Development of a novel mammalian cell surface antibody display platform

Abstract

Antibody display systems have been successfully applied to screen, select and characterize antibody fragments. These systems typically use prokaryotic organisms such as phage and bacteria or lower eukaryotic organisms, such as yeast. These organisms possess either no or different post-translational modification functions from mammalian cells and prefer to display small antibody fragments instead of full-length IgGs. We report here a novel mammalian cell-based antibody display platform that displays full-length functional antibodies on the surface of mammalian cells. Through recombinase-mediated DNA integration, each host cell contains one copy of the gene of interest in the genome. Utilizing a hot-spot integration site, the expression levels of the gene of interest are high and comparable between clones, ensuring a high signal to noise ratio. Coupled with fluorescence-activated cell sorting (FACS) technology, our platform is high throughput and can distinguish antibodies with very high antigen binding affinities directly on the cell surface. Single-round FACS can enrich high affinity antibodies by more than 500 fold. Antibodies with significantly improved neutralizing activity have been identified from a randomly mutagenized library, demonstrating the power of this platform in screening and selecting antibody therapeutics.

Figure 1

Schematic illustration of Homologous Recombination by the Flp-In^™ System. Two expression cassettes for heavy and light chains were inserted into pcDNA5/FRT to form vector FV. Vector pOG44 contained the Flp recombinase gene. After co-transfection of both vectors FV and pOG44, the Flp recombinase expressed from pOG44 catalyzes a homologous recombination event between the FRT sites of the Flp-In host cell and the vector FV. Then the whole vector is inserted by homologous integration into the genome of the host cell only at the location of FRT.

Figure 2

Western blot analysis of antibody expression in single cell clones. The selected single stable-cell clones were expanded in selection medium which was replaced with serum-free medium. Day 4 serum-free conditioned medium (CM) from each clone was collected and analyzed by SDS-PAGE. After electrophoretic separation, the proteins were transferred onto PVDF membrane, stained by anti human-Fc monoclonal antibody and visualized by SEC technology (Amersham). (A) FCHO cell clones stably transfected by a mixture of both vectors FV and pcDNA5/FRT-Fc-Fusion. (B) Regular CHO cell clones stably transfected by vector pcDNA5/FRT-FC-Fusion. (C) FCHO cell clones stably transfected by vector pcDNA5/FRT-FC-Fusion.

Figure 3

Schematic illustration of mammalian display system. The PDGFR trans-membrane domain was C-terminally fused to the heavy chain constant region in vector FV to form the antibody display vector FVTM. After transfection of the vector FVTM into Flp-In cells, both heavy and light chains were expressed, assembled and displayed on the cell surface.

Figure 4

Flow cytometry analysis of antibodies displayed on the cell surface. Flp-In 293 cells were transiently transfected by FVTM containing the same human antibody gene (B–E) or a library of antibody genes (F). Cells were labeled 48-hrs post transfection with fluorescence-conjugated antibodies and/or antigen and then analyzed by flow cytometry. Kappa-PE, Phycoerythrin (PE)-conjugated mouse anti-human Kappa chain antibody; FC-FITC, FITC-conjugated mouse anti-human IgG antibody; Antigen-FITC, FITC-conjugated specific antigen. The x- and y-axis indicate the fluorescence intensities of FITC and PE fluorophores respectively. The cells carrying vector without antibody genes were used as the control (A).

Figure 5

Characteristics of antibody display on cell surface. FCHO cells stably transfected with antibody display vectors directing the surface expression of antibodies were co-stained with PE-conjugated anti-kappa chain antibody and FITC-conjugated antigen and then analyzed by flow cytometry. The concentration of anti-kappa chain antibody was fixed and the concentration of the antigen was titrated in a range of 6.25 pM–62.5 nM. Neg is the negative control and the cells were stained by neither antibody nor antigen. (A) Dose dependent binding of antigen to antibody-displaying cells. The X-axis indicates the fluorescence strength of FITC and Y-axis represents the fluorescence strength of PE. (B) Analysis of cell populations displaying antibodies with different affinities. The RBA for each antigen concentration was calculated and the data for both Ab-1 and Ab-2 are presented as antigen dose dependent binding curves.

Figure 6

Enrichment of cells displaying high affinity antibodies. (A) A mixture of cell populations displaying Ab-1 and Ab-2 were combined with parental untransfected Flp-In CHO cells at a ratio of 50:50. The ratios of Ab-1 and Ab-2 in final cell suspension were 1:10, 1:100 or 1:1,000 as indicated. The final cell mixtures were double stained by 100 nM of FITC-antigen and PE-anti-kappa chain antibody and then analyzed by flow cytometry. (B) 0.17% of the cells displaying Ab-1 were sorted from the mixture by FACS for expansion and re-analysis. (C) The sorted cells were re-analyzed by flow cytometry after 10 days expansion.

Figure 7

Strategy of combinatorial mutagenesis of VH domain. Fragments 1, 2 and 3 were amplified separately using primer pairs as indicated, and the three fragments were assembled to form the mutagenized full-length heavy chain library using sequence overlap extension PCR with primers P1 and P6. The library was then inserted into the display vector using the HindIII and BamHI cloning sites.

Figure 8

Sorting of the library for high affinity clones. An FCHO cell pool stably transfected with mutant library was double-labeled using both PE-conjugated anti-Lambda chain antibody and FITC-conjugated OX40 ligand (100 nM) and analyzed by flow cytometry. Cells located in gate R3 (0.1%) were single-cell sorted into a 96-well plate with one cell per well and expanded for further analysis.