Genomic ascertainment refines the population prevalence and phenotype associations of germline TERT variants

Background Individuals with deleterious germline variants in telomerase (encoded by TERT ) are at high risk of pulmonary fibrosis (PF), liver disease (LD), bone marrow failure, head/neck squamous cell carcinoma (HNSCC), and other complications. Clinically ascertained cohorts are subject to biased ascertainment of affected individuals and families with early onset, complex disease. This study used genomic ascertainment to determine prevalence of deleterious TERT variants and the associated risks of telomere biology disorder (TBD) phenotypes. Methods TERT variants were extracted from exome sequencing data in UK Biobank (UKB, 469,377 exomes) and Geisinger-Regeneron DiscovEHR collaboration (DiscovEHR, 170,503 exomes). Cases were individuals with heterozygous deleterious TERT variants classified per modified ACMG/AMP criteria. Controls excluded rare, exonic, and splice site TERT variants. Electronic health records were assessed for TBD-related ICD10 codes. Case-control comparisons were done separately for UKB and DiscovEHR, accounting for smoking, alcohol, BMI, sex, genetic ancestry, and birthyear. Results There were 64 deleterious TERT variants in 203 UKB participants and 32 in 138 DiscovEHR participants. Prevalence estimates were 1:2,312 in UKB and 1:1,235 in DiscovEHR. Risk of PF was significantly increased for TERT heterozygotes compared with controls (UKB: odds ratio [OR] 10.9 95% confidence interval [CI]: 6.11-19.45, p < 0.001; DiscovEHR: OR 6.64, 95% CI: 3.31-13.3, p < 0.001). Deleterious TERT variants were also associated with higher risk of LD (OR 2.4, 95% CI:1.25-4.59, p = 0.038) and HNSCC (OR 5.75, 95% CI: 2.36-14.02, p = 0.001) in UKB but not in DiscovEHR. Conclusions Rare deleterious TERT variants were present in 0.04 to 0.08% of the populations studied. Genomic ascertainment shows a predominance of pulmonary phenotypes, with PF associated with deleterious TERT variants in both cohorts. Associations with LD and HNSCC were observed only in UKB, possibly reflecting cohort differences including healthy volunteer bias. These findings illustrate the clinical importance of genetic diagnosis, and need for TBD-specific clinical trials.