What we’re working on right now….
Small cell carcinoma of the ovary, hypercalcemic type (SCCOHT) is the most common undifferentiated ovarian malignancy in women below age 40, with a mean diagnosis age of 23.9 years. It is an extremely aggressive tumor and, due to lack of effective treatment, long-term survival rates are poor. Between 2012 and 2014 we performed whole exome sequencing on 3 families with multiple members affected by SCCOHT and discovered that inherited mutations in SMARCA4 were segregating with the disease in these families. Sequencing of additional tumors, familial and sporadic in origin, revealed that all had loss of SMARCA4 expression resulting either from mutations or loss of heterozygosity, identifying SCCOHT as a monogenic disease. This discovery was published in Nature Genetics in May 2014.
In collaboration with Dr. Sidong Huang from the Biochemistry department at McGill, we are using SCCOHT cell lines as a model system to uncover synthetic lethal interactions with SMARCA4. So far we uncovered a critical vulnerability of SMARCA4 loss to CDK4/6 inhibitors, which is conserved in SCCOHT and lung cancer where SMARCA4 is also mutated. These findings were published in Nature Communications in January 2019 and, based on our findings, a CDK4/6 inhibitor is now in clinical trial for SMARCA4 mutant cancers (NCT03297606). Keeping in mind that single agent therapies usually result in notable levels of drug resistance development, we are continuing our investigations to identify additional molecular vulnerabilities that could be exploited through combination treatments with CDK4/6 inhibitors to improve outcomes for women at risk for, or affected by, SCCOHT.
Breast cancer is the most prevalent cancer among women, and approximately 10% of all breast cancers are hereditary. The two best known breast cancer susceptibility genes, BRCA1 and BRCA2, were identified over 25 years ago, but mutations in these genes account for ~40% of inherited breast cancers.
Since then, other genes including ATM, BARD1, BLM, BRIP1, CHEK2, NBS1, PALB2, PTEN, RINT1, TP53 and XRCC2 have been proposed as candidate breast cancer susceptibility genes and together account for another ~10% of hereditary breast cancer families. Despite these advances, the genetic variants responsible for the other 50% of hereditary cases are yet to be identified.
Worldwide, efforts continue to identify the missing breast cancer susceptibility alleles, but it is becoming clear that the remainder of the “missing heritability” for breast cancer in outbred populations cannot be attributed to a small number of undiscovered moderate to high penetrance breast cancer susceptibility genes, and progress has been slow. In our lab, we have on-going research projects that approach the problem from different angles.
Founder mutations in the French-Canadian population
Certain populations, like the French-Canadian population of Quebec or the Ashkenazi Jewish population, are notable for the presence of founder mutations that are the result of their presence in a small number of original settlers that, aided by relative genetic isolation, spread through the population in subsequent generations. Our team has contributed significantly to the identification, characterization and clinical application of founder mutations in BRCA1, BRCA2 and PALB2 in French Canadians, which greatly facilitated genetic screening of at-risk families in Quebec using low-cost, targeted screening assays for many years. In this project, we take advantage of the founder status of the Quebec population to use the latest next generation sequencing technologies to search for predisposing genetic lesions in high risk French-Canadian breast cancer families without identified mutations in known breast cancer susceptibility genes.
Leveraging mutational signatures to profile breast tumours
It has been recognized for several years now that BRCA-related breast cancers are deficient in homologous DNA repair, which results in tumors with unstable genomes filled with mutations that remain unrepaired. These mutations form a footprint, or signature, which can be recognized bioinformatically from tumour sequencing data. In collaboration with Dr. Paz Polak at Mount Sinai Hospital, we are exploring strategies to use the rich information provided by mutation signatures of tumours to 1) identify driver variants in novel genes that cause breast cancer predisposition; 2) clarify the pathogenic potential of variants of unknown significance in known breast cancer predisposition genes; and 3) identify breast tumours that are BRCA-like and might benefit from targeted treatments with PARP inhibitors even in the absence of clearly identified germline mutations in DNA repair genes.
In 2009, the association of germline DICER1 mutations with a distinctive human disease syndrome (OMIM 601200) involving familial pleuropulmonary blastoma was reported. Since then, germline DICER1 mutations have also been described in cystic nephroma, embryonal rhabdomyosarcomas, ovarian sex cord stromal tumors (especially Sertoli-Leydig cell tumors), nasal chondromesenchymal hamartoma, ciliary body medulloepithelioma, multinodular goiter, differentiated thyroid carcinoma, Wilms tumor, pituitary blastoma and pineoblastoma, among others. Since 2009, our group contributed several manuscripts to the description of the DICER1 mutations and associated phenotypes in affected families.
More recently, our work on this syndrome has extended to the identification of somatic mutations in DICER1 tumors. There is mounting evidence to support the notion that, unlike the situation with oncogenes or classical tumor suppressor genes, DICER1-related tumor development relies on a delicate balance of germ-line and somatic lesions that compromise the microRNA processing function of the enzyme while in most cases avoiding its complete loss. Several on-going projects in our laboratory using patient tumors and model systems aim at elucidating the molecular events required for tumor development in carriers of hereditary DICER1 mutations.
We have started the first Canadian research network for the study of DICER1 Syndrome. With the purpose of collecting more information about DICER1 Syndrome, we have started a website where we share information about DICER1 Syndrome. Our database is being shared with interested parties free of cost.
Families with multiple members affected by rare or unusual cancer phenotypes, or individuals with multiple primary tumours, provide valuable opportunities to study the genetic factors underlying these disorders. Such families and individuals occasionally come to the attention of our research group through the clinical genetics work of our PI. One such case was a patient with multiple gastrointestinal lesions who proved to be a carrier of a biallelic mutation in BUB1B. Our discoveries about his genetic lesion provided the basis of our subsequent research projects on spindle assembly checkpoint genes. Another example was our work on 3 families with familial SCCOHT. Our findings from these families showed that genetic factors predisposing to rare hereditary versions of these tumors can also be
involved in the etiology of the more frequent, sporadic versions of the disease, resulting in another long-term research project.
Identifying predisposing mutations in a single family or individual can be challenging, but the advent of next generation sequencing technologies has greatly facilitated the process. In collaboration with research groups at our local core facility, the Centre d’Expertise et de Services Génome Québec, we have worked to optimize exome sequencing of formalin-fixed, paraffin-embedded tumors. We are now able to compare the exomes from the germline and the tumours of individual patients, thus enabling us to observe the genetic changes involved in the molecular evolution of tumour cells. We collect families with rare hereditary cancer phenotypes from the clinic and from the literature that are then investigated in our laboratory. We had several recent success stories, including the discovery that FGFR1 abnormalities play a role in both familial and sporadic neuroepithelial tumors commonly referred to as DNETs, and that a variant in the microRNA processing gene DGCR8 predisposes to familial multinodular goitre with schwannomatosis.