The landscape of somatic mutation in normal colorectal epithelial cells

Abstract

The colorectal adenoma-carcinoma sequence has provided a paradigmatic framework for understanding the successive somatic genetic changes and consequent clonal expansions leading to cancer. As for most cancer types, however, understanding of the earliest phases of colorectal neoplastic change, which may occur in morphologically normal tissue, is comparatively limited. Here, we whole genome sequenced hundreds of normal crypts from 42 individuals. Signatures of multiple mutational processes were revealed, some ubiquitous and continuous, others only found in some individuals, in some crypts or during certain periods of life. Likely driver mutations were present in ~1% of normal colorectal crypts in middle-aged individuals, indicating that adenomas and carcinomas are rare outcomes of a pervasive process of neoplastic change across morphologically normal colorectal epithelium. Colorectal cancers exhibit substantially elevated mutation burdens relative to normal cells. Sequencing normal colorectal cells provides quantitative insights into the genomic and clonal evolution of cancerdriver mutations, which conceivably are morphologically indistinguishable from normal cells, are similarly unclear. In large part, these deficiencies are due to the technical challenge of identifying somatic mutations in normal tissues, which are composed of myriad microscopic cell clones. Several different approaches have been adopted to address this 4-14 , revealing signatures of common somatic mutational processes in normal cells of the small and large intestine, liver, blood, skin, and nervous system. Thus far, however, studies have not been of sufficient scale to characterise variation in signature activity or detect less frequent processes 4-14. Remarkably high proportions of normal skin, oesophageal, and endometrial epithelial cells have been shown to be members of clones already carrying driver mutations 10,11,15,16 , and large mutant clones have been detected in blood 17-20. The extent of this phenomenon in the colon, an organ with a high cancer incidence, has not been investigated. Colonic epithelium is a contiguous cell sheet organised into ~15,000,000 crypts each composed of ~2,000 cells 21. Towards the base of each crypt resides a small number of stem cells ancestral to the maturing and differentiated cells in the crypt 22. These stem cells stochastically replace one another through a process of neutral drift 23,24 such that all stem cells, and thus all cells, in a crypt derive from a single ancestor stem cell that existed in recent years 25-27. The somatic mutations that were present in this ancestor are thus found in all ~2,000 descendant cells and can be revealed by DNA sequencing of an individual crypt. These stem cells are thought to be the cells of origin of colorectal cancers 28. To characterise the earliest stages of colorectal carcinogenesis, somatic mutation burdens, mutational signatures, clonal dynamics, and the frequency of driver mutations in normal colorectal epithelium were explored by sequencing individual colorect...

Elevated somatic mutation burdens in normal human cells due to defective DNA polymerases

“…Small insertion and deletion (ID) mutation rates in normal intestinal crypts were also elevated in individuals with germline POLE/POLD1 mutations, with rates of 13/year (POLE L424V), 44/year (POLD1 S478N) and 12/year (POLD1 D316N and POLD1 L474P) (linear mixed-effects model 95% C.I., 10-16, 35-53, 9-16, P=10 -10 ,P=10 -13 and P=10 -9 respectively). These are all substantially above the expected rate of 1/year in individuals without POLE or POLD1 mutations 26 (Fig. 1c and 1d) (Methods, Supplementary Code).…”

“…In both brain and oesophagus, a further novel mutational signature (SBS10e) was present, characterised predominantly by C>A substitutions at TCG and TCA trinucleotides (Fig 3c). The mechanism underlying this tissue-limited signature is not known and this signature has not previously been observed in extensive surveys of human cancer and several normal tissue types 7,9,23,[26][27][28][29]41 .…”

“…Nine have been previously reported; SBS1, SBS5, SBS10a, SBS10b, SBS17b, SBS28, SBS35, SBSA and SBSB. SBS1 (characterised by C>T substitutions at NCG trinucleotides and likely due to deamination of 5-methylcytosine) and SBS5 (of unknown etiology) are found in all normal intestinal crypts from normal individuals where they accumulate in a more-or-less linear manner with age 7,9,26,40 . SBSA and SBSB are found in some normal intestinal crypts from some normal individuals and are predominantly acquired during childhood 26,41 .…”

“…SBS1 (characterised by C>T substitutions at NCG trinucleotides and likely due to deamination of 5-methylcytosine) and SBS5 (of unknown etiology) are found in all normal intestinal crypts from normal individuals where they accumulate in a more-or-less linear manner with age 7,9,26,40 . SBSA and SBSB are found in some normal intestinal crypts from some normal individuals and are predominantly acquired during childhood 26,41 . SBSA is likely due to colibactin, a mutagenic product of a strain of E.coli sometimes present in the colon microbiome 26,42 .…”

“…SBSA and SBSB are found in some normal intestinal crypts from some normal individuals and are predominantly acquired during childhood 26,41 . SBSA is likely due to colibactin, a mutagenic product of a strain of E.coli sometimes present in the colon microbiome 26,42 . SBS10a, SBS10b and SBS28 were previously found in the subsets of colorectal, endometrial and other cancer types with somatically acquired POLE mutations 7,9 (Fig.…”