The results obtained during the two years postdoc fellowship (01/09/17-31/08/2019) are explained in detail:
We have evaluated RNF43 and ZNRF3 mutational status in 40 tumor samples from 20 Lynch syndrome patients, including intestinal adenomas, colorectal carcinomas and endometrial tumors. This study indicated that Lynch syndrome tumors presented a high frequency of mutations in RNF43 and ZNRF3 genes, particularly in CRC lesions but not in benign malignancies (e.g. adenomas) or other tumors developed by the same patient (e.g. endometrial cancer). More importantly, RNF43G659fs mutation was presented in 9 out of 10 CRC samples analyzed. This mutation is placed in a codifying microsatellite with 7 consecutive guanines, a location previously described by us and Levi Garraway´s group as the most prevalent hotspot in sporadic hypermutant CRC tumors. In order to confirm our results we collected 25 CRC tumor samples from 25 new Lynch Syndrome patients and studied by haloplex the mutational status of a panel of 347 tumor suppressor and oncogenes, including RNF43 and ZNRF3 genes.
On the other site, we established cultures of 3D organoids derived from normal human colon tissue as well as SV40 T antigen immortalized human colonocytes (HCEC). We generated 293T expressing soluble R-Spondin 1 (293T-RSPO1mCh) and 293T expressing noggin (293-nogginmCh) essential to produce conditioned media to grow our organoids. Immunochemistry staining of phalloidin and Ki67 showed the structure and proliferation rate of our organoids. In addition, we generated a biobank of human organoids derived from colon samples from the Hospital with more than 15 samples.
We used CRISPR/Cas9 genome editing technology to knock out MSH2, MSH6 or MLH1 genes in HCEC and intestinal organoids. Exome sequencing from HCEC-MSH2ko and HCEC-MSH6ko cells showed high acquisition of frameshift and non-silent mutations compared with controls.
Moreover we also established and characterized the MSH2ko mouse model that recapitulates Lynch syndrome. This mouse model resulted in intestinal lesions at 6-8 months that led to invasive intestinal tumors at 8-12 months and death of the mice between 10-16 months (median of survival 11.5 months). The treatment of MSH2ko bearing tumor mice (positive signal by ultrasonography) with porcupine inhibitor WNT974 or tankyrase inhibitor (G007-LK), depending on the upstream or downstream mutations on Wnt signaling pathway, in combination with the checkpoint inhibitor PD-1 did not resulted in any statistically significant decrease in number or size of the tumors in treated mice compared with controls. The analysis of immune infiltration of the tumor did not show any statistically significant difference in the percentage of CD3, CD4 or CD8 positive cells between treated and control mice.
Taking into the account the intra and inter tumor heterogeneity of this complex mouse model that may be making difficult the interpretation of our results, we decided to establish syngeneic models derived from a single MSH2ko tumor. We generated an organoid biobank of hypermutant intestinal tumors with more than 40 samples that was used to establish 4 syngeneic models. Syngeneic models were genotyped and characterized showing differences in terms of mutations, immune infiltration and Wnt signaling activation. Then, we performed two experiments with two different syngeneic models with high beta-catenin signal. Porcupine inhibitor WNT974 or tankyrase inhibitor (G007-LK) treatment, depending on the original upstream or downstream mutations on Wnt signaling pathway, in combination with the checkpoint inhibitor PD-1 did not induce statistically significant antitumor effect. In addition, flow cytometry studies from the tumor samples did not show increase in immune infiltration or activation of immune cells upon treatment.
