The scFv libraries are then transfected into yeast cells and labeled with either soluble OX40 or lysate from cells expressing OX40
The scFv libraries are then transfected into yeast cells and labeled with either soluble OX40 or lysate from cells expressing OX40. or OX40-expressing cells and OX40-encoding DNA vector. We found that all tested methods produce a unique diversity of scFv binders. However, when we reformatted forty-one of these scFv as full-length monoclonal antibodies (mAbs), we observed that mAbs identified using soluble antigen immunization with cell lysate sorting always bound cell surface OX40, whereas other methods had significant Dynemicin A false positive rates. Antibodies identified using soluble antigen immunization and cell lysate sorting were also significantly more likely to activate OX40 in a cellular assay. Our data suggest that sorting with OX40 protein in cell lysate is usually more likely than other methods to retain the epitopes required for antibody-mediated OX40 agonism. Keywords: OX40, humanized mouse antibody repertoires, deep sequencing, yeast display 1. Introduction Many antibody drugs bind to disease Dynemicin A targets expressed on cell surfaces. For example, antibodies may bind to the surface of tumor cells and induce antibody-dependent cellular cytotoxicity (ADCC). Conventionally, antibody drug discovery groups use either hybridomas [1] or phage display [2] to discover antibody drugs. Hybridomas are typically screened for cell surface binders using enzyme-linked immunosorbent assays (ELISAs) in 96-well plates [3]. Hybridoma methods, therefore, require expensive robotics to screen thousands of antibody candidates. Phage display has a much higher throughput, because billions-diverse phage libraries can be panned against cells affixed to well plates [4]. However, most therapeutic antibodies have been discovered in mice [5], perhaps due to difficulties with developability of artificial antibodies, such as low solubility binders discovered in phage display [6,7]. Recently, we invented a novel method for screening millions-diverse antibody repertoires using microfluidics, yeast display, and deep sequencing [8,9,10]. Our method leverages the developability advantages of naturally Dynemicin A paired antibodies with the massively parallel throughput of display technologies. Other groups later further validated our work with comparable methods [11,12]. However, our previously published methods required soluble antigen for both mouse immunization and fluorescence-activated cell sorting (FACS). This limitation excluded the possibility of using the method to identify antibodies against multi-pass transmembrane proteins, such as G-protein coupled receptors. Additionally, the requirement for soluble protein may lead to antibodies directed against spurious epitopes not present in the native conformation on the surface of target cells. OX40, or tumor necrosis factor receptor superfamily member 4 (TNFRSF4), is usually a costimulatory immune receptor transiently expressed on T cells which upregulates T cell activity upon binding to its ligand, OX40L. Therapeutic agonism of OX40 may increase T cell differentiation and tumor killing functions [13]. Agonism requires a ligand binding to OX40 in a way that generates complexes of crosslinked OX40 molecules on cell surfaces [14]. Although the crystal structure of OX40 binding to OX40L has been resolved [15], the specific epitopes required for agonism are not well understood. Development of novel therapeutic antibodies would benefit from a method that generates large panels of antibodies directed against a variety of OX40 epitopes that are bioavailable at the cell surface. To improve OX40 antibody discovery, we adapted our previously published methods [8,9,10] to test different immunization methods (cells versus soluble antigen) and different antibody selection methods (cell lysate versus soluble antigen). The cell lysate selection method was adapted from prior work [16,17], specifically by using a peptide tag rather than biotin to label the cell lysate. We synthesized forty-one monoclonal antibodies (mAbs) from the various methods and found that soluble OX40 soluble antigen immunization followed by sorting with cell lysate was most likely to identify antibodies that bind cell surface antigen and yielded more antibodies that activate OX40 in cellular assays. 2. Materials and Methods 2.1. Mouse Immunization and Sample Preparation All mouse Dynemicin A work CD350 was performed at Antibody Solutions (Sunnyvale, CA, USA) and overseen by a licensed veterinarian. All experiments were performed using mice from Trianni (San Francisco, CA,.