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

From a prognostic perspective, aberrant PRSS1 expression correlates significantly with the cetuximab response rate in patients with mCRC. We found that patients with better cetuximab responses had significantly lower serum PRSS1 levels than patients with cetuximab resistance and that the speed of cetuximab cleavage was also much slower in patients with good responses than that in patients with worse responses to cetuximab treatment, further demonstrating that PRSS1 may cause mAb resistance. We also found that the serum PRSS1 level increased after cetuximab administration and verified that anti-EGFR drugs induced PRSS1 secretion events specific to drug responses or resistance. Furthermore, ELISA showed that patients with mCRC had significantly higher serum PRSS1 levels than healthy individuals ( fig. S7A ) and that patients with cetuximab resistance had significantly higher serum PRSS1 levels than cetuximabresponsive patients (Fig. 6B) . Kaplan-Meier curves consistently demonstrated much worse survival for patients with high PRSS1 expression than that for patients with very low PRSS1 expression (Fig. 6D) , indicating that PRSS1 levels are clearly related to poor responses to cetuximab in mCRC. Analyses of patients with other cancers who received antibody monotherapy revealed that all patients with low PRSS1 or PRSS3 expression had longer PFS than patients with high PRSS1 or PRSS3 expression regardless of whether the patients had HNSCC (Fig. 6E) , colorectal adenocarcinoma (Fig. 6F), glioma (fig. S8C ), or ovarian serous cystadenocarcinoma ( fig. S8E ). Among these patients, no significant difference in OS was noted between the high-and low-PRSS1 or PRSS3 expression groups ( fig. S8, A, B , D, and F). Together, our findings indicated that high PRSS expression levels were associated only with poor responses to mAb treatment and not with a poor overall cancer prognosis. Last, the preliminary validation of our findings in the serum of patients with mCRC suggested that the level of secreted PRSS1 correlated with the response to cetuximab treatment ( fig. S8G and data file S1), indicating that PRSS1 may be a better predictive marker of the cetuximab response than tumor markers such as CEA and CA19-9, although more work is needed to assess the full potential of this biomarker.