LAUSR

Pancreatic cancer is a devastating disease ranked as the 4th leading cause of cancer-related deaths in the United States, and its incidence rate is increasing according to the latest statistics. The overall survival rates for patients with pancreatic cancer have not significantly improved over the past thirty years (Siegel et al., 2012; Simard et al., 2012). One of the reasons for the high mortality rates is the high resistance of pancreatic cancer to chemotherapy and radiation. Most patients are diagnosed at late stages of the disease. Approximately 15%–20% of patients diagnosed with pancreatic cancer are eligible for surgical resection, and 85% of these patients eventually experience relapse and ultimately cancer-related death (Siegel et al., 2012). In recent years, increasing evidence indicates that the fibro-inflammatory stroma is a source of cellular and molecular components contributing to tumor progression and metastasis (Feig et al., 2012; Waghray et al., 2013). Importantly, increased levels of stroma are positively related to a poor prognosis (Erkan et al., 2008). Despite the broader understanding of pancreatic cancer biology, gemcitabine, a chemotherapeutic approved for pancreatic cancer treatment approximately twenty years ago, still remains the standard of care (Burris et al., 1997). Thus, the development of novel treatment strategies for this devastating disease is urgently needed. Immunotherapy based on T cells modified with a chimeric antigen receptor (CAR) has been demonstrated to be a promising strategy for cancer treatment. CAR T cells specifically recognize tumor-associated antigens and eliminate tumor cells in a non-major histocompatibility complexrestricted manner. Several pilot clinical trials using CAR T cells have recently been reported to have promising clinical outcomes, even in solid tumors (Brown et al., 2016; Kershaw et al., 2013). Mesothelin (MSLN) is a membrane protein that is overexpressed in many cancer types, including pancreatic cancers, and is expressed only at low levels on normal peritoneal, pleural, and pericardial mesothelial surfaces (Chang and Pastan, 1996). Previously, several types of MSLN-targeted CAR-T cells were developed and have been found to have impressive antitumor activities in mesothelioma and ovarian cancer models (Carpenito et al., 2009; Lanitis et al., 2012). However, there are no reports on the antitumor activities of anti-MSLNCAR-T cells toward pancreatic tumor xenograft models. No study has yet examined the use of CAR T cells in PDX models of pancreatic cancer. Therefore, it is necessary perform a preclinical evaluation of novel CAR T cells as a treatment for pancreatic cancer in PDX models. In this study, we developed a novel fully human anti-mesothelin antibody. To investigate the binding properties of anti-MSLN antibody, we fist established the MSLN-overexpressed cell lines CHO-K1-MSLN and PANC-1-MSLN. The expression of mesothelin in these two established cell lines was confirmed by Western blotting (Fig. 1A). The fully human anti-MSLN antibody was screened from a fully human naïve antibody library by using phage display technology. The binding specificity of the anti-mesothelin antibody was tested on CHO-K1-MSLN and PANC-1-MSLN cells. The scFv proteins of anti-mesothelin antibody were produced transiently in FreeStyleTM 293F cells and purified by protein A affinity chromatography (Fig. S1). The results in Fig. 1C indicated that P1A6E and P3F2 scFv bound specifically to MSLN-expressing cells but not to cells without MSLN expression. Additionally, we compared the two fully human antibodies P1A6E and P3F2 with the SS1 and C10 antibodies. SS1 and C10 have a high binding affinity to mesothelin (Chowdhury and Pastan, 1999), and SS1 has also been found to be safe in patients when administered as a recombinant immunotoxin (Hassan et al., 2007). The results indicated that P1A6E and P3F2 had a significantly higher binding affinity than SS1 and C10 to MSLN-expressing cells (MFI value in PANC-1-MSLN cells: scFvP1A6E: 327.5, scFv-P3F2: 308.8, scFv-SS1: 48.9 and scFvC10: 46.8; MFI value in CHO-K1-MSLN cells: scFv-P1A6E: 452.3, scFv-P3F2: 445.1, scFv-SS1: 65.5 and scFv-C10: 80.2). The mean fluorescence intensity (MFI) of different scFv proteins bound cells as determined by flow cytometric analysis is shown in Fig. 1B. To test the affinity of antibody binding to mesothelin, we used Biacore Surface Plasmon resonance (SPR). The binding sensorgrams were collected at 25°C. The data were double-referenced by using