Analyzing the Hypersensitivity of MMR-deficient Colorectal Cancers to mTOR Inhibition and the Response of Cancer Stem Cells

Project Summary Mutations in DNA mismatch repair (MMR) genes are causative in Lynch Syndrome (LS) and defective MMR drives tumorigenesis in a significant proportion (15 - 20%) of sporadic colorectal cancers. The loss of MMR not only increases genomic mutation rates, but also results in the resistance of tumors to conventional chemotherapeutic agents. To study the how the loss of MMR causes colorectal tumorigenesis and affects the response of intestinal tumors to treatment we developed a novel mouse line, termed VCMsh2Thu, carrying a “humanized” TgfβRII allele (Thu) that is susceptible to coding frameshift mutations during MMR-driven intestinal tumorigenesis. VCMsh2Thu is the first mouse model that develops colorectal cancers (CRCs) closely mimicking the histopathologic features of CRCs in LS patients. In preliminary studies, we found that while rapamycin only had a cytostatic effect on MMR-proficient (MMR+) CRC cell lines, it exerted a strong cytotoxic effect on MMR- deficient (MMR-) CRC cell lines. Importantly, rapamycin treatment induced the rapid and persistent regression of MMR(-) small and large intestinal tumors in our LS mouse models as well as of three MMR(-) patient-derived CRC xenografts (PDXs), but not of MMR(+) CRC PDXs. Interestingly, mutations in oncogenes such as KRAS or PI3KCA and tumor suppressor genes including APC or TRP53 so far did not affect the rapamycin response in these PDX models. The striking response of MMR(-) tumors was due to the failure to efficiently repair rapamycin-induced oxidative DNA damage, which resulted in autophagic cell death of differentiated tumor cells. We also observed that MMR(-) tumors contain treatment resistant “cancer stem cell niches” consisting of Lgr5+ and Bmi1+ cancer stem cells (CSCs) as well as Paneth cells that accumulate during rapamycin treatment. Importantly, preliminary studies in MMR(-) CRC organoids showed that rapamycin is effectively exported from CSCs and that inhibition of the P-glycoprotein transporter (P-gp) could provide a novel strategy for their elimination. In addition, similar to CRCs in human patients, the CRCs in VCMsh2Thu mice are characterized by increased expression of the stem cell marker Aldh1a3, suggesting an important role for Aldh1a3+ CSCs in MMR- tumorigenesis. A novel Aldh1a3-Cre knock-in allele (AC) allows us to inactivate MMR in Aldh1a3+ stem cells and study the effects on intestinal tumorigenesis and the rapamycin response. Based on these findings, in this application we propose to study the effect of mTOR inhibition on the prevention and treatment of CRCs in VCMsh2Thu preclinical mice and a panel of clinically well-annotated MMR(-) PDX models. In addition, we will test a novel strategy involving P-gp inhibition for the elimination of CSCs in MMR(-) mouse and human CRCs. Finally, we will perform an unbiased analysis of the CSC populations that persist during rapamycin treatment in mouse and human MMR(-) CRCs by single cell RNAseq (scRNAseq) analysis.