Although most cancers are sporadic, a considerable proportion is due to genetic susceptibility. Many cancer types display moderate to high heritability, but predisposition genes have been established for only a few. In this thesis project, we have utilized the data in the Finnish Cancer Registry (FCR) and the Population Information System (PIS) databases to identify cancer patients with a potential genetic predisposition. For most of the patients in the FCR archival tissue material in the form of paraffin-embedded formalin-fixed (FFPE) blocks is available in the pathological department of their local hospital, which makes genetic studies on selected cases possible. The Finnish population is due to its demographic history more homogenous than most populations, and thus optimal for studies on germline predisposition to disease. This thesis includes studies on two cancer types currently lacking well-established susceptibility genes: lung adenocarcinoma (LUAD), and angioimmunoblastic T-cell lymphoma (AITL). In addition, my thesis work examines esophageal squamous cell carcinoma (ESCC) and hereditary diffuse gastric cancer (HDGC) susceptibility. A fraction of these cancers can be attributed to germline mutations in RHBDF2 and CDH1, respectively. We utilized exome sequencing in all studies with the aim to discover novel susceptibility genes for the different phenotypes. Gastric cancer is the second leading cause of cancer mortality. Three distinct hereditary gastric cancer syndromes are recognized: HDGC, familial intestinal gastric cancer, and gastric adenocarcinoma and proximal polyposis of the stomach. Up to one third of HDGC families have heterozygous germline mutations in CDH1, which encodes E-cadherin. Most cases showing familial aggregation do not, however, have a known genetic basis. The aim of the first study was to find the putative cancer-predisposing gene defect in an HDGC family that had previously been tested negative for CDH1 mutations. We identified three candidate variants: p.Glu1313Lys in INSR, p.Arg81Pro in FBXO24 and p.Pro1146Leu in DOT1L. INSR is of special interest as insulin signaling has been shown to modulate E-cadherin glycosylation and stability. ESCC represents the most common subtype of esophageal cancer, a common and often fatal malignancy. Although ESCC shows familial clustering and moderate to high heritability, few studies on moderate to high penetrance predisposition exist. We have previously found familial aggregation of classic Kaposi sarcoma by performing family name at birth and municipality at birth based clustering of all cancer cases in the FCR. In study II, we used the same clustering algorithm to find potential familial cases of ESCC. We managed to find a geographical population subset that showed enrichment of clustered ESCC patients. The finding may indicate that a susceptibility variant is enriched in this population subset. We collected FFPE tissue material and exome sequenced a total of 30 ESCC cases. Six variants passed filtering and analysis criteria, the most frequent being a nonsense mutation in DNAH9 detected in four unrelated patients. One patient's tumor showed loss of the wild type allele of DNAH9, suggesting a tumor-suppressive function. The other candidate genes GKAP1, BAG1, NFX1, FCSK, and DDOST harbored missense variants. A rare variant in EP300, a gene that has previously been implicated in ESCC carcinogenesis, was discovered to segregate in three affected individuals in a single family. Lung cancer is the most common cancer in the world and the leading cause of cancer death in both men and women. It is well established that the primary cause of the malignancy is cigarette smoke, however, other factors are involved as only approximately 15% of smokers develop lung cancer and an estimated 10-25% of all lung cancers occur in never smokers. Gender influences the risk of lung cancer and women are overrepresented among never smokers with the disease. Because of a possible difference in the molecular basis, we chose to study never-smoker women with LUAD in study III. We selected the youngest patients listed in the FCR, who are more likely to harbor a predisposing germline mutation. We managed to collect normal FFPE tissue material from 21 never-smoker women who had been diagnosed with lung adenocarcinoma before the age of 45. Potentially pathogenic variants were found in eight Cancer Gene Census germline genes: BRCA1, BRCA2, ERCC4, EXT1, HNF1 A, PTCH1, SMARCB1, and TP53. The variants in TP53, BRCA1, and BRCA2 are very likely to have contributed to the early-onset lung cancer in the respective patients. This supports the idea that lung adenocarcinoma can be a component of certain cancer predisposition syndromes. In addition, fifteen genes displayed potentially pathogenic mutations in at least two patients: ABCC10, ATP7B, CACNA1S, CFTR, CLIP4, COL6A1, COL6A6, GCN1, GJB6, RYR1, SCN7A, SEC24A, SP100, TTN, and USH2A. Some of these candidate genes may explain a subset of female lung adenocarcinoma. AITL is a subtype of peripheral T-cell lymphoma with a bleak prognosis. Somatic driver mutations have been found in TET2, IDH2, DNMT3A, RHOA, FYN, PLCG1, and CD28, whereas germline susceptibility to AITL has to our knowledge not been studied. In study IV we performed an exome-wide rare variant analysis in 23 AITL patients. No germline mutations were found in the driver genes, indicating that they are not frequently involved in genetic AITL predisposition. Potentially pathogenic variants present in at least two patients and showing significant enrichment in our sample set were found in ten genes: POLK, PRKCB, ZNF676, PRRC2B, PCDHGB6, GNL3L, TTC36, OTOG, OSGEPL1, and RASSF9. The most significantly enriched variants were p.Lys469Ter in a splice variant of POLK and p.Pro588His in PRKCB. These two are intriguing candidates, as Polk deficient mice display a spontaneous mutator phenotype and PRKCB has been found somatically mutated in 33% in adult T-cell lymphoma, another form of peripheral T-cell lymphoma. These studies show that rarely performed next-generation sequencing of DNA extracted from FFPE tissue samples is well suited for research and should be utilized to a greater extent. We found plausible candidate genes for the phenotypes under study. However, due to relatively small sample sets, our findings need to be genetically validated in independent sample series and through functional studies to prove causality.
|Tila||Julkaistu - 2020|
|OKM-julkaisutyyppi||G5 Tohtorinväitöskirja (artikkeli)|
LisätietojaM1 - 89 s. + liitteet
- 3111 Biolääketieteet