Guided selection of an anti-gamma-seminoprotein human Fab for antibody directed enzyme prodrug therapy of prostate cancer

Abstract

Background: The HAMA response is a major challenge when murine antibodies are repeatedly administered for antibody directed enzyme prodrug therapy in vivo. In this study we have achieved humanization of the anti-gamma-seminoprotein E(4)B(7) murine mAb by guided selection.

Methods: Using optimal Ig Fab primers, human Fd and CL gene repertoires were amplified by RT-PCR from PBMCs of prostate cancer patients. The human Lc gene repertoire was first paired with the murine Fd gene of E(4)B(7) mAb to construct a pComb3X hybrid Fab display library. This hybrid library was screened with purified gamma-seminoprotein antigen. The human Fd gene repertoire was then paired with the selected human Lc to construct a fully human Fab library. After four more rounds of panning, completely human Fab antibodies specific for gamma-seminoprotein were selected and further identified.

Results: First, using the E(4)B(7) Fd gene as a template, light chain shuffling was achieved by panning the hybrid library. Then, using the selected Lc as a template, a human Fab antibody against gamma-seminoprotein was produced through heavy chain Fd shuffling. Western blotting, ELISA, and flow cytometry results demonstrated that the resulting human Fab antibody resembled the parental E(4)B(7) mAb in that they both recognized the same epitope with similar affinities. Fluorescent cell staining and immunohistochemistry analysis further confirmed that this newly constructed human anti-gamma-seminoprotein Fab antibody indeed specifically bound prostate cancer cells and tissue.

Conclusions: Through guided-selection, we successfully produced a human anti-gamma-seminoprotein Fab antibody. This work lays the foundation for optimal antibody-directed enzyme prodrug therapy of prostate cancer using a fully human Fab antibody.

Fig. 1

Gel electrophoresis of PCR products amplified using the optimal family-specific primers. Aliquots of cDNA derived from the PBMCs of prostate cancer patients were amplified by PCR. The five Fd 3′ primers (specific for IgG1 (1–5), IgG2 (6–10), IgG3 (11–15), IgG4 (16–20), and IgM (21–25)) were used along with the same Fd 5′ primer set (HV1-HV5). The three Lc 3′ primers (specific for Igκ (26–30), Igλ1 (31–40), and Igλ2 (41–50)) were used with the same Lc 5′ primer set (KV1-KV5, LV1-LV10) as shown. The first lane of each panel (M) contains molecular weight markers (prominent bands 2,000, 1,000, 750, 500, 250 and 100 bp)

Fig. 2

Restriction enzyme digestion analysis of Fab libraries. Recombinant gene percentage of randomly selected clones from the human Lc (1–10) and Fd (11–20) libraries were determined by restriction with Xho I + Spe I and Sac I + Xba I, respectively. The first lane of each panel (M) contains molecular weight markers (prominent bands 2,000, 1,000, 750, 500, 250, and 100 bp)

Fig. 3

Detection of anti-γ-sm specific binding of soluble PIII-fusion Fab antibody by ELISA. Binding was determined by indirect-ELISA to γ-sm, GST, BSA, and OVA. Positive control: anti-γ-sm murine mAb E₄B₇; Non-related control: anti-hepatoma murine mAb HAb18; Negative control: PBS

Fig. 4

Analysis of the purified anti-γ-sm human Fab antibody product by SDS-PAGE and western blot. a Purification of the anti-γ-sm human Fab by affinity chromatography using a cross-linked protein G column. M Protein standards, L column load, FT column flow-through, b 10% SDS–PAGE analysis of purified anti-γ-sm human Fab. 1 reducing conditions, 2 non-reducing conditions. c 12% SDS–PAGE analysis of the purified γ-sm antigen. 3 γ-sm, untreated, 4 γ-sm, treated with PNGase F and Endo-H. d Western blot of untreated (5) and treated (6) γ-sm antigen with purified anti-γ-sm human Fab antibody used as the primary antibody

Fig. 5

Binding characteristics of anti-γ-sm human Fab antibody as measured by ELISA. a Inhibitory ELISA: serial dilutions of γ-sm were incubated with fixed concentrations of human Fab (or mAb E₄B₇). These pre-incubated mixtures were applied for the given time periods to ELISA plates containing immobilized γ-sm. The binding percentage was determined relative to the reactivity of the same number of Fab (or mAb) with immobilized γ-sm in the absence of competing antigen. Data points are the mean of two independent determinations. b Competitive inhibitory ELISA: mAb E₄B₇ at different concentrations (a = 1/1, b = 1/10, c = 1/100) competed with the Fab for binding to γ-sm, whereas the HAb18 mAb (d) did not change affect binding (e)

Fig. 6

Immunofluorescent staining of different cells with the newly selected anti-γ-sm human Fab antibody. Purified anti-γ-sm human Fab antibody (first column) and anti-hepatoma mAb HAb18 (second column) were used separately as the primary antibody (1:500 dilution, PBS as the control in the third column) to bind LnCap cells (first row), PC3M cells (second row), HHCC cells (third row), and NIH3T3 cells (fourth row). FITC-labeled goat anti-human or -mouse IgG (1:2,000 dilution) were used as corresponding secondary antibodies to visualize the stained cells ( × 400)

Fig. 7

Competive binding analysis of anti-γ-sm human Fab antibody and the parent mAb E₄B₇ by the FCM. LnCap (left column) and PC3M (right column) cells were labeled with the selected anti-γ-sm human Fab antibody at a constant concentration (1 μg/ml) followed by addition of E₄B₇ mAb at different concentrations (01, 1 μg/ml; 02, 0.6 μg/ml; 03, 0.2 μg/ml; 04, 0 μg/ml) for competitive binding (PBS served as the negative control)

Fig. 8

Immunohistochemical staining with selected anti-γ-sm human Fab antibody on HGPIN and prostatic adenocarcinoma. Strong staining was observed in high-grade prostatic intraepithelial neoplasia (A2), especially in Gleason pattern 3 (B2) and Gleason pattern 4 prostate cancer cases (C2) (A1, B1, and C1, HE staining × 200)