Final Report Summary - FINDING CRC GENES (Identification of new colorectal cancer susceptibility genes) Colorectal cancer (CRC) is usually sporadic, i.e. appears in individuals mostly due to environmental factors. Nonetheless, around 30% of CRCs are familial, i.e. they appear in several individuals of the same family, probably due to a mixture of environmental and genetic factors. From these, some families show a strong aggregation of CRC as a cause of the inheritance of genetic mutations that pass from one generation to the next one in a family. In this case, we name it hereditary CRC. Although several genes are known to cause CRC when mutated, there is still a big proportion of familial cases that remain without a known genetic predisposing cause. Knowledge of the genetic mutation causing the disease in a family, allows us to offer genetic testing in the at-risk individuals. The result of this testing permits intensifying early detection and prevention measures only in those individuals that carry the pathogenic mutation. The overall aim of this proposed project is to discover new CRC susceptibility genes or variants to better understand the colorectal carcinogenesis process and, at the same time, improve the clinical monitoring of families affected by CRC. This general aim has been tackled through several complementary subprojects, all of them through the study of CRC families in which genetic testing was done, but no mutation was identified in the known CRC predisposing genes. Below, we briefly summarize the most important results and conclusions achieved, as well as their putative impact on our society. 1. Molecular characterization of a French-Canadian founder MSH6 mutation in Quebec: MSH6 is a known gene that predisposes to Lynch syndrome (LS), a CRC syndrome with other extra-colonic malignancies such as endometrial cancer, ovarian cancer and others. Interestingly, we found that a mutation in this gene (c.10C>T; p.Q4X) appeared in 11 French-Canadian (FC) families from Quebec presenting clinical features of LS. Additionally, 16 out of 6433 newborns in Quebec were found to carry this mutation. These two pieces of evidence suggested to us the mutation could be one of the most common LS-causing mutations and likely had a common founder origin. To prove this latter conjecture, we aimed to determine if the mutation was a founder mutation in the FC population of Quebec, and to characterize its clinical implications in this population. By haplotyping 14 microsatellite markers surrounding the mutation, in 23 carriers, 15 non-carriers of the family and 178 FC controls, we concluded that a common haplotype of 1.61-3.42Mb was shared by the carrier individuals, strongly suggesting that the mutation had occurred as a unique event in a single founder individual 400-600 years ago. This hypothesis is consistent with a mutation that arrived with the first French Quebec founders. Regarding the clinicopathological features of the mutation we can conclude that, in line with previous studies done on MSH6 mutations compared to other mismatch repair genes mutations, we see an increased risk of EC in our MSH6-mutation carrier families. Finally, through the screening of the mutation in 381 FC endometrial cancer cases, 187 FC CRC cases and 6433 newborns, we estimated that the mutation increases 7.5 times the risk for endometrial cancer and 2.2 times for CRC. Overall, these results will have important implications for genetic counselling in the FC population of Quebec. 2. Identification of a novel POLE mutation through whole-exome sequencing in a family with an unusual polyposis phenotype: One of our main goals was to study families with unknown genetic cause of the disease but with a striking phenotype of CRC. To this aim, we selected and collected samples from 9 families either with a strong history of CRC or polyposis in the family, or young individuals diagnosed with CRC before age 25y. Whole-exome sequencing (WES) permits to sequence all the coding regions of the genome at the same time, giving us the opportunity to look beyond the already known CRC genes that are tested in the molecular diagnosis routing. In one of these families, presenting with an unusual polyposis phenotype, with or without the appearance of CRC and with multiple tumors in one of the affected siblings (including ovarian and breast cancers), we identified a novel variant in POLE gene (T278K). This gene encodes for the catalytic subunit of the DNA polymerase ε and has been recently described as a CRC susceptibility gene. We have used several approaches to assess the pathogenicity of the variant found in our family: i) functional analyses in yeast have shown that T278K interferes in the protein function; ii) analysis of the tumors of the patients of this family show that they are hypermutated; and iii) harbor a specific mutational signature associated to POLE disruption. Overall, we can conclude that the POLE variant T278K is the cause of the disease in this family, the members of which will be able to benefit from genetic testing and counselling thus improving their clinical management.3. Germline mutations in CDC20 result in familial malignant ovarian germ cell tumors and aberrant mitotic progression: Although initially ovarian tumors were not the focus of this project, in a parallel study of malignant ovarian germ cell tumors (mOGCT), we found two variants in CDC20 gene, which is part of the spindle-assembly checkpoint (SAC) pathway. Mutations in other genes of this pathway such as BUBR1, BUB1 and BUB3, have been related to CRC predisposition. Additionally, one of the two mOGCT families with a CDC20 mutation, developed a CRC. Therefore, we decided to proceed with this exciting study in the framework of the Marie Curie fellowship. In contrast to male germ cell tumors (testicular cancer), female malignant germ cell tumors are rare, and almost nothing is known about their genetic etiology. To address this gap in knowledge, we collected five families with two or more cases of familial malignant ovarian germ cell tumors (there are only 8 families like this published). Exome sequencing of these five families uncovered two likely deleterious missense variants in CDC20 (c.452T>G, g.43359993T>G, p.Ler151Arg and c.993C>G, g.43360877C>G, p.Asn331Lys) in two of the five families. No other variants of note were common to any other familial malignant ovarian germ cell tumors. The variants segregate with cancer in these families. CDC20 is essential for mitotic progression and its interaction with the spindle assembly checkpoint (SAC) prevents aneuploidy in daughter cells. In silico predictors indicate the variants are likely pathogenic; for p.N331K structural computational analyses anticipate an impairment of CDC20 interaction with SAC partner, BUBR1. We in vitro studies in cellular models (HeLa cells, human fibroblasts, yeast) and demonstrated that both variants compromise mitotic arrest when the SAC is active and likely are the underlying cause of the tumors in the families. Thus, we identify CDC20 as the first gene that predisposes to familial malignant ovarian germ cell tumors. Families carrying such variants could benefit from appropriate counselling and screening programs. The latter might not be restricted to the ovaries as pathogenic variants in CDC20 could also potentially increased risk of neoplasia at other sites, as illustrated by one the families in this study that presents with other malignancies.4. RNAseq as a novel approach to tackle the missing heritability in colorectal cancer: Regardless the advances made so far, currently identified CRC predisposing genes only contribute to a small proportion of the expected predisposition to CRC. Thus, we face an important component of missing heritability that may in part be explained by a polygenic model of the disease, where rare moderate-penetrance and common low-penetrance alleles represent a continuum to the risk of CRC. Additionally, the presence of variants in non-coding regions that may be affecting gene expression or splicing and which escape from DNA-scanning strategies, also represent a promising field to explore for novel targets that could explain part of the missing heritability in CRC. Decreased gene expression has proven to drive cancer. In a similar manner, imbalanced expression between the two alleles of a gene (allele specific expression – ASE) can also underlie a pathogenic defect. Nowadays, high-throughput RNA sequencing (RNA-seq) has become a mature technology capable of analyzing the transcriptome in high-depth. Recently, researchers have used RNA-seq to better understand the regulatory effects driving colorectal carcinogenesis, focusing on the somatic level of the disease through large subsets of matched normal and tumoral sporadic cases. With this subproject, we used RNA-seq at the germline level to analyze 96 familial CRC cases lacking a known pathogenic defect and compared them to publicly available controls. Although the experimental part has been done, the bioinformatics analyses are still undergoing for this subproject. Briefly, we aim to analyse the ASE of the CRC cases and find out if there are any differences in any genes compared to the 344 publicly available controls that could explain part of the missing heritability of CRC and open the door to new RNA-based strategies to discover CRC mechanisms.
