The mutational landscape of human somatic and germline cells

Luiza Moore, Alex Cagan, Tim Coorens, Matthew D.C. Neville, Rashesh Sanghvi, Mathijs A. Sanders, Thomas R.W. Oliver, Daniel Leongamornlert, Peter Ellis, Ayesha Noorani, Thomas J Mitchell, Timothy M Butler, Yvette Hooks, Anne Y Warren, Mette Jorgensen, Kevin J. Dawson, Andrew Menzies, Laura ONeill, Calli Latimer, Mabel Teng, Ruben VanBoxtel, Christine A. Iacobuzio-Donahue, Inigo Martincorena, Rakesh Heer, Peter Campbell, Rebecca C. Fitzgerald, Michael Stratton, Raheleh Rahbari

Posted on: 11 December 2020

Preprint posted on 26 November 2020

Article now published in Yearbook of Paediatric Endocrinology at http://dx.doi.org/10.1530/ey.19.14.16

Through the looking glass: examining mutation rate in normal human tissues.

Selected by Kerryn Elliott

Categories: bioinformatics, cancer biology, cell biology, developmental biology, genetics, genomics

Background

Our bodies accumulate DNA damage as we age, due to imperfect cell divisions and other external influences such as UV light. These mutations occur at different rates in different tissues, and are usually dependent on the number of cell divisions. In recent years the amount of mutations present in “normal tissues” had been shown to be variable between sample types, and includes a surprisingly high number of mutations in genes which are thought to drive cancer (driver genes), which previously had only been found in tumour samples. In order to understand how this mutation rate varies between different tissues and importantly between somatic and germline tissues, the authors present an impressive #panbodymap where they combine histology and genomics to microdissect tiny regions of normal human tissues and perform whole genome sequencing to determine how different cell types accumulate mutations over time.

Main findings

The authors used laser capture microscopy to dissect samples from 29 microscopic structures in three individuals, and then performed whole genome sequencing on more than 500 samples (Figure 1). These whole genomes were then used to analyse mutational signatures, telomere length, clonal structures, cancer associated mutations and mutation rate in all tissues sampled.

Figure 1. Histological sections and variant allele frequencies (VAF) from the 29 tissues sampled. The black outline indicates the sectioned area. A high VAF indicates a more clonal population. Figure 1b made available by a CC-BY-NC-ND 4.0 International license.

Of the most notable findings, the clonality of the mutations, represented by a higher variant allele frequency (Fig 1) and the total mutation burden varied considerably between the tissue types, with the crypts in the intestine showing the highest mutation rate at 52 singe base substitutions (SBS) per year, and the spermatagonial stem cells showing the lowest mutation rate at 2.6 SBS/year.

In order to determine how these mutations arose, mutational signatures, which are indicative of various mutational processes that are active in the genome, were calculated. The most abundant mutational signatures in normal tissues were SBS1, which occurs due to deamination of 5-methyl cytosine as we age, and SBS5/SBS40 which accumulate with age with an as yet unknown mechanism (Fig 2). Although these two signatures were present in every tissue, the amount of each was variable between tissue type.

Other dominant signatures detected which were restricted to certain sample types were the UV damage-induced SBS7, found in all skin samples and hair follicles, but not in skin sebaceous glands, and SBS18, which is caused by oxidative damage and was found in 19/29 sample types and was a major source of mutations in intestinal crypt stem cells. Interestingly SBS 2 and 13, which are thought to result from the mutagenic properties of APOBEC deaminases where present only in a subset of intestinal samples. Additionally, telomere length, which is also considered a hallmark of aging, was also found to vary between tissues with the prostate showing the shortest telomere length and the thyroid follicle and seminiferous tubule in the testis having the longest telomere length.

Figure 2. Mutational signature analyses across the cohort showing the contribution of each signature in each tissue. Note that SBS1 (cyan) and SBS5/40 (blue) dominate. Figure 3a made available by a CC-BY-NC-ND 4.0 International license.

One of the more striking discoveries was the decreased mutation rate in the testis. In order to further investigate this, the authors sequenced an additional 162 microbiopsies from 11 men. SBS1 was used to correlate the number of mutations to the number of cell divisions, suggesting that sperm stem cells only go through a handful of cell divisions per year, as compared to around a hundred divisions per year as seen in colon crypts (Fig 3).

Figure 3. Model depicting the number of SBS1 mutations per cell division for each cell division. Figure 2d made available by a CC-BY-NC-ND 4.0 International license.

Finally, although previous studies of normal tissues revealed the presence of mutations in key driver genes, most of the tissues sampled in this study had no driver mutations, although there were a handful detected in some tissues.

Why I chose this paper

My research focuses on cancer genomics, and mutational signatures have been an interest. I have followed the papers from the Sanger on mutations in normal tissues for many years now with fascination as they have been a real game changer in how we think about mutations in cancer. It has been great to see this compilation of cell types in one paper!

Questions for the authors

The low mutation rate for the testis was very interesting, but this is expected as the germline cells should be protected from mutation by increased repair systems. How would you further explore this phenomenon?

Why do you think there were very few driver gene mutations detected in this study?

The contribution from SBS1 and SBS 5/40 is intriguing in its differences. Was there any pattern to the cancers which had more or less of either?

Can you postulate why APOBEC deaminases would only be detectable in the intestinal crypts? Is it purely because that is where you see the most mutations?

Finally, what do you think was the most surprising finding?

doi: https://doi.org/10.1242/prelights.25983

Read preprint