Identification of anti-CA125 antibody responses in ovarian cancer patients by a novel deep sequence-coupled biopanning platform

High-grade epithelial ovarian cancer kills more women than any other gynecologic cancer and is rarely diagnosed at an early stage. We sought to identify tumor-associated antigens (TAA) as candidate diagnostic and/or immunotherapeutic targets by taking advantage of tumor autoantibody responses in individuals with ovarian cancer. Plasma-derived IgG from a pool of five patients with advanced ovarian cancer was subjected to iterative biopanning using a library of bacteriophage MS2 viruslike particles (MS2-VLPs) displaying diverse short random peptides. After two rounds of biopanning, we analyzed the selectant population of MS2-VLPs by Ion Torrent deep sequencing. One of the top 25 most abundant peptides identified (DISGTNTSRA) had sequence similarity to cancer antigen 125 (CA125/MUC16), a well-known ovarian cancer-associated antigen. Mice immunized with MS2-DISGTNTSRA generated antibodies that cross-reacted with purified soluble CA125 from ovarian cancer cells but not membrane-bound CA125, indicating that the DISGTNTSRA peptide was a CA125/MUC16 peptide mimic of soluble CA125. Preoperative ovarian cancer patient plasma (n = 100) was assessed for anti-DISGTNTSRA, anti-CA125, and CA125. Patients with normal CA125 (<35 IU/mL) at the time of diagnosis had significantly more antibodies to DISGTNTSRA and to CA125 than those patients who had high CA125 (>35 IU/mL). A statistically significant survival advantage was observed for patients who had either normal CA125 and/or higher concentrations of antibodies to CA125 at the time of diagnosis. These data show the feasibility of using deep sequence-coupled biopanning to identify TAA autoantibody responses from cancer patient plasma and suggest a possible antibody-mediated mechanism for low CA125 plasma concentrations in some ovarian cancer patients.