Discovery of a novel highly specific, fully human PSCA antibody and its application as an antibody-drug conjugate in prostate cancer

Abstract

Prostate stem cell antigen (PSCA) is expressed in all stages of prostate cancer, including in advanced androgen-independent tumors and bone metastasis. PSCA may associate with prostate carcinogenesis and lineage plasticity in prostate cancer. PSCA is also a promising theranostic marker for a variety of other solid tumors, including pancreatic adenocarcinoma and renal cell carcinoma. Here, we identified a novel fully human PSCA antibody using phage display methodology. The structure-based affinity maturation yielded a high-affinity binder, F12, which is highly specific and does not bind to 6,000 human membrane proteins based on a membrane proteome array assay. F12 targets PSCA amino acids 63-69 as tested by the peptide scanning microarray, and it cross-reacts with the murine PSCA. IgG1 F12 efficiently internalizes into PSCA-expressing tumor cells. The antimitotic reagent monomethyl auristatin E (MMAE)-conjugated IgG1 F12 (ADC, F12-MMAE) exhibits dose-dependent efficacy and specificity in a human prostate cancer PC-3-PSCA xenograft NSG mouse model. This is a first reported ADC based on a fully human PSCA antibody and MMAE that is characterized in a xenograft murine model, which warrants further optimizations and investigations in additional preclinical tumor models, including prostate and other solid tumors.

Keywords: Antibody drug conjugate; PSCA; fully human antibody; monomethyl auristatin E; prostate cancer; tumor xenograft mouse model.

Figure 1.

Discovery and characterization of the PSCA antibody fab G7.(a) fab G7 binding to recombinant PSCA-Fc protein measured by ELISA. Bovine serum albumin (BSA) was used as a negative control. Experiments were performed in duplicate and the error bars denote ± SD, n = 2. (b) Kinetics of fab G7 binding to PSCA-Fc, as measured by Blitz. (c) Fab G7 binding to PSCA positive (PC-3-PSCA, Du-145-PSCA and HT1376 cells) and PSCA negative cells (PC-3, Du-145 and CHO-K1 cells) as tested by flow cytometry. An irrelevant fab (anti-SARS-CoV-2 fab ab1) was used as the isotype control. Fab G7 at the concentration of 500 nM was incubated with cells. (d-g) competition of fab G7 and fab F12 binding to HT1376 cell surface-associated PSCA by the recombinant PSCA-Fc protein (d and f) and by the murine PSCA antibody 7F5 (e and g). 500 nM of fab G7 or 200 nM of F12 was incubated with cells in the presence of gradient concentration of competitors. The bound fab G7 or F12 was detected by the pe-conjugated anti-flag tag antibody.

Figure 2.

Characterization of the affinity-maturated PSCA antibody F12. (a) the structure model of fab G7, represented as the cartoon model using PyMoL. The light chain and heavy chain CDR3 and FR2 hydrophobic residues are depicted as cyan and red sticks. The orange dotted lines highlight hydrophobic patches in antibody paratopes. (b) Evaluation of binding affinity to PSCA-Fc for affinity-enhanced clones by ELISA. Fab G7 is the parental clone. (c) The kinetics of fab F12 binding to PSCA-Fc, as measured by Blitz. (d) MPA assay to evaluate F12 binding specificity. IgG1 F12 was tested for binding to as many as 6,000 human transmembrane proteins that are transgenically expressed in HEK293 cells in a high throughput manner. (e) Epitope mapping of F12 by the conformational peptide scanning. The PSCA protein was scanned by cyclic peptides with 7, 10 and 13 amino acid length with peptide-peptide overlaps of 6, 9 and 12 amino acids. The conformational PSCA peptide microarray was framed by the HA control peptides. F12 binding was detected by the goat anti-human IgG (H+L) DyLight680 (red color), and the array outmost HA tag peptide was detected by the anti-ha (12CA5) DyLight800 (green color). (f) IgG1 F12 binding to murine PSCA recombinant protein by ELISA. Detection was achieved by hrp-conjugated anti-human fc antibody. Experiments were performed in duplicate and the error bars denote ± SD, n = 2.

Figure 3.

The internalization of IgG1 F12 into PSCA expressing cancer cells. (a). The scheme of the pH sensitive dye that emits fluorescence only at acidic endosome. (b-d). Evaluation of internalization of IgG1 F12 by using the pH sensitive dye and antibody conjugate. IgG12 F12 was conjugated with pHab dye by the lysine amine-nhs coupling. Different concentrations of F12-pHab were incubated with PSCA+ or – PC-3 and Du-145 cells for 12 or 24 hrs at either 37°C (B and C) or 4°C (d). After washing with 3× PBS, the cell fluorescence was recorded by the fluorescent plate reader, and the signal was normalized to the background. F12 internalization into PC-3 and Du-145 cells was presented in (b) and (c) respectively. (e) Comparison of IgG1 F12 internalization to that of benchmark antibody IgG1 m276, which targets CD276 (B7-H3). Both WT PC-3 and Du-145 cells intrinsically express CD276. (f) The internalization of F12-pHab conjugate can be outcompeted by the naked IgG1 F12 antibody on both PC-3-PSCA and Du-145-psca cells in a concentration dependent manner. The two-way ANOVA followed by Tukey correction was used for statistical analysis. ns: p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 4.

Generation and characterization of ADC IgG1 F12-MMAE. (a) the scheme of ADC IgG1 F12-MMAE. MMAE was conjugated onto IgG1 F12 by the nhs-amine coupling through a cleavable dipeptidyl linker and the PAB spacer. (b) The intact LC/MS spectrum of IgG1 F12-MMAE to determine DAR. The ADC exhibits a heterogenous distribution of DAR with an average DAR of 1.98. (c and d). The quality controls of ADC by SDS-PAGE and ELISA. (c). F12-MMAE and Ab1-mmae HC and LC integrity were checked on SDS-PAGE. (d).The binding to PSCA antigen by F12 was ascertained after conjugation with MMAE. (e and f) the in vitro cytotoxicity of IgG1 F12-MMAE, Ab1-mmae with comparison with the naked IgG1 F12 and Ab1 antibody and the linker-payload combination small molecule (vc-mmae). Compounds were incubated with PC-3-PSCA or PC-3 cells for 4-5 days. The cell viability was detected by the Promega CellTiter-Glo® luminescent cell viability assay.

Figure 5.

The in vivo pharmacokinetics (PK) and tolerability test. (a and b) PK of IgG1 F12 and ADC IgG1 F12-MMAE in NSG mice (n = 4). Blood were collected at the indicated time point, and total antibody (a) and ADC (b) were quantified by indirect ELISA; (c) the PK parameters including clearance rate, τ1/2 and AUC for IgG1 F12-MMAE and naked IgG1 F12 were obtained by fitting the plots of sera antibody concentration vs. time post injection using the noncompartmental pharmacokinetics software PK solutions 2.0. (d) Body weight change after a high single dose administration of naked antibody IgG1 F12 or ADC at 20 mg/kg (n = 3); (e) blood ALT level at 15 days post injection of a single dose of IgG1 F12 or ADC at 20 mg/kg. The two-way ANOVA followed by Tukey correction was used for statistical analysis. ns: p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 6.

In vivo therapeutic effects of ADC IgG1 F12-MMAE in prostate cancer PC-3-PSCA xenograft mouse model.(a) inhibition of tumor growth by F12-MMAE at the dose of 1 mg/kg and 3 mg/kg, 2XQ1W. The NSG mice were s.C. grafted with PC-3-PSCA cells (n = 5). ADC was intraperitoneally administered. Ab1-mmae was used as the isotype control. Tumor sizes were monitored twice a week; (b) the mouse survival across different treatment groups in (A). Mice were euthanized when the tumor volume exceeded 1000 mm3 or if the animals showed any signs of suffering. (c) The therapeutic efficacy of F12-MMAE when dosed at 6 mg/kg, 2XQ1W. After ADC washout, tumor relapsed. When tumors regrew to 140 mm³(38 days post-first injection), a single injection of 3 mg/kg of ADC was administered to evaluate the response of ADC to relapsed tumors. (d) The mouse survival across different treatment groups in (c). The two-way ANOVA followed by Tukey correction was used for statistical analysis of tumor size. The comparison of mouse survival curves was done by the log-rank test. ns: p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001.