Sage Journals: Discover world-class research

Abstract

Small-animal tumor models are essential for developing translational therapeutic strategies in oncology research, with imaging having an increasingly important role. Magnetic resonance imaging (MRI) offers tumor localization, volumetric measurement, and the potential for advanced physiologic imaging but is less well suited to high-throughput studies and has limited capacity to assess early tumor growth. Bioluminescence imaging (BLI) identifies tumors early, monitors tumor growth, and efficiently measures response to therapeutic intervention. Generally, BLI signals have been found to correlate well with magnetic resonance measurements of tumor volume. However, in our studies of small-animal models of malignant brain tumors, we have observed specific instances in which BLI data do not correlate with corresponding MRIs. These observations led us to hypothesize that use of BLI and MRI together, rather than in isolation, would allow more effective and efficient measures of tumor growth in preclinical studies. Herein we describe combining BLI and MRI studies to characterize tumor growth in a mouse model of glioblastoma. The results led us to suggest a cost-effective, multimodality strategy for selecting cohorts of animals with similar tumor growth patterns that improves the accuracy of longitudinal in vivo measurements of tumor growth and treatment response in preclinical therapeutic studies.

IN ONCOLOGY RESEARCH, small-animal models are critical for the development of effective therapeutic strategies. The majority of rodent studies involve mice, owing in large part to the ability to readily manipulate the murine genome. Mice are now used routinely to probe the effects of individual genes on tumor growth and development and in preclinical studies of new therapeutics. Methods to accurately monitor tumor growth and response to therapy in vivo are a key to the effectiveness of these preclinical investigations. This is particularly true for brain tumors. Intracranial models of brain tumors, in which tumor cell lines are implanted directly into the brain^1–4 or mice that are genetically engineered to spontaneously initiate tumor growth,^5–7 are the most clinically relevant because the tumor grows directly within the microenvironment of the brain itself. However, use of these models requires the development and application of methods to quantitatively measure intracranial tumor growth.

Noninvasive small-animal imaging can provide a solution to this challenging problem.^1,2,8–12 Magnetic resonance imaging (MRI) offers excellent anatomic tumor localization and accurate volumetric measurement^1,9,11 and is well suited to longitudinal studies of tumor growth and therapeutic response. Experiments such as diffusion-weighted imaging,^13–16 dynamic contrast imaging,^17–21 and magnetic resonance spectroscopy^22,23 afford insights into tumor biology extending beyond anatomic measures. In vivo bioluminescence is a valuable tool for monitoring the growth of small-animal tumor models, characterizing signaling pathways, and quantifying response to therapeutic intervention in large cohorts of animals,^24–27 and its use has increased dramatically over the past 10 years. The utility of bioluminescence imaging (BLI) in neuro-oncology research has been discussed,^28,29 and BLI has proven particularly useful for studying small-animal models of brain tumors.^4,10,30,31

One important factor in the success of small-animal preclinical studies is the selection of cohorts of animals with similar tumor growth patterns. Because of its ability to detect growing tumors early following implantation, BLI can be a valuable tool for this animal selection. MRI can also play an important role as solid tumors of 1 to 2 mm in diameter can be detected and their growth measured quantitatively. Although a number of studies have suggested a strong correlation between bioluminescence measurements and magnetic resonance–based calculations of tumor volumes,^3,4,30,32 only a small number of studies comparing BLI and MRI for characterizing tumor growth and response to therapy have been described.^3,4

A second important factor in the success of small-animal preclinical studies is having a surrogate, in vivo, marker for both tumor growth and for tumor regression in response to successful therapeutic interventions. Many investigators have suggested that BLI would be useful for testing investigational therapeutic agents. However, this use is dependent on confirming that stabilization or regression of the BLI signal corresponds to a stabilization or diminution of tumor size.

In our laboratory, we are particularly interested in the diagnosis, staging, and treatment of high-grade brain tumors and have significant experience with malignant murine brain tumors using the DBT-FG-cell model. Although not a perfect model of human glioblastoma, the DBT-FG-cell model has imaging, growth, and histopathologic features that are similar to high-grade gliomas,¹ shows some degree of cellular infiltration, and displays genetic changes (eg, loss of p53 and PTEN) similar to those seen in human tumors. In our experience with this model, we have observed specific instances in which BLI data did not directly correlate with corresponding anatomic MRIs. In this work, we describe a comprehensive, combined MRI-BLI study of tumor growth in the DBT-FG-cell model. Based on this study, we identify cautions that must be exercised in the interpretation of BLI or MRI data alone in the context of tumor growth and therapeutic response and describe a cost-effective, multimodality imaging strategy for selecting cohorts of animals with similar tumor growth patterns.

Materials and Methods

Luciferase-Expressing Tumor Cell Line

The DBT glioblastoma cell line was originally generated from an intracranial tumor induced in an adult mouse after intracerebral injection of a Rous sarcoma virus. The DBT cell line used in these experiments was generously donated by Dr. Michael Lai, Department of Microbiology, University of Southern California. Transduction and maintenance of DBT-FG cells, a derivative tumor cell line transduced with a lentivirus for stable expression of both firefly luciferase and green fluorescent protein (GFP), have been described.^33,34 Tumor cells were passaged for up to 8 weeks, after which fresh cells were reinnoculated from frozen stock into media and cultured to minimize loss of the luciferase signal.

Intracranial Tumor Injections

Preparation of Cells for Injection

Feeding medium was aspirated from culture flasks containing DBT-FG cells grown to approximately 70% confluence. Cells were washed with Dulbecco's Modified Eagle's Medium (DMEM), and then 3 mL of 0.5% trypsin–ethylenediaminetetraacetic acid was added over 1 minute. After the cells were suspended, trypsin was inactivated by adding 7 mL of DMEM containing 10% fetal calf serum. The suspended cells were centrifuged at 250 g for 10 minutes at 4°C and resuspended in 2 mL of DMEM. An aliquot of cells was aspirated, stained with trypan blue dye, and counted with a hemocytometer. The cells were recentrifuged and suspended in serum-free DMEM to a concentration of 10,000 DBT-FG cells/μL.

Animal Surgery

Adult female BALB/c mice weighing approximately 25 g were used in this study. All animal protocols were approved by the Washington University Division of Comparative Medicine and met or exceeded American Association for the Accreditation of Laboratory Animal Care International and National Institutes of Health standards. The mice were anesthetized by intraperitoneal injection of 25 mg/kg ketamine, 5 mg/kg xylazine, and 2.5 mg/kg acepromazine prior to intracranial implantation of DBT-FG cells. After a midline scalp incision, a craniostomy was created over the cortex with a 1 mm cutting burr. Mice were secured in a stereotactic frame, and 5 μL of the DBT-FG tumor cell suspension (≈50,000 cells) was aspirated into a Hamilton syringe attached to the frame, which was then passed into the brain. Tumor cells were injected into the striatum at a site 2 mm lateral and 2 mm posterior to the bregma and 4 mm deep to the cortical surface. After the needle was removed, the craniostomy was sealed with bone wax and the scalp was closed with superglue. Following surgery, mice were housed in nonbarrier animal facilities for the duration of the imaging experiments. Although animals with large brain tumors were more likely to die under anesthesia than healthy animals, mice were successfully and routinely imaged at multiple time points with both MRI and BLI throughout this study.

In Vivo BLI

Mice were injected intraperitoneally with 150 mg/kg of D-luciferin in phosphate-buffered saline 10 minutes prior to imaging. Imaging was performed using a charge-coupled device camera (IVIS 100, Xenogen-Caliper, Alameda, CA; exposure times 1–120 s, binning 2–8, field of view [FOV] 15 cm, f/stop 1, no filter). Mice were anesthetized with isoflurane (2% vaporized in O₂) prior to and during imaging. Total photon flux (photons/s) was measured from a fixed region of interest over the skull using Living Image and IgorPro software (Wavemetrics, Portland, OR) as previously described.³⁵

Magnetic Resonance Imaging

Images were collected in an Oxford Instruments (Oxford, UK) 4.7-Tesla magnet (33 cm, clear bore) equipped with 15 cm inner diameter, actively shielded gradient coils (maximum gradient 18 G/cm; rise time 200 μs). The magnet/gradients are interfaced with a Varian (Palo Alto, CA) INOVA console, and data were collected using a 1.5 cm OD surface coil (receive) and a 9 cm ID Helmholtz coil (transmit). Before the imaging experiments, mice were anesthetized with isoflurane/O₂ (4% [v/v]) and maintained on isoflurane/O₂ (1.5% [v/v]) throughout the experiments.

Mice were injected intraperitoneally with 500 μL Omniscan (Gadodiamide, GE Healthcare, Princeton, NJ) contrast agent, diluted 1:10 in sterile saline, 15 minutes prior to being placed in the magnet. T₁-weighted, gradient-echo multislice images (coronal and transaxial views) were collected (repetition time [TR] = 0.125 s, echo time [TE] = 0.0025 s, FOV [transaxial] = 1.5 × 1.5 cm², FOV [coronal] = 3 × 3 cm², slice thickness = 0.5 mm). In addition, T₂-weighted, multislice spin-echo transaxial images were collected for each animal (TR = 1.5 s; TE = 0.05 s; FOV = 1.5 × 1.5 cm²; slice thickness = 0.5 mm).

Quantitative Assessment of Tumor Volume (MRI)

For MRI studies, tumor volumes were measured by manually segmenting tumors using either Varian's ImageBrowser software or the public domain program ImageJ (<http://rsbweb.nih.gov/ij>). T₂-weighted images were sometimes used to help clarify tumor margins. Anatomic abnormalities such as hemorrhage, hydrocephalus, or tumor necrosis were noted, and their presence was correlated with optical signals.

Histologic Analysis

Following the last day of imaging, mice were heavily sedated with isoflurane and perfused with 4% paraformaldehyde. Brains were removed from the mice and placed in 4% paraformaldehyde for 24 hours at 4°C. The brains were next placed in 70% alcohol for 48 hours and then cut into 2 mm sections for paraffin fixation. The paraffin embedded specimens were cut, again in the coronal plane, into 50 μm sections. These fixed specimens were stained with hematoxylin and eosin and were correlated with corresponding MRIs.

Results and Discussion

MRI is effective for measuring tumor growth, and therapeutic response and calculations of intracranial tumor volumes from MRIs correlate well with postmortem tumor volumes.^11,36 However, MRI is not well suited to high-throughput studies and cannot detect tumor growth early after implantation. For volumetric assays, BLI offers a valuable method for identifying tumors early after implantation,³⁷ monitoring tumor growth,³⁴ and measuring response to intervention.^38,39 Recently, the use of BLI for monitoring the effects of antivascular therapy in a breast cancer xenograft model, emphasizing the importance of luciferin delivery to tumor, was described.⁴⁰ Previous studies have also documented the linearity of BLI signal measured in vivo with respect to luciferase activity measured ex vivo.^34,41 Compared with MRI or other methods for assessing tumor growth in vivo, BLI experiments are quick and easy to perform and offer a relatively inexpensive and efficient method for high-throughput studies.

In this study, four groups of 10 mice each were implanted with 50,000 DBT-FG cells constitutively expressing firefly luciferase and then serially imaged using both MRI and BLI. The first objective of the study, using the first 20 mice, was to establish a correlation between bioluminescence signal and tumor volume using BLI and MRI over the time course of tumor growth. The second objective, using the second 20 mice, was to determine whether early bioluminescence signal accurately predicted subsequent (exponential) tumor growth and whether, conversely, those tumors showing exponential growth on MRI had longitudinal BLI signals that corresponded well to the magnetic resonance growth curves.

Although the majority of the mice showed excellent correlation between tumor volumes as measured by MRI and quantitative BLI signal, particularly at the early stages of tumor growth (Figure 1), we found that at the extremes (very low BLI signal or very large tumor volumes), this correlation was much weaker. We identified two particular situations in which mice without MRI-visible tumors had a strong BLI signal: mice with hydrocephalus (Figure 2) and mice with intracranial hemorrhage (Figure 3). Significant tumor was not visible on MRIs of any of these animals. Mice (n = 2) with hydrocephalus survived to 13 and 14 days post–tumor injection and had BLI growth curves consistent with exponential tumor growth. These mice deteriorated clinically in a similar fashion to animals with growing tumor. One mouse with an intracranial hematoma failed to deteriorate clinically and displayed a BLI signal that initially increased exponentially in intensity and declined in the late stages of the experiment.

We noticed an additional discrepancy between BLI signal and MRI measurement of tumor growth. At late stages of growth in animals with a growing solid tumor, MRI measurements showed steadily increasing volumes, whereas the optical signal tended to plateau. A similar effect, with in vivo BLI signal plateauing in large tumors, was observed in a study of mouse liver.⁴² This late plateau is seen in the correlations between BLI signal and tumor volumes in Figure 1. A representative example of this discordance is demonstrated longitudinally and anatomically in Figure 4. This change is attributed to development of hemorrhage and necrosis within the tumor bed, a finding that was seen on both anatomic MRIs (Figure 4, C and D) and in histologic specimens (Figure 5). Thus, BLI may reflect viable cell mass,^34,41,43,44 whereas MRI reflects overall tumor volume.

Figure 1.

Bioluminescence imaging (BLI) optical signal intensities are plotted versus their corresponding magnetic resonance–determined tumor volumes (in murine brain tumors generated by injection of luciferase-bearing DBT-FG cells). While in the early stages of tumor growth, when tumors are less than 40 mm³, there is a strong correlation between tumor volume and BLI signal (R² = .78); a significant loss of this correlation is seen once the tumors grow beyond 40 mm³ (R² = .13).

Figure 2.

Left, Single slice from a series of transaxial, T₂-weighted spin-echo magnetic resonance images of a mouse injected with DBT-FG tumor cells. The mouse has a high bioluminescence signal (2.3 × 10⁸ photons/s) but no significant solid tumor burden. Rather, this mouse displays extensive hydrocephalus, and the high bioluminescence imaging (BLI) signal is likely due to the presence of tumor cells in the cerebrospinal fluid. Right, BLI from this same animal showing a strong bioluminescence signal, indicative of viable, growing tumor cells. The signal is visualized diffusely over the brain, which may be indicative of the absence of a focal tumor source.

One of the major strengths of preclinical studies in animals is the ability to perfuse organs and perform subsequent histology and immunohistochemistry (IHC) staining. Such correlative histology/IHC provides insight into the cellular changes occurring within tumor tissue and can help validate the results of imaging studies. Figure 5 illustrates examples of hematoxylin-eosin stains from brains imaged in these studies. Figure 5A shows tissue from a region of brain tumor; the tissue is characterized by high cellular density and several unusually large nuclei. This highly cellular tumor, with numerous, densely packed viable cells, is the setting where BLI signal would likely correlate well with magnetic resonance volume. The presence of intraventricular tumor is illustrated in Figure 5B. Growth of intraventricular tumor, either within the cerebrospinal fluid as isolated cells or in small clusters causing ventricular obstruction and hydrocephalus, offers an anatomic and biologic explanation for the high BLI signal in the absence of intraparenchymal tumor seen in Figure 2. Figure 5C shows tissue from a hemorrhagic region of brain, whereas necrotic tumor tissue is illustrated in Figure 5D. Each of these changes occurs late in the time course of tumor growth and may show focally enhanced absorption of BLI signal or indicate a loss of viable tumor cells.

Figure 3.

Left, Single slice from a series of transaxial, T₂-weighted spin-echo magnetic resonance images of a mouse injected with DBT-FG tumor cells. The mouse has a high optical signal (5.3 × 10⁷ photons/s) within the intraparenchymal hemorrhage (dark mass in the upper right portion of the brain) at the site of injection but no visible intracranial tumor. Right, Bioluminescence imaging from this same animal, showing a strong bioluminescence signal, indicative of viable, growing tumor cells.

In a combined BLI-MRI study using another cohort of 20 mice, we tested the hypothesis that the presence and growth of BLI signals at early time points would be predictive of subsequent exponential tumor growth and that those tumors that grew in a standard exponential pattern would all have BLI signals that longitudinally matched that growth. Serial BLI and MRI experiments were performed on all mice. After tumors had grown to the maximally tolerated size, animals whose magnetic resonance–detected tumor growth was exponential were identified. The optical signals for each of these animals were reviewed retrospectively. We had expected that mice with exponential tumor growth by MRI would have optical growth curves similar to one another. However, we found a few examples of spontaneous diminution of luciferase signal (a finding noted in previous studies of luciferase tumor models⁴⁵) that confounded our ability to reliably predict tumor growth from optical signal curves alone (Figure 6). Although the number of animals with spontaneous regression of BLI signal was small (n = 2), if this study had been a preclinical therapeutic response trial, these two animals, with progressive exponential growth in tumor size, would have, by BLI alone, been described as “responders.” This observation demonstrates the value of using BLI and MRI in combination to best evaluate tumor regression in preclinical therapeutic trials.

Our successes in using combined BLI and MRI to measure tumor growth, along with the complementary information available with each imaging method, lead us to suggest a strategy for combining these two imaging modalities to identify and select animals for preclinical therapeutic studies. This multimodality strategy, designed to identify animals with similar tumor growth rates while limiting the impact of confounding factors such as hemorrhage, necrosis, and loss of luciferase signal, is summarized in the flowchart in Figure 7. Methodologically, we propose beginning by injecting a group of animals larger than the intended study sample size with a luciferase-bearing tumor cell line (eg, for an intended sample size of 6 to 8 mice, we typically inject 10 mice). Animals are assessed shortly after implantation and serially using BLI semiweekly or at appropriate time points based on the anticipated growth rate of the chosen tumor cell line. Regular measurement of BLI signal permits us to match photon outputs across mice, thereby allowing us to select cohorts of animals with similar amounts of viable tumor cell mass. In addition, all animals with a visible BLI signal undergo standard anatomic MRI early in the course of the study to identify confounding anatomic features such as hemorrhage or hydrocephalus. Following multiple BLIs and the single MRI, animals with abnormal tumor growth patterns or unusual anatomic findings are excluded from subsequent preclinical experiments. In our experience, approximately 20% of animals initially injected with tumor cells would be eliminated from later preclinical studies by the strategy outlined in Figure 7.

Figure 4.

A, Exponential growth of a DBT-FG cell tumor as measured by magnetic resonance volumetry. B, Corresponding optical signal, measured on a logarithmic scale, for the same animal. The optical signal plateaus following the day 15 measurement and drops modestly at later time points. The imaging time points at postimplantation days 15 and 19 correlate to the magnetic resonance images shown in Figure 5, C and D. C, Single slice from a series of transaxial, T₁-weighted, gadolinium-enhanced, gradient echo images of a mouse injected with DBT-FG tumor cells at postoperative day 15. Note the significant enhancing solid tumor in the right hemisphere. D, Single slice from a series of transaxial, T₁-weighted, gadolinium-enhanced, gradient echo images of the same mouse at postoperative day 19. The enhancing solid tumor has increased in size significantly and is now accompanied by significant hemorrhage and necrotic material.

Figure 5.

Hematoxylin-eosin–stained slides of sections from mouse brain tumors: (A) tumor tissue (×40), showing high cellular density and several abnormally large nuclei; (B) intraventricular tumor (×10), the presence of which can obstruct the ventricle, thereby leading to the development of hydrocephalus within the brain; (C) hemorrhagic tumor tissue (×10); (D) necrotic tumor tissue (×40).

Figure 6.

Bioluminescence imaging signal from a longitudinal study of 10 mice, all of whom, based on serial anatomic magnetic resonance imaging studies, showed exponential growth of solid tumor without confounding hemorrhage or hydrocephalus. Each symbol represents data for one of the ten mice in this study. Two outliers, whose tumor growth was significant but whose optical signal declined early in the time course of growth, are noted.

The timing of the steps suggested in Figure 7 will vary depending on the growth rate of the tumor cell line, the number of cells injected, and the elapsed time following injection of cells prior to initiation of treatment. In our previously published imaging study of carmustine chemotherapy in DBT-FG-cell tumors, tumors typically had grown to a volume of 20 to 30 mm³ at the time that treatment was initiated.¹ The combination of anatomic MRI and BLI summarized in Figure 7 provides a multidisciplinary approach for selecting a subset of animals with similar patterns of tumor growth for subsequent imaging or therapeutic studies while eliminating the confounding effects of hemorrhage or hydrocephalus. Although we performed this study using the DBT-FG-cell tumor model, the methodology can be readily extended to include other glioma cell lines (eg, U87 cells) that can be modified to stably express luciferase.³⁴

Although performing the MRI adds a step to these studies, we have found that this additional set of scans does not significantly complicate the overall logistics of these experiments, nor, in centers where small-animal imaging is available, does it add dramatically to the cost. Indeed, by enabling efficient selection of cohorts of animals with similar tumor growth that are free of confounding anatomic features, we predict that this multimodal approach will ultimately decrease both the number of animals and the number of scans needed to obtain meaningful results. Further, we suggest that adding periodic MRIs to longitudinal BLI at later stages of therapeutic studies will increase the role of imaging in assessing treatment efficacy by allowing measurement of both remaining viable tumor cell mass (BLI) and residual tumor volume/mass, including intratumoral regions of hemorrhage and necrosis (MRI). The two imaging modalities have the capacity to function synergistically, and each can offset the potential challenges of the other: BLI can identify early tumor growth and improve efficiency, and MRI can identify anatomic anomalies (hemorrhage, hydrocephalus, necrosis) and catch the rare instances of spontaneous BLI diminution. The combination of the two modalities will further increase the likelihood that useful information will be gleaned from imaging-based preclinical studies, allowing efficient translation to stage 1 human clinical trials.

Figure 7.

Schematic representation of proposed strategy for combining bioluminescent optical imaging (BLI) and anatomic magnetic resonance (MR) imaging to select cohorts of animals for preclinical trials.

Footnotes

Acknowledgments

We thank Dr. Neha Dahiya for assistance in reviewing the histology results.

Financial disclosure of authors: This work was supported by a National Institutes of Health/National Cancer Institute (NIH/NCI) Small Animal Imaging Resource Program grant (U24 CA83060), an NIH/NCI In vivo Cellular and Molecular Imaging Center grant (P50 CA094056), and the Alvin J. Siteman Cancer Center at Washington University in St. Louis, an NCI Comprehensive Cancer Center (P30 CA91842).

Financial disclosure of reviewers: None reported.

References

Jost

Wanebo

Song

. Small animal imaging as a tool for in vivo monitoring in a murine model for glioblastoma. Neurosurgery 2007;60:360–70.

Nelson

Algon

Munasinghe

. Magnetic resonance imaging of patched heterozygous and xenografted mouse brain tumors. J Neurooncol 2003;62:259–67.

Rhemtulla

Stegman

Cardozo

. Rapid and quantitative assessment of cancer treatment response using in vivo bioluminescence imaging. Neoplasia 2000;2:491–5.

Szentirmai

Baker

Lin

. Noninvasive bioluminescence imaging of luciferase expressing intracranial U87 xenografts: correlation with magnetic resonance imaging determined tumor volume and longitudinal use in assessing tumor growth and antiangiogenic treatment effect. Neurosurgery 2006;58:365–72.

Kwon

Zhao

Chen

. Pten haploinsufficiency accelerates formation of high-grade astrocytomas. Cancer Res 2008;68:3286–94.

McConville

Hambardzumyan

Moody

. Magnetic resonance imaging determination of tumor grade and early response to temozolomide in a genetically engineered mouse model of glioma. Clin Cancer Res 2007;13:2897–904.

Zhu

Guignard

Zhao

. Early inactivation of p53 tumor suppressor gene cooperating with NF1 loss induces malignant astrocytoma. Cancer Cell 2005;8:119–30.

Adzamli

Yablonskiy

Chicoine

. Albumin-binding MR blood pool agents as MRI contrast agents in an intracranial mouse glioma model. Magn Reson Med 2003;49:586–90.

Bock

Zadeh

Davidson

. High-resolution longitudinal screening with magnetic resonance imaging on a murine brain cancer model. Neoplasia 2003;5:546–54.

10.

Rustamzadeh

Hall

Todhunter

. Intracranial therapy of glioblastoma with the fusion protein DTIL13 in immunodeficient mice. Int J Cancer 2006;118:2594–601.

11.

Schmidt

Ziu

Schmidt

. Volume reconstruction techniques improve the correlation between histological and in vivo tumor volume measurements in mouse models of human gliomas. J Neurooncol 2004;68:207–15.

12.

Sherburn

Wanebo

Kim

. Gliomas in rodent whisker barrel cortex: a new tumor model. J Neurosurg 1999;91:814–21.

13.

Chenevert

Stegman

Taylor

. Diffusion MRI: early surrogate marker of therapeutic efficacy. J Natl Cancer Inst 2000;92:2029–36.

14.

Hall

Moffat

Stojanovska

. Therapeutic efficacy of DTI-015 using diffusion magnetic resonance imaging as an early surrogate marker. Clin Cancer Res 2004;10:7852–9.

15.

Moffat

Chenevert

Lawrence

. Functional diffusion map: a noninvasive MRI biomarker for stratification of clinical brain tumor response. Proc Natl Acad Sci U S A 2005;102:5524–9.

16.

Yablonskiy

Bretthorst

Ackerman

JJH

. Statistical model for diffusion attenuated MR signal. Magn Reson Med 2003;50:664–9.

17.

Ah-See

Makris

Taylor

. Early changes in functional dynamic magnetic resonance imaging predict for pathologic response to neoadjuvant chemotherapy in primary breast cancer. Clin Cancer Res 2008;14:6580–9.

18.

Barrett

Brechbiel

Bernardo

. MRI of tumor angiogenesis. J Magn Reson Imaging 2007;26:235–49.

19.

Huang

Morris

. The magnetic resonance shutter speed discriminates vascular properties of malignant and benign breast tumors in vivo. Proc Natl Acad Sci U S A 2008;105:17943–8.

20.

Jackson

O'Connor

Parker

. Imaging tumor vascular heterogeneity and angiogenesis using dynamic contrast-enhanced magnetic resonance imaging. Clin Cancer Res 2007;13:3449–59.

21.

Knopp

Giesel

Marcos

. Dynamic contrast-enhanced magnetic resonance imaging in oncology. Top Magn Reson Imaging 2001;12:301–8.

22.

Bernsen

HJJA

Heerschap

Kogel

. Image-guided 1H NMR spectroscopical and histological characterization of a human brain tumor model in the nude rat; a new approach to monitor changes in tumor metabolism. J Neurooncol 1992;13:119–30.

23.

Howe

Barton

Cudlip

. Metabolic profiles of human brain tumors using quantitative in vivo 1H magnetic resonance spectroscopy. Magn Reson Med 2003;49:223–32.

24.

Contag

Jenkins

Contag

. Use of reporter genes for optical measurements of neoplastic disease in vivo. Neoplasia 2000;2:41–51.

25.

Edinger

Cao

Hornig

. Advancing animal models of neoplasia through in vivo bioluminescence imaging. Eur J Cancer 2002;38:2128–36.

26.

Jenkins

Hornig

Oei

. Bioluminescent human breast cancer cell lines that permit rapid and sensitive in vivo detection of mammary tumors and multiple metastases in immune deficient mice. Breast Cancer Res 2005;7:R444–54.

27.

Gross

Piwnica-Worms

. Spying on cancer: molecular imaging in vivo with genetically encoded reporters. Cancer Cell 2005;7:5–15.

28.

Paroo

Bollinger

Braasch

. Validating bioluminescence imaging as a high throughput, quantitative modality for assessing tumor burden. Mol Imaging 2004;3:117–24.

29.

Soling

Rainov

. Biouminescence imaging in vivo: application to cancer research. Expert Opin Biol Ther 2003;3:1163–72.

30.

Rubin

Kung

Klein

. A small-molecule antagonist of CXCR4 inhibits intracranial growth of primary brain tumors. Proc Natl Acad Sci U S A 2003;100:13513–8.

31.

Soling

Theiss

Jungmichel

. A dual function fusion protein of herpes simplex virus type 1 thymidine kinase and firefly luciferase for noninvasive in vivo imaging of gene therapy in malignant glioma. Genet Vaccines Ther 2004;2:7.

32.

Dinca

Sarkaria

Schroeder

. Bioluminescence monitoring of intracranial glioblastoma xenograft: response to primary and salvage temozolomide therapy. J Neurosurg 2007;107:610–6.

33.

Yuan

Siegel

Choi

. Transglutaminase 2 inhibitor, KCC009, disrupts fibronectin assembly in the extracellular matrix and sensitizes orthotopic glioblastomas to chemotherapy. Oncogene 2007;26:2563–73.

34.

Goldhoff

Warrington

Limbrick

Jr . Targeted inhibition of cyclic AMP phosphodiesterase-4 promotes brain tumor regression. Clin Cancer Res 2008;14:7717–25.

35.

Gross

Piwnica-Worms

. Monitoring proteasome activity in cellulo and in living animals by bioluminescent imaging: technical considerations for design and use of genetically encoded reporters. Methods Enzymol 2005;399:512–30.

36.

Cornelissen

Kersemans

Staelens

. Comparison between 1 T MRI and non-MRI based volumetry in inoculated tumours in mice. Br J Radiol 2005;78:338–42.

37.

Edinger

Sweeney

Tucker

. Non-invasive assessment of tumor proliferation in animal models. Neoplasia 1999;1:303–10.

38.

Uhrbom

Nerio

Holland

. Dissecting tumor maintenance requirements using bioluminescence imaging of cell proliferation in a mouse glioma model. Nat Med 2004;10:1257–60.

39.

Vooijs

Jonkers

Lyons

. Noninvasive imaging of spontaneous retinoblastoma pathway-dependent tumors in mice. Cancer Res 2002;62:1862–7.

40.

Zhao

Richer

Antich

. Antivascular effects of combretastatin A4 phosphate in breast cancer xenograft assessed using dynamic bioluminescence imaging and confirmed by MRI. FASEB J 2008;22:2445–51.

41.

Contag

Bachmann

. Advances in in vivo bioluminescence imaging of gene expression. Annu Rev Biomed Eng 2002;4:235–60.

42.

Sarraf-Yazdi

Dewhirst

. Use of in vivo bioluminescence imaging to predict hepatic tumor burden in mice. J Surg Res 2004;120:249–55.

43.

Piwnica-Worms

Schuster

Garbow

. Molecular imaging of host-pathogen interactions in intact small animals. Cell Microbiol 2004;6:319–31.

44.

El Hilali

Rubio

Blanco

. Different effect of paclitaxel on primary tumor mass, tumor cell contents, and metastases for four experimental human prostate tumors expressing luciferase. Clin Cancer Res 2005;11:1253–8.

45.

Brutkiewicz

Mendonca

Stantz

. The expression level of luciferase within tumour cells can alter tumour growth upon in vivo bioluminescence imaging. Luminescence 2007;22:221–8.

Measuring Brain Tumor Growth: Combined Bioluminescence Imaging–Magnetic Resonance Imaging Strategy

Abstract

Materials and Methods

Luciferase-Expressing Tumor Cell Line

Intracranial Tumor Injections

Preparation of Cells for Injection

Animal Surgery

In Vivo BLI

Magnetic Resonance Imaging

Quantitative Assessment of Tumor Volume (MRI)

Histologic Analysis

Results and Discussion

Footnotes

Acknowledgments

References