PRSS contributes to cetuximab resistance in colorectal cancer

Tan, Zhaoli; Gao, Lihua; Wang, Yan; Yin, Huihui; Xi, Yongyi; Wu, Xiaojie; Shao, Yong; Qiu, Weiyi; Du, Peng; Shen, Wenlong; Fu, Ling; Jia, Ru; Zhao, Chuanhua; Zhang, Yun; Zhao, Zhihu; Sun, Zhiwei; Chen, Hongxing; Hu, Xianwen; Xu, Jianming; Wang, Youliang

As cetuximab is mainly and extensively used to treat patients with mCRC, we used ELISA to examine PRSS1 levels in the serum of healthy individuals (n = 64) and patients with mCRC (n = 156) to determine the clinical significance of the PRSS1 level in mCRC ( Fig. 6A and data file S1). The serum PRSS1 levels of the patients (average, 59.43 ng/ml) were significantly higher than those of the healthy individuals (average, 34.00 ng/ml) ( fig. S7A ). Of the 156 patients, 52 received cetuximab monotherapy ( Fig. 6A and data file S1). The serum PRSS1 levels of the patients with primary cetuximab resistance (n = 20; average, 94.57 ng/ml) were significantly higher than those of the cetuximab-responsive patients (n = 32; average, 59.31 ng/ml) (Fig. 6B) . We also found that the patient serum PRSS1 levels before cetuximab treatment were significantly higher than those after cetuximab treatment (when disease progressed) ( Fig. 6C and data file S1). However, a difference was not observed between responders and nonresponders treated with chemotherapy alone ( fig. S7B and data file S1), and no difference was observed between the PRSS1 levels before chemotherapy and the PRSS1 levels after chemotherapy ( fig. S7C and data file S1). On the basis of whether the PRSS1 level exceeded the average PRSS1 level of all patients, we divided the patients receiving cetuximab monotherapy into two groups: high and low PRSS1 expression groups. Kaplan-Meier curves for these patients consistently demonstrated much worse overall survival (OS) for the patients with high PRSS1 than that for the patients with low PRSS1 median survival, 83 days versus 206 days, respectively; P = 0.009; Fig. 6D ), indicating that a high PRSS1 level was clearly related to a poor cetuximab response in mCRC. We conducted further analyses to determine whether the prognostic impact of PRSS1 expression is independent of other clinical variables. Fig. 2I , which was normalized to the heavy chain before cleavage. All values are the means ± SD from three independent experiments. Differences in growth were determined using Student's t test and by calculating subsequent P values. Data are the means ± SD (n = 5 cultures); **P < 0.01 and ***P < 0.001. Error bars represent the SD. Data shown are from experiments that were repeated three times with similar results. We pooled the patients who received cetuximab monotherapy with available PFS data (n = 52) for univariate and multivariate analyses of factors affecting PFS (table S2) . Multivariate analysis showed that the effect of PRSS1 expression on PFS was independent of other clinical variables (table S2) .