Periodic Reporting for period 2 - HYPER-INSIGHT (Hypermutated tumors: insight into genome maintenance and cancer vulnerabilities provided by an extreme burden of somatic mutations)
Reporting period: 2019-08-01 to 2021-01-31
DNA is the molecule of heredity: it encodes all the genetic information that is necessary for cells to survive, grow, divide and differentiate into tissues that make up organisms. When this information stored in DNA is changed, this is referred to as mutation. In humans and in animals more generally, mutations that happen in germline cells can lead to disease phenotypes in the offspring, while mutations in somatic cells can cause tumors (or, they may potentially contribute to aging of tissues). In many cancers, and potentially in some healthy cells, a large number of mutations can accumulate; this may happen, for instance, because a DNA repair system has failed, which is a common risk factor for cancer. The HYPER-INSIGHT project is interested in what happens to such cells after they accumulate a very large number of mutations -- a phenomenon called hypermutation. This can help us learn how the human cells copy and repair DNA, which may have important implications for cancer research and for evolutionary biology research. Additionally, cancer cells undergoing hypermutation might have a particularly strong dependency on certain genes on which normal cells do not depend on so strongly. We will also search for such genes, which represent vulnerabilities of cancer cells and might be potentially used therapeutically, to selectively target cancer cells while sparing healthy tissues. Because cancer is an increasingly heavy burden on aging societies of the developed world, we anticipate that our project and related projects that look for novel avenues for tumor therapy will in the long run bring societal benefit. Additionally, understanding the biological mechanisms of mutagenesis may further help with prevention of cancer and with gaining better insight into the mechanisms underlying genetic diseases.
Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far
We have examined genome sequences of thousands of tumors (of various cancer types) to search for unusual patterns in how mutations are distributed in the genome. This has led to a finding of the so-called ‘mutation fog’: a clustering pattern that reflects how sometimes the cellular DNA repair systems can, erroneously, start introducing mutations instead of correcting them (published as Mas-Ponte & Supek, Nature Genetics, 2020). This may help explain the differences in cancer propensity across individuals and across human tissues. Next, we have examined a different kind of pattern: the density of mutations distributed across large swaths (a million nucleotides) of the human chromosomes. Remarkably this pattern allows a very accurate classification of tumors into tissues and subtypes, starting only from the mutational pattern (published as Salvadores, Mas-Ponte and Supek, PLOS Computational BIology, 2019), without necessity to consult e.g. gene expression data -- a standard tool for cancer typing. This opens new avenues for diagnosing cancer type by DNA sequencing from ‘liquid biopsies’ (cancer cells or cancer DNA in blood plasma), and for classifying the metastatic ‘cancers of unknown origin’. Furthermore, we have examined how increased mutation burden of so-called ‘nonsense’ mutations affects human cells including tumors, which have a special mechanism (NMD) that degrades genetic messages containing nonsense mutations. We found that this NMD mechanism dampens the results of cancer immunotherapy in many cancer patients, and that it can be predicted from the genome sequence of tumors which patients those are (Lindeboom, Vermeulen, Lehner & Supek, 2019, Nature Genetics). Our data strongly suggest pharmacological NMD inhibition could be used to potentiate tumor immunotherapy, guided by genomic markers. Finally, we have analysed diverse mutation patterns (together with gene expression patterns and epigenomic patterns) occurring in tumor cell lines -- an experimental model of tumor biology, which is used by many labs around the world to test new cancer therapies. We found, suprisingly, that approximately 6% of all cancer cell lines may originate from a different tissue than originally thought, as classified by the mutational and gene regulation patterns, using a machine learning method (Salvadores, Fuster and Supek, 2020, Science Advances). This has important implications for experimentally testing new cancer therapies in cancer cell lines, because the tissue-of-origin has a strong effect on how the cancer cell lines respond to drugs. Overall, our work suggests that examining mutational patterns in human cancer cells can provide new insight into carcinogenesis and also new avenues for treating cancer.
Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)
In the second part of the HYPER-INSIGHT project, we will aim to push the boundary of the state of-the-art approaches for computational cancer genomics methods for mutational analysis. Our goal is to further the understanding of fundamental biological mechanisms of DNA copying and repair, and also to propose further therapeutic vulnerabilities of cancerous cells. Firstly, we are particularly interested in using mutational patterns observed in some rare inherited tumors to be able to measure the DNA replication program of tumor cells -- which parts of chromosomes replicate earlier and which later, and how this changes across cancers. This has important implications for understanding how errors occur during DNA copying, and therefor also for occurrence of heritable diseases and cancer (that often result from such errors). Secondly, we are interested in how mutation density in various segments of chromosomes varies across individuals and/or across various cancer types, and whether this mutation rate variation has a genetic basis. This might have implications to understanding why certain oncogenes or tumor suppressor genes get activated in different individuals, depending on which chromosomal segment the gene lies in. Thirdly, we are interested in examining genetic interactions (meaning, statistically unexpected co-occurrence or mutual exclusivity) between DNA mutational patterns and DNA copy number changes. We are hoping such genetic interactions may reveal new mechanisms of carcinogenesis and/or drug resistance that are important for hypermutating tumoral cells. Fourth, we are continuing our experimentation on various models of hypermutating cells examined on cancer cell lines, such as to be able to support a variety of predictions from computational cancer genomics with experimental data, adding further support to the multitude of evidence we will collect about hypermutating cells. In conclusion, this project aims to further our understanding of how mutation in human cells drives cancer and other genetic disease.