Novel strategy for selection of monoclonal antibodies against highly conserved antigens: phage library panning against ephrin-B2 displayed on yeast

Abstract

Ephrin-B2 is predominately expressed in endothelium of arterial origin, involved in developmental angiogenesis and neovasculature formation through its interaction with EphB4. Despite its importance in physiology and pathological conditions, it has been challenging to produce monoclonal antibodies against ephrin-B2 due to its high conservation in sequence throughout human and rodents. Using a novel approach for antibody selection by panning a phage library of human antibody against antigens displayed in yeast, we have isolated high affinity antibodies against ephrin-B2. The function of one high affinity binder (named as 'EC8') was manifested in its ability to inhibit ephrin-B2 interaction with EphB4, to cross-react with murine ephrin-B2, and to induce internalization into ephrin-B2 expressing cells. EC8 was also compatible with immunoprecipitation and detection of ephrin-B2 expression in the tissue after standard chemical fixation procedure. Consistent with previous reports on ephrin-B2 induction in some epithelial tumors and tumor-associated vasculatures, EC8 specifically detected ephrin-B2 in tumors as well as the vasculature within and outside of the tumors. We envision that monoclonal antibody developed in this study may be used as a reagent to probe ephrin-B2 distribution in normal as well as in pathological conditions and to antagonize ephrin-B2 interaction with EphB4 for basic science and therapeutic applications.

Figure 1. Selection, validation, and sequence of ephrin-B2-specific human single-chain antibodies.

(A) A schematic diagram of phage panning against antigens expressed in yeast display system . (B) Immunofluorescence flow cytometry measurements of protein and phage binding to yeast cells. Surface-displayed ephrin-B2 was detected by the binding of anti-Myc antibody (‘Myc’) as well as recombinant human EphB4-Fc (‘EphB4’) to yeast cells (top panel). Progressive enrichment of phage clones from first three rounds of panning (denoted as ‘1 st’, ‘2 nd’ and ‘3 rd’) was detected by antibody against His tag (bottom panel). Histograms drawn in shaded area and solid lines indicate antibody binding to uninduced and induced yeast cells, respectively. The percentage of phage clones with positive binding is indicated. (C) SDS-PAGE of scFv-EB1 (lane ‘1’) and scFv-EC8 (lane ‘2’). (D) Ephrin-B2 specific scFv binding to irrelevant yeast cells, yeast cells with expression of ephrin-B2 ectodomain, 293T cells, and 293T cells with transient expression of full-length ephrin-B2. Shown are the histograms of cells labeled with secondary antibody with (solid line) and without (shaded area) ephrin-B2 specific scFv as primary antibody. (E) Sequence alignment of scFv-EA6, scFv-EB1, and scFv-EC8. Complementarity determining regions (CDR), the beginning of immunoglobulin variable heavy (VH) and variable light (VL) chain domains, and the linker connecting VH and VL are noted. * indicates amino acids differ between scFvs.

Figure 2. Conversion of scFv-EC8 as a fusion to immunoglobulin Fc and functional test.

(A) SDS-PAGE images of EC8 resolved under reducing (R) and non-reducing (NR) conditions. (B) EC8 (solid line) binding to yeast cells, 293T with stable expression of ephrin-B2, and murine ovarian epithelium. The binding of isotype control is shown in shaded area. (C) Flow cytometry measurements of EC8 binding to 293T cells are shown in filled squares. First order Langmuir adsorption model was used to fit the data to estimate equilibrium dissociation constant (K_D). (D) Conformation speficifity of EC8 against eprhin-B2 was examined by flow cytometry with (‘+’) or without (‘−’) incubating cells either in 6 M guanidine hydrocholoride (‘GnHCl’) for 20 min or in elevated temperature at 80°C (‘Heat’) for 10 min. (E) Western blot image of immunoprecipitated ephrin-B2 from 293T cells with (‘+’) or without (‘−’) EC8 antibody, detected by rabbit ephrin-B2 polyclonal antibody. (F) Flow cytometry measurements of EC8 binding to ephrin-B1 and ephrin-A5 displayed on yeast. Labeling of uninduced yeast cells is shown in shaded histograms. (G&H) Competition assay. Relative binding of EphB4 (100 nM) to yeast cells expressing ephrin-B2, preincubated with varying concentrations of EC8, was measured by flow cytometry. Affinity purified human IgG was included as isotype control. n = 3 independent measurements. (I) Confocal microscopic images of surface-bound EC8 and internalized ones before and after membrane permeabilization of 293T cells. Scale bar = 10 µm.

Figure 3. Detection of ephrin-B2 expression in human cancer cell lines and tumor xenograft in mice.

(A) Flow cytometry measurements of EC8 binding to COLO205 and HCT116 cells (solid line) in comparison to the isotype control (shaded area). CHO cells with no ephrin-B2 expression were also included for comparison. (B) RT-PCR detection of ephrin-B2 expression in different cell lines. (C) Immunostaining of ephrin-B2 on human colon cancer xenografted in mice. Control denotes immunostaining without EC8 as a primary antibody. Tumor and stromal cells were indicated with arrowhead and arrow, respectively. Circle indicates murine endothelium stained with EC8. Scale bar = 20 µm.

Figure 4. Detection of ephrin-B2 expression in human tissue arrays.

(A&B) Immunostaining of ephrin-B2 expression in human tumor tissue arrays using EC8. Control denotes immunostaining without EC8 as a primary antibody. Tumor and stromal cells were indicated with arrowhead and arrow, respectively. Scale bar = 20 µm. PAC = Papillary Adenocarcinoma; AC = Adenocarcinoma; BR = Bronchus; AL = Alveoli; SC = Squamous Cell Carcinoma; SAC = Serous Adenocarcinoma. IDC = Nonspecific Infiltrating Duct Carcinoma. (C) Immunofluorescence staining on human colon tumor tissue demonstrating that EC8 (red) detects ephrin-B2 expressions in both cancer cells and tumor-associated vasculature (green). Blow up views of the two areas indicated with dashed box are shown in the right panel. Scale bar = 100 µm.