Integrative Analysis of Germline Rare Variants in Clear and Non-Clear Cell Renal Cell Carcinoma.

ABSTRACT

IMPORTANCE RCC encompasses a set of histologically distinct cancers with a high estimated genetic heritability, of which only a portion is currently explained. Previous rare germline variant studies in RCC have usually pooled clear and non-clear cell RCCs and have not adequately accounted for population stratification that may significantly impact the interpretation and discovery of certain candidate risk genes. OBJECTIVE: To evaluate the enrichment of germline PVs in established cancer-predisposing genes (CPGs) in clear cell and non-clear cell RCC patients compared to cancer-free controls using approaches that account for population stratification and to identify unconventional types of germline RCC risk variants that confer an increased risk of developing RCC. DESIGN SETTING AND PARTICIPANTS: In 1,436 unselected RCC patients with sufficient data quality, we systematically identified rare germline PVs, cryptic splice variants, and copy number variants (CNVs). From this unselected cohort, 1,356 patients were ancestry-matched with 16,512 cancer-free controls, and gene-level enrichment of rare germline PVs were assessed in 143 CPGs, followed by an investigation of somatic events in matching tumor samples. MAIN OUTCOMES AND MEASURES: Gene-level burden of rare germline PVs, identification of secondary somatic events accompanying the germline PVs, and characterization of less-explored types of rare germline PVs in RCC patients. RESULTS: In clear cell RCC (n = 976 patients), patients exhibited significantly higher prevalence of PVs in VHL compared to controls (OR 39.1, 95% CI 7.01-218.07, p-value4.95e-05, q-value0.00584). In non-clear cell RCC (n = 380 patients), patients carried enriched burden of PVs in FH (OR 77.9, 95% CI 18.68-324.97, p-value1.55e-08, q-value 1.83e-06) and MET (OR 1.98e11, 95% CI 0-inf, p-value 2.07e-05, q-value 3.50e-07). In a CHEK2-focused analysis with European cases and controls, clear cell RCC patients (n=906 European patients) harbored nominal enrichment of the previously reported low-penetrance CHEK2 variants, p.Ile157Thr (OR1.84, 95% CI 1.00-3.36, p-value0.049) and p.Ser428Phe (OR5.20, 95% CI 1.00-26.40, p-value0.045) while non-clear cell RCC patients (n=295 European patients) exhibited nominal enrichment of CHEK2 LOF germline PVs (OR 3.51, 95% CI 1.10-11.10, p-value 0.033). RCC patients with germline PVs in FH, MET, and VHL exhibited significantly earlier age of cancer onset compared to patients without any germline PVs in CPGs (Mean 46.0 vs 60.2 years old, Tukey adjusted p-value < 0.0001), and more than half had secondary somatic events affecting the same gene (n=10/15, 66.7%, 95% CI 38.7-87.0%). Conversely, patients with rare germline PVs in CHEK2 exhibited a similar age of disease onset to patients without any identified germline PVs in CPGs (Mean 60.1 vs 60.2 years old, Tukey adjusted p-value 0.99), and only 30.4% of the patients carried secondary somatic events in CHEK2 (n=7/23, 95% CI 14.1-53.0%). Finally, rare pathogenic germline cryptic splice variants underexplored in RCC were identified in SDHA and TSC1, and rare pathogenic germline CNVs were found in 18 patients, including CNVs in FH, SDHA, and VHL. CONCLUSIONS AND RELEVANCE This systematic analysis supports the existing link between several RCC risk genes and elevated RCC risk manifesting in earlier age of RCC onset. Our analysis calls for caution when assessing the role of germline PVs in CHEK2 due to the burden of founder variants with varying population frequency in different ancestry groups. It also broadens the definition of the RCC germline landscape of pathogenicity to incorporate previously understudied types of germline variants, such as cryptic splice variants and CNVs.