A Novel Fully Human Antibody targeting Extracellular Domain of PSMA Inhibits Tumor Growth in Prostate Cancer

Prostate cancer is the most commonly diagnosed malignancy and the second leading cause of cancer-related death in men (1). Localized prostate cancer can be treated with surgery or radiation, and recurrent disease can be temporarily controlled with androgen ablation. However, nearly 30%–40% prostate cancer will eventually become hormone refractory and then rapidly progress (2, 3). And hormone-refractory or androgen-independent prostate cancers are largely resistant to conventional chemotherapy (4). Thus, new treatment strategies for prostate cancer are urgently needed.

Antibody-based therapy is one of most attractive strategies in cancer therapy, because of the high binding affinity and specificity of an antibody. Because of the outstanding effectiveness in clinical cancer treatment, antibody-based antitumor strategies have been awarded as top one of science's breakthrough in 2013 (5). However, for successful therapies, it is of vital importance to find ideal targets and to generate novel antibodies that have high binding affinity and specificity (6, 7).

PSMA (prostate-specific membrane antigen) is a type II transmembrane protein, which contains three parts: a 19 amino acid internal portion, a 24 amino acid transmembrane portion, and a 707 amino acid external portion. PSMA is specifically expressed on prostate epithelial cells and strongly upregulated in all prostate cancers and its expression is much higher in poorly differentiated, metastatic, and hormone refractory cases (8–11). In addition, PSMA has also been found to express in tumor neovasculature in various kinds of cancer types, such as lung cancer, breast cancer, colon cancer, renal cell carcinoma, bladder cancer, thyroid cancer etc., but not in the normal vascular endothelial cells (12–17). Thus, PSMA has been regarded as an ideal target for not only prostate cancer, but also the cancer types with vascular PSMA expression (18–20).

More than a dozen of PSMA-specific antibodies have been developed and evaluated in the diagnosis and therapy of prostate cancer and other cancer types. However, most of these antibodies are mouse mAbs or humanized antibodies, which may bring safety concerns when applied in human. The antitumor effect of these antibodies is still under evaluation, and some did not show expected efficacy. For example, in a pilot trial of radio-labeled humanized J591 for the treatment of castrate metastatic prostate cancer, only 1 in 14 patients had a >50% decline in PSA level (21). Thus, more human PSMA antibodies with high binding affinity or therapeutic efficacy are still needed.

In our previous study, we have obtained a PSMA-specific single-chain variable fragment (scFv; named gy1) from a large yeast display naïve human scFv library, which can specifically bind with the extracellular domain of PSMA (22). In this study, we reconstructed this scFv into a fully human antibody (named PSMAb) and evaluated its characterization both in vitro and in vivo. We showed that PSMAb can specifically bind with and internalize into PSMA⁺ cells, and its binding affinity is at nanomolar level. We confirmed the specific localization and antitumor effect of PSMAb on PSMA⁺ tumors in vivo. We also explored the mechanism how PSMAb inhibits the growth of PSMA⁺ tumors. Our results confirmed that PSMAb deserves further study for PSMA-targeted therapy for prostate cancer and other cancer types with vascular PSMA expression.

Human prostate cancer cell lines C4-2, LNCaP, PC-3, and DU-145 were obtained from the ATCC. All cells were authenticated through STR profiling and confirmed that they were not contaminated by Mycoplasma. PC-3-PSMA⁺ and PC-3-PSMA⁻ cells that stably express PSMA and luciferase or luciferase only were prepared in our laboratory. Cells were maintained in RPMI1640 or DMEM or F-12K or Eagle Minimum Essential Medium (Gibco Life Technologies) supplemented with 10% FBS (Gibco Life Technologies) and 1% penicillin–streptomycin (Invitrogen Life Technologies), and cultured at 37°C with 5% CO₂ and 95% relative humidity in a humidified incubator. Suspension Chinese hamster ovary (CHO-S) cell was maintained in Chemically-defined CHO (CD CHO) medium supplemented with 40 mL/L 8 mmol/L l-glutamine in Erlenmeyer flasks (Falcon) and cultured at 37°C with 8% CO₂ in a Humidified Orbital Shaker (Thermo Fisher Scientific) rotating at 125–135 rpm.

The coding sequences for the heavy and light chain variable region were compared in IgBlast database to find the sequence with highest similarity, and corresponding constant region sequences of IgG1 were obtained from the website: http://www.imgt.org. The sequences of variable regions were then jointed with corresponding constant regions, optimized, and synthesized. The two sequences encoding the full heavy and light chain were cloned into the bicistronic eukaryotic expression vector Lh1, which was constructed in our laboratory. Successful construction of the PSMAb-expressing vector was confirmed by DNA sequencing. The full amino acid sequence of PSMAb is shown in Supplementary Fig. S1. The PSMAb-expressing vector was then transiently transfected into CHO-S cells using FreeStyle MAX Transfection Reagent (Invitrogen, Life Technologies). At day 7 after transfection, the supernatants were collected and centrifuged at 4°C at 5,000 rpm for 20 minutes, before they were filtered by 0.45-μm filter. The supernatant was then mixed with same volume of binding buffer (20 mmol/L sodium phosphate, pH 7.0). PSMAb was then purified using HiTrap rProtein A Fast Flow 5 mL Column (GE Healthcare) and AKTA FPLC System (GE Healthcare) according to the manufacturer's protocol. Characterization of PSMAb was analyzed by size exclusion chromatography-high performance liquid chromatography (SEC-HPLC). Briefly, PSMAb at the concentration of 1.0 mg/mL was filtered through 220-nm PES filter membrane, before it was loaded and passed through a P3000 column at the speed of 0.7 mL/minutes, and monitored at 280 nm by Viscotek TDAmax.

For SDS-PAGE analysis, the culture supernatants before purification and purified PSMAb either in denatured or nondenatured condition were separated by 12% SDS-PAGE and the gel was stained with Coomassie blue and destained with destaining buffer. For Western blot analysis, the SDS-PAGE separated proteins were transferred onto polyvinylidene difluoride membrane. The membrane was then blocked with 5% nonfat milk diluted in PBS for 2 hours at room temperature before being incubated with horseradish peroxide (HRP)-conjugated anti-human IgG Fc antibody (ab97225, Abcam) overnight at 4°C. The antibody bound proteins were visualized using a MultiImage Light Cabinet Filter Positions (Alpha Innotech Corporation).

Cells were harvested, washed, and resuspended in FACS buffer (PBS containing 0.2% BSA and 0.05% sodium azide) at the density of 1 × 10⁶ cells/mL. Cells were then incubated with 200 μL 50 nmol/L PSMAb for 30 minutes at 4°C, followed with washing and incubation with 100 μL PE-conjugated anti-human IgG Fc antibody (HP6017, BioLegend) for 30 minutes at 4°C in darkness. As positive control, these cells were also incubated with PE-conjugated anti-human PSMA IgG (LNI-17, BioLegend) for 30 minutes at 4°C in darkness. FACS buffer and human IgG (R&D System) were used as negative and isotype control. After being washed with FACS buffer, cells were analyzed by flow cytometry (BD Bioscience). For competitive binding assay, different amount of PSMA protein was incubated with PSMAb for 2 hours at room temperature before it was applied for the incubation with PSMA⁺ C4-2 cells and examined by flow cytometry.

PSMA⁺ C4-2 cells were plated in 96-well plates, which have been incubated with 2% gelatin (Sigma-Aldrich) for 30 minutes at 37°C, at the density of 1 × 10⁴ cells per well and cultured overnight. On the second day, cells were washed with cold culture medium, then incubated with serially diluted PSMAb or control IgG for 1 hour at 4°C before being washed with cold medium for 3 times, and incubated with HRP-conjugated goat anti-human IgG Fc antibody (Jackson ImmunoResearch) for 1 hour at 4°C. Cells were washed again with cold medium for 3 times, followed with PBS washing for another 3 times, and colorimetric signals were developed by the incubation with 3, 3′, 5, 5′-tetramethylbenzidine (eBioscience) for 30 minutes and stopped by 2 mol/L H₂SO₄ for 15 minutes. Absorbance was then measured at 450 nm using a Sunrise Microplate Reader (Tecan) and the K_d value was calculated using nonlinear regression analysis of a one-site binding hyperbola equation by GraphPad Prism 6.0 software.

C4-2 and PC-3 cells were plated in 24-well plates with coverslips at the density of 5 × 10⁴ cells per well and cultured overnight. On the second day, cells were incubated with 200 μL 100 nmol/L PSMAb or control human IgG for 2 hours at 37°C before being washed and fixed with 4% paraformaldehyde for 20 minutes. Cells were then permeabilized through incubation with 0.5% Triton X-100 in PBS for 10 minutes at room temperature. Then cells were incubated with 5% BSA for 1 hour at 37°C for blocking before the incubation with FITC-conjugated goat anti-human IgG (Santa Cruz Biotechnology) for 1 hour at 4°C in darkness. Cells were then stained with 4′,6-diamidino-2-phenylindole (DAPI) to visualize the nuclei. Finally, cells were washed with PBS and mounted on slides. Slides were observed and pictures were captured under laser scanning confocal microscope (FluoView FV1000, Olympus).

Thermostability of PSMAb was analyzed using VP-DSC Capillary Cell Microcalorimeter from MicroCal/Malvern Instruments, with scans from 10 to 100°C at a rate of 1.5°C/minute. Dialyzed PSMAb was diluted to 0.9 mg/mL and thoroughly degassed before being loaded into the calorimetric cell. The reference cell was filled with dialysis buffer. Data collection and processing were performed using the software included with the instruments.

BALB/c nude mice (4- to 6-weeks old, male, body weight 20–30 g) were purchased from the animal center in Fourth Military Medical University (Xi'an, China) and maintained under specific pathogen-free conditions. All the animal work was performed according to the protocol approved by the Guidelines for the Care and Use of Laboratory Animals of Fourth Military Medical University. The xenograft tumor models were developed by injecting 5 × 10⁶ firefly luciferase–expressing PC-3-PSMA⁺ or PC-3-PSMA⁻ cells in 0.2 mL PBS subcutaneously in the right flank of each mouse. Imaging study was performed when the tumor volume reached 100 mm³. For imaging study, PSMAb was labeled with near IR dye using IRDye800CW Labeling Kit (Li-Cor Biosciences). Extra dye was removed through dialysis. For each mouse, 0.2 μmol/kg of the IRDye800CW-labeled PSMAb was intravenously injected, and the mouse was anesthetized at indicated time points and the IRDye800CW fluorescence was monitored in a real-time manner under the Xenogen IVIS kinetic imaging system at an excitation wavelength of 745 nm. To analyze the biodistribution of PSMAb in different tissues, mice were sacrificed at 48 hours after IRDye800CW-labeled PSMAb injection and different tissues were isolated and fluorescence intensities were analyzed. Fluorescence intensities were calculated using living image software and presented as photon flux (p/s/cm²/sr).

For treatment study, 16 nude mice were injected with 5 × 10⁶ luciferase–expressing PC-3-PSMA⁺ cells in the left flank. When tumor volume reached 50–150 mm³, these mice were randomly divided into four groups (PBS, IgG 5 mg/kg, PSMAb 1 mg/kg, and PSMAb 5 mg/kg). PSMAb, control IgG, or PBS was injected into the tail veins of the tumor-bearing mice; each mouse was treated twice a week, for three weeks. Tumor diameters were measured with calipers every 3–5 days and tumor volumes were calculated using the formula 0.5 × long axis × short axis² and presented as the mean ± S.E. Tumor growth was also observed using bioluminescence imaging (BLI) every week. At the end of the experiment, mice were sacrificed and tumors were excised before tumor pictures were taken. Tumor tissues were collected to prepare paraffin sections for terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining to examine cell apoptosis, IHC staining of PSMA and Ki-67 to evaluate cell proliferation, and hematoxylin and eosin (H&E) staining. Important organs like heart, liver, spleen, lung, and kidney tissues were also collected to prepare paraffin sections for H&E staining. Serum was harvested to examine liver toxicity by checking the ALT (alanine transaminase) and AST (aspartate transaminase) level, and kidney toxicity by checking the BUN (blood urea nitrogen) and Cr (serum creatinine) level. For the treatment study using mice with both PC-3-PSMA⁻ and PC-3-PSMA⁺ xenografts, 12 nude mice were injected with 5 × 10⁶ luciferase–expressing PC-3-PSMA⁻ and PC-3-PSMA⁺ cells in the left and right flank, respectively. These mice were randomly divided into three groups (PBS, IgG 5 mg/kg, and PSMAb 5 mg/kg). PSMAb, control IgG, or PBS was injected into the tail veins of the tumor-bearing mice; each mouse was treated once a week, for 2 weeks. Tumor growth was observed using BLI every week; tumor diameters were measured with calipers, and tumor volumes were calculated every 3–5 days.

IHC staining was performed according to a previous protocol with minor modifications. Briefly, slides were deparaffinized in xylene before rehydration in a graded alcohol series, then treated with 3% H₂O₂ for 10 minutes to block endogenous peroxidase, followed by incubation with goat serum for 30 minutes at room temperature to block nonspecific binding. Slides were then incubated with anti-PSMA antibody (1:100, Cell Signaling Technology) at 4°C overnight in a humidified box, and then washed and incubated with HRP-conjugated anti-Rabbit antibody (Dako) for 30 minutes at room temperature. PSMA expression was visualized using 3,3′-diaminobenzidine chromogen staining for 2–3 minutes, followed by counterstaining with hematoxylin. After being dehydrated and cleared with Xylene, slides were mounted with mounting solution and images were captured under a microscope. Numbers of Ki-67–positive cells in different groups were determined through analyzing five random fields in each group under 400 × magnification and data were shown as mean ± S.D.

For the detection of apoptosis, paraffin-embedded sections of tumor tissues were labeled using a fluorescein-TUNEL Kit from Roche according to the manufacturer's instruction. Slides were mounted with Vectashield mounting solution containing DAPI to stain the nuclei. The slides were viewed and images were captured by laser scanning confocal microscopy (FluoView FV1000, Olympus).

Human or mouse peripheral blood mononuclear cells were isolated from human blood or spleens of BALB/c mice, and then natural killer (NK; effector, E) cells were isolated using MojoSort human or mouse NK Cell Isolation Magnetic Beads (BioLegend). C4-2 (target, T) cells were trypsinized, rinsed, suspended in PBS, and stained with CFSE (5-(and 6)-Carboxyfluorescein diacetate, succinimidyl ester) at 37°C for 15 minutes. Then C4-2 cells were washed with PBS and adjusted to the density of 2 × 10⁵/mL, different target/effector (E/T) ratio of cells were mixed and incubated at 37°C for 4 hours in the presence of PSMAb or control IgG before they were stained with propidium iodide (PI) for 15 minutes at 4°C in darkness to label the dying cells. Then cell cytotoxicity was measured using flow cytometry.

Complement-dependent cytotoxicity (CDC) was evaluated using CCK (cell counting kit)-8 staining. C4-2 and PC-3 cells were trypsinized and seeded in 96-well plates at the density of 1 × 10⁴ cells/mL and cultured overnight. On the second day, cells were incubated with different concentrations of PSMAb or control human IgG at 37°C for 2 hours in the presence or absence of guinea pig serum (1:20). Then CCK-8 was added and incubated at 37°C for 1 hour before absorbance was measured at 450 nm using a Sunrise Microplate Reader (Tecan).

Statistical analysis was performed using IBM SPSS statistical software (version 20.0). Quantitative data were presented as the mean ± S.D. or the mean ± S.E. and analyzed by ANOVA for repeated measurement data followed by post-hoc test. P < 0.05 was considered to be statistically significant.

The coding sequences for the heavy and light chain variable region were jointed with corresponding constant region of human IgG1 heavy chain and lambda light chain and synthesized, before they were constructed into the bicistronic eukaryotic expression vector Lh1. The construction, expression, and assembly of PSMAb are shown in Fig. 1A, and the full amino acid sequence of PSMAb is provided in Supplementary Fig. S1. The PSMAb-expressing plasmid was transiently transfected into CHO-S cells, one of the most widely used cell lines for antibody expression and production (23). Supernatants were collected at day 7 after transfection and PSMAb protein was purified by Protein A affinity chromatography. SDS-PAGE showed that in denatured condition, the molecular weight of heavy chain and light chain of PSMAb was around 55 and 25 kDa, respectively, whereas in nondenatured condition, the molecular weight of the whole PSMAb was 150 kDa (Fig. 1B). Expression and purification of PSMAb was also confirmed by Western blot analysis (Fig. 1C). Characterization of PSMAb was analyzed by SEC-HPLC, and results showed that PSMAb mainly existed as monomers (Fig. 1D). These results indicate that PSMAb was successfully expressed and purified.

To examine whether PSMAb can specifically bind with PSMA⁺ cancer cells, we used several prostate cancer cell lines, including C4-2, LNCaP, PC-3, and DU-145 cells. A commercial anti-PSMA antibody (LNI-17) was used to confirm the expression of PSMA in these cells. Results showed that PSMAb can specifically bind with C4-2 and LNCaP cells, which were demonstrated to have positive PSMA expression by the commercial anti-PSMA antibody. PSMAb did not bind with PSMA⁻ PC-3 and DU-145 cells (Fig. 2A). To evaluate the binding affinity of PSMAb, we performed cellular ELISA, and results showed that the binding affinity of PSMAb with C4-2 cells is pretty high, with the K_d value 0.48 nmol/L (Fig. 2B).

We then investigated whether PSMAb can internalize into PSMA⁺ cells. For this purpose, PSMAb and control human IgG were incubated with PSMA⁺ C4-2 cells or PSMA⁻ PC-3 cells for 2 hours at 37°C before immunofluorescent staining. Results showed that fluorescent signal can be clearly detected in C4-2 cells, but not in PC-3 cells (Fig. 2C). These results demonstrate that PSMAb can specifically bind with and internalize into PSMA⁺ cancer cells.

To evaluate the binding specificity of PSMAb, we performed competitive binding assay, in which different amount of PSMA protein was used to bind with PSMAb before it was incubated with PSMA⁺ C4-2 cells and examined by flow cytometry. Results showed that the binding of PSMAb with C4-2 cells gradually decreased after its incubation with increasing amount of PSMA protein, indicating that PSMAb has pretty good binding specificity with PSMA (Supplementary Fig. S2A).

Then we analyzed the thermostability of PSMAb using differential scanning calorimetry (DSC) assay. Result showed that the T_m (midpoint transition temperature) of PSMAb was 74.36°C (Supplementary Fig. S2B). Because PSMAb has traditional IgG1 structure, the T_m value indicated that PSMAb had pretty good thermostability. These results demonstrate that PSMAb has an ideal binding specificity and thermostability, which ensures its further application for in vivo study.

To investigate whether PSMAb can specifically enrich in PSMA⁺ prostate cancer tissues in vivo, we established PSMA⁺ and PSMA⁻ xenograft mouse models using luciferase-expressing PC-3-PSMA⁻ and PC-3-PSMA⁺ cells. PSMA expression in these two cells was confirmed by flow cytometry after they were incubated with PSMAb and PE-conjugated anti-human IgG Fc antibody (Fig. 3A). Stable luciferase expression was confirmed in both mouse models by Xenogen IVIS kinetic imaging system. PSMAb was labeled with IRDye800CW-NHS ester before it was injected into these mice through tail vein. The location of the near IR dye–labeled PSMAb was examined using whole-body near IR fluorescence imaging at different time points. Results showed that IRDye800CW-labeled PSMAb distributed rapidly throughout the whole body at 4 hours after injection and can be detected in PSMA⁺ tumor tissues from 24 hours. Then the IRDye800CW-labeled PSMAb was gradually cleared from the body due to metabolism but could still be observed in PSMA⁺ tumor tissues, even at 126 hours after injection. Specific accumulation was not observed in PSMA⁻ tumor tissues (Fig. 3B).

To further examine the biodistribution of PSMAb in different tissues, mice were sacrificed at 48 hours after IRDye800CW-labeled PSMAb injection and different tissues were isolated and fluorescence intensities were analyzed. Results showed that in mice with PC-3-PSMA⁺ xenograft, strong fluorescent signal can be detected in tumor tissue, relative weak signal can be detected in liver, while only negligible signal can be detected in other tissues. In contrast, in mice with PC-3-PSMA⁻ xenograft, fluorescent signal can only be detected in liver and negligible signal can be detected in tumor tissue and other tissues (Fig. 3C and D). These results indicate that PSMAb can specifically enrich in PSMA⁺ tumor tissues in vivo.

To examine whether PSMAb can inhibit tumor growth in vivo, PSMAb (1 mg/kg or 5 mg/kg), control IgG (5 mg/kg), or PBS was injected into mice with PC-3-PSMA⁺ xenografts, tumor growth was observed through BLI every week, and tumor volume was measured and tumor growth curve was generated. As shown by BLI observation and growth curve, tumor growth was inhibited by the treatment of PSMAb, compared with the control group, with the 5 mg/kg group showing better growth inhibition than the 1 mg/kg group (Fig. 4A–C). At the end of the experiment, mice were sacrificed and tumor tissues were excised to make paraffin sections for H&E staining, IHC staining for Ki-67 and PSMA, and TUNEL staining; important organs were also excised to make paraffin sections for H&E staining; serum was collected to examine liver and kidney function through the detection of ALT, AST, BUN, and Cr level. Results of H&E staining of the tumor tissues showed that the morphology was impaired in PSMAb-treated groups (Fig. 4D). IHC staining of Ki-67 showed that the number of Ki-67–positive cells decreased in PSMAb-treated group, especially in the 5 mg/kg group, indicating that cell proliferation was significantly inhibited in these xenografts. IHC staining of PSMA showed that PSMA expression decreased in PSMAb-treated group, yet still can be detected after treatment, indicating that repeated treatments may provide further inhibition on tumor growth (Fig. 4D and E). Results of TUNEL staining showed that fluorescent signal can be profoundly found in PSMAb-treated groups, with 5 mg/kg group showing more apoptosis than 1 mg/kg group. No obvious fluorescent signal was found in PBS or control IgG group (Fig. 4F). No obvious tissue toxicity was found in the important organs like heart, liver, spleen, lung, and kidney, as shown by the relative normal structure through H&E staining (Fig. 4G). Results of liver and kidney function showed that, in PSMAb-treated groups, the levels of ALT and AST were slightly increased, but has no statistical significance, and the levels of BUN and Cr did not increase (Fig. 4H). These results indicate that PSMAb has antitumor effect when applied in vivo, and the application of PSMAb did not bring obvious toxicity.

To further examine whether PSMAb can specifically inhibit the growth of PSMA⁺ tumors in vivo, we established a xenograft model in which PC-3-PSMA⁺ and PC-3-PSMA⁻ cells were injected in each side of the mouse. Then PSMAb (5 mg/kg), control IgG (5 mg/kg), or PBS was intravenously injected into these mice; tumor growth was observed through BLI observation every week, and tumor volume was measured every 3–5 days until the end of the experiment. BLI data showed that the growth of PSMA⁺ tumor was inhibited, whereas the PSMA⁻ tumor kept growing. The specific inhibition on PSMA⁺ tumor growth was also confirmed by growth curve. The tumor volume of PSMA⁺ tumors decreased more than 90% percent, whereas PSMA⁻ tumor did not show any reduction, compared with the control IgG group (Fig. 5A and B). These results demonstrate that PSMAb can specifically inhibit the growth of PSMA⁺ tumor, but not PSMA⁻ tumor.

To explore the mechanism how PSMAb inhibits PSMA⁺ tumor growth, we examined ADCC and CDC effect, because PSMAb contains traditional IgG1 constant region. For ADCC experiment, human or mouse NK cells were isolated and used as effector cells, and C4-2 cells labeled with CFSE was used as target cells. NK cells and C4-2 cells in different E/T ratios were incubated for 4 hours before they were stained with PI and examined by flow cytometry. Results showed that in the presence of PSMAb, both human NK cells and mouse NK cells can induce obvious cytotoxicity to C4-2 cells, and stronger cytotoxicity was detected in the groups with higher E/T ratios (Fig. 6A and B).

For CDC experiment, C4-2 or PC-3 cells were used as target cells, and CCK-8 staining was used to examine cell death. Target cells were incubated with different concentrations of PSMAb or control IgG in the presence of guinea pig serum for 4 hours before CCK-8 staining. Results showed that with the increase of PSMAb concentration, the survival of C4-2 cells decreased, whereas no obvious cell death was observed in the group of control IgG. No obvious cell death can be found in PC-3 cells (Fig. 6C). In addition, in the absence of guinea pig serum, no cell death was observed in both C4-2 and PC-3 cells (Fig. 6D). These results indicate that PSMAb may inhibit PSMA⁺ tumor growth through ADCC and/or CDC effect.

In recent years, antibody-mediated–targeted therapy is becoming more and more attractive in cancer therapy. However, for successful cancer therapy, it is of vital importance to identify specific tumor targets and to discover novel antibodies that can specifically recognize these targets.

PSMA is a highly prostate-restricted protein, and its expression has been shown to be upregulated in prostate cancer and correlated with tumor grade, pathologic stage, recurrence, and increase of PSA (8–11). In addition, PSMA has also been shown to be specifically expressed in the neovasculature of variable cancer types (12–17). Thus, PSMA is considered to be an ideal target for the therapy of not only prostate cancer, but also the cancer types with vascular PSMA expression (18–20).

So far, more than a dozen of PSMA-targeting antibodies have been developed. On the basis of the position of the epitopes they recognize, these antibodies can be divided into two categories. The first category is those that bind with the intracellular domain of PSMA, such as 7E11-C5.3 and PM2J004.5. These antibodies cannot bind with live cells, yet they can be used for the IHC staining of PSMA⁺ tissues (24, 25). The second category is those that bind with the extracellular domain of PSMA, such as J591, J415, 3E11, 3/A12, and 5D3. These antibodies can bind with live cells; thus, they are considered to be more suitable for PSMA-targeted therapy (26–29).

Some of these antibodies have been applied for diagnosis or therapy such as radiotherapy or ADC (antibody–drug conjugation) therapy in prostate cancer (30, 31). For example, 89Zr-labeled-J591, 89Zr-7E11, and 64Cu-labeled 3/A12 mAbs have been applied for targeted immunoPET in mouse models, all of which displayed high tumor-to-background tissue contrast, and thus they were considered to be used to quantify PSMA expression or to monitor tumor response in vivo (32–34). For radiotherapy, J591 mAb has been most intensively studied. Phase I studies confirmed the safety and biologic activity of ¹⁷⁷Lu-labeled J591 and ⁹⁰Y-labeled J591 in patients with androgen-independent prostate cancer (35, 36). In phase II study, single dose of ¹⁷⁷Lu-labeled J591 was found to be able to differentially reduce PSA level and improve survival in different patients with metastatic castration–resistant prostate cancer (37). In addition, the safety and specific targeting of ¹¹¹In-labeled J591 has also been confirmed in phase I trials in advanced solid tumors with vascular PSMA expression (38, 39).

However, these antibodies were mouse mAbs, and their murine origin may cause serious side effects when applied in human, thus severely restrains their further application in clinic. To reduce immunogenicity, humanized mAbs, for example, humanized J591 was generated, yet until now the humanized antibody has only been evaluated in tumor imaging in animal model and in ADC study, in which hJ591 was conjugated with DM1 (Maytansinoid-1) and named MLN2704. In addition, results of clinical trials have shown that MLN2704 has limited activity and the deconjugation of DM1 may lead to neurotoxicity (40–42). It is possible that the mouse origin of the variable region of the humanized antibody may also cause immunogenic side effect.

Thus, fully human antibodies are safer because they have negligible immunogenicity. There is one group that generated a fully human PSMA-specific antibody using XenoMice, in which the murine Ig gene locus was replaced with human Ig gene locus (43). This antibody has been applied in the study of ADC, in which the antibody was conjugated with monomethyl auristatin E, and showed potent antitumor effect in mouse models and patient-derived xenograft models (44–46). This antibody was also applied in the study of antibody–photoabsorber conjugate, in which the antibody was conjugated with the photoabsorber, IR700DX, and also showed antitumor effect in xenograft nude mice model (47). However, their antitumor effect still needs to be evaluated in clinic.

Besides XenoMice, antibody libraries provide another feasible way to directly isolate fully human antibodies (48), and yeast display antibody library is one of the platforms for human antibody isolation. In our previous study, we have constructed a naïve human scFv library with large capacity (1 × 10¹¹) and have obtained a PSMA-specific scFv (named gy1) by screening the library using the extracellular domain of PSMA. We have also confirmed that this scFv has high binding affinity and can specifically internalize into PSMA⁺ cancer cells through the endosome–lysosome pathway (22). In this study, we reconstructed the PSMA scFv into a fully human antibody PSMAb and confirmed that PSMAb can specifically bind with and internalize into PSMA⁺ cells. The affinity of PSMAb is 0.48 nmol/L, which is much higher than the intensively studied J591 antibody, whose affinity with native PSMA is 1.4 nmol/L as examined by ELISA, and 0.75 nmol/L when examined by surface plasmon resonance (49). In addition, we also confirmed that PSMAb can specifically enrich in PSMA⁺ tumors and significantly inhibit the growth of PSMA⁺ tumors in vivo.

To explore the potential mechanism how PSMAb inhibits tumor growth, we performed ADCC and CDC assay, and found that PSMAb can induce ADCC effect in the presence of NK cells, and can induce CDC effect in the presence of complement. Thus, we assumed that PSMAb may inhibit tumor growth through ADCC and/or CDC mechanisms. There is another report that showed that PSMA antibody has antitumor effect in vivo, and they found that the epitope (residues 638–657) in PSMA is important, only antibodies that specifically recognize this epitope have antitumor effect (50). However, whether our PSMAb also recognize the same epitope has not been identified and further study is still needed to find out the detailed mechanism how PSMAb inhibits tumor growth.

Besides therapeutic application, PSMAb may also be applicable in PET imaging for the diagnosis of prostate cancer and other cancer types with vascular PSMA expression. The applicability of PSMAb to be used in ADC study is also guaranteed, because PSMAb has pretty high binding affinity, and can be specifically internalized into PSMA⁺ cells. However, further investigations are still needed to explore its potential application in these fields.

In our animal study, no obvious toxicity was observed when PSMAb was applied in vivo, shown by the relative normal liver and kidney function and normal structure of important organs shown by H&E staining. According to phase I trials, the application of ¹¹¹In-labeled humanized J591 resulted in minimal toxicity in patients, however, hepatic saturation seemed to occur by 60 mg (38, 39). In addition, the application of MLN2704, the ADC that is composed of humanized J591 and DM1, induced toxicities in 20% of patients, such as elevated hepatic transaminases, fatigue, nausea, diarrhea, and neurotoxicity (41, 42). It is for sure that formal safety studies are needed before the application of PSMAb in patients.

In summary, we have successfully reconstructed the anti-PSMA scFv gy1 into a fully human antibody PSMAb, and we have demonstrated that PSMAb can specifically bind with PSMA⁺ cells and has high binding affinity. We have also showed that PSMAb can specifically enrich in PSMA⁺ tumors and inhibit their growth in vivo. And PSMAb may inhibit tumor growth through ADCC and/or CDC mechanisms. All these results indicate that the novel fully human antibody, PSMAb, deserve further study to be used in PSMA-targeted diagnosis and therapy for prostate cancer and other cancer types with vascular PSMA expression.

No potential conflicts of interest were disclosed.

Conception and design: J. Wu, Y. Han, A. Zhao, W. Qin, W. Wen

Development of methodology: D. Han, Y. Han, F. Yang, A.-G. Yang, A. Zhao

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Wu, D. Jiao, L. Zhang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Wu, D. Han, S. Shi, G. Zheng, M. Wei, G. Li

Writing, review, and/or revision of the manuscript: J. Wu, W. Wen

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Shi, G. Li, P. Xie, A.-G. Yang, W. Qin, W. Wen

Study supervision: Q. Zhang, W. Qin, W. Wen

This work was partly supported by grants from the National Natural Science Foundation of China (grant no. 81372225, to W. Wen, and grant nos. 81372771 and 81772734, to W. Qin) and the Natural Science Foundation of Shaanxi Province (2018SF-215, to G. Zheng). We thank Mrs. Yunxin Cao and Mr. Jintao Hu (Department of Immunology, Fourth Military Medical University, China) for flow cytometry analysis; Dr. Changhong Shi (Laboratory Animal Center, Fourth Military Medical University, China) for the xenograft model experiment; Dr. Yi Wan (Department of Health Services, School of Public Health, Fourth Military Medical University, China) for statistical analysis; and Dr. Min Yang for DSC and SEC-HPLC analysis (Malvern panalytical, Shanghai, China).

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.