Abstract
The αv integrins, which act as cell adhesion molecules, are closely involved with tumor invasion and angiogenesis. In particular, αvβ3 integrin, which is specifically expressed on proliferating endothelial cells and tumor cells, is a logical target for development of a radiotracer method to assess angiogenesis and anti-angiogenic therapy. In this study, a dimeric cyclic RGD peptide E[c(RGDyK)]2 was labeled with 18F (t1/2 = 109.7 min) by using a prosthetic 4-[18F]fluorobenzoyl moiety to the amino group of the glutamate. The resulting [18F]FB-E[c(RGDyK)]2, with high specific activity (200–250 GBq/μmol at the end of synthesis), was administered to subcutaneous U87MG glioblastoma xenograft models for micro-PET and autoradiographic imaging as well as direct tissue sampling to assess tumor targeting efficacy and in vivo kinetics of this PET tracer. The dimeric RGD peptide demonstrated significantly higher tumor uptake and prolonged tumor retention in comparison with a monomeric RGD peptide analog [18F]FB-c(RGDyK). The dimeric RGD peptide had predominant renal excretion, whereas the monomeric analog was excreted primarily through the biliary route. Micro-PET imaging 1 hr after injection of the dimeric RGD peptide exhibited tumor to contralateral background ratio of 9.5 ± 0.8. The synergistic effect of polyvalency and improved pharmacokinetics may be responsible for the superior imaging characteristics of [18F]FB-E[c(RGDyK)]2.
Introduction
Tumor angiogenesis, the formation of new blood vessels or the growth of blood vessels between a tumor and surrounding tissue, is a complex process regulated through a precise balance of pro-angiogenic and anti-angiogenic molecules, which involves complex interactions between extracellular matrix molecules, proteolytic enzymes, and cell adhesion molecules on endothelial cells (ECs) [1,2]. As most solid tumors are angiogenesis dependent, the ability to develop molecular imaging methods for evaluating angiogenesis is of vital importance to understand the complex process and to monitor target specific anti-angiogenic treatment efficacy. Inhibition of the activity of any of the key regulators in the angiogenic cascade is expected to block tumor angiogenesis. Integrin adhesion receptors regulate cell survival, proliferation, motility, and differentiation through their ability to transduce signals into and out of the cells and to engage in reciprocal interactions with other cellular receptors [3]. The cell surface adhesion and signaling receptor αvβ3 integrin, which is highly expressed in tumor cells of various origin and activated ECs in growing tumors but not on normal cells and quiescent ECs, has been identified as a marker for angiogenic vascular tissue and potentially can be used as a target for diagnostic and therapeutic agents aimed at tumor angiogenesis [4].
Two αv-integrin antagonists—vitaxin, a humanized antibody [5], and cilengitide [6], a cyclic peptide mimicking the RGD ligand recognition peptidic domain common to αv-integrin ligands—are in phase II clinical trials [4]. Due to their better pharmacokinetics and oral bioavailability, small synthetic heterocyclic inhibitors are also being developed as successors to peptide-derived agents [7]. Other strategies include radiolabeled cyclic RGD peptides for internal radiotherapy [8], and combination of integrin antagonists and radioimmunotherapy for enhanced therapeutic synergy [9]. The ability to rapidly and accurately detect tumor growth and metastasis is of great importance for implementation of tailored therapeutic regimens [10]. We and others have recently labeled cyclic RGD peptides with 18F [11–17] and 64Cu [14] for positron emission tomography (PET) imaging of αvβ3-integrin expression in murine tumor models.
The conjugation labeling of small cyclic RGD peptides with the prosthetic labeling group 4-[18F]fluorobenzoyl significantly increased lipophilicity of the peptide and led to unfavorable hepatobiliary excretion and rapid tumor washout of the resulting PET tracer [11,12,14]. The clinical potential of this type of radiotracer is thus limited due to unfavorable biodistribution, including very high activity accumulation in the lower abdomen [11,12,14]. Haubner et al. [15,16] inserted a sugar moiety between the 2-[18F]fluoropropionate moiety and the ε-amino group of the cyclic RGD peptide c(RGDfK) lysine residue. The increased hydrophilicity upon glycosylation, rapid excretion, adequate metabolic stability, and the low estimated radiation dose of this radiopharmaceutical endows it with potential for application in human studies. We applied a different approach to improve the in vivo pharmaokinetic profile of 18F-labeled RGD peptide through PEGylation [13]. The 18F-labeled PEG-RGD peptide has significantly improved tumor retention relative to [18F]FB-RGD without compromising the desirably rapid clearance of radioactivity from liver and kidneys. Additionally, decreased biliary excretion minimized intestinal retention of the activity and increased tumor-to-nontumor ratios [13].
Specific recognition of a peptide by a given receptor depends on the specific configuration of the peptide side-chain with regard to its ability to fit in the receptor binding sites. A multimeric peptide with two or more monomeric units connected via semi-rigid molecular linkers may achieve improved binding avidity and specificity through multiple, weakly cooperative interactions [18]. It was recently reported that dimeric and multimeric cyclic RGD peptides have higher receptor binding affinity in vitro [19] and better tumor retention in vivo [19,20]. This is presumably due to polyvalency, which gives rise to an enhanced binding and steric stabilization [21]. The aim of the study presented here was to compare the tumor targeting and in vivo kinetics of the 18F-labeled dimeric RGD peptide [18F]FB-E[c(RGDyK)]2 to that of its monomeric analog, [18F]FB-c(RGDyK).
Materials and Methods
Materials
All reagents, unless otherwise specified, were of analytical grade and commercially obtained. Cyclic RGD peptide c(RGDyK) was synthesized via solution cyclization of the fully protected linear pentapeptide H-Gly-Asp(OtBu)-
[19F]FB-E[c(RGDyK)]2
As a reference standard, [19F]FB-E[c(RGDyK)]2 containing the 4-[19F]fluorobenzoyl moiety was synthesized by coupling E[c(RGDyK)]2 with N-succinimidyl 4-fluorobenzoate (SFB) under slightly basic condition. In brief, SFB 2.4 mg (10 μmol) in 500 μL acetonitrile was added to E[c(RGDyK)]2 2.7 mg (2 μmol) dissolved in Na2HPO4 buffer 1500 μL (pH = 8.7) and allowed to stand at room temperature for 2 hr. The reaction was quenched by adding 50 μL TFA and followed by semi-preparative HPLC purification. Matrix-assisted laser desorption/ionization–time of flight mass spectrometry (MALDI-TOF) of [19F]FB-E[c(RGDyK)]2: m/z = 1472.9 for [M + H]+ (C66H90FN19O19); the retention time was 17.3 min from analytical HPLC; the purity was 99%, and the yield was 76%.
Radiolabeling E[c(RGDyK)]2 with 18F
E[c(RGDyK)]2 was labeled with 18F by conjugation coupling with N-succinimidyl 4-[18F]fluorobenzoate ([18F]SFB). [18F]SFB with a specific activity of 200–250 GBq/μmol [14] was dissolved in acetonitrile (1 mL) and added to E[c(RGDyK)]2 (200 μg, 0.15 μmol) dissolved in Na2HPO4 buffer (1 mL, pH = 8.7). The reaction was continued for 30 min at 45°C until most of SFB had reacted according to radio-TLC (eluent CH2Cl2/EtOAc = 4:1). Final purification was accomplished by C18 reversed-phase chromatography (detection modes: radioactivity and UV at 218 nm). HPLC fractions containing the activity were combined and evaporated with a stream of argon to remove acetonitrile. The pH of the resultant aqueous solution was adjusted to 7.0 with 0.1 N NaOH. This solution was passed through an activated Waters C18 Sep-Pak cartridge, washed with water and eluted with 200 μL portions of 80% ethanol. The ethanol fractions were pooled and evaporated to a small volume. The activity was reconstituted in phosphate-buffered saline and passed through a 0.22-μm Millipore filter into a sterile multidose vial for use in animal experiments.
Receptor Binding Assay
Primary human brain capillary endothelial cells (HBCECs) were isolated, characterized, and grown in RPMI-1640 medium with 10% FBS in 5% CO2 at 37°C as previously described [24]; the cells were used between passages 5 and 14. 125I-c(RGDyK) (specific activity 1700 Ci/mmol) was prepared via the chloramine-T method according to a reported procedure [21]. For competitive integrin receptor binding assay studies, Costar 24-well plates were seeded with 2 × 105 BCEC cells/well for 2 hr and then rinsed twice with binding buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 2 mM CaCl2, 1 mM MgCl2, 1 mM MnCl2, and 0.1% bovine serum albumin) [24]. The plate was incubated with 125I-c(RGDyK) in the presence of increasing concentrations of different RGD peptide analogues (0–1000 nM). The total incubation volume was adjusted to 400 μL. After the cells were incubated for 2 hr at 4°C, the supernatant was removed, and the cells were washed twice with cold binding buffer and then solubilized with 0.5% SDS. Radioactivity was determined using a NaI(Tl) gamma counter (Packard, Meriden, CT). The IC50 values were calculated by fitting the data by nonlinear regression using GraphPad Prism (GraphPad Software, San Diego, CA). Experiments were carried out with triplicate samples.
Tumor Models
Animal experiments were conducted under a protocol approved by the USC Institutional Animal Care and Use Committee (IACUC). Female athymic nude mice (nu/nu) obtained from Harlan (Indianapolis, IN) at 4–5 weeks of age were injected subcutaneously in the right front leg with 2 × 106 U87MG glioblastoma cells suspended in 150 μL Eagle's minimum essential medium (EMEM). When the tumors reached 0.4–0.6 cm in diameter (14–18 days after implant), the mice were used for biodistribution and micro-PET imaging experiments.
Biodistribution
Nude mice bearing subcutaneously xenografted human glioblastoma U87MG tumors were injected intravenously with approximately 740 kBq (20 μCi) of [18F] FB-E[c(RGDyK)]2. Animals were euthanized at 30 min, 1 hr, 2 hr, and 4 hr postinjection. Blood, tumor, and the major organs and tissues were collected, wet-weighed, and counted in a gamma-counter (Packard). The percent-injected dose per gram (%ID/g) was determined for each sample. For each mouse, radioactivity of the tissue samples was calibrated against a known aliquot of the injectate and normalized to body mass of 20 g. Values are quoted as mean ± standard deviation (SD) for a group of four animals. Receptor-mediated localization of the radiotracer was investigated by intravenous injection of 10 mg/kg of c(RGDyK) mixed with [18F]FB-E [c(RGDyK)]2 into four mice bearing the U87MG tumor. The mice in this group were sacrificed at 1 hr pi.
Micro-PET Imaging Studies
PET imaging was performed with a micro-PET R4 rodent model scanner (Concorde Microsystems, Knoxville, TN). The scanner has a computer-controlled bed, and 10.8 cm transaxial, 8 cm axial field of view (FOV), and an intrinsic resolution of 1.85 mm at the center of the FOV. It has no septa and operates exclusively in 3-D list mode. The raw list mode data were 3-D histogrammed without smoothing into single-frame sonogram files and reconstructed into images via the Ordered Subsets Expectation Maximization (OSEM) method. The spatial resolution obtained with this reconstruction method ranged from 1.9 to 2.3 mm. Regions of interest were drawn around areas of maximum tracer uptake. The mice were injected with 150 μCi of [18F]FB-E[c(RGDyK)]2 via the tail vein, anesthetized with ketamine/xylazine at 1 hr pi and then centered in the FOV of the micro-PET. Mice were scanned for 15 min, and the raw data were framed into one static frame without attenuation correction. The micro-PET was calibrated in terms of absolute activity concentration (nCi/cc) by imaging a plastic bottle approximating the dimensions of a mouse body and filled with a known concentration of 18F.
Whole-Body Autoradiography
Autoradiography was performed using a Packard Cyclone Storage Phosphor Screen system (Downers Grove, IL) and a Bright 5030/WD/MR cryomicrotome (Hacker Instruments, Fairfield, NJ). In order to correlate the results obtained from micro-PET and whole-body autoradiography, tumor-bearing mice were sacrificed under ketamine/xylazine anesthesia at 1 hr pi, followed by micro-PET imaging and subsequently cryosectioning for autoradiography. This procedure was followed in order to eliminate the possibility of change in tracer biodistribution between micro-PET and autoradiography. Immediately after micro-PET scanning, the euthanized mouse was frozen in dry ice and isopropyl alcohol bath for 2 min. The body was then embedded in 4% carboxymethyl cellulose (CMC) (Aldrich, Milwaukee, WI) in water mixture within a stainless steel mold. The mold was placed in the dry ice and isopropyl alcohol bath for 5 min and then into a −20°C freezer for 1 hr. The walls of the mold were then removed, and the frozen block was mounted in the cryomicrotome. The block was cut into 50 μM sections, and desired sections were digitally photographed and then captured for autoradiography. The sections were transferred into a chilled film cassette containing a Super Resolution screen (spatial resolution 0.1 mm; Packard) and kept there overnight at −20°C. Screens were laser-scanned with the Packard Cyclone.
Statistical Analysis
The data are expressed as mean ± SD. Means were compared using one-way analysis of variance (ANOVA) and Student's t test. A p value of < .05 was considered significant.
Results
Radiosynthesis
Labeling of dimeric RGD peptide E[c(RGDyK)]2 using [18F]SFB as a prosthetic group resulted in [18F]FB-E[c(RGDyK)]2 (Figure 1) of high radiochemical purity (> 99%) according to analytical HPLC. The decay- corrected yield of [18F]FB-PEG-RGD, counted from [18F] SFB, ranged from 20% to 30% (in three preparations). The specific radioactivity of [18F]SFB was estimated by radio-HPLC to be 200–250 GBq/μmol at the end of synthesis (EOS) (or 300–400 GBq/μmol at end of bombardment [EOB]), based upon 9.25 GBq (250 mCi) of starting [18F]F−. For micro-PET studies, the injected dose of 5.55 MBq (150 μCi) contained approximately 30–40 ng of [18F]FB-E[c(RGDyK)]2. For biodistribution studies, the injected dose of 0.74 MBq (20 μCi) was composed of 4–5 ng of radioactive 18F-labeled RGD dimer.
Receptor Binding Studies
All RGD peptide analogues tested in the displacement assays showed the expected sigmoid curves (Figure 2). The dimeric RGD peptide E[c(RGDyK)]2 showed higher binding affinity (IC50 = 2.3 ± 0.7 nM) for HBCECs as compared to the monomeric analogue c(RGDyK) (IC50 = 3.5 ± 0.3 nM). Modification of the RGD peptides with the 4-fluorobenzoyl group somewhat decreased their binding avidity, with the IC50 values for FB-c(RGDyK) and FB-E[c(RGDyK)]2 being 16.9 ± 0.5 nM and 6.7 ± 0.5 nM, respectively.
Biodistribution
The biodistribution of 18F-labeled dimeric RGD peptide ([18F]FB-E[c(RGDyK)]2) in female nude mice bearing subcutaneous U87MG glioblastoma tumors are shown in Figure 3. Data were obtained at 30 min, 1 hr, 2 hr, and 4 hr postinjection. Tumor weights ranged from 0.3 to 0.5 g. The tracer showed rapid blood clearance, with initial blood radioactivity concentration of 0.81 ± 0.09 %ID/g at 30 min pi, and only 0.09 ±0.1 %ID/g at 4 hr pi. Tumor uptake was rapid and high (6.79 ± 0.05 %ID/g at 30 min pi) with some tumor washout observed with time (uptake declined to 4.27 ± 1.04 and 2.04 ± 0.5 %ID/g, at 2 and 4 hr pi, respectively). Tumor-to-blood activity concentration ratio peaked at 4 hr (23 ± 2). Systemic clearance occurred predominately by renal excretion. Biodistribution data for [18F]FB-E[c(RGDyK)]2 and [18F]FB-RGD [12,14] are compared in Figure 4. Tumor uptake of dimeric RGD was significantly higher than that of monomeric RGD at all time points (p < .001). At 2 hr after tracer administration, tumor uptake of [18F]FB-E[c(RGDyK)]2 was 4.27 ± 1.04 %ID/g compared with only 1.56 ± 0.35 %ID/g for [18F]FB-RGD. Liver accumulation of monomeric RGD peptide was higher than that of dimeric RGD peptide (p < .001) at 30 min (2.50 ± 0.18 %ID/g for [18F]FB-RGD and 1.34 ± 0.18 %ID/g for [18F]FB-E[c(RGDyK)]2), but the difference diminished with time. Renal uptake of the dimeric RGD peptide was significantly higher than that of the monomeric RGD peptide at all time points examined. Consequently, the tumor-to-kidney ratio for the monomeric RGD peptide was higher than for the dimeric RGD peptide (2.8 ± 0.4 for [18F]FB-RGD and 1.4 ± 0.2 for [18F]FB-E[c(RGDyK)]2 at 2 hr pi, respectively). It is also noteworthy that intestinal uptake of the dimeric RGD peptide was significantly lower than that of the monomeric RGD peptide. Coinjection of [18F]FB-E[c(RGDyK)]2 with 10 mg/kg of c(RGDyK) resulted in decreased uptake at 1 hr pi in all dissected tissues except the kidneys. The most pronounced reduction of uptake (4.27 ± 1.04 to 0.45 ± 0.11 %ID/g) occurred in tumor (Figure 5). A similar phenomenon has also been observed with other radiolabeled RGD peptides [8,11–16,22].
Schematic structure of [18F]FB-E[c(RGDyK)]2 (MW = 1473). 18F-labeling was carried out via acylation of the amino group at the glutamate residue.
In vitro displacement of 125I-c(RGDyK) by c(RGDyK) (▪), FB-c(RGDyK) (▾), E[c(RGDyK)]2 (•), and FB-[c(RGDyK)]2 (△) in HBCECs. Values are the mean of three determinations.
Micro-PET and Autoradiography
A 2-D projection micro-PET image (15 min static single frame beginning at 60 min after intravenous injection of 150 μCi of [18F]FB-E[c(RGDyK)]2) of a mouse bearing U87MG tumor on the right front leg is shown in Figure 6A. The tumor was clearly visible, with high contrast to contralateral background (ratio: 9.5 ± 0.8). Prominent uptake was also observed in the gallbladder and urinary bladder. Liver uptake was significantly lower than tumor uptake (p < .001), and activity accumulation in the kidneys was slightly higher than in tumor, which agreed well with the data obtained from direct tissue sampling. Notice that, at the time of imaging, the activity in the intestinal tracts was very low, resulting in a very high-quality image and favorable tumor-to-nontumor ratios. As expected, no activity accumulation was observed in the normal brain, presumably due to a low level of αv-integrin expression in the intact blood–brain barrier [21,25]. As a comparison, 2-D projection micro-PET image in the presence of [18F]FB-RGD (15 min static scan at 1 hr postinjection of 150 μCi of monomeric RGD peptide radiotracer) is shown in Figure 6B. U87MG tumor on the right hind leg was visible with contrast to contralateral background, however, excessive activity accumulation in the intestinal tracts eliminated its possibility to visualize tumors in the lower abdomen [12,14].
Quantification of tumor and major organ activity accumulation in micro-PET scans was performed on coronal views using regions of interest (ROIs) that encompassed entire organs. Uptake in the U87MG tumor, liver, and kidneys was calculated as 4.4 ± 0.6, 0.9 ± 0.1, and 3.6 ± 1.2 %ID/g, respectively. These results were in good agreement with the biodistribution data obtained at 1 hr pi except for the tumor, which was underestimated due to partial volume effect. For one mouse, micro-PET images obtained following injection of [18F]FB-E[c(RGDyK)]2 were compared with subsequent high-resolution (0.1 mm) autoradiography. Figure 7 compares a coronal micro-PET slice (thickness 1.2 mm) including the U87MG xenograft (Figure 7, left) with a corresponding autoradiographic section (thickness 0.05 mm, Figure 7, right); correlations between the regions of high uptake observed in the micro-PET scan and autoradiography were clearly demonstrated.
Discussion
This study demonstrates that 18F-labeled dimeric RGD peptide [18F]FB-E[c(RGDyK)]2 has significantly improved tumor targeting efficacy, in particular, higher tumor uptake and prolonged tumor retention, compared to the monomeric RGD peptide analogue, [18F]FB-c(RGDyK). Imaging of αvβ3-integrin positive U87MG glioblastoma tumor in an intact mouse using a high-resolution micro-PET scanner resulted in strong contrast between tumor and normal tissues at 1 hr after injection of [18F]FB-E[c(RGDyK)]2.
Biodistribution of [18F]FB-E[c(RGDyK)]2 in athymic nude mice bearing subcutaneous U87MG glioblastoma tumors (% ID/g: % injected dose per gram).
Comparison of the biodistribution data of [18F]FB-E[c(RGDyK)]2 (□) and [18F]FB-c(RGDyK) (▪) in athymic nude mice with subcutaneous U87MG tumors.
Dysangiogenesis is involved in most of the major human pathologies [25]. Angiogenic properties of tumors are considered to be critical determinants of the potential for growth and metastasis. Angiogenic therapies have received a great deal of attention in recent years; methods to identify the angiogenic phenotypes of tumors and evaluate the efficacy of these therapies would be of great significance. Integrin receptors are implicated in many pathological processes, such as osteoporosis, misregulated angiogenesis (e.g., rheumatoid arthritis or retinopathy), thrombosis, and inflammation, as well as tumor growth and tumor metastasis [26]. Overexpression of αvβ3 integrin on the cell surfaces of tumor cells and activated ECs compared with resting ECs suggests the application of suitably labeled αvβ3 antagonists for imaging of receptor density levels in αvβ3-positive tumors. Potential antagonists include antibodies, cyclic RGD peptides, peptidomimetics of the confined RGD sequence, or small molecules. However, genetic ablation of αv integrins in mice has been associated with enhanced pathological angiogenesis [27]. This seemingly paradoxical finding implies that the exact role of αv in tumor angiogenesis is yet to be established. It is possible that αv integrins are negative regulators of angiogenesis and that the drugs targeting them may be acting as agonists rather than antagonists [28].
Glioblastoma multiforme, the most malignant astroglial-derived tumor, grows as an adherent mass and locally invades normal brain [29]. Both brain capillary and brain tumor cells express high levels of αvβ3 and αvβ5 integrins in vitro and in vivo [30], as also observed in human brain tumor specimens [31]. EMD 121974, c(RGDf[NMe]V), an αv-integrin antagonist, detaches both the αv-integrin-expressing brain capillary and brain tumor cells from the matrix proteins vitronectin and tenascin, resulting in significant apoptosis of both cell types [32]. Also, daily administration of EMD 121974 inhibited growth of a glioblastoma (U87MG) cell line xenotransplanted into the forebrain of nude mice resulting in increased survival [32]. Suitably labeled RGD peptide may be used to visualize brain tumor growth, quantify integrin receptor expression, and follow anti-integrin treatment efficacy, as well as to develop more potent anti-integrin drugs based on the tracer targeting ability and in vivo kinetics.
Biodistribution of [18F]FB-E[c(RGDyK)]2 (A) and [18F]FB-c(RGDyK) (B) in the absence (control) and presence (block) of 10 mg/kg nonradioactive monomeric RGD peptide c(RGDyK) at 1 h pi in athymic nude mice with subcutaneous U87MG tumors.
(A) 2-D projection of U87MG bearing mouse 60 min after injection of 150 μCi of [18F]FB-E[c(RGDyK)]2 (15 min static image). Activity in the kidneys and urinary bladder reflects the renal excretion of the radiolabeled peptide. (B) 2-D projection of U87MG tumor-bearing mouse 60 min after injection of 150 μCi of [18F]FB-c(RGDyK) (15 min static image).
Because of the very precise design of their configuration, modification and labeling of peptidomimetics and small molecule antagonists of αvβ3 integrin usually eliminate the receptor binding characteristics of such ligands. Furthermore, the lipophilicity of these types of ligands further limits their applicability for diagnostic imaging of αvβ3-integrin expression. On the other hand, cyclic RGD peptides, with suitable metabolic stability and nanomolar binding affinity for αvβ3 integrin, are ideal for development of suitably labeled probes for tumor visualization, αvβ3-integrin quantification in vivo, and assessment of anti-integrin treatment efficacy. Cyclic RGD peptides have been labeled with 18F [11–17] and 64Cu [14] for PET, and 111In and 99mTc [8,19,33,34] for SPECT imaging studies. In addition, antibodies have been attached to the surface of microbubbles and liposomes for contrast-enhanced ultrasound [35] and magnetic resonance imaging [36] of tumor angiogenesis, respectively.
Although γ-emitting radiolabels are more readily available and in general have longer half-lives relative to β+-emitting radionuclides, PET cameras have much higher sensitivity, better spatial resolution, and more accurate attenuation correction, and therefore better quantitative accuracy, than SPECT scanners [37]. The acquisition of higher count statistics is particularly valuable for detecting the fewest possible cells per unit volume with the least amount of radioactivity. We expect PET imaging with the 18F-labeled dimeric RGD peptide described here to provide higher sensitivity and resolution for early tumor detection and assessment of tumor treatment efficacy than SPECT imaging with 111In- and 99mTc-labeled peptide analogs.
18F-labeled monomeric RGD peptides, labeled either via electrophilic [17] or nucleophilic [11–14] substitution, have encountered the problems of rapid tumor washout and unfavorably high hepatobiliary excretion. Insertion of a sugar amino acid between the F-18 label and the RGD ligand increased the hydrophilicity of the peptide, which brought about improved in vivo pharmacokinetics of the resulting PET tracer. An alternative to this sialidation/glycosylation approach is to introduce a poly(ethylene glycol) moiety into the original lipophilic radiotracer ([18F]FB-RGD). The PEGylated peptide [18F]FB-PEG-RGD has improved bioavailability, increased stability toward enzymatic degradation and solubility under physiological conditions compared with [18F]FB-RGD [13]. In this study, dimeric RGD peptide E[c(RGDyK)]2 labeled with 18F through the 4-fluorobenzoyl moiety showed both renal and hepatobiliary excretion pathways, resulting in significant activity accumulation in the gallbladder, kidneys, and urinary bladder (Figure 6B). The affinity and specificity of the radiotracer for αv integrins and relatively slow blood clearance of the radiotracer might be responsible for the prominent tumor retention.
Left: Micro-PET scan (15 min single frame) of U87MG tumor-bearing mouse sacrificed 1 hr after injection of 150 μCi of [18F]FB-E[c(RGDyK)]2. Right: Digital autoradiograph of the section containing tumor. Tumors were visualized in the right upper leg (arrows).
It has been proposed by several groups that the receptor binding characteristics of dimeric and multimeric RGD peptides would be better than that of monomeric RGD peptide based upon polyvalency [18]. Simultaneous binding of multiple RGD motifs within a single molecule to one integrin or binding of multiple integrins to a single multimeric RGD peptide induces cooperativity, entropically enhanced affinity, and steric stabilization [18]. Immobilized human placental αvβ3-receptor ELISA assays have indicated a more avid binding of the dimer than the monomeric RGD analogue [19]. Dimeric RGD peptide E[c(RGDfK)]2 has thus been conjugated with 1,4,7,10-tetraazadodecane-N,N′,N¶ime;,N-tetraacetic acid (DOTA) and labeled with 90Y and 111In [22], and with hydrazinonicotinamide (HYNIC) for 99mTc [19]. It has been demonstrated that uptake of the dimer is higher than that of the monomer in most organs and tissues, but kidney uptake is also higher for the dimer [19]. Because kidneys received the highest radiation dose among normal tissues and seems to be the dose-limiting organ for the dimeric RGD peptide tracer, further modification of the radioligand is thus needed to obviate too high a renal dose.
Maintenance of integrin receptor binding affinity was previously tested in HBCECs, which are known to express high levels of αvβ3 and αvβ5 integrins [30]. Dimeric RGD peptide revealed higher integrin receptor affinity than the monomeric analog. Fluorobenzoyl group labeling decreased the receptor binding affinity in vitro for both monomeric and dimeric RGD peptides, which might be attributed to the added hindrance upon substitution of the primary amines. In the current study, higher uptake of [18F]FB-E[c(RGDyK)]2 as compared to [18F]FB-c(RGDyK) (Figure 4) might be attributed to more effective reabsorption of the dimeric RGD peptide and its major metabolite into the negatively charged proximal renal tubular cells than that of the monomeric RGD analogue [8,19]. Whether the increased tumor targeting of dimeric RGD peptide was due to cooperative binding or pharmacological effects remains to be determined since there is no direct evidence about whether the spatial arrangement and separation of two RGD moieties is optimal for multiple cell surface integrin binding.
Receptor-specific tumor uptake of [18F]FB-E[c (RGDyK)]2 was demonstrated by effective blocking of activity accumulation in U87MG glioblastoma xenograft resulting from coinjection of the radiotracer with monomeric RGD peptide c(RGDyK), which reportedly has high affinity and specificity for αvβ3 integrin [22]. It is unclear at this stage whether the visualization of the tumor is du to enhanced integrin expression on the neovasculature or enhanced integrin expression on tumor cells. Further validation of the noninvasive PET imaging requires correlation of the magnitude of tumor uptake with receptor density, which can be determined independently using antibodies that are specific to human αvβ3 integrin (U87MG tumor cells) and to mouse αv and/or β3 integrin (ECs). In order to address the ultimate goal of imaging in patients, model systems with intermediate integrin expression and negative controls should also be studied in order to represent the full extent of human presentations. Since PEGylation of monomeric RGD peptide significantly improved in vivo kinetics of the RGD peptide without compromising tumor-targeting efficacy [13,22], we expect that PEGylation may further enhance integrin targeting and improve the biodistribution of the dimeric RGD peptide tracer.
Footnotes
Acknowledgments
This work was carried out in part with contributions from NIBIB grant R21 EB001785 (to X. C.), ACS grant ACS-IRG-580007-42 (to X. C.), the Wright Foundation (to X. C.), DOD BCRP Concept Award DAMD17-03-1-0752 (to X. C.), DOD BCRP IDEA Award BC030012 (to X. C.), and NCI grant P20 CA86532 (to P. S. C.), The USC cyclotron team, particularly Joseph Cook and Luis Pedroza, are acknowledged for radionuclide production.