in the other hind leg showed no significant change by T1-weighted imaging. In each of the scans, a hypointense region surrounded by a hyperintense region was evident. It is assumed that the hypointense region is the location of the bolus injection. There was a gradual decrease in the hyperintense regions with time with persistence of the hypointense areas. The percent increase in signal intensity of the hyperintense region relative to unenhanced muscle was: 337%, 272%, 206%, and 180% at 0, 6, 24, and 48 ht postinjection, respectively. The continued hyperintense regions around the bolus over time may arise from the incorporation of the novel contrast agent into membrane structures since Gd-DTPA-Chol is based on cholesterol, a constituent of cell membranes. Further investigation by histology and atomic absorbance will be done to verify the structures in the hyper- and hypointense regions.
A.H. Herlihy, None.
Abstract ID: 124
Recent advances in imaging technology have brought high-resolution imaging into the practical laboratory setting. As a result, the number of micro-imaging applications has increased rapidly. Imaging processes at the cellular level is one avenue that is being hotly pursued. To image an individual cell or group of cells, a high-contrast particle can be inserted into the cell, which is then introduced into the animal. How well the cells are visualized in the image is dependent upon the amount of metallic label present in each element of the image volume, on the size of the contrast-enhanced region, and on the signal and noise characteristics of the surrounding regions of the image volume. We have modeled the noise characteristics of a micro-CT scanner. Using this model, we can calculate the contrast-to-noise ratio (CNR) for a labeled cell in a water or tissue-like background. Our model allows us to predict the amount of label that would be required in a voxel for visualization at a given X-ray dose and image resolution. Volumetric data were manually edited to simulate the introduction of labeled cells into a known location in the image; this allows us to investigate the challenges of detecting and visualizing a single labeled cell or groups of labeled cells within a complex background. Images of synthesized data will be presented. Our predictions will allow us to deduce the appropriate imaging parameters, including the amount of label needed in a region of interest, prior to exposing an animal to radiation. This a priori knowledge will be invaluable in planning cellular imaging protocols and experiments. Our CNR predictions could also be applied to other imaging modalities, such as micro-MR, micro-US, and PET.
N.L. Ford, None.
Abstract ID: 125
For spatial and quantitative kinetic analysis of protein-protein interactions (PPIs) in living mammalian cells, a method was developed, in which PPI-induced complementation of split Renilla luciferase triggers spontaneous emission of luminescence, with a cell membrane permeable substrate, colenterazine. Consequently, unlike conventional complement enzymes that lead to stable diffusive fluorescent products, this split Renilla luciferase complementation readout is capable of locating the PPIs with emission of bioluminescence only at the sites and time of their occurrence in living cells. We previously developed split luciferase system based on protein splicing system for detecting protein-protein interactions, which provides an integrated magnitude of the interaction to be detected. This system made it possible to detect any PPIs, especially useful for long-lived PPIs. But the spliced luciferase was stable and diffusive, and we cannot specify when and where the PPIs occur. The principle of the present split Renilla luciferase complementation strategy is shown in Figure 1. This split Renilla luciferase complementation readout was shown to work for locating a PPI between tyrosine-phosphorylated peptide (Y941) of IRS-1 and SH2 domain of PI3K among insulin signaling pathways in living Chinese hamster ovary cells overexpressing human insulin receptors (CHO-HIR). It was thereby found that the insulin-stimulated interaction occurred near the plasma membrane in the cytosol.
A. Kaihara, None.
Abstract ID: 126
Development of biological markers to better visualize intracranial gliomas will be important in the diagnosis and treatment of these invasive and deadly tumors. Due to an apparent lack of membrane-associated phospholipases, NM404, a second-generation phospholipid ether analog, has demonstrated striking tumor avidity in 20/20 tumor models in rodents. The agent has never been formally evaluated in a brain tumor model, however. The aim of this study was to examine the tumor avidity of 125I-NM404 in an intracranial rat glioma model. 125I-NM404 was radio-iodinated via isotope exchange with Na125I, solubilized in aqueous 2% Tween-20, and injected (5–20 μCi/200 g rat) into normal, C6 glioma (3–5 mm in diameter) tumor-bearing, and sham-operated rats. On Days 1 (n = 1), 2 (n = 1), and 4 (n = 2), animals were scanned with MRI and subsequently with a modified Bioscan AR2000 radio-TLC scanner and their brain were excised for histological analysis. Normal brain, blood, kidney, liver, spleen, thyroid, and tumor tissue were excised, weighed, and radioactivity counted in a gamma-counter. NM404 radioactivity was minimal in normal brain tissue of control or sham animals. Tumor/brain ratios (% inj dose/g) in C6-bearing rats were 10.5, 12.2, and 6.7 at 24, 48, and 96 hr, respectively. In all cases, glioma uptake corresponded to the tumor location as examined with MRI and histology. These preliminary results suggest that further studies, including dosimetry, are warranted to fully characterize the imaging and therapeutic potential of NM404 in glioma.
Photo (left) of excised rat brain with Glioma (arrow) and fused NM404 Bioscan image (right).
J. Weichert, None.
Abstract ID: 128
The scope of this project is to develop novel 18F-peptides to probe integrin expression in vivo using small-animal imaging. Integrins are cell surface heterodimeric membrane glyocoproteins expressed on most nucleated cells. The best studied integrin to date is αvβ3, the vitronectin receptor expressed on melanoma, glioblastoma, smooth muscle, and osteoclasts, which most interestingly is up-regulated on endothelial cells during angiogenesis. It is anticipated that 18F-radiolabeled peptides could be useful in quantifying integrin expression and play a key role in monitoring responses to therapy. The main objectives of this work are to combine three methodologies to allow high-throughput screening of radiolabeled peptides in vivo. The one-bead one-compound methodology is being used to identify novel peptide sequences associated with integrin expression, a rapid solid-phase approach is being used to radiolabel the peptides with a positron emitter, and micro-PET II is being used to image the localization of the radiolabeled peptides in a mouse model. Peptides currently synthesized are small (5 and 6mers) cyclic peptides containing an RGD sequence. The peptides were assembled on PEG-PS resin attached via an aspartate side chain. Cyclization and radiolabeling was performed on the resin. 4-[18F]Fluorobenzoic acid was conjugated either to the N-α or N-å amino group using the in situ activating agent HATU. Unfortunately, low radiolabeling efficiencies (4–20%) were observed probably due to steric hinderance. Peptides were screened for efficacy and selectivity using both ELISA and cell adhesion assays. Several peptides were identified with IC50 values of 5–20 nM for αvβ3. Cyclo(RGDyK) has been tested in vivo and preliminary images were obtained. In conclusion, preliminary in vivo studies look promising and current screening experiments are in progress to identify the optimum probe to image αvβ3 expression.
MIP image 30 minutes post injection of cyclo [18F] RODyK.
J.L. Sutcliffe-Goulden, None.
Abstract ID: 129
Magnetic relaxation switch (MRS) biosensors are superparamagnetic nanocomposites, which can be synthesized to recognize a variety of molecular targets such as DNA, RNA, and proteins in solution. Molecular target recognition by MRS nanoparticles induces nanoassembly formation and a switch in the T2 relaxation time. Using biotin-CLIO, an MRS avidin biosensor, we show that nanoassembly formation restructures the local field gradients due to the nanoparticles resulting in T2 but not T1 relaxation changes. We then show that the relative avidin concentration is directly proportional to the relaxation time ratio T1/T2. A general method for molecular target detection is then developed using iminobiotin-CLIO (Kd ˜10−9 M at pH 7.4) as a model system of reversible molecular interactions. Dual reporting of nanoparticle concentration by 1/T1 and nanoassembly formation by T1/T2 enabled quantitative detection of neutravidin. This new class of sensing nanoparticles will enable the development of rapid homogeneous assays in microfluidic systems or bioreactors and possibly as in vivo biosensors.
T.J. OLoughlin, None.
Abstract ID: 130
Reverse-engineered targeting agents are being developed to deliver imaging agents to a specific population of cells within the body. These multimeric complexes are designed to contain different ligands, which will bind to a cell with the corresponding complement of receptors. Targeting of an imaging or therapeutic agent via attachment to a multimeric ligand complex would theoretically improve the binding specificity and affinity for the target cells while decreasing interaction with nontarget cells. The specific binding of the ligands is currently being tested in competitive binding assays using an Eu-labeled ligand as the competing agent. The first phase evaluation of homomeric ligand complexes containing multiple residues of low-affinity, truncated melanotropin (MSH) has been completed. By increasing the number of attached ligands from 1 to 2 to 3, the binding of the complex is enhanced though cooperative affinity. The Hill coefficients for the monomers, dimers, and trimers increase from 0.76 to 1.12 to 1.35, respectively. Second-phase studies are underway in which the structure of the ligand complex has been altered with the hypothesis that this will enhance binding affinity to a greater degree. Additionally, studies are currently underway to determine the binding affinity and Hill coefficient of a heteromeric ligand complex containing MSH and enkephalin directed against cells engineered to express defined combinations of complementary receptors. The long-term goal of this project is to reverse engineer agents that are directed against endogenous epitope combinations.
H.L. Handl, None.
Abstract ID: 131
The development of specific molecular reporter probes is limited by compartmental delivery barriers, lack of high affinity in monovalent receptor-probe interactions, and lack of efficient coupling chemistries to generate suitable multivalent systems. The purpose was to synthesize and test a multivalent conjugate designed for specific targeting of E-selectin up-regulated on vascular endothelium in inflammatory diseases. Two linear peptides containing DITWDQLWDLMK (sequence1), and IELLQAR (sequence2), which have been reported to bind to human E-selectin, were synthesized and coupled to a nanoparticle consisting of an iron oxide core coated with cross-linked FITC-labeled aminodextran. Peptides were coupled C- and N-terminally using neutral or negatively charged spacer amino acids by reacting the particle with N-hydroxysuccinimide esters to yield thioether and disulfide linkages, respectively, to the terminal cysteine of the peptide (6–27 peptides/nanoparticle). The nanoparticle conjugates were incubated on stimulated human umbilical vein endothelial cells, using unstimulated cells and scambeled-peptide conjugates as controls. Amounts of bound conjugate were compared using a competitive anti-FITC immunoassay, FACS, and confocal microscopy. Particles containing sequence1 showed no selective binding, however, sequence2 produced 6–11× higher fluorescence for stimulated cells. The peptide containing sequence2 with N-terminal attachment and negatively charged linker arm (S4) showed the highest affinity to E-selectin. During 4-hr incubations with 80–320 fM conjugate, S4-binding was 80–1200× higher compared to neutral linker arm peptide, and 200–10000× higher compared to C-terminally attached peptide. Maximal uptake of S4-conjugate was 112 mmol/cell. Unstimulated cells showed at least 80× less uptake. Stimulated cells after incubation were brightly fluorescent both in the epifluorescent and confocal applications. S4-uptake was time- and concentration-dependent, and saturable at 4°C, suggesting receptor-mediated endocytosis, which was confirmed by the intracellular location of the conjugate in confocal microscopy. Successful receptor-specific imaging of E-selectin has been accomplished in vitro using a multivalent peptide-nanoparticle conjugate which can be used as a combined optical/MR contrast agent.
M.A. Funovics, None.
Abstract ID: 132
We are developing novel probes for imaging receptor expression using in vivo optical techniques. One potential target, epidermal growth factor (EGF) receptor, is up-regulated in many forms of cancer. In light of this, we have synthesized and validated a Renilla Luciferase (RLuc)/EGF fusion protein that should aid in the study of the EGF receptor system in vivo. Using PCR, the pelB leader sequence was attached to the 5′ end of hrluc (Promega), and a sequence encoding the 53 amino acids of mature human EGF was extended onto the 3′ end. The product (hrluc-egf) was cloned into an expression plasmid such that a 6xHis tag was attached to the 3′ end of the fusion. The construct was expressed in Escherichia coli LMG194 cells, released from the periplasm by osmotic shock, purified by nickel affinity chromatography, desalted on a PD-10 column, and stored in 1% human serum albumin. Specificity of the fusion protein was assessed by comparing RLuc-EGF and RLuc binding on A431 human epidermoid carcinoma cells (high EGFR expresser) and NIH 3T3 murine fibroblasts (negative control). After removal of unbound protein and addition of coelenterazine (Renilla substrate), only RLuc-EGF on A431 cells showed luciferase activity. Specificity was further confirmed by a competitive binding experiment on A431 cells, with EGF inhibiting RLuc-EGF binding in a dose-dependent manner. RLuc-EGF and RLuc injected into control mice (three each) showed renal activity, indicating this may be a route of clearance. These results demonstrate production of a functional RLuc-EGF fusion protein. Evaluation of this bifunctional protein for tumor imaging in living animals is in progress. This fusion protein strategy can be extended to other proteins with high-affinity cell surface targets, along with other reporters including fluorescent proteins, and will provide a new class of reporter probes for molecular imaging.
A.M. Loening, None.
Abstract ID: 133
We have previously validated noninvasive imaging of a multi-modality double-fusion reporter vector (Ray et. al., Cancer Res., 2003, 63:1160) carrying a mutant herpes simplex virus thymidine kinase (sr39tk) PET reporter gene and a renilla luciferase (rl) bioluminescence optical reporter gene in living mice. Here we report a triple fusion reporter construct combining a synthetic renilla luciferase (hrl), a monomeric red fluorescence protein (mRFP1) and a truncated version of sr39tk (ttk), which can be imaged in living cells using a fluorescence microscope and in living mice with both an optical cooled charged couple device (CCD) camera (fluorescence and bioluminescence) and microPET. hrl, mRFP1, and ttk genes were PCR amplified and ligated in frame in the pCDNA3.1 vector to generate the cmv-hrl-mrfp-ttk fusion gene. This fusion construct was assessed for HRL, RFP, and TK activity in three different cell lines. 10 × 106 of 293T cells transiently expressing the CMV-hrl–mrfp1–ttk fusion, CMV-ttk, CMV-hrl and CMV-mRFP1 vectors were implanted subcutaneously at four sites of four nude mice and imaged next day using a CCD camera both for fluorescence and bioluminescence and by micro-PET. A lentiviral vector carrying the triple fusion gene was constructed and stably expressing 293T cells infected with lenti-virus were isolated by FACS sorting. The hrl-mrfp1–ttk fusion construct shows fluorescence and very good ttk and moderate hrl activities in N2a, 293T, and A375M cells. Fluorescence microscopy reveals predominant cytoplasmic and nuclear localization of this triple fusion reporter protein. Implanted cells expressing the CMV-hrl–mrfp1–ttk fusion can be imaged by three different modalities in a single living mouse. This hrl-mrfp1–ttk triple fusion gene can be used to image reporter gene expression noninvasively and repeatedly from living cells to living mice.
P. Ray, None.
Abstract ID: 134
Guanylate cyclase-C (GC-C) is a cell-surface receptor expressed on the apical surface of normal intestinal epithelial cells and colorectal cancer cells. This G-protein coupled receptor specifically binds the peptide hormones guanylin, uroguanyln, and their molecular mimic, the human Escherichia coli STh(1–19) peptide. Radioligand binding assays with human prostate cancer PC-3 cells and human pancreatic CFPAC-1 cells showed that STh(1–19) and In-DOTA-Phe-STh(1–19) compete with 125I-STh(1–19) for sites and exhibit Kd values between 1 and 10 nM. Covalent cross-linking of 125I-STh(1–19) to the STh binding moiety reveals that this 125I-STh(1–19)-“receptor-like” binding moiety complex migrates as a broad band of approximately 125 kDa on SDS-PAGE. In contrast, the 125I-STh(1–19) cross-linked to the GC-C receptor migrates with a broad band of approximately 160 kDa. In addition, guanylin, uroguanylin, or the truncated STh(6–18) was not able to inhibit binding of 125I-STh(1–19). Guanylin, uroguanylin, and STh(6–18) peptides inhibit 125I-STh(1–19) binding to GC-C receptors on human colon cancer T-84 cells and agonistically stimulate cGMP production in these cells. Even though In-DOTA-Phe-STh(1–19) binds to this “receptor-type” moiety with high affinity on the CFPAC-1 and PC-3 cells, stimulation of cGMP in these cells was not observed, whereas NO or the atropeptins, CNP and ANP, markedly increased cGMP. These studies show that 125I-STh(1–19) and 111In-DOTA-Phe19STh(1–19) bind selectively with high affinity to a 125-kDa “receptor-like” moiety on the surface of the CFPAC-1 and PC-3 cells. Furthermore, this novel binding moiety is not the GC-C receptor for guanylin and uroguanylin. These results indicate that radiolabeled conjugates of Phe19-STh(1–19) can be used for in vivo targeting of this “receptor-like” entity expressed on pancreatic and prostate tumors.
W.A. Volkert, None.
Abstract ID: 135
Membrane potential dependent enhanced uptake of phosphonium salts, including [3H]tetraphenylphosphonium (TPP), in tumor cells, suggests the potential use of phosphonium salts as tracers for tumor imaging. In this study, we characterize the tumor specificity of [3H]TPP in vitro and in vivo, and compare it to FDG. [3H]TPP and [3H]FDG accumulation was compared in cell culture with a variety of cell lines in different glucose concentrations. Normal biodistribution and tumor uptake were assessed using nude mice with or without subcutaneous xenograft tumors (HeLa). To compare the accumulation of TPP and FDG in metastatic tumor, scid mice were tail vein injected with human melanoma cell lines (A375-fl). To characterize the accumulation of TPP and FDG in inflammation, an inflammatory reaction was induced by subcutaneous injection of Complete Freund's Adjuvant in the left hind paw of Sprague-Dawley rats. Mitochondrial membrane potential data from a separate study and the current TPP uptake data showed good correlation. [3H]TPP accumulation was significantly greater than [3H]FDG for glucose ≥100 mg/dl. Normal biodistribution of [3H]TPP showed low uptake in normal tissues, but high accumulation in the heart and kidneys. In vivo subcutaneous tumor showed highest accumulation of [3H]TPP at a size of 5 mm and accumulation decreased as the tumor grew. Interestingly, [3H]TPP accumulation in metastatic tumors was higher than that of [18F]FDG in lung and liver. [3H]TPP accumulation in inflammatory tissue was significantly lower than that of [18F]FDG TPP accumulation is greater than that of FDG in cell culture and in metastatic melanoma models in living mice. The sensitive tumor accumulation of TPP with less propensity for inflammatory regions warrants further use of radiolabeled phosphonium analogs for tumor imaging in living subjects.
J. Min, None.
Abstract ID: 136
Growing use of PET instrumentation in preclinical and clinical applications has generated interest in expanding the selection of positron-emitting radioisotopes beyond 18F. Especially promising are positron emitters that can be classified as “crossover agents,” meaning they are PET isotopes of elements that also have single photon emitting isotopes for which radiopharmaceutical research data have been accumulated. Examples of such elements and their respective (PET, SPECT) isotopes are: iodine (124I, 123, 125, 131I), copper (60,62,64Cu, 67Cu), and gallium (66,68Ga,67Ga). Current PET imaging devices produce superior spatial resolution, detection efficiency, and quantitative analysis compared with SPECT hardware. The present work deals with the other relevant issues that impact SPECT and PET image quality in crossover agents. A list of crossover agents has been compiled, and relevant data on isotope production, availability, half-life, and positron energy are tabulated. The method of production (i.e., the projectile particle, energy, and target in the cyclotron) determines the type of trace contamination carried by the radioisotope. Radioactive contaminants associated with the various meansofproduction have influenceonimage quality through dead time and random coincidences in PET. Detected photon energy spectra for known crossover agents are presented. High-energy photons in the decay scheme can reduce contrast through collimator septal penetration and spectral spill-down in SPECT. The specific radiopharmaceuticals that employ crossover agents are examined for optimum imaging time postinjection, biological and physical half-lives, injected radioactivity, and radiation dose to the patient. The clinical applications and specific imaging tasks are addressed. For pharmaceuticals having imaging tasks requiring fine spatial resolution, positron emission with relatively high positron energy (and hence positron range) have spatial resolution limitations imposed by the positron range in tissue. Crossover agents can also be employed as dualpurpose imaging/therapy radiopharmaceuticals, and the chemical elements and their associated radiopharmaceuticals currently under investigation are presented.
D.J. Wagenaar, Siemens Medical Solutions USA, Inc. 5.
Abstract ID: 137
A genetically engineered anti-carcinoembryonic antigen (CEA) antibody fragment (diabody, VL-linker-VH) was previously constructed from the murine anti-CEA T84.66 antibody. Tumor targeting, imaging, and biodistribution studies of radio-iodinated or radiometal-labeled anti-CEA diabody demonstrated rapid tumor uptake, fast clearance from the circulation, and favorable properties for use as an imaging agent, when evaluated in nude mice bearing CEA-expressing LS174T xenografts. In the present work, a cysteine residue has been introduced into the anti-CEA diabody at the C-terminal preceded by two glycines (GGC), in order to provide thiol groups for site-specific conjugation and radiolabeling. The Cys-diabody has been expressed, purified, and characterized. It was shown to exist exclusively as a disulfide-bonded dimer. When radio-iodinated, it retained high binding to CEA and demonstrated tumor targeting and biodistribution properties identical to the noncovalent diabody, reaching 10.02% ID/g in tumor at 4 hr postinjection. Furthermore, following reduction of the disulfide bond, the Cys-diabody could be chemically modified using a thiol-specific bifunctional chelating agent, resulting in attachment of DOTA for radiometal labeling. Micro-PET imaging studies using anti-CEA DOTA-Cys-diabodies radiolabeled with Cu-64 demonstrated specific targeting to CEA-positive xenografts in mice at 4 and 18 hr postinjection. The Cys-diabody has been further modified by the introduction of a threonine at the N-terminus and a methionine in the linker between the variable domains to provide reactive amino- and thioether groups, respectively. The ThrMetCys-diabody has been expressed, and has been shown to bind to CEA by ELISA. This approach will allow modification using up to three orthogonal conjugation chemistries, enabling the addition of multiple functions to this tumor targeting protein. The introduction of reactive residues in a diabody for site-specific labeling allows control over the location and stoichiometry of the conjugation process and should enhance the uniformity and quality of antibody fragments as radiopharmaceuticals.
T. Olafsen, None.
Abstract ID: 138
Microarray analysis can identify molecular targets for the development of new molecular imaging agents. We chose to test this hypothesis in a mouse model for human head and neck tumors (squamous cell carcinoma, SCCVII). Genomic analysis was performed by sampling tissues based on the contrast-enhanced MRI imaging patterns. Contrast-enhanced (CE) and nonenhanced (NE) regions of this tumor model were sampled and genomic analysis was performed. Platelet derived growth factor receptor alpha was found to be up-regulated in the CE regions of the tumor as compared with the NE areas. Radiolabeled antibody (I-125) against PDGF-R alpha was used to image the distribution of this receptor within the tumor. A blocking experiment using unlabeled antibody was performed to confirm the specificity of the target. We observed a correlation between the contrast-enhanced MR image of these tumors and the gamma image of the PDGF-R. Autoradiography was used to confirm these correlations. In conclusion, molecular imaging using gamma scintilligraphy and radiolabeled antibodies is a rapid, powerful tool to confirm molecular targets identified by genomic analysis in vivo.
S. Guccione, None.
Abstract ID: 139
T.A. Beresten, None.
Abstract ID: 140
T.A. Beresten, None.
Abstract ID: 141
Pretargeting—specifically placing artificial receptors on target cells in vivo and capturing small probe molecules at these sites—can lead to highly specific images with little background. Pioneering work in this area has involved the strong avidin-biotin binding pair and weaker antibody-hapten binding pairs; our objectives are to develop a humanized system with properties superior to both. We have solved the crystal structure of the rare-earth-DOTA binding antibody 2D12.5 and have used it to identify preferred residues on the protein for engineering mutants capable of forming a permanent bond to a captured rare-earth-DOTA ligand. We have designed and expressed a collection of chimeric antibody fragments incorporating site mutations and tested them for reactivity with complementary rare-earth-DOTA molecules. The relative reactivity of each mutant depends on accessibility to the ligand molecule in the binding site. Details of the molecular engineering and irreversible reactivity of each antibody mutant will be described. The class of metallic elements that can bind to this antibody is rich in useful physical properties. For example, the nuclear emissions of yttrium, lutetium, and several of the other rare earths are suitable for radioimaging and/or therapy. The luminescence properties of terbium and europium feature long excited-state lifetimes and sharp emission lines, and the paramagnetism of gadolinium and other lanthanides has led to contrast agents for magnetic resonance imaging. With its ability to bind all of the rare earths with high affinity, this system represents a molecular docking station with broad applications for imaging and therapy.
T.M. Corneillie, None.
Abstract ID: 142
Somatostatin analogues such as octreotide are useful in targeting radiolabels to somatostatin receptor-positive tumors, for imaging and therapy. However, the effectiveness, particularly in therapeutic applications, is limited by the balance between target and kidney uptake of these molecules. This limitation might be improved either by facilitating kidney clearance or by enhancing tumor retention of radiolabels delivered by these molecules. Tumor retention of radioactivity is limited by receptor recycling mechanisms that involve vesicular packaging of receptor-ligand complexes and result in radioligand externalization. Taking advantage of the abundance of vesicular encapsulated proteases, intracellular cleavage and deposition of the radiochelate might be effected; since radiochelates do not clear subcellular compartments well, this frustration of the receptor-ligand recycling pathways should enhance tumor retention. We have synthetically incorporated proteolytically labile peptide sequences into Tyr3-octreotate conjugates. Enzymatic cleavage was evaluated in vitro using coumarin-conjugated peptides incubated with cathepsins and evaluated by TLC analysis. The results indicate that the metabolizable peptides cleave at rates comparable to that of the positive control CBZ-Arg-Arg-aminomethylcoumarin, while the coumarinyl-octreotate control remained intact. We have synthesized DOTA-Tyr3-octreotate conjugates incorporating the selected peptide sequences and have examined uptake and retention of the radiolabel in receptor-positive AR42J cells in culture. Compared to the DOTA-Tyr3-octreotate control, the cleavable peptides exhibited similar uptake kinetics but nearly twofold better retention in AR42J cells; thin layer chromatographic analysis of cell lysates confirmed intracellular cleavage. The insertion of these cleavable sequences into radiolabeled DOTA-Tyr3-octreotate may be an effective means to maintain radioactivity in tumor cells. Radioactivity biodistribution of Cu-64-labeled DOTA-peptides in tumor-bearing mice is being evaluated by micro-PET.
P.A. Whetstone, None.
Abstract ID: 143
The development of receptor-targeted contrast agents for MRI using derivatized IgG class antibody (Ab) technology has been a longstanding, but largely unrealized goal. Limitations of existing IgG-based approaches include steric hindrance in accessing the antigens, nonspecific binding, low sensitivity of relaxation agents, limited number of receptors available, random incorporation of chelates, and reduced affinity of the derivatized Ab toward the targeted ligand or receptor. Recently, we pioneered a technology of recombinant chimeric proteins consisting of variable fragment (Fv) antibodies as targeting domains and metal coordination sites as reporting domains [Malecki et al., 2002, PNAS 99: 213–219]. The objective of this work is development of receptor-targeted MRI contrast agents using bioengineered recombinant antibodies with Gd coordination sites as relaxation enhancers. The antibodies against transferrin receptor (TfR) were bioengineered using two complementary routes: (1) natural immunization, affinity purification, and generation of fragments followed by conjugation of chelates and (2) preparation of the libraries with expression in yeast Pichia pastoris and bacteria Escherichia coli, followed by selection through biopanning and affinity purification. The antibodies were derivatized with Gd ions. Affinity testing was performed using Western blotting, surface plasmon resonance, ELISA, and immunofluorescence. Incorporation of Gd was verified using energy-dispersive X-ray spectral imaging. The antibodies retained affinity towards TfR even at high levels of derivatization. Relaxivity of water protons was ˜200 mM−1 s−1 at 9.4 T. The high relaxivity resulted in MRI contrast changes at antibody concentrations as little as 0.1 uM, which is sufficient for imaging of receptors and ligands in vivo. The Fv antibodies have much smaller size (29 kDa) than natural IgG (155 kDa), which improves their hydrodynamic diameter and volumetric density. Absence of the Fc fragment eliminates nonspecific binding. The high level of derivatization results in enhancement of relaxivity, while maintaining high affinity and specificity towards the target.
M. Malecki, None.
Abstract ID: 144
HSV1-tk reporter gene and a radiolabeled nucleoside, as a reporter probe, is the best developed system for the nuclear molecular gene imaging. Several nucleoside analogues were tested as reporter probes for HSV1-tk, a well-established reporter gene, undergoing clinical trials. The best combination of signal-to-background ratio and total accumulation was shown for five substituted 2′-fluoro-1-β-
M. Doubrovin, None.
Abstract ID: 145
The overexpression of peptide receptors has been utilized for several years in the diagnosis of neuroendocrine tumors. Targeting tumors by peptide conjugates is one of the possible ways to achieve high specificity in molecular imaging. We aim at modifying somatostatin receptor agonists by adding sequence tags which lead to different intracellular trafficking of these conjugates. Receptor binding was studied by inhibition assays. Internalization and subcellular localization of cyanin dye conjugates were examined in permanent cell lines, in cells expressing a receptor-EGFP fusion protein and in human primary neuroendocrine cells by confocal laser microscopy. Binding, activation, and internalization studies were conducted by laser microscopy or biochemical assays. Fluorescence of native somatostatin dye conjugate was rapidly eliminated from the cultured cells. Experiments using several well-known analogues (e.g., octreotide, octreotate) showed that internalization of ligand and receptor can differ greatly depending on the receptor subtype as well as the structure of the ligand. The addition of sequence tags targeting various intracellular compartments resulted in a variety of different responses. Most of these structure-function relationships of these ligands with regard to receptor activation and endocytosis will be discussed. Improving peptide ligands for tumor targeting can be done by rational design or combinatorial approaches. The effect of amino acid substitutions on the properties of the resulting analogues is still not predictable.
C. Grotzinger, None.
Poster Session 02: Advances in Imaging Instrumentation, Informatics, and Reporters
Abstract ID: 146
We have developed imaging technology based on the stimulation of thermoacoustic emission using visible, infrared, radio waves, or microwaves. These thermoacoustic waves are created whenever such electromagnetic energy is absorbed in biologic tissue. We call this imaging technology “thermoacoustic computed tomography (TCT).” The images we have produced are 3-D in nature, and offer improvements in spatial resolution and sensitivity, when compared to competing optical imaging modalities, such as diffusion optical tomography or fluorescence imaging. The purpose of this presentation is to introduce the molecular imaging community to this imaging technology, to review the underlying imaging principles, and to present initial applications of TCT to optical imaging in mice. We will discuss the conversion of absorbed optical energy into an acoustic wave via thermal heating and subsequent thermal expansion of tissue. We will then develop a theoretical basis for image formation based on an approximation to the Radon transform. Lastly, we will describe a prototype instrument for mapping the optical absorption of near-infrared light in three dimensions in mice, and report the results of initial imaging experiments. We have fabricated and tested a prototype TCT scanner for imaging mice, which uses a tunable, pulsed, near-infrared laser to stimulate thermoacoustic waves, and an array of 128 ultrasound transducers to detect them. Using this device, we have achieved isotropic spatial resolution of 0.35 mm in tissue-mimicking, optical-scattering phantoms. We have visualized the vasculature, and internal organs of athymic mice with submillimeter resolution and excellent signal to noise. We have also demonstrated that it is possible to perform volume-localized, in vivo optical spectroscopy using this device. Based on our initial experience, we believe that TCT can image optical absorption in small animals with submillimeter spatial resolution, and can be used to map the distribution of dye-labeled molecular probes with excellent spatial resolution.
R.A. Kruger, OptoSonics, Inc. 4, 5.
Abstract ID: 147
Labeling cells with Gd-based contrast agents provides a means to assess, in vivo, the migration potential of implanted cells with MR imaging. These type of studies require high spatial resolution, on the order of 50 μm, and high contrast of the implanted cells against the host tissue. The high spatial resolution is attainable by imaging at magnetic field of high strength. However, the relaxivity of Gd-based contrast agents decreases as the field strength increases above 3 T, leading to a decrease of contrast enhancement efficiency. The concentration of Gd contrast agents necessary to compensate for this drop in relaxivity may be in a toxic range for the cell (K. Bahk, PhD Thesis, 2002). We have developed a unique methodology (H. Zhang et al., ISMRM 7: 340, 1999), paramagnetic relaxation enhancement in off-resonance rotating frame, to overcome this relaxivity loss without increasing the contrast agent's concentration. This method modifies the field dependence of a contrast agent's relaxivity and maintains the relaxation efficiency at low contrast agent concentration. We report here the relaxation behavior of Gd-DTPA labeled C6/LacZ glioma cells in off-resonance rotating frame as a function of Gd-DTPA concentration and off-resonance pulse parameters. Glioma cells were labeled with high efficiency by electroporation technique and their intracellular Gd concentration was quantified with ICP-AES. MR experiments were carried out at 9.4 T for three different intracellular concentration of Gd-DTPA. We show that the paramagnetic relaxation enhancement in the rotating frame provides sufficient contrast even at the low concentration, which will permit detection of cell migration after subsequent cell divisions.
H. Zhang, None.
Abstract ID: 148
We report on the working principle, characteristics, and applications of an ultra-high rep. rate (up to 110 MHz), picosecond-gated (down to 50 ps), gain modulated (up to 1 GHz), intensified CCD camera system. This camera is ideally suited for time-resolved optical imaging in conjunction with high rep. rate pulsed lasers. We will describe the use of the camera in various applications such as imaging through scattering media, fluorescence lifetime imaging microscopy (FLIM), dynamic FRET, time-gated Raman imaging and spectroscopy, and time-resolved single molecule and quantum dot imaging and microscopy. Further developments of the intensifier technology using quadrant photocathodes will also be discussed.
R. Ahuja, None.
Abstract ID: 149
This paper presents a design study for a small animal SPECT imager that could provide a resolution of a few hundred micorns in vivo for imaging with I-125 radiotracers. It is based on the use of a very high resolution scintillation camera using the newly developed electron bombarded CCD (EBCCD), as shown in Figure 1. The operating characteristics of the device can be found elsewhere [1]. When coupled to a columnar grown CsI array or a thin and continuous scintillation crystal, it is capable of providing an intrinsic spatial resolution of <100 μm. In order to make use of the high resolution, we propose to use a multiple pinhole/coded aperture as the collimator because the conventional single pinhole becomes extremely inefficient due to the small pinhole size. To evaluate the performance of the SPECT camera with various multiple pinhole configurations, we adapted an analytical approach for calculating performance indices such as resolution-variance tradeoff, pixelwise signal-to-noise ratio (SNR), and the variance on integrated region-of-interest (ROI). This approach is based on recent development in theoretical understanding of the maximum a posteriori (MAP) reconstruction algorithms. It enables one to perform a multivariate detector optimization with reasonable accuracy without doing computationally expensive Monte Carlo simulations. The results from this study will be presented. The performance of the EBCCD-based camera will also be compared with a position sensitive photo multiplier tube (PSPMT)-based SPECT camera
Schematic of the EBCCD tube
Abstract ID: 150
We have developed imaging technology based on the stimulation of thermoacoustic waves by the absorption of near-infrared light in biologic tissues. This imaging technology, thermoacoustic computed tomography (TCT), is described in a companion presentation. Especially exciting is the potential of using TCT to localize dye-labeled, molecular probes, which target specific diseases, such as cancer, in vivo. When optical dyes absorb light, electrons are raised to excited molecular states. These electrons rapidly dissipate this energy in two predominant ways: “nonradiative” collisions with other molecules, producing heat, and “radiative” emission back to lower energy states (fluorescence). The fraction of the energy is dissipated as heat produces thermoacoustic waves. For near-infrared dyes, where the quantum yield of fluorescence is often low (10–15%), 85–90% of the absorbed energy is available for thermoacoustic emissions. Another common feature of optical dyes is that their optical absorption falls precipitously at wavelengths slightly longer than that of peak absorption. This suggests that a viable strategy for detecting the presence of a dye is to make “differential” thermoacoustic measurements at the wavelength of peak absorption and at a slightly longer wavelength. We have used this approach to detect an infrared dye (ICG) in a set of imaging experiments. We imaged dye concentrations from 5 to 640 nM in 1.14-mm diameter plastic tubes embedded in a tissue-mimicking optical scattering phantom. The response of our TCT imaging system was linear over this entire range of concentrations. The lowest concentration we could visually detect was 5 nM. Our phantom studies suggest that we can detect ICG, an infrared dye with an extinction coefficient of 250,000 M−1 cm−1 at 783 nm, at a concentration of 5 nM within a 1-pl volume [(1.14 mm diameter tube) × (1.1 mm thick slice)]. This corresponds to a limiting sensitivity of 5 fmol.
R.A. Kruger, OptoSonics, Inc. 4, 5.
Abstract ID: 151
Bioluminescence images are typically integrated for minutes rather than the 33 msec in typical video imaging. As a result, cosmic rays are often detected that can interfere with measurements of signal size as well as background estimation. We are investigating computer algorithms for removing these very high valued, random events. Our experiments include a rat nonviral gene delivery system for the skin giving little light scatter and a peaked response, and an in vitro cell culture experiment with very low signal levels. These experiments particularly stress algorithms for cosmic ray removal where the goal is to remove cosmic events while leaving signal processing untouched. We investigated multiple image processing algorithms including two morphological image processing schemes, an out-of-range smoothing algorithm, and adaptive median filtering with three different techniques for pixel substitution. Evaluation was done using synthetic images created by adding cosmic ray images obtained by integrating several minutes without the presence of signal-to-signal images having no obvious cosmic events. We purposely included difficult cases where a cosmic ray was superimposed on a signal of interest. Evaluation consisted of counting cosmic ray removed and of subtracting the corrected image from the signal image to determine any potential loss of signal. The adaptive median filtering algorithm was judged best. It was fast, did not require parameter adjustment for different signal shapes, removed all cosmic rays, and proved effective for cases which had cosmic events overlapped with the desired signal. Importantly, integrated signals from our rat nonviral gene delivery experiments were changed by only 0.17 ± 0.01%. Other less desirable algorithms degraded the integrated signal much more.
J.J. Derakhshan, None.
Abstract ID: 152
Advances in the development of coherent terahertz sources and detectors have allowed this region of the spectrum to be accessed with good signal-to-noise properties. This has allowed corresponding advances in both recognition of substances through absorption, which gives a molecular signature, and reflection with its active imaging potential. We have been investigating the applications of terahertz imaging and spectroscopy in a variety of situations having the common feature of a change of molecular signature. We illustrate how different polymorphs of common drugs can be distinguished by their spectroscopic response in both powder and tablet form. The nature and distribution of excipients and incipients in tablets are both accessible to noninvasive analysis and results will be presented illustrating how these factors can be determined. Comparison of terahertz results with other modalities of both imaging and spectroscopy will be presented, for example, use of Raman transitions. In a similar manner, the applications of the technology to identification of hazardous materials presenting a security problem will be presented. It is shown that in the terahertz regime, a distinction is made possible by the unique features of the absorption spectrum. A change of material properties reflects the terahertz waves and the phase coherence allows 3-D imaging to be constructed with depth information. We use the skin as a model system and present further results on investigations of basal cell carcinoma, both in vivo and ex vivo, and demonstrate that the change in absorption properties allows identification of the cancerous region and its distribution as a function of depth. Other tissue type results will be shown and the potential of such imaging and spectroscopy for detection and noninvasive analysis of epithelial carcinomas will be discussed.
M. Pepper, TeraView 5.
Abstract ID: 153
Current management of visceral cancers relies on the accurate identification of metastatic disease. Since many tumors spread by way of lymphatics, assessment of the draining lymph node nearest the site of a primary malignancy has both prognostic and therapeutic implications. At present, preoperative lymphatic mapping of visceral cancers is limited given their inaccessibility. Our goal was to provide the oncologic surgeon with an interactive intraoperative imaging tool utilizing optical, noninvasive technology. Such a system would allow instantaneous imaging of lymphatic flow and draining visceral lymph nodes as well as provide real-time visual feedback for image-guided localization and dissection. We have previously proved the principle of using NIR light for intraoperative imaging in small and large animal model systems. Our system simultaneously displays color video and NIR fluorescence images of the surgical field. The introduction of type-II quantum dots (QDs) into solid visceral organs allows visualization of draining lymphatic channels and nodes. In porcine model systems, we demonstrate that 200 pmol of 860 nm NIR QDs injected into small bowel and lung parenchyma quickly and accurately maps lymphatic drainage and the primary draining lymph node. Injection into the mesenteric side of the mid-ileum results in fluorescence of a single lymph node at the root of the small bowel mesentery. Injection of QDs into the hilar region of the left upper lobe of the lung resulted in identification of a hilar intraparenchymal node, which was excised minimally invasively. Histologic analysis in all cases confirmed the presence of nodal tissue. We report the use of a noninvasive intraoperative NIR fluorescence imaging system for mapping visceral lymphatic drainage and guiding excision of the primary draining node. Such a system is useful for solid organs not amenable to preoperative intracavitary staging and will likely be valuable in laparoscopic and thoracoscopic staging procedures.
J.V. Frangioni, None.
Abstract ID: 154
GFP-labeled in vivo tumor models and GFP-based transgenic systems have proven to be powerful research tools, revealing details of tumor growth, angiogenesis, and metastatic spread, for example. However, achieving useful images can be challenging if the label is buried in deep structures (because of scattering and absorption processes) or if bright autofluorescence is present. Spectral imaging generates a series of images at multiple wavelengths, associating an optical spectrum with every pixel. We investigated whether this approach, using liquid crystal tunable filters and mathematical unmixing algorithms, could usefully be applied to in vivo imaging. Visible-range LCTFs typically have a 10- to 30-nm-wide window that can be electronically tuned, with 1-nm precision, to lie anywhere in the range 420 to 720 nm. Such devices have been shown to be useful for multicolor FISH, and for resolving multiple fluorescent reagents with overlapping emission spectra. In addition, they have permitted the elimination of interfering autofluorescence in microscope-based applications. Using LCTF-based imaging, we examined mouse and zebrafish model systems and found that spectral imaging could enhance the detection and quantitation of GFP-labeled tissues. Standard GFP imaging equipment was modified by replacing the usual color camera with a cooled monochrome camera and an LCTF positioned in front of a conventional macro lens. A series of images was taken every 10 nm from 500 to 650 nm. Using predefined or experiment-specific GFP and autofluorescence spectra, and appropriate algorithms, the original image could be unmixed to yield a GFP signal uncontaminated by autofluorescence. In one model system, this approach clearly revealed GFP signals emanating through skin from tumors located in the mouse lung. Similar results were also achieved imaging zebrafish embryos and adult fish expressing transgenic GFP-fusion proteins in various locations. Thus, we have demonstrated the general utility of spectral imaging for in vivo fluorescence-based studies.
R. Levenson, CRI, Inc. 5.
Abstract ID: 155
Variants of Tetrahymena ribozyme have been shown to perform self-splicing and trans-splicing in mammalian cells, but the low splicing efficiency of these ribozymes and the low production of final proteins largely impair their promise for RNA repair of mutated genes in genetic disorders. An in vivo directed ribozyme evolution would facilitate the discovery of new active ribozyme variants, but its implementation requires a functional assay that works in intact mammalian cells and is compatible with high-throughput screening. Splicing-dependent reporter gene assays have been exploited to report ribozyme activity previously, including firefly luciferase and β-galactosidase, but they are either unable to be performed in intact living cells or lack sensitivity for single-cell resolution. Here we report a design of a reporter ribozyme construct in which the self-splicing Tetrahymena thermophila group I intron ribozyme is inserted into the open reading frame of the mRNA of a bacterial enzyme, TEM-1 β-lactamase. The splicing activity can be readily detected in cell lysates quantitatively and sensitively using our newly prepared fluorogenic substrate CC1 and visualized in single living cells using a ratiometric fluorescent substrate CCF2/AM. This reporter system was applied to screen variants of the Tetrahymena ribozyme for improved efficiency, and led to identification of variants with up to 4-fold more final activity than the native sequence. The compatibility of this system with high-throughput screening with flow cytometry should permit a direct and facile selection of desired ribozyme variants in mammalian cells.
A fluorescent image of Cos-1 cells transfected with a ribozyme reporter construct. Blue cells indicated the occurrence of self-splicing, and green cells were non-transfected and thus had no splicing activity.
J. Rao, None
Abstract ID: 156
MR imaging has a high potential for application in molecular imaging. Contrast agents are the key component to detect diseases in an early stage. We focused on contrast agents based on iron oxides. These agents significantly change the T2* relaxation time. Therefore, the distribution of these iron oxides is usually determined by measuring T2*-weighted images. We present a novel method to perform a fast T2* relaxation time mapping with the future aim of quantifying the concentration of the agent. Furthermore, this imaging technique allows the simultaneous mapping of T2* values and the imaging of morphology without lengthening the measurement time. The method is based on a radial multi-echo readout: After slice excitation, a number of echoes are measured under different projection angles. A high-resolution image is reconstructed from all gradient echoes, while a series of low-resolution images is reconstructed from the separate echoes of increasing echo times. From these low-resolution images, the T2* decay is determined by fitting an exponential decay to each pixel versus the echo time. The accuracy and speed of different methods for T2* calculation were compared with each other (direct calculation vs. fitting routines). First investigations were performed on a clinical 1.5-T system by using phantoms with known concentrations of Resovist (USPIO, Schering). Figure 1 shows the correlation between the measured T2* and the concentration. Hence, the method is well suited for a rapid quantification of contrast agents. The method can furthermore be applied for the near real-time tracking of targeted contrast agents and labeled stem cells, in particular, due to the high speed of data acquisition and a rapid T2* calculation.
H. Dahnke, Philips Research 5.
Abstract ID: 157
Increasing numbers of new optical probes for molecular imaging require imaging tools which allow fast probe evaluation. The aim of this study was to test a new fluorescence reflectance imaging (FRI) system covering the visible to the near-infrared (NIR) spectrum. The system consists of a high-pressure mercury-vapour lamp, which delivers photons in the blue and green light range. The excitation in the NIR spectrum is implemented by LEDs. Four different emission filters allow for spectral separation of emitted photons. Images are acquired by a CCD camera and processed on a PC-based system. System parameters such as sensitivity, depth penetration, channel selectivity, and spectral separation were tested in vitro and in vivo using different fluorochromes and fluorescent proteins. The current set up of emission filters shows complete spectral separation in green, red, and Cy7 NIR channel and a minor cross-excitation (30%) between the Cy5.5 and Cy7 NIR channels. Dilution series showed a detection threshold for tested fluorochromes in picomolar range. As little as 1000 cells expressing GFP could be detected in phantom studies. In vivo experiments showed sensitive differentiation between a Cy5.5 labeled molecular contrast agent and tumoral GFP expression within the same animal. The new multichannel FR imaging system allows rapid and sensitive imaging of most commonly used fluorochromes or fluorescent proteins at four different wavelengths. Rapid data acquisition and versatile wavelength handling are ideal for in vivo and ex vivo applications to test novel optical contrast media.
A. Wall, University of Muenster 5; Institut of Laser and Medical Technology, Berlin 6; Siemens Medical Solutions 6.
Abstract ID: 158
A number of different fluorescent probes targeted to and/or activated by specific molecular events have been developed to detect and characterize disease processes in vivo. We report on the development of methods for quantitative 3-D imaging of fluorescence in living subjects, in order to more fully realize the potential of fluorescence molecular imaging. A parallel-plate imaging chamber has been developed that allows the acquisition of fluorescence reflectance images (FRI), as well as transmittance images from multiple illumination positions that can be reconstructed into a quantitative calibrated map of the concentration of active fluorochrome within the subject (fluorescence-mediated tomography, FMT). Source and detector geometries have been rigorously optimized using singular value analysis to determine the sampling configurations that yield datasets with maximal information content. A 3-D camera has been integrated into the system to allow measurement of the subject surface geometry and improve reconstruction models as well as data visualization. These improvements in FMT hardware and software have pushed its quantitation accuracy to better than 5% error margins and spatial resolution to the submillimeter range. To showcase the quantitation capacity of the method in tissues, we present measurements from mice implanted with known fluorochrome concentrations. We further report on the study of fluorescent molecular reporters in mouse tumor models. FMT offers a unique means of visualizing and measuring biochemical activity in vivo, and is developing as a significant technique in the arsenal of molecular imaging.
E.E. Graves, None.
Abstract ID: 159
The integrin αvβ3 is an adhesion ligand that has been shown to be highly expressed on metastatic tumors and is recognized as a potential binding site for targeted imaging of angiogenesis (the development of new blood vessels required for tumor growth). In this work, the application of targeted contrast agents to enhance ultrasound detection of angiogenic cell lines is demonstrated. Flow cytometry is first used to demonstrate expression of αvβ3 on A375m melanoma and human umbilical vein endothelial cells (HUVEC). Stimulation with either TNF or PMA is used to enhance integrin activation. Cells are grown in a monolayer on relatively nonechogenic plastic membranes, are exposed to targeted or control contrast agents, and are then are washed with PBS to remove nonadherent bubbles. Tested agents include two cyclic-RGD peptide agents, consisting of 1 and 5 wt.% targeting ligand, one agent targeted with the anti-αvβ3 antibody LM609, and two control agents. Specificity of the agents is evaluated through blocking with scrambled peptide, free RGD peptide, and LM609 antibody. Acoustic studies with a simple ultrasound system illustrate a backscatter intensity increase from αvβ3 expressing cells exposed to the targeted contrast agent ranging from 3- to 20-fold depending on ligand type, cell type, and cell stimulation. Additionally, blocking with free antibody or free peptide reduced the signal intensity to near the level of a nontargeted control. The acoustic signal correlates with the optically determined number of bubbles bound per cell for concentrations less than one bubble per cell. Frequency domain analysis demonstrates that adherent targeted bubbles exhibit a mean echo frequency approximately 2 MHz lower than those from free agents (as determined within the −20 dB bandwidth of the detection transducer), which suggests that adherent contrast agents may be differentiable from free-floating microbubbles.
P. Dayton, None.
Abstract ID: 160
We have optimized firefly luciferase protein fragment complementation to produce a robust and broadly applicable bioluminescence imaging method, luciferase complementation imaging (LCI), by screening combined incremental truncation libraries of N- and C-terminal fragments of luciferase fused to the FRB domain of mTOR and FKBP-12, respectively. In intact HEK-293 cells, the optimized FRB-NLuc/CLuc-FKBP LCI pair reconstituted luciferase activity upon single-site binding of rapamycin (Kd = 1.5 nM) in an FK506-competitive manner (Ki = 4.2 nM) (Figure 1), quantitatively reproducing known binding constants for small molecule-induced FRB/FKBP association. In mice bearing implants of cells expressing the FRB-NLuc/CLuc-FKBP LCI pair, repetitive LCI showed dose- and time-dependent luciferase activity induced by rapamycin with a maximal in vivo signal-to-background ratio of 23:1. Furthermore, optimized LCI successfully reported association of other known protein pairs. For example, in cells expressing a Cdc25C-NLuc/CLuc-14-3-3e LCI pair, the protein kinase inhibitor UCN-01 produced decreased bioluminescence corresponding to decreased S216 phosphorylation of the Cdc25C-NLuc fusion. In addition, cells expressing a STAT1-NLuc/CLuc-STAT1 LCI pair reported strong and specific STAT1/STAT1 association, but in this case, independent of interferon-induced protein phosphorylation. Thus, LCI provides a platform for detection of regulated and small molecule-induced protein-protein interactions in intact cells and living animals and should enable a wide range of applications in cell biology, drug discovery, chemical genetics, and proteomics research.
K.E. Luker, None.
Abstract ID: 161
With the increasing use of multimodality imaging for research in the molecular imaging community, a need has arisen for software tools that can seamlessly handle the datasets generated by these studies (e.g., PET/CT, SPECT/MRI). AMIDE (Amide's a Medical Image Data Examiner) has been developed for the purpose of providing a free and open source tool for viewing and analyzing both single and multi-modality volumetric datasets. Central to the package's ability to simultaneously display multiple datasets and regions of interest (ROIs), is the automatic, on-demand data reslicing implemented within the program. No constraints are imposed by the dimensions, anisotropy, or voxel sizes of the data. Instead, the user is divorced from the underlying digital representation of the data and is allowed to shift, rotate, and view multiple datasets at arbitrary angles and directions, with the program directly interpolating from the original data as needed. Additional features include fully 3-D ROIs (ellipses, cylinders, boxes, and isocontours), multislice viewing, and volume rendering. Data importing is done through the (X)MedCon library, which supports DICOM 3.0, ECAT 6.4/7.2, Acr/Nema 2.0, Analyze (SPM), InterFile 3.3, and Concorde formats. Validation has been performed by comparing the output of AMIDE with that of several existing software packages. AMIDE is released under the terms of the GNU General Public License (GPL), with source code and binaries for Linux, Mac OS X, and Windows available at http://amide.sf.net. Within our institute, AMIDE has been used in a variety of research studies, including following metastatic tumor models with PET/CT and testing antibody imaging agents using SPECT/CT, and is continuously refined based on feedback from these and other studies.
A.M. Loening, None.
Abstract ID: 162
Fluorescence proteins (FPs) have become essential reporter molecules for different biomedical applications. Most frequently, engineered FPs are detected by epifluorescence, confocal (microscopy) or reflectance (whole animal) imaging. The latter has been shown to be useful in detecting and following tumors in vivo, particularly those implanted near the surface or in surgically exposed organs. However, reflectance imaging has inherent limitations, as it does not retrieve depth information nor allow absolute quantification of fluorescence activity. This is mostly due to strong nonlinear light attenuation and propagation in biological tissues. In this study, we present a set of technologies, originally developed for fluorescent tomography in the near-infrared region and their adaptation to tomographic imaging of proteins fluorescing in the visible range. The tomographic method employed a multi-angle, multi-projection illumination scheme from an Ar+ laser while detection was achieved using a highly sensitive CCD camera. Reconstruction is based on novel photon propagation models developed specifically for visible light. We show that GFP expression is possible in deep tissues quantitatively. We also show that the detection of GFP expression in a mouse subcutaneous tumor model (using Gli36 human glioma cells) transferred through an HSV/EBV hybrid amplicon vector is feasible. We conclude that tomographic imaging in the visible range is possible and that the technique can be used to image gene expression.
G. Zacharakis, None.
Abstract ID: 163
Misonidazole and its derivatives labeled either with gamma or positron emitters are known to bind selectively to hypoxic cancer cells. These ligands undergo a rather complex chain of biochemical reactions within the cell and are even subject to partial washout after reduction. To describe their kinetic behavior complex, compartmental models were proposed in the literature [1]. Due to the excessive noise content of nuclear images, larger macroscopic regions need to be averaged to obtain stable kinetic parameters. However, regional averaging may mitigate the diagnostic value since hypoxic cell clusters may alternate with normoxic and necrotic cells on small scale. To overcome this limitation we aim at a protocol, which allows a pixel-wise kinetic modeling. Prior to kinetic modeling, we estimate systematic and statistical errors anticipated for each pixel and point in time using a purpose designed simulation environment. It is based on the GEANT 4 Monte Carlo package [2] and comprises tools to define kinetic phantoms and tools to simulate time series of PET images for a specific scanner geometry and its detector properties. This additional knowledge is then utilized to stabilize the pixel individual kinetic modeling process. The strategy is tested against data from patients with non-small cell lung cancer. The arterial input function is extracted noninvasively from the reconstructed image of the left ventricle of the heart and used as reference for the tracer concentration at the tissue of interest. We investigate whether a pixel-wise extraction of kinetic parameters of tumor oxygenation for lung cancer patients is feasible if a simulation-based optimization is used.
Abstract ID: 164
To characterize the trabecular bone structure in three dimensions, micro-CT has emerged as the leading technique. No special sample preparation is needed and objects up to several centimeters can be imaged with a spatial resolution down to 5 μm. While visualizations are very helpful to get a first glimpse of the bone samples, powerful evaluation tools are needed to fully characterize important structural features of cancellous and cortical bone. A complete 3-D evaluation would include the calculation of indices such as bone volume fraction, trabeculae number, thickness and separation, degree of anisotropy, connectivity density, structure model index (rod vs. plate characteristic), bone surface ratios, and so on. Moreover, this evaluation can be made more precise by defining an irregularly shaped volume of interest, enabling the assessment of a complete bone or trabecular or cortical compartment. The new Scanco in vivo micro-CT ‘vivaCT-40’ delivers images of rats and mice with a voxel size of 10 to 40 μm with a slice thickness of also 10 to 40 μm. For a typical measurement site such as the proximal tibia, a set of 255 slices (6 mm in axial length) is obtained in 7 min. In a first study with OVX rats, the same animal was scanned once every week, and changes in bone volume fraction and bone architecture were measured with very high accuracy (reproducibility better than 0.2%). 3-D analysis revealed that bone volume fraction dropped more than 50%, with a concomitant decrease in trabeculae number of 40% and a slight reduction in thickness of 25%. The changes in bone structure can thus be directly followed in the same animal, a breakthrough in drug testing and screening.
A. Laib, None.
Abstract ID: 165
The NIH ATLAS II small animal PET scanner is a high-sensitivity, dual-layer (GSO/LGSO) phoswich PET scanner designed for rodent imaging. ATLAS II, functionally identical to ATLAS I at the NIH, has been in use at Johns Hopkins University in collaborative studies with the NIH since August 2002 and more than 150 rodent imaging studies have been performed to date. Here we report the results of an independent test of ATLAS II in which we evaluated spatial resolution, sensitivity, and count rate performance. Radial and tangential resolutions were measured by Gaussian fitting of the reconstructed image profile of an F-18 line source (ID = 0.5 mm) at the system center and 2 cm off-axis. The system sensitivity along the central axis was measured using an F-18 point source. Count rate performance was measured with two different C-11 filled cylindrical phantoms simulating a typical rat body (ID = 5 cm, length = 15 cm) and mouse body (ID = 3 cm, length = 7.5 cm). Images were acquired with an energy window of 100–700 keV, a range found to be operationally useful in actual imaging studies. Radial and tangential resolution determined from the reconstructed images of the line source with the 3-D OSEM algorithm were 1.36 mm at center and 1.98 mm (radial), 2.13 mm (tangential) at 2 cm off-axis. Sensitivity was >2.0% after correcting for positron escape from the source. Maximal noise equivalent count rate was 10.3 kcps at 52.2 MBq total activity for the rat phantom and 16.0 kcps at 30.4 MBq for the mouse phantom. These characteristics are comparable to those of the prototype scanner. The results of this preliminary field test of the ATLAS scanner are consistent with performance measurements made on the original system and are consistent with the requirements of rodent imaging studies.
J. Lee, None.
Abstract ID: 166
Molecular imaging is an important technology to solve biological and medical questions in the 21th century. This is best realized via in vivo imaging of biological processes in small animals. Thus, a special high-resolution imager dedicated for small animals is required. We recently installed a high-resolution animal positron emission tomography (PET) scanner (micro-PET R4) for doing in vivo molecular imaging of gene expression. This work describes the performance evaluation of our micro-PET R4 scanner. The micro-PET R4 scanner is a dedicated PET for studies of rodents. The system is composed of 96 detector modules, each with an 88 array of 2.12.110 mm3 lutetium oxyorthosilicate (LSO) crystals, arranged as 32 crystal rings and 14.8 cm in diameter. The detector crystals are coupled to a Hamamatsu R5900-C8 position sensitive photomultiplier tube (PS-PMT) via a 10-cm-long optical fiber bundle. The system operates in 3-D mode without interplane septa, acquiring data in list mode. A number of scanner parameters such as sensitivity, spatial resolution, and energy resolution were determined in this work. In the center of field of view (FOV), a maximal sensitivity of 21.04 cps/kBq was calculated from a measurement with a germanium-68 point source with an energy window of 250–750 keV. A spatial resolution of 2.03 mm (FORB + 2D-FBP)/1.61 mm (FORB + 2D-OSEM) full width at half maximum (FWHM) in the tangential direction and 2.07 mm (2D-FBP)/1.65 mm (2D-OSEM) FWHM in the radial direction was measured in the center with a 0.28 mm diameter 18F-FDG line source. The energy resolution of the scanner was measured across all crystals ranging from 13.9% to around 34.6% with a mean of 18.45%. Images of a high-resolution phantom and from rats and mice studies demonstrate good performance of the scanner. The results show that the micro-PET R4 is a suitable PET scanner for small animals.
J. Chen, None.
Abstract ID: 167
H. Alfke, None.
Abstract ID: 168
The image quality attainable in a mouse micro-CT scan is limited by the radiation dose that the mouse can receive without affecting the underlying biology that is being studied. If a micro-CT scanner gives images of high quality but at an unacceptably large radiation dose to the mouse, then it is of little practical use, especially for longitudinal studies, where the same animal would be imaged multiple times. In order to better quantify the doses that a mouse receives in a micro-CT scan, 1 mm3 thermoluminescent dosimeters (TLDs) were placed in a cylindrical ‘mouse-like’ polycarbonate phantom and irradiated using a micro-CT system. The phantom was constructed from a stack of 3 cm diameter, 3-mm-thick polycarbonate disks, with each disk containing a series of holes to allow placement of the TLDs at specific depths in the phantom. Dose measurements were made for tungsten anode X-ray settings of 40 and 50 kVp and for molybdenum anode settings of 30 kVp. To validate phantom measurements, three TLDs were implanted into a mouse ex vivo and irradiated with the 40 kVp tungsten X-ray setting. All irradiations were performed using an 800-view CT scan with a 360° orbit. The dose measured in the phantom as a function of depth is shown in the figure. The results show that both the dose and the dose gradient from center to edge increase for lower energy X-ray beams. The TLDs implanted in the mouse recorded dose levels from 24.8 to 32 cGy for the 800-view scan. The results suggest that for typical 200 view screening scans, the dose to the mouse would be in the range of 6–8 cGy.
A.L. Goertzen, None.
Abstract ID: 170
The Large-Scale Digital Cell Analysis System (LSDCAS) is a collection of automated microscope systems and analysis techniques dedicated to performing live cell measurements under controlled laboratory conditions. The overall philosophy underlying LSDCAS development is to provide the biomedical research community at large with a new set of tools for the analysis of living populations of cells. LSDCAS is capable of imaging thousands of living cells for a period of a few days up to weeks in a single experiment. In addition to its native data formats, digital motion pictures (MPEG movies) are produced and archived to DVD-ROM for later analysis and display. LSDCAS is comprised of over 30,000 lines of programming code in C++. For our current configuration, LSDCAS produces peak data acquisition rates of about 40 gigabytes per week, thus requiring rather significant computer and analytical resources. At this time, LSDCAS consists of two automated microscope systems, a dedicated, highly available Storage Area Network Data Center (2 quad Itanium servers—2 terabytes FC raid storage—30-slot DLT tape library), 15 analysis/acquisition workstations, and tissue culture support apparatus. All software runs on the Linux operating system; DEC alpha, itanium, and 386-based processor hardware is supported. Following the completion of its documentation this summer, the source code will be made available as an Open Source project under the GNU General Public License. LSDCAS is being used in the study of cell motility, measurement of intracellular pro-oxidant species, and the determination of phenotypic changes observed using adenovirus-mediated gene expression systems. Other applications of LSDCAS technology are under development within interdisciplinary collaborative interactions at the University of Iowa. It is expected that the analytical capabilities of LSDCAS will continue to grow in the near future. (Support: 1R33 CA94801 and 3R01 CA74899 [NIH].)
M.A. Mackey, None.
Abstract ID: 171
To understand the role of vascular development in brain growth and function, we need to see morphology at the capillary level and be capable of identifying molecular and cellular processes as they occur. This requires imaging at high levels of resolution, which are best achieved with microscopy. But histological sections provide a small and incomplete view of the elaborate structures. This limits assessment of to small “snapshots” rather than providing a more complete understanding. Therefore, to visualize vascular development, we are building new tools that allow us to build 3-D reconstructions of capillary architecture from 2-D images of histological sections. We have successfully imaged the vasculature of the brain of a mouse and reassembled the sections into a 3-D volume. Specimens are prepared by vascular perfusion of the FITC-conjugated lectin, L. esculentum, followed by fixative. A vibratome is used to obtain sections that range in thickness from 30 to 75 μm. Images are then collected by fluorescence microscopy and montage digital photography. An example image is shown in Figure 1. The deformations that occur from sectioning are large in relationship to the vessel features that we are aligning. Therefore, we have developed a registration method that models the deformations and precisely aligns corresponding features. Vessels are identified automatically and correspondences between endpoints on neighboring layers are identified either manually or automatically. These correspondences are used to compute multi-layer free-form deformations that warp the sections such that features are aligned into a common coordinate system. We have visualized the resulting volume data using consumer graphics hardware. We believe these tools will be useful for other applications such as tumor imaging.
M.L. Gleicher, None.
Abstract ID: 172
Reactive oxygen species are important for many natural cellular processes and also play a role in cell death. Studying the balance between pro-oxidant/anti-oxidant activities in living cells would provide an important tool for understanding mechanisms involving these short-lived species in biological systems. The Large-Scale Digital Cell Analysis System (LSDCAS), a multi-microscope computer-controlled data acquisition and analysis system, was designed to analyze large numbers of living cells under a variety of experimental conditions. We have recently added fluorescent image acquisition hardware and software to LSDCAS, thus extending its capabilities as an analytical tool. As an inaugural project employing this new analytical capability, we have used the oxidant-sensitive fluorescent probe C-400 to measure intracellular pro-oxidant levels on a cell-by-cell basis in populations of living cells. In the development of this technology, initial experiments determined machine and cell culture parameters in order to minimize cell perturbation. To determine the concentration dependence of probe uptake and efflux kinetics, the oxidant-insensitive probe C-369 was used. Quantification of the probe uptake and efflux rates as a function of concentration allowed for the development of models of these transport processes which are required for similar measurements using the oxidant-sensitive probe C-400. Under the control of LSDCAS, balanced salt solution containing the probe was applied to the cell culture in a 35-mm Petri dish contained in a dual-ported perfusion chamber. After a 5-min labeling period, the probe solution was rinsed once and full culture medium was returned to the culture dish. We are currently working to determine the long-term toxicity of these and other fluorescent compounds, and to develop mathematical models for quantitatively relating the rate of pro-oxidant production to the eventual loss of clonogenicity on a cell-by-cell basis. (Support: 1R33 CA94801 [NIH].)
M.A. Mackey, None.
Abstract ID: 173
Iodine-125 (I-125) is used as in biomedical research employing quantitative autoradiography for in vitro studies. It is commercially available bound to nucleic acids, antibodies, and other ligands for molecular biology research. I-125 emits 27–35 keV photons. Though these photons are not energetic enough for human nuclear medicine studies, their energies are sufficient for mouse imaging. We have developed gamma cameras capable of imaging the biodistribution of I-125 for a small animal SPECT system. The system is being developed to acquire SPECT imaging data of I-125 labeled molecules in un-anesthetized mice to avoid affects of anesthesia on neurological and physiological processes during the study. Initial development of this imaging system focuses on brain studies. An infrared-based position tracking apparatus has been built. The tracking apparatus is integrated into a SPECT gantry and faces a transparent tubular “burrow” in which the animal is placed during imaging. Retro-reflective markers glued to the mouse's head are illuminated by infrared LEDs. Compact I-125 imaging detectors based on position-sensitive photomultiplier tubes were designed and built for the SPECT imaging system. The use of parallel hole and multi-pinhole collimation for improved sensitivity with high resolution is being investigated. Tracking has been tested with live mice. Time-correlated list mode data containing the tracking coordinates and the gamma imaging data acquired from the gamma cameras have been obtained of a moving radioactive phantom. Initial phantom studies and modeling studies utilizing multi-pinhole masks have been underway. We have successfully designed and built a novel I-125 imaging system for imaging unrestrained mice. The system is being validated. Tests are planned to acquire time-correlated list mode data of a mouse's head position and the gamma-ray imaging data. Tracer uptake in the animal's head will be reconstructed from the time-correlated tracking and gamma-ray datasets.
A.G. Weisenberger, None.
Abstract ID: 174
Imaging bioluminescent signals from luciferase-tagged cell lines in small animals is an increasingly used research tool. Images obtained in such work are a diffuse projection on the surface of the bioluminescent sources located deeper inside. Extremely accurate measurements of surface light intensity can be used to quantify some properties including relative source growth, but absolute measurements of source size require additional information such as the depth of the source. Depth measurements can also be used to provide information on the location of the cells (e.g. in which organ a metastatic lesion developed). We present here a relatively simple method based on the use of spectroscopic information to estimate the depth of the bioluminescent cells. The method takes advantage of the fact that the bioluminescent signal is emitted from 500 to 700 nm, therefore spanning through a region of the spectrum where there are major contrasts in tissue optical properties. Tissue absorption, mainly due to hemoglobin in that part of the spectrum, drops dramatically from the green/yellow (500–580 nm) to the red wavelengths (600–750 nm). As a result, the bioluminescent signal observed at the surface of the animal is dependent on both the wavelength and the thickness of tissue it travels through. In vivo measurements obtained from mice injected with bioluminescent cells show that our spectral imaging technique using relatively simple analytical expressions and assuming a flat subject surface allows to determine the depth of the signal from two or more images acquired at different wavelengths. The signal can then be absolutely quantified, therefore providing the biologist with information on the absolute number of cells. Further development of this method is currently being pursued, including the use of structured light imaging to derive the 3-D profile of the surface examined. This should lead to a more accurate estimate of signal depth.
O. Coquoz, None.
Abstract ID: 175
We investigated the usefulness of intermittent contrast-enhanced ultrasonography (US) in the detection of prostate cancer. Nineteen patients with an elevated prostate-specific antigen level (≥1.25 ng/ml and a free-to-total PSA < 18%) or an abnormal digital rectal examination were enrolled in the study. We used an intravenous US contrast agent (SonoVue). Continuous gray-scale, intermittent grayscale, phase inversion gray-scale, and color and power Doppler US of the prostate were performed. Sonographic findings were correlated with biopsy results. After US contrast agent administration, we found significant enhancement on both gray-scale and Doppler images (p < .01). In three isoechoic tumors, we only detected focal enhancement using intermittent imaging. Focal areas of enhancement were identified in only one patient (5%) without cancer. Intermittent contrast-enhanced US of the prostate seems to be useful for selective enhancement of malignant prostatic tissue. Therefore, this technique may be useful for targeted biopsies.
A. Klauser, None.
Abstract ID: 176
Molecular and functional targeting of tumors with MRI, optical, and radioisotope imaging is well established. High-speed X-ray CT also has the potential to observe the transport of administrated contrast agents enabling quantification of tissue physiology in highly vascularized tumors. A sequential computed tomography study was performed with 11 mice using a high-speed clinical CT scanner (Mx8000 IDT, Philips Medical System). Following an infusion of 200 ml contrast through the tail vein, a rapid imaging sequence with 1.25-sec intervals was recorded for 4 min to follow the arterial phase. The sequence was continued for 5 min with 20-sec intervals during the wash out. The measured contrast kinetics in organs (kidney and liver) was fit with a compartmental model for physiologic parameter estimation. Different contrast concentrations and varying injection rates were employed in repeat studies to investigate the dependency of measurement results on imaging protocols. Simulation using compartment modeling was also performed to predict the contrast dynamics using elected physiological parameters. The comparison of measured contrast kinetics and simulation was used to examine the suitability of the selected compartmental model configuration. This study indicates that mouse physiology can be quantified with the fast CT providing 500 msec temporal resolution and 1.25-sec sampling intervals. A good correspondence between simulation and experimental findings was observed. No significant differences (p > .6) were found in the measured parameters (blood flow, capillary permeability, and relative compartment volumes) as long as the injection rate is above 1 ml/min and contrast concentration above 200 mg I/ml. This study reveals minimum requirements for temporal resolution and signal-to-noise ratio for quantitative CT perfusion studies in mice. The observation of tissue physiology heterogeneity is restricted by the limited spatial resolution of clinical CT. The study provides a rational basis for designing and developing new functional micro-CT systems dedicated to small animals.
Y. Liang, None.
Abstract ID: 177
3-D tomographic imaging of fluorescent-labeled probes in small animals is a potentially powerful modality for molecular imaging. However, the strongly absorbing and scattering nature of biological tissue makes the inverse problem highly ill posed. Since scatter and absorption are frequency dependent, the spectrum of light escaping from the animal will vary with the depth of the emission source. Consequently, spectral information should help to resolve source depth and hence reduce the degree of ill-posedness of the tomographic reconstruction problem. We propose to exploit this property using the hyperspectral imaging system described in Ref. [1]. Our measurement model assumes that multiple view angles will allow spectral data acquisition over most of the skin surface. Forward computations are currently based on a homogeneous tissue model and solved using a slab approximation with an extrapolated boundary condition. Extensions to an inhomogeneous model based on a volumetric mouse atlas are being developed: the atlas will be warped to the shape of the mouse by fitting to a grid pattern projected onto the mouse as shown in Ref. [1]. To reduce the dimensionality of the spectral data, we project onto a set of basis functions. The forward matrices for these basis functions are computed efficiently using a combination of on-the-fly and look-up tables. Images are reconstructed using a positively constrained conjugate gradient method. Simulation studies show that the addition of spectral information does indeed improve our ability to resolve the depth of fluorescent sources in a 3-D mouse model. Additional practical advantages of the hyperspectral system include reduction of autofluorescence through spectral subspace projection techniques and the possibility of distinguishing multiple fluorophores through their spectral characteristics.
R.M. Leahy, Concorde Microsystems 2.
Abstract ID: 178
Magnetic resonance histology (MRH) is a newly emerging area which is being used to explore the morphologic effects of genetic alterations in mice. The results of the following study show that very high quality microimages of the mouse anatomy can be achieved with relatively short scan times using innovative strategies for specimen preparation (Johnson et al.) in tandem with new tools and concepts that we have developed for imaging mice on a clinical 1.5-T scanner with a customized microimaging system. Mice (˜15 g) were fixed by cardiac perfusion with a mixture of 10% buffered formalin and gadopentetate dimeglumine. The whole fixed mouse was placed in a 50-ml tube and immersed in fluorinert. Imaging was performed at 1.5 T using a custom-built high-powered gradient coil insert (inner diameter 12 cm, peak slew rate 2000 mT/m/sec and maximum gradient strength 600 mT/m), a custom-built solenoid RF coil (3 cm inner diameter, 4 cm length), and a modified 3-D steady-state free precession (SSFP) imaging sequence, known as 3DFIESTA. We have developed methods for phenotypic imaging at 1.5 T, with a clinical whole body scanner, which is an unconventional approach to microimaging. The optimized 3DFIESTA pulse sequence produced high SNR and CNR image data and together with the customized MR hardware has proven to provide a highly efficient method for high-resolution MRM. High-throughput phenotyping is not yet an established application of MRI, and research in this area has focused on much higher static fields than 1.5 T as well as on novel multiple-mouse imaging concepts. Our results open the door to the use of lower field clinical imagers, which generally have greater capacity for parallel receive channels, and therefore are more suited to ultra-high throughput imaging applications.
P. Foster-Gareau, None.
Abstract ID: 179
Single-photon radiotracer techniques have many advantages over other molecular imaging techniques. A number of isotopes are readily available with half-lives of hours to months so that biological processes evolving over those time scales, including cell trafficking and tumor antibody uptake, can be studied. New specially designed detectors can discriminate signals from multiple isotopes, thus multiple biological functions can be observed simultaneously. Since localization of tracers relies on absorptive collimators, the sensitivity/spatial resolution trade-off can be easily optimized by interchanging collimators to suit specific applications. We developed a dedicated small animal single-photon radiotracer imaging system that includes tomographic capabilities for volumetric tracer concentration measurement. We also integrated a micro-CT to improve tracer anatomical localization. The system is used with an animal in a horizontal stationary position, with imaging components rotating around the animal. Our radionuclide detector is based on a 58 × 58 × 6 mm thick, 2.2 × 2.2 mm2 pitch, segmented NaI(Tl) crystal array combined with a position sensitive photomultiplier tube readout, where the spatial resolution of the detector is defined by the 2.2-mm crystal pitch to achieve high-resolution imaging. It is optimized to image radiation in an energy range from 20 to 250 keV, which includes radiation from I-125, Tl-201, Tc-99m, I-123, In-111, and others. CT components consist of a microfocus X-ray tube and a 50-μm pixel, GOS phosphor, CMOS readout detector. The SPECT spatial resolution was measured with different imaging geometries. It can approach 0.5-mm FWHM with a 0.5-mm pinhole. The sensitivity is greater than 60 cps/microCi for I-125 with a parallel-hole collimator. Distribution of I-125 tagged antibodies was measured 48 and 72 hr postinjection. Dual isotope imaging of a mouse with I-125 and Tc-99m agents indicated no visually detectable down scatter. Co-registered CT images greatly enhanced localization of radiotracers. This dedicated small animal system provides powerful capabilities for molecular imaging.
K. Iwata, Gamma Medica, Inc. 5.
Abstract ID: 180
Two-dimensional imaging of bioluminescent markers in vivo has been shown to be an effective means of monitoring biological activity and making quantitative measurements of some properties including relative source growth. Three-dimensional localization of markers for absolute source size and position within the living tissue would provide improved capability to quantify biological function. A novel tomographic technique for 3-D localization and quantification of in vivo bioluminescent molecular reporters in small animals has been developed. Data are acquired as a series of 2-D images of animal subjects lying horizontally on a stage, at multiple viewing angles. This method of imaging provides ease of animal handling and relatively high throughput for investigators. Surface radiance of photons emitted from deep within the tissue can be modeled by the diffusion equation and boundary conditions set by the topography of the animal surface. We reconstruct the animal surface topography in order to account for the conditions at the boundaries. At multiple viewing angles, images of line patterns projected onto the animal are analyzed for topography by extracting phase displacement information of the lines with respect to a constant line frequency. The reconstructed surfaces from the multiple viewing angles are registered and stitched together into one continuous surface defining the animal topography. An approximation to the Green's function for the partial-current boundary condition models the propagation of the photons within living tissue to the surface. Distributions and quantification of the photon emitting reporters are derived using bound-constrained inverse methods. Simulations and measurements in which light sources are embedded in tissue phantoms have demonstrated that this tomographic method successfully localizes source distributions to within 2–3 mm resolution. Our results suggest that the diffuse luminescence tomographic technique presents a valuable tool for monitoring activity tagged by bioluminescent reporters.
C. Kuo, Xenogen Corporation 4, 5.
Abstract ID: 181
Multiplexed fluorescent imaging of 2-D tissue slices is generally limited to about three different fluorophores using traditional techniques based on filters. With a hyperspectral imaging device, it is conceivable to acquire overlapping spectra and to deconvolve the different components, thus allowing a greater number of different dyes to be imaged. The advent of quantum dots, with narrower emission spectra, opens up the possibly of much higher degrees of multiplexing than was previously possible. The system is based on an imaging spectrograph, ImSpector, coupled to a cooled CCD camera. A line in the object space is imaged into a 2-D spatial-wavelength image. By scanning the spectrograph across the object, one can therefore acquire a 3-D cube with two spatial dimensions and wavelength. In the present apparatus, excitation is obtained with 640 nm and/or 532 nm laser light. Microarrays and tissue slices have been imaged to evaluate spatial resolution capabilities. A slice of a mouse expressing GFP superimposed with a narrow phantom containing Vybrant DiD dye was also imaged for dual-band imaging. The capability to deconvolve different fluorophores was studied by mixing, in different proportions, Vybrant DiD and Vybrant DiR dyes, which have partially overlapping spectra. In the figure one can see a measured spectrum with the presence of both Vybrant DiD and Vybrant DiR dyes. In black is a fit of the data with the two components measured separately. From the fit one can extract the proportions of the different dyes. A hyperspectral imaging device capable of multi-fluorophore imaging from microarrays to tissue slices has been presented. Some first images and spectral deconvolutions have given promising results.
G. Zavattini, None.
Abstract ID:
3-D near infrared in vivo fluorescence imaging of mice is an emerging technique. Recent work [V. Ntziachristos, 2002] has shown that 3-D reconstruction of localized fluorescence is possible by measuring the fluorescence light intensity on the surface of the mouse as a function of illumination. Since the absorption coefficient of tissue is wavelength dependent, here we propose to add spectral information to the data to improve the reconstruction. The system is based on an imaging spectrograph, ImSpector, coupled to a CCD camera, which can image a line in the object space into a 2-D (x, λ) image. By scanning the spectrograph across the object, a 3-D cube (x,y, λ) can be acquired. The animal is illuminated from underneath across a matrix of points, and the emitted fluorescence light is collected from above. The excitation light (640 nm) is filtered with a notch filter. The system is mounted in a light-tight black box. A first set of measurements was made to assess the spectral variations of the detected light as a function of thickness of traversed tissue. Small fluorescent phantoms, with Vybrant DiD dye, were placed between slices of red meat. From the spectra, the ratio, R, of two wavelength bands was taken and plotted as a function of depth. The ratio R showed a significant and almost linear dependence on depth. A first in vivo study (in collaboration with M. Cole, J. Ross, N. Van Bruggen from Genentech) showing fluorescently labeled antibody accumulating in a tumor is shown in the figure. A hyperspectral imaging system has been assembled. First images have been acquired, demonstrating the validity of using spectral data for source depth information in 3-D imaging.
G. Zavattini, None.
Abstract ID: 183
Clinical imaging equipment is routinely used to detect and diagnose cancer in patients. Advances in molecular biology and genomics have led to the development of numerous mouse models for the study of disease. Clinical instruments which are currently available lack both the sensitivity and spatial resolution to image anatomy and function in small laboratory animals. Recently, several special purpose laboratory microimaging instruments have become available. X-ray micro-computed tomography (micro-CT) provides volumetric imaging which is well suited to contrast-enhanced imaging. Micro-CT systems designed for both in vitro and in vivo imaging have a number of applications in cancer research. The proliferation of small vessels, which characterize angiogenesis, may be studied by filling the vasculature of a tissue specimen with contrast agent. This technique allows vessel structures less than 10 μm in size to be resolved. Three-dimensional quantitative parameters of vessel morphology may be extracted from high-resolution volumetric image data. In vivo micro-CT provides noninvasive mouse imaging at resolutions approaching 50 μm. The ‘leakiness' of vessels within tumors may be exploited through the use of extravascular contrast agents by imaging the increased pooling of contrast agent within cancerous tissues. Small animal CT systems capable of volume acquisitions on the order of 1 sec are in development and provide the potential for dynamic perfusion CT studies in small animals where tumor size, blood volume, blood flow, and permeability may be quantified noninvasively. Small animal micro-CT has multiple applications in cancer research through high-resolution anatomical studies of tissue microvasculature and functional studies of tumors through dynamic perfusion imaging. These microimaging techniques have significant applications in the study of tumor biology, and the development and evaluation of new therapeutics.
M.M. Thornton, General Electric Medical Systems 5.
Abstract ID: 184
We have developed a website entitled “MI-Central” http://www.mi-central.org, a molecular imaging portal website, as a resource for the molecular imaging community. The goal is to provide useful information for investigators already in the field, as well as educational material for those new to the field. The web pages were designed in HTML, database backed and include text, graphics as well as animation segments. A user can learn about various imaging technologies (e.g., small animal PET, SPECT, CT, MRI, optical, ultrasound, etc.). A user can also learn about various subtopics in molecular imaging (e.g., molecular biology, gene therapy, etc.). The user can take advantage of an online glossary when they come across an unfamiliar term. A “Hot Papers” section listing relevant published papers dating back to 1996 is available. When possible, links to the home pages of those publications are available so that the user may read the entire article. Industry and supplier links can be found under the “Instrumentation” section, which includes technology vendors and suppliers of laboratory materials needed for molecular imaging. Links to sources of both federal and private funding can be found in the “Funding” section. A list of all past and upcoming molecular imaging meetings and research groups involved in molecular imaging is also provided. This molecular imaging portal should be a useful source of information for the entire field. Although still in its infancy, rapidly increasing content and ease of use should lead to a growing user base.
J. Strommer, None.
Abstract ID: 185
Confocal microscopy is a powerful method of optical sectioning in biological tissue, and can be used for in vivo molecular and cellular fluorescence imaging in small animal models if the lenses and scanning mechanisms can be made sufficiently small. We are developing a dual-axes architecture that uses separate low numerical aperture lenses oriented with the illumination and collection axes crossed at an angle q to the midline. This configuration produces high axial resolution, and offers long working distance, large field of view, and reduced noise from scattered light. With fiber coupling and the scanning mirror located in the postobjective position, this architecture can be scaled down in size to millimeter dimensions with micro-electromechanical systems (MEMS) technology. We have constructed a benchtop dual-axes prototype, collected fluorescence images from several transgenic models tagged with GFP in 90 sec per image, and measured the image signal-to-noise ratio (SNR) and contrast. This system has a resolution of approximately 3 to 4 μm. We found the fluorescence SNR from the following locations: cytoplasm of DSH-293 cells of 13 ± 5; Purkinje cell bodies, molecular layer, and internal granule layer in mouse cerebellum of 13.1 ± 0.5, 4.4 ± 0.3, 4.3 ± 0.3, respectively; cytoplasm of muscle of 4.8 ± 0.5; and stripes of expressed engrailed of 15 ± 3. Also, we measure the fluorescence contrast ratio between adjacent locations: cytoplasm and nuclei of DSH-293 and muscle cells of 1.8 and 3.5, respectively; Purkinje cell bodies and the molecular and granular layers of cerebellum of 3.0 and 3.1, respectively; and stripe of engrailed and cytoplasm of 115. The images shown demonstrate the proof of concept for use of the dual-axes confocal architecture to collect fluorescence images with high axial resolution, SNR, and contrast for future in vivo molecular and cellular imaging.
T.D. Wang, None.
Abstract ID: 186
In vivo imaging of bioluminescent reporters has become an established technology in the recent years for studying cancer, infectious disease, and gene expression in laboratory animals. The next step in improving the quantitative aspects of this technology is to develop instrumentation and software that will allow researchers to account for the effects of photon absorption and scattering in tissue. A full multi-view 3-D diffuse tomography imaging system is one approach to improving this technology and such systems are in development. However, methods to incorporate diffuse tomography algorithms in a simpler, higher-throughput, single-view 2-D imaging system, would be beneficial. We have recently developed an advanced 2-D imaging instrument designed for high throughput that has improvements in the optics, increased filtering and fluorescence capability, and incorporates a new technology that allows partial 3-D reconstructions from a single-view image. A custom lens features an extended range of field of views and includes user-accessible filter wheels that accept up to 22 filters—enough to accommodate a large set of fluorescence filters and another set of filters for performing spectral imaging studies on bioluminescent samples. A scanning laser galvanometer is incorporated into the design for projecting a structured light pattern onto the subject. Analysis of the displacement of these lines allows the surface topography of the subject facing the imaging system to be reconstructed. Diffuse tomography codes have been developed to utilize the single-view surface topography and luminescence measurements to reconstruct the location of internal light sources. A description of the instrument and examples of phantom mouse reconstructions will be presented.
B.W. Rice, Xenogen Coporation 4, 5.
Abstract ID: 187
To design, construct, and evaluate in vivo an interventional catheter-based imaging sensor for the intravital monitoring of molecularly sensitive near-infrared fluorescent (NIRF) probes and optical marker genes. An imaging device based on a miniaturized fiberoptic sensor (MIFS) was built in which the image created by a 2.7F fiberoptic catheter is relayed through a dichroic mirror, bandpass filter, and on two independent cameras, permitting simultaneous recording of white light and fluorescent images. Spatial resolution and spectral transmissions were measured, and sensitivity was determined on a dilution series of a near-infrared (NIR) fluorophore. In vivo testing was performed on nude mice bearing intraperitoneal tumors expressing green fluorescent protein (GFP), and on a mouse model of ovarian carcinoma with enzyme-activatable NIR probes sensitive to tumoral activity of the protease cathepsin B. The 18-G catheter showed a resolution of seven line-pairs per millimeter, and a detection limit for the fluorochrome Cy5.5 of 1–10pmol. Detection of endogeneous GFP gene expression was feasible in tumor nodules of <1 mm in size, with tumor SI of 153.26 ± 26.45 (mean ± SD) compared to SI from adjacent nontumoral tissue of 36.73 ± 11.69 (mean ± SD; p <.008). Similarly, activation of the NIR probe by tumoral cathepsin B could be detected in the peritoneal tumor seeds of the ovarian cancer model with tumor SI of 246.33 ± 7.77 (mean ± SD) compared to SI from adjacent nontumoral tissue of 41.56 ± 18.64 (mean ± SD; p <.001). Combined NIR/white light images consistently revealed more tumors than white light examinations alone. The proposed minimally invasive optical imaging system offers a method to obtain simultaneous white light and fluorescent images at high sensitivity, allowing the superimposition of anatomic and molecular information in real time. In vivo MIFS imaging of gene expression, enzyme activity, and potentially other molecular events is feasible using both a transvascular approach as well as direct interventional access to several organs and body cavities.
M.A. Funovics, None.
Abstract ID: 188
A high-resolution single photon emission computed tomography (SPECT) system has been developed to image unrestrained, nonanesthetized small laboratory animals. An optical-based animal position tracking apparatus has been developed that measures the position of the animal during the SPECT scan so motion-induced blurring can be corrected prior to SPECT image reconstruction. This system is being developed for animal studies where anesthesia and/or physical restraint can alter the physiologic process being monitored (e.g., brain function). The objective of this work is to measure, in real time, the 3-D pose of a nonanesthetized, unrestrained small animal to submillimeter accuracy during a SPECT scan. This pose data are used to correct the SPECT data before reconstruction. The SPECT volume is registered to a micro-CT volume of the same animal. The optical tracking system employs infrared (IR) strobes, retro-reflective markers, and a pair of stereo CMOS cameras to measure the 3-D pose of the animal's head. Also, the cameras and laser profilers are used to construct a surface of the animal's exterior. The surface is registered to a standard surface (micro-CT scan) to determine the 3-D transformation parameters. The system tracks the position of the subject to within 0.06 mm at a rate of more than 10 measurements per second. Initial experiments have been completed with the laser profiling system to generate surfaces of the animal's exterior, and surface registration algorithms are under development. An optical animal tracking system has been developed that measures the 3-D pose of an animal to submillimeter accuracy. This tracking system operates in parallel with a small animal micro-SPECT imaging system, allowing correction of motion-induced blurring prior to image reconstruction. This system has important applications in physiologic imaging in both animals and humans.
S.S. Gleason, None.
Abstract ID: 189
Dynamic PET can be utilized for quantitative measurements of biological parameters, such as metabolism, flow, or proliferation rates. The description of the underlying biological mechanisms is based on compartmental models, which can be translated into mathematical equations. Fitting these equations with the dynamic PET measurements yields estimates for the biological parameters. The parameter estimates are corrupted by a variety of error sources. One class of errors originates in the description of the biology. A compartmental description is only an incomplete model of the true biological mechanisms. Most models make some assumptions on certain reaction rates that are not completely valid. PET instrumentation is responsible for another class of errors. It involves systematic errors such as scatter and partial volume effects as well as statistical errors such as quantum noise. We will present a general framework for à-posteriori estimation of accuracy and reproducibility of the kinetic models. With appropriate input the framework will determine whether the biological inaccuracies or the instrumentation determines the overall measurement errors. The framework can subsequently be applied to optimize the parameters for image acquisition and data processing. A simulation software based on Geant4 is utilized for the assessment of the systematic errors. The statistical errors (noise) can be modeled as Poisson noise in the sinograms. For the estimation of the errors in the biological parameters, the error propagation of the reconstruction and the parameter fit has to be considered. The choice of reconstruction parameters and the temporal sampling scheme have a strong influence on the amount of error propagation. A careful analysis of the various error sources is the prerequisite for the optimization of acquisition and processing protocols in order to get a good trade-off between accuracy and reproducibility [1].
Poster Session 03: Novel Targets: Targeting Pathways in Signaling Networks
Abstract ID: 190
We used real-time assays on live cells to understand the kinetics of cellular responses to induced thermal stress. Using the stress-inducible promoter from a gene encoding a 70-kDa heat shock protein (Hsp70) to express luciferase as an indicator of thermal stress in living cells, we studied temporal responses to a thermal gradient over a range of temperatures and durations. A stable cell line (NIH 3T3) was generated using this reporter construct, and luciferase activity, measured with a cooled CCD camera, was used as an indicator of cellular response to thermal stress over time durations of 5 sec to 20 min. Following each of the thermal stress conditions, luciferase expression was assessed in live cells every 2 hr for the first 10 hr and at 24 hr. The levels of luciferase expression reflected thermal stress patterns up to a threshold temperature, beyond which we define a transition zone within a very narrow temperature window (beginning at 51.5 °C), where luciferase activity is lost and cells are still viable. At higher temperatures, over 50% of cells die (53.2 °C). Cells treated at both of these higher temperatures recover the ability to bioluminesce upon a secondary stress. In general, higher temperatures were tolerated for shorter durations and vice versa. Longer durations and higher temperatures resulted in delayed expression of the reporter gene. Our results indicate that there is an upper and lower thermal threshold for Hsp70 gene expression depending on the duration and magnitude of the thermal stressor. The Hsp70-luc reporter gene construct may be useful for the assessment and prediction of collateral thermal damage created during treatment of biological tissue with lasers and other sources of heat, particularly within this reporter-defined transition zone.
C.E. O'Connell-Rodwell, None.
Abstract ID: 191
In vivo bioluminescent imaging (BLI) was utilized to monitor heme oxygenase-1 (HO-1) expression in response to oral infection with Salmonella typhimurium in HO-1-luc reporter transgenic FVB juvenile and adult mice. HO-1 is the rate-limiting enzyme in the degradation pathway of heme producing biliverdin, carbon monoxide, and iron. There are 2 well-described isozymes of HO, the constitutive HO-2 and inducible HO-1. HO-1 is induced by stress, LPS, heme, and heavy metals and is reported to mediate antiinflammatory processes. However, the exact role of HO-1 in mediating host response to infection is unknown. Therefore, we sought to characterize and compare the expression patterns of HO-1 of juvenile (a more susceptible age group) and adult mice in response to oral inoculation with a wild-type strain of S. typhimurium, SL1344. HO-1 expression patterns during the full disease course were determined by BLI and comparisons of abdominal signals between infected mice and age-matched controls were performed. HO activity assays on excised tissues were performed at time points as guided by in vivo expression patterns. In 2- and 6-week-old HO-1-luc reporter mice, BLI revealed that HO-1 expression increases late in infection at 7 and 10 days post infection, respectively. This increase correlated with a significant increase in HO activity in the liver and spleen of 2-* and 6-week-old infected mice compared to age-matched control mice (p ≤ .05, *except p = .06 for spleen; n = 4-6). Interestingly, there was a decrease in HO activity in the jejunum and cecum in both infected age groups. The expression kinetics of HO-1, as revealed by BLI, appears to demonstrate that HO-1 is activated as a late response to infection in both juvenile and adult mice.
S.M. Burns, None.
Abstract ID: 192
Interactions between circulating cancer cells and surrounding vasculature significantly influence the metastatic process. The galectins, a family of beta-galactoside carbohydrate-binding animal lectins, have a variety of cellular roles and have been implicated in metastatic events. Observations from our laboratory indicate that galectin-3 (gal-3) is involved in carbohydrate-mediated adhesion and interacts with the carbohydrate portion of the tumor specific Thomsen-Friedenreich (TF) antigen. The TF antigen is present on up to 90% of adenocarcinomas including those of the breast, prostate, and ovary, and its expression correlates with the metastatic state of the tumor. We propose that tumor cell-associated TF antigen mediates endothelial cell surface gal-3 expression and secretion into the surrounding vasculature and that the interaction of these antigens promotes cancer cell adhesion and metastasis. Hence, the ability to detect and examine the migration and interplay of TF and gal-3 may lead to a better understanding of tumor metastasis. Furthermore, development of agents that specifically bind to TF and/or gal-3 may act to image this process and may provide the foundation for development of a new class of antimetastatic agents. While carbohydrate molecules have been designed to block gal-3 function, peptide-based mimetics would have advantages of affinity and ease of synthesis. We have used bacteriophage (phage) display to isolate a series of tumor targeting peptides. High affinity (low nanomolar kDa) specific peptides that bind TF or gal-3 have been selected from phage libraries displaying 6 or 15 amino acid in length foreign peptides. The synthesized peptides effectively bind to and image human cancer cells in culture and those that bind at the gal-3–TF interaction domain function to block human MDA-435 breast cancer cell adhesion. Not only the peptides, but also the phage displaying the targeting peptides can be used to image tumors and vascular-tumor cell interactions in vitro and in vivo.
S.L. Deutscher, None.
Abstract ID: 193
Decorating cell surfaces with functional groups that react selectively with small chemical probes is useful for a variety of imaging and drug delivery strategies. Our laboratory has previously demonstrated that cells in tissue culture incorporate the abiotic azide group into cell surface glycoconjugates via metabolism of unnatural carbohydrate precursors. Moreover, these cell surface azides can be selectively reacted with phosphine-based probes in a process termed the Staudinger ligation. This chemoselective reaction forms a covalent link between the cell surface azide and the exogenously delivered phosphine reagent. We report here the extension of this work in vivo. Specifically, mice were administered an azide-bearing carbohydrate precursor (Ac4ManNAz), and the metabolic incorporation of the unnatural functional group was observed in various murine organs by treating isolated cell populations or tissue homogenates with a phosphine probe. The presence of azide groups on murine splenocytes was also detected by administering the phosphine probe to mice previously treated with the unnatural azido sugar, and our results demonstrate that the Staudinger ligation proceeds selectively in vivo. We are currently exploring the in vivo Staudinger ligation with radiolabeled phosphine reagents for noninvasive imaging of metabolically incorporated azide groups in mice.
J.A. Prescher, None.
Abstract ID: 194
Live cell imaging is an important new technology for the study of transient effects of changes in gene expression on cell behavior. We have modified the Large-Scale Digital Cell Analysis System (LSDCAS) to provide these analysis capabilities in large populations of living cells. To accomplish this goal, we developed software needed for experiments using green fluorescent protein (GFP)-tagged dominant negative (Ad5CMVp53Δ175-GFP) and wild-type (Ad5CMVp53-GFP) adenovirus vectors to study the role of p53 in radiation-induced mitotic catastrophe. These expression systems coexpress GFP as well as the p53 variant genes, so the infected cells are identifiable as green when illuminated with blue light. Using LSDCAS, fluorescent and phase-contrast images were alternately acquired for 3 days using LSDCAS at 5-min intervals. LSDCAS has the capability to change microscope optics from phase contrast to fluorescence detection without manual intervention. For visualization of the manipulations of gene expression, the fluorescence signals were overlaid in green over the phase contrast image, which was colored in red. These experiments were designed to test the hypothesis that p53 is an important regulator of mitotic catastrophe, one of the mechanisms of cell death following irradiation. Experiments using dominant-negative p53 gene expression to attenuate gene activity lead to an increase in the yield of mitotic catastrophe in mouse embryo fibroblasts expressing normal p53. When normal p53 was delivered to mouse embryo cells containing an endogenous mutant p53 gene, the incidence of mitotic catastrophe was markedly reduced. Thus, these results suggest that p53 plays a role in the regulation of mitotic catastrophe following radiation. In combination with our biochemical and cytological studies, these findings indicate a novel role for p53 in radiation-induced cell killing. Further experiments will extend this technology to other gene expression systems studying other phenotypic endpoints. Support: 1R33 CA94801 and 3R01 CA74899 (NIH).
M.A. Mackey, None.
Abstract ID: 195
Understanding of the tumor microenvironment is important to attempt to design new strategies for development of new chemotherapies for the treatment of cancer. Tumor hypoxia is one condition that may influence many factors, including efficacy of drug treatment, the potential for metastasis, and induction of an aggressive tumor phenotype. Currently immunohistological staining of hypoxic regions of tumors is possible with a 2-nitroimidazole compound such as EF5. Noninvasive in vivo imaging of hypoxia is possible using a reporter gene driven by a hypoxia-inducible promoter. A series of promoters have been created using hypoxia response elements (HREs) from genes upregulated by hypoxia-inducible factor, HIF. Mammalian expression vectors were designed using multiple repeats of the HREs from Carbonic Anhydrase IX (CA9), Phosphoglycerate Kinase (PGK), or Vascular Endothelial Growth Factor (VEGF), a minimal CMV promoter and a firefly luciferase reporter. The induction of luciferase activity using these promoters can be optimized by selection of the cell line in which they are expressed. Quantitative RT-PCR was used to determine upregulation of CA9, PGK, and VEGF in several cell lines exposed to hypoxic conditions. Cell lines were transfected with the luciferase-HRE vectors, CA9, PGK, PGK/CA9, and VEGF. Cells were exposed to normoxia or hypoxia, lysed, and assayed for light production. Cell lines with the greatest increase of mRNA for a specific gene also had the greatest induction of light using the corresponding promoter after exposure to hypoxia. Using this information, cell line and vector combinations can be chosen for optimal induction and increased sensitivity of a reporter for in vivo imaging of tumor hypoxia.
B. Baggett, None.
Abstract ID: 196
The high metabolic rate of malignant tumors is a common characteristic observed in many types of cancer. This occurs in the proliferating tumors in an effort to maintain supply and demand needs. Often, an overexpression of glucose transporters (GLUTs) is seen as a characteristic of the tumor malignancy. We have manipulated this high glycolytic rate for targeted delivery of contrast agents to the tumors. In an effort to locate and identify viable tumor areas, 2-deoxglucose was labeled with fluorescent dyes in the NIR region of the spectrum. The uptake of the fluorescent-labeled glucose can act as an indicator of metabolic activity within the tumor. The demand for early tumor detection techniques has propelled such new technologies in the field of optics. In this current study, monitoring the delivery ability of NIR-labeled 2-deoxyglucose has opened a new realm of molecular beacons and detection. Using 2 fluorophores, Cypate (Achilefu) and Cy7 (Zheng) the uptake of the delivery vehicle by the tumor was observed by their strong fluorescence within the tumor region. Applying several imaging methods including the low-temperature cryo-imager, the phased-array system, and a CCD camera system, we were able to collect several images for comparative data analysis. Successes with NIR imaging in cancer detection have fueled efforts to further develop methods of delivery. In Figure 1, a comparison of ICG and a control tumor are used to demonstrate the successful delivery of the Cypate beacon to the tumor.
D.M. Blessington, None.
Abstract ID: 197
The dopamine 2 receptor (D2R) signal pathway regulates gene expression by phosphorylation of proteins including cAMP response element-binding protein (CREB), a transcription factor. In this study, we developed a reporter strategy using GAL4 fusion CREB to assess the phosphorylation of CREB, one of the targets of the D2R signal transduction pathway. We used 3 plasmids: GAL4 fusion transcription factor (pCMV-CREB), firefly reporter with GAL4 binding sites (pG5-FLUC), and D2R plasmid (pCMV-D2R). Group 1 293T cells were transiently transfected with pCMV-CREB and pG5-FLUC, and Group 2 cells were transfected with all 3 plasmids. Transfected cells were stimulated with different concentrations of dopamine (0–200 μM). For animal studies, Groups 1 and 2 cells (1 × 106) were subcutaneously injected on the left and the right thigh of 6 nude mice, respectively. Dopamine stimulation was performed with intraperitoneal injection of 
J. Min, None.
Abstract ID: 198
Protein-protein interactions control many cellular processes and are an important area of study in molecular biology. We previously reported an inducible yeast two-hybrid system and split protein complementation systems with firefly luciferase and synthetic renilla luciferase [1–3] to study protein-protein interactions in cell lines and in living animals. In this study we used 2 proteins that heterodimerize in the presence of rapamycin leading to complemented synthetic renilla luciferase (hrluc) activity. Plasmid constructs with N- and C- portions of split synthetic renilla luciferase (hrluc) gene, alone, or linked to genes encoding the proteins FRB and its heterodimerizer FKBP12, were made using a specific linker. The efficiency of rapamycin-mediated dimerization of the proteins was indirectly measured by estimating the recovered hrluc activity in cell culture and in living animals, using transiently transfected 293T cells. Escalating rapamycin doses (up to 10 nM) resulted in increasing hrluc activity in transiently transfected 293T cells. The activity observed without rapamycin was 6±2 relative light units (RLU)/μg protein/min and was similar to that in mock-transfected cells. The cells exposed to 10 nM rapamycin for 24 hrs showedsignificant increase(p <.001)inhrluc activity(4000±250 RLU/μg protein/min). The activities observed in mice (N = 3) with cell implants were similar to background activity in the control group (N = 3) and in animals injected with 10μg ip rapamycin. The animals that received 25 μg and 50 μg of rapamycin showed hrluc activity that was 2-fold and 3.5-fold greater than in control mice respectively. The described heterodimerization system can be used to specifically modulate protein function in cells and image this modulation in living animals. These studies will have important implications for studying the effects of small drugs designed to modulate protein-protein interactions.
R. Paulmurugan, None.
Abstract ID: 199
The binding of growth factors to specific receptor tyrosine kinases activates the phosphoinositide 3-kinase (PI3K) and the serine-threonine kinase Akt. Akt promotes cell survival and proliferation by directly phosphorylating and inhibiting members of the FOXO subfamily of forkhead transcription factors. To image PI3K-AKT pathway activity noninvasively, we developed a retroviral vector where the HSV-tk1Nes/GFP fusion reporter gene was placed under control of a newly designed artificial forkhead-specific enhancer element. PC3 prostate carcinoma cells were transduced with this reporter vector. After treatment of the cells with PI3K inhibitor (LY294002), the GFP-positive cell population was selected by FACS. This population was sensitive as to LY294002, as to a Src-selective kinase inhibitor (PP1) and to starvation, which suggests down-regulation of AKT activity. This down-regulation resulted in nuclear translocation of endogenous forkhead factors leading to activation of the TKnesGFP reporter gene expression. Activation of the reporter gene was assessed by fluorescence microscopy, FACS analysis and radiotracer assay. Under LY294002 treatment, the FIAU/TdR ratio was 0.529±0.005, as compared to 0.09±0.001 in nontreated cells. SCID mice bearing PC3 subcutaneous xenografts with or without cis-FOXO reporter system were used for in vivo experiments. One group of animals (N = X) was treated with LY294002 over 1 week, and the control group was treated with the vehicle. MicroPET images at 2 hr post [18F]FEAU administration demonstrated significant accumulation of [18F]FEAU in PC3/cis-FOXO/TKnesGFP tumors but not in the wild-type PC3 tumors. Minimal accumulation of [18F]FEAU was observed in the DMSO-treated control group of mice. We have developed a novel PI3K-Akt–FOXO-specific TKGFP reporter system. Our results demonstrate that noninvasive imaging of FOXO-mediated transcriptional regulation of genes is feasible using the TKnesGFP-based cis-reporter system. This approach can be used for in vivo studies of the involvement of this pathway in cancer and for the assessment of novel anticancer drugs.
I.S. Serganova, None.
Abstract ID: 200
Expression of the human heat shock protein 70 (HSP70) gene is induced by a wide range of physiological stress, including exposure to heat shock and the ancymycin drug, 17AAG, which binds to HSP90 and leads to the degradation of certain signaling proteins (e.g., Raf and HER2). We developed a reporter system to image HSP70 activation under conditions of stress and drug treatment, using a recently constructed dual-reporter gene—the human sodium iodide symporter (hNIS)–green fluorescent protein (GFP) IRES-linked dual-reporter system—for both nuclear and optical imaging. A retroviral vector (pQHNIG70) containing an inducible and constitutive pHSP70 promoter, driving the hNIS-IRES-EGFP dual-reporter gene (NIG), was recently constructed (pHSP70-NIG). RG2 cells were transduced with the vector, and GFP positive cells (RG2-pHSP70-NIG) were selected by FACS after heat shock, and then maintained under normal growth conditions. In vitro studies were performed over 4 days following heat shock (42 °C for 60 min) and exposure to 0.5 nM to 1 mM concentrations of 17-AAG. Response was monitored by fluorescence imaging, FACS analysis, and immunocytochemistry for HSP protein. Radiotracer studies and in vivo imaging studies are pending. Following heat shock stimulation, pHSP70-NIG expression monitored by GFP imaging and FACS increases between 3 and 48 hr, reaching a peak at 24 hr, and then slowly declines to baseline levels level after 48–72 hr. Immunocytochemistry showed high expression of HSP70 protein at 24 hr compared to that at baseline (pre-heat shock) and in wild-type RG2 cells. Increases in NIG fluorescence monitored by FACS were seen as early as 6 hr following a 3-hr exposure of RG2-pHSP70-NIG cells to 17AAG. The pHSP70-NIG dual-reporter gene is sensitive to heat shock and 17-AAG in vitro, and is likely to be a useful in vivo reporter system for imaging stress responses and specific drug treatment effects in tumor and normal tissues.
J. Che, None.
Abstract ID: 201
Homodimeric protein-protein interactions provide a powerful mechanism in biology to regulate cellular functions, but they are particularly challenging to study in vivo. We used a split synthetic renilla luciferase (hrluc) protein-fragment-assisted complementation-based bioluminescence assay [1] to quantitatively measure and visualize real time protein homodimerization in mammalian cells and in living mice. The well-studied mutant herpes simplex virus type-1 sr39 thymidine kinase (TK) enzyme was used as a model homodimeric protein. We produced 4 different protein chimeras of TK fused to N-terminal (N-hrluc) or C-terminal (C-hrluc) fragments of hrluc, either upstream or downstream of TK. Transient cotransfection strategies in 293T cells led to homodimerization of the TK chimera combination containing N-hrluc and C-hrluc in the upstream and downstream positions respectively (tail-to-head homodimer) and to successful complementation of hrluc. This was monitored using luminometry (68-fold increase, p <.01, above background light emission in RLU/min/mg protein) and by CCD camera imaging of living mice implanted with ex vivo transfected cells (2.6-fold increase, p <.01, above background light emission in maximum photons/sec/cm2/sr). We then devised an experimental strategy to both simulate and image inhibition of TK homodimerization. We constructed variant TK chimeras (possessing extra N-hrluc or C-hrluc “spacers”) to simulate the effects of hypothetical dimerization inhibitor drugs. Luminometry and CCD camera imaging of living mice were used to also monitor the resulting incremental lack of association of the tail-to-head TK homodimer. Ex vivo substrate (8-3H Penciclovir) accumulation assays in 293T cells expressing the TK chimeric proteins showed viable TK enzyme, with no cells possessing net accumulation less than 56% of control cells expressing nonchimeric TK. Application of this potentially generalizable assay to screen for molecules that promote or disrupt ubiquitous homodimeric protein interactions could serve not only as an invaluable tool to understand biological networks but can also be applied to drug discovery and validation.
T.F. Massoud, None.
Abstract ID: 202
Bone-morphogenetic protein 4 (BMP4) is an important regulatory molecule that functions throughout vertebrate development starting as early as gastrulation in mesoderm induction, continuing through neural tube and limb formation, and later stages such as bone induction, tooth, and skin development. To visualize the neonatal regulation of mouse BMP4, we utilized a transgenic mouse in which the transgene is comprised of the BMP4 promoter driving expression of a firefly luciferase gene as a reporter. Neonatal mice were injected with luciferin and imaged daily between 2 and 14 days of life using a cooled CCD camera. Gene expression in newborn animals on Days 1–2 could be visualized after injecting the maternal mouse with luciferin, as luciferin is transmitted to the neonates through the breast milk. On Days 1–5, the entire mouse is luminescent, with the strongest signal coming from the spinal column, the top of the head, and the nose. During Days 5–6, the signal preponderance spreads to areas of the face and extends from the spinal column ventrally. During Days 7–8, the signal in the spinal column diminishes while becoming more uniform throughout the body. During Week 2, the body signal is stable, whereas the hindlimbs and ears greatly increase in luminescence. After Week 2, the signal decreases throughout the animal except for the tail and tips of the ears, which continue to luminesce throughout adult life at a low level. We conclude that BMP4 is spatiotemporally regulated in the neonatal mouse and appears to be expressed strongest in body regions experiencing their maximal growth phases.
J. Ravindran, None.
Abstract ID: 203
An increase of phosphocholine in cancer is closely related to malignant transformation, invasion, and metastasis. We are investigating pathways in the choline phospholipid cycle, which contribute to the elevation of phosphocholine and total choline using 1H NMR spectroscopy. Here, we used a choline kinase inhibitor, hemicholinium-3 (HC-3), to determine the effect of choline kinase inhibition on the choline phospholipid metabolites of malignant and nonmalignant human mammary epithelial cell lines. Parallel assays were performed to determine the effect oncell viability. The human breast cancer cell line MDA-MB-231 and the nonmalignant human mammary epithelial cell line MCF-12A were studied. Cells were exposed to fresh media with/without 5 mM HC-3 for 24 hr, and perchloric acid (PCA) cell extracts were obtained. Water-suppressed proton spectra of PCA extracts were acquired on a Bruker 500 MHz NMR spectrometer. A significant decrease of cell viability was observed when both MDA-MB-231 and MCF-12A cells were treated with 5 mM or higher concentrations of HC-3. Proton spectra of MDA-MB-231 breast cancer cells were typical of malignant cells with high phosphocholine (PC) and low glycerophosphocholine (GPC) but when treated with 5mM HC-3, PC and total choline[choline(CHO)+PC + GPC] decreased and the PC/GPC ratio significantly decreased. These data demonstrate that the high phosphocholine level of cancer cells is largely dependent upon choline kinase and support the concept of developing effective targeting of choline kinase in cancer cells as a therapeutic strategy. Acknowledgements: This work was supported by P-20 CA86346. We thank Dr. V.P. Chacko for expert assistance.
(A) Proton NMR spectra of cell extracts obtained from comparable cell numbers of control and HC-3 (5 mM) treated MDA-MB-231 cells. Spectra are scaled to display comparable signal to noise. 
N. Mori, None.
Poster Session 04: Use of Clinically Approved Molecular Imaging Agents
Abstract ID: 204
Over the past decade, transplantation of stem or progenitor cells has been proposed as a revolutionary new technique for the treatment of malfunctioning endogenous cell populations. The introduction of cells (myoblasts) that can remodel damaged or dysfunctional myocardium to perform cardiac work once again is of great interest. However, noninvasive monitoring of the myoblasts biodistribution in tissue is required for further validation. MR imaging offers both near-cellular (25-50 μm) resolutions and whole-body imaging capability. The isolation and culture of pig skeletal myoblasts were previously described. The 2 suspensions (30 ml lipofectamine in 0.5 μl of Opti-MEM and 100 μg total iron's Endorem in of Opti-MEM 0.5 ml) were mixed, and the suspension was added to the dishes containing myoblasts. Cells were then labeled, collected, and fixed in 2% buffered formaldehyde in PBS with 1.5% (w/v) gelatin. Gelatin (3%, w/v) was added to a solution of Gd-DTPA (1:200 solution) in PBS, and 0.4 ml of the solution was put in culture insert. We then added 10 μl of each cell suspension to a culture insert. Finally, the labeled cells were covered by the same gelatin/Gd-DTPA solution served as samples for experiments. Samples of unlabeled cells were used as controls. High-resolution MR imaging was performed on the myoblasts samples using a GE 1.5 T Signa CVi MRI (General Electric Medical Systems, Milwaukee, USA). The imaging protocol used a 3-D SPGR sequence and a round surface coil (inner diameter 1.5 cm). Our results have shown that the lipofectamine labeling method used produced in 100% labeled myoblasts with rapid SPIO uptake. Single myoblasts could be detected by high-resolution MRI. Unlabeled cells could not be detected in any of the control samples used. The technique of high-resolution MR imaging detecting labeled myoblasts may provide biological insight into stem cells engraftment in vivo.
Z. Zhang, None.
Abstract ID: 205
This study investigates the correlation between the temporal changes in T1/T2-weighted contrast-enhanced MRI (CE-MRI), histological evaluation (H&E staining), and genomic analysis using oligonucleotide microarrays in a murine squamous cell carcinoma (SCC VII) tumor models of head and neck cancer. Adult mice with SCC VII cells implanted were imaged using CE-MRI. Different stages of tumor growth were defined based on changes in the T1/T2-weighted MR imaging patterns. The contrast-enhancing (CE) and nonenhancing (NE) regions of the tumors were biopsied for microarray analysis and H&E staining. Tumors with no differential contrast enhancement were used as controls. We have found that distinct temporal stages of tumor progression can be defined using T1/T2-weighted CE-MRI and microarray analysis. The early stage tumors show a homogeneous CE pattern in the T1/T2 MR images with no significant differential gene expression from the center and periphery of the tumor. The more advanced tumors that show discrete CE region in the postcontrast T1-weighted MR images have drastic differential gene expression profiles from the CE and NE regions. The gene expression profiles of the late stage tumors continued to change as T2-weighted MR signal intensity increased. H&E showed that the samples obtained from the periphery and center of the early stage tumors, and the CE and NE regions from the more advanced stage tumors were similar. The region from the late stage tumors, however, showed characteristic signs of necrosis. The results demonstrate that temporal changes in T1- and T2-weighted CE-MRI are related to distinct gene expression profiles, and histological analysis may not be sufficient to detect these detailed changes. As tumors progress, discrete regions of postcontrast T1 enhancement are identified and these regions have distinct gene expression patterns despite similar histological features. In late-stage tumors, regions of T2 signal changes are observed, which correspond to tissue necrosis.
Y. Yang, None.
Abstract ID: 206
A novel approach to measure gene expression with submillimeter accuracy based on the sodium iodide symporter reporter gene, NIS, is presented. Cells infected with adenovirus incorporating NIS acquire the ability to take up iodide or pertechnetate ion and allow for thin-tissue autoradiographs in which optical density is a function of gene expression. Digitized autoradiographs are stacked and displayed in 3 dimensions to yield volumes of gene expression. By using standards of known activity, histograms of optical density are used to produce quantitative gene expression distributions. Quantitative improvements to standard adenovirus-based gene therapies, such as (1) the use of replicative versus replication-defective, (2) the use of increased viral dose, and (3) the use of improved vehicles can be measured and a head-to-head comparison with the standard preparation can be made. We present the results for each (n ≥ 3). Under sterile surgical conditions, 1 lobe of the dog prostate is administered the standard virus preparation (3 × 1011 Ad5-CD/TKrep-NIS viral particles, a replication competent virus that incorporates the 2 suicide genes cytosine deaminase and thymidine kinase). In the other prostate lobe, the test virus preparation is delivered. Days following viral delivery, dogs were administered a clinically relevant dose of radioisotope and following at least 2 hr of uptake, and prostates were removed, frozen, and processed for autoradiographs. An unexpected conclusion from the studies is that the spread of the virus was limited in the transverse direction by the connective tissue separating glands of the prostate effectively causing increases in viral spread to occur in the axial directions. The technique can be applied to any gene therapy approach that allows for incorporation of the NIS gene and provides approximately 85 μm precision in the in-plane field limited by the resolution of the film scanner (300 dpi) and 12 μm precision in the axial direction limited by the thickness of the frozen prostate section.
S.L. Brown, None.
Abstract ID: 207
To evaluate the capacity of human monocytes to phagocytose iron oxide based magnetic resonance (MR) contrast agents of different particle sizes, concentrations, and incubation times in order to provide an optimized depiction of the labeled cells with MR imaging. Human monocytes were isolated from blood samples and incubated in vitro with superparamagnetic iron oxide particles (SPIO): Ferumoxides (Endorem, Guerbet, France and Feridex, Advanced Magnetics, USA) and Ferucarbotran (Resovist, Schering AG, Germany), as well as ultrasmall SPIO (USPIO): Ferumoxtran (Sinerem, Guerbet and Combidex, Advanced Magnetics) and SHU555C (Resovist S, Schering AG). The influence of different particles sizes (SPIO vs. USPIO), varying iron oxide concentrations (5–2000 μg iron/ml) and of varying incubation times (1–24 hr) on cellular iron oxide uptake was investigated. The intracellular iron content was measured by atomic emission absorption spectrometry. In addition, MR signal intensities and T1 and T2 relaxation rates of labeled cells and nonlabeled controls were determined with a clinical 1.5 T MR scanner (Philips ACS Intera, The Best, Netherlands). A significantly higher cellular iron oxide uptake was found after incubation with SPIO compared to USPIO (p < .05). The ionic SPIO Ferucarbotran revealed the highest uptake of all iron oxides. In vitro incubations of the cells with contrast agent concentrations, which exceeded those reached in plasma after intravenous injection of clinical doses, increased the cellular iron uptake significantly (p < .05), and these differences could be depicted and quantified with MR. However, incubation with high iron oxide concentrations (Ferumoxides > 100 μg iron/ml and Ferucarbotran > 500 μg iron/ml) increased apoptosis of monocytes. SPIO are better suited than USPIO for labeling of human monocytes. The capacity of human monocytes to phagocytose iron oxides is more than 10 times higher than the in vivo plasma concentration allows. Therefore, an ex vivo labeling prior to injection may improve the in vivo depiction of human monocytes with MR.
S. Metz, None.
Abstract ID: 208
Prostatic secretory function and ductal morphology are lost in prostate cancer. Dynamic contrast-enhanced MR imaging (DCE MRI) shows higher enhancement in cancer versus normal tissues, attributed to larger extravascular-extracellular space in which the contrast agent accumulates. The MR apparent diffusion coefficient (ADC) is lower in cancer versus normal, perhaps due to less free space from increased cancer cellularity. MR spectroscopic imaging (MRSI) shows decreased citrate in cancer due to ductal and secretory function losses. This work hypothesizes that Gadolinium-DTPA does not penetrate normal prostatic ducts, which are surrounded by epithelial and basal cell layers. Thus, DCE MRI will not reflect ductal volume, whereas ADC will. Twelve untreated patients were scanned at 1.5 T. The DCE MRI used a 5-slice, 2-D-FSPGR with a single-dose bolus injection of Gd-DTPA at 8 sec/stack, spanning >9 min. The diffusion used a single-shot FSE with a b value = 600. ROIs were drawn on coil-corrected, aligned, T2-weighted images in normal peripheral zone, cancerous peripheral zone, central gland, stromal benign prostatic hyperplasia (BPH), and glandular BPH, as determined by MRI, MRSI, and, when available, biopsy. Peak enhancements and the ADCs were calculated and compared. While taken individually, the results confirm previous findings. When combined, the peak enhancement/ADC values completely separated the ductal tissues from the low-ductal tissues; all normal peripheral zone and glandular BPH (ductal) tissues were <0.096, while all central gland and stromal BPH tissues were >0.096. Within patients, central gland always had a higher peak and lower ADC versus normal peripheral zone. The ductal tissues all had high citrate, whereas the low-ductal tissues had low citrate. Cancer values varied, perhaps due to their varying amounts of ducts. The results support the theory that Gd-DTPA does not penetrate the normal prostatic ducts and that the combined DCE MRI/ADC can reflect ducts important for assessing prostatic function and cancer aggressiveness.
S.M. Noworolski, None.
Abstract ID: 209
M. Doubrovin, None.
Poster Session 05: Combination Instruments and Multifunctional Probes
Abstract ID: 210
One of the major limitations of multiphoton excited fluorescence microscopy is the low image rate due to point-by-point scanning. We have recently developed a novel multifocal multiphoton microscope based on a beam divider, which allows imaging at frame rates up to 3.5 kHz. We report on the working principle and characteristics of the beam divider that splits an incoming laser beam into an array of up to 256 beamlets, which are scanned simultaneously in the focal plane. The beam divider is based on exclusively flat optics, is free from aberrations, and has >90% light throughput. Due to intrinsic time multiplexing of the beamlets, there is no cross-talk. The technology reported here has several advantages over the alternative microlens array-based setups. The beam divider can be used either in single point or line scan mode. The line scan mode allows convenient coupling to spectrographs enabling 3-D spectral sectioning. Applications are high-resolution 3-D imaging of intracellular dynamics, particle tracking, anisotropy imaging, calcium signaling, vesicular trafficking, and protein-protein interactions. Several representative data obtained using this technology will be presented.
R. Ahuja, None.
Abstract ID: 211
Multimodality imaging is increasingly required for adequate diagnosis and treatment. In surgery, techniques for reliable intraoperative imaging and image-guided dissection are quite limited. Our goal was to provide the surgeon with real-time, intraoperative imaging using optical, noninvasive technology. Such a system would allow instantaneous imaging of organ vasculature, sentinel lymph nodes, and regional organ ischemia. Additionally, this system would permit direct, real-time image guidance in surgical localization and dissection. Invisible near-infrared (NIR) light penetrates relatively deeply into living tissue, and background autofluorescence in the NIR region is extremely low. We have previously proven the principle of using NIR light for intraoperative imaging in small animal model systems. We have now developed a portable, intraoperative NIR fluorescence system for use in large animal surgery. The system has a 20-cm-diameter field of view and maintains an NIR excitation fluence rate of 5 mW/cm2. It also simultaneously displays color video and NIR fluorescence images of the surgical field. By introducing exogenous, nonisotopic NIR fluorophores, anatomy and function can be visualized in real time. In porcine model systems, we demonstrate that several polysulfonated heptamethine indocyanines, including the FDA-approved agent indocyanine green (CardioGreen(tm)), permit real-time assessment of coronary, small bowel, gastric, and colonic vasculature, as well as delineate areas of regional organ ischemia. For each fluorophore, we determined its organ uptake, clearance, arteriovenous gradient, and vascular adherence properties. In a variety of simulated vascular occlusion models, we obtained high-resolution images reliably documenting impaired blood flow. Furthermore, in a subset of animals, we confirmed imaging results with colored microsphere analysis in regions of ischemia distal to a vascular occlusion. This novel, noninvasive, intraoperative NIR fluorescence imaging system is capable of providing the surgeon with highly sensitive intraoperative images and real-time visual feedback to complement surgical planning, approach, and dissection.
J.V. Frangioni, None.
Abstract ID: 212
Introduction of suicide gene, herpes simplex virus type 1 thymidine kinase (HSV1-tk), in tumor cells has provided a useful method for tumor gene therapy and imaging gene expression [1]. Several
S. Choi, None.
Abstract ID: 213
Folate receptor is a glycosylphosphatidylinositol (GPI)-anchored, folate binding protein overexpressed in various types of human tumors. It functions by capturing folate to feed rapidly dividing cancer cells. Thus, folate receptors provide highly selective sites that differentiate tumor cells. The folate conjugates are delivered into the cell via endocytosis, in which a folate molecule bound to the receptor is first delivered into the cell as a group, then the folate is released and the receptor is recycled back to the surface to capture more folate. In this study, we report a method to prepare nanometer-sized fluorescent superparamagnetic folate-conjugated iron oxide (IO) colloids for targeted delivery to folate-receptor-expressing cells (KB). In the present study, we employed nanometer-sized hydrophilic dextran-coated iron oxide particles as bare colloids to ensure long circulation, and we subsequently tethered folic acid and fluorescent imaging agents onto the iron oxide surface via coupling reactions. The particle size was measured using light scattering photon correlation spectroscopy. Particle size distributions of colloids after different stages of surface conjugation showed that folate-FITC–IO colloid has a slightly broader distribution and larger size (26 nm) than NH2-conjugated IO colloid (22 nm). The magnetization curve exhibits a superparamagnetic behavior that is highly desirable for magnetic imaging. Flow cytometry analysis indicated incorporation of folate-FITC–IO colloids approximately 97% and 17% into the KB and A549 (folate-receptor negative) cells, respectively. FITC-IO colloids showed 6–8% uptakes in both cells. The resultant nanoparticle colloids are readily internalized into folate-receptor-expressing tumor cells, thus demonstrating their suitability for targeted delivery and diagnostic imaging.
S. Choi, None.
Abstract ID: 215
The use of light-emitting probes, such as firefly luciferase or fluorescent proteins as reporters of gene expression in living cells, is a well-established technique for the study of biological activity. Both bioluminescent and fluorescent reporters can be used for in vivo imaging, and each has different biological applications. The ability to track light-emitting cells in live laboratory animals has a wide range of applications in biomedical research. At Xenogen Corporation (Alameda, CA), instrumentation to optimize the sensitivity of detection for both types of reporters is being developed. Upgrades to the existing IVIS™ Imaging System Series 100 technology allow for the detection of bioluminescence and continuous wave fluorescence in the wavelength range of 400–950 nm. The system is absolutely calibrated, allowing for a quantitative comparison of bioluminescent and fluorescent signals in the same instrument. Background tissue luminescence and autofluorescence limit the detection sensitivity for both bioluminescent and fluorescent reporters. Detection of bioluminescence has an advantage because it does not require an external light source for excitation; therefore, the background emission levels are significantly lower. Even though fluorescent probes can be several orders of magnitude brighter than bioluminescent probes, autofluorescence makes it difficult to benefit from their brighter signal levels. Therefore, without background autofluorescent subtraction techniques, bioluminescent probes tend to have better detection sensitivity. In this work, the dual imaging system is presented. In vitro and in vivo comparisons of cell lines tagged with both bioluminescent and fluorescent reporters are shown. Background bioluminescent and fluorescent emission data, and techniques to reduce autofluorescence background using background filters, are presented.
T.L. Troy, None.
Abstract ID: 217
V. Ponomarev, None.
Abstract ID: 218
Bone metastases are a significant cause of morbidity in cancer patients. By the time these metastases are detectable by diagnostic imaging they often cause great pain, and treatment regimes are often too late to do significant good. If imaging tests can be developed to detect these lesions at an earlier stage, it is possible that treatments can be initiated that will alleviate suffering. We have developed a novel in vivo bone invasion model using human neuroblastoma cells in nude mice and begun to develop a potential therapeutic regime using bisphosphonate drugs, which inhibit the formation of lesions. We used micro-PET and high-resolution radiographic imaging to begin to define the imaging modalities and parameters that allow the earliest detection of these lesions. We injected human neuroblastoma cells into the femurs of nude mice, which generated radiographically apparent osteolytic lesions in 2–3 weeks in untreated animals. Treatment delayed osteolytic lesion development. Radiographic studies were conducted on treated and untreated mice over time. Radiographs were graded according to a 4-point scale, and histology performed to validate the imaging assessment. Initial micro-PET FDG and fluorine studies were done on one treated mouse to compare sensitivity of PET to radiography for lesion detection. Histologic and radiographic grades were in good agreement for high-grade lesions, as expected, but relatively poor for low-grade lesions. The micro-PET FDG study showed increased signal in the region of tumor growth although this lesion was not radiographically apparent. Our model will be useful for assessing treatment and establishing early means of imaging detection and characterization. Further PET studies need to be performed to fully characterize the potential of this modality for this use.
Preliminary PET imaging data. Image (A) is mouse R-45 with FDG imaging agent. Note strong signal from right femur. (Arrow), where tumor cells are implanted, compared to left femur. Image (B) was made of the same mouse using a Flourine imaging agent. Note increased signal in metaphysial areas of the joints (arrows). This mouse was treated with the bis-phosphonate drug, and had no evidence of lytic bone lesion formation on plane film radiography as seen in image (C). The arrow in image (C) is on the same side as the arrow in image A. This preliminary result shows the utility of PET imaging to detect the presence of tumor cells in out orthotopic implantation model.
R.A. Moats, None.
Poster Session 06: Advancing Molecular Therapies through Imaging: Genetic and Cell-based Therapies
Abstract ID: 219
Elevated phosphocholine (PC) and total choline (tCho) levels typically detected in cancer cells have been linked to malignant transformation, invasion, and metastasis. Increased choline kinase (CK) activity and expression is correlated to increased malignancy qualifying it as a unique target for cancer cells. We developed a small interfering RNA (siRNA)-based approach to down-regulate CK. This novel technique is a powerful tool for knocking down gene expression of specific proteins. The effect of siRNA specific for CK (siRNA-ck) on PC and tCho levels in malignant and nonmalignant human mammary epithelial cells was monitored using 1H NMR spectroscopy. MDA-MB-231 human breast cancer cells and nonmalignant human mammary epithelial cells (HMECs) MCF-12A were treated with siRNA-ck. Reverse transcription-polymerase chain reaction (RT-PCR) analysis of CK mRNA was performed following siRNA-ck treatment. Water-soluble as well as lipid extracts were obtained from control and siRNA-ck-treated cells and analyzed using 1H NMR spectroscopy. RT-PCR analysis of CK mRNA following 48 hr siRNA-ck treatment resulted in a threefold reduction of CK mRNA in MDA-MB-231 cells. High CK mRNA levels were detected in untreated MDA-MB-231 cells compared with untreated MCF-12A cells. SiRNA-ck treatment of MDA-MB-231 cells significantly decreased PC and tCho levels by 38% and 53%, respectively. No significant changes in CK mRNA levels and choline phospholipid metabolites were detected after siRNA-ck treatment in nonmalignant MCF-12A cells. A significant increase of the triglyceride signal was observed in MDA-MD-231 and MCF-12A cells. These data demonstrate that CK expression mediates the increased PC and tCho levels observed in breast cancer cells. The increase of the triglyceride signal is consistent with the possibility of inducing differentiation in these cells by inhibiting CK. 1H MRSI is uniquely useful in developing and applying cancer therapies utilizing siRNA technology to target this critical enzyme.
K. Glunde, None.
Abstract ID: 220
Regulated gene expression can be achieved by use of a bidirectional tetracycline-regulated promoter system (Gossen et al., 1992). Bidirectional controlled gene expression should serve proportional co-expression of a PET marker gene and any gene of interest. To investigate whether noninvasive assessment of regulated expression of a PET marker gene is possible. The PET marker gene HSV-1-tk39 (Black et al., 1996) was introduced into an existing HSV/EBV/TET-amplicon vector. The vector HET-6-tk39, consists of the HSV-1 amplicon backbone, the EBV orip, a mutant EBNA-1 gene (Sena-Esteves et al., 1999), and the Tet-responsive element. To analyze regulation, human Gli36ΔEGFR glioma cells were infected with HET-6-tk39 at an MOI = 0.3 ± doxycycline. 24–72 hr after infection RFP-positive cells (RPC) were counted and the ratio ± doxycycline was determined. A single-cell analysis of relative RFP-expression per cell was performed by means of MPI-Tool imaging software (Jacobs et al., 2003). [18F]FHBG-PET imaging was performed to correlate quantitative data in culture and in vivo. In cell culture, the mean ratio of RPC with and without doxycycline was 494 ± 213, ranging from 145- to 796-fold induction depending on the duration of induction. The relative level of RFP-fluorescence per cell was 196 ± 55 relative fluorescent units (RFU) in the presence and 22 ± 6 RFU in the absence of doxycycline, representing a 9-fold induction with respect to RFP-fluorescence intensity per single cell. Preliminary data in vivo demonstrate that HET-6-tk39 regulated gene expression can be assessed by [18F]FHBG-PET. Proportionality between quantitative inducible gene expression in cell culture and in vivo is currently being assessed. An HET-amplicon vector was engineered to regulate the expression of a PET marker gene and a proportionally co-expressed gene of interest over a wide range of expression levels. This type of HSV-1 amplicon vector should serve noninvasive assessment of any co-regulated therapeutic gene of interest substituting the rfp gene.
A. Winkeler, None.
Abstract ID: 221
MRI can be used as a tool to monitor iron-labeled stem cell engraftment. We test the hypothesis that in an allogeneic transplantation model, mesenchymal stem cells labeled with micron scale iron particles preferentially home to myocardial injury after intravenous administration and undergo site-specific differentiation. Myocardial infarction was induced in 10 Sprague-Dawley rats by surgical occlusion of the left coronary artery followed by reperfusion and intravenous administration of either allogeneic iron-labeled mesenchymal stem cells derived from ACI rats (n = 5) or free iron particles (n = 5). Serial cardiac MRI exams were performed at 1 day, 5 days, and 6 weeks. The presence of the iron-labeled stem cells was confirmed using confocal microscopy and immunohistochemistry for markers of cardiomyogenic differentiation. The total number of cells was counted in both infarcted and noninfarcted tissue and compared to the total number of engrafted IFP-labeled cells. Iron-labeled mesenchymal stem cells were detected using MRI at 1 day, 5 days, and 6 weeks in 4/5 of the treated animals. IFP-labeled MSCs preferentially home to the site of myocardial injury, with a significantly higher labeled cell to total cell ratio within the infarct zone compared to normal myocardium (0.62 vs. 0.16, p <.001). At 6 weeks, the engrafted mesenchymal stem cells expressed sarcomeric and gap junction proteins. This study demonstrates the feasibility of using iron fluorophores combined with MRI to track the homing and preferential engraftment of allogeneic mesenchymal stem cells as they traffick to sites of tissue injury and undergo milieu specific differentiation.
J.M. Sorger, Osiris Therapeutics 6.
Abstract ID: 222
Evaluation of gene expression changes of muscle tissue cells after treatment with focused ultrasound (FUS) using an optical imaging system. MRI determination of the tissue changes using a diffusion-weighted STEAM sequence. Muscle tissue of 25 Balb/c mice were transfected with a plasmid containing the CMV-promoter and Luciferase as reporter gene. Ten days after the transfection FUS (Focus Surgery, Indianapolis) was applied for 20 sec in a continuous wave mode using five different energy levels (4133, 3067, 2157, 1401, 802 W/cm2). After the application of FUS, an MRI scan (1.5 T GE Signa) of all animals using a diffusion-weighted STEAM sequence (TR = 6000 msec, TE = 30 msec, NEX = 8, ä = 9 msec, Δ = 170 msec, b values = 50, 250, 500, 750, 1000 [s/mm2]) was performed. Gene expression activity over a time period of 2 weeks was determined by using an optical imaging system (Xenogen). The diffusion coefficient values and the T2 times varied depending upon the applied energy level. At an energy level of 4133 W/cm2, the mean diffusion coefficients of the treated tissue was D = (1.30 ± 0.03) × 10−3 mm2/sec and the mean T2 time 117.49 ± 6.74 msec showing a 83% reduction of Luciferase reporter gene activity. However, at an energy level of 2157 W/cm2, the diffusion coefficient had a mean value of D = (0.98 ± 0.03) × 10−3 mm2/sec and the tissue a mean T2 time of 63.91 ± 3.90 msec, an increase gene expression of 410% was seen. At an energy level of 802 W/cm2, the mean diffusion coeffient was D = (0.84 ± 0.04) × 10−3 mm2/sec and the T2 = 46.34 ± 1.18 msec still leading to a higher gene expression of 250% compared to the reporter gene activity prior to the application of FUS [D = (0.68 ± 0.04) × 10−3 mm2/sec; T2 = (36.3403 ± 1.1827) msec]. FUS can induce gene expression in muscle tissue. The diffusion coefficient D (mm2/sec) and the T2 time (msec) of muscle tissue correlates with the reporter gene activity of the transfected muscle tissue.
W. Hundt, None.
Abstract ID: 223
One of the limitations of early-phase clinical trials in gene therapy is the difficulty in detecting the level and duration of gene expression on administration of the delivery vector. In this study, we investigated the application of human sodium iodide symporter (hNIS) as a reporter gene to image in vivo gene transfer and expression to normal tissues in real time. For the lung imaging, 3 × 108 PFU Ad-hNIS or Ad-Bgl II was administered intratracheally to Cotton rats twice within a 48-hr time interval. Forty-eight hours after the second infection 1 mCi 99mTcO4− was injected intraperitoneally. Images were acquired by gamma camera scintigraphy at different time points following administration of the tracer. One hour after injection of the radiotracer, the rats' individual lungs were clearly visible in the images from Ad-hNIS infected animals while those of Ad-Bgl II infected animal were undetectable. We also wished to determine the duration of gene expression in the adenovirally infected pulmonary tissue. Our results also showed that after 14 days we still could detect signal from the lungs of the Ad-hNIS infected animals. For the study of gene transfer to the skeletal muscle, 2 × 108 PFU Ad-hNIS or Ad-Bgl II was injected intramascularly on contralateral legs of a Fisher rat. Forty-eight hours after infection, 1 mCi 99mTcO4−− was administrated via tail vein. The images were acquired by gamma camera scintigrapghy at different time points. The Ad-hNIS infected skeletal muscle cells rapidly accumulated pertechnetate, and showed much more activity than that of Ad-Bgl II infected cells in the opposite leg. NIS is a highly effective reporter gene in lung tissue and skeletal muscle cells. After infection with Ad-hNIS, these tissues could be visualized noninvasively in real time. It is possible to monitor the distribution and duration of gene expression using hNIS. (Supported by NIH grant CA91709.).
M.M. Graham, None.
Abstract ID: 224
Recent studies demonstrate the feasibility of therapetic radionuclide uptake enhancement by gene therapy, using the sodium-iodide symporter (NIS) gene, and I-131 in cancer cells. p53 is one of the major determinants of intrinsic cellular sensitivity to the cytotoxic effects of ionizing radiation. To evaluate the role of p53 in enhanced radioiodine gene therapy using the NIS gene, we investigated the cytotoxic effect of the NIS and p53 genes by cotransfection into cancer cells. The NIS gene was transfected to a wild-type (wt) p53 expressing cancer cell line (SK-Hep1), and a mutant (mt) p53 expressing cancer cell line (ARO), using liposomes. A recombinant adenovirus-5 vector containing a CMV promoter and wt p53 cDNA was used to facilitate p53 expression. The cells were treated with I-131 and Re-188, and the survival rate of each cell line was determined using a clonogenic assay. The NIS-transfected SK-Hep1-N and ARO-N cells accumulated up to 150 and 109 times more radioiodine than did nontransfected cells. The survival rate of SK-Hep1-N cells treated with I-131 (0.5 mCi/5 ml) and Re-188 (1 mCi/0.5 ml) were 46.3% and 24.4%, respectively. The survival rate of untransfected SK-Hep1 cells treated with I-131 and Re-188 was not changed compared to untreated cells. There was no difference between the survival rates of untransfected mt p53 expressing ARO cells and NIS-transfected ARO-N cells after treatment with these radionuclides, suggesting a role of p53 in radionuclide sensitivity mediated by NIS. SK-Hep1-N cells, treated with I-131 and Re-188, expressed 3.9 and 4.7 times more p53 than nontreated cells. When ARO-N cells containing wt p53 were treated with radionuclides, apoptotic cells increased to three times compared to ARO-N. Thus, the NIS gene, transfected into cancer cell lines, increased susceptibility to beta-emitter radionuclides, and cotransfection with a wt p53 gene enhanced the cytotoxic effect.
J. Lee, None.
Abstract ID: 225
Monogenic, inflammatory, and other skin diseases have the potential to be treated using gene therapy techniques. However, expression of therapeutic genes in skin is typically lost within 7 days and may be insufficient for treatment. We are investigating systems for sustained delivery of plasmid DNA to prolong gene expression. A gene delivery system consisting of agarose hydrogel and compacted DNA has been developed, and the gene encoding firefly luciferase is being used to evaluate the in vivo efficacy of the delivery system. The conventional method to assess luciferase activity uses tissue lysates and a luminometer to measure light production. Alternatively, bioluminescent imaging (BLI) can be performed on a live animal and does not require tissue harvesting. In this study, BLI was validated against luminometer measurements and used to monitor the time course of expression from intradermal injections of compacted DNA either in solution or in a hydrogel delivery system. When the luminometer and BLI were compared using solutions of 5 to 500 pg/ml of recombinant luciferase, a linear relationship was found. The systems were then compared using rodents injected intradermally with 1 to 50 mg of compacted luciferase plasmid. For luminometer measurements, tissue was harvested, crushed, and lysed. For BLI, the rodents were anesthetized, injected with luciferin, and imaged. Both systems were able to detect the luciferase activity, but BLI was shown to be much more effective. With BLI, results can be obtained within minutes without the lengthy process of preparing tissue, and the time course of gene expression can be studied in the same animal. BLI results have demonstrated that with compacted DNA in the hydrogel, significant gene expression in the skin has been measured beyond the 7 days previously reported. (This research was funded by the John Glenn Biomedical Engineering Consortium and by NSF grant CAREER-BES9874631.)
P.J. Cheung, None.
Abstract ID: 226
Many approaches, such as high-intensity focused ultrasound (HIFU), have been used with variable levels of success in increasing transfection when coupled with a direct intratumor injection. Pulsed HIFU, which deposits mainly mechanical rather than thermal energy in tissues, has been shown to increase endothelial permeability, intraparenchymal diffusion of macromolecules, and cell membrane permeability. We hypothesized that pulsed HIFU can be used to increase the delivery and expression of plasmid DNA, in tumors, using an intravenous injection in mice. Female CH3 mice were injected subcutaneously with SCC7 tumor cells in both flanks. Cells were allowed to grow until the tumors were 1.0 cm in diameter. The mice were treated with pulsed HIFU delivered by a 1-MHz transducer set at 20.5 W or a total acoustic power of 1380 W/cm2 per focal zone. Tail vein injection of 5 μg/g of endotoxin free plasmid DNA containing a CMV promoter and GFP genes was administered to the mice. The mice were returned to their cages and kept for 24 and 72 hr to allow for GFP expression. The mice were sacrificed and observed histologically for plasmid transfection through the expression of the GFP. Western blot analysis was performed on the tumor tissues to quantitate GFP expression. All mice showed a significant increase in protein expression in the HIFU-treated tumors, at both time points, compared to control tumors, which demonstrated negligible GFP expression. Histological evaluation of the liver, spleen, and kidney showed no tissue damage and minimal GFP expression. In conclusion, treating tumors with pulsed HIFU before the intravenous injection of plasmid DNA increases transfection and expression of desired proteins in targeted tumor cells. This minimally invasive form of targeted delivery can play an important role in the delivery of macromolecular agents to tissues of interest.
K.M. Dittmar, None.
Abstract ID: 227
V. Ponomarev, None.
Abstract ID: 228
Studies in plasminogen activator inhibitor type-1 (PAI-1) null mice have disclosed a requirement for this protease inhibitor in tumor invasion and angiogenesis. PAI-1 is the major regulator of urokinase PA and, thereby, of plasmin generation. PAI-1 affects tumor cell and endothelial cell motility by controlling the extent and localization of matrix barrier proteolysis as well as by influencing cell-to-matrix adhesion mediated by integrins and the urokinase PA receptor. To evaluate relationships between PAI-1 expression and cellular migratory activity, the proximal 806 bp of the human PAI-1 promoter were cloned upstream of an insert encoding a PAI-1-GFP chimeric protein. De novo synthesized PAI-1-GFP was visualized in situ by fluorescence microscopy. Gene activation and the subsequent deposition of PAI-1-GFP into cellular migration tracks was assessed using a planar (2-D) scrape wound assay to initiate PAI-1 transcription and cell motility. Comparative analysis of PAI-1-GFP chimera expression with that of the endogenous PA1–1 gene (at the mRNA and protein levels) revealed similar kinetics indicating that the PAI-1-GFP construct provided an easily accessible “reporter” system to evaluate PAI-1 promoter function within the real time of gene activation. Such PAI-1 expressing cells were highly motile in the 2-D planar model of induced migration. The basal and TGF-beta-stimulated migration rates of PAI-1-GFP transfectants were significantly greater that control GFP-only transfectants under the respective culture conditions. Both responses could be attenuated by incubation with PAI-1 neutralizing antibodies or by secondary transfection with PAI-1 antisense constructs confirming the importance of PAI-1 in cell locomotion. These findings illustrate the usefulness of this approach to image transcriptional and behavioral response in single living cells using GFP reporters under control of promoter sequences from genes implicated in cellular invasive traits. (This work was supported by grants from the NIH (GM57242) and the Department of the Army (DAMD17-98-1-8015 and DAMD17-00-1-0124).)
P.J. Higgins, None.
Abstract ID: 229
Cell surface receptors offer a unique avenue for targeted delivery of diagnostic and therapeutic agents. In particular, considering the fact that rapidly proliferating tumors overexpress transferrin receptors (TfR). Glioma can be one of many examples where such receptor-targeted agents are of critical importance. In this regard, the most essential limitations of IgG-based receptor targeting in vivo include steric hindrance, nonspecific binding, and reduced affinity of derivatives. Recently, bioengineered recombinant single-chain variable fragment (scFv) antibodies open a new frontier in image-guided therapy involving receptor targeting (Malecki et al., 2002, PNAS 99: 213–219). The objective of this work is to demonstrate the feasibility of using scFv-based MRI contrast agents to create signal changes in a clinically relevant model of tumor receptor targeting. The anti-TfR antibodies were bioengineered by preparation of the libraries, expression in yeast Pichia pastoris, followed by selection through biopanning and affinity purification. The efficiency of labeling of the TfR was tested in C6 glioma and fibroblast cell cultures. Affinity was verified through Western blotting and immunofluorescence. Sub-cellular localization of Gd derivatives of IgG, Fab, and Fv was determined with atomic resolution by means of electron energy loss spectral imaging (EEL-SI). Qualitative and quantitative differences in image contrast were demonstrated between the C6 glioma and fibroblast cells on inversion-recovery T1 weighted imaging following labeling. Measurements of T1 during magnetic resonance imaging at 1.5 T of cells labeled with our scFv antibodies indicated significant enhancement of proton relaxation times in TfR overexpressing gliomas (1443 msec) relative to normal fibroblasts (1587 msec), and unlabeled cells (1750 msec). We demonstrated that scFv antibodies bioengineered with Gd binding domains are capable of labeling glioma cells in vitro. These scFv-based receptor targeting contrast agents also create a clinically relevant change in relaxivity by MRI at 1.5 T.
M. Malecki, None.
Poster Session 07: Drug Discovery and Development
Abstract ID: 230
The power of multivalent binding in biological systems is a familiar concept that holds great potential for drug design (for a review, see Whitesides et al., 1998. Angew. Chem. Int. Ed. 37:2754–2794). In theory, targeting cell surface receptors with multivalent ligands can improve the specificity and affinity of the drug for a given cellular target. We have previously shown that dimerization of cyclic RGD peptide antagonists to the integrin avb3 (vitronectin receptor) provides higher levels of tumor uptake and retention (Rajopadhye et al., 2000. J. Nucl. Med. 41:259P). This dimer effect was also noted in more recent work with peptidomimetic avb3 antagonists based on the quinolone carboxamide core (Harris et al., in press. Cancer Biother. Radiopharm.). In this study, we have investigated the source of this dimer effect in a series of eight quinolone carboxamide dimers having the general structure shown below, where Csa is the pharmacokinetic modifier cysteic acid. Changing the number of Csa and PEG repeat units provided us with dimers having spacer lengths of 24,112 atoms. The dimers were derivatized with DOTA for labeling with In-111 and evaluated in the c-Neu Oncomouse tumor model. All dimers displayed high affinity for integrin avb3, and good selectivity relative to integrin aIIbb3. Excretion was predominantly renal. In contrast to the monomers, all dimers showed an increase in tumor uptake between 2 and 24 hr PI. At 24 hr PI, the tumor uptake of all dimers was superior to the quinolone monomers, exceeding 14% ID/g for two of the dimers. The potential of these dimers as tumor imaging and radiotherapeutic agents will be discussed.
T.D. Harris, None.
Abstract ID: 231
Metalloporphyrins (Mps) potently inhibit heme oxygenase (HO), the rate-limiting step in the catabolism of heme to bilirubin. This enzyme is therefore a logical target for chemoprevention of neonatal jaundice. However, some Mps increase transcription of the inducible HO isozyme (HO-1) and may counteract clinical utility. Our objective was to study the intestinal absorption of selected Mps and then assess their effects on in vivo HO-1 gene expression in target and nontarget tissues of adult HO-1-luc transgenic mice using bioluminescent imaging (BLI). 30 μmol/kg of tin mesoporphyrin (SnMP), zinc protoporphyrin (ZnPP), zinc bis glycol porphyrin (ZnBG) or vehicle (control) was directly injected into the surgically exposed proximal small intestine. HO-1 promoter activity was assessed, in vivo, at various time points via BLI). HO activity inhibition (measure of efficacy) was assessed by measuring in vivo heme degradation rates (VeCO) as well as HO enzyme activity levels in excised tissues (spleen, liver, and intestine). After intestinal SnMP administration, HO-1 transcription increased up to 4-fold by 48 hr. In contrast, ZnPP and ZnBG minimally affected transcription with a return to baseline by 6 hr. VeCO levels decreased 20% only after SnMP administration and lasted beyond 48 hr. Splenic HO activity was not inhibited by any Mp. Liver HO activity was significantly inhibited up to 60% and lasted beyond 48 hr after SnMP administration, and up to 30% after 3 hr of ZnBG administration with a return to baseline by 24 hr. Intestinal HO activity was inhibited after administration of each Mp. In summary, luminally injected Mps are absorbed from the intestine at different rates and affect HO-1 gene expression and enzyme activity in a tissue- and time-dependent manner. Therefore, we conclude that BLI can assess HO-1 gene expression and is a useful tool to evaluate safe and efficacious compounds for controlling HO-1 activity with utility in the treatment of pathologic jaundice.
R.J. Wong, None.
Abstract ID: 232
Previously, we described a transgenic mouse model [FVB/N-Tg(CYP3A4-luc)Xen] using a reporter construct consisting of 13 kilobases (kb) of the human CYP3A4 promoter driving the firefly luciferase gene in FVB/N mouse background. Here we report regulation of this same CYP3A4-luc reporter gene in a transgenic rat, and in a transgenic mouse with a different genetic background. Basal reporter expression and responses to several xenobiotics in the transgenic CD-1 mice [Crl:CD-1(ICR)BR-TgN(CYP3A4-luc)Xen] were similar to those in the transgenic FVB/N mice with greater induction in male compared with female mice. In contrast to the mouse models, transgenic rats [Tac:N(SD)-TgN(CYP3A4-luc)Xen] basal expression of the luciferase reporter was higher in females than in males. Responses to some compounds were much greater in the rat than in mouse and the kinetics of induction was different from with the peak at 12 hr following drug injection compared with a peak at 6 hr in mouse. In transgenic rats, PCN, clotrimazole, and dexamethasone induced the reporter strongly; while phenobarbital and nifedipine were moderate inducers. Rifampicin and pregnenolone had little effect in rat or mice. Our results suggest that the human CYP3A4 promoter functions differently in transgenic mice and rats, confirming the species differences in regulation of drug metabolizing genes.
W. Zhang, Xenogen Corp. 5.
Abstract ID: 233
The purpose of this study was to identify peptides from a phage peptide library which bind with high affinity to a mutant EGF receptor, known to be present in a number of cancers. Using phage libraries, available commercially, in vivo selection was performed in mice with tumors in one thigh that express the mutant EGF receptor. Elution strategies were incorporated to remove phage of lower binding affinity, thus to select a peptide with high affinity to this mutant receptor. A phage library expressing a diversity of 109 linear 12-mer peptides was administered intravenously to mice bearing tumors, about 1011 phage per mouse. After 2 hr the tumors were removed and homogenized into a single cell suspension in PBS on ice. To remove loosely bound, low-affinity phage, the cell suspension was incubated three times in PBS for 15 min and pelleted. Then 0.2 M glycine pH 2.2 was added for 2 min, removed and a second aliquot was added for 5 min to remove more tightly bound phage. These phage were reinjected and after five rounds of in vivo selection a common peptide appeared in round 4 and again in round 5. The peptide was synthesized with incorporation of a Gly-Gly–Gly-Cys for radiolabeling with 99mTc in preparation for imaging studies. Screening of this 99mTc-peptide against human tumors demonstrated higher activity bound (significant in 4 out of 7, p ≤ .05) than to a phage peptide selected by standard procedures. When administered to mice the tumor to muscle ratio for this 99mTc-peptide was 4.6 versus 2.4 (n = 3) for the standard peptide. This study shows that one can obtain peptides with enhanced binding by altering elution strategies. The peptides described in this study may be useful as agents in the diagnosis of cancers and as agents for specific delivery of radionuclides to cancer cells for radiotherapy.
M. Rusckowski, None.
Abstract ID: 234
Heme oxygenase (HO) catabolizes heme to produce iron, carbon monoxide (CO), and biliverdin, which is immediately reduced to form bilirubin. Biliverdin and bilirubin are known potent antioxidants. Therefore, the induction of HO-1 is thought to be protective against oxidative stress and, recently, is believed to be also involved in antiinflammatory mechanisms. Here, we investigate the effect of dexamethasone (DEX), a glucocorticoid widely used for the treatment of inflammatory diseases, on basal HO-1 expression and induction of HO-1 transcription by cadmium chloride (CdCl2), a known transcriptional HO-1 inducer and source of oxidative stress. A mouse fibroblast cell line (NIH3T3), expressing luciferase from the full-length mouse HO-1 promoter (HO-1-luc), was treated with DEX (1 μM), CdCl2 (10 μM), or both. Transgenic mice (FVB), expressing HO-1-luc, were injected intraperitoneally with DEX (1 mg/kg) or CdCl2 (20 μmol/kg), or both. HO-1 promoter transcriptional activity was assessed by luciferase activity using bioluminescent imaging. HO-1 protein and total HO enzyme activity of cell lysates or liver sonicates were assessed by ELISA and by CO production using gas chromatography, respectively. In vitro results showed that DEX treatment alone did not affect HO-1 expression, HO-1 protein, nor HO activity. Twenty-four hours post-CdCl2 treatment, HO-1 expression, HO-1 protein, and HO activity increased 4.6-, 1.8-, and 2.5-fold, respectively. Treatment with DEX and CdCl2 together significantly decreased HO-1 induction observed with CdCl2 alone. In vivo results showed that DEX treatment alone did not affect HO-1 expression. Twelve hours post-CdCl2 treatment, HO-1 expression, HO-1 protein, and HO activity increased 449-, 1.5-, and 4.0-fold, respectively. Treatment with DEX and CdCl2 together significantly decreased HO-1 induction observed with CdCl2 treatment alone. Treatment with DEX suppressed the HO-1 induction observed with CdCl2 treatment both in vitro and in vivo and suggests that it may attenuate HO activity and consequently, reduce the protective effects of HO-1 induction under conditions of oxidative stress.
M. Mizobuchi, None.
Abstract ID: 235
Transgenic male mice expressing luciferase (luc) under the control of the kidney androgen regulated protein (KAP) promoter were treated daily with anti-androgenic compounds or vehicle for 15 days. The test compounds, cyproterone acetate, vinclozolin, and flutamide, were administered subcutaneously over a range of doses to determine the sensitivity of the KAP-luc reporter to changes in androgen status. Endpoints evaluated included in vivo biophotonic imaging, final body weight, organ weights (liver, kidney, testes, epididymies, accessory sex gland unit [ASG], and seminal vesicles), protein luciferase assays, and serum hormone concentrations. Biophotonic images were acquired from each animal throughout the experiment using a highly light sensitive low-light imaging system (IVIS Imaging System). These imaging results were correlated with traditional endpoints of body and organ weights as well as serum hormone concentrations. KAP-luc signals were observed at high levels in kidneys, epididymis, testes, and seminal vesicles in nontreated animals. Following treatment with anti-androgens, luciferase signal significantly decreased in the intact male mouse in vivo and by measuring luciferase activity in homogenized organ extracts. The decrease in epididymal and seminal vesicle weight confirmed the action of the anti-androgens. In vivo imaging documented significant changes in luciferase expression within the first few days of the experiment, indicative of the anti-androgenic activity of test compounds. Biophotonic imaging may provide a useful approach for tracking noninvasively the effects of endocrine disruptors in specific tissues.
S.E. Malstrom, Xenogen Corporation 5.
Abstract ID: 236
In studies of radiolabeled antisense oligomers, we have observed in tumor cells in culture an increased accumulation of radioactivity on antisense oligomers compared to sense, scrambled, or random control oligomers in eight cell lines, primarily with 99 mTc but also with 35 S as radiolabel, with uniform phosphorothioate DNAs, phosphorodiamidate morpholinos (MORFs), and locked nucleic acids (LNAs) oligomers and with and without cationic peptides and liposomes as carriers. In four cancer cell lines, we have reduced or eliminated the increased accumulations by increasing the DNA concentration. We have now extended these studies to targeting of MDR1 mRNA controlling the expression of Pgp and multidrug resistance. The uniform phosphorothioate DNA targeted the AUG start codon was used naked along with the sense DNA control. KB-G2 was transfected to overexpress MDR1 mRNA from its parent KB-31 while TCO-1 is also reported to overexpress Pgp. As a marker of Pgp expression, the uptake of 99 mTc-sestamibi (MIBI) in KB-31 cells was unaffected by incubation at therapeutic dosages with either DNA while uptake was significantly elevated in KB-G2 cells by antisense DNA. Thus, the antisense DNA was able to interfere with Pgp expression, probably by an antisense mechanism. Both DNAs were radiolabeled with 99 mTc via MAG3 and cellular uptake measured after 24 hr of incubation in all three cell types. A significant increased accumulation of antisense versus control DNA was observed in all cells when incubated with DNA at low nM concentrations, whereas this increased accumulation decreased or disappeared at high nM concentrations as expected for specific binding. These results suggest that Pgp expression may be a useful target for antisense imaging in MDR+ cancers. Perhaps more importantly, the inescapable conclusion from the weight of this evidence is that we have almost certainly observed accumulations of 99 mTc-oligomers in cells in culture by an antisense mechanism.
D.J. Hnatowich, None.
Abstract ID: 237
Joint space narrowing (JSN) is considered an important index for assessing the degree of osteoarthritis (OA) in human clinical practices; however, an accurate and robust method for measuring JSN in animal models of OA has not been reported. In this study, an automated tool to quantify JSN was applied to the characterization of disease progression in a rat model of chemically induced OA over a 2-month period using micro-CT. OA was induced in rats by a single intra-articular injection of 1 mg monosodium iodoacetate (MIA) into the right knee. Six animals were imaged using in vivo micro-CT at six time points over a period of 2 months post-MIA injection. A computer algorithm was developed to segment the surfaces of the tibial plateaus and femoral condyles from the 3-D micro-CT dataset, and then calculate the minimum distance between these two surfaces as a measurement of JSN (Figure 1). The time course evaluation indicated that the 3-D joint space decreased from 500 μm (Day 1) to about 250 μm at the end of 2 months, with the greatest reduction taking place during the first 14 days post-MIA (Figure 2). Based on these findings, we conclude that the application of micro-CT and the development of automated quantification tools provide a potential biomarker and/or efficacy end point for assessing OA in animal models. The method is robust, reproducible, sensitive to disease progression, scalable to other animals, and has the potential to be a valuable tool for use in drug discovery and development.
L. Chen, None.
Abstract ID: 238
Ionizing radiation (IR) and other stress factors can substantially change gene expression. At low IR exposures (2 cGy), we have demonstrated that clusterin (CLU) protein and its transcript levels are induced in human MCF-7 breast cancer cells, making it the only known gene to respond in this way at such low exposures. Our studies should improve understanding of low-dose IR effects and determine the feasibility of using CLU as a biodosimeter for the retrospective assessment of biological effects of exposure. Currently, we are developing a cellular biodosimetry system utilizing the CLU promoter-luciferase genetic construct and a bioluminescent imaging (BLI) system. MCF-7 cells were stably transfected with a DNA vector containing the human CLU promoter-luciferase cassette and a neomycin selectable resistance gene. An MCF-7 clone 1403 was isolated, in which the regulation of luciferase protein expression mimicked the endogenous CLU gene. Using a luminometer, statistically significant induction was detected at ≥50 cGy. In contrast, significant changes were detected using BLI after only 10 cGy. Detection required optimization of integration time, cosmic ray removal, binning, cell density, cell buffers, and growth factors. Induction responded in a dose-dependent manner. To investigate the potential for detection in a xenograft mouse model, luciferase activity was examined in large cell aggregates, rather than in single cell monolayers of MCF-7 1403 cells. Under these experimental conditions, we were able to reliably detect changes in luciferase expression induced by 10 cGy. In summary, bioluminescent imaging is more sensitive than standard luminometer assays; clusterin production is a promising concept for biodosimetry; and further studies will focus on the generation of mouse xenograft and transgenic mouse model systems using IR-inducible CLU promoter-luciferase gene expression. (This work was supported by a grant from NASA to D.L.W. and by DOE grant DE-FG-022179 to DA.B.)
D. Klokov, None.
Abstract ID: 239
Noninvasive quantification of HSV-1 and cellular thymidine kinase gene expression relies on the determination of the accumulation rates of specific marker substrates such as [124I]FIAU or [18F]FHBG for HSV-1-tk and [18F]FLT for cellular tk expression. Our objective was to normalize HSV-1-TK activity to cellular TK activity noninvasively in vivo. Rat F98 glioma cells were retrovirally transduced with a tkgfp fusion gene (TG17). F98-TG17 cells were implanted into the right striatum of cats (n = 3). Wild-type F98wt cells on the contralateral side served as controls. MRI and multitracer PET imaging was performed after systemic administration of [18F]FDG (166.5±66.6 MBq), [11C]MET (504.7±70.6 MBq), [18F]FLT (351.5±37.0 MBq), and [18F]FHBG (298.5±23.1 MBq). To study the kinetics of tracer accumulation and wash-out, a series of emission scans were acquired until 7 hr after tracer administration. Thereafter, animals were killed for coregistrative histology. Intracranial rat F98 gliomas in the cat are readily delineated by [11C]MET-PET, [18F]FLT-PET, and MRI but not by [18F]FDG-PET because of high cortical background activity. Tumors showed relatively more [18F]FLT uptake than [11C]MET uptake, the tumor-to-background (T/B) ratios ranging between 1.0 and 1.9 ([11C]MET) and between 1.4 and 3.7 ([18F]FLT). Transduced F98-TG17 tumors accumulated [18F]FHBG, the T/B ratios ranging between 3.0 and 3.7 (0.009–0.025%ID/g). At 182–327 min after [18F]FHBG administration, some background activity appeared in negative control F98wt tumors, with T/B ratios ranging between 2.2 and 2.8 (0.008–0.015%ID/g). [18F]FHBG/[18F]FLT uptake ratios, normalizing viral TK activity to cellular TK activity and, hence, proliferative activity, were significantly higher in F98-TG17 gliomas (0.9–2.7) than in F98wt tumors (0.74–0.77). These data indicate that background [18F]FHBG activity in nontransduced gliomas necessitates a conventional kinetic analysis and/or normalization to endogenous cellular tk gene expression to quantitate and localize transduced HSV-1-TK expression. Supported in part by MSWF 516–400 002 99, ZMMK-TV46 and DFG-Ja 981/1–2.
R. Graf, None.
Abstract ID: 240
MicroCT scanners capable of sub-20-micron spatial resolutions have now been introduced that afford ultra-high-resolution potential to more effectively monitor in vivo tumor growth and development in live small animal models. Relatively long image acquisition times, however, preclude the use of conventional water-soluble contrast agents in small animal models. Studies were initiated to compare the efficacy of BP20, a new long-acting vascular CT contrast agent, for noninvasively assessing vascular integrity in anesthetized live mice. BP20 is a surface-modified iodinated radiopaque microemulsion (50 mg I/ml) with a mean particle diameter of 70–80 nm currently undergoing preclinical development. Following injection into normal and tumor-bearing mice, the agent remains in the blood for over 2 hr prior to subsequent hepatobiliary recognition and elimination. BP20 (0.4 ml/20 g bw) was injected intravenously via tail vein into normal Balb/C (6) or alternatively into ApcMin/+ mice (6) bearing mammary carcinomas and the animals scanned on an ImTek microCT scanner (43 KVP, 410 ìA, 390 steps). Following imaging, animals were euthanized and tumors dissected and photographed for image correlation. Subsequent 2-D and 3-D surface-rendered images were constructed using Amira (V3.0) software for comparison with gross pathology. BP20 afforded striking vascular enhancement of normal vessels including, for example, the splenic vein, as well as extratumoral feeder vessels in live mice. Moreover, it is now possible to virtually “fly” through the lumen of these opacified blood vessels and associated branches less than 1 mm in diameter. To our knowledge, this work represents the first report of utilizing a microCT compatible vascular contrast agent for high-resolution 3-D vascular imaging of tumor vessels in live mice. This may prove extremely useful for monitoring vessel patency and atherosclerosis as well as tumor response to antiangiogenic agents acting at the vascular level.
J.P. Weichert, None.
Abstract ID: 241
Our laboratory has been developing a number of targeted molecular imaging agents, some of which are photodynamic therapy (PDT) agents. A number of tumors overexpress low-density lipoprotein receptors (LDLr). Imaging of LDLr by near-infrared (NIR) optical imaging is achieved by utilizing LDL labeled with cholesterol oleate (CO) conjugated to tricarbocyanine, pyropheophorbide (Pyro-CO), or bacteriochlorophyll (BChl-CO) and also by NIR imaging of apoE-labeled liposomes containing indocyanine green. Pyro-CO and BChl-CO can be used for PDT of tumor such as B16 melanoma and human G2 hepatoma (HepG2). Accumulation of these agents in B16 and HepB2 cells has been confirmed by confocal microscopy and in frozen tumor sections by low-temperature NIR reflectance imaging (Quistorff et al., 1985). Labeled LDL is cleared from the blood in 30 min. Labeling of LDL with gadolinium chelates for MRI detection has also been achieved. Delivery of NIR dyes and PDT agents via GLUT1 and GLUT3 transporters has been achieved with 2-[2-deoxyglucose] adducts of tricarbocyanines, Pyro, and BChl. Localization of these adducts in 9 L glioma cells has been confirmed by confocal microscopy and flow cytometry. Uptake is inhibited by
G. Zheng, None.
Abstract ID: 242
Measuring drug biodistribution is typically done invasively. The trend, however, is toward imaging studies, such as PET, to perform these functions. Optics offers much higher sensitivity to low concentrations of agents as compared to the other imaging modalities. Optical agents can be coupled to small molecules, proteins, and antibodies in order to quantitate tissue levels and image biodistribution. We set out to quantify target agent concentration and distribution. We performed this in two ways: (a) First, an optically active, strongly absorbing agent, Tirapazamine (tpz), was injected into animal subjects with s.q. tumors. The tumor was monitored either noninvasively as well invasively using FireFly™, a portable spectroscopy system. (b) Second, optical contrast agent distribution was studied in animal models. After injection of two agents, one renally excreted and the other hepatically metabolized, images were collected using the Palomar™ room-light fluorescence imaging system. Results show that (a) tpz concentration and time course was collected, analyzed, and displayed over 2 hr, and matched results from a prior study requiring hundreds of animals and months of tissue analysis (Figure A—Noninvasive Intratumoral Tpz Concentration). (b) Optical imaging clearly differentiated the different metabolism and excretion routes, even by eye, in real-time hepatic uptake, and renal excretion can be distinguished (Figure B—Renal Excretion; Figure C—Hepatic Uptake and Elimination). We conclude that in vivo pharmacokinetics is feasible using either optical spectroscopy or optical imaging, or both combined.
Noninvasive pharmacokinetics and biodistribution using optical imaging and spectroscopy.
P.G. Maxim, Spectros 1.
Poster Session 08: Detecting Endogenous Molecules and Signatures
Abstract ID: 243
A number of activatable imaging probe technologies that can be used to sense molecular targets and functions in vivo have been described. The goal of the current study was to extend magnetic relaxation switch (MRS) technology to sense and image telomerase activity. Telomerase is a ribonucleoprotein with a reverse transcriptase activity, bearing the unique feature that its RNA template is an integrated constituent of the enzymatic complex. Increased telomerase activity is found in the majority of malignant tumors but is absent from or significantly lower in normal cells. The capability to quantitate telomerase activity in a high throughput fashion in vitro or even in vivo would be helpful in many biomedical applications. We have previously developed an MRS-based telomerase sensor capable of sensing at attomol telomeric DNA levels. Here, we show the quantitative nature of the assay, validate intrinsic amplification, and validate it against accepted gold standards. We show that telomerase levels in human and mammalian cancer tissues can be reliably assayed with MRS considerably faster than with conventional assays. Furthermore, we show that the sensors can be used to test telomerase regulation and the efficacy of novel telomerase inhibitors. These studies for the basis for future developments of in vivo telomerase sensors.
J. Grimm, None.
Abstract ID: 244
Noninvasive imaging of gene transfer to specific cells/tissues in vivo should allow for optimization of gene therapy. Taking advantage of the ability of HIV-1-based lentiviral vectors to integrate stably into the host genome and transduce quiescent cells; we have developed a Two-Step Transcriptional Amplification (TSTA) system based lentiviral vector to image prostate-specific gene expression in living animals. The TSTA construct, PSEBCVP2-G5-fl was cloned into the third generation self-inactivating (SIN) lentivector, PTK 134. The SIN lentiviral vectors (LV-TSTA) were prepared by cotransfection into 293 T cells. For in vivo studies, LV-TSTA was injected into male, nude mice (intratumorally) carrying LAPC-9 tumor xenografts. The mice were imaged four days postinjection using a cooled charge-coupled device (CCD) camera using
M. Iyer, None.
Abstract ID: 245
Computed tomography (CT) with water-soluble iodinated contrast media is a commonly used modality for the detection, staging, and management of liver disease. The primary type of information obtained from these procedures, however, is anatomical in nature. The ability to provide functional information correlating the physiological status of the liver to contrast enhancement profiles would provide a significant tool in the management of liver pathology. We have designed and characterized a polyiodinated triglyceride (ITG) analog formulated in an oil-in-water lipid emulsion (LE) that provides hepatocyte-selective imaging agents for use in CT. The ITG is incorporated into the lipophilic core of a synthetic CMR-like lipid emulsion for localization to the hepatic parenchyma from a peripheral intravenous administration. Because the uptake and subsequent clearance of the ITG molecule is determined by the metabolic status of extracellular and intracellular liver lipases, the ITG-LE technology provides the potential for evaluation of liver function as well as for liver anatomy by CT. Moreover, the metabolites of the ITG are capable of providing contrast enhancement of the entire biliary tract during their clearance from the body. The imaging efficacy of the ITG-LE formulation in preclinical animal models of diffuse and focal liver disease will be presented. Several applications of the technology will be presented including visualization of liver cancer (primary and metastatic), contrast enhancement profiles in functionally compromised liver, and characterization of lesion margins in the liver following ablation procedures. The ITG-LE technology represents a powerful imaging platform capable of providing both anatomical and functional evaluations of the entire hepatobiliary system following a single intravenous administration.
D.A. Bakan, MetaProbe, Inc. 5.
Abstract ID: 246
We examined the expression and regulation of the rat insulin promoter (RIP) in transgenic mice by following expression of the firefly luciferase (luc) gene with real-time noninvasive detection technology using the IVIS™ Imaging System. Transgenic mice expressing the RIP-luc construct specifically in the pancreas allowed in vivo examination of RIP-luc transcriptional regulation resulting from whole animal metabolic alterations. RIP-luc expression was tracked in transgenic mice fed a chow or high-fat diet, during an extended fast, and following treatment with streptozoticin. The imaging results from these studies were correlated with body weight, blood glucose and insulin levels, and pancreatic protein luciferase activity. During high-fat feeding, RIP-luc expression was found to increase slightly over that of chow-fed animals. Pancreatic RIP-luc expression, blood insulin levels and pancreatic luciferase activity were reduced following streptozoticin treatment while blood glucose levels increased. The RIP-luc transgenic model can function as a useful tool for in vivo studies of pancreatic β-cell health and function.
S.J. Smith, Xenogen Corporation 5.
Abstract ID: 247
The genetically engineered anti-carcinoembryonic (CEA) diabody, a dimer of the single-chain Fv from the murine anti-CEA T84.66 antibody, has previously demonstrated high-level tumor targeting, fast blood clearance, and usefulness in imaging applications. When radiolabeled, the anti-CEA diabody can be used for positron emission tomography imaging of athymic mice with CEA-positive LS174T xenografts. We are developing a technique for optical imaging of tumors expressing CEA by fusing an anti-CEA diabody with the bioluminescence protein Renilla luciferase (hRLuc). hRLuc catalyzes the oxidation of its substrate coelenterazine producing visible light of 480 nm and has been used for optical imaging in living mice (Bhaumik and Gambhir, 2002). Genes encoding the 25-kDa anti-CEA diabody and the 36-kDa hRLuc were fused via a 5 or 18 amino acid linker. The fusion protein was expressed using bacterial cells and purified by metal affinity chromatography. A bioluminescence ELISA assay demonstrated that the fusion protein can simultaneously bind to recombinant CEA and oxidize coelenterazine to emit light. Western blots using antibodies specific for both the anti-CEA diabody and hRLuc identified the same protein band corresponding to a molecular weight of 61 kDa. Mammalian expression is being explored and may offer higher yields and easier purification. This bifunctional fusion protein should be suitable for in vivo optical imaging of CEA-expressing xenografts in athymic mice using a cooled charge-coupled device camera. Optical imaging approaches with novel probes will be useful in the development and evaluation of preclinical models of disease, and may have clinical applications in breast imaging, endoscopy, and intraoperative imaging.
K.M. Venisnik, None.
Abstract ID: 248
Mouse models are increasingly being used in biomedical research to understand disease processes. With the availability of genetically defined murine models, functional imaging modalities such as MRI and PET have become valuable tools to interrogate phenotypic changes repeatedly and noninvasively. Imaging modalities often require that the animal be immobilized, often times with anesthesia, which may, in turn, change important physiologic parameters, such as core body temperature and blood flow. Changes in the oxygenation and blood flow in tumors (squamous cell carcinoma (SCCVII)) implanted in mice, maintained at core temperatures of 30 or 37 °C were monitored using blood oxygen level dependent (BOLD) MRI and compared to tumor oxygen levels obtained using an oxygen-sensitive Eppendorf electrode. Tumor-bearing mice were imaged by BOLD MRI, while first breathing air and then carbogen (95% O2, 5% CO2) for 15-min intervals, respectively, at a core temperature of 30 °C. After an equilibration period, the identical regimen was conducted with the same animal maintained at 37 °C. This procedure was repeated with additional mice starting at 37 °C followed by imaging at 30 °C. Likewise, oxygen electrode measurements of the tumor were determined at core temperatures of 30 and 37 °C. Both techniques showed that tumors in animals maintained at 30 °C were significantly more hypoxic than at 37 °C. The findings of the study stress the importance of maintaining normal core temperature when assessing tumor oxygen status using functional imaging modalities or oxygen-sensitive electrodes.
K. Reijnders, None.
Abstract ID: 249
Coelenterazine, the imidazolopyrazine substrate for Renilla luciferase (Rluc), shares structural properties of other compounds transported by the multidrug resistance MDR1 P-glycoprotein (Pgp). To test the hypothesis that Pgp mediates efflux of coelenterazine, human KB 3–1 cells and KB 8-5-11 cell lines, expressing nonimmunodetectable and high levels of MDR1 Pgp, respectively, were stably transfected with codon humanized Rluc and coelenterazine-mediated bioluminescence assayed in intact cells. Despite comparable Rluc protein levels, KB 3-1 Rluc cells showed high bioluminescence, while KB 8-5-11 Rluc cells demonstrated 10-fold lower bioluminescence. Light emission from KB 8-5-11 Rluc cells, but not KB 3–1 Rluc cells, was reversed to control levels by the potent Pgp modulators GF120918, PSC 833, and Cyclosporin A, and unaffected by nontargeted agents. In addition, coelenterazines f, h, and hcp were avidly transported by Pgp, whereas coelenterazines n and cp were not. Following systemic administration of native coelenterazine in a nude mouse tumor xenograft model, compared to control, a GF120918-reversible low level bioluminescence was readily detected in Pgp expressing tumors in vivo. Thus, coelenterazine is a Pgp transport substrate, thereby providing new insight into mechanisms of cell permeation of this important molecule in marine organisms as well as enabling a novel method for noninvasive bioluminescence imaging of Pgp transport activity and its inhibition in intact cells and in living animals. In light of our results, indiscriminate use of coelenterazine-dependent enzymes such as Rluc and the free-calcium reporter aequorin would be suspect in intact cells or transgenic animals where Pgp status is not known.
A. Pichler, None.
Abstract ID: 250
One of the early hallmarks of tumorigenesis in the majority of epithelial cell adenocarcinomas is the overexpression and underglycosylation of the MUC-1 antigen (uMUC-1). As a result of underglycosylation, core peptide epitopes on uMUC-1, which are cryptic in the nonneoplastic state, become exposed and can be targeted for imaging and therapeutic purposes. Cellular distribution of uMUC-1 is allocated to the cytoplasm and to the cell surface. In this study, we synthesized and tested a multimodal imaging probe, crosslinked iron oxide (CLIO)-EPPT, which recognizes uMUC-1 antigen and utilizes both optical and MR imaging modalities. The probe consists of Cy5.5-modified CLIO nanoparticles conjugated to an FITC-labeled synthetic EPPT peptide specifically recognizing uMUC-1. On the average, CLIO-EPPT probe contained 7 EPPT molecules and 1 Cy5.5 molecule per CLIO particle. R1 and R2 relaxivities were similar to that of the parental compound CLIO-NH2. The specificity of CLIO-EPPT probe was assessed using a panel of uMUC-1 positive human adenocarcinoma cell lines and uMUC-1 negative cell lines as controls. uMUC-1 expression was verified by real-time PCR. Furthermore, we confirmed the differential availability of uMUC-1 epitopes on adenocarcinoma cells versus control cells by immunostaining with a monoclonal antibody to the peptide core of the antigen. Cellular accumulation of CLIO-EPPT probe showed significantly higher amounts of the probe in uMUC-1 positive adenocarcinoma cells compared to the control cell lines (p <.05) (Figure 1). We also confirmed specific accumulation of the probe in uMUC-1 positive cell lines by fluorescence microscopy. In summary, this study demonstrates the specificity of a novel multimodal imaging probe, which targets uMUC-1 for epithelial adenocarcinoma cells. The potential of CLIO-EPPT probe as an in vivo imaging tool is described in the accompanying abstract.
Z. Medarova, None.
Abstract ID: 251
Underglycosylated mucin-1 antigen (uMUC-1) is one of the early hallmarks of tumorogenesis in a wide variety of tumors. It is overexpressed in almost all human epithelial cell adenocarcinomas, including more than 90% of breast cancers, pancreatic, colorectal, lung, prostate, colon, and gastric carcinomas. Moreover, uMUC-1 expression has been demonstrated in nonepithelial cancers (multiple myeloma, some B-cell non-Hodgkin lymphomas). The goal of this study was to test in vivo recently developed multimodal imaging probe specifically recognizing uMUC-1-expressing tumors in animal models. The probe consists of crosslinked iron oxide nanoparticles (CLIO, MR imaging agent), modified with Cy5.5 fluorochrome (NIRF imaging agent) and has EPPT peptides, specifically recognizing uMUC-1, attached to its dextran coat. The probe designated as CLIO-EPPT was tested in vitro on uMUC-1-positive (+) and uMUC-1-negative (–) human cell lines and in vivo in the mouse model of human cancer. For in vivo studies, mice (nu/nu) were injected with uMUC-1+ and uMUC-1– human tumors in the opposite flanks. The biodistribution of the probe showed its preferential accumulation in uMUC-1+ tumors. Initial imaging experiments of intravenously injected probe in vivo utilizing both optical (Figure 1) and MR imaging (Figure 2) modalities showed strong signal in uMUC-1+ tumors, and virtually no signal in uMUC-1– tumors. Ex vivo NIRF imaging of excised tumors confirmed in vivo data. Fluorescence microscopy of frozen tumor sections corroborated our in vivo imaging data that CLIO-EPPT specifically accumulated in uMUC-1+ tumors and did not accumulate in control tumors. This study could possibly lead to the ability to detect and follow the early progression of cancer.
A. Moore, None.
Abstract ID: 252
We synthesized and tested two groups of contrast agents based on iron-oxide core nanoparticles (SPIO) and different macromolecules labeled with a large number of gadolinium complexes. As a model system, we used the HER-2/neu cell surface receptor that is expressed in approximately 25% of breast cancers and other forms of human cancer and is important in breast cancer prognosis, and as a target for immunotherapy with humanized monoclonal antibody (mAb), Herceptin. Overexpression of the receptor is associated with increased rates of relapse and mortality in breast cancer patients. The panel of contrast agents tested included streptavidin-SPIO conjugates, avidin–(GdDTPA)N, and biotinylated albumin(GdDTPA)N, histone(GdDTPA)N and fourth- and fifth-generation PAMAM denrimer(GdDTPA)N. Preliminary studies were performed with a panel of human breast cancer cells including AU565, BT-474, MCF-7, and MDA-MB-231 lines pretargeted with biotinylated Herceptin. Cell expression of the HER-2/neu receptors was assessed with Western blot, immunofluorescence, and FACS analysis. Stability of the mAb/contrast complex on the cell surface was determined with confocal microscopy. For MR imaging, labeled cells were immobilized in agarose gel layers in 5-mm NMR tube and imaged using standard T1 and T2 protocols using 400 MHz NMR spectrometer. All contrast agents generated strong contrast in immobilized HER-2/neu-expressing cells; SPIO particles produced the highest negative T2 contrast. T1 positive contrast generated by Gd-based agents was proportional to the number of gadolinium atoms per macromolecule carrier and to the expression level of the HER-2/neu receptor on the cell plasma membrane.
D. Artemov, None.
Abstract ID: 253
Molecular profiling of breast cancer has led to the identification of candidate diagnostic and therapeutic targets. Discoveries from these genome-scale approaches may have applicability in the imaging of breast cancer. To characterize the expression levels of target molecules in breast cancer cells, quantitative evaluation was performed using clinical samples. Fresh frozen sections were prepared from human breast cancer tissues obtained from 29 patients and laser capture microdissection (LCM) was performed to obtain normal glandular cells and cancer cells. Total RNA extracted from LCM captured cells were amplified and reverse transcribed. PCR primers and a fluorescent hybridization probe for the transferrin receptor (TfR), gastrin-releasing peptide receptor, urokinase plasminogen activator receptor, and matrix metalloproteinase 1, 2, 9, and 13 transcripts were designed. Quantitative real-time PCR was carried out in an ABI 7900 Prism Sequence Detector (Applied Biosystems), and the quantification of each mRNA expression was analyzed with standard curve method. The chi-square test was used to determine the significance of the results. A p value < .05 was considered to be statistically significant. Each sample derived from cancer tissue was compared to matched normal breast tissue from the same patient. Thirty-nine samples (ductal carcinoma in situ 17, invasive carcinoma 22) were evaluated. TfR mRNA was overexpressed in approximately 75% of the cancers tested and overexpressed at imageable levels (>5 times over expression relative to normal breast tissue) in 41% (16/39) of the samples. Preliminary data on other imageable targets suggested a similar degree of overexpression, 35–50%, within the patient population. Molecular analysis of LCM cells from breast cancer tissues can be used to determine the degree of overexpression within an individual and the degree of overexpression within a patient population. Studies like these will be necessary to validate these and future molecular diagnostic targets.
F. Sato, None.
Abstract ID: 254
The goal of our project is to characterize and establish for the first time the sources of endogenous fluorescence contrast in ethylnitrosourea (ENU) induced mammary hyperplasias, malignant tumors, and normal tissue in genetically altered female mice. A promising technology for the detection of hyperplasia is fluorescence spectroscopy. No other nondestructive technology available at comparable cost can equal fluorescence spectroscopy in sensitivity. Moreover, fluorescence spectroscopy can probe a large number of endogenous molecules already present in breast tissue, thus providing a wealth of biochemical and morphological information related to disease progression. Additionally, advances in fiber-optic sensors and detectors enable fluorescence spectroscopy to be performed rapidly and remotely from biological tissues in vivo. The project is based upon two central hypotheses: (1) There exists a systematic and significant difference in the endogenous fluorescence of the normal mammary gland, hyperplasias and malignant tumors. (2) The differences in the endogenous fluorescence of normal, hyperplastic, and malignant breast tissues is attributed to one or more of the following fluorophores: tryptophan, reduced nicotinamide adenine dinucleotide, flavin adenine dinucleotide, and collagen. We have set out to identify the macroscopic fluorescence signatures of normal mammary tissues, late hyperplasias, and malignant tumors in the mouse mammary gland and to correlate these findings with histology. The macroscopic fluorescence signatures were measured from the exposed mammary glands of FVBxB6 Min/+ and wild-type mice in vivo using a fiber-optic probe, coupled to a fast, multiwavelength optical spectrometer. The mammary glands were then collected for histological analysis and to allow correlation of spectral data with histology. The fluorescence spectral data was analyzed using a multivariate statistical algorithm developed by our group. Preliminary observations of the excitation-emission wavelengths at which maximal differences are observed in the fluorescence of the different tissue types will be presented.
A.R. Moser, None.
Abstract ID: 255
Vascular endothelial growth factor (VEGF) is an important protein regulating neovascularization. We have developed a transgenic mouse model in which firefly luciferase (fl) gene expression can be used to indirectly monitor endogenous VEGF gene expression. We developed a two-step transcriptional amplification (TSTA) plasmid, pVEGF-VP2G2-fl in which the human VEGF promoter drives a GAL4-VP16 fusion protein, which, in turn, activates the fl reporter gene. Transient transfection studies were used to validate that TSTA-based fl expression was significantly higher than a direct one-step approach and that fl expression increased under hypoxia. The fl activity of HeLa cells stably transfected with pVEGF-VP2G2-fl correlated very well (r2 = .97) with endogenous VEGF protein levels measured by ELISA when the cells were grown under hypoxia for various times. Transgenic mice bearing pVEGF-VP2G2-fl DNA construct were developed. For wound studies, 5-mm-diameter wounds were made on the back of the mouse (N = 3). For tumor studies, BR44 cells isolated from FVB mammary cancer were implanted on the transgenic mice. The mice were imaged at 0, 5, 9, 15, and 20 days after injury/implantation. The background FL activity of the transgenic mouse varied between 1 and 7.7 max (×105 photons/sec/cm2/sr) over time. The FL signal emitted from the wound region increased over time, peaking at Day 15 (1.7 × 107 photons/sec/cm2/sr), a 19-fold increase over the signal at Day 0. The xenograft tumors implanted in the transgenic mice also lead to significantly higher fl expression in the tissues surrounding the tumor as compared to other sites. These results demonstrate the feasibility of using the TSTA amplification strategy to monitor VEGF gene expression in transgenic models and should prove useful under various environmental/pharmaceutical interventions.
Y. Wang, None.
Abstract ID: 256
Regulation of glutamate/glutamine homeostasis is a multicellular process, that enables normal neuro-excitation by release of glutamate into the synaptic space, and also provides for its rapid cellular reuptake. In neurodegenerative diseases and brain trauma, prolonged, excess glutamate in the synaptic space can trigger a toxic cascade, via NMDA/AMPA kainate receptors leading to cell death. In vivo detection of changes in glutamate/glutamine along with markers for cellular populations, and intra- and extracellular environments would be useful. In this study, we examined in vivo detectability of glutamate/glutamine homeostasis and markers for cell population and environment using high field MR spectroscopy. Data were collected on a Signa 3T scanner (GE Medical Systems, Milwaukee, WI), using TE-averaged PRESS. Glutamate was measured at 2.35 ppm (Glu2.35) and glutamine was measured at 3.75 ppm (Glx3.75), once glutamate contribution as subtracted. N-Acetyl metabolites resonating at 2.02 ppm (NAA and NAAG) were quantified as a marker for neuronal content, and myoinositol was quantified as a marker for glial content. Choline and creatine were also measured. T2's were measured by single exponential fit. T1's were estimated from partial saturation using a TR 8000 reference scan where partial saturation was estimated as S=So(1-2e)(TR-TEef/2)/T1+e(-TR/T1)). TEeff is the echo time for mean signal in the TE-averaged sample. This method was also simulated using GAMMA at field strengths from 0.5 to 7 T. MR spectroscopy at 3 T and above can detect changes of 10% or more in glutamate, and with comeasured metabolite markers, it may be able to distinguish between gliosis, immune response, and changes in glutamate/glutamine homeostasis not associated with cell population changes. T1 and the T2 measurement inherent in the TE-averaging method are important for quantification and likely for detecting cellular environment changes.
R.E. Hurd, GE Medical Systems 5.
Abstract ID: 257
MR spectroscopic imaging (MRSI) has become a powerful clinical and research tool to characterize prostate cancers based on endogenous cellular metabolite levels and to detect early response to therapy. Virtually all prior studies have been performed at 1.5 T. The goal of this study was to develop 3-T techniques for prostate MRSI data at higher SNR and spatial resolution than 1.5-T studies. Also, we applied single-voxel, j-resolved MRS to determine the j-modulation of citrate and polyamines at 3 T. Because the frequency range for the metabolites of interest (choline, creatine, polyamines and citrate) are doubled at 3 T, new rf pulses were required for observing the metabolites and suppressing water and lipid resonances. Also, new reduced peak power rf pulses were necessary for 3-T body coil excitation. Novel phase-modulated dual-band spectral spatial pulses were designed and implemented, which excited water at a 100-fold reduction, fully excited the range of metabolites, and excluded the excitation of lipid resonances. The new 3-T MRSI sequence was applied in 5 patient and volunteer studies on a GE 3-T scanner using body coil excitation and reception with a MedRad 3-T endorectal coil. The 3-D MRSI data was acquired in 16 min at 0.15 cc (half typical 1.5-T resolution). The 2-D j-resolved data was acquired from a ˜4-cc volume with 32 steps of 10-msec intervals from TE = 35–345 msec and demonstrated different j-modulation at 3 T than 1.5 T. 3-D MRSI was acquired throughout the gland at TE = 95, 130, and 260 msec. The 95-msec TE data showed maximal inverted citrate and greatly improved separation of choline, polymanine, and creatine resonances than 1.5-T data. This study demonstrated the feasibility of obtaining 3-D MRSI data from prostate cancer patients at 3 T and the ability of obtaining high-spatial-resolution MR spectra throughout the prostate in vivo with improved discrimination of choline, creatine, and polyamine resonances.
A. Chen, None.
Abstract ID: 258
In previous in vivo magnetic resonance spectroscopic imaging and ex vivo high-resolution magic angle spinning (HR-MAS) spectroscopy studies, elevation of the composite or “total” choline resonance has been shown to correlate with prostate cancer presence and aggressiveness. The individual choline compounds phosphocholine (PC), glycerophosphocholine (GPC), and free choline (Cho) contain information about phospholipid membrane synthesis and degradation, and, consequently, proliferative status. However, even at 11.5 T (500 MHz), it is difficult to resolve the individual choline head group resonances of Cho (3.21 ppm), PC (3.23 ppm), and GPC (3.24 ppm) in 1-D HR-MAS spectra. Because the choline side chain (CH2-CH2) protons are adjacent to different substituents, namely, hydroxyl (Cho), phospho (PC), and glycerophospho (GPC) moieties, their resonances are further separated in frequency. By applying a technique that spreads the spectral data into two dimensions, the individual (CH2–CH2) crosspeaks can be completely resolved. In this study, 2-D Total Correlation Spectroscopy (TOCSY) experiments were performed under HR-MAS conditions to identify and quantify individual choline side chain (CH2–CH2) crosspeaks for Cho, PC, and GPC in 10 benign and 10 malignant prostate tissues. Data were acquired at 1 °C and a 2250-Hz spin rate, using a Varian INOVA spectrometer equipped with a 4-mm gHX nanoprobe. Solutions containing varying concentrations of Cho, PC, and GPC were analyzed to optimize experimental conditions and validate the method. Degradation studies were performed to determine the optimum experimental duration and to correct for metabolic degradation during the course of the TOCSY experiments. TOCSY parameters were optimized for quantitative detection and rotor synchronized adiabatic pulses were used to preserve magnetization transfer. 2-D Crosspeaks were volume integrated and referenced using a diagonal peak. Concentrations of choline containing compounds will be compared for cancer versus healthy prostate tissues, and correlated with immunohistochemical markers for proliferation (MIB-1) and apoptosis (TUNEL).
A.S. Zektzer, None.
Abstract ID: 259
Because of the accessibility of the luminal endothelial cell surface to the circulation, the vasculature is a logical target for drug or gene delivery. Transcytotic vesicles (caveolae) are dynamic structures on the surface of the endothelium capable of budding from the plasma membrane for endocytosis and/or transcytosis of blood macromolecules. We purified luminal endothelial cell plasma membranes and its caveolae directly from various organs and resolved the membrane proteins by SDS-PAGE and 2-D gel electrophoresis to reveal extensive heterogeneity of cell surface protein expression between organs. Tissue-specific proteins were identified by mass spectrometry and database searching. Using antibodies to appropriate specific polypeptides, we have confirmed the tissue specificity of the proteins. One such antibody, designated TX3.833, recognizes a lung-specific protein expressed quite selectively in caveolae of microvascular endothelium only in lung. Biodistribution analysis and whole-body imaging using TX3.833 injected intravenously show the accessibility of the recognized antigen with significant and selective lung tissue accumulation. Dynamic imaging shows rapid uptake and targeting of the iodinated TX3.833 antibody to the lung region within minutes after tail vein injection, which can be competed with the addition of “cold” TX3.833 antibody but not “cold” control antibody. Targeting endothelial caveolae maybe a worthy strategy for directed molecular imaging and targeting in vivo. (Supported by NIH Grants HL58216 and HL52766.)
P. Oh, None.
Abstract ID: 260
In vitro glucose consumption in tumor cell lines correlate with elevated HIF-1α levels under hypoxic conditions. PET provides the opportunity to study glucose metabolism in vivo using the glucose analog 18FDG. A majority of tumors as seen from PET imaging studies are characterized by high rates of glycolysis. Tumors expressing elevated aerobic glycolysis (Warburg Effect) are highly aggressive and frequently associated with poor prognosis in patient recovery. It is not yet fully understood what mechanism drives this phenotype. The recent discovery of HIF-1 (hypoxia-inducible factor) has generated interest because it becomes activated in low oxygen conditions associated with tumors and subsequently triggers the production of survival factors including genes responsible for upregulating glucose metabolism. The relationship between HIF-1α and increased glycolysis implicates HIF-1α as the potential mediator triggering the glycolytic phenotype in progressive tumor development. We examined glucose uptake and lactate and proton production of HIF-1α-expressing tumor cell lines under normal and hypoxic conditions. Nonmetastatic MCF-7 cells exhibited a twofold increase in glycolytic activity under hypoxia, whereas highly metastatic MDA-mb-435 cells were glycolytic regardless of oxygen conditions. RCC4 cells transfected to express von Hippel-Lindau protein, a cytoplasmic factor that targets HIF-1α for degradation, also responded to hypoxia by doubling glycolysis. Mock transfected RCC4 counterparts, however, maintained similar levels of glycolysis under normoxic and hypoxic conditions. This suggests that HIF-1α expression under hypoxia drives increased glycolysis, and when expressed under normoxia, it sustains aerobic glycolysis.
I.F. Robey, None.
Abstract ID: 261
Spliceosome-mediated RNA trans-splicing (SMaRT) is a novel platform technology to image or reprogram gene at the level of mRNA splicing. Here, we report the development of SMaRT technology for noninvasive, real-time, optical imaging of gene expression in living animals. The synthetic Renilla luciferase (hRluc) reporter gene has been split into 2 components. The target (Luc-HPVT3) bearing 5′ exon and a pre-trans-splicing molecule (Luc-PTM37) consisting of 3′ exon of hRluc were both engineered. In vivo delivery of PTM37 was performed by tail vein injection of transferrin-polyethylineamine PTM37 complexes (Tf-PEI-PTM37) (Hildebrandt, 2003) into nude mice (n = 3) bearing N2a cells, transiently transfected with Luc-HPVT3 (N2a-HPVT3 cells) implanted at various subcutaneous (5 × 106 cells) sites or injected intravenously (1 × 106 cells) via tail vein. Luc-HPVT3 target and Luc-PTM37 were also complexed separately with DOTAP/cholesterol and injected via tail vein in different nude mice (n = 3) [Mol Ther. 6(4):555–562, 2002]. The mice were imaged after 24 hr, using a CCD camera, after tail vein injection of the substrate coelenterazine (60 μg/mouse). Through its binding domain, the Luc-PTM37 base pairs with the target Luc-HPVT3 intron sequence bearing the 5′ exon and trans-splices the 3′ exon, thereby repairing the hRluc pre-mRNA and subsequently restoring the enzymatic activity. Restored signal is seen from tumor and liver sites of living animals under the CCD camera. Tail vain injection of Tf-PEI-PTM37 complex shows target specificity for N2a-HPVT3 cells implanted at various sites in living mice with a statistically significant (p <.05) signal, as compared to mice not injected with Tf-PEI-PTM37 complex. Injection of both target-PTM/DOTAP/cholesterol complexes shows significant (p <.05) imageable signals from lungs as compared to mice injected with the target/DOTAP/cholesterol complex alone. Repeated PTM37 injections shows increase in signal restoration. This study demonstrates imaging of mRNA repair in living mice and should eventually allow imaging of endogenous mRNA in living subjects. This approach has the potential to be generalized to other imaging modalities and to various mRNA targets.
S. Bhaumik, None.
Abstract ID: 262
Dopamine (DA) is released in the striatum in response to drugs of abuse and is critical for drug reward, which reinforces repeated drug-taking behavior. Individual differences in response to drugs may depend on the dynamics of the DA response such as latency to respond, time to peak response, and magnitude and persistence of the response. Much of what we know about neurotransmitter responses to drugs of abuse is based on semiquantitative microdialysis in small animal models of drug abuse. Recent advances in PET now allow for sufficient spatial resolution to image regional variations in binding of neuroreceptor ligands in discrete brain regions (e.g., striatum) of rodents. DA readily competes with some DA-ergic PET tracers, thus altering the dynamics of the measured PET signal. Raclopride is one such tracer that is reproducibly displaced by changes in endogenous DA and, thus, is ideal for our purpose. We present a new PET-based method to track with high-temporal-resolution dynamic changes in endogenous DA release in response to stimuli. We have shown via hundreds of computer simulations that it is possible—with a proper data acquisition, current knowledge of raclopride kinetics in brain tissue, specifications of our small animal PET scanner, and some prior information regarding physiological parameters—to reconstruct a rat's striatal DA response to intravenous alcohol in fine time increments. The keys to our method are (a) modeling the competition between endogenous DA and an exogenous PET tracer, (b) simultaneous estimation of kinetic parameters from multiple PET studies, and (c) application of uniquely appropriate constraints to the data fitting process.
We believe that our method will allow researchers in numerous areas of neuroscience to investigate how information is encoded by different temporal patterns of neurotransmitter release and how these patterns may be altered in disease or even encode the susceptibility to disease.
E.D. Morris, None.
Abstract ID: 263
Glioblastoma multiforme (GBM) is a primary brain tumor with poor prognosis with the overall 5-year survival rate of less than 6%. Recent studies have shown both intra- and intertumor genetic and protein heterogeneity, making sampling of tissue for further analysis a critical parameter to determine molecular targets in human GBM. In this study, we attempted to better characterize GBM by investigating protein expression profiles differences based on MRI differences. A diffusible contrast agent Gadolinium-DTPA (Gd) was used to delineate contrast-enhanced (CE) and nonenhanced (NE) regions of gliomas on surgical planning MRI scans. Total protein was extracted from snap-frozen, freshly resected tumor samples guided by MRI characteristics. To improve reproducibility and reliability of differential protein expression analysis, we utilize the two-dimensional differential gel electrophoresis to analyze the extracted protein samples. Protein extracts from CE and NE regions were labeled with Cy3 or Cy5 fluorescent dyes, respectively, and were mixed before electrophoretic separation. Our results show that while the majority (>60%) of the detected proteins did not change more than 1.5-fold, similar amount of protein is either up- or down-regulated between CE and NE regions. This number is significantly higher than our previous findings using traditional two-dimensional gel electrophoresis. The difference may reflect the improved sensitivity and reproducibility of the DIGE technology. We are currently identifying these differentially expressed proteins. Combining sophisticated imaging tools such as MRI with selective tumor sampling methods and new proteomics technology, specific patterns of protein expression can be identified in GBM. The ability to identify spatial tumor-specific proteins correlated with imaging characteristics also provides exciting opportunities in the design of improved diagnostic and therapeutic agents.
G. Shi, None.
Abstract ID: 264
Metabolic imaging at high resolution can be a problem for a technique such as PET, however, optical imaging has significant advantages over the PET in the sense that it can measure metabolic activity with the following three metrics: (1) relative deoxygenation of hemoglobin with respect to nearby normal tissues, (2) alterations of the redox state of the mitochondrial matrix state as measured by fluorescence of flavoprotein and NADH, (3) increased uptake of NIR-labeled fluorescent glucose. The availability of forms of fluorescent labeled glucose adds to the localized metabolic activity activation available from hemoglobin deoxygenation in the vascular bed and from changes of the mitochondrial matrix space redox state. Of particular interest is pyropheophorbide-labeled glucose, which absorbs near the FIR window and has translational possibilities. Furthermore, the uptake of fluorescent glucose indicates the delivery of reducing equivalents to the mitochondrial space, presumably through the citric acid cycle and causes increased NAD reduction in a manner proportional to its uptake over a wide range, that is, enhanced glycolytic activity affords more highly reduced state of the mitochondrial matrix. Combination of these three markers can afford much improved measures of the metabolic state of cancers and will, in the case of hemoglobin deoxygenation and fluorescent glucose uptake, afford high levels of recognition of cancerous metabolic states in human tissues. While these results are currently based largely upon high-resolution 3-D cryoimaging in vitro, transferability to human cancers by the deoxygenation of hemoglobin is currently in use in many laboratories for detection of human cancers and can complemented by NIR-labeled glucose. The measurement of NADH in the mitochondrial matrix space may also soon be possible through a special multiphoton technology. Thus, identification of human cancers by NIR imaging signals of altered metabolic states may become an important part of human cancer diagnosis.
B. Chance, None.
Poster Session 09: Imaging and Therapy Combinations
Abstract ID: 265
SU11657 (Sugen Inc.) inhibits signaling via class III and V receptor tyrosine kinases with both antiangiogenic and antitumor effects. [18F]FDG, [18F]FLT and [13N]NH3 (for perfusion) have been used previously as tracers for PET imaging. Here, microPET and CCD optical bioluminescence imaging were chosen to monitor efficacy of SU11657 for treatment of C6 rat glioma cells, which stably express firefly luciferase (C6-fluc) in a mouse xenograft model. Three groups of mice for each tracer were inoculated with C6-fluc cells. Scanning was started after the tumor volume reached ˜200 mm3. Before microPET scanning, each mouse was scanned using CCD with tail-vein injection of
Y. Wu, None.
Abstract ID: 266
In vivo delivery and targeting of macromolecular pharmaceutical agents has remained a major challenge primarily because of vascular and interstitial tissue integrity barriers. Our laboratory utilizes pulsed focused ultrasound to physically alter vascular permeability as well as interstitial barriers to macromolecules in a targeted and transitory fashion. We hypothesized that the equivalent energy level of focused ultrasound treatment would induce gene expression changes in tumors that could potentially be exploited for inducible molecular targeting of pharmaceuticals or gene delivery, as well as for molecular imaging. C3H mice were subcutaneously injected with squamous cell carcinoma VII cells in each flank, and the tumors were allowed to grow up to one centimeter. Focused ultrasound was then applied at nonablative energy levels to one of the entire tumors on each mouse using a dual imaging/therapeutic ultrasound system. The effect of the ultrasound treatment on gene expression was then analyzed by comparing each treated tumor with the nontreated tumor on the same mouse at different time points posttreatment using cDNA microarrays. Microarray analysis of 10k genes revealed consistently up-regulated expression of 23 genes throughout many time points including those related to stress response, induction of angiogenesis, and inflammation. Some of these diverse gene products include heat shock proteins (HSP 70, HSP 40), resistin-like alpha, cysteine-rich protein 61, and complement component 3. Our results show that by using a single external energy source (ultrasound), both physical (vascular/interstitial) and potentially molecular (gene up-regulation) levels of targeting are achieved simultaneously. In this way, we have developed a noninvasive double targeting delivery system for safe and accurate drug and gene delivery in vivo.
M.D. Ringler, None.
Abstract ID: 267
Genetically engineered antibody fragments are recognized as promising vehicles for delivery of imaging and therapeutic agents to tumor sites in vivo. The objective of this study was to design a fragment with improved tumor activity, scFv-Fc [(scFv-CH2–CH3γ1)2, 110 kDa], compared to previously studied diabody [(scFv)2, 55 kDa] and minibody [(scFv-CH3)2, 80 kDa]. The serum persistence of intact IgG and fragments containing intact Fc region is controlled by interactions with the protective FcRn receptor. Here, we developed scFv-Fc variants with mutations in the Fc/FcRn binding site to modulate their clearance rates. The anti-carcinoembryonic antigen (CEA) parental scFv-Fc and five mutants (scFv-Fc I253A, H310A, H435Q, H435R, and H310A/H435Q) were engineered by molecular cloning and site-specific PCR-based mutagenesis. Expression levels in the murine NS0 myeloma cell line ranged from 15 to 40 mg/ml culture supernatant. The protein purification scheme encompassed anion exchange, hydroxyapatite and a second round of anion exchange chromatography, resulting in more than 90% protein purity confirmed by SDS-PAGE and size exclusion chromatography. Effective in vitro antigen binding was demonstrated by a competition assay. Biodistribution studies in Balb/c mice using 125I- or 131I-labeled fragments revealed blood clearance rates from the longest to the shortest as follows: wild type > H435R > H435Q > I253A > H310A > H310A/H435Q. The terminal half-lives for the mutant scFv-Fcs ranged from 83.4 to 7.96 hr. Additionally, 124I-labeled H435Q and H310A/H435Q variants were evaluated in LS174T-xenografted athymic mice by microPET, revealing localization to the CEA-positive xenografts and low activity in normal organs. Attenuating the Fc/FcRn interaction provides a way of controlling the antibody fragments serum half-lives without compromising expression, binding, and tumor uptake. The mutant constructs I253A and H310A, intermediate in blood activity, could potentially be used in therapy. The double mutant, with its rapid clearance, suggests suitability for tumor imaging.
V.E. Kenanova, None.
Abstract ID: 268
Regulated gene expression can be achieved by use of a progesterone-inducible promoter system (Wang et al., 1994). The HSV-1–tkgfp fusion gene (Jacobs et al., 2003) was used to study coregulated TKGFP expression in culture and in vivo. To investigate whether noninvasive assessment of regulated HSV-1–tkgfp gene expression is possible, a regulated HSV-1 amplicon vector was generated by the use of the GeneSwitch system placing the HSV-1–tkgfp gene under transcriptional control of a mifepristone-regulated minimal synthetic GAL4-E1b promoter. To analyze regulation, human Gli36ÄEGFR glioma cells were infected with HSV-Switch-TG17 at various MOIs (0.03–1.0) in the absence or presence of the progesterone analogon, mifepristone. At 24 h after infection, GFP-positive cells (GPC) were counted and the ratio of GPC in the presence and absence of mifepristone was determined. A single-cell analysis of relative GFP-expression per cell was performed by means of MPI-Tool imaging software (Jacobs et al., 2003). [18F]FHBG-PET imaging was performed to correlate quantitative data in culture and in vivo. In the cell culture, the mean ratio of GPC with and without mifepristone was 14.1 ± 8.8, ranging from 3.6 to 35.0, independent of the MOI. The relative level of GFP fluorescence was 51.7 ± 15.5 relative fluorescence units (RFU) in the presence and 14.5 ± 4.4 RFU in the absence of mifepristone, representing a 3.6-fold induction with respect to GFP fluorescence intensity per single cell. Preliminary data in vivo demonstrate that HSV-Switch-TG17-regulated gene expression can be noninvasively assessed by [18F]FHBG-PET. Proportionality between quantitative inducible gene expression in cell culture and in vivo is currently being assessed. An inducible HSV-1 amplicon vector was engineered to regulate the universal HSV-1-tkgfp PET marker gene. The level of induction in cell culture varied depending on the assay (3.6- to 35-fold). This type of HSV-1 amplicon vector should serve noninvasive assessment of coregulated gene expression. Supported in part by MSWF 516–400 002 99, ZMMK-TV46, and DFG-Ja 981/1-2.
A. Winkeler, None.
Abstract ID: 269
The development of molecular imaging probes for monitoring beta-amyloid (Aβ) requires an understanding of the dynamics of Aβ production, aggregation, clearance, and trafficking as well their relationship to PET imaging. The objective of this study is to develop a computational simulation model of Aβ monitoring using PET. The model was tuned by adjusting the rate of production, clearance, and transport from different brain regions (hippocampus and cortex) both to and from the CSF and plasma. Parameters were adjusted to reproduce published levels of Aβ (fits were within 10%). Runs of the model were set up to examine both the long-term and short-term Aβ response to therapies. Time-activity curves for cortex, hippocampus, and plasma were generated using a simple PK/PD model. A modular Monte Carlo PET detector system model previously developed at GE was configured to model a small animal PET imaging system. Two candidate therapies were simulated: gamma-secretase inhibitors and passive immunotherapy directed against Aβ. The first showed a dramatic decrease in monomer concentration, while the decrease in total brain Aβ was delayed by the time constant of the reversible polymerization process. Simulated images taken 24 hr following therapy confirms a 37% reduction in free Aβ monomer versus a 3% reduction of total Aβ, including fibrils. The simulation of immunotherapy reproduced the previously published rise in plasma Aβ concentrations, yet, at levels below those experimentally measured (DeMattos et al., 2001). These results demonstrate the feasibility for using computational methods to predict posttherapy Aβ concentrations and to combine this with PET simulations for future imaging optimization. Although these data are promising, current models of Aβ production, aggregation, clearance, and trafficking are insufficient to accurately predict experimental measurements and additional improvements are necessary to realize its full potential.
M.C. Montalto, General Electric 5.
Abstract ID: 270
The mutant Herpes Simplex Virus 1 thymidine kinase (HSV1-sr39tk) positron emission tomography (PET) reporter gene (PRG)/suicide therapeutic transgene (TG) can be specifically imaged by the PET reporter probe [18F]FHBG in living subjects. We monitored [18F]FHBG accumulation into stable HSV1-sr39tk expressing C6 glioma (C6sr39) tumor xenografts for 6 weeks, including 14 days of Ganciclovir (GCV) treatment (100 mg/kg ip) and 3 weeks follow-up. We also attempted to image p53 TG by linking its expression to HSV1-sr39tk, using the two indirect covector administration and double promoter approaches (Ray et. al., 2001). On the average, 14 days of GCV treatment caused 81% regression of C6sr39 tumor volumes and 99% decrease in [18F]FHBG accumulation. Whereas, tumor volumes were 104% larger 3 weeks after halting GCV treatment than just prior to starting GCV treatment, [18F]FHBG accumulation was 91% less. GCV exposure causes elimination of [18F]FHBG accumulating C6sr39 cells and selects for regrowth of tumors unable to accumulate [18F]FHBG, lacking HSV1-sr39TK enzyme activity. The covector administration approach was evaluated using 2 identical replication deficient adenoviruses, 1 carrying the p53 TG (Ad-CMV-p53) and the other carrying the HSV1-sr39tk PRG (Ad-CMV-HSV1-sr39tk). We constructed a plasmid vector, consisting of p53 and HSV1-sr39tk, regulated by separate cytomegalovirus (CMV) promoters (CMVp53_CMVHSV1sr39tk) to evaluate the double identical promoter approach. To evaluate in vivo indirect imaging of p53, we coinjected various equivalent titers of Ad-CMV-p53 and Ad-CMV-HSV1-sr39tk (0-20 × 10 pfu) into the tail vein of nude mice (N = 5). Expression of p53 and HSV1-sr39tk correlated (r2 = .90 and .98) in dual adenovirus-infected and CMVp53_CMVHSV1sr39tk-transfected PC3 (p53 null) cells. There was a good correlation (r2 = .89) between [18F]FHBG (%ID/g) activity and the amount of p53 protein in coinfected livers. The dual identical vector and dual identical promoter bicistronic vector delivery methods can be used to link p53 and HSV1-sr39tk expression.
S.S. Yaghoubi, None.
Abstract ID: 271
Our institution has developed a local cancer therapy by direct implantation of drug delivery devices (e.g., polymer millirods) inside solid tumors using image-guided interventional procedures. In this treatment, understanding the drug transport from the delivery device into tumor tissues is critical to provide important pharmacokinetic information to evaluate the therapeutic efficacy. In this study, we describe the development of computed tomography (CT) as a noninvasive imaging technique to measure the dynamics of carboplatin release and distribution in rabbit livers in vivo. The presence of platinum atoms attenuates X-rays and generates the CT contrast without the necessity to radiolabel the drug. The CT study was carried out on a Phillips Mx8000 scanner. First, the acquisition conditions were optimized to achieve the highest sensitivity of detection. The sensitivity limit for carboplatin was determined to be 0.17 mg/cc. Second, the release kinetics of carboplatin from polymer millirods were measured in rabbit livers in vivo and compared to that in phosphate buffers in vitro. Quantitative image analysis showed similar release kinetics in the two environments. Third, dynamics of drug distribution in liver tissues were analyzed over time and compared with results from chemical analysis by atomic adsorption spectroscopy. Results from this study demonstrate the feasibility of CT as a noninvasive imaging technique to monitor drug release and tissue distribution in vivo. These data can be further used to optimize the implant design for a safe and effective local drug therapy.
J. Gao, None.
Abstract ID: 272
Sensitive, noninvasive in vivo imaging strategies are desirable for the early monitoring of treatment response in rheumatoid arthritis (RA). Such technologies would be valuable to initiate early appropriate therapy and dosing, judge clinical outcome, and allow more effective drug development. Furthermore, molecular reporters could serve as indicators for stage and severity of the disease, prognosis, and enable early diagnosis. Matrix degrading enzymes like certain cathepsins and matrix metalloproteinases are highly up-regulated in RA, leading to the destruction of inflamed joints. Therefore, these enzymes are suitable targets for in vivo imaging of early treatment response. Using a cathepsin B activatable, near-infrared fluorescence (NIRF) imaging probe, we examined whether methotrexate (MTX) treatment would alter cathepsin B activity. Our results show that 24 hr after intravenous injection of the reporter, inflamed paws of arthritic mice (n = 10 mice) showed sevenfold higher fluorescence intensity than paws of healthy mice (n = 6 mice). In MTX-treated animals (35 mg MTX/kg 48 hr prior injection of the probe, n = 10 mice), a significantly (p <.001) lower fluorescence signal (inflamed paws: 50%, inflamed toes: 70%) was observed. Histologically, fluorescence of the cathepsin B probe and cathepsin B antibody staining are localized in both major cell types in the hyperplastic synovia, synovial fibroblasts, and synovial macrophages. In conclusion, cathepsin-B-activatable NIRF imaging probes are suitable reporters for imaging of the early treatment response of antirheumatic drugs such as MTX. Moreover, the data show for the first time that MTX treatment results in a decrease of cathepsin B activity in the inflamed synovial tissue.
A. Wunder, None.
Abstract ID: 273
High-intensity focused ultrasound (HIFU) can be used to alter vascular permeability and interstitial barriers in a targeted and transitory fashion without resulting in tissue destruction. This technology has been utilized to target delivery of macromolecular agents, chemotherapeutics, and gene therapy. In order for HIFU to be universally implemented in the targeted delivery of a variety of agents in diverse tissue types, methods are needed to monitor the ultrasound-induced vascular changes. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) allows the noninvasive evaluation of the functional status of microvasculature. DCE-MRI allows the semi quantitative assessment of blood flow, and vascular permeability. C57/BL6 mice were implanted with mouse colon adenocarcinoma (MC38) cells subcutaneously. After tumors reached approximately 1.5 cm in diameter, HIFU was applied using a dual imaging/therapeutic ultrasound system. Mice were divided into three groups: no treatment, low-intensity FU treatment, and high-intensity FU treatment. Mice were imaged by DCE-MRI before and after HIFU treatment. Image-processing software allowed us to derive quantitative parameters for vessel density/flow (A) and vascular permeability (k21). Mean k21 values were increased by both low- and high-intensity FU treatments relative to the no-treatment group. Doubling the intensity of ultrasound treatment did not result in further increases in vascular permeability. We have demonstrated that DCE-MRI can detect pulsed HIFU-induced changes in k21, which has been shown to correlate with vascular permeability. Therefore, DCE-MRI may be an effective tool to noninvasively characterize the effects of HIFU in vivo and ultimately may guide its implementation.
G. Choy, None.
Abstract ID: 274
Radiation treatment planning that incorporates molecular image data, such as positron emission tomography (PET), relies on reconstructed images of radiopharmaceutical uptake to help delineate cancerous tissues for radiation targeting. PET image quality, however, depends on the type of reconstruction method used. The purpose of this study was to quantitatively compare target volumes derived from PET images reconstructed using analytical filtered backprojection (FBP) and statistical ordered subset-expectation maximization (OS-EM). The statistical nature of OS-EM allows for modeling of the physics of the imaging process in the reconstruction software, and thus to reduced image artifacts. These nontraditional volumes are also compared to more typical X-ray computed tomography (CT)-based target volumes. Five patients were imaged using CT and 18F FDG-PET prior to radiotherapy. PET data were reconstructed using FBP and OS-EM. Both CT and PET image data sets were then registered using a mutual information-based algorithm. Target contours were outlined on the CT data followed by the FBP and OS-EM PET image sets. Target volumes were then quantified using commercially available software. The average OS-EM target volume was 19.8 cm3 (range 3.9–37.9 cm3) and FBP volume was 16.9 cm3 (range 5.5–31.4 cm3). The OS-EM target volumes were, on the average, 17% larger than FBP. The average CT volume of the same target structures was 17.4 cm3 (range 5.8–30.6 cm3). A slight reduction in streak artifacts was seen in the OS-EM images when compared to FBP. In conclusion, differences were seen in target shapes and overall target volumes when comparing PET images reconstructed using FBP and OS-EM. These differences may translate to more accurate target coverage as a result of the reduction in image artifacts that occurs when using OS-EM, and thus to an improved ability to target cancerous tissue while avoiding surrounding normal structures.
C. Scarfone, None.
Abstract ID: 275
Polyethylenimine (PEI) derivatives are polycationic nonviral vectors for gene transfer. Nuclear imaging can play an important role in providing the information on biodistribution of polycation/DNA complex in living state. In this viewpoint, we thought that monitoring in vivo biodistribution of gene delivery vector could be approached using with nuclear imaging modality. The asialoglycoprotein receptor mediates internalization of proteins bearing galactose-terminated oligosaccharide moieties into hepatocytes. The aim of this study was to deliver the adequate amount of gene for treatment and to image galactosylated PEI nonviral vector in hepatoma-bearing transgenic mouse. For nuclear imaging of galactosylated PEI derivative, we modified galactosylated PEI-PEG with nicotinyl hydrazine derivative to HYNIC-PEI–PEG-Gal. Enhanced green fluorescent protein (eGFP) gene was complexed with HYNIC-PEG–PEI-Gal and then labeled with Tc-99m. Labeling efficiency of Tc-99m HYNIC-PEG–PEI-Gal/DNA complexes was determined with ITLC. DNA binding was studied by means of agarose gel retardation assays. The nuclear imaging was acquired after administration of Tc-99m HYNIC-PEG–PEI-Gal/DNA complexes via tail vein of transgenic mice at 10, 30, and 60 min. Using fluorescence microscopy, transgene expression was observed at 2 days after injection. HYNIC-PEG–PEI-Gal completely retarded DNA migration at a weight ratio of 9.7. Labeling efficiency of Tc-99m HYNIC-PEG–PEI-Gal/DNA complexes determined by ITLC was above 95% until 6 hr. Tc-99m HYNIC-PEG–PEI-Gal/DNA showed intense hepatic uptake in transgenic mouse. eGFP gene expression was mainly observed at the periphery of the tumor containing proliferating cells but appeared lower in normal hepatocyte. These data demonstrated that Tc-99m HYNIC-PEG–PEI-Gal/DNA complexes delivered gene into hepatoma cells successfully and acquired nuclear imaging for in vivo biodistribution of these complexes.
H. Jeong, None.
Abstract ID: 276
Retinoids are natural derivatives of vitamin A, which play important roles in modulating tumor-cell growth through the regulation of differentiation, thus supporting the potential use of retinoids in cancer therapy and prevention. To develop a molecular imaging method for monitoring the intranucleus response of retinoic acid (RA), we examined dual imaging reporter gene system, sodium/iodide symporter (NIS) and luciferase in cancer cell lines. The NIS/luciferase dual reporter gene was placed under the control of an artificial cis-acting retinoic acid enhancer (RARE) derived from pRARE-TA-SEAP (Clontech). NIS and luciferase genes were linked with IRES (internal ribosome entry site), named as pRARE-N/L and transfected to human hepatocellular carcinoma cell line, SK-HEP1 and human embryonal kidney cell line, HEK-293, which were RA-responsive cell lines with lipofectamine (Invitrogen). After treatment of RA (all-trans RA, 9-cis RA) with serial doses (up to 1 μM) for 48 hr, I-125 uptake and luminescence intensity were measured. Uptake of I-125 and luminescence intensity reached the maximum at 48 hr of treatment of RA. The pRARE-N/L-transfected cells in case of RA treatment accumulated up to about 3 times higher than did RA nontreated cells. Luminescence intensity increased as the same pattern according to increased doses of RA. Radioiodide uptake and luminescence intensity correlated well with RA, dose-dependent manner (n = 8, r2 = .948 and .968 in SK-HEP1- and HEK-293-transfected pRARE-N/L, respectively). To confirm that the increase of reporter gene expression was triggered by RARE mediated mechanism, a construct containing p53 response element (p53RE-N/L) instead of RARE was examined as a negative control. Radioiodide uptake of p53RE-N/L-transfected cell was not changed by RA treatment. This NIS/luciferase imaging reporter system is sufficiently sensitive to monitor RA response. The cis-enhancer imaging reporter system with RARE may be useful for the evaluation of RA response such as cellular differentiation and chemoprevention.
M. So, None.
Abstract ID: 277
Estrogen receptors (ER) are ligand-dependent transcription factors whose activity is modulated either by estrogens or by alternative signaling pathway. ERs interact via specific DNA-binding domain, estrogen responsive element (ERE), in promoter region of sensitive genes. This binding leads to an initiation of gene expression and hormonal effects. To determine the estrogen receptor activity, we developed a molecular imaging system using sodium iodide symporter (NIS) as reporter gene under the control of ERE. The hNIS reporter gene was placed under the control of an artificial ERE derived from pERE-TA-SEAP (Clontech) and named as pERE-NIS. pERE-NIS was transferred to MCF-7, human breast cancer cells, which highly expressed ER-α with lipofectamine (Invitrogen). Stably transfected MCF7/pERE-NIS cells were generated by selection with G418 for 2 weeks. After treatment of 17β-estradiol (E2) and tamoxifen with serial doses, the uptake of I-125 was measured for the determination of NIS expression. The inhibition study of NIS was performed with 1 mM potassium perchlorate. The MCF7/pERE-NIS treated with E2 accumulated I-125 up to 3–4 times higher than did nontreated cells. NIS expressions were increased the same pattern according to increasing doses of E2. MCF7/pERE-NIS treated with tamoxifen also accumulated I-125 up to 3 times higher than did nontreated cells. Potassium perchlorate inhibited the I-125 uptake completely. The pERE-NIS reporter system is sufficiently sensitive for monitoring estrogen receptor activity. Therefore, cis-enhancer reporter systems with ERE are applicable to the development of novel selective estrogen receptor modulator with low toxicity and high efficacy.
J. Kang, None.
Abstract ID: 278
L19 is a single chain antibody fragment (scFv) specific for the ED-B domain of a spliced variant of human fibronectin. ED-B fibronectin is an extracellular matrix protein linked to angiogenesis and expressed in solid tumors and their metastases as well as in tissue undergoing physiologic remodeling. It is an attractive antigen for targeting approaches in a variety of cancer indications as deduced from immunohistochemical analyses. L19 was labeled with I-123 or Tc-99m after introduction of a peptidergic chelator into the primary sequence. The affinity of these constructs lies in the low nanomolar range and maximum accumulation in a syngeneic teratocarcinoma xenograft (F9) mouse model is 8–10%ID/g (3 hr pi). ScFvs, because of their rapid accumulation in tumor tissue and fast blood clearance, are well suited to isotopes with short half-lives, such as F-18 (2 hr), Tc-99m (6 hr), and I-123 (13 hr) commonly used for imaging. Images with I-123- and Tc-99m-labeled L19 showed very good tumor-to-background ratios in a variety of xenograft mouse models at 5 and 24 hr. The radiotherapeutic efficacy depends on the therapeutic window, defined as the maximal dose that can be delivered to the tumor without causing severe side effects. Whereas the dose limiting organ for full size antibodies is usually the bone marrow, the use of smaller proteins and peptides (MW < 60 kDa) is limited by renal toxicity. The protein-radioisotope complexes are taken up and retained by the proximal tubular cells especially when labeled via lanthanide chelators. Size and format of antibody fragments, the choice of isotopes (e.g., I-131, Re-188, Y-90), and the linking/chelating technology are all variables influencing radiotherapeutic efficacy. We are currently investigating these parameters to determine the optimal construction of L19 for radiotherapy. First, therapeutic results with an I-131-labeled L19 derivative look promising.
L.M. Dinkelborg, None.
Abstract ID: 279
The ubiquitin-proteasome pathway is the central mediator of regulated proteolysis in cells, and defects in this pathway are associated with cancer and neurodegenerative diseases. To noninvasively monitor 26S proteasome function and inhibition in living animals, we developed a tetra-ubiquitin-luciferase reporter for bioluminescence imaging. The reporter was degraded rapidly in the proteasome under steady-state conditions, resulting in significantly reduced bioluminescence relative to unfused firefly luciferase. The ubiquitin-luciferase reporter was stabilized in a dose- and time-dependent manner in response to reversible and irreversible proteasome inhibitors. After one dose of the chemotherapeutic proteasome inhibitor bortezomib (PS-341), bioluminescence imaging of living mice bearing tumor xenografts showed that proteasome function was blocked within 30 min and returned to nearly baseline by 46 hr. However, after a 2-week regimen of bortezomib, imaging of target tumors showed significantly enhanced proteasome inhibition that no longer returned to baseline. The ubiquitin-luciferase reporter enables repetitive analysis of 26S proteasome activity in specific tissues in vivo. The ubiquitin-luciferase reporter should facilitate development and validation of proteasome inhibitors in mouse models as well as investigations of the ubiquitin-proteasome pathway in disease pathogenesis.
G.D. Luker, None.
Abstract ID: 280
In vivo MR spectroscopic imaging (MRSI) is a noninvasive technique that provides information about metabolic characteristics in tissue. For brain tumors, lactate is an excellent marker of anaerobic metabolism due to hypoxia and poor perfusion within the lesion and mobile lipids are elevated following degradation of membrane phospholipids. Reliable estimation of lactate and lipid levels are important for evaluating prognosis and guiding treatment planning for glioma patients. In conventional 1H-MRSI data, lactate and lipid peaks overlap within the region from 0.9 to 1.3 ppm and may be hard to distinguish. The goal of this study is to utilize a lactate-edited MRSI method for robustly measuring in vivo lactate and lipid levels in clinical brain tumor studies. Sixteen newly diagnosed glioma patients (6 low-grade and 10 high-grade) were studied prior to treatment. MR data were acquired on a 1.5-T GE scanner. J-difference lactate edited 3-D MRSI using Point Resolved Spectral Selection (PRESS) localization and Band Selective Inversion with Gradient Dephasing (BASING) pulses was performed with a reduced k-space sampling technique (nominal voxel size 1 cc, TR 1 sec and TE 144 msec). Spectra data were quantified automatically using software developed in our lab. Lactate and lipid resonances with height more than four times of the noise standard deviation were defined as significant. Significantly elevated lactate was observed in 3 out of 6 low-grade lesions (volumes ranging from 2.55 to 24.52 cc), suggesting that anaerobic metabolism cannot be considered as being uniquely characteristic of high-grade gliomas. None of the low-grade lesions had significantly elevated lipid peaks. Most of the high-grade lesions had elevated lipid (volumes ranging from 3.65 to 42.75 cc). This suggests that elevated lipid is the more reliable indicator of malignancy in glioma than lactate. Further studies will examine the relationship of the spatial distributions of lactate, lipid, and perfusion data, and will investigate the prognostic significance of these variables.
X. Li, None.
Abstract ID: 281
We have investigated the potential of 124I labeling of adenovirus to study its biodistribution as a function of time in mice by microPET and bioluminescence imaging. A recombinant adenovirus carrying the firefly luciferase reporter gene (Ad-CMV-fluc) was labeled with 124I by iodogen method. A total of 8 × 1011 viral particles of Ad-CMV-fluc were incubated with 1 mCi of 124I (specific activity ˜60 Ci/mg) at 37°C for 1 hr to obtain a 95% radiochemical yield. 124I-Adenovirus (0.1 mCi; 2 × 1010 viral particles) purified by size exclusion chromatography was tail-vein injected in four nude mice, and whole-body dynamic microPET images were obtained. Prior administration of potassium iodide (drinking water) and potassium perchlorate (gastric lavage) blocked thyroid and stomach uptake of free iodine. The radiochemical purity of 124I-adenovirus determined by ITLC was 99.0%, and it was stable for 48 hr at room temperature. Following injection of the 124I-adenovirus, activity was seen exclusively in the heart during the first 15 sec and subsequently in the lungs and liver. After 5 min, activity was seen predominantly in the liver while uptake in the heart and lung recedes. Radioactivity in the bladder, probably due to excretion of dissociated free iodine, was seen as early as 5 min. Optical imaging using a cooled CCD camera showed fluc expression in the liver (5.4e+06 ± 1.2e+06 photons/sec/cm2/sr) even 3 weeks after 124I-adenovirus injection, which was comparable to mice injected with nonlabeled adenovirus indicating the viability of labeled virus. This study demonstrates stable labeling of adenovirus with 124I using the standard iodogen method for microPET imaging without loss of virus viability or infectivity. Extensions of this work should help to better optimize viral trafficking studies for gene therapy.
G. Sundaresan, None.
Abstract ID: 282
The third leading cause of contrast nephropathy results from new, hospital-acquired etiologies of acute renal failure. This is due to increased use of radio-opaque contrast media (CM) during interventional techniques and diagnostic imaging. A carbonic anhydrase inhibitor (CAI) is being investigated as a preventative therapeutic agent of contrast-induced renal damage. Potential mechanisms of CAI indicate modification in contrast media perfusion and wash out. By implementing dynamic CT methods to track kidney blood flow and contrast media retention, the acute and chronic toxic effects of CM can be monitored and the protective capabilities of CAI can be better understood. To test this hypothesis, three groups of rats are studied: a negative control group on a normal diet, a positive control group on a modified diet that induces vulnerability to CM nephropathy, and a treatment group (induced vulnerability) receiving the drug acetazolamide. Serum creatinine levels are measured to monitor renal function. Preliminary results reveal a measurable increase in plasma serum creatinine levels in the control group after CM application. These results are also reflected in GFR values. Dynamic CT studies is being performed to exam its capability in determining kidney physiology. Both dynamic and high-resolution CT scans provide the quantitative assessment of blood kinematics, such as perfusion, blood volume, and tissue permeability, and the anatomical information to better delineate juxta-medullary and cortical nephrons. Physiological defects and CM washout rates will be examined and correlated with kidney anatomy to determine whether renal failure is a localized or systemic phenomenon and to quantify the protective effects of CAI.
K.M. Stantz, None.
Abstract ID: 283
Expression of wild-type p53 (wtp53) has been suggested to kill tumor cells by oxidative stress (Nature. 389:300, 1997) and 2-deoxyglucose (2DG) has been suggested to enhance tumor cell killing by agents that induce oxidative stress (Cancer Res. 63:in press, 2003). These observations have led to the hypothesis that 2DG could enhance tumor cell killing induced by wtp53. In the current study, PC-3 and DU-145 human prostate carcinoma cells (both lacking functional p53) were treated with 2DG combined with expression of wtp53, using a replication incompetent adenoviral gene transfer vector (Adp53). In both cell lines, a 24-hr pretreatment with 20 mM 2DG significantly enhanced (p <.05) clonogenic cell killing induced by exposure to 4 hr of Adp53 at a multiplicity of infection of 50 plaque forming units (pfu)/cell in cell culture experiments. Enhancement of cell killing by 2DG + Adp53 was more pronounced in DU-145 cells that consumed 8.3 ± 0.3 (imol glucose/1E6 cells/24 hr, relative to PC-3 cells that consumed 4.4 ± 0.3 p,mol glucose/1E6 cells/24 hr. Furthermore, accumulation of glutathione disulfide was monitored as an index of oxidative stress and was found to be greater in DU-145 cells treated with 2DG + Adp53, relative to PC-3 treated with 2DG + Adp53. Finally, when these cancer cells were grown as solid tumors (0.5 × 0.5 cm) in nude mice, the DU-145 tumors were found to be more sensitive to growth inhibition induced by intratumoral injection of 250 million pfu of Adp53 combined with intraperitoneal injection of 0.5 g/kg 2DG every other day for 2 weeks. These results support the hypothesis that 2DG enhances tumor cell killing by wtp53 as well as suggesting this enhancement may involve oxidative stress and be proportional to glucose consumption. (Supported by NIH P20-CA91709.)
D.R. Spitz, None.
Abstract ID: 284
Recent studies suggest that imaging with radiolabeled annexin V, an in vivo marker of cell death, may provide an early indication of the success or failure of anticancer therapy on a patient-by-patient basis within several days after the start of treatment. The utility of this measure, however, requires the determination of optimal time window(s) after the start of treatment, in which to perform annexin V imaging. To address this issue, we obtained simultaneous measurements of tumor burden and radiolabeled annexin V uptake following a single dose of chemotherapy in an orthotopic, luciferase-expressing murine lymphoma model using a combination of in vivo bioluminescence imaging and small animal single photon emission tomography. Multimodality imaging revealed the temporal patterns of tumor cell loss and annexin V uptake that will be useful for developing radiolabeled annexin V as marker of therapeutic efficacy in the oncology clinic. (NIH-HL-61717)
S.J. Mandl, None.
Abstract ID: 285
We have developed a targeted nanoparticle (NP) for both imaging angiogenesis and the delivery of genetic material to the endothelium involved in angiogenesis. The presence of the integrin αvβ3 on angiogenic vessels was utilized to target these vessels for antiangiogenic therapy in tumors using an integrin antagonist (IA) that binds to αvβ3 with high avidity. The integrin antagonist is covalently attached to the surface of a 60-nm NP that can also bind genetic material (plasmid DNA) electrostatically. We have previously shown successful targeted delivery of a dominant negative plasmid that contains the Raf mutant gene to the tumor endothelium of human melanoma xenograft in nude mice. Tumors were completely eliminated in 80% of the animals and reduced dramatically in size in the other 20% using only one dose of (ATPμ-Raf)IA-NPs (1 μg/g plasmid and 10 μg/g IA-NP). This antiangiogenic therapy, however, is successful only within a window of opportunity when the targeted receptor is highly up-regulated on the proliferating tumor endothelium. Optical imaging studies confirm the spatial and temporal changes in the presence of the integrins and gene expression in these tumors (Figure 2b). Angiogenesis imaging can therefore be used to define vascular proliferation patterns in tumors and screen for the presence of target receptors such as αvβ3 prior to vascular-targeted gene therapy. We were able to screen for tumors that will respond to vascular-targeted gene delivery using molecular imaging methods. Therefore, angiogenesis imaging will be a powerful technique to set the stage for developing clinical criterion for vascular-targeted gene therapy in human oncology.
S. Guccione, None.
Abstract ID: 286
Tumor necrosis factor related apoptosis-inducing ligand (TRAIL) induces apoptosis in neoplastic cells, and its controlled induction could be useful in tumor therapy. We have developed a vector delivery system that simultaneously allows a specific viral protease to control the secretion of TRAIL and to perform real time near infrared imaging (NIRF) in vivo. The two variants of TRAIL, a secreted one with the cDNA encoding the extracellular domain of TRAIL (aa114–281) and the other with an additional C-terminal fusion of HSV-1 protease substrate sequence and an endoplasmic reticulum (ER) retention sequence were cloned into HSV amplicons and packaged into helper virus-free vectors. Human glioma cells (Gli36) infected in culture with the HSV amplicons bearing these two different variants of TRAIL showed that the ER-tagged form had a significantly reduced apoptotic effect as compared to the secreted form. The cDNA encoding the active domain of HSV-1 protease (aa 1-245) fused to the extracellular domain of hftL3 was cloned into HSV amplicons and packaged into helper virus-free vectors. Coinfection of the ER tagged TRAIL ad HSV-1 protease resulted in the release of TRAIL from ER and the induction of apoptosis. We furthermore show that expression of the HSV-1 protease could be directly imaged in vivo by NIRF imaging following intratumoral injection of the HSV-amplicon vector encoding it. The simultaneous regulation of TRAIL-mediated apoptosis and NIRF imaging of protease activity should be widely useful for controlled induction of apoptosis and for real-time noninvasive imaging of gene expression in vivo.
K. Shah, None.
Abstract ID: 287
The standard imaging modality in radiotherapy treatment planning is computed tomography (CT). Positron emission tomography (PET) provides molecular data that may alter radiation target volumes. This study assessed the accuracy of registration of independently acquired CT and 18F-FDG PET images and PET's impact on CT-based target volumes in patients with head and neck cancer. Five patients were custom fitted with head and neck immobilization devices, which were worn during both the CT and PET scans and fixed with external fiducial markers containing CT contrast and FDG. CT and PET images were acquired on separate, dedicated scanners and registered using internal structures in a mutual information-based algorithm. Accuracy was assessed by comparing structure-based registration with fiducial-based registration; a target registration error (TRE) was calculated. Gross tumor volume (GTV) was defined with contours on the CT and PET image sets. A new contour was then drawn on the CT image set that incorporated pertinent data from the PET contours. CT volumes were compared to CT/PET volumes. Accuracy was calculated as the mean and standard deviation of all TREs in the image set. The average local TRE was 3.13 ± 0.66 mm. Based on the comparison of CT and CT/PET contours, 4 of the 5 GTVs were altered as a result of the PET contouring information. The 4 CT/PET GTVs were larger than the original CT GTV by an average of 169% (range 17.4–425). Four of five patients who received PET in addition to CT as part of their work-up had significantly altered GTVs based on the additional information provided by PET; the final radiotherapy treatment plan dose distribution was also significantly changed. The inclusion of PET in the contouring process provides metabolic data that may improve the ability of conformal intensity-modulated radiotherapy (IMRT) to optimize dose distribution and improve loco-regional tumor control.
W.C. Lavely, None.
Abstract ID: 288
M. Subbarayan, None.
Abstract ID: 289
The margin of brain tumor has been routinely visualized and defined with the brain region exhibiting extra MR signal enhancement after a dose of intravenous injection of contrast agent, such as Gd-DTPA. The same information about the leaky BBB associated with tumor has been used to plan its biopsy and resection as well as radiation therapy. Abnormal signals detected by the magnetic resonance spectroscopic imaging (MRSI) reflects the deviation in concentration of brain metabolites from that of normal, this provides additional information regarding to neurochemistry that can be used to characterize an unknown lesion noninvasively. In order to compare the margins defined by the 2 different imaging methods, 6 brain tumor patients were studied simultaneously with both imaging schemes. The MRSI tumor margin was delineated by the increased choline signal areas. The preliminary result showed generally that there was a significant morphological difference in terms of the margin defined by MRSI and CE-MRI. This difference in margin can not be fully counted by the spatial resolution difference between the 2 techniques. MR-guided neurobiopsies showed that the elevated choline signal has been found to be consistent with the existence of rapid tumor cell proliferation in the corresponding area. The actual neurobiopsy guided with the spectroscopic imaging method demonstrated that it could provide the valuable information that is complimentary to the existing examinations in delineating the tumor margin especially in the case involving nonenhancing tumor.
H. Liu, None.
Abstract ID: 290
Currently available therapeutics for malignant glioma have limited curative potential. Hence, we have applied, for the first time, glioma-targeted immunotherapy utilizing genetically modified T-cells. The development of this novel therapeutic approach has been facilitated by the engineering of an Interleukin-13 (E13Y)-zetakine chimeric immunoreceptor (IL-13Zet), which when expressed by genetically modified primary human cytotoxic T lymphocytes (CTLs), engages the glioma-restricted IL-13 cytokine receptor (IL-13Rα2), resulting in activation of glioma-specific CTL cytolytic and cytokine effector mechanisms. We hypothesize that CNS regional or systemic adoptive therapy utilizing these glioma-specific CTL clones will yield good responses compared to studies employing nonspecific effector cells. Correlative laboratory work will allow us to identify surrogate CSF cytokine markers of tumor burden and in situ T-cell activation/persistence, to screen for systemic immune responses directed against the zetakine, and to assess the incidence of tumor escape by selection of IL-13Rα2 antigen-loss glioma variants. Insights gained from these studies, along with the imaging studies highlighted here, will facilitate the design and implementation of subsequent work employing optimized T-cell dosing, imaging modalities, and surrogate endpoints. Currently, we are working in mouse models to develop imaging and surrogate marker analytic capabilities for quantifying in vivo antitumor activity of adoptively transferred IL-13Zet re-directed CTL clones. We have selected imaging modalities based on their ability to accurately assess tumor burden in this setting. We have used bioluminescence and MRI to image both the human 251 glioma cell line in vivo and a more tumorogenic subclone, 251T, as well as the T-cells themselves, and are assessing growth of the tumors and location of T-cells post injection. An update on the progress of the in vivo imaging in these studies will be provided.
R.A. Moats, None.
Abstract ID: 291
We are developing agents for the real-time guidance of surgical procedures directed at tumors, in this case ovarian cancer. Folate receptor (FR) is overexpressed in certain cancers, particularly ovarian cancer. One of us (PSL) developed a type-II FR targeting moiety, and this agent was previously developed into an injectable radiopharmaceutical imaging agent. As this agent is a small molecule, it rapidly equilibrates in tissues, with tumor/blood ratios of 400:1 as the agent is renally excreted. There are also native FR binding sites in the kidney, but little elsewhere. We set out to develop an optical agent targeting FR for real-time imaging. An FR-targeting moiety was coupled to a Cy-Dye (Cy-7) agent. Nude mice were injected with KB cells, a human nasopharyngeal cell line that constitutively overexpresses FR. Tumors were injected subcutaneously and intraperitoneally with 1 μmol/kg and allowed to grow for 10–21 days. FR-Cy Dye was injected either intravenously or intraperitoneally. Imaging was performed with the Palomar™ room-light imaging system, developed for OR use. After injection, images were taken for 2 hr, then intermittently every 12–24 hr for up to 3 weeks. Images at injection show a flush (Figure A) over agent throughout the body. By 1 hr, the kidneys and tumor glow strongly (Figure B). By 1–7 days, the tumor and kidney show tumor/background ratios of 20:1. In animals without tumor, only the kidneys light up (Figure C). We developed an agent targeting FR receptor. This agent will be investigated as a candidate for human use.
Real-time images of the folate Receptor (FR) targeting agent talen using the real-time room-light Palomar™ system.
D.A. Benaron, Stanford University School of Medicine AND Spectros Corporation 1, 4, 5.
Abstract ID: 292
Recent developments in nanoparticle technology have shown that these may play an important role in the treatment of tumors by acting as specialized and targeted therapeutic and diagnostic agents. In this presentation, a novel dynamic polyacrylamide SPIO (pa-SPIO) NP platform for imaging of brain tumors by MRI is demonstrated. These particles have extremely large spin-spin relaxivities between 600 and 800 L sec−1mmol−1. The dynamic NP platform is different from other SPIOs in that the iron oxide particles are not coated but are chemically incorporated into the polymeric matrix. Furthermore, by addition of different sized PEG, attachments to the surface of the NPs can be used to vary their plasma half-life. The flexibility within the formulation of these NPs suggests that these particles could be used as a platform technology for targeted NP contrast agents. Pa-SPIOs were injected intravenously into Fischer-344 rats bearing intracerebral 9L gliomas with the aim of understanding the pharmacokinetic uptake of NPs within the animal model vasculature, brain, and tumor. Pa-SPIOs with 4 different sized PEG extensions were used in order to determine the optimized composition for maximum uptake in the tumor and longest half-life in the blood. In conclusion, it was found that by increasing the length of PEG attachments to the NPs, the plasma half-life could be extended from ˜30 min (no PEG) to ˜3 hr (10 kDa PEG). This platform technology can be easily adapted for molecular targeting of contrast and therapeutic agents as a molecular imaging agent.
Gradient-echo MRI time course of the rat brain containing a 9L glioma, at 0 s, 30 s, 2 min, 15 min and 1 hr following intravenous injection of PA-SPIO NPs incorporating (a) no PEG and PEG extension of 10 kD.
B.D. Ross, Molecular Therapeutics, Inc. 4.
Abstract ID: 293
Development of novel suicide cancer gene therapy paradigms is the goal of numerous laboratories. Assessment of therapeutic enhancements achieved through genetic manipulations of gene constructs requires a significant amount of effort to quantitate the biological outcome. Molecular imaging could provide early and quantitative feedback of comparing 1 gene construct to modifications of the construct in order to rapidly optimize the therapeutic effectiveness. The aim of this study was to evaluate whether MRI and 19F MRS could be used to quantitatively asses the therapeutic efficacy of 2 different approaches to yeast cytosine deaminase (yCD) gene therapy of the rat 9L glioma. Two separates Table 9 Lglioma cell lines that expressed yCD and yCD fused to uracil phosphoribosyltransferase (yCD-UPRT) were established in cell culture. These clonal lines were implanted subcutaneously into the flank of nude mice for analysis by 19F MRS and intracranially into Fischer-344 rats for studies using T2- and diffusion-weighted MRI. In the case of 19F MRS, a single intraperitoneal dose of 5-fluorocytosine (5FC: 1 g/kg) was given, and the 19F signal was obtained at 20-min intervals for 6 hr to quantitate the specific activity of the fusion constructs. MRS studies revealed that the conversion of the prodrug (5FC) to therapeutic product (5-fluorouracil) was dramatically enhanced for the yCD-UPRT genetic construct over the yCD construct. The more efficient observed conversion of the prodrug to chemotherapeutic product by the yCD-UPRT modification should in practice translate to a greatly enhanced therapeutic outcome. In order to investigate this possibility, we studied the therapeutics effectiveness of 5FC treatment on each of these cell lines using diffusion MRI. Diffusion maps revealed that the yCD-UPRT fusion construct produced a significantly greater and more efficient cell killing than the yCD construct. In conclusion, molecular imaging provided noninvasive insights for evaluating this cancer gene therapy paradigm.
B.D. Ross, Molecular Therapeutics, Inc. 4.
Abstract ID: 294
An experimental cancer gene therapy model was employed to develop noninvasive imaging procedure using radiolabeled FIAU as an enzyme substrate for monitoring retroviral vector-mediated HSV1-tk transgene expression. [131I]FIAU was prepared by a no-carrier-added (NCA) synthesis process and lyophilized to give “hot kits.” The stability of [131I]FIAU in the form of lyophilized powder (the “hot kit”) was much better than in the normal saline solution. In cellular uptake, the NG4TL4-STK cells accumulated more radioactivity than NG4TL4 cells in all conditions and increased with time up to 8 hr. The kinetic profile of the cellular uptake of NCA [131I]FIAU formulated from the lyophilized “hot kit” or from the stock solution is qualitatively similar. For animal model cancer gene therapy studies, FVB/N mice were inoculated subcutaneously with the HSV1-tk(+) and tk(–) sarcoma cells into the flank to produce tumors. Biodistribution studies showed that tumor/blood ratios were 2, 3.5, 8.2, and 386.8 at 1, 4, 8, and 24 hr post injection, respectively, for the HSV1-tk(+) tumors; and 0.5, 0.5, 0.7, and 5.4, respectively, for the HSV1-tk(–) tumors. The significant difference in radioactivity accumulation was revealed among the HSV1-tk(+) tumors, the tk(–) tumor, and other tissues. The highest activity retention was observed 24 hr post injection in HSV1-tk(+) tumors (9.67±3.89%ID/g) compared with that in HSV1-tk(–) tumors (0.48±0.19%ID/g). After 7 consecutive daily treatment of prodrug ganciclovir, planar gamma camera imaging showed HSV1-tk(+) tumor regression at Day 4, and complete tumor regression at Day 7. These results clearly demonstrate that the simplified NCA synthesis process developed in this study is reliable, and the [131I]FIAU product is useful for in vivo radio-image monitoring of HSV1-tk gene transfer, expression, and therapy.
W. Deng, None.
Abstract ID: 295
The goal of the following studies was to develop a molecular imaging method that can measure vascularity of specific regions within a tumor and therefore regional response to antiangiogenic therapy. The hypothesis of this work was that tumor perfusion as measured by arterial spin tagging MRI (AST-MRI) would vary depending on tumor vascularity. To test these hypotheses, 9L cell lines expressing low, wild-type, and high levels of VEGF were transplanted intracerebrally into Fischer-344 rats (6 animals per cell line). Tumor diffusion, growth rate, and tumor perfusion were then measured using an MRI protocol that included, T2-weighted imaging, diffusion-weighted imaging, and AST-MRI. The results of this imaging were then correlated to both histology and VEGF levels as determined by Northern blot analysis. The results of the diffusion MRI showed that despite different levels of VEGF within the tumor types, the cellularity was unchanged. However, both tumor perfusion and tumor growth rates were significantly different among the 3 tumor genotypes. The 3 cell lines varied with respect to in vivo doubling times and the high VEGF tumors also had dramatically increased mean tumor blood flow (400 ± 80 ml/100 g/min) and blood flow heterogeneity over the wild-type (26 ± 17 ml/100 g/min), while the low VEGF tumors had a mild but significant increase in blood flow (70 ± 25 ml/100 g/min) but no significant difference in perfusion heterogeneity. Because tumor perfusion as measured by AST-MRI has the ability to measure the degree and heterogeneity of vascularization within tumors, it has the potential to become an extremely important and powerful surrogate marker for determining efficacy of antiangiogenic therapy.
B.D. Ross, Molecular Therapeutics, Inc. 4.
Abstract ID: 296
Tc-99m recombinant human annexin (Tc-rh-annexin) has been shown to bind to phosphatidylserine (PS) in vitro and in vivo. PS exposure on the exterior of the cell membrane double layer is considered an invariant accompaniment of apoptosis. A radiopharmaceutical kit incorporating hydrazinonicotinamide (hynic) linked to recombinant human annexin V as a chelator of technetium has been developed. Tc-99m hynic-annexin has been studied in Phases I and II clinical trials. Studies in subjects with Stage IIIB or IV non-small cell lung cancer at least 1 cm in diameter suggest occasional PS expression prior to initiation of anticancer therapy. This finding suggests varying degrees of apoptosis in late-stage tumors and comforts with the results of prior biopsy studies in this disease. Increased posttherapy PS expression, as indicated by increased annexin localization within 24–48 hr of initial exposure to chemotherapy, has been seen only in NSCLC subjects who have demonstrated an objective response to therapy (greater than 50% reduction in tumor size). Posttherapy NSCLC tumor PS expression appears to be heterogeneous, suggesting significant variability within the tumor mass in responsive late-stage disease. Studies in 12 subjects with follicular cell lymphoma treated with ionizing radiation (2 doses of 2 Gy) have shown early (24–48 hr) posttherapy increases in PS expression in clinical responders, but not in a single nonresponder. These 2 studies suggest that PS expression, as indicated by Tc-hynic-annexin localization, occurs in at least some responders to chemotherapy or radiation therapy and may be a clinically useful biomarker of early response.
A.M. Green, North American Scientific, Inc. 4, 5.
Poster Session 10: Mechanisms of Oncogenesis and Disease Progression in Tumor Models Abstract ID: 297
Malignant human mammary epithelial cells (HMECs) typically exhibit elevated phosphocholine (PC) and low glycerophosphocholine (GPC), whereas low PC and high GPC levels are typical of nonmalignant HMECs. To discover the molecular origins of the differences in choline metabolism between malignant and nonmalignant HMECs, we performed 1H and 13C NMR spectroscopy of cells labeled with [1,2-13C]-choline. Microarray-based gene expression analyses of malignant and nonmalignant HMECs were performed in parallel studies. Nonmalignant MCF-12A HMECs and malignant MDA-MB-231 HMECs were incubated with [1,2-13C]-choline to identify pathways of choline metabolism. Cell extract fractions were analyzed by 13C and 1H NMR spectroscopy. Microarray gene expression analyses of MCF-12A and MDA-MB-231 cells were performed with the Human Genome U133 GeneChip Set (Affymetrix). High PC levels were evident in malignant MDA-MB-231 HMECs, together with depletion of most of the free intracellular choline (Cho). In contrast, nonmalignant MCF-12A HMECs contained similar amounts of PC and Cho. 13C Enrichment of the PC pool was significantly increased in long-term versus short-term [1,2-13C]-choline exposure in malignant HMECs but not in nonmalignant HMECs. Microarray analysis revealed that breast cancer cells overexpressed choline kinase and phospholipase C, but underex-pressed lysophospholipase 1, cytosolic calcium-dependent phospholipase A2 and phospholipase D1. Malignant HMECs exhibit a higher rate of choline phosphorylation by choline kinase compared to normal HMECs, which is consistent with the overexpression of choline kinase detected in the micorarray analyses. The reduction of fractional 13C enrichment of PC in the short-term experiments in malignant HMECs may be due to a higher phospholipase C activity, again consistent with overexpression of phospholipase C. Underexpression of lysophospholipase 1 and phospholipase A2 may account for the decreased GPC levels detected in malignant HMECs. The combined NMR spectroscopy and microarray gene expression analysis approach has provided further insight into the molecular mechanisms underlying the aberrant choline metabolism of cancer cells.
K. Glunde, None.
Abstract ID: 298
To investigate the diagnostic value of arabinogalactan (AG)-coated ultrasmall iron oxide (USPIO)-enhanced MR imaging for the detection of hepatocellular carcinoma in transgenic mice expressing the Hepatitis B virus X-protein, compared to USPIO-enhanced MR imaging. Eight transgenic mice with hepatocellular carcinoma were enrolled in this study. All animals underwent evaluation with precontrast and postcontrast MR imaging. The mice were allocated into 1 of 2 groups: USPIO group (n = 4) and AG-USPIO group (n = 4). The MR examination protocol included T1-weighted spin-echo (SE, TR/TE = 400/15 msec) and T2-weighted spin-echo (TR/TE = 4000/54 msec) before and 1 hr after administration of USPIO and AG-USPIO. Gold standard was provided by findings at autopsy. For the quantitative analysis, we measured the signal intensity of the mass, liver, and noise with region of interest technique and we calculated SNR of the mass and liver, and lesion to liver contrast-to-noise ratio (CNR). For quantitative analysis, 2 experienced abdominal radiologists compared 2 kinds of image in terms of lesion conspicuity and delineation. For quantitative analysis, AG-USPIO-enhanced MR images showed the better lesion to liver CNR than USPIO-enhanced MR images (16.7 vs. 9.1, P = 0.44) and similar parenchymal signal loss ratio of the liver (66% vs. 64% P = 0.8). For qualitative analysis, AG-USPIO-enhanced MR images showed better lesion conspicuity and delineation than USPIO. This experimental study suggests that MR receptor imaging using USPIO particles targeted to asialoglycoprotein receptors could be better than USPIO for the detection of hepatocellular carcinoma.
J. Lee, None.
Abstract ID: 299
Metastatic melanoma mouse models have been well established in the literature. These models have typically focused on the development of lung metastases and have been characterized by various histopathology techniques. Our objective in this study was to characterize both the time progression and extent of melanoma metastases in mice by using noninvasive micro-CT imaging correlated with histopathology. Eight-week-old female C57BL/6 mice were injected intravenously with B16-F10 melanoma cells via the tail vein. The mice were then serially imaged under anesthesia using an ImTek MicroCAT II scanner. Mice were necropsied either when adverse clinical signs developed or at Postinjection Day 24, and tissue samples were collected for histopathology. At necropsy and on histopathology, metastatic lesions were confirmed in several different tissues including lung, brain, and bone. On histopathology, lung metastases varied from multiple micrometastases to large (>4 mm diameter) discrete tumors. Examples of bone metastases included tumors in the mandible and iliac crest, which varied in histologic appearance. Using micro-CT, the development and progression of metastatic tumors were followed. For lung metastases, lesions were discerned within 14 days post injection, and the progression followed until necropsy, shown below. Osteolytic bone degradation was observed in a mandibular metastasis. Additional lesions were not detected by micro-CT imaging either due to the small size of the lesion or inadequate soft tissue contrast between tumor and normal tissue. Micro-CT imaging was found effective in detection and characterization of lesions produced by this metastatic melanoma model. Micro-CT imaging can be coupled with SPECT imaging using radiolabeled biomolecules that selectively target B16-F10 tumor cells via the α-MSH receptor, combining micro-CT anatomical detail with functional SPECT images to follow metastatic tumor development.
C.T. Winkelmann, None.
Abstract ID: 300
We compared the in vitro cellular metabolism and the in vivo biodistribution of radiolabeled acetate (AC), choline (CH), and fluorodeoxyglucose (FDG) in human prostate cancer xenografts. The in vitro cellular incorporation of [14C]-AC, [14C]-CH, and FDG were determined after incubation of these tracers with androgen-independent (PC3) and androgen-sensitive (CWR22) human prostate cancer cells. The in vivo biodistribution studies were performed after subcutaneous injection of PC3 and CWR22 cells in castrated and noncastrated athymic male mice. Biodistribution studies were performed after intravenous administration of 1.25 μCi AC and CH and 25 μCi FDG. Animals were sacrificed at 5, 10, and 20 min for AC and CH and after 30, 60, and 90 min for FDG. Autoradiography was performed 20 min after intravenous administration of 1.25 μCi AC and CH. Dynamic MicroPET imaging was performed for 60 min immediately after intravenous administration of 200 μCi FDG followed by autoradiography. The in vitro study demonstrated higher cellular uptake ratio in PC3 than in CWR22 for both AC (3:2) and FDG (1.8:0.3) but not for CH (1.7:4). The in vivo biodistribution studies in noncastrated mice showed higher tumor-to-muscle (T/M) uptake ratios in PC3 than in CWR22 for AC (1.6:0.9 at 10 min) and similar ratio for CH (1.2:1.7 at 10 min) and FDG (2.9:2.8 at 60 min). Autoradiography and MicroPET demonstrated higher T/M FDG uptake ratio for PC3 than for CWR22 in the castrated mice in comparison to the noncastrated mice (4:1 vs. 2:1, respectively). Either AC or FDG may be suitable for PET imaging of advanced prostate cancer while CH may be appropriate for imaging androgen-responsive early disease. The level of FDG uptake in the androgen-sensitive tumor is directly dependant on the androgen presence and stimulation. (This work was supported by a grant from The Wright Foundation.)
X.S. Chen, None.
Abstract ID: 301
Noninvasive molecular functional imaging is critical in early diagnosis, designing individualized therapy, and monitoring early therapeutic response for cancer patients. Near-infrared (NIR) optical imaging offers unique advantages due to its high sensitivity, versatile fluorescent probe design, and the ability to provide real-time data. Herein, we report the in vivo NIR optical imaging of human tumors in ectopic or orthotopic xenograft models. A NIR dye Cy5.5 (excitation/emission, 675/694 nm) peptide conjugate was used in the imaging. Human breast cancer cells MDA-MB-468, SKBr3, human fibrosarcoma cells HT1080, and human Kaposi's sarcoma KS1767 were inoculated subcutaneously. Human prostate cancer cell PC3 was inoculated intratibially. Human glioma cell U87 was inoculated intracranially. In vivo imaging studies were performed on day 3 after the inoculation of PC3 cells or when tumor size from 2 mm to 1.2 cm for the other tumor models. The imaging properties were assessed using continuous wave fluorescence imaging accomplished via an intensified charge coupled device camera. Images were obtained 24 hr following the tail vein injection of contrast agent at an equivalent dose of 1–15 nmol of Cy5.5. Excitation light was delivered via an expanded laser diode beam to the surface of the mouse such that the entire mouse was illuminated. In vivo NIR imaging showed selective accumulation of the contrast agent in the tumor after the conjugate injection. Our data demonstrate that NIR noninvasive optical imaging can detect small tumors regardless of the site of inoculation. Such an imaging technique maybe used in early tumor detection. (Supported by NIH grants R01 EB00174, Prostate Cancer SPORE Career Development Award CA90271, and Texas Advanced Technology Program).
S. Ke, None.
Abstract ID: 302
MLN2704 is an investigational drug currently in Phase I clinical trials and is designed to deliver the maytansinoid chemotherapeutic agent DM-1 directly to prostate cancer cells through the targeting of MLN591, a deimmunized monoclonal antibody to prostate-specific membrane antigen (PSMA). Here, we have extended our preclinical evaluation of MLN2704 in a novel model of prostate cancer bone metastasis. Our model is based on the 22Rv1 prostate cancer cell line, which is an androgen-independent derivative of the CWR22 prostate cancer xenograft. We first engineered PSMA-positive 22Rv1 cells to express firefly luciferase. Clones were selected and evaluated for PSMA and luciferase expression as well as sensitivity to MLN2704. In vitro analysis indicated that sensitivity to MLN2704 was correlated with PSMA expression levels. One luciferase-expressing clone 22Rv1-luc1.17 was selected for analysis in vivo. A total of 1.5×105 22Rv1-luc1.17 cells were injected into the right tibiae of immunocompromised mice. Luminescence signal, detected with Xenogen IVIS™, increased over time and endpoint tumor volume, as assessed by MRI, was proportional to luminescence. Radiographic and histologic analysis revealed that intraosseus 22Rv1-luc1.17 tumors developed osteoblastic or mixed osteoblastic/osteolytic pathology within 10 weeks after injection. While osteoblastic lesions are a typical clinical finding associated with bony metastases in prostate cancer, this feature of the disease has been difficult to reproduce in animal models and has hindered biological understanding and drug development in this area. Next, we assessed MLN2704 antitumor efficacy in this model. Utilizing the IVIS™ to monitor tumor growth, we found that a dose of 240 μg/kg (DM1 equivalents) MLN2704 delivered on a q3d schedule significantly delayed intraosseus tumor growth compared to a nonspecific antibody conjugate, which demonstrated no efficacy in this model. Micro-CT analysis of treated mice revealed that MLN2704 treatment also reduced the presence of osteoblastic lesions.
M.D. Henry, Millennium Pharmaceuticals, Inc. 4, 5.
Abstract ID: 303
Solid tumor growth is restricted by its ability to acquire new vasculature in a process termed angiogenesis. Vascular endothelial growth factor (VEGF), the most potent mediator of both physiologic and pathologic angiogenesis, is currently under investigation as a target for antiangiogenic cancer therapy. In order to further current understanding of tumor-induced VEGF expression, this research utilizes a novel transgenic mouse model (VEGFp-GFPLuc [VGL]) possessing a human VEGF promoter driving expression of a luciferase reporter gene. This model allows real-time, noninvasive visualization of both tumor and host VEGF promoter activity in vivo by monitoring luciferase luminescence. In order to examine host VEGF expression, mouse mammary tumor virus (MMTV) tumor cells were orthotopically injected into the VGL model. Tumor VEGF expression was examined by injecting MMTV tumor cells with the VGL transgene (MS+ cells) into wild-type mice. Combined VEGF expression was examined by injecting MS+ tumor cells into the VGL model. In addition, crossing VGL and MMTV transgenic mice yielded a spontaneous tumor model (MMTV/VGL) also representing combined VEGF response. All mice were imaged periodically for luciferase expression over the duration of tumor progression. Correction for VEGF expression due to wound healing and growth was determined by control experiments. Results indicate VEGF expression increases with tumor volume in both the combined and tumor models. Host VEGF expression decreases during tumor progression suggesting tumor cells are responsible for the majority of VEGF production. The balance between tumor and host VEGF expression may elucidate important information on tumor-host interaction as well as mechanisms that facilitate tumor survival. The VGL model is validated as a useful tool for investigating VEGF expression related to tumor growth and progression.
S. Faley, None.
Abstract ID: 304
Due to its inability to be metabolized and cleared by membrane-associated phospholipases, NM404, a second-generation phospholipid ether analog, is undergoing evaluation as a potential tumor-selective diapeutic agent. Although radioiodinated NM404 has displayed striking tumor avidity in 20 out of 20 spontaneous and animal/human xenograft tumor models examined, it has never been evaluated in a pancreatic tumor model. The aim of this study was to examine the tumor avidity of NM404 in the c-myc mouse pancreatic adenocarcinoma model, which is known to produce invasive tumors with mixed acinar/ductal phenotype. NM404 recently received IND approval for diagnostic evaluation in both human prostate and lung cancer patients. Moreover, due to its apparent universal success in imaging a variety of tumor types, it is also undergoing extensive preclinical evaluation in breast, colon, and glioma cancer models. NM404 was radioiodinated with 125I via isotope exchange, solubilized in aqueous 2% Tween-20, and injected (15 μCi/20 g mouse) into c-myc mice (6) bearing pancreatic tumors Anesthetized mice were scanned for up to 21 days post NM404 injection with a modified Bioscan AR2000 radio-TLC scanner and with an ImTek microCT scanner for anatomic comparison and image fusion analysis. At sacrifice, tissues were imaged ex vivo, lesions were excised, weighed, and radioactivity quantitated. Tumors were sectioned and submitted for histologic classification. Initial imaging results have shown striking uptake and prolonged retention (>14 days) of NM404 in spontaneous pancreatic adenocarcinomas ranging from 5–12 mm in diameter. NM404 appears to be selectively retained in pancreatic tumors in this c-myc mouse model, thus, further extending the tumor selectivity of this agent. These findings suggest that NM404 holds considerable promise for the detection and characterization of a variety of cancers including pancreatic adenocarcinoma.
J.P. Weichert, None.
Abstract ID: 305
Due to its inability to be metabolized and cleared by membrane associated phospholipases, NM404, a second-generation phospholipid ether analog, is undergoing evaluation as a potential tumor-selective imaging agent. Radioiodinated NM404 has displayed striking tumor avidity in 20/20 animal and human xenograft and spontaneous tumor models examined. The aim of this study was to examine the tumor avidity of NM404 in the spontaneous ApcMin/+ mouse mammary carcinoma model. This model is unique in that tumors ranging developmentally from hyperplasias to carcinomas are formed in the same animal, thus allowing evaluation of the specificity of NM404 for cancer cells. NM404 recently received IND approval for diagnostic evaluation in human prostate and lung cancer patients. Due to its apparent universal success in imaging a variety of tumor types, it is now undergoing extensive preclinical evaluation in lung, pancreas and breast cancer models. NM404 was radioiodinated with 125I, solubilized in aqueous 2% Tween-20, and injected (15 μCi/20 g mouse) into 8 female ApcMin/+ mice. Anesthetized mice were scanned for up to 21 days post injection in a modified Bioscan AR2000 radio-TLC scanner and an ImTek microCT scanner for anatomic comparison and image fusion analysis. At sacrifice, tissues were imaged ex vivo, lesions were excised, weighed, and radioactivity quantitated. Lesion samples were submitted for histologic classification. Initial imaging results have shown striking uptake and prolonged retention (>21 days) of NM404 in mammary adenocarcinomas and squamous cell carcinomas ranging from 2–15 mm in diameter, but little or no uptake in mammary focal alveolar hyperplasias. NM404 appears to be selectively retained in malignant tumor cells based on preliminary results in this endogenous mouse tumor model. These findings suggest that NM404 holds considerable promise for the detection, characterization, and possibly treatment of both primary and metastatic breast cancer.
J.P. Weichert, None.
Abstract ID: 306
Small-animal imaging in mouse models of human cancer promises more precise and efficient analysis of tumor progression, and reduction of the number of animals needed for statistical power in preclinical therapeutic intervention trials. Of the various imaging modalities adapted to small-animal imaging, positron emission tomography (microPET) holds particular promise for imaging mouse models of cancer because of its high sensitivity and the quantitative nature of the tomographic data. The study reported here monitors the longitudinal development of disease in a mouse mammary tumor model derived from transgenic mice constructed with the polyoma-virus middle-T oncogene. The course of disease mimics the neoplastic progression of breast carcinoma from preinvasive ductal carcinoma in situ to invasive carcinoma with metastatic potential. Critical to these studies is precise assessment of neoplastic progression over time. The goal of the study was to evaluate in vivo microPET imaging as a surrogate for standard histological measures of disease proliferation and progression that require sacrificing the animal and therefore preclude the possibility of longitudinal assessment. A total of 10 mice were imaged weekly over a 13-week period using 18FDG as the imaging agent. Animals were scanned on a dedicated small-animal microPET system with a spatial resolution of approximately 2 mm. Starting at Week 4, individual animals were sacrificed immediately following imaging and their mammary fat pads were harvested for histological evaluation. The progression of disease is evident in the microPET images and FDG uptake correlates with histological evaluation. We can quantitatively track the proliferation of disease using the volume of functionally activated tissue across multiple time points. Furthermore, the normalized uptake of tracer appears to mark the transition from in situ to invasive carcinoma. We conclude that microPET imaging can be used to track disease progression in this mouse model, and these data are motivation for future work monitoring therapeutic interventions.
C.K. Abbey, None.
Abstract ID: 307
HGF/SF and its receptor Met play an important role in normal cellular processes as well as in tumorigenicity and metastasis. We constructed GFP-Met and DsRed-HGF/SF chimeric proteins and generated transient and stable transfected cells as well as transgenic mice. Both proteins retained their biological activity. LSM 510 with a Meta detector was used to characterize the spatial distribution and association of Met and HGF/SF. Intravital imaging of live transgenic mice showed expression in epithelial and endothelial tissues. Enhanced fluorescence was mainly observed in sebaceous glands. High-resolution intravital analysis revealed GFP-Met membrane sub-cellular localization. At the age of 4–6 months, male transgenic mice developed angiosarcomas and/or sebaceous gland adenocarcinomas in their lower abdominal area. GFP-Met is highly expressed in the tumors. Intravital FRAP analysis revealed attenuated GFP-Met trafficking in sebaceous glands. Moreover, degradation of the receptor was evident 1 min following intravenous injection of HGF/SF. Ultrasound functional molecular imaging of Met, recently developed in our lab, demonstrated increasing contrast media signal intensity in the tumor upon HGF/SF treatment. Here, we demonstrate that direct molecular imaging of a tyrosine kinase growth factor receptor in living cells and mice can provide new insights into the molecular mechanisms of their activity. These novel techniques could help identify alterations in protein expression, isolate cells that undergo the early events of transformation and metastasis in vivo, and help in understanding the mechanisms of Met-HGF/SF-induced transformation and metastasis. This work was supported in part by a research grant from the NIH (P50CA93990).
S. Moshitch-Moshkovitz, None.
Abstract ID: 308
We have established a dual-color fluorescence imaging model of tumor-host interaction based on a red fluorescent protein (RFP)-expressing tumor growing in green fluorescent protein (GFP) transgenic mice. This model enabled visualization of the tumor-stroma interaction including tumor angiogenesis and infiltration of lymphocytes in the tumor. Transgenic mice, expressing the GFP under the control of a chicken beta-actin promoter and cytomegalovirus enhancer, were used as the host. All of the tissues from this transgenic line, with the exception of erythrocytes and hair, fluoresce green under blue excitation light. B16F0 mouse melanoma cells were transduced with the pLNCX2-DsRed-2-RFP plasmid. The B16F0-RFP tumor and GFP-expressing stroma could be clearly imaged simultaneously in excised tissue. Dual-color imaging enabled resolution of the tumor cells and the host tissues down to the single cell level. Tumor stroma include fibroblast cells, tumor infiltrating lymphocytes, blood vessels, and capillaries, all expressing GFP. GFP stromal cells were readily distinguished from the RFP-expressing tumor cells. This dual-color fluorescence imaging system should facilitate studies for understanding tumor-host interaction during tumor growth and tumor angiogenesis. The dual-color system also provides a powerful tool to analyze and isolate tumor infiltrating lymphocytes and other host stromal cells interacting with the tumor for therapeutic and diagnostic/analytic purposes.
R.M. Hoffman, None.
Abstract ID: 309
We previously reported that the uptake of tracer neutral amino acids, 1-aminocyclopentane carboxylic acid (ACPC) and a-amino-isobutyric acid (AIB), is increased in brain tumors and that “facilitated” ACPC transport is upregulated across C6 glioma capillaries. However, the precise mechanism of this phenomenon remains unclear. In order to clarify this issue and understand the biodistribution of 2 neutral amino acid transporters (A and L) in the vasculature, we performed an imaging and biochemical assessment of these transporters. Double label ([14 C]-AIB or [14C]-ACPC and [67Ga]-DTPA) quantitative autoradiographic studies were performed on 2 functionally different intracerebral RG2 gliomas (RG2E and RG2D). We compared the facilitated transport (deltaK1) of ACPC and AIB in tumor tissue and in normal brain, where deltaK1AA = K1AA – (“pore diffusion” + “membrane diffusion”). Immunochemical and Western blot analysis for both transporters were performed on the cryosections as well as on an endothelial cell line (bEND.3) and tumor cell lines. Passive vascular permeability in both RG2 gliomas was low and contributed minimally to the flux of AIB and ACPC across tumor capillaries. deltaK1ACPC and deltaK1AIB in RG2E were 29.3±7.8 and 23.2±6.0 ml/min/g, respectively, and these values were significantly higher than contralateral RG2D gliomas and corresponding normal cortex. Western blot and immunochemistry confirmed that the “A” and “L” transporters were expressed in endothelial cells in tissues as well as bEND.3 and RG2 cell lines. These results suggest that: (1) the “A” transporter is expressed on the luminal surface of RG2 gliomas; (2) the relatively small increase in ACPC “facilitated” transport in RG2 gliomas compared to cortex can be explained by “A” transporter expression in tumor vessels and the affinity of ACPC to both the “L” and “A” transport systems (“L” > “A”). These data support the use of nonmetabolized amino acids that are predominantly transported by the “A” transporter for imaging brain tumors.
T. Miyagawa, None.
Abstract ID: 310
The significance of lymphatics in the metastatic cascade has largely been eclipsed by the focus on angiogenesis and vascular dissemination. Recent research, however, is gradually defining mechanisms underlying lymphangiogenesis and lymphatic metastases. Among the unresolved questions is whether cancer cells in a primary tumor are transported to remote sites via patent, intratumoral lymphatic channels. Recently, we presented a noninvasive technique for identifying potential lymphatic channels in a human breast cancer model in vivo, using functional MRI. In this study, we present an approach for identifying “functional” lymphatic channels in a metastatic human breast cancer model (MDA-MB-231), using a macromolecular probe (biotinylated albumin-GdDTPA) in conjunction with fluorescence microscopy. Four tumor-bearing (>200 mm3) mice were anesthetized and 60 mg/ml of biotinylated albumin-GdDTPA was administered intravenously. Since lymphatic drainage of macromolecules in tumors is a slow event («hours), 3 hr was allowed to elapse, after which animals were sacrificed and were tumors excised and fixed in Carnoy's solution. Paraffin-embedded tumor sections were double-stained for lymphatic vessels (LYVE-1) and biotinylated albumin-GdDTPA (streptavidin-TRITC) and counterstained with hematoxylin. Lymphatic vessels stained with LYVE-1, with or without albumin-GdDTPA present, were detected in tumors (Fig. 1). The presence of albumin-GdDTPA demonstrates functional vessels, whereas its absence is indicative of nonfunctional lymphatic vessels. These data demonstrate, for the first time, the presence of functional lymphatics at distances greater than 100 μm from the tumor periphery. The functional assays described here take us a step closer to understanding the role of lymphangiogenesis and lymphatic in metastasis. Acknowledgements: This work was supported by NIH R01 CA90471.
(a) Fluorescence image showing LYVE-1 (green) and Albumin-GdDTPA (red), and (b) Hematoxylin counterstained image [40 ×, LV=lymphatic vessel, BV=blood vessel].
A.P. Pathak, None.
Poster Session 11: Cell Migration, Stem Cells, and Immunology
Abstract ID: 311
Migration and infiltration of inflammatory cells is an underlying mechanism in all inflammatory diseases. We aimed to image inflammation by magnetic resonance (MR) using superparamagnetic contrast agents. Ultrasmall iron oxide nanoparticles (USPIO) are iron-based contrast agents that accumulate in phagocytic cells and cause considerable T2 shortening. We addressed 2 questions: (1) Is the time course of T2 shortening in an inflammatory lesion after USPIO injection consistent with contrast uptake by trafficking phagocytes? (2) Where does the contrast-labeling of phagocytes occur? Oxazolone-induced contact hypersensitivity was induced in the ears of 4–6-week-old BALB/c mice. At 18–24 hr after the challenge, ear thickness was measured, and baseline MR imaging of the ear was performed on a 9.4-T horizontal-bore MR scanner. T2 measurements were performed. Intravenous USPIO were administered. Imaging data were fitted to a T2 decay model and regions of interest analyzed and T2 values were plotted over time. Evans blue was used to assess endothelial integrity. Plasma and purified leucocytes were analyzed to determine the process of labeling by imaging, histology, and iron measurements. Maximal swelling was observed 18–24 hr post challenge. In mice given USPIO, there was significant T2 shortening, maximal around Day 4 post challenge. Control T2 values were not significantly altered (Figure 1). Histology confirmed labeling of infiltrating cells, and evidence that cellular labeling occurred in the extravascular compartment. Change in T2 measurement provides a robust method for serial studies of USPIO uptake by phagocytes in vivo, as it is independent of coil/animal registration and has the potential to provide concentration information. Further studies are needed to define the precise mechanism of uptake. Cell-specific imaging of inflammatory lesions has the potential to be an important investigative tool for inflammatory diseases.
P. Reynolds, None.
Abstract ID: 312
Stem cell transplantation is a promising approach for the therapy of Parkinson's disease (PD). However, the mechanisms of differentiation, migration and long term survival of the transplanted stem cells are still not clear. In this study we labeled murine embryonic stem cells with iron-oxide particles (USPIO) for their detection by in vivo Molecular Magnetic Resonance Imaging (MRI). We further investigated whether the magnetic labelling of cells with iron-oxide-particles in vitro results in an increase of oxidative stress. We showed that the incubation of cells with iron-oxide-particles results in a transient increase of oxidative stress which is reduced to control levels one day after particle incubation. The augmentation of oxidative stress is closely linked to the incubation of the cells with iron-oxide particles. The embryonic stem cells were labeled with USPIO-particles in vitro and transplanted in the striatum of rats. In vivo MRI studies were conducted using clinically applied field strengths (1,5 T). The labeled stem cells (1×105) led to a significant contrast in T2* weighted images. Six weeks after transplantation the labeled stem cells are still detectable. The detection at 1,5 T requires around 2000 cells. This first results show that a non-invasive tracking of transplanted embryonic stem cells by molecular MRI is a promising approach in the visualisation of distinct cell populations.
A. Stroh, None.
Abstract ID: 313
The purpose of this study was to determine whether intravenously administered ferumoxides (FE) labeled human mesenchymal stem cells (hMSC) can be delivered to a target and held in place within the rat liver by using an external magnetic field placed on the abdomen. FE complexed with poly-l-lysine (PLL) was incubated with hMSC overnight. After collecting FE-PLL labeled hMSC, about 1 million cells were administered IV into tail veins of 6–8 weeks old nude rats. 50% of the rats wore a custom designed jacket with an EM (0.24 Tesla, over liver) during IV injection of labeled cells and EM remained in place until MRI exams or until euthanasia. MRI was performed at 1.5 Tesla using a phase array wrist coil in all rats using PD/T2 FSE and GRE sequences weekly from day 1 to day 29, post infusion of labeled cells. Pairs of rats with and without EM were euthanized weekly for determining iron concentration and number of accumulated hMSC in the liver. Rats injected with unlabeled hMSCs served as control. ROI analysis of MRI of liver and external standard was performed. There was abrupt change of MRI signal intensity (SI) of the liver after injecting labeled hMSCs. The decrease in signal intensity change was more in rats with placement of EM. Prussian blue staining showed more iron labeled cells in the liver of rats with jacket, and IHC with anti HLA-1 positively identified the cells as human cells. Determination of iron in the liver tissues showed more iron in rats with jacket at all time point. Targeted delivery and retention of FE-PLL labeled hMSCs in liver was facilitated by using an EM. Magnetically labeled cells can be delivered and retained in a specific location by using EM, and this approach may be used for targeted cell or gene therapy.
A.S. Arbab, None.
Abstract ID: 314
Endothelial precursor cells (EPCs) contribute significantly to tumor neovasculature (1), and are being investigated as angiogenesis-selective gene-targeting vectors. It is important to determine the distribution of EPCs containing a therapeutic gene in order to measure potential effects and to optimize delivery. For detection by MRI, cells were labeled with a SPIO/transfection agent complex (2,3), which localizes in endosomes in the cytoplasm. Our purpose is to determine if MRI can be used to detect incorporation into vasculature and distribution in tumor of systemically administered EPCs labeled with SPIO/TA complex. RT2 rat glioma cells (5×105) were transplanted by intracranial injection into 5 week old SCID mice. Sca1+ cells isolated from the marrow of Swiss mice were labeled by incubating with SPIO (Ferumoxides, Berlex)-Poly-L-lysine complex. Labeled (L-) or unlabeled (U-) EPCs were injected into tumor-bearing mice i.v. Mice were imaged at 7T from 4–12 days following infusion. In vivo 2D gradient echo and RARE images were acquired, and T2*W 3D gradient echoes for fixed brains. In mice at day 5 post tumor/EPC transplants, a hypointense area in the tumor was visible on MRI of the L-EPC mice and not in unlabeled mice. In mice imaged serially, a hypointense region within the tumor at 6 days progressed to a hypointense ring surrounding the tumor at 10 days and was seen only in L-EPC animals. All mice developed evidence of hemorrhage by 12–14 days. Histology demonstrated iron-labeled cells in the tumor. MRI 5–12 days after infusion showed hypointense regions that changed over time, within and around the brain tumors of mice given L-EPCs. Histology identified SPIO-labeled cells in the tumor and showed that EPCs incorporate in the neo-vasculature (1). These results indicate we can directly image EPCs in angiogenic vessels as well as follow a gene-delivery vector by MRI.
S.A. Anderson, None.
Abstract ID: 315
Tracking the fate of cells in vivo is crucial to elucidate cellular mechanisms and functions. Our goal was to develop a biocompatible superparamagnetic nanoparticle for highly efficient in vivo detection of systemically administered cells by MRI and to apply MR cell tracking to a relevant adoptive transfer mouse model. Based on principles of multivalency we developed a novel Tat peptide derivatized superparamagnetic nanoparticle, CLIO-HD, and optimized its cellular uptake-efficiency in CD8+ T cells. Labeling with CLIO-HD did not affect T cell physiology as assessed via in vitro cytotoxicity and cell killing assays, and a physiological flow model of leukocyte-endothelial interactions. CLIO-HD labeled ovalbumin-specific (OT-I) CD8+ T cell recruitment in an antigen-specific murine tumor model (B16OVA vs. B16F0) was analyzed by MRI (8.5T) and the results confirmed by immunohistochemistry, autoradiography and biodistribution studies. CLIO-HD labeling efficiency was optimized 200-fold, and labeling was stable in cells for at least 5 days. This allowed the 3D visualization of antigen-specific T cell recruitment to, and departure from target tumors dynamically and quantitatively in live intact animals with both very high sensitivity (<3 cells/voxel) and high spatial resolution. Using this approach, we demonstrated that the homing pattern of OT-I CD8+ effector T cells (CTLs) to tumors is highly heterogeneous and that sequential adoptive transfers result in infiltration of CTLs into different tumoral compartments, potentially resulting in a more efficient anti-tumor attack The current study is the first to demonstrate that the recruitment of systemically injected cells can be imaged by MRI in live mice, and holds promise for the further investigation of cellular behavior in vivo and in the evaluation of novel cell based therapies.
M.F. Kircher, None.
Abstract ID: 316
Current methods for molecular and cellular imaging involve introducing superparamagnetic species into cells and tracking them using MRI. Micron size, fluorescent, superparamagnetic magnetite particles are efficiently endocytosed by many cell types and their large size makes them effective T2* agents (Hinds KA, et al, Blood, 2003, online and in press). Here we demonstrate, for the first time, that single particles in single cells can be detected using MRI and that the enhancement effect of single particles is as high as several hundred microns. Several cell types (hepatocytes, fibroblasts, MSC's and others) were lightly labeled with 0.96 micron diameter particles, co-labeled with a FITC analog, by simple incubation. Fig. 1 shows a high resolution (50 microns) MRI of live hepatocytes on a culture dish harboring single particles as corroborated by confocal microscopy. The dipolar field of the single particles is clearly visible. Mouse embryos were injected at the single cell stage with hundreds of similar 1.63 micron particles and allowed to develop to E11.5, where they were then removed and fixed. Fig. 2 shows an MRI and fluorescent stereo-microscope image of a paraffin-embedded slice, with expansions shown below, where only a single magnetite particle is detected by fluorescence and registered to the MRI. Advantages of using these particles are conditions where it may be difficult to get large labeling efficiency, for instance, in-vivo labeling of rare cell types. Additionally, cell division can not dilute the contrast below a minimum observable threshold because only single particles are required to produce robust artifacts.
E.M. Shapiro, None.
Abstract ID: 317
Cell tracking techniques are essential for studying physiological and pathological cell-host interactions in-vivo. Previous efforts have been made to tag cells with MR-visible iron-oxide particles. The aim of this study was to analyze the impact of hydrodynamic diameter and surface-coating of clinically available superparamagnetic iron-oxide particles (SPIOs) on cell labeling efficiency in the presence or absence of a polycationictransfection medium (TM). Different cancer cell lines (CLL185, HTB56, DU4475, HT1080) were tagged using carboxydextran-coated SPIOs of different diameters (65, 46, 21, 17 nm) and a dextrancoated SPIO (diameter: 150 nm). Cells were incubated in iron containing medium (0.01–1.00 mg iron/ml), with or without TM. Iron-uptake was visualized by light-microscopy after Prussian-Blue-staining and quantified by Atomic Emission Spectroscopy (AES). MR-signal characteristics and the number of detectable cells were analyzed at 1.5–and 3.0–Tesla using T2-weighted Gradient-Echo- and Turbo-Spin-Echo- sequences. For all experiments intracellular iron-uptake could be demonstrated by light-microscopy and MRI. Larger particle size resulted in an improved cellular iron uptake (e.g.: [1.00 mg Fe/ml]: 65 nm: 2.23 μg/100,000 cells; 17 nm: 1.46 μg/100,000 cells; p<0.05). Cellular uptake of dextran-coated particles was virtually identical to large carboxydextran-coated particles (e.g.: [1.00 mg Fe/ml]: 65 nm: 2.23 μg/100,000; 150 nm: 2.02 μg/100,000 cells; p > 0.05). The presence of the TM significantly increased the iron load of cells up to 2.7-fold (p<0.05). As little as 10,000 cells were readily detectable with clinically available MR-techniques. Efficient tagging of mammalian cells can be performed with clinically available MR-agents. Large particle diameter and the use of a transfection medium significantly enhances cellular iron-oxide uptake.
L. Matuszewski, None.
Abstract ID: 318
Purpose of this study was to evaluate the ability of Gadophrin-2 to trace intravenously injected human hematopoietic progenitor cells in a murine xenotransplant model with optical imaging (OI) and MR imaging. Hematopoietic progenitor cells were labeled with Gadophrin-2 (Schering AG, Berlin, Germany), a metalloporphrin with a high r1 relaxivity (19.8±0.1 l*s*mmol-1*s-1; 20MHz, 37°) and fluorescence (excitation: 499 nm, maximal emission: 617 nm). The labeled cells were evaluated in vitro with electron microscopy, spectrometry (ICP-AES) and viability tests. Then, 1000000–300000000 labeled cells were injected into 14 athymic Balb/c mice and the in vivo cell distribution was evaluated with MR and OI before and 1, 4, 24 and 48 hours (h) after intravenous injection (p.i.). 3 additional non-treated mice served as controls. The contrast agent effect was quantified for MR by calculating DSI(%)-data. After completion of in vivo imaging studies, OI and fluorescence microscopy of excised organs was performed. Intracellular cytoplasmatic uptake of Gadophrin-2 was proven with electron microscopy. Cell viability of the labeled progenitor cells was not significantly impaired compared to non-labeled controls. Spectrometry determined an uptake of 31.555 nmol Gd per 1000000 cells. After intravenous injection into nude mice, the in vivo migration of Gadophrin-2-labeled cells could be depicted with MR and OI: At 1 and 4 h p.i., the transplanted cells mainly distributed to lung, liver and spleen and, 24 h p.i., also to the bone marrow. OI of postmortem excised organs showed a steadily declining fluorescence in lung, liver and spleen, but increasing fluorescence of the bone marrow with increasing time p.i.. Fluorescence microscopy confirmed the distribution of contrast agent containing transplanted human cells to the murine organs. Gadophrin-2 is a bifunctional contrast agent suited to label hematopietic progenitor cells and to trace the distribution of the labeled cells in a xenotransplant model with OI and MR imaging.
H.E. Daldrup-Link, None.
Abstract ID: 319
Recent developments in stem cell and gene therapy will require methods to monitor cell migration and integration repeatedly and non-invasively with a high temporal and spatial resolution in vivo. The purpose of this study was to visualize the different migration behaviour of phosphatase and tensin homolog-deficient (PTEN) neuronal stem cells in comparison with non-modified (PTEN+) neuronal stem cells, in vivo. In neuronal stem cells mRNA and protein expression of PTEN was downregulated by RNA-interference. Subsequently cells were labeled with SPIOs using cationic lipofection. Neuronal stem cells were implanted intracerebral: 10 mice (NMRI) received 1×104 PTEN− cells another 10 mice received 1×104 PTEN+ cells. Magnetic resonance imaging (MRI) (T2 weighted 2D spinecho sequences: TR 3704, TE 90, field of view 100, Matrix size 512, slice thickness 0.5 mm, 8 acquisitions) was performed after 1 h, 1 week and 2 weeks using a clinical 1.5 Tesla scanner. MRI results were confirmed by histopathologic analysis. In the PTEN- group MRI displayed a circumscribed, sharply, round shaped decrease in signal intensity (SI) after 1 hour. After 1 to 2 weeks the sharp form became more and more blurred with spread decrease in signal intensity. In comparison MRI of PTEN+ cells displayed a sharp, well circumscribed round shape of decrease in SI up to 2 weeks. Histopathologic analysis showed a more diffuse distribution of the PTEN− cells. Trypan blue exclusion tests and differentiation assays of labeled neuronal stem cells revealed that cell labeling did not alter the cell viability or biology. Genetically modified neuronal stem cells transplanted intracerebral in mice could be detected by MRI. Moreover increased migration behaviour of PTEN− neuronal stem cells could be visualized with a 1.5 Tesla MR-scanner confirmed by histopathologic analysis. Therefore this approach seems to be applicable to monitor stem cell and gene therapy in a clinical setting.
M. Rudelius, None.
Abstract ID: 320
To determine effects of interferons (IFNs) on replication and tropism of HSV-1, we infected wild-type and congenic mice lacking IFNαβ, IFNγ, or IFNαβγ receptors with a virus (KOSdlux) that expresses luciferase from the ICP8 promoter. We monitored localization and relative amounts of virus by bioluminescence imaging (BLI). Following foot pad or corneal infection, relative levels of transmitted light quantified by BLI at infection sites were in rank order of IFNαβγ R−/− > IFNαβ R−/− > IFNγ R−/− ≈ wild-type mice. In IFNγ R−/− and wild-type mice, transient dissemination of virus beyond the site of infection occasionally occurred only after footpad injections. Spread of virus to regional lymph nodes, spleen, liver, and spinal cord occurred in IFNαβ R−/− mice, but these animals survived. HSV-1 infection in IFNαβγ R−/− animals produced quantitatively greatest and most widespread dissemination of virus to visceral organs and the nervous system, and these mice invariably died after ocular or footpad infection. In the absence of IFNαβ receptors, deletion of IFNγ receptors has a profound effect on susceptibility to HSV-1, although loss of IFNγ receptors alone minimally affects pathogenesis. These data demonstrate different effects of type I and II interferons in limiting systemic dissemination of HSV-1 and show that BLI is an important tool for studying viral pathogenesis in vivo.
G.D. Luker, None.
Abstract ID: 321
The ability to non-invasively track the migration and possibly proliferation of neural stem cells would have significant clinical and research implications. We produced stable F3 human neural progenitor cells (F3-hNIS) transfected with human sodium/iodide symporter (hNIS) for non-invasive in vivo tracking. The expression patterns of hNIS gene in F3-NIS according to the culture passage were examined. F3 human neural stem cells isolated from human telencephalon (14–16 weeks) and immortalized with v-myc were obtained from Dr. SU Kim (Ajou University, Suwon, Korea). hNIS and hygromycin resistance genes were linked with IRES (Internal Ribosome Entry Site) under control of CMV promoter (pIRES, Clontech Co.). This construct was transfected to F3 with Lipofectamine (Invitorgen Co.) to make stable F3-NIS by selection with hygromycin for 2 weeks. hNIS expression was confirmed by I-125 uptake assay and RT-PCR analysis with specific oligonucleotide primer of hNIS. Iodide uptake analysis showed that F3-NIS accumulated up to 12.3, 7.7, 13.2, and 1.1 times higher iodide than non-transfected F3 cells in passage-1, 2, 3, and 4 cell lines, respectively. Potassium perchlorate, the inhibitor of NIS, inhibited completely this uptake of iodide. hNIS expression in F3-NIS diminished gradually according to cell passage and was almost silenced after 4 passages. RT-PCR analysis showed that hNIS expression was also decreased with increasing cellular passage. Good correlation was found between iodide uptake and hNIS expression. These results suggest that hNIS-transfected F3 neural stem cells might undergo a change in its biological characteristics with increasing cellular passage. Gene expression of exogenously transferred reporter gene in human stem cell might have been affected by the epigenetic promoter modulation as cells passed by their generation.
Y. Kim, None.
Abstract ID: 322
There is an increasing interest in the use of MRI to track cells in vivo to monitor inflammatory processes and stem cell therapies. Mono-crystalline-iron-oxide nanoparticles (MIONs), linked to Tat peptide, liposomes and antibodies, have been utilised to enhance “tagging” of cells in order to make them “MRI visible”. However, although successful, it is not possible to determine the labelling uniformity of the cells. This is important as in vivo cell tracking of labelled cells, especially in cell-based therapy, assumes that all injected cells are equally tagged and therefore equally visible to MRI. Here we propose the use of automated-magnetic-cell-sorting (auto-MACS) to separate cells according to their relative levels of tagged MION to allow more precise quantitative MRI measurements to be made with respect to cell numbers recruited to an anatomical region. Rat granulocytes were isolated using a modified bone-marrow extraction and Percoll gradient procedure. Cells were then incubated with either MION-linked-antibodies or MION-Tat peptide. Cells were then passed through an auto-MACS system. Several discrete pools of cells were titrated, suggesting significant differences in the relative levels of paramagnetic labelling in both conjugation methods. Furthermore, by varying the relative separation procedure of the auto-MACS (ie slow vs fast), further sub-populations of labelled cells could be achieved. MRI contrast measurements of the cells also suggested good conjugation of MION using both techniques. In conclusion, cell tagging with MION-Tat peptide and MION-linked-antibodies under normal circumstances is non-uniform which may lead to misinterpretation of in vivo cell tracking measurements made by MRI. The use of auto-MACS can therefore be used to ensure that uniformly labelled populations of cells with maximum MRI visibility are employed in any in vivo MRI contrast measurement. This will allow for accurate quantification of cell migration and recruitment following their administration.
S. Hotee, GlaxoSmithKline 1.
Abstract ID: 323
At least for some human cancers, the process of metastasis appears to be mediated by the invasive character of cancer cells. The effects of therapeutic treatments on tumor cell motility should reflect the treatment's effectiveness in the inhibition of cancer metastasis. We have used the Large Scale Digital Cell Analysis System (LSDCAS) to quantify cell motility in the human glioblastoma cell line U87-MG following ionizing radiation exposures to test the hypothesis that ionizing radiation exposure leads to a marked increase in cell motility in this cell line. Digital sequences obtained from cells after varying doses of radiation were analyzed for motility using LSDCAS software. These experiments used standard T25 cell culture flasks for the growth of large numbers of cells which were sampled at random using 50 microscope fields per experimental group. Images were acquired at 5 min. intervals, and microscope fields contained, on average, 5 cells each. Image segmentation was performed automatically through the application of pixel connectivity techniques. Statistical analysis of the time-dependence of the cumulative distance traveled by the centroid of the borders of individual cells over a three-day interval following treatment was used to determine the distribution of speeds in each cell population. Our results showed that there was no significant effect of radiation on cell motility, based on the analysis of 50–100 cells from each treatment group (i.e., after 0, 5, 10, 20 Gy Γ-ray exposures). Other experiments, using attachment of cells to fibronectin- and collagen-coated tissue culture flasks, demonstrated the sensitivity of our technique to detect changes in cell motility. Thus, at least for the time interval of observation and for the dose levels studied, no significant effect of radiation on cell motility was found. (Support: 1R33 CA94801 [NIH])
M.A. Mackey, None.
Abstract ID: 324
Cellular imaging is a newly emerging field with the potential to allow visualization of early events in inflammation, at the cellular level. To accomplish this in our lab, an unconventional approach was taken, where new tools and concepts that enable MRM at 1.5T were developed and tested including custom-built hardware and optimized pulse sequences (3DFIESTA). We have investigated two interesting rodent models where inappropriate macrophage activity results in lesion progression, a clip compression model of spinal cord injury (SCI) and experimental autoimmune encephalomyelitis (EAE). Rats were administered an iron oxide-based contrast agent (Feridex), which is taken up by these phagocytic cells, at different stages of the disease process to monitor inflammatory events. Imaging was performed at 1.5T with a custom-built gradient insert and solenoid RF coils. A 3DFIESTA (SSFP) imaging pulse sequence was optimized for single cell imaging with a THP-1 cell model system (Fig. 1). Under these conditions, this sequence produces contrast which is very sensitive to the superparamagnetic effects of iron oxide particles and also allows the differentiation between this iron and iron associated with hemorrhage in early SCI. Using this combination of novel strategies for cellular and microimaging we were able to detect subtle pathologic features, and the presence of iron labeled cells, over the time-course of acute SCI and EAE (Figs. 2 and 3).
Iron loaded cells in gel MR and Fluorescent overlay SCI after Feridex injection
P. Foster-Gareau, None.
Abstract ID: 325
Cell trafficking imaging studies in cancer need to evolve from simple trafficking studies to studies of immune cells in various activated states.methods are needed in which imaging signal allows one to determine if t cells are activated and actually destroying target tumor cells. granzyme b expression is an indication that the t cell has been activated and is ready to destroy its target. Therefore we used the murine granzyme b promoter (pgranb) to drive the expression of the firefly luciferase (fluc) reporter gene (pgranb-fluc vector) in primary splenocytes in cell culture. mouse splenocytes (2×107) were transfected with 15 μg pgranb-fluc vector using deae-dextran and then activated with 1 μg/ml of plate bound anti-cd3 and anti-cd28 antibodies or with plate bound anti-cd3 antibody and 5μg/ml concavalin a (cona). cd3/cd28 induced a 20–30 fold activation over un-stimulated control and cd3/cona induced a 2–3 fold activation over un-stimulated control. therefore, pgranb can be used as a marker for t cell activation. additional studies are under way to look at activation in vivo and reporter gene amplification including the two-step transcriptional amplification (tsta) strategy. future work will include utilizing pgranb-fluc as well as other t cell activation markers (il-2 and il-2rα) in various cell trafficking models including transgenic models.
M. Patel, None.
Abstract ID: 326
Metabolic disorders and genetic diseases of the brain can potentially be cured by transplantation of neural stem cells. A major challenge towards development of stem cell therapy is to establish a methodology for non-invasive monitoring of the migration of donor cells in the brain. GFP and luciferase are suitable for optical imaging of superficial organs, and HSV-tk requires binding to PET detectable ligands that may not cross the blood brain barrier. Cells labeled with super paramagnetic iron oxide particles (SPIO) can be detected using magnetic resonance imaging with high spatial resolution. In this study, the utility of MRI for stem cell tracking in neonatal and adult brains was evaluated. SPIO labeled C17.2 mouse neural progenitor cells were implanted either in the ventricles of neonatal SCID mice or in the hippocampus and cortex of adult C3H mice. In vivo MR imaging was performed on a 4.7 T horizontal bore magnet using a gradient echo imaging sequence. Neonatal mice were imaged seven weeks after stem cell implantation. Adult mice were serially imaged at one day, and one, two, three and four weeks after C17.2 cell injections. At the end of in vivo experiments, formalin fixed brains were imaged ex vivo using a 9.4T vertical bore magnet. Cells labeled with SPIO particles appeared as hypo-intense regions on MRI images. Neonatal mice showed a widespread migration of the stem cells extending from the olfactory bulbs to the cerebellum, while adult mice showed local migration of these cells in the cortex. Labeled cells were detected in tissue sections by Prussian blue staining for iron. There was a high degree of correlation between ex vivo MR images and Prussian blue staining. These studies point towards the potential of tracking stem cell migration in vivo which might assist in developing and monitoring stem cell based therapeutic approaches.
H. Poptani, None.
Abstract ID: 327
The ability to track and enumerate cell populations as they move is of interest in animal models of disease, and in human diagnostics. As optical methods are more sensitive than other in vivo methods for rare cells, and as injected contrast approaches are potentially more transferable to patients than genetically engineered approaches, we set out to develop tools for tracking cells in animals and human subjects. We developed two methods. First, injectable contrast with a tissue-penetrating fluorescent signal were targeted to lymphocytes (SPX-Ly-036). Second, injectable contrast with a tissue-opaque fluorescent signal were targeted to cells of prostate origin (SPX-Pr-045). The targeting molecules varied (e.g., MNM-591 for prostate). The fluorescent reporters were cyanine dyes (Cy Dyes 5 and 7). We found that in animal models, lymphocytes (L) could be seen in subcutaneous lymph nodes (Fig. A—Labeled WBC's in Lymph Nodes) and at sites of paw (P) inflammation (Fig B—Labeled WBC's at Site of Inflammation). In ex vivo models (not shown), circulating tumor cell load could be estimated. Both of these approaches are being evaluated for transfer to human subjects. Figure. Labeled Lymphocytes Tracked In Vivo. WBC's labeled via injected contrast are seen at (A) lymph nodes and (B) Sites of Inflammation in a Mouse.
Labeled Lymphocytes Tracked In Vivo. WBC's labeled via injected contrast are seen at (A) lymph nodes and (B) Sites of Inflammation in a Mouse
D.A. Benaron, Spectros Corporation 1, 4, 5.
Poster Session 12: Angiogenesis and Apoptosis
Abstract ID: 328
Plasminogen activator inhibitor type-1 (PAI-1), the major regulator of plasmin generation, is expressed by migrating keratinocytes during wound healing but not in the intact epidermis. Plasmin is utilized by invading cells to locomote through complex matrix/stromal tissue barriers and essential for epidermal wound repair. PAI-1 synthesis must be controlled in this process as deficient (as in the PAI1-/- mouse) or excessive PAI-1 production is incompatible with maintenance of a matrix “scaffold” structure permissive for epithelial migration or formation of functional neocapillaries. To view the time course of induced PAI-1 transcription during stimulated cell motility a molecular genetic approach was coupled with an in vitro model of epidermal injury repair. An injury-responsive reporter was constructed consisting of 800 bp of the human PAI-1 promoter cloned upstream of a GFP insert. PAI-1 transcription in keratinocyte transfectants was confirmed by Northern blotting and, at the single cell level, by GFP reporter activity after the switch from a sessile to motile phenotype. GFP expression was activated (as was the endogenous gene) within the real time of induced motility and only in the migrating cohort. To evaluate this relationship more directly, a second vector was developed in which 800 bp of the PAI-1 promoter was cloned upstream of an insert encoding a PAI-1-GFP chimeric protein. De novo-synthesized PAI-1-GFP was tracked in situ by fluorescence microscopy. Such PAI-1 expressing cells were highly motile in monolayer scrape-wound migration assays. The PAI-1-GFP protein was deposited almost exclusively into cellular migration “tracks”. Invasive traits could be “rescued” in PAI-1-/- keratinocytes, moreover, upon transfection with PAI-1/PAI-1-GFP expression vectors. The level of chimeric protein expressed correlated with cellular migratory activity. These findings illustrate the usefulness of this approach to image transcriptional and migratory events in single cells using promoters from genes implicated in cellular invasive traits. (Supported by NIH grant GM57242.)
P.J. Higgins, None.
Abstract ID: 329
Integrin αvβ3, which binds several ligands via an RGD sequence, is overexpressed in activated endothelial cells and various solid tumor cells. The involvement of this cell adhesion receptor in cell-cell and cell-matrix interactions during tumor angiogenesis represents an attractive target for imaging and treating tumor angiogenesis. The aim of this study is to label RGD peptide with 18F and 64Cu for PET imaging of breast cancer. Cyclic RGD peptide c(RGDyK) was labeled with 18F via a prosthetic 4-[18F]fluorobenzoyl moiety. c(RGDyK) was conjugated with DOTA and further radiolabeled with 64Cu. Human breast cancer MDA-MB-435 orthotopically implanted into female nude mice MFPs was studied with microPET, quantitative autoradiography (QAR) and direct tissue sampling to assess tumor targeting efficacy and in vivo kinetics of the radiotracers. 125I-RGD revealed good tumor accumulation and consistent retention, with tumor uptake of 3.1±0.2%ID/g and tumor-to-blood ratio of 9±1 at 1 hr post injection. [18F]FB-RGD showed rapid blood clearance, fast tumor accumulation and rapid washout rate, resulting in tumor uptake of 1.3±0.1%ID/g and tumor-to-blood ratio of 5.0±0.5 at 1 hr post injection. 64Cu-DOTA-RGD had the highest tumor uptake at 1 hr (3.3±0.2%ID/g) and tumor/muscle ratio of 9±1 as confirmed from direct tissue sampling, QAR and microPET. Specificity of the radiotracer was demonstrated via dose-dependent blockade of tracer uptake by the αv-integrin antagonist. In summary, it was shown that both 18F- and 64Cu-labeled RGD peptides may be useful for noninvasively imaging αvβ3 integrin expression. Potentially, the 64Cu-labeled RGD peptide can be used to determine dosimetry and tumor response to the same ligand labeled with therapeutic amounts of 67Cu for αvβ3 integrin-mediated radiotherapy. (Support in part from 5P20 CA86532, ACS-IRG-58007-42, and R24 CA86307).
X. Chen, None.
Abstract ID: 330
Oncology clinical trials frequently employ a wide range of imaging techniques (e.g., PET, MRI, and CT) in an attempt to evaluate changes in the levels of angiogenesis. However, as effective antiangiogenic therapy often does not lead to tumor mass/volume reduction, conventional measurements of response may be insensitive. Therefore, direct clinical imaging of angiogenesis via the targeting of receptors expressed selectively on angiogenic endothelial cells is a more specific approach for cancer diagnosis, and could be used for therapy monitoring. The receptor studied in this work has limited tissue distribution, with highest levels of expression achieved in growth factor activated endothelium during angiogenesis. The chosen vectors are peptides containing an RGD (Arg-Gly–Asp) motif in a configuration that allows high affinity and specific binding to the receptor. Based on these peptides, several chelate-peptide conjugates have been synthesized and screened, with 99 mTc-NC100692 demonstrating good synthesis, acceptable radiolabelling, pharmacokinetics, and in vivo efficacy. The objective of this work was to evaluate the ability of 99m Tc-NC100692 to monitor changes in tumor growth as a result of treatment using the therapeutic agent paclitaxel. Therefore, it was necessary to determine the amount of angiogenesis in experimental tumors (MVD/mm2) and correlate this to biodistribution data (%ID). Initial experiments were carried out using the Lewis lung subcutaneous model and low doses of paclitaxel (5 and 10 mg/kg). At these dose levels, tumor weight was unaffected, while the level of angiogenesis decreased. %ID 99mTc-NC100692 decreased by up to 51% (compared to the control group). Comparison data with 14 C-FDG demonstrated that FDG uptake was maintained in all groups (14 C-FDG decrease ˜5%). In conclusion, the results demonstrate that changes in 99m Tc-NC100692 uptake reflect a therapeutic effect before standard techniques that rely on changes in size (e.g., CT) and also suggest that 99m Tc-NC100692 will be more specific in antiangiogenic therapy monitoring than FDG.
M.S. Morrison, Amersham Health 5.
Abstract ID: 331
The VEGFR2 gene is transcriptionally regulated during angiogenesis. The ability to monitor and quantify VEGFR2 expression in vivo may facilitate a better understanding of the role of VEGFR2 in different states. Here, we describe a transgenic mouse, Vegfr2-luc, in which a luciferase reporter is under control of the murine VEGFR2 promoter. In adult mice, luciferase activity was highest in lung and uterus, intermediate in heart, skin, and kidney, and lower in other tissues. Luciferase expression in these tissues correlated with endogenous VEGFR2 mRNA expression. In a cutaneous wound-healing model, Vegfr2-luc expression was induced in the wound tissue. Histological and immunohistochemical studies showed significant macrophage infiltration into the wound and induction of Vegfr2-luc expression in both endothelial and stromal cells. Dexamethasone significantly suppressed both Vegfr2-luc expression and macrophage infiltration into the wound, resulting in delayed healing and impaired angiogenesis. In a skin hypersensitivity reaction produced by treatment with oxazolone, Vegfr2-luc expression was induced in the ear. Treatment by dexamethasone markedly suppressed Vegfr2-luc expression and leukocyte infiltration in the ear and was correlated with reduced dermal edema and epidermal hyperplasia. The Vegfr2-luc model will be valuable in monitoring the ability of drugs to affect angiogenesis in vivo.
N. Zhang, None.
Abstract ID: 332
Transglutaminases form a family of enzymes that have evolved for covalent crosslinking of proteins. The crosslinking activity can serve disparate biological processes depending on the location of the target protein. Extracellular activation of tissue transglutaminase (tTG) contributes to stabilization of the extracellular matrix (ECM) and promotes cell-substrate interaction. To generate new matrix, tumors as well as newly formed angiogenic blood vessels, are known to elicit wound-healing responses from the host tissues resulting in formation of granulation tissue at the advancing margins of the tumors. In addition, the enzymatic activity of tTG is a critical component of clotting. The goal of this study is to develop contrast material for specific imaging of tTG activity in vitro and in vivo, and to investigate whether this material can be utilized to outline the boundaries of malignant tissue and blood clots by MRI and optical imaging. MCF7 human breast carcinoma cells were cultured as multicellular tumor spheroid. Spheroids were incubated with biotin pentylamine or a biotinylated tTG peptide-substrate, and then were incubated with avidin-FITC for detection. Fixed spheroids were embedded in paraffin blocks and sectioned serially at 4 ìm thickness. Fluorescence microscopy of spheroid histological sections showed that the biotinylated pentylamine was crosslinked by tTG to the viable rim of the spheroid. In addition, the slides were stained for tTG by anti-tTG monoclonal antibody. The slides were examined by Optiphot2 microscope (Nikon, Japan); tTG was detected in the cells and in the ECM in the viable rim in correlation with the sites of crosslinking activity. Thus, optical imaging can be used for monitoring tTG activity in spheroids. Experiments are underway to generate MRI/optical multimodality compatible probes.
G. Mazooz, Varian Biosynergy 1.
Abstract ID: 333
Apoptosis plays an important role in many diseases and is a critical component in assessing efficacy of many antiproliferative treatments. Methods to noninvasively image and quantitate apoptosis in vivo would thus be an important adjunct to modern medial treatment. One approach to image apoptotic cells is by labeling annexin V, a high-affinity phosphatidylserine binder. We have recently validated a near-infrared fluorescent (NIRF)-labeled annexin V as a reporter for imaging drug-induced apoptosis in tumors. Here, we examine methotrexate (MTX)-induced apoptosis in a mouse model of rheumatoid arthritis. Our results demonstrate that 3 hr after intravenous injection of the NIRF-labeled annexin V, inflamed paws of MTX-treated arthritic mice (35 mg MTX/kg 48 hr prior injection of the probe, n = 3 mice) showed a sevenfold higher fluorescence intensity than arthritic paws of untreated mice (n = 3 mice), and a fourfold higher fluorescence intensity than nonarthritic paws of MTX-treated mice (p < .001 each). Using histology, we observed significant NIRF signal in the hyperplastic synovia of MTX-treated mice. These results were confirmed by conventional TUNEL staining for apoptosis in histology sections. In conclusion, NIRF-labeled annexin V is a suitable reporter for noninvasive, in vivo imaging of drug-induced apoptosis in arthritis. Using this reporter, we have been able to demonstrate for the first time in vivo that MTX induces apoptosis in the inflamed synovial tissue.
A. Wunder, None.
Abstract ID: 334
Nitric oxide (NO) is an cellular messenger involved in a variety of conditions and acts as a fundamental determinant of vascular homeostasis and angiogenesis. NO has also been shown to stimulate glucose metabolism in certain cells, but this has not been evaluated in endothelial cells. We thus investigated the relation between NO and deoxyglucose uptake in endothelial cells using sodium nitroprusside (SNP), a major NO donor. HUVEC cells were treated with various concentrations of SNP for 24 hr, or with 1 mM SNP for variable times. Some cells were treated with the NO synthase inhibitor, 
J. Paik, None.
Abstract ID: 335
Vascular endothelial growth factor (VEGF) is a potent permeability factor which is frequently upregulated in cancer cells. Increased expression of VEGF has been linked to increased metastases. VEGF production is also increased under hypoxic conditions, frequently observed in tumors. We determined the effects of VEGF overexpression on the invasion of PC-3 cells in contact with human umbilical vein endothelial cells under conditions of normoxia and hypoxia. Stably transfected PC-3 cells overexpressing full length VEGF were created using human pHu VEGF.21 (from Genentech) ligated into the expression vector pCR3.1. These studies were performed with our Metabolic Boyden Chamber Invasion Assay [1]. Cancer cell invasion was quantified from changes in profiles of intracellular water along the sample, obtained by diffusion-weighted 1H NMR [1]. The Invasion Index at Day 2 for PC-3 cells and PC-3VEGFc2 cells under well-oxygenated or hypoxic conditions is shown in Figure 1. PC-3 cells overexpressing VEGF exhibited increased invasion compared to wild-type PC-3 under oxygenated as well as hypoxic conditions. This increased invasion may be mediated in part through the HUVEC, since endothelial cells will degrade basement membrane in response to angiogenic stimuli. These results demonstrate that cancer cells overexpressing VEGF in the presence of endothelial cells may be capable of increased invasion which may explain increased rates of metastasis observed in patients with cancers secreting high levels of VEGF.
Figure 1. Comparison of the invasion indices obtained for PC-3 and PC-3VEGFc2 cells in the presence of HUVEC after ˜48 hr under well-oxygenated or hypoxic conditions. PC-3VEGFc2 exhibited increased invasion compared to wild-type PC-3 under both conditions.
E. Ackerstaff, None.
Abstract ID: 336
F. Frauscher, None.
Abstract ID: 337
Tissue hypoxia is one of the most important factors in tumor neo-angiogenesis. A transcriptional activator has been identified as the major mediator of cellular hypoxic responses, named hypoxia-inducible factor 1 (HIF-1). Under hypoxic conditions, HIF-1 binds to the hypoxia-responsive element (HRE) in the regulatory region of its target genes and initiates transcription. Recently, we proposed an approach for monitoring hypoxia-induced up-regulation of different endogenous genes in situ and in vivo by noninvasive imaging using a hypoxia-sensitive reporter system. Based on this system, we have developed a vector with a new reporter gene (TKnesGFPLuc) placed under control of hypoxia-responsive element, and we have shown that the vector can be effectively used in identification of the inhibitors for HIF-1α-induced transcriptional activity. To demonstrate the response of the new fusion reporter to hypoxic conditions, we transduced C6 and RG2 cell lines with the vector. Hypoxia was induced by exposing cells to various concentrations of CoCl2 (1–500 ìM) or 2% O2 during 24 hr. Expression of the TKnesGFPLuc reporter gene was equally activated either by oxygen deficiency or addition of CoCl2 (50–500 ìM), and activation of the reporter gene was easily assessed by fluorescence microscopy, FACS, and bioluminescent analysis. To demonstrate the efficacy of this reporter system for detection of HIF-1 signaling inhibitors, we used a recently developed compound 3-(5′-hydroxymethyl-2′-furyl)-1-benzylindazole (YC-1), which inhibits HIF-1 activity in vitro by blocking HIF-1α expression at the posttranscriptional level. When tumor cells carrying the HRE/TGL reporter system were treated with YC-1 during hypoxia the level of HIF-1-mediated transcriptional activation of the reporter system was significantly inhibited as compared to nontreated hypoxic cells. These results demonstrate that the current HRE/TGL reporter system can be used to study anti-HIF-1 drugs in vitro, and that PET and optical imaging could be used for in vivo assessment of novel drugs.
I.S. Serganova, None.
Abstract ID: 338
Tumor microenvironmental factors, including those arising from angiogenic processes, alter endothelial cell surface protein expression. Accessible to the circulation, the tumor vascular endothelium is a logical target for drug or gene delivery. Transcytotic vesicles (caveolae) on the surface of the endothelium provide a pathway for overcoming the endothelial cell barrier for delivery to underlying tumor cells. Luminal endothelial cell plasma membranes and their caveolae directly from normal and tumor-bearing tissue were isolated and resolved the membrane proteins by SDS-PAGE and 2-D gel electrophoresis to create vascular endothelial protein (VEP) maps. These VEP maps reveal extensive heterogeneity of cell surface protein expression between tissues as well as apparent tumor-induced markers. Mass spectrometry and database searching identified the tumor-associated targets. Using antibodies to the appropriate specific polypeptides, we have confirmed the tumor induction of the proteins. One of the tumor-induced proteins, designated TE3, is expressed in endothelial caveolae of solid tumors but not normal organs. Biodistribution analysis and whole-body imaging using iodinated TE3 antibodies injected intravenously shows the accessibility of TE3 with significant and selective tumor tissue accumulation at the site of the metastatic tumor lesions with little accumulation in the surrounding “normal” tissue. Targeting the endothelium and its caveolae may be worthy strategies for tumor targeting and imaging in vivo. (Supported by NIH Grants CA97528 and CA83989.)
P. Oh, None.
Abstract ID: 339
We have recently demonstrated that radioiodine angiostatin (AS) has potential for tumor imaging (EJNM and Mol Imaging, 2003), but also found that in vivo instability of the present form makes it difficult to perform delayed imaging when blood pool activity is reduced. In this study, we hypothesized that AS alternatively labeled to have greater in vivo stability would retain specific endothelial targeting, but would allow higher tumor contrast. Human angiostatin radioiodine labeled by the Bolton-Hunter method (I-BH-AS) was compared to that by the previous lactoperoxidase method (I-AS). Competitive binding studies performed on HUVEC cells confirmed specific endothelial targeting of both radiotracers with significant and dose-dependent inhibition of binding with excess cold AS. Polyacrylamide gel analysis showed both radiotracers to remain stable in serum in vitro for 24 hr. However, 24-hr biodistribution in rats suggested that in vivo stability was relatively poor for 125I-AS, but appeared to be better for 125I-BH-AS. Tumor uptake of radioiodine AS was evaluated in Balb/C nude mice, bearing human colon cancer subcutaneous xenografts on the flank. Biodistribution of 125I-BH-AS in these mice demonstrated significant improvements in tumor to nontumor uptake ratios at 24 hr. Tumor/lung ratios were increased from 0.86 ± 0.01 to 1.72 ± 0.26 (p <.05), tumor/liver ratios from 0.88 ± 0.10 to 1.67 ± 0.33 (p = 0.08), and tumor/muscle ratios from 3.56 ± 0.13 to 8.36 ± 1.05 (p <.05) at 8 and 24 hr post injection, respectively. While both 125I-AS and 125I-BH-AS had similar uptake ratios at 8 hr, 125I-BH-AS had significantly higher tumor uptake ratios at 24 hr. Thus, radioiodine AS labeled with the Bolton-Hunter method retains specific endothelial binding while accomplishing better in vivo stability and higher tumor contrast. This could allow improved utility of radioiodine labeled AS as a tumor imaging agent.
S. Song, None.
Abstract ID: 340
The met protooncogene product (Met tyrosine kinase growth factor receptor) and its ligand, hepatocyte growth factor scatter factor (HGF/SF), mediate mitogenicity, tumorigenicity, and angiogenesis. HGF/SF and Met play a significant role in the pathogenesis of a wide variety of cancers and may serve as potential targets for cancer prognosis and therapy. We have shown previously that in vitro activation of Met by HGF/SF increases oxygen consumption. We recently demonstrated functional molecular imaging of Met receptor activity in tumors expressing high levels of Met. Organs and tumors expressing Met have substantial alteration in blood oxygenation levels as measured by blood oxygenation level-dependent (BOLD) MRI. No significant alteration was observed in tumors or organs that do not express Met. The extent of the MRI signal alteration correlated with HGF/SF doses. To better understand the physiological mechanism of HGF/SF-induced alteration of hemodynamics, color Doppler, and real-time perfusion contrast media (CM) ultrasound were used. Mice bearing mouse and human breast tumors were injected intravenously with saline and CM as a baseline, followed by HGF/SF or saline intravenous injection, followed by injections of CM after 5, 20, and 35 min. The results demonstrate increasing CM signal intensity upon HGF/SF injection (peak at 20 min). HGF/SF increases the area having blood vessels with significant blood flow as well as increasing the blood volume in the tumor by opening new blood vessels with slower blood flow. The hemodynamic effects of Met were correlated with the angiogenic status of the blood vessels using immunofluorescent and confocal analysis. This novel functional molecular imaging is a powerful tool in the understanding the metabolic activity induced by Met and can serve as a tool for the detection and prognosis of a wide spectrum of tumors. This work was supported in part by a research grant from the NIH (P50CA93990).
G. Tsarfaty, None.
Poster Session 13: Imaging in the Neurosciences
Abstract ID: 341
EMD 121974, a cyclic RGD peptide inhibitor of αv-integrins, is currently in phase I/II clinical trails for patients with progressive or recurrent gliomas. Initial results have been favorable. However, current PET imaging of brain tumor based on metabolic radiotracers such as FDG has limited usefulness for assessing the efficacy of antiangiogenic treatment of brain tumors using αv-integrin antagonists. In this study, we evaluated a 18F-labeled RGD peptide [18F]FB-RGD for imaging αv-integrin expression in brain tumors. Cyclic RGD peptide was labeled with 18F via N-succinimidyl-4-[18F]fluorobenzoate. The radiotracer was evaluated in vivo for its tumor targeting efficacy in both subcutaneous and orthotopic U87MG brain tumor models. The [18F]FB-RGD was produced in 20–25% decay corrected yield with specific activity of 230 GBq/μmol. The tracer had very rapid blood and renal clearance. Tumor-to-muscle ratio of around 5 (30 min post injection) was observed from biodistribution, microPET, and quantitative autoradiography imaging of the subcutaneous model. Dynamic microPET imaging of the orthotopic model demonstrated high tumor-to-brain contrast. The tumor uptake reached early maximum and then slowly washed out. Receptor specific localization of the tracer was shown in a blocking experiment in the presence of excess unlabelled RGD peptide. Longitudinal microPET imaging study found disproportionate uptake in larger tumors. Blocking experiments utilizing excess of unlabeled RGD peptide demonstrated receptor specific localization of the tracer. In conclusion, [18F]FB-RGD, a novel radiotracer for αv-integrin expression, can be used to image brain tumor growth, spread, and angiogenesis. We are currently in the process of correlating tumor uptake with receptor expression level and evaluating the use of kinetic modeling with PET-derived data for [18F]FB-RGD in our brain tumor model. (Supported in part from 5P20 CA86532 and ACS-IRG-58007-42).
X. Chen, None.
Abstract ID: 342
Antagonists of the mGluR5 receptor may exhibit anxiolytic properties. To better understand the interactions of antagonists with mGluR5 receptors in vivo, we have developed radiotracers with high affinity for the receptor and methods for the synthesis of PET radiotracers incorporating either C-11 or F-18. For in vitro characterization, we synthesized [3H]-methoxyPyEP to characterize the binding properties to the rat and rhesus brain mGluR5 and to provide information on the characteristics of the receptor in brain. [3H]-MethoxyPyEP binds with high affinity (Kd = 3.4 nM: k+1 = 4 × 107 M−1 min−1, k–1 = 0.11 min−1) to receptor from rat brain regions of interest (Bmax in fmol/mg tissue): caudate/putamen, 83.4; frontal cortex, 64.7; hippocampus, 54.9; thalamus, 39.6; hypothalamus, 21.0. In the rat, binding to cerebellum, while apparent, did not provide significant values for Bmax. The Bmax/Kd ratio (an indicator of potential in vivo imaging capability) in the striatum, hippocampus, and cortex is sufficient for in vivo imaging. Similar results were obtained using rhesus monkey regions of interest. The mGluR5 antagonist MPEP blocked [3H]-methoxyPyEP binding with Ki = 0.3 nM, while mGluR1 and mGluR2 agonists/antagonists, and a variety of agonists and antagonists for other neuroreceptor systems (e.g., NMDA, mAChR, a 1-AdR, etc.) were ineffective. Autoradiographic images were obtained with [3H]-methoxyPyEP in rat brain sections and with [18F]-fluoromethoxyPyEP in rhesus monkey brain sections. Results are consistent with the tissue homogenate studies. The in vivo distribution of [18F]-fluoromethoxyPyEP in mice gave results consistent with the brain distribution of this receptor. Finally, in vivo imaging of mGluR5 receptors in rat using the MPEP analog [11C]methoxyMTEP provided specific binding to the corpus striatum and cortex, consistent with the in vitro result. Radiotracer time-activity curves were characterized by rapid wash-in and fast wash-out, and the receptor-specific binding (demonstrated by blockade with 1 mpk iv MPEP) gave a usable signal-to-noise ratio of 2.
S. Patel, None.
Abstract ID: 343
Myelinogenesis is a complex tightly regulated process between the myelinating cell and the axon. A novel transgenic mouse line was established to study myelination in the central nervous system (CNS). The expression of a conditionally toxic gene, the herpes simplex thymidine kinase (HSV1-TK), was controlled by myelin basic protein gene promoter (MBP). Myelin formation in our transgenic mice is completely controlled by daily subcutaneous administration of Ganciclovir (GCV). The present transgenic mouse represents a useful model to study white matter abnormalities in the CNS by MRI, to understand the basic mechanisms of myelin formation, and to perform cell transplantation for myelin reparation. The most severe phenotype, characterized by gradual severe tremors and ataxia, was obtained when GCV was given daily from Day 1 after birth to the end of the third week. Magnetic resonance imaging (MRI) in vivo was performed. T2-weighted images of developing mice at 2, 4, 6, and 8 weeks postnatal were recorded to evaluate the severity of the dysmyelination and to detect an eventual reparation of the myelin in the brain. Histological tissue slices at Week 4 showed a small population of remaining oligodendrocyte in GCV-treated brain. Cerebral white matter particularly in the corpus callosum and in cerebellum showed a severe hypomyelination. The present transgenic mouse represents a powerful model to explore the normal and abnormal myelination during development and in adult CNS using MRI and other noninvasive imaging techniques.
S. Ghandour, None.
Abstract ID: 344
A program is currently underway to enhance our ability to utilize endogenous iron-dependent contrast in MRI for diagnosis and monitoring important neurodegenerative diseases such as Alzheimer's disease. One goal of this program is to correlate histochemical and elemental iron detection to T2-weighted MR images of iron rich brain tissues. Perl's staining and inductively coupled plasma techniques are being used to establish iron concentrations in gross tissue specimens. The results are being correlated with T2-weighted MR images taken at 1.5, 3.0, and 7 T. Establishing the basic correlations between these unrelated technologies is required to ensure clear understanding of the resulting relationships. The magnetic properties of both tissue samples and defined pure compounds are being measured and correlated with the T2 relaxation times measured in the MR images. Therefore, correlations between the iron concentrations in tissue specimens and pure compounds are being established by colorimetric changes on histochemical staining and correlated with SQUID magnetometry and MR imaging of these materials. Iron-containing materials from various sources are being utilized. These sources include standard biological reference materials including human and bovine liver, iron oxide nanoparticles, ferritin, and ferric EDTA.
K.M. Fish, General Electric 5.
Abstract ID: 345
A series of F-18-labeled compounds targeting Aβ plaques are proposed as positron emission computed tomography (PET) imaging agents, which may be used for direct measurement of the Aβ burden relating to pathological states of Alzheimer's disease (AD). Previously, a series of novel Aβ aggregate-specific ligands IMPY and its related derivatives were prepared. In in vitro binding assay, IMPY competed with [125I]TZDM binding to Aβ40 aggregates showing an excellent binding affinity (Ki = 15.0 ± 5.0 nM). A comparable binding affinity value was obtained for the corresponding bromo derivative (Ki = 10.3 ± 1.2 nM). Based on the structure-activity data on IMPY, a fluoroethyl derivative was designed as potential PET tracer. The “cold” FE-BrIMPY showed an excellent binding affinity (Ki = 9.1 nM; Aβ40 aggregates in vitro binding assay using [125I]TZDM as the ligand). In addition, [18F]FE-BrIMPY labeled Aβ plaques of postmortem brain sections of AD patients. Same high specific binding of Aβ plaques in the brain sections of transgenic mouse was observed by in vitro labeling. More importantly, the compound penetrated the blood-brain barrier after an intravenous injection into normal mice, suggesting that the compound is suitably lipophilic. Taken together, preliminary data suggest that this is a potentially useful ligand for mapping Aβ plaques for AD.
S. Choi, None.
Abstract ID: 346
To develop potential single photon emission computerized tomography imaging agents for detecting β-amyloid (Aβ) plaques in the brain, small molecules were evaluated in order to meet the challenge of selecting agents with the desirable properties. A series of N-phenylpyrazoles containing a simple pyrazole group were synthesized. Bromo and iodo derivatives, 3 and 4 (MIPA), showed good in vitro binding affinities (Ki = 35 and 19 nM, respectively, using Aβ40 aggregates and [125I]TZDM). Radioiodinated MIPA was successfully prepared via the standard iododestannylation reaction, which gave a good labeling yield (50–60%) and a high radiochemical purity (95%). Using in vitro autoradiography, [125I]MIPA showed an excellent amyloid plaque labeling on sections from postmortem brain samples of Alzheimer's disease (AD) and those of transgenic mice specially engineered to overexpress Aβ plaques. A good correlation was observed between AD brain homogenate binding and autoradiography of brain labeled with [125I]MIPA. In vivo biodistribution study in normal mice showed that [125I]MIPA displayed a high initial brain uptake (2.2% ID at 2 min) and a fast washout (<0.2% ID at 1 hr post injection). Ex vivo autoradiographic study of [125I]MIPA exhibited a clear labeling for plaque-like structures in the same transgenic mice. The results clearly suggest that MIPA may be useful as a molecular imaging agent to study amyloidogenesis in the brain. (Supported by ISOA and NIH.)
M. Kung, None.
Abstract ID: 347
Postmortem brain samples of patients of Alzheimer's disease (AD) contain β-amyloid (Aβ) plaques, which are associated with pathogenesis of the disease. There is a strong impetus to develop specific Aβ-aggregate-binding ligands as in vivo imaging agents for diagnosis and monitoring progression of AD. We have prepared a thioflavin derivative, IMPY, which can be radiolabeled with I-125/I-123 for binding or SPECT imaging studies. IMPY displayed a good binding affinity for preformed synthetic Aβ40 aggregates (Ki = 15 nM vs. [125I]TZDM), and the radiolabeled form showed selective plaque labeling of postmortem AD brain sections. Further characterization of IMPY binding to plaque-like structures was evaluated in transgenic PS1-APP (TT) mice, an animal model for AD. Distinct in vitro plaque labeling of frozen TT mouse brain sections were observed with [125I]IMPY. A significant amount of plaques present in the cortical, hippocampal and entorhinal regions of the TT mouse brain can be detected with [125I]IMPY via ex vivo autoradiography at 30 min after an intravenous injection. The plaque labeling is consistent with Thioflavin-S staining. In contrast, the uptake of [125I]IMPY in the age-matched control mouse brain is minimal. Tissue homogenate binding further confirmed the plaque level in various brain regions of TT mouse correlating well with the density of the Aβ deposition. Using tissue dissection technique, [125I]IMPY showed a moderate increase in cortical region of TT mice as compared to the age-matched control. We are currently investigating metabolism of [125I]IMPY in the brain and in vivo blocking. (Supported by NIH, Institute for the Study of Aging and AstraZeneca Inc.)
M. Kung, None.
Abstract ID: 348
Exposure of young adult mice to a diet containing a copper chelator, cuprizone, induces a synchronous consistent demyelination in the CNS by 4 weeks. This demyelination can be reversible after removal of cuprizone or persistent when the treatment is maintained for longer period. This useful model is adapted to study demyelination in vivo. The aim of the present study is to characterize demyelination using MRI of at various stages of de/remyelination with correlation with histopathological analysis and to follow the myelin damage and repair in mice brain. Fourteen C57BL/6J 2-month-old mice were used, cuprizone was administrated at 0.2 wt.% mixed with diet. MRI was performed with a MR scanner operating at 4.7 T. We followed the evolution of lesions using T2- and T1-weighted sequences before and after injection of gadolinium from 4 to 10 weeks after starting treatment. Histopathological studies were undertaken in parallel on mice sacrificed at 6 weeks. Antibodies against myelin proteins (MBP, MOG) and an oligodendrocyte marker, carbonic anhydrase II (CA II), were used. In addition, a texture analysis study in several parts of the brain was applied. Within 4 weeks, we observed hyperintense regions in T2-weighted images principally in the dorsal hippocampus area and cerebellum reflecting a significant cytotoxic edema, which culminates at 7 weeks with only slight focal blood-brain barrier breakdown. This hypersignal declines partially and spontaneously over the following 3 weeks. Edema was completely disappeared in half of the animals 3 months after cuprizone removal. In agreement with MRI data, the histopathology of treated mice sacrificed at 6 weeks showed a massive demyelination in cerebellum, corpus callosum, and olfactory bulb associated with oligodendrocytes death. In summary, the demyelination model by cuprizone is a valid experimental model in which lesion formation can be investigated by in vivo MRI and texture analysis.
W. Jalabi, None.
Abstract ID: 350
Glutamate is an excitatory neurotransmitter derived from the Krebs cycle. Under pathological conditions, excess glutamate in the synaptic space can trigger a toxic cascade, via NMDA/AMPA kainate receptors leading to cell death. Histopathological studies have linked axonal injury to the presence of glutamate-producing immune cells in the proximity of multiple sclerosis (MS) lesions. Detection of elevated glutamate levels will provide a fundamental explanation of axonal injury and accumulation of clinical disability in MS. We demonstrate the feasibility of unobstructed single resonance detection of glutamate using proton magnetic resonance spectroscopy at 3 T and examine the relationship between glutamate levels in MS lesions and normal tissue. Data were acquired on a GE Signa 3.0T scanner (Menlo Park, CA). Figure 1 shows the detection of glutamate at 2.35 ppm using a TE-averaged PRESS acquisition. Spectra were derived from an 8-cc voxel in a parietal white matter region or surrounding an active/acute MS lesion. Given that glutamate is concentrated in neurons, the potential impact of glial/neuronal cell population differences was avoided by normalizing glutamate by NAA, a neuronal marker. Figure 2 highlights a 35.5% Glu2.35/NAA increase (0.145–0.107/0.107), between gadolinium-enhancing (active) lesions and normal white matter. T1 hypointense (chronic) lesions do not differ from normal white matter. These results support the link between active inflammatory lesions and altered glutamate metabolism. Excess glutamate present in active MS lesions may derive from immune cells such as macrophages, which could release glutamate in the intercellular space. Glutamate from the cerebro-spinal fluid warrants investigation.
R. Srinivasan, None.
Abstract ID: 351
Estrogens or antiestrogens are currently used by millions of women, but the interaction of these hormonal agents with brain estrogen receptors (ER) in vivo has not been fully characterized to date. The goal of the present study was to assess the level and regional distribution of brain ER occupancy in rats chronically exposed to 17b-estradiol (E2) or Tamoxifen (TAM). For that purpose, female ovariectomized Sprague-Dawley rats were implanted with pellets containing either placebo (OVX), E2, or TAM (n = 4/group). After 21 days, rats were injected with 11b-Methoxy-16a-[125I]-iodo-estradiol ([125I]MIE2), killed 2 hr later, and brains processed for quantitative film autoradiography. ER occupancy was measured in the prefrontal cortex (FCX), bed nucleus of the stria terminalis (BSTMA, BSTMP), preoptic area (MPA), hypothalamus (ARC, VMHVL, PVA), amygdala (LaDL, MeA, CoA), hippocampus (HIPPOd, HIPPOv), and central grey (CG). The striatum (ST) was used to quantify the nonspecific binding. For each region, mean radioactive density was computed and the specific binding (region ST) calculated. The OVX group displayed high binding density in MPA, BSTMP, MeA, VMHVL, ARC > PVA, ACo, CG, BSTMA, LaDL > HIPPOv, HIPPOd > FCX, in agreement with the known distribution of brain ER. This confirms that [125I]MIE2 is a suitable tracer for in vivo studies of brain ER. Binding for the E2 group was very low and homogenous for all regions, indicative of maximal occupancy of all ER subtypes (mean inhibition (SD) = 86.6% (17.5)). By contrast, ER occupancy in the TAM group showed a marked regional distribution (ANOVA, p <.0001), and inhibition ranged from 40.5% (15.6) in VMHVL to 84.6% (4.5) in CoA. These results suggest that the pattern of ER occupancy by TAM is different from that of E2 and appears to be influenced by the regional proportion between ERalpha and ERbeta subtypes.
D. Pareto, None.
Abstract ID: 352
Our goal was to develop a magneto-optical nanoparticle that would serve as a preoperative MRI contrast agent for brain tumor delineation and intraoperative optical probe for tumor visualization. A long-circulating, superparamagnetic nanoparticle (CLIO) was conjugated to the near-infrared fluorescent (NIRF) dye Cy5.5 (Cy5.5–CLIO), yielding on the average 1.2 dyes per particle. GFP-transfected 9L-gliosarcoma cells were stereotactically implanted into Fischer 344 rats (n = 9). When tumors had reached 4–5 mm in diameter, Cy5.5–CLIO was injected (15 mg/kg iv), and MRI (T2-weighted spin-echo) of the tumors was performed after 24 hr. Craniotomy was performed, simulating an intraoperative scenario. Surface reflectance fluorescent images were acquired with a custom-built optical imaging system using filters for Cy5.5 and GFP. Brain tumors (n = 5) were excised, sectioned, and examined by fluorescence microscopy (epifluorescence and laser-scanning confocal), H&E, and DAB-enhanced Prussian blue staining. MRI of the brain tumors post injection of Cy5.5–CLIO showed a reduction of signal intensity with excellent accuracy in delineating tumor from brain tissue, as verified by correlative histology. In the simulation of an intraoperative setting, the surface reflectance fluorescence in the Cy5.5 channel provided accurate delineation of tumor extent as compared to GFP fluorescence as a gold standard. Histological examination of tumors showed an excellent correlation between NIRF signal, GFP fluorescence, and H&E staining. Cy5.5–CLIO is a dual-modality MR/optical agent that allows for both preoperative and intraoperative brain tumor delineation. The magnetic properties allow for preoperative assessment of tumor localization and borders by MRI. The fluorescent properties of the probe allow for the intraoperative visualization of tumor borders by optical imaging without the need of MRI and therefore might offer a novel tool for improved accuracy of brain tumor resection.
M.F. Kircher, None.
Abstract ID: 353
The serotonin transporter (SERT) is located on neuron processes and axon terminals and terminates the action of serotonin by reuptake. SERT alterations are reported in several psychiatric illnesses and SERT sites are the targets of SSRIs. We characterized in vivo SERT binding using [C-11]DASB (0.2–1.4 mCi iv). Studies were performed in adult male Sprague-Dawley rats. Animals were anesthetized (2% isoflurane), ventilated, and arterial blood gases monitored and maintained. Scan duration was 90 min. Regional cerebral blood flow (rCBF) was determined using [O-15]H2O during 10 min scans. Regional brain radioactivity and time-course was determined by micro-PET (Concorde Microsystems, R4). The localization of [C-11]DASB in brain in vivo corresponds to the known distribution of the SERT. Specific binding (region: cerebellum) decreased with increase in unlabelled tracer dose. The selective SERT antagonist sertraline (10 mg/kg ip) reduced binding ratios to 1:1 with the cerebellum. The peak value of radioactivity immediately following tracer administration correlates with CBF in the region of interest (r =.79, p = .02). The findings indicate: (1) [C-11]DASB is selective for labeling the SERT in rat brain in vivo; (2) SERT distribution of SERT in brain can be resolved by micro-PET; and (3) the initial peak in [C-11]DASB concentration in brain regions of interest corresponds to rCBF. Full quantification using kinetic modeling with an input function requires receptor occupancy be less than 10%, a condition challenging to meet in rat under these circumstances. Because of the regional differences in rCBF, Scatchard analyses to determine the amount of receptor occupancy using pseudo-equilibrium methods are difficult to interpret. Finally, arterial input, metabolite correction, physiological maintenance, scatter from surrounding regions, and specific binding in ‘reference region’ present additional technical difficulties. We will present some possible solutions to these problems. (Work supported by MH62185 and MH001997; Alan Wilson provided the DASB standards.)
M.D. Underwood, None.
Abstract ID: 354
Ca2+/calmodulin-dependent protein kinase type II (CaMKII) is a multisubunit protein kinase enriched in the postsynaptic density. Its activity is pivotal for synaptic plasticity including hippocampal long-term potentiation (LTP). A Ca2+-calmodulin complex is formed by an influx of Ca2+ through NMDA receptor activation and this complex will relieve CaMKII autoinhibition by binding the CaMKII auto-inhibitory/regulatory domain and thereby exposing the catalytic core to substrates. This initial activation is followed by autophosphorylation that renders the kinase constitutively active even in the absence of Ca2+. This autophosphorylation is catalyzed by mutual intersubunit reactions, therefore, activated CaMKII stays activated until all subunits are dephosphorylated. This sustained activity of CaMKII has been suggested as an underlying mechanism for LTP. In view of this, we have initiated a study to monitor the CaMKII activity in living neurons using fluorescence resonance energy transfer (FRET) technique in combination with two-photon laser scanning microscopy. We will first discuss the methodological aspects in nonneuronal cell lines and then report some preliminary data obtained from neurons. This study has a potential to be used to optically label a synapse which undergoes synaptic plasticity in vivo once we can transgenically express it in animals.
Y. Hayashi, None.
Poster Session 14: Cardiovascular Imaging
Abstract ID: 355
Thyroid hormone (TH) is involved in the regulation of many physiological and developmental processes, and plays a major role in the cardiovascular system. The specific role of thyroid receptors (TR) in the regulation of cardiac function is well documented in adult mice, but molecular biologists are interested to see how these receptors act at the embryonic and fetal stage, to further understand the biological mechanisms involved. The aim of this study was to evaluate the cardiovascular function, in utero, in 129Sv mouse fetuses lacking the alpha subtype of TR (TR-α0/0). Fifteen TR-α0/0 pregnant mice were studied between embryonic day (ED) 9.5 and 19.5 using a commercially available ultrasound machine (ATL HDI 5000) and a linear transducer (7–15 MHz). Studies were performed with mice under general anaesthesia. For each fetus, the following parameters were obtained from pulsed-waved Doppler time-velocity curves guided with color Doppler images both in the thoracic aorta and umbilical artery: heart rate, acceleration time (AT), deceleration time (DT), ejection time (ET), nonejection time (NET). Results were compared to values obtained in wild-type (WT) fetuses of the same strain. Heart rate was higher in TR-α0/0 fetuses than in WT fetuses, which is the opposite from what is observed in adults. AT and DT were found to be significantly higher in TR-α0/0 fetuses than in WT fetuses in the second half of pregnancy, both in the aorta and umbilical artery. Ultrasonography allows a noninvasive assessment of the cardiovascular function at the fetal stage and provides new insights into the physiology of this system. Additional elements on the role of TR-α in cardiovascular physiology are provided here, with interesting differences observed compared to the situation in adult mice. Several hypothesis are proposed to explain these differences, which can be explored using molecular biology techniques.
W. Mai, None.
Abstract ID: 356
Vascular endothelial cell growth factor (VEGF) plays a central role in the neovascularization under physiologic and pathologic conditions such as embryonic development, wound healing, diabetes, and tumor growth. An animal model capable of monitoring VEGF expression provides a powerful system to elucidate a role of VEGF in physiologic and pathological tissue growth. To visualize the VEGF promoter activity in living animals, we generated a transgenic mouse, VGL (VEGFp-GFPLuc), in which a GFP-luciferase fusion protein (GFPLuc) is expressed under the control of the human VEGF-A promoter. RT-PCR analysis and in vivo bioluminescence imaging of the VGL mice at Postnatal Day 7 exhibited a high level of a reporter gene transcription and intense bioluminescence in a variety of organs known to express VEGF, including brain, lung, spleen, and intestine. On the other hand, adult VGL mice showed lower reporter gene transcription and very little bioluminescence in the organs. To know whether the VGL mice can be used to follow VEGF expression in tumor growth, we crossed VGL mice with transgenic mammary tumor mice (MMTV-PymT). Consequently, we observed intense and increasing bioluminescence in the mammary tumors developed in VGL mice. Further, in situ hybridization with the tumor tissues and hypoxic culture experiment using isolated tumor cells revealed spatially and temporally correlated expression of VEGF and reporter gene in the tumor cells. GFP fluorescence was undetectable in the tumor cells although a GFPLuc protein of appropriate size was detected by Western blot. Furthermore, skin wounding experiment using adult VGL mice revealed induced bioluminescence in the lesion with a peak at Day 7. These results demonstrate that in vivo bioluminescence imaging permits highly sensitive longitudinal monitoring of VEGF induction during neovascularization and suggested that it may be effectively used for testing treatment strategies for diseases associated with angiogenesis.
T. Takahashi, None.
Abstract ID: 357
Tissue factor (TF) is a transmembrane glycoprotein involved in a number of important physiological and pathological processes including thrombosis, angiogenesis, and cell signaling. Molecular imaging of this receptor on cells could provide insight into the mechanism of several disease processes, but its low concentration on cell surfaces and the weak signal amplification capability of traditional paramagnetic contrast agents have limited the utility of MRI for this purpose. Accordingly, we have developed an MR molecular imaging contrast agent platform consisting of liquid perfluorocarbon nanoparticles that can deliver up to 90,000 Gd-DTPA ions to each binding site. Nanoparticles were targeted to TF receptors on cultured porcine aortic smooth muscle cell monolayers by sequential exposure to excess biotinylated anti-TF antibody, avidin, and then biotinylated perfluorocarbon nanoparticle emulsion. Nontargeted cell cultures were treated similarly except for exposure to the biotinylated antibody, whereas control cell cultures were left untreated. The cell monolayers were then imaged at 1.5 T to obtain relaxation rates and contrast-to-noise ratios (CNR). Although both the targeted (T) and nontargeted (NT) cell layers exhibited a statistically higher longitudinal relaxation rate than did the untreated (UT) cell layers (see figure), targeted binding resulted in a 3.7 times greater increase, indicating successful targeting to TF receptors. For a series of spin-echo sequences with varied repetition times, this produced a maximal CNR between the targeted and untreated cells of 17.2, while the nontargeted cells exhibited a CNR <5. The high paramagnetic payload of these particles provides the signal amplification necessary to visualize sparsely concentrated epitopes on the surface of cell monolayers, and provides opportunities for understanding their role in disease states such as tumor angiogenesis and postangioplasty restenosis.
A.M. Morawski, Philips Medical Systems 1.
Abstract ID: 359
The purpose of the study was to demonstrate the feasibility of XMR guidance for left ventricular (LV) catheterization of transendocardial delivery of extracellular tissue markers. This XMR approach is a potential modality for direct delivery of gene therapy and multipotential cells to the myocardium. Experiments were performed in six dogs using a novel hybrid interventional XMR system. The initial advancement of a guide-wire and catheter into the heart was performed under X-ray fluoroscopic guidance. MR fluoroscopy for passive catheter tracking was done with the acquisition of 10 images/sec using steady-state free precession (bFFE) sequence. A steerable dual lumen catheter was used to deliver 0.5, 1.0, and 2.0 ml of Gd-DTPA-BMA and Evans blue dye via injections through endocardial surface of LV. The deposition of the solution was done using dual-inversion recovery T1-enhanced black blood sequence. Gd-enhanced region on MRI and blue-stained tissue were measured and compared. LV catheterization via arterial access was feasible under both X-ray and MR fluoroscopy. MRI showed the selected myocardial site of injection of the solutions. Intramyocardial injection of injectate caused no significant changes in heart rate or blood pressure. The sites of injections were bright on T1-sequence, but varied with the volume of injectate. The size of gadolinium-enhanced regions were 1.5 ± 0.6% of LV after 0.5 ml and 7.0 ± 0.5% after 2.0.ml and on blue dye 2.3 ± 0.6% and 8.3 ± 0.4%, respectively, (<.05, paired t test). The MR compatible catheter can be navigated inside the LV under MR guidance to “hit the target.” The gadolinium oxide labels on the tip of the catheter were heplful for trucking the catheter inside the aorta and the LV. Transendocardial delivery of potential therapeutic solutions is feasible under XMR guidance. The injection catheter can be navigated under MR guidance to “hit the target.” The distribution of the extracellular markers in myocardium was volume dependent.
M. Saeed, None.
Abstract ID: 360
Angiogenesis is an early feature of plaque development in atherosclerosis, which is not detectable by noninvasive clinical imaging. We utilized targeted paramagnetic nanoparticles to detect expression of αíô3-integrin, an angiogenic marker, with routine MRI at 1.5 Tesla. Rabbits were fed either 1% cholesterol (n = 16) or standard chow (n = 4) for 80 days. Cholesterol-fed rabbits received an intravenous dose of either (1) αíô3-targeted paramagnetic nanoparticles (n = 6), (2) nontargeted paramagnetic nanoparticles (n = 4), or (3) pretreatment with αíô3-targeted nonparamagnetic nanoparticles followed by αíô3-targeted paramagnetic nanoparticles (n = 6). All standard chow rabbits received αíô3-targeted paramagnetic nanoparticles. Signal enhancement was measured from black-blood MRI of the abdominal aorta acquired before and 2 hr after nanoparticle injection. Delayed enhancement (20 min) with Gd-DTPA (0.2 mmol/kg) was determined following the 2-hr timepoint. Lumenal and vessel wall areas of the abdominal aorta were equivalent for cholesterol-fed and control diet animals. Averaged over the entire abdominal aorta, αíô3-targeted nanoparticles (T) produced higher MRI signal enhancement (p <.05) in cholesterol-fed animals compared with nontargeted nanoparticles (NT) or control diet animals (Figure). In the competition group (C), blunted MRI enhancement confirmed the specificity of αíô3-targeted nanoparticles. Delayed enhancement with Gd-DTPA, a nonspecific MRI contrast agent, could not distinguish early atherosclerotic changes of the vaso vasorum. In conclusion, molecular imaging with αíô3-targeted paramagnetic nanoparticles may provide a sensitive non-invasive method for defining the burden and evolution of atherosclerosis in susceptible individuals.
P.M. Winter, None.
Abstract ID: 361
Atherosclerotic plaque rupture is widely recognized as the proximate cause of vascular thrombosis and resultant myocardial infarction and stroke. Fast multislice CT is emerging as a useful noninvasive, high-resolution modality for diagnostic coronary and carotid luminal studies. However, angiographic techniques cannot distinguish microthrombi formation within microfissures of vulnerable plaques, which often herald serious clinical sequelae. The objective of this study was to demonstrate the first, fibrin-targeted molecular imaging contrast agent that could be used to recognize early plaque disruption with multislice CT. In Experiment 1, inherent radio-opacity of the nanoparticles was demonstrated by imaging suspensions serially diluted from 20% (i.e., undiluted) to 0.02% w/v with saline. The nanoparticles were found to be radio-opaque down to a concentration of 0.2% w/v when imaged with a Philips AcQSim-CT scanner. In Experiment 2, anti-fibrin monoclonal antibodies were coupled through avidin-biotin cross-linking to nanoparticles and bound to plasma clots (n = 3) suspended in saline on suture in vitro. Fibrin-targeted nanoparticles provided marked radio-opaque contrast in sharp contradistinction to the feature-poor images of nontargeted control clots (n = 3, Figure). Grayscale contrast enhancement was greater (p <.05) in the targeted clots (209.7 ± 4.9) versus control clots exposed to the nontargeted nanoparticles (160.2 ± 1.3). Contrast-to-noise ratios were markedly greater (p <.05) for the targeted thrombus (22.3 ± 2.2) than for the nontargeted control agent (5.1 ± 0.2). These results provide the first demonstration of a fibrin-specific molecular imaging agent designed to increase the sensitivity of thrombus detection with CT. Molecular imaging with radio-opaque contrast nanoparticles may allow early plaque ruptures to be localized and treated during routine cardiac catheterization or intervention.
H.P. Shukla, Philips Medical Systems 5.
Abstract ID: 362
MRI-based vascular intervention techniques will require novel MR-compatible technology, whether applied locally or ideally systemically, to inhibit restenosis. Paramagnetic nanoparticles directed against avB3-integrins, expressed on VSMC, and collagen III, present within the extracellular matrix, provide opportune targets within the media. This objective requires penetration, specific recognition of biochemical signatures, and sustained persistence of the nanoparticles following balloon injury. Carotid arteries of six pigs were injured bilaterally by balloon-overstretch injury (5 × 20 mm) and treated with avB3-integrin or collagen III targeted nanoparticles or saline. Contrast was administered locally immediately postinjury and penetrated through the stretch-created microfractures. In each pig, one carotid served as a control and the contralateral vessel received the targeted treatment. After 15 min of treatment, unbound nanoparticles were washed from the injured tissues and circulation was re-established. Porcine carotids underwent T1-weighted MRI at 1.5 T to demonstrate penetration, binding, and retention of the targeted particles. MR carotid angiograms were indistinguishable between control and targeted treatments (Figure). Both avB3-targeted and collagen III-targeted paramagnetic nanoparticles revealed similar asymmetric three-dimensional patterns of mural injury, which were ˜50% longer (31 ± 5 mm) than the balloon length with an average injury volume of 955 ± 234 mm3. No injury patterns were discernible for the contralateral control vessels. Contrast-to-noise ratio of avB3-targeted nanoparticles (13.8 ± 5.2) was 4-fold greater (p <.05) than the collagen III-targeted formulations (3.3 ± 0.3), suggesting a greater retention of the integrin-targeted agent. These results demonstrate that paramagnetic nanoparticles can penetrate and bind cellular and extracellular biochemical epitopes within the media following balloon-stretch injury. Paramagnetic nanoparticles targeted to media components after angioplasty could interrogate biological responses, quantify microfracture severity, and provide a platform for local anti-restenotic therapy.
T. Cyrus, None.
Abstract ID: 363
Cell therapy shows potential promise for treating ischemic heart disease. However, current understanding of cell survival relies on postmortem histology, which precludes longitudinal analysis. In this study, we hypothesize that cells marked with reporter genes and transplanted into the myocardium can be followed noninvasively using optical bioluminescence and micro positron emission tomography (micro-PET) imaging. Embryonic rat H9c2 cardiomyoblasts were transfected with adenovirus carrying cytomegalovirus promoter driving herpes simplex type 1 mutant thymidine kinase (Ad-CMV-HSV1-sr39tk) or firefly luciferase (Ad-CMV-Fluc) reporter genes. Study rats (n = 30) were injected with transfected cells (3 × 10) and control rats (n = 5) with saline via aseptic lateral thoracotomy. Optical and micro-PET imaging were acquired using 
J.C. Wu, None.
Abstract ID: 364
Noninvasive imaging of therapeutic gene expression should be useful for evaluating the effectiveness of cardiac gene therapy. We have previously demonstrated the feasibility of using micro positron emission tomography (micro-PET) to monitor the expression of a single PET reporter gene in rat myocardium (Wu et al., 2002. Circulation July 9; 106(2): 180-183). We have now developed a bicistronic adenoviral vector (Ad.CMV.D2R80a.IRES.HSV1sr39tk) for linking the expressions of two PET reporter genes—a mutant dopamine-2 receptor (D2R80a) and a mutant herpes simplex virus type 1 thymidine kinase (HSV1-sr39tk)—via an internal ribosomal entry site (IRES) cis-element. To verify that the two reporter gene expressions are coupled, we injected various titers up to 2 × 109 pfu of Ad.CMV.D2R80a.IRES.HSV1sr39tk directly into the anterolateral myocardial wall of experimental nude rats (n = 7) or saline into control rats (n = 6). Two to three days after injection, we used micro-PET to quantify the D2R80a and HSV1-sr39tk-dependent myocardial accumulations of [18F]FESP (3.14 ± 0.05 mCi) and [18F]FHBG (2.92 ± 0.26 mCi), respectively. Weekly [18F]FESP and [18 F]FHBG scans thereafter were done to assess the long-term expressions of the two reporter genes. In experimental rats injected with titers greater than 1 × 10 pfu of Ad.CMV.D2R80a.IRES.HSV1sr39tk, [18F]FESP and [18F]FHBG scans on Days 2 and 3, respectively, revealed significant (p <.05) tracer uptake (0.22 ± 0.06%ID/g FESP; 0.18 ± 0.02%ID/g FHBG), compared to negligible uptake in control rats (0.062 ± 0.03%ID/g FESP; 0.016 ± 0.008%ID/g FHBG). Viral titers lower than 1 × 108 pfu were undetectable. Longitudinal [18F]FESP and [18F]FHBG scans of experimental rats collectively showed a strong correlation (r2 = .76) between the expression of the two reporter genes. These results suggest that an IRES-based bicistronic adenoviral vector can be reliably used in conjunction with PET technology to indirectly monitor therapeutic gene (e.g., VEGF) expression in murine, and potentially human, myocardium.
I.Y. Chen, None.
Abstract ID: 365
Peripheral vascular arterial insufficiency secondary to atherosclerosis is a common condition which may result in claudication or limb loss. Currently, the most definitive clinical diagnostic tool used intraoperatively to determine peripheral vascular disease (PVD) is angiography using fluoroscopy. While having relatively good resolution, X-ray angiography is not without undesirable effects. Safety issues such as exposure to ionizing radiation as well as the need for specialized handling, eyeware, and clothing, make X-ray angiography a somewhat cumbersome technology. Fluorescence imaging is an alternative to X-ray angiography that offers the advantages of abscence of ionizing radiation, less bulk, and no requirement for specialized equipments and clothing. We utilized a fluorescence imaging system (SPY 2000 imaging system) to visualize the arterial vasculature of PVD patients during bypass surgery. An exogenous fluorophore, indocyanine green (ICG), was utilized with the Spy 2000 camera device to track its distribution in vivo. Our findings revealed that ICG excited with laser energy in the near-infrared region (806 nm) emitts a signal at 830 nm that can be captured as it passes through the arterial vessels. In addition, results from this feasibility study indicate that the method provides qualitative flow information on the integrity of the newly grafted vessels. Our data also suggest that the method may provide information on distal flow runoff.
B. Ramjiawan, None.
Abstract ID: 366
Atherosclerotic plaque imaging using magnetic resonance (MR) contrast agents may be important in determining its vulnerability and potential for rupture. Gadofluorines, a new class of contrast agents, based on a macrocyclic gadolinium complex, or based on a Gd-DO3A derivative, form micelles in aqueous solution and with a potential for specific plaque enhancement. The objective of the study was to investigate a new contrast agent, Gadofluorine M (Schering), on plaques using Apolipoprotein (Apo) E mice. Fifteen-month-old Apo E mice (n = 4) and an age-matched wild type (WT) (n = 4) group underwent in vivo MR microscopy (MRM) of the abdominal aorta using a 9.4-T MR system. Pre- and post-contrast-enhanced (CE) (24 hr) MRM was performed using a T1W black blood sequence. Sixteen contiguous 500-μm-thick slices with an in-plane resolution of 93 mm were acquired in 30 min. Gadofluorine M with the fluorescent-labeled Cy3.5 (100 mmol/kg) was injected via the tail vein. After imaging the mice, the aorta was isolated. Confocal fluorescence microscopy and immunostaining were performed. MRM images of the matched (pre and 24 hr-post) slices were used for analysis. In the Apo E mouse group, there was a significant increase in contrast-to-noise ratio (CNR) for wall/lumen and wall/muscle in the post-CE versus pre-CE images (paired t test, p <.05). There was no increase in CNR for wall/lumen and wall/muscle for the WT group. The ratio of the post to the precontrast signal intensity of the wall (normalized to muscle) was 2.25 (enhancement by 125%) in Apo E versus 0.91 for the WT groups. Confocal fluorescence microscopy and immunostaining showed specific uptake in the plaque area. Gadofluorine M showed good contrast enhancement in plaques of ApoE mice. CE MRM with Gadofluorine M may play a role in the targeted molecular imaging of plaques, and improve detection and definition.
J.S. Aguinaldo, None.
Abstract ID: 367
Delayed contrast-enhanced (DCE) magnetic resonance (MR) imaging offers a method to noninvasively visualize infarct location and size. In addition, the ability to magnetically label mesenchymal stem cells (MSCs) offers a method to track the fate of MSCs used to regenerate myocardial tissue. Our objective was to test the feasibility of MR fluoroscopy to deliver MR-labeled MSCs for precise targeting to the infarct in a canine myocardial infarction (MI) model. Mongrel dogs (n = 3) were subjected to a coronary artery balloon occlusion for 90 min followed by reperfusion to create MI. MSCs isolated from bone marrow and magnetically labeled with 25 μg Fe/ml Feridex and 375 ng/ml poly-
D.L. Kraitchman, Osiris Therapeutics, Inc. 1; Berlex Laboratories, Inc. 1; Schering AG 1.
Poster Session 15: Late Breaking
Abstract ID: 368
Luminescent imaging has been used to determine the in vitro and in vivo efficacy of topotecan, taxol, doxorubicin, and cisplatin against human breast carcinoma MCF-7 Luc and prostate carcinoma DU-145 Luc cells transfected with Photinus pyralis luciferase. Tumor cells were transfected with plasmid pcDNA-3.1-Luc. In vitro experiments demonstrated dose-response drug activity against both cell lines using both luminescent imaging and luminometry of extracted cellular ATP. In vivo experiments were performed using athymic nude mice inoculated intraperitoneally with 5 × 106 MCF-7 Luc cells or subcutaneously with 5 × 106 DU-145 Luc cells and then treated with topotecan, taxol, doxorubicin, or cisplatin. Tumor progression and regression were monitored for 27 days after treatment. Animals inoculated subcutaneously with DU-145 Luc cells and treated with these cytotoxic agents demonstrated significant tumor growth and regression as measured with calipers and luminescent imaging. High correlation was observed between caliper and imaging results. Similarly, tumor progression and regression were measurable using luminescent imaging for untreated and treated mice inoculated intraperironeally with MCF-7 Luc cells. These data indicate that luminescent imaging is a useful tool for evaluating anti-cancer drugs in vivo and may prove to be particularly useful for the development of novel agents, particularly those targeting deep tissue and metastatic lesions.
A. Aller, None.
Abstract ID: 369
Anti-angiogenic therapy of gliomas using EMD 121974, a cyclic RGD antagonist of cell adhesion receptor αv-integrins, showed favorable results. The ability to visualize and quantify αv-integrin expression level will play a pivotal role in visualizing brain tumor growth, quantifying integrin receptor expression, and following anti-integrin treatment efficacy, as well as in developing more potent anti-integrin drugs. Recently, we labeled cyclic RGD peptide c(RGDyK) with 18F through a prosthetic labeling group 4-[18F]fluorobenzoyl. The radiotracer was able to detect U87MG glioblastoma in both subcutaneous and orthotopic models. Weekly micro-PET scans to follow brain tumor growth revealed positive correlation between the magnitude of uptake and tumor growth rate. However, fast tumor washout and unfavorable hepatobiliary excretion of this tracer limited its potential clinical application. In this study, we modified the radiotracer by inserting a poly(ethylene glycol) (PEG) moiety (M.W. = 3,400) between the RGD moiety and the fluorine-18 label. Cyclic RGD peptide is coupled with t-Boc protected PEG-NHS ester followed by TFA cleavage, after which the N-terminal is conjugation labeled with 18F by reacting with N-succinimidyl 4-[18F]fluorobenzoate. The reaction is complete within 30 min and radiochemical yield is about 20–30% based upon the activated ester [18F]SFB. Specific activity is over 100 GBq/umol. The new tracer revealed fast blood clearance, resulting in very high tumor-to-background ratio in the subcutaneous U87MG model (tumor: 1.9 ± 0.3, liver: 0.6 ± 0.1, kidney 1.2 ± 0.2, and intestines: 0.4 ± 0.1%ID/g at 4 hr postinjection). This tracer also accumulated in orthotopic U87MG tumor, with tumor-to-brain ratio of 10 ± 1 at 1 hr postinjection, although the brain tumor SUV is less than the subcutaneous tumor. In conclusion, PEGylated RGD peptide resulted in sustained blood concentration, prolonged tumor retention, and improved in vivo kinetics.
X. Chen, None.
Abstract ID: 370
The hyperpolarization technique can increase 10,000 times the signal sensitivity of NMR for 3He and 12Xe and make instantaneous imaging of the cavity gases. This technology is expected as practical uses of medical equipment that makes it possible to carry out instantaneous diagnostic of lung functions and of medical technologies for diagnosis and prevention of cerebral infarction based on high-accuracy and highspeed imaging of the blood flow in the brain. An automated batch-type apparatus for hyperpolarization, including ultra pure gas application system and using high-power diode laser arrays was developed. The stray field of the horizontal superconducting magnet (2T) was used for the optical pumping processes. In this apparatus, small pieces of the alkaline metal Rb and Xe and N2 gas with pressure of around 1.2 atm are sealed in a cylinder cell of 60 mm diameter and 100-mm-length made of Pyrex glass with a flat window. The cell is heated to 100°C in a magnetic field of around 100 G and irradiated by 30 W semiconductor laser light of 794.7 ± 1 nm in wavelength, circularly polarized through a quarter-wave plate. We achieved 1–4% polarization of 12Xe for Xe (98%) and N2 (2%) gas mixture. We have prepared a surface coil for 12Xe frequency at 2 T (˜23.56 MHz) with the matching circuits and performed MRI experiments. MRI images of hyperpolarized Xe gas extracted in a syringe were acquired. The faster acquisition sequences, for example, EPI sequence is essential technology for the high-accuracy and high-speed imaging of the blood flow in the brain. The time resolution of acquisition was enhanced from about 60 sec for the GE sequence (TE = 3.7 msec, 64 × 64) to about 500 msec for the four-shots segmented EPI sequence (number of segments = 4, TE = 9.68 msec, 64 × 64). The research should facilitate work on NMR/MRI to shorten measuring times, produce more diverse information, and increase accuracy.
M. Hattori, None.
Abstract ID: 371
Coelenterazine is the substrate for Renilla luciferase (Rluc), a bioluminescent reporter protein recently used by optical imaging in living mice. So far there are more than 10 coelenterazine analogs commercially available. This research is aimed to determine the analogs best suitable for cell culture and animal studies, based on their signal intensity, background bioluminescent intensity, influx and efflux across membrane, and biodistribution in mouse. In vitro study, HeLa cells were stably transfected with constitutively expressed Rluc (pRL-CMV). Bioluminescence was detected with a cooled charged-coupled device camera. Rluc reactions with the analogs were monitored in PBS, DMEM, or serum itself. Coelenterazine -f and -h showed the highest signal intensity in PBS (peak at 15 to 20 min), while coelenterazine -hcp, -cp, and -ip produced stronger background bioluminescence in serum (peak at 5 min). In mouse model, liver cells were transfected with humanized Rluc gene (phRL-CMV) by tail vein injection. Coelenterazine analogs were administrated into mice by interperitoneal or intravenous and whole body images were collected immediately for 30 min. Compared with intraperitoneal coelenterazine injection, mice with intravenous injection gave higher light intensity (10-fold stronger in first minute) but also faster declination rate. Among six analogs we tested in mice, coelenterazine -native, -h, and -f showed relative higher intensity. Both Rluc and firefly luciferase (Fluc) could be imaged in the same liver sequentially without cross-reaction between substrates. To test whether P-glycoprotein (Pgp) mediates efflux of coelenterazine in liver, four analogs (-native, -h, -n, and -cp) were tested in wild type or Pgp knock out mice. Preliminary data implied that there might be more factors involved in influx or efflux of coelenterzine besides Pgp. These data suggest that under optimized condition, Fluc and Rluc dual-luciferase system is feasible to be used in studying two biological processes in vitro and in vivo.
H. Zhao, None.
Abstract ID: 372
Mesenchymal stem cells (MSCs) found in human bone marrow are able to proliferate extensively in vitro without loss of differentiation potential. They represent a novel cellular therapy of inherited or degenerative diseases. We have demonstrated that culture-expanded autologous and allogeneic MSCs can be safely infused into humans and the preliminary results showed that MSCs facilitate hematopoietic engraftment and reduce GVHD. However, there is insufficient information about in vivo distribution and survival of intravenously infused MSCs and their contribution to normal tissue function. In this study, we tested the in vitro feasibility of using Na+/I− symporter (NIS) gene as a reporter gene for the eventual imaging of MSCs' distribution and persistence in vivo, simply with radioiodide or equivalent (such as the pertechnetate) as tracer. Human MSC (hMSC) cultures were established from bone marrow aspirate specimens, and were transduced with the NIS using lentiviral vector wpts.hNIS. To determine if hNIS gene transferred occurred, DNA was isolated from 500 K transduced hMSC and PCR was done using the wpt.hNIS primers. A positive band was seen at 650 bp location only in NIS-hMSCs, but not in untransduced hMSCs, indicating gene transfer. NIS protein in transduced cells was evaluated by Western blot analysis. In addition, each NIS-hMSC clone was tested for NIS function by an in vitro RAIU assay, which includes uptake, inhibition, and efflux. The radioiodide uptake increased 30+ fold in NIS-hMSC comparing to wt hMSC. The amount of uptake was proportional to the number of the cells. Such uptake was inhibited completely by NaClO4, a known NIS specific inhibitor. The in vitro half lives (retention) of radioiodide in NIS-hMCS was 7 min. Due to the fact that in vivo half lives of iodide retention is always longer, this efflux rate will be fine for in vivo tracking MSC infusion on animal/human by repeatedly imaging.
Z. Lee, None.
Abstract ID: 373
A.A. Tzika, None.
Abstract ID: 374
Tumor necrosis factor-α (TNFα) is a major cytokine with cytotoxic and cytostatic activities against various human cancer epithelial cells and has marked effects on permeability and procoagulant activity on tumor-associated neovasculature. In order to explore the potential application of vascular targeted delivery of TNFα to angiogenic endothelium as a novel therapy, we transfected a TNFα expressing gene with 5′ forward signal peptide coding sequence under the CMV promoter into a mouse microvascular endothelial cell line. A murine squamous cell carcinoma SCCVII cell line and a TNFα resistance PMEL cell line were used for positive and negative controls, respectively. DAPI stained nuclear morphology combined with terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) method was used to observe and quantify cell deaths. In order to further address the cell death effect by TNFα, a single dual promoters expression plasmid was constructed in which the TNFα and GFP was driven by an independent strong promoter, and was transfected into the same cell lines. We found that TNFα transfection could lead to significant cultured endothelial cell death after 48 hr transfection in a dose-dependent manner (20.2 ± 3.5%, 0.375 mg/cm2; 29.4 ± 3.9% 0.75 mg/cm2). This was statistically significant (p <.05) as compared to the same amount of vector only (5.8 ± 2.1%) or lipofectamine only (2.4 ± 1.5%) controls. We found not only the apoptotic signals were co-incident and co-localized with the EGFP signals, also the number of apoptotic cells in TNFα transfection was more than the number of transfected cells (9.2 ± 2.8%, as judged by GFP expression) after 48 hr of the transfection, strongly suggesting a bystander killing effect of TNFα transfection and expression. Our data suggest that TNFα gene expression has a potential cytotoxic effect on endothelial as well as cancer epithelial cells. These data are critically important for our pursuit of vascular targeted gene therapy for neoplasms using TNFα.
J. Xie, None.
Abstract ID: 375
Optically labeled receptor-binding agents offer high observational sensitivity, as well as the ability to measure receptor density in target tissues. Our objective was to demonstrate receptor-binding properties of an optically labeled receptor-binding diagnostic agent, Cy5.5-DTPA-galactosyl-dextran. The optical reporter Cy5.5 was covalently coupled to DTPA-galactosyl-dextran or DTPA-dextran. DMSO was added to the N-succinimidyl-ester of the dye (0.15 mg). The Cy5.5 solution was added dropwise to a DMSO solution of DTPA-galactosyl-dextran or DTPA-dextran (90 nmol dextran/ml). After one hour the product was purified (G25) to yield approximately 8 Cy5.5s per dextran T70. Male BALBc mice (10 mg/kg ketamine, 1 mg/kg acepromizin ip) were imaged in the SAMI imaging system (Advanced Research Technologies). Cy5.5-DTPA-dextran or Cy5.5-DTPA-galactosyl-dextran was administered (24 nmol/kg) via the tail vein. With the animals supine on a heated scan plate, a rectangular area covering the heart and liver was selected using a top-viewing digital camera. The rectangle was raster scanned in 2-mm steps using appropriate excitation/emission wavelengths at 22.7 mW in 25 sec. Scans were repeated continuously for 30 min. ROIs over the heart and liver were used to generate curves of fluorescence intensity. The receptor-binding agent, Cy5.5-DTPA-galactosyl-dextran, displayed hepatic accumulation similar to [99mTc]DPTA-galactosyl-dextran. At 2.5 min, a typical liver ROI contained 4500 kcpm and peaked within 20 min at 10,000 kcpm. When compared to the radiolabel receptor-binding agent, Cy5.5-DTPA-galactosyl-dextran displayed similar pharmacokinetic properties with an approximately 100-fold sensitivity—typical nuclear count rates using a pinhole collimator were 100 kcpm. The time-intensity curves for Cy5.5-DTPA-dextran did not demonstrate hepatic uptake. This study suggests three potential roles for optical imaging of receptor-binding diagnostic agents. First is the use of optical reporters during the initial development, when receptor density is unknown. Second, the increased sensitivity should permit the implementation of tomography. Third is the estimation of target receptor concentration via kinetic modeling.
R.F. Mattrey, None.
Abstract ID: 376
Adenoviral vectors have been applied for gene therapy of numerous diseases. Evaluation of vector targeting strategies has been accomplished by reporter systems; however, these techniques cannot assess the true physical biodistribution of viral particles which does not absolutely correlate with reporter gene expression. In the use of replicative adenovirus for cancer therapy, direct dynamic monitoring of viral replication and localization of progeny virions are essential to further understand and develop these agents. To address these issues, we hypothesized that adenovirus packaged with an expressed fluorescent protein label during replication would give virions a fluorescent property useful for their detection in targeting purposes. Moreover, this genetic labeling system can be applied to replicative adenoviral agents to monitor replication and spread. To achieve such labeling, we fused EGFP to the C-terminus of the external capsid cement protein IX of human adenovirus type 5. Progeny yield was decreased but remained on the same order of magnitude as that of control. Cytopathic effect and viral binding were not affected. After double CsCl ultracentifugation, fractionation of the bottom and top viral bands revealed major viral DNA and fluorescent peaks for the bottom band and minor peaks for the top band. Western blot analysis confirmed high amounts of IX-EGFP in both the bottom and top bands colocalizing with the major hexon capsid protein. Ad-IX-EGFP could be visualized by fluorescence microscopy, detected by FACS to assess viral binding, and tracked during infection. Moreover, the labeled virus was detectable in frozen sections down to a single particle following injection in vivo. We have extended our strategy to label adenovirus with red fluorescent protein as well as luciferase and are currently characterizing these viruses. We believe that our virus labeling system has great implications for evaluating vector targeting, viral replication, and adenovirus biology.
L.P. Le, None.
Abstract ID: 377
Development of posttraumatic knee arthrosis is one of the most common conditions managed by orthopedic surgeons worldwide. New orthopedic theory, based on the principle of tissue homeostasis, indicates the loss of osseous homeostasis (manifested by persistently positive technetium 99m bone scans) precedes the development of overt arthrosis, and therefore identifies joints “at risk” of arthrosis at a time when radiographs and even MRI bone signal are still normal. The purpose of this study was to document possible development of early degenerative arthrosis of the knee as a function of osseous homeostasis characteristics. Pre- and postoperative Tc99m-MDP bone scans and standing AP radiographs of adult patients with symptomatic meniscal tears undergoing arthroscopic partial medial meniscectomy were tracked. All patients had preoperative confirmation of meniscal injury on MRI. Forty-two patients (25 males, 17 females, ages 22–49) with 46 symptomatic knees were included. All 46 knees manifested loss of osseous homeostasis (positive technetium scan) and normal radiographs preoperatively. Eighty percent (37/46) manifested resolution or substantial improvement of the abnormal bone scan activity 18 months postoperatively without development of degenerative changes. Six of the nine knees whose abnormal bone scans persisted or worsened postoperatively demonstrated early degenerative changes on radiographs. The conclusions drawn from these data are: (1) Persistent loss of osseous homeostasis precedes and therefore can identify joints “at risk” of development of early degenerative changes in knees with meniscal injury; (2) Loss of osseous homeostasis reflects the pathophysiologic effect on living bone of damage of the meniscal cartilages at a time when both radiographs and MR indicate normalcy of bone; (3) Future advances in orthopaedic surgical diagnosis and treatment can be expected with development of molecular imaging techniques that can manifest soft tissue homeostasis characteristics of musculoskeletal structures such as articular cartilage, menisci, ligaments, tendons, synovium, and muscle.
S.F. Dye, None.
Abstract ID: 378
The use of nonreplicating adenoviral vectors as delivery vehicles in gene therapy has been seriously hampered by safety concerns and the inability to efficiently deliver the virus to the target tissue. Typically greater than 99% of the virus will be removed from the circulating blood supply within several minutes of delivery, with the majority of the virus filtered from the blood in the liver. Viral infection within the liver may then occur with resultant toxicity. We have examined the biodistribution of viral infection over time following delivery by different routes of administration. The use of a nonreplicating luciferase expressing adenovirus vector and subsequent whole animal imaging of bioluminescence production has allowed us to examine significant differences in the targeting of the virus. It was shown that the biodistribution of viral gene expression over time was dependent on both the delivery mechanism employed (intravenous verses intraperitoneal) and the dose of the virus used. These results demonstrate that it is possible to retain the virus in the circulation for longer periods and to alter the targeting profile by simply adjusting the dose and method of delivery used. The reasons for these differences in the distribution of viral infection and the implications for the use of adenovirus in gene therapy will be discussed.
S.H. Thorne, None.
Abstract ID: 379
A growing variety of cell surface markers is intimately associated with onset and progression of human disease. Growth factors and antibodies that recognize these markers can serve as targeting vehicles for imaging radionuclides and other imaging agents. Importantly, imaging of multiple markers might lead to better characterization of each patient, and therefore, to development of personalized treatment regiments and to a rational selection of patients for experimental therapeutics. Lack of efficient technology for “loading” imaging agents onto proteins prevents a widespread use of targeted imaging. Current loading strategies are based on random chemical modification of targeting proteins and share common problems such as (i) partial inactivation of functional domains; (ii) heterogeneity of final products; and (iii) customized conjugation procedure for every targeting protein. An attractive alternative would be assembly of targeting imaging complexes that include a universal radiolabeled adapter protein that can be noncovalently “docked” to any targeting protein expressed with a corresponding “docking tag.” In addition to being nondestructive, this approach allows the use of multiple existing and newly discovered targeting proteins in a rapid and uniform fashion. We have recently developed adapter/docking tag system based on interactions between an 18-127aa fragment of human RNase I and a 1-15aa fragment of RNase I fused to a targeting protein. The docking system was used to assemble vascular endothelial growth factor (VEGF) driven 99mTc-labeled complexes for imaging of vasculature in tumor-bearing mice. We found that these complexes selectively and specifically image tumor neovasculature in mouse tumors as small as 2–3 mm in diameter. We propose the use of standardized adapter/docking tag system as an alternative to direct derivatization of any targeting proteins with chelators for imaging and, potentially, therapeutic radionuclides. Availability of multiple imaging proteins might have tremendous implications for the development and evaluation of novel anti-cancer and, specifically, anti-angiogenic therapies.
J.M. Backer, Sibtech, Inc. 4, 5.
Abstract ID: 380
Light passing through tissues is absorbed, scattered, reflected, and refracted; the problems of image reconstruction using laser CT (CTLM) are therefore much greater than in X-ray CT. These problems delayed the development of optical imaging, but have been solved with a level of accuracy permitting absorption and/or fluorescence imaging in human patients. Using appropriate molecules or fluorophores, CTLM provides CT and 3-D images of anatomic, physiologic, and metabolic processes, and can characterize the effects of interventional and pharmacologic therapies. CTLM is mechanically similar to X-ray CT but the X-ray tube has been replaced by a laser and there are two rings of detectors, one configured for absorbed light, from an external laser source, and the other for fluorescence, from a fluorophore within the tissue.
E.N. Milne, Imaging Diagnostic Systems Inc. 2.