PET Imaging of VEGFR with a Novel 64Cu-Labeled Peptide.

Vascular endothelial growth factor receptors (VEGFRs) are well recognized as significant biomarkers of tumor angiogenesis. Herein, we have developed a first-of-its-kind peptide-based VEGFR positron emission tomography (PET) tracer. The novel [64Cu]VEGF125-136 peptide possessed satisfactory radio-characteristics and showed good specificity for the visualization of VEGFR in various mouse models, in which the tumor-specific radioactivity uptake was highly correlated to the VEGFR expression level. Moreover, the tracer showed high tumor uptake (ca. 5.89 %ID/g at 20 min postinjection in B16F10 mice) and excellent pharmacokinetics, achieving the maximum imaging quality within 1 h after injection. These features convey [64Cu]VEGF125-136 as a promising, clinically translatable PET tracer for the imaging of tumor angiogenesis.

Introduction

Tumor angiogenesis is a pivotal process for tumor growth and metastasis.[1,2] Vascular endothelial growth factor (VEGF)/VEGF receptor (VEGFR) signaling is primarily responsible for tumor angiogenesis and, as such, has been considered the principal target pathway for antiangiogenic therapy.[3,4] A member of the VEGF family, VEGF-A[5,6] acts through two VEGFRs, VEGFR-1 and VEGFR-2.[6,7] VEGFR-1/2 are type II transmembrane tyrosine kinase receptors that are expressed mainly on endothelial cells and dominate VEGF-induced microvascular permeability, endothelial cell proliferation, and the functions of cellular invasion.[8] In several types of cancer, the overexpression of VEGFR-1/2 in the tumor vasculature has been linked to tumor progression and poor prognosis, making them attractive biomarkers for antiangiogenic therapy via targeted drugs.[9,10]

Currently, more than ten VEGFR-targeted drugs are approved by the U.S. Food and Drug Administration (FDA) for the treatment of various tumors,[11−13] such as apatinib[14] and cabozantinib.[15] Even so, the overall response rate of the therapeutics is highly variable, most likely due to the temporal/spatial heterogeneity of the VEGFR expression levels amongst individual patients.[16] Therefore, methods for the noninvasive detection and quantification of VEGFR-1/2 expression are particularly significant for the selection of VEGFR-1/2-positive patients and the evaluation of therapeutic response following VEGFR-1/2–targeted therapies.

Several molecular imaging agents have been applied to detect and quantify VEGFR in angiogenic tumors, such as radiotracers for use in single-photon emission computed tomography (SPECT) and positron emission tomography (PET).[17−36] For example, Chen et al. developed PET radiotracers based on VEGF-A isoforms for imaging VEGFR.[18,23] More recently, several VEGFR-2-specific monoclonal antibodies labeled with copper-64 or zirconium-89 have shown encouraging results for the selective imaging of VEGFR-2.[27,33] Despite these advances, mAb-based imaging agents have shown unfavorable pharmacokinetics (PK) and high nonspecific uptake. Besides, several research efforts have been put toward affibody and peptide-based radiotracers.[35,36] These molecules showed satisfactory PK and image contrast, but their imaging mode is based on SPECT, which significantly limits signal sensitivity and image quality. Compared to SPECT, PET exhibits several advantages, including improved image quality, greater interpretive certainty, higher diagnostic accuracy, and lower patient dosimetry.[37] In the quest to detect VEGFR expression with high spatiotemporal accuracy, we reasoned that the most desirable imaging agents are peptide-based PET tracers that possess high specificity and preferable PK profiles.

VEGF125–136 (QKRKRKKSRYKS) is a 12-amino-acid peptide encoded by exon 6 of VEGF-A that was first identified as an effective inhibitor to VEGFR in 2001.[38] This peptide is composed of hydrophilic natural amino acids, lending it good solubility and easy synthesis. Based on these attributes, pioneer studies reported the labeling of VEGF125–136 with 188Re for the SPECT imaging of VEGFR-2.[24] However, this tracer resulted in SPECT images with low quality (including low image resolution and low image contrast), largely deviating from the clinical demands. In this context, it is conceivable that significant improvements could be made to the imaging capacity of VEGF125–136-based tracers by properly reforming VEGF125–136 with PET functionality. In this work, we have developed a VEGF125–136-based PET tracer named [64Cu]VEGF125–136 that displayed good PK and considerable stability in vivo, resulting in satisfactory imaging quality in the tested tumor mouse models (Scheme ).

Scheme 1

Development of the First VEGFR-2 PET Tracer Based on the VEGF125–136 Peptide Labeled with 64Cu

The tracer predominantly bound to VEGFR-2, as the tracer’s uptake in the tumor is associated with the expression level of VEGFR-2.

Results and Discussion

The chelator 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) was selected for 64Cu labeling (Figure A). To avoid any DOTA-based perturbation to the mode of VEGF125–136 binding, we added a spacer (poly(ethylene glycol)3 (PEG3)) between the peptide and DOTA. The labeling operation was conducted according to the method reported in a previous study.[39] The resulting overall synthetic processes, including 64Cu labeling, purification, and formulation, were completed in 30 ± 5 min (n = 7). The radiochemical yield (RCY), molar activity, radiochemical purity, and high-performance liquid chromatography (HPLC) retention time of [64Cu]VEGF125–136 are summarized in Table . The high molar activity and radiochemical purity (Figure B) made this PET tracer highly suitable for automated radiosynthesis. Furthermore, [64Cu]VEGF125–136 showed good stability with no decomposition or 64Cu detachment after 12 h of incubation in saline at 37 °C (Figure C). The tracer also showed considerable stability in mouse serum, as ∼45% of intact tracers were retained after 30 min of incubation with serum (Figures S1 and S2).

Figure 1

(A) Chemical structure of DOTA–PEG3–VEGF125–136. (B) Purity determination for [64Cu]VEGF125–136 by analytical RP-HPLC. (C) Competition-binding curve of [64Cu]VEGF125–136 to U87MG cells. Log of concentration of competitor compounds (VEGF125–136 or DOTA–PEG3–VEGF125–136) versus the percentage of the uptake of the radiolabeled molecules.

Table 1

Quality Control Results for [64Cu]VEGF125–136a

tracer	[64Cu]VEGF125–136
radiochemical yield (%)	>95
molar activity (GBq μmol–1)	74.3 ± 3.8
radiochemical purity (%)b	>98
retention time (tR)	6.53 min

Radiochemical yield (RCY), molar activity, and radiochemical purity of the as-prepared tracer. Data are expressed as mean ± standard deviation (SD) (n = 7).

The radiochemical purity was determined by HPLC with the conditions as follows: column, YMC-Triat-C18 column (4.6 mm i.d. × 150 mm, 5 μm); solvent gradient, 10–90% acetonitrile (0.1% trifluoroacetic acid (TFA)), 20 min; flow rate, 1 mL/min.

Next, we evaluated the influence of PEG3–DOTA on the binding affinity of VEGF125–136 to VEGFR., and were thus used to compare the binding affinities of VEGF125–136 and DOTA–PEG3–VEGF125–136 using a competitive binding assay.[32] As expected, DOTA–PEG3–VEGF125–136 exhibited a similar IC50 (302.7 ± 4.9 nM) to that of VEGF125–136 (277.3 ± 6.19 nM, Figure C), indicating that negligible perturbation of PEG3–DOTA occurred to the binding mode of VEGF125–136.

We further examined the imaging capacity of this radiotracer in vivo. Three tumor models, glioblastoma U87MG, murine melanoma B16F10, and breast cancer MDA-231, were selected for PET studies based on their differing levels of VEGFR expression.[32] The tumor bearing mice (n = 3, respectively) were imaged with 1 h dynamic PET scans. Representative decay-corrected axial and coronal images at 40 min after intravenous (iv) injection are shown in Figures A and S3. It was apparent that the accumulative uptakes of the tracer were variable in the different tumors, with the order: B16F10 > U87MG > MDA-231. Quantitative PET analysis of the activity in tumor and muscle tissues is shown in Figure B–D. In the B16F10 model, the tracer quickly accumulated in the tumor and the radioactivity peaked at 10 min postinjection (p.i.). Moreover, the activity in the tumor was stably maintained at 3.1 ± 0.2 %ID/g from 10 to 60 min p.i. Contrastingly, in the muscle tissue, the radioactivity peaked within 5 min and then quickly decreased, with a ratio of tumor to muscle at 60 min (Rt/m,60min) of 4.55 ± 0.83. In the U87MG tumor model, although there was a high peak uptake (3.3 %ID/g) at 10 min p.i., the tracer was rapidly washed out from the tumor, suggesting that there was lower VEGFR expression at the tumor site. Significantly different from the B16F10 and U87MG tumors, the MDA-231 tumor showed a low uptake of the tracer within the whole monitoring process, exhibiting no statistical significance in the radioactivity in muscle.

Figure 2

Small animal PET imaging studies. (A) Representative axial and coronal PET images of mice bearing the B16F10, U87MG, and MDA-231 tumors at 40 min after injection of [64Cu]VEGF125–136. (B–D) Time–activity curves of radioactivities in tumor and muscle after iv injection of [64Cu]VEGF125–136.

To validate this observation in microPET and evaluate the tracers’ PK, we further performed an ex vivo biodistribution assay. As shown in Table , we observed a rapid and high accumulation of the tracer in the tumor 20 min after injection (5.89 ± 2.58 %ID/g). We observed only a slight decrease at 60 min after injection (4.45 ± 0.61 %ID/g), coinciding with the findings obtained by PET imaging. Furthermore, [64Cu]VEGF125–136 displayed rapid blood clearance, with 3.18 ± 0.14 %ID/g observed at 20 min p.i., which then dropped to 0.88 ± 0.12 %ID/g at 60 min p.i. Hence, the tumor-to-blood ratio increased from 1.84 ± 0.77 at 20 min to 5.11 ± 1.02 at 60 min. Similarly, the tumor-to-muscle ratio reached its maximal value at 15.13 ± 2.20 at 60 min. The radioactivity was primarily excreted through the kidney–bladder route, as was evidenced by the extremely high radioactivity observed in the kidney and bladder tissues. All other major organs examined in this study showed low uptake and fast clearance, further demonstrating the high specificity and good PK of this tracer. The results obtained from biodistribution studies and PET scans were highly consistent in the B16F10 model, confirming the robustness of the tracer for visualization of VEGFR in tumorigenesis.

Table 2

Biodistribution of [64Cu]VEGF125–136 in B16F10 Tumor Bearing Mice at 20 min and 1 h Postinjection

	20 min	60 min
tumor	5.89 ± 2.58	4.45 ± 0.61
blood	3.18 ± 0.14	0.88 ± 0.12
heart	1.70 ± 0.28	0.59 ± 0.04
kidney	110.12 ± 12.72	156.97 ± 8.63
liver	4.58 ± 0.60	3.93 ± 0.42
lung	4.38 ± 1.02	1.73 ± 0.27
thymus	1.07 ± 0.24	0.51 ± 0.10
pancreas	1.03 ± 0.16	0.39 ± 0.02
spleen	1.86 ± 0.45	0.97 ± 0.10
small intestine	2.18 ± 0.36	1.83 ± 0.39
intestinal lymph node	2.62 ± 0.41	1.52 ± 0.72
muscle	0.95 ± 0.14	0.29 ± 0.01
bone	2.59 ± 0.76	1.78 ± 0.11
testis	0.80 ± 0.07	0.39 ± 0.05
stomach	2.01 ± 0.24	0.97 ± 0.10
brain	0.17 ± 0.02	0.06 ± 0.01
urinary bladder	10.89 ± 3.32	14.62 ± 5.81

To verify the specific binding of [64Cu]VEGF125–136 to VEGFR, we carried out autoradiography (ARG) in sagittal B16F10 tumor sections (Figure A). The results of the normal experiment showed a high level of binding of the tracer to the tumor section (Figure A); however, the majority of the binding was successfully blocked by DOTA–PEG3–VEGF125–136 (>90% reduction, Figure B). These results suggested that [64Cu]VEGF125–136 was exclusively bound to VEGFR. To rationalize the variable uptake of the tracer in different tumor models, we then set out to detect the VEGFR expression level in each tumor using immunofluorescence staining. The representative images of VEGFR-1/2 staining of B16F10, U87MG, and MDA-231 tumor tissue slices are shown in Figures C and S4. The results revealed that the strongest VEGFR-1/2 staining was in the B16F10 tumors and that relatively weaker staining appeared in U87MG tumors, and the weakest staining was observed in MDA-231 tumors. These findings were also highly consistent with the results obtained from PET imaging, demonstrating that the [64Cu]VEGF125–136 uptake was well correlated to the expression of the VEGFR protein.

Figure 3

(A) In vitro ARG of [64Cu]VEGF125–136 in B16F10 tumor sections. (B) Quantification of radioactivity under normal and blocking conditions. (C) Immunofluorescence staining of VEGFR-1/2 in frozen B16F10, U87MG, and MDA-231 tumor slices.

Here, we have described the development of a first-of-its-kind peptide-based VEGFR PET tracer. By leveraging the VEGFR binding sequence (VEGF125–136) reported previously, we successfully reformed the peptide with PET functionality. We effectively demonstrated the specificity of the tracer to VEGFR using a competitive cell-binding assay and an ARG experiment. We then validated the robustness of the tracer to map VEGFR in living animals by performing PET imaging and ex vivo biodistribution analysis, showing that the uptake of the tracer was closely correlated to the expression of VEGFR-1/2. Despite success, our study also shows some shortcomings. The peptide VEGF125–136 was demonstrated to selectively bind to VEGFR-2;[38] however, in this study, we were unable to observe such selectivity in in vivo tumor models. Considering that the selective imaging of VEGFR subtype is more valuable for precision medicine, we will carry out further study to develop more specific radiotracers based on this lead peptide. Strategies such as β-amino-acid replacement and peptide cyclization would be effective tools to increase the selectivity.

Conclusions

In summary, [64Cu]VEGF125–136 could be considered as a desirable and clinically translatable tracer due to the following merits. First, the peptide (precursor) is easily synthesized and its radiolabeling can be conveniently scaled up using automated synthesis modules. Second, the maximum uptake in the tumor and the corresponding satisfactory image contrast can be achieved in 1 h after injection (for data exceeding 1 h, see Figure S5), making it more acceptable for clinical translation. What is more, the fast excretion of the tracer from nonspecific organs minimizes unwanted exposure to radioactivity (estimated half-life in blood is about 22 min). In sum, the novel tracer [64Cu]VEGF125–136 is a state-of-the-art PET imaging agent for VEGFR. It has great potential for broad application in the future, including aiding in evaluating the efficacy of newly developed VEGF(R)-targeted cancer therapeutics. It may also be significantly beneficial for patient stratification and monitoring the response to treatment in VEGFR-targeted cancer therapy.

Materials and Methods

Chemical Reagents and Equipments

The peptide PEG3-VEGF125–136 (QKRKRKKSRYKS) was synthesized by a solid-phase peptide synthesis (SPPS) protocol using a semiautomated peptide synthesizer. All chemicals were purchased from Wako Pure Chemical Industries (Osaka, Japan) or Sigma-Aldrich (St. Louis, MI). 64Cu was produced at the National Institute of Radiological Science (NIRS, Chiba, Japan) with 98% radionuclidic purity in house. All radio-HPLC analysis for the tracers was performed using a JASCO HPLC system (JASCO, Tokyo, Japan) coupled with a YMC-Triat-C18 column (4.6 mm i.d. × 150 mm, 5 μm, Waters, Milford, MA). A flow rate of 1 mL/min was used. The gradient started with 90% solvent A (0.1% trifluoroacetic acid [TFA] in water) and 10% solvent B (0.1% TFA in acetonitrile [MeCN]), and 20 min later, ended with 0% solvent A and 100% solvent B. Effluent radioactivity was measured using a NaI (TI) scintillation detector system (Ohyo Koken Kogyo, Tokyo, Japan). A Wizard 2480autogamma counter (PerkinElmer, Waltham, MA) was used to measure radioactivity, as expressed in counts of radioactivity per minute (CPM), accumulating in cells and animal tissues. A dose calibrator (IGC-7 Curiemeter; Aloka, Tokyo, Japan) was used for the other radioactivity measurements.

Preparation of DOTA–PEG3–VEGF125–136

The synthetic procedure for DOTA–PEG3–VEGF125–136 is illustrated in Scheme S1. Here, we only introduce the method for DOTA conjugation (2). DOTA was preactivated by N-(3-dimethylaminopropyl)-N-ethylcarbodiimide hydrochloride (EDC·HCl, 2 mg) at a molar ratio for DOTA/EDC·HCl/N-hydroxysuccinimide (NHS) of 10:5:4 in dimethyl sulfoxide (DMSO, 2 mL) for 3 h. The PEG3-VEGF125–136 resin (100 mg; 0.37 mmol/g) was suspended in 1 mL of N-methyl-2-pyrrolidone (NMP). Then the activated DOTA as well as 100 μL of DIEA was co-added to the resin. The reaction mixture was gently bubbled by N2 gas for 2 h. Then the DOTA–PEG3–VEGF125–136 was cleaved from the resin and purified by HPLC. The collected product eluents were lyophilized and dissolved in H2O at a concentration of 1 mg/mL for use in radiolabeling reactions.

64Cu Radiolabeling

DOTA–PEG3–VEGF125–136 peptide was radiolabeled with 64CuCl2. The labeling method is the same as that we previously reported. Specifically, 10 μg of DOTA–PEG3–VEGF125–136 in 10 μL of 0.1 M sodium acetate buffer (pH = 4.1) was reacted with ∼370 MBq (10 mCi) of the neutralized 64CuCl2 solution at 80 °C for 10 min. After incubation, the reaction mixture was purified with RP-HPLC.

In Vitro Stability Assays

The stability of the tracer in saline and mouse serum was examined. Ten microliters (∼100 μCi) of tracers were added to 90 μL of mouse serum (freshly prepared) and incubated at 37 °C with slight agitation for a designated time. Aliquots were taken out at the designated time point, and 100 μL of acetonitrile and water (1:1, v/v) was added. The suspension was centrifuged for 10 min at 10 000 rpm. The supernatant was analyzed by analytical HPLC. For HPLC, the analysis used an additional guard column (Phenomenex Security Guard 3.00 mm i.d.) to protect the C-18 column. For stability in saline, the tracers were incubated in saline (>90%, v), and then aliquots were removed at each time point for HPLC analysis.

Cell Line and Animals

B16F10 human melanoma cells, U87MG human glioblastoma cells, and MDA-231 human breast cancer cells were purchased from the American Type Culture Collection and grown in DMEM with 10% fetal bovine serum at 37 °C with 5% CO2. Cells were used for in vitro or in vivo experiments when they reached approximately 75% confluence.

Male C57BL/6 Jms and male BALB/c nude–/– mice (7-week-old) were purchased from Japan SLC (Shizuoka, Japan). All animals received humane care, and the Animal Ethics Committee of the National Institute of Radiological Sciences approved all experiments. All experiments were carried out according to the recommendations of the Committee for the Care and Use of Laboratory Animals, National Institute of Radiological Sciences. The B16F10 tumor bearing models were established using C57BL/6 J mice via a left flank subcutaneous injection of B16F10 cells (1 × 106 cells per mouse) one week before the PET imaging experiment. The U87MG and MDA-231 tumor bearing models were established using BALB/c nude–/– mice via a left hind leg or left flank subcutaneous injection of U87MG cells (5 × 106 cells per mouse) and MDA-231 cells (5 × 106 cells per mouse), respectively. Tumor bearing mice were used for studies when tumor diameters reached 3–5 mm.

Cell-Based Competitive Binding Assay

The [64Cu]VEGF125–136 cell uptake assay was performed using U87MG that overexpress VEGFR, as demonstrated by previous studies. The day before the experiment, cells were seeded in 12-well plates at a concentration of 1 × 105 cells per well in the appropriate growth medium (see the previous section). The binding inhibition assay was performed using U87MG cells. DOTA–PEG3–VEGFR125–136 or VEGFR125–136 (final concentration: 0, 10, 100, 1000, or 10 000 nmol/L) and the radiotracers (74 kBq/1 mL) in the medium were added to each well, and the plates were incubated at 37 °C for 1 h. After incubation, the cells were washed three times with ice-cold phosphate-buffered saline (PBS) and trypsinized. The cells were harvested into a microfuge tube and spun down at 1500 rpm in a microcentrifuge. Cell pellet associated radioactivity was measured using a γ counter (Perkin-Elmer Packard). All experiments were performed in triplicate.

Small Animal PET Study

PET scans were conducted using an Inveon PET scanner (Siemens Medical Solutions, Knoxville, TN), which provides 159 transaxial slices with 0.796 mm (center-to-center) spacing, a 10 cm transaxial field of view, and a 12.7 cm axial field of view. All list-mode acquisition data were sorted into three-dimensional (3D) sinograms, which were then Fourier rebinned into two-dimensional (2D) sinograms (frames × min: 4 × 1, 8 × 2, 8 × 5). PET dynamic images were reconstructed with filtered back projection using a Hanning filter and a Nyquist cutoff of 0.5 cycles/pixel. Tumor bearing mice were kept in the prone position under anesthesia with 1–2% (v/v) isoflurane during the scan. The tracers (8–17 MBq/200–500 μL) in saline were injected via a preinstalled tail vein catheter. Immediately after the injection, a dynamic scan in 3D list mode was acquired for 60 min. Maximum intensity projection images were obtained for tumor bearing mice. PET dynamic images were reconstructed by filtered back projection using a Hanning filter with a Nyquist cutoff of 0.5 cycles/pixel, which was summed using analysis software (ASIPro VM, Siemens Medical Solutions). Volumes of interest, including the tumors and muscle, were placed using the ASIPro software. The radioactivity was decay-corrected for the injection time and expressed as the percent of the total injection dose per gram tissue (%ID/g).

Ex Vivo Biodistribution

The tracer in saline (1.78 ± 0.11 MBq/0.1 mL) was administrated to B16F10 melanoma bearing C57BL/6 Jms mice via the tail vein. The mice were subsequently killed by cervical dislocation at 20 min, 60 min, 6 h, and 22 h after [64Cu]VEGF125–136 injection (n = 3). Major organs, including the heart, liver, lung, spleen, pancreas, kidneys, stomach, muscle, small intestine, intestinal lymph node, testis, muscle, bone, tumor, and blood, were quickly harvested and weighed. The radioactivity in these organs was measured using the autogamma counter. The results are expressed as the percentage of the injected dose per gram of wet tissue (%ID/g). All radioactivity measurements were decay corrected.

In Vitro Autoradiography

The B16F10 tumor was cut into 20 μm sections and stored at 80 °C until they were used for the experiment. The tumor sections were preincubated with Tris-HCl buffer (pH 7.4, 50 mM), MgCl2 (1.2 mM), and CaCl2 (2 mM) solution for 20 min at ambient temperature, followed by incubation with [64Cu]VEGF125–136 (37 MBq/200 mL) for 30 min at ambient temperature. For blocking studies, unlabeled DOTA–PEG3–VEGFR125–136 (10 mM) was added to the incubation solution in advance to determine the specificity of radioligand binding. After incubation, tumor sections were washed with Tris buffer two times for 2 min and dipped in cold distilled water for 10 s. The tumor sections were dried with cold air then placed on imaging plates (BASMS2025, GE Healthcare, NJ) for 20 min. Autoradiograms were obtained and ROIs were carefully drawn with the reference of naked-eye observation. Radioactivity was expressed as photostimulated luminescence values per unit area (PSL/mm2) and measured by a Bio-Imaging analyzer system (BAS5000, Fujifilm).

Immunofluorescence Staining

Frozen tissue slices of 10 μm thickness were fixed with cold acetone for 10 min and dried in the air for 30 min. After being rinsed with PBS and blocked with 10% donkey serum for 30 min at 25 °C, the slices were incubated with 10 nM rabbit antimouse VEGFR-2 antibody (Abcam, used in 1:150 dilution) overnight at 4 °C. After being washed with PBS, tissue slices were stained with Alexa 594-labeled donkey antirabbit IgG for 2 h at 25 °C. All slides were mounted in the medium for fluorescence with 2-(4-amidinophenyl)-1H-indole-6-carboxamidine (DAPI) for nuclear counterstaining (Vector Laboratories) and observed under a Zeiss Axiovert microscope.