Background. The human telomere consists of a hexamer repeat motif, TTAGGG, at the 3’ and 5’ end of chromosomes preserving the genomic integrity and chromosomal stability. In turn, genomic instability is a hallmark of cancer - implicating telomere disturbance. Prostate cancer (PCa), the most diagnosed male associated cancer in Australia, shows significant ancestral disparity, with men of African ancestry at the greatest risk for aggressive disease presentation and mortality. Most recently, our team demonstrated that compared with European Australians, African derived prostate tumours present with significant genomic instabilities. Yet, no study has explored the association of telomere length (TL) with PCa ancestral disparities.
Objectives. Here, using patient-matched tumour-blood whole genome sequencing data generated for 117 treatment-naïve men of African versus 62 of European ancestries, we sought firstly to establish the reliability and validity of computational TL measurement techniques, and secondly to determine if estimated TL is correlated with PCa features of genomic instability and associated ancestral disparity.
Methods. TL estimates were derived from short-read whole genome data using three informatic tools, namely TelSeq, Telomerecat and Computel using high-performance computation (Artemis). Statistical tests were conducted between TL and genomic or clinical variables in R and STATA.
Results. Compared to Telomerecat, we found TelSeq and Computel to generate stable and highly correlated in both blood (rho=0.831, P<2.2e-16) and tumour (rho=0.923, P<2.2e-16) TL estimates. Continuing with a dual-tooled approach, we show African compared with European PCa patients (age from 40 to 85) to present with longer blood (3361.977 base pair versus 3095.169 base pair) and tumour TLs (3752.258 base pair versus 3184.965 base pair). Notably, compared with European patients, shortened tumour TL was significantly associated with aggressive PCa in African men (P=0.00283) and further statistically negatively correlated with genomic instabilities including percentage genome alteration, tumour mutational burden, somatic single-nucleotide variant, somatic indel, somatic structural variant, and gained regions.
Conclusion. Demonstrating accuracy and stability for TelSeq and Computel for tumour-relevant TL prediction, we provide evidence that TL is not only ancestrally determined, but that shortened tumour TL has a predictive value as a potential biomarker for aggressive PCa and future precision oncology of African men.