Tumor periphery and lymph nodes of tumor-induced lymphangiogenesis often abundantly express VEGFR-3. In our previous study, we identified a 5-amino acid peptide named TMVP1, which binds specifically to VEGFR-3. The objective of this study was to develop a novel 68Ga-labeled TMVP1 for VEGFR-3 PET imaging and to investigate its safety, biodistribution, and tumor-localizing efficacy in xenograft tumor models and a small cohort of patients with recurrent ovarian and cervical cancer.
The DOTA-conjugated TMVP1 peptide was labeled with radionuclide 68Ga. SPR and saturation binding assays were used for the receptor-binding studies. Gynecologic xenograft tumors were employed for small-animal PET imaging and biodistribution of 68Ga-DOTA-TMVP1 in vivo. In the clinical study, 5 healthy volunteers and 8 patients with gynecologic cancer underwent whole-body PET/CT after being injected with 68Ga-DOTA-TMVP1.
DOTA-TMVP1 was successfully labeled with 68Ga. LECs showed higher binding capacity with 68Ga-DOTA-TMVP1 than LEC(shVEGFR-3) and human umbilical vein endothelial cells. In mice with subcutaneous C33-A and SKOV-3 xenografts, the tracer was rapidly eliminated through the kidney to the bladder, and the small-animal PET/CT helped to clearly visualize the tumors. In patients with recurrent ovarian cancer and cervical cancer, tracer accumulation well above the background level was demonstrated in most identified sites of disease; especially with recurrent endodermal sinus tumors, the diagnostic value of 68Ga-DOTA-TMVP1 was comparable with that of 18F-FDG PET/CT.
68Ga-DOTA-TMVP1 is a potential PET tracer for imaging VEGFR-3 with favorable pharmacokinetics.
VEGFR-3 is a potential target for developing new therapeutic drugs, several kinds of VEGFR-3 targeting agents have been developed, including VEGFR-3 mAbs, anticalins, and small pharmaceutical molecules. TMVP1 was identified as a VEGFR-3–binding peptide, which when labeled with radionuclide 68Ga, may be useful as a noninvasive radiotracer for assessing of VEGFR-3 levels in primary and metastatic tumors. 68Ga-DOTA-TMVP1 PET imaging may provide dynamic management design for VEGFR-3–based molecular therapy.
Most deaths from cancer occur as a result of metastasis. Lymphangiogenesis plays an important role in many solid tumors, including cervical cancer, breast cancer, and gastric cancer etc (1, 2). New lymphatic vessels in the tumor microenvironment correlate with tumor progression and metastasis (3). Therefore, blocking of tumor lymphangiogenesis has been regarded as a suitable therapeutic strategy (4, 5). However, imaging and quantifying new lymphatic vessels in vivo has historically been a challenge.
VEGFR-3 has been identified as one of the main lymphatic specific markers, whose expression is restricted to endothelial cells of lymphatic vessels in adults and is also found in many solid tumors (6, 7). A Vegfr3-reporter transgenic mouse model demonstrated that VEGFR-3 expressions were abundant in tumor-induced lymphangiogenesis at the tumor periphery and in lymph nodes (6, 8). In recent years, the VEGF-C/VEGFR-3 signaling pathways have been confirmed to play an important role in mediating the formation of solid tumors and in the intratumoral lymphangiogenesis process (9–11). These findings suggest that VEGFR-3 is a potential target for developing new tumor-induced lymphangiogenesis imaging agents and therapeutic drugs.
We previously identified a novel peptide, TMVP1 (LARGR), which can bind specifically to VEGFR-3. We carefully assessed the accuracy and specificity of TMVP1 binding to VEGFR-3 in vitro and in vivo. A VEGFR-3 PET radioligand represents a noninvasive tool for serial imaging of lymphatic vessel density in both primary and metastatic tumors. Herein, TMVP1 peptide was labeled with radionuclide 68Ga to generate 68Ga-DOTA-TMVP1, and the biologic profile and targeting effect of 68Ga-DOTA-TMVP1 in vitro and in tumor-bearing mouse models were evaluated (12, 13). In this study we further report on the safety, biodistribution, and tumor-targeting potential of 68Ga-DOTA-TMVP1 in patients with recurrent ovarian cancer and cervical.
Materials and Methods
Human cervical adenocarcinoma (C33-A) or human ovarian adenocarcinoma (SKOV-3) cells were purchased from ATCC and cultured according to their guidelines. C33-A and SKOV-3 cells were authenticated at the China Center for Type culture collection using short tandem repeat DNA profiling. Cells were used for experiments within 20 passages. Human normal lymphatic endothelial cells (LEC) were purchased from ScienCell and human umbilical vein endothelial cells (HUVEC) were purchased from Procell Life Science and Technology, and cultured in endothelial cell medium (ScienCell) with 5% FBS and endothelial growth medium supplements.
Preparation of 68Ga-DOTA-TMVP1
The TMVP1 peptide was presented as the general structure GCGXXXXXGC [XXXXX represented as the core amino acids (LARGR)], flanked by two cysteine residues to allow disulfide linkage and loop formation as our previous design (14). For labeling with radionuclide 68Ga, DOTA was employed as a bifunctional chelator (15, 16). DOTA-GGG(CGLARGRGC) was synthesized by WuXi AppTec Ltd.; 68Ga-DOTA-TMVP1 was obtained as previously reported, and subsequent quality control was performed as detailed in the Supplemental Materials and Methods (17, 18).
In vitro stability and the plasma protein binding measurement
In vitro assays
To achieve a VEGFR-3 knockdown, LECs were transfected with Lentiviral particles containing shVEGFR-3 or control, and incubated for 48–72 hours. To analyze knockdown efficiencies, immunofluorescence and Western blot analysis were performed. Cell uptake studies were conducted on LEC, LEC(shVEGFR-3), and HUVEC by incubating them with 148 KBq of 68Ga-DOTA-TMVP1 at 30, 60, and 120 minutes. For the internalization assay, a previously published protocol was followed (20). For blocking study, the above cell lines were incubated with 148 KBq of 68Ga-DOTA-TMVP1 with or without 100-fold excess DOTA-TMVP1 for 60 minutes at 37°C.
Saturation binding assays
Saturation binding experiment was performed in LECs, on the day of the experiment the cells were incubated at 37°C in 5% CO2 atmosphere for 90 minutes in the presence of increasing concentration of 1–40 nmol/L with or without 100-fold excess DOTA-TMVP1. Then the cells were washed three times with 0.4 mL of chilled PBS and harvested in 0.5 mL of 0.1 mol/L NaOH. The cell suspensions were collected and measured in the Radio-immune Gamma Counter. Specific binding was calculated by total binding minus nonspecific binding at each concentration. The KD value was calculated using GraphPad prism software (21–23).
Peptide–protein interaction assays
Surface plasmon resonance (SPR) was used for the receptor-binding assay. The cold gallium complex (natGa-DOTA-TMVP1) were synthesized and analyzed by high-performance liquid chromatography (HPLC) and LC/MS (detailed in the Supplementary Materials and Methods). Then OpenSPR Instrument (Nicoya Lifesciences Inc.) was used to determine the equilibrium binding constant (KD) of natGa-DOTA-TMVP1 with VEGFR-3. All the steps were performed as detailed in Supplementary Materials and Methods (24, 25).
All animal experiments were approved by the Hubei Institute Animal Research Committee and carried out following the Guide for the Care and Use of Laboratory of Tongji Hospital. Female BALB/c nude mice were obtained from Beijing HFK BioScience Co. Ltd. C33-A or SKOV-3 cells were subcutaneously injected into the axillary fat pad of athymic mice to establish the tumor-bearing model. In approximately 4 weeks, the xenograft masses reached a size of approximately 100 mm2.
Small-animal PET imaging
For dynamic PET scans, tumor-bearing mice were injected intravenously with 3.7 MBq of 68Ga-DOTA-TMVP1 under isoflurane anesthesia. The acquisition was performed at 30, 60, and 120 minutes with an Inveon PET Scanner (Siemens Preclinical Solutions). Images were reconstructed using the 2-dimensional ordered-subsets expectation maximum algorithm without attenuation or scatter correction. Regions of interest over xenograft tumors and contralateral tissues were measured for ratios of tumor site Bq/mL to the contralateral site (T/N ratios) on whole-body coronal images for each PET scan using vendor software (ASI Pro 22.214.171.124; Preclinical Solutions, Siemens; ref. 19). The blocking study was further conducted by coinjection with 1 mg/kg DOTA-TMVP1 peptide.
In vivo biodistribution
The tumor uptake of 68Ga-DOTA-TMVP1 was determined using the C33-A tumor-bearing BALB/c nude mice. A 1.85 MBq of 68Ga-DOTA-TMVP1 was injected by tail vein into the xenograft model. At 30, 60, and 120 minutes postinjection, the mice were euthanized by cervical dislocation, and blood samples were withdrawn immediately from the retro-orbital sinus using capillary tubes. The organs of interest were collected and weighed. Radioactivity was measured by the Radio-immune Gamma Counter. Organ and tissue uptakes were calculated as a percentage of the injected dose per gram of tissue mass (% ID/g). Tumor-to-normal tissue (T/NT) ratios are reported as an average for each timepoint.
Volunteers and patient recruiting
This first-in-human study was approved by the Institutional Review Board of Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, and Peking Union Medical College (No: ZS-817; Beijing, China), and was conducted in accordance with the Declaration of Helsinki. The trial was registered with chictr.org (ChiCTR-DOD-15006474). All subjects gave written informed consent to participate. Five volunteers, 5 patients with recurrent ovarian cancer and 3 patients with recurrent cervical cancer underwent PET/CT imaging with each bed position lasting for 2 minutes on average 40 ± 10 minutes after injection of 68Ga-DOTA-TMVP1 (102 ± 28 MBq; range, 70–162 MBq). The acquisition field covered from the top of the skull to the middle of the femur with six or seven bed positions, depending on the height of the patient. Corresponding 18F-FDG PET/CT image was acquired with the same patients within 1 week.
Tumor samples were fixed with 4% neutral buffered formalin and embedded in paraffin. The IHC-P protocol was followed as the basic steps (SP-9001; ZSGB-BIO). Anti-VEGF Receptor 3 antibody (Ab27278; Abcam) was diluted at 1:200. The number of VEGFR3+ vessels per 200× field at the xenograft site was counted for six fields. Lymphatic vessel density (number of VEGFR3+ vessels/field) was quantified in terms of mean ± SD.
Quantitative data are expressed as means ± SD. Means were compared using Student t test and P < 0.05 indicates significance (*, P < 0.05; **, P < 0.01).
Radiolabeling and stability
DOTA-TMVP1 was labeled successfully with 68Ga (Fig. 1A), and the RCP of 68Ga-DOTA-TMVP1 was always greater than 98%, which was confirmed by reversed-phase HPLC (RP-HPLC) and instant thin-layer chromatography–silica gel (SG) (Fig. 1B). 68Ga-DOTA-TMVP1 also showed good stability, it was stable in saline, cysteine, and human serum for at least 3 hours (Fig. 1C). As shown in Fig. 1D, the plasma protein binding rate was about 20%–30% at different time points.
Affinity analysis of natGa-DOTA-TMVP1 associated with VEGFR-3 protein
The LC/MS analysis of natGa-DOTA-TMVP1 confirmed the presence of a single main complex at 757.5 [m/z] 2+, which was in agreement with the calculated value (MW = 1,515.7). The binding kinetics between natGa-DOTA-TMVP1 and VEGFR-3 were determined using the SPR. The results showed that KD value was 6.73 × 10−6 mol/L in the SPR assay (Fig. 2A).
Cellular uptake, internalization, and competitive binding assay
Time dependent cellular uptake of 68Ga-DOTA-TMVP1 was studied in LEC, LEC(shVEGFR-3), and HUVEC (Fig. 2B). Immunofluorescence and Western blot analysis confirmed the lower expression of VEGFR-3 in LEC(shVEGFR-3) and HUVECs than LECs (Supplementary Fig. S1). The uptake of 68Ga-DOTA-TMVP1 in LECs at 60 minutes (3.02% ± 0.51% of added doses) was higher than that in HUVECs (0.87% ± 0.09%; **, P < 0.01) and LECs(shVEGFR-3) (1.01% ± 0.21%; *, P < 0.05) at different time points. The internalization assay showed steady internalization of approximately 50%–65% of the attached radioactivity in the LECs (Fig. 2C). Similarly, high and specific binding was observed for 68Ga-DOTA-TMVP1, and cellular uptake was inhibited by excess DOTA-TMVP1 at 60 minutes (Fig. 2D; *, P < 0.05). Saturation binding assays were performed using the LECs cells, the KD value of 68Ga-DOTA-TMVP1 was found to be 8.49 ± 1.94 nmol/L (Supplementary Fig. S2).
Small-animal PET imaging
We performed micro-PET imaging in C33-A and SKOV-3 tumor–bearing mice to evaluate the potential of 68Ga-DOTA-TMVP1 for molecular imaging of VEGFR-3. Micro-PET whole-body coronal images of mice demonstrated a low background for the whole-body imaging (Fig. 3). A focal area of 68Ga-DOTA-TMVP1 uptake was detected in the C33-A and SKOV-3 tumors, and the tumor-to-background contrast increased over time in the xenograft model after tracer administration; T/N was 3.58 ± 0.31 and 4.98 ± 0.28 in the C33-A and SKOV-3 tumor models at 120 minutes, respectively (Fig. 3B). The specificity of 68Ga-DOTA-TMVP1 uptake in the tumor was further confirmed in blocking studies, in which excess TMVP1 was coinjected along with the radiotracer. The accumulation of 68Ga-DOTA-TMVP1 in tumors was reduced by 80% in two tumor models compared with mice that received the radiotracer alone at 60 minutes (Fig. 3C). In addition, the radiotracer accumulated mainly in the kidney and bladder, which reflected the renal-mediated clearance of the radiotracer. IHC demonstrated that positive VEGFR3+ vessels were abundant in C33-A and SKOV-3 tumor tissues (Fig. 3D and E).
In vivo biodistribution
The biodistribution of 68Ga-DOTA-TMVP1 was evaluated in the athymic mice bearing C33-A tumors. %ID/g was quantified at different time after the injection of radiotracer in different organs with or without VEGFR-3 receptors blocking. In accordance with micro-PET imaging, the radiotracer was predominantly cleared by the urinary system, the kidney had the highest uptake compared with other organs at the same timepoint (Fig. 4A). Tracer accumulation in C33-A tumors plateaued between 30 and 60 minutes after tracer administration, and both T/Blood and T/Muscle had the highest ratio at 60 minutes (Fig. 4B). The in vivo binding specificity of radiotracer was confirmed by preinjection with a blocking dose of cold TMVP1. Accumulation of 68Ga-DOTA-TMVP1 in C33-A tumors was reduced by 70% compared with mice that received the radiotracer alone at 60 minutes (Fig. 4C).
Safety and PET/CT imaging
Between June 2015 and July 2016, five healthy volunteers and 5 patients with recurrent ovarian cancer and 3 patients with cervical cancer completed the study protocol. The patients received 68Ga-DOTA-TMVP1 and then underwent PET/CT imaging, and a corresponding 18F-FDG PET/CT was performed within 1 week. Patient characteristics are summarized in Supplementary Table S1.
No subjective effects were reported by any of five healthy volunteers and 8 patients with cancer after injection of 68Ga-DOTA-TMVP1. No adverse events or apparent changes in vital signs or clinical laboratory test results occurred after injection. Figure 5 and Supplementary Table S2 show the biodistribution of 68Ga-DOTA-TMVP1 in patients and healthy volunteers. The tracer was rapidly cleared from the blood, which allowed imaging at early time points. Tracer elimination occurred through the renal system, with high accumulation in the kidney, ureter, and bladder. Weak uptake was seen in glandular tissues such as the thyroid, pituitary, salivary glands, lacrimal glands, and sweat glands. In addition, the whole-body blood pool also showed moderate accumulation.
All 5 patients with recurrent ovarian cancer in this study had at least one recurrent lesion as diagnosed by 18F-FDG PET/CT and rising CA125 levels, and 68Ga-DOTA-TMVP1 PET/CT identified eight of 10 metastatic lesions with SUVmax ranging from 1.2 to 2.6 and SUVmean ranging from 0.9 to 1.9 (Supplementary Table S3). In 1 of the 8 patients with multiple metastases, including that in bone, pelvic, and liver, 68Ga-DOTA-TMVP1 PET/CT detected multiple pelvic and bone metastasis lesions (Fig. 5A). Three of the 5 patients with recurrent ovarian cancer showed that the 18F-FDG findings were more prominent than the 68Ga-DOTA-TMVP1 findings. However, in 2 patients with endodermal sinus tumor, the 68Ga-DOTA-TMVP1 PET/CT finding was equivalent to the 18F-FDG PET/CT finding (Fig. 6A and B). VEGFR-3 was stained positively in the surgical tumor tissues by IHC, which mainly located to the new lymphatic vessels (Fig. 6C).
The 3 patients with recurrent cervical cancer also underwent 68Ga-DOTA-TMVP1 and 18F-FDG PET/CT. Only one 68Ga-DOTA-TMVP1 PET/CT finding showed the lymph node metastatic lesion (Fig. 4B), concordant with the 18F-FDG PET/CT finding.
Recent studies have explored different targeted treatment modalities for cancers, with angiogenesis and lymphangiogenesis being the most studied targets for inhibition of tumor growth and metastasis (26). Several antiangiogenic inhibitors were developed for targeting sprout angiogenesis, with proven clinical benefit in various types of cancers (27, 28). Bevacizumab (Avastin), a humanized mAb that targets VEGF-A, has been registered for the treatment of advanced disease of colon, non–small cell lung, breast, and renal cancers and refractory glioblastoma (29, 30). However, because of a lack of lymphatic-specific markers, the study of lymphatics has historically lagged behind that of hemangiogenesis. Targeting drugs aimed at blocking lymphangiogenesis are rarely developed in cancer therapy (26). VEGFR-3 is one of the new lymphatic-specific markers, and VEGFR-3–targeting agents such as VEGFR-3 human IgG subclass 1 mAb LY3022856 and AD0157 have now been developed (31, 32).
In our current study, for developing VEGFR-3–targeting drugs, a five peptide sequence of TMVP1 was obtained by displaying the extracellular domain of recombinant active human protein VEGFR-3 to a random peptide library. TMVP1 is an attractive scaffold for PET imaging with high target affinity and rapid tumor uptake. Following the rationale, we successfully labeled TMVP1 with 68Ga according to the published protocols. Using two tumor xenograft models, we demonstrated that the radiotracer could target the tumor tissues as analyzed by the micro-PET imaging and investigated the biodistribution in vivo after injection of 68Ga-DOTA-TMVP1. 68Ga-DOTA-TMVP1 also showed favorable pharmacokinetics and a favorable dosimetry profile. Tracer elimination occurred through the renal system, and the tracer was cleared rapidly at early time points (30–60 minutes after injection). Taken together, the above results support the promising PET imaging prospects of 68Ga-DOTA-TMVP1.
In earlier studies, several kinds of VEGFR-3–targeting agents have been developed, including VEGFR-3 mAbs, anticalins, peptides, and small pharmaceutical molecules. Huhtala and colleagues had shown that the 111In-labeled anti-VEGFR-3 antibody mF4-31C1 had a strong accumulation of the radioactivity in the tumor area, 48 hours postinjection (33). A typical feature of antibodies is their slow drug metabolism, unlike small pharmaceutical molecules and peptides in vivo. Shi and colleagues screened peptide III (WHWLPNLRHYAS) targeting the extracellular fragment of recombinant human VEGFR-3/Flt-4 through a phage-displayed random peptide library (34), but the efficacy of VEGFR-3 imaging was lower than that of 68Ga-DOTA-TMVP1.
In the context of PET/CT imaging in the clinic, 68Ga-DOTA-TMVP1 had been tested in a first-in-man study in 5 patients with recurrent ovarian cancer and 3 patients with recurrent cervical cancer. No subjective effects were reported by any of five healthy volunteers and 8 patients with cancer after being injected with 68Ga-DOTA-TMVP1. Furthermore, the new tracer could localize to most of the recurrent multi metastatic lesions, as seen in the representative scans in Figs. 5 and 6. In most patients, the 18F-FDG findings were more prominent than the 68Ga-DOTA-TMVP1 findings. Impressively, in the two endodermal sinus tumor patients, 68Ga-DOTA-TMVP1 PET/CT finding was equivalent to 18F-FDG PET/CT finding (Fig. 6A and B). Therefore, the imaging agent for tumor detection is possibly not useful, 68Ga-DOTA-TMVP1 may be useful as a noninvasive radiotracer for assessing VEGFR-3 levels.
We consider the unavailability of enough patient samples and the lack of detection of VEGFR-3 expression in tumor tissues as the major limitations of our study; the latter because the patients were not suitable for surgical treatment. Only basic information on biodistribution and radiopharmaceutical localization could be collected. However, given the encouraging preliminary results of this pilot study, we plan to evaluate the correction of 68Ga-DOTA-TMVP1 imaging and VEGFR-3 expression in patients with primary ovarian cancer.
The KD values of natGa-DOTA-TMVP1 in this study, as measured by SPR assays, were in the range of 10−6 and 68Ga-DOTA-TMVP1 was moderately also up taken in tumors with a short retention time in blood (35). To improve the binding affinity, stability, and biodistribution of TMVP1, several strategies can be used to modify TMVP1, including PEGylation, lipidization, and multimerization (36, 37). These modifications may decrease renal clearance, enhance water solubility, and improve delivery efficiency (38). Future studies should focus on modifications of TMVP1 or the screening of novel peptides for VEGFR-3.
68Ga-DOTA-TMVP1 is a safe PET tracer with a favorable biodistribution. 68Ga-DOTA-TMVP1 could localize to most of the multi metastatic lesions in patients with recurrent ovarian and cervical cancer. Therefore, 68Ga-DOTA-TMVP1 may be useful as a noninvasive radiotracer for assessing VEGFR-3 levels. To the best of our knowledge, ours is the first clinical study to implicate the potential application of 68Ga-DOTA-TMVP1 in VEGFR-3 PET imaging, however, further research is required for evaluating the correlation of 68Ga-DOTA-TMVP1 imaging and VEGFR-3 expression.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: Fang Li, L. Xi
Development of methodology: J. Cai, Fang Li, L. Xi
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Fei Li, Z. Zhang, X. Chen, Y. Zhou, Q. Dong
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Fei Li
Writing, review, and/or revision of the manuscript: Fei Li, Y. Zhou, Fang Li
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Fei Li, Z. Zhang, X. Ma
Study supervision: Fang Li, L. Xi
We acknowledge the staff of the Department of Nuclear Medicine, Peking Union Medical College Hospital, for their help. The authors thank Yuan Yuan, Danmei Yan, Yun Dai, and Heng Cao for their help during the revision process. The authors thank Wuhan Yanjin Biotechnology Co., Ltd. for providing LSPR molecular interaction services. This work was supported by the grants from National Natural Science Foundation of China (81601526, 81802608, and 81472444).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.