Use of mutational signatures and spectra to determine the extent to which prior therapy drives the onset of colorectal and endometrial cancer

Abstract

The mutagenic effects of genotoxic damage, caused from DNA-damaging chemotherapy (CT) and radiotherapy (RT), has been a long-term concern of anticancer treatment. Accumulating epidemiological evidence suggests there is a small increase of secondary cancers, including colorectal (CRC) and endometrial cancer (EC), a period after receiving genotoxic treatment for a first malignancy. CRC is the 4th most common cancer within the UK and is the 2nd biggest cancer killer; EC is the most commonly diagnosed gynaecological cancer, resulting in ~1,370 deaths per year. Recent advances in sequencing technology means established mutational signatures within a genome can now be identified and used to determine the extent in which prior genotoxic therapy drives the onset of cancer. Therefore, this study utilized next-generation (NGS) and whole genome sequencing (WGS) technologies, to identify both mutational signatures and unique CRC/EC driver gene mutations which may have originated from therapy exposure.

This was investigated within the 100K genomes project, and CORGI (Colorectal Tumour Gene Identification) cohorts. Variant callers strelka2 and smCounter2 produced VCFs (for unique driver mutation identification), and signatures within both cohorts were produced by inputting these VCFs into SigProfilerExtractor tool. Within the CORGI cohort, signature analysis identified de novo decomposed signature SBS96A, composed of SBS5, SBS87 and SBS54 COSMIC signatures – clock-like, thiopurine CT treatment, and possible sequencing artefact respectively. Additionally, driver gene analysis identified a handful of uncommon and/or rare pathogenic mutations in AT rich regions – primarily within APC, in a T>A C>T context, reflective of an alkylating agent. Within the GEL (Genomics England Limited) 100KGP cohort two signatures of interest were identified within RT groups: SBS58 and SBS89 – an artefact signature and unknown aetiology respectively. In EC, a significant increase of DBS were identified within the CT 10 year group.

Conclusively, the above findings provide molecular evidence for the role of prior genotoxic therapy in the growth +/ onset of subsequent cancers. However, both GEL and CORGI cohorts lack power, because of their small size and limited clinical data, weakening the strength of these hypotheses. Consequently, observations could be due to a confounding variable. It is likely sample size and power will increase in future, as more genomes are sequenced as part of the personalised medicine drive within the NHS. With this additional supporting evidence, these preliminary findings may be enhanced and theoretically confirm genotoxic therapy’s influence in the onset of secondary cancer in latter life.