Mutational heterogeneity in human cancers origin and consequences pdf

Posted on Monday, May 17, 2021 3:37:15 PM Posted by Tioconsonnter - 17.05.2021 and pdf, english pdf 1 Comments

mutational heterogeneity in human cancers origin and consequences pdf

File Name: mutational heterogeneity in human cancers origin and consequences .zip

Size: 2405Kb

Published: 17.05.2021

International Journal of Biological Sciences. Journal of Cancer. Journal of Genomics. Global reach, higher impact. Journal of Genomics - Submit manuscript now

The repertoire of mutational signatures in human cancer

Although not all somatic mutations are cancer drivers, their mutational signatures, i. However, the mechanisms underlying nearly one-third of all mutational signatures are not yet understood. The signatures with established etiology and those with hitherto unknown origin appear to have some differences in strand bias, GC content and nucleotide context diversity. While nucleotide contexts might be adequate to establish etiologies of some mutational signatures, in other cases additional features, such as broader epi genomic contexts, including chromatin, replication timing, processivity and local mutational patterns, may help fully understand the underlying DNA damage and repair processes.

Nonetheless, remarkable progress in characterization of mutational signatures has provided fundamental insights into the biology of cancer, informed disease etiology and opened up new opportunities for cancer prevention, risk management, and therapeutic decision making. Genomic instability is a hallmark of all cancers. Cancer genomes typically harbor 10 3 —10 5 somatic point mutations, along with other classes of genomic alterations, including insertions and deletions InDels , copy number variations, rearrangements and ploidy changes 1 , 2.

While a vast majority of somatic mutations are not oncogenic drivers, their patterns of genetic changes and associated contexts can provide insights into past exposure to mutagens, mechanisms of DNA damage and repair defects, and extent of genomic instability, which in turn can guide rational strategies for cancer prevention, risk management and therapeutic decision making 3—5.

More recently, computational deconvolution of mutational patterns in somatic genomes has provided complementary and unbiased insight into the genome-wide consequences of these mutagenic processes in vivo in human tissues. Here, we first describe the computationally derived mutational signatures, emerging bioinformatics resources for analysis of the signatures and characteristics of signatures of known and unknown etiologies; we then discuss the emerging approaches for broader context-guided assessment of somatic mutations, mechanistic inference of the signatures and future direction.

Inferences of mutational signatures using different approaches. A Targeted mutagenesis using selected agents and sequencing of clonally derived cell populations to identify corresponding mutational signatures. B Perturbation of selected cellular processes in model systems e. C Data-driven approaches to identify mutational signatures of exo- and endogenous mutagenic processes. While the schematics above are shown for single base substitution SBS signatures, similar approaches have also been adopted for doublet base substitution DBS , small InDel and genomic rearrangement signatures.

Consequences of exposure to carcinogenic agents were known even in ancient civilizations. One of the oldest descriptions of cancer is found in an Egyptian papyrus dated about BC.

After the industrial revolution, coal tar was prescribed for medical purposes in the s, but later it was suspected to cause cancer in animals. In , Yamagiwa Katsusaburo and Koichi Ichikawa experimentally showed that coal tar can induce tumors on rabbits' ears, which could be one of the early systematic experiments demonstrating chemically induced carcinogenesis.

However, the idea that carcinogens cause DNA damage did not arise until the s, and the now accepted paradigm of cancer development that cancer is a genetic disease that progresses via mutagenesis began to take shape in the s.

Initially, a number of reporter assays were used to investigate mutagenic processes in cell lines and model systems 10— However, these were relatively low throughput and did not capture all aspects of the complexity of environmental exposure and deficiency in genome maintenance that are characteristics of the mutational landscapes of human tumors and nonmalignant somatic cells. More recently, whole genome sequencing and whole exome sequencing of thousands of cancer genomes have provided an opportunity to examine mutation patterns in cancer genomes using data-driven approaches and infer their likely etiologies Figure 1C.

A majority of computationally inferred signatures match with the mutation profile characteristics of known mutagenic processes.

These include environmental carcinogens [e. Recently, SBS17 has been attributed to 5-fluorouracil treatment These findings have provided insights into cancer etiology and influenced treatment options. Analysis of mutational patterns has also unveiled novel mutagenic processes and established their etiologies [e.

It appears that some mutational signatures typically arise progressively during aging processes in normal somatic cells [e. The PCAWG mutational signature analyses and COSMIC catalog of mutational signatures provide an excellent discourse of the latest mutational signatures, their etiologies and nucleotide-level characteristics 18 , A number of computational resources have been developed for extraction, interpretation and annotation of mutational signatures from large-scale somatic mutation data.

WTSI 28 and Emu 29 were among the first available to identify mutational signatures from somatic mutation data in cancer genomes. Since then, a number of additional tools such as SomaticSignatures 30 , SigProfiler 19 , SignatureAnalyzer 19 , sigfit 31 , Helmsman 32 , maftools 33 , signeR 34 and others have been developed.

These tools use probabilistic approaches and NMF to process and extract mutational signatures de novo from cancer genomic data. The number of mutational signatures present in somatic genomes is not known a priori ; some tools can automatically estimate an optimal number of signatures [e. EMu 29 , maftools 33 ]. Appropriate null models are critical for meaningful discovery of mutational signatures from genomic data.

Together, these tools provide a rich resource for signature discovery. In the future, it might be appropriate to have a DREAM Challenge-type community-driven systematic study to compare and benchmark performance of these tools on an open platform.

A number of computational methods such as deconstructSigs 37 , sigfit 31 , MutationalPatterns 38 , decompTumor2Sig 39 , etc. Some of them [e. SignatureEstimation 40 , SigsPack 41 ] further allow estimation of confidence intervals for each identified signature in a somatic genome.

It appears that analyzing somatic mutations in their genomic context and local patterns can provide additional critical insights. TensorSignatures 45 has been recently developed based on an overdispersed statistical model incorporating mutational catalogues, transcription and replication strand bias, and kataegis, leading to more robust extraction of mutation signatures. SigMa 46 and recently StickySig 47 model statistical dependencies among neighboring mutations to characterize strand coordination, and other genomic and nongenomic factors that influence the activity of mutation signatures.

Such efforts are exciting and contributing to the broader understanding of the patterns of the mutational signatures in the genome. For example, it appears that some signatures e. Other resources have been developed to link the mutational signatures with tumor evolution and therapeutic strategies. MutaGene offers a maximum likelihood approach to predict the likely etiology of individual mutations, which can be used to infer the likely mutagenic process behind individual driver mutations in a cancer Palimpsest 51 and trackSig 52 can provide clonality inferences for mutational signatures, which can inform how mutagenic processes change during the course of tumor progression.

Structural variation signatures 53 and HRDetect 54 can identify homologous recombination HR deficiency in human tumors, which could be targeted clinically. Several other tools can predict signatures e. Some signatures e. Are the known and unknown signatures somewhat different? Are there certain characteristics or lack thereof that helped decipher the known signatures, and could those provide potential informed guidance while investigating the signatures of hitherto unknown etiologies?

The COSMIC mutational signatures were identified using NMF, which is a mathematical technique for blind source separation, resolving an original matrix into a product of two matrices with lower dimensions It has an inherent clustering property such that it implicitly and parsimoniously groups the original dataset into a smaller set of relatively homogeneous subgroups.

Thus, if mutations of a given etiology are sparse and clustered, i. Indeed, many well-established endo- and exogenous mutational signatures have these properties.

Moreover, many of these signatures are associated with external mutagenic exposure or oncogenic mutations in DNA repair pathways that result in a specific and substantial burden of associated mutations in tumor genomes, which are associated with clinical variables. For instance, smoking and UV exposure cause an excessive burden of somatic mutations with distinct substitution patterns in lung and skin cancer subtypes.

Some DNA repair and genome maintenance defects also result in distinct nucleotide-level changes and manifest in tissue-dependent manner.

For instance, tumors with mismatch repair defects e. POLD1 and POLE mutations are relatively common in colon tumors, which lead to distinct substitution biases and up to one to two orders of magnitude more mutations in affected tumors compared to other tumors of the same subtypes; notably, a minority of tumors possessing defects in both mismatch repair and DNA polymerase functions show intricate signatures defined by SBS14 and SBS20 58 , The other major subset of interpretable mutational signatures of endogenous origin often involves sporadic, burst-like activity of specific mutations at distinct contexts e.

Indeed, these signatures were among the first to be identified. Rigorous examination, curation and validation by the broader collaborative scientific community in general, and the COSMIC initiative in particular, have helped establish etiologies of more complex signatures The success of this approach is exemplified by deciphering complex signatures e.

We further examined whether there are other quantitative differences between the known and hitherto unknown signatures. When the signatures were projected on a PCA plot using these meta-features, the unknown signatures partially segregated from the known signatures and the overall differences were qualitatively similar to those observed above. Feature selection using LASSO indicated that transcriptional strand bias and GC content are associated with differences between signatures of known and unknown etiologies Figure 2C ; the unknown signatures on average have weak transcriptional strand bias and lower GC content.

Both DBS and InDel signatures are recent, such that differences between the known and hitherto unknown signatures may be superficial, and etiologies of many of them could be established in the near future. In all cases, random forest mean decrease in Gini index and mean decrease in accuracy, which indicate feature importance, also showed comparable patterns.

However, compositions of the major signatures are broadly consistent across COSMIC versions 18 , indicating that these are usually distinct and stable It is also possible that signatures of basal genome maintenance, DNA damage and repair processes that are operative in most somatic cells during development and aging are inter-related 59 , and thus harder to isolate.

Moreover, crosstalk between multiple genome maintenance processes might lead to complex signatures. It is also possible that mutations arising from dose-dependent or reduced activities of genome maintenance processes might be harder to pinpoint than those associated with oncogenic mutations in DNA replication and repair-related genes.

Multidisciplinary efforts from the scientific community are addressing these open questions from different angles, and their innovative approaches are expected to provide new insights into origins, higher order patterns and consequences of the mutational signatures, as we discuss in the following sections.

DNA damage and repair depend on local nucleotide sequences, as well as broader genomic, epigenomic and nuclear contexts. Mechanisms underlying some mutational signatures might be sufficiently explained by their nucleotide contexts alone [e. Transcriptional strand biases can inform whether transcription-coupled DNA damage and repair processes could contribute to the signature of interest.

Likewise, replication strand bias can help predict whether replication of continuous strands and Okazaki fragments or other associated factors could potentially contribute to a signature of interest.

Emerging strategies for investigating characteristics of mutation signatures. Approaches such as analyses of strand bias, context preference, local patterns of mutations, cooperativity and correspondence with laboratory-generated mutation signature can potentially help provide additional mechanistic insights into the mutational signatures, including those of hitherto unclear origin.

Only representative examples are shown. Chromatin and nuclear contexts can influence mutagenesis and DNA repair pathway choices 61—63 , such that certain signatures may show context-specific enrichment. To narrow down likely mechanisms for a signature of interest, under-representation of the signature in certain contexts and absence of corresponding context-associated biases can help exclude unlikely possibilities. For instance, using chromatin and replication timing data it was shown that SBS8 is uncommon in gene-rich euchromatin regions, and likely arises in late and fast replicating regions due to uncorrected replication errors during tumor progression Local associative patterns of mutations can suggest potential cooperative processes driving the mutation signature s.

Signatures attributed to the same underlying mutagenic processes may correlate within and between individuals. Similarly, SBS8 and SBS40 show comparable trinucleotide frequencies and similar context preferences, and may be related It is possible that attributes of other unknown signatures could be predicted from their association with known signatures.

Moreover, correspondence analysis between mutational signatures generated in engineered model systems with those in human tumors can help establish etiologies of specific signatures 64 , 65 , as also discussed below. In light of these observations, it is unlikely that a single strategy will be necessary and sufficient to explain all signatures, and those signatures that are not sufficiently explainable by trinucleotide contexts alone could benefit from analyses of broader contexts and patterns.

Previous works on mutational landscape of tumor genomes and mutational signatures have primarily analyzed epigenomic contexts from closely related cell types or those contexts that are cell type invariant 8 , 48 , 63 , 66 , Unfortunately, data on epigenome and replication profile are limited to reference cell lines and tissues 68 , 69 , and it remains technically challenging to obtain similar high-quality data from primary cell types, especially from rare cell populations from normal or tumor tissues, which may have genetic and nongenetic heterogeneity.

Emerging single-cell assays are enabling multi-omics profiling on primary cell populations 70 , 71 , allowing for integrating relevant epigenomic and mutation data directly from the cell populations of interest, which may provide valuable insights about mutagenic processes in somatic cells in vivo. Reporter assays in well-characterized cell lines and model organisms can validate selected mutational signatures and provide mechanistic insights.

After the development of massively parallel sequencing technologies, targeted or whole genome sequencing of clonally derived cell populations has been used to complement reporter assays and determine genome-wide consequences of the mutagenic processes at base-pair level resolution.

Epigenetic heterogeneity in cancer

The incidence and mortality of lung cancer have increased steadily worldwide in the past several decades 1. In the previously published literature, we confirmed a number of factors associated with LUAD 3. However, for LUAD, we still have a lot of unknown. For this reason, we need better tools to describe the characteristics of LUAD and predict the prognosis. Several studies have shown that tumors are considered to be derived from the conversion of multiple genes in normal cells 4. This mutation exists in the entire process of tumor formation while these cells will continue to mutate in the tumor 4.

Cancer is a group of diseases involving abnormal cell growth with the potential to invade or spread to other parts of the body. The risk of developing certain cancers can be reduced by not smoking, maintaining a healthy weight, limiting alcohol intake, eating plenty of vegetables , fruits , and whole grains , vaccination against certain infectious diseases, limiting consumption of processed meat and red meat , and limiting exposure to direct sunlight. In , about Greek physicians Hippocrates and Galen, among others, noted similarity of crabs to some tumors with swollen veins. The word was introduced in English in the modern medical sense around

Thank you for visiting nature. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser or turn off compatibility mode in Internet Explorer. In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript. Somatic mutations in cancer genomes are caused by multiple mutational processes, each of which generates a characteristic mutational signature 1.

Mutational Heterogeneity in Human Cancers : Origin and Consequences

Although not all somatic mutations are cancer drivers, their mutational signatures, i. However, the mechanisms underlying nearly one-third of all mutational signatures are not yet understood. The signatures with established etiology and those with hitherto unknown origin appear to have some differences in strand bias, GC content and nucleotide context diversity. While nucleotide contexts might be adequate to establish etiologies of some mutational signatures, in other cases additional features, such as broader epi genomic contexts, including chromatin, replication timing, processivity and local mutational patterns, may help fully understand the underlying DNA damage and repair processes.

Despite numerous advances in cell biology, genetics, and developmental biology, cancer origin has been attributed to genetic mechanisms primarily involving mutations. Embryologists have expressed timidly cancer embryological origin with little success in leveraging the discussion that cancer could involve a set of conventional cellular processes used to build the embryo during morphogenesis. The concept of environment is often used with a broad scope and includes all nongenetic factors such as diet, lifestyle, and infectious agents. At the current juncture of the XXI century, cancer disease should not be dissociated from the environment and external stimuli, which are considered as the causes of most human cancers [ 1 , 2 ]. Also, the epigenetics consolidated a formal theory of carcinogenesis [ 6 ] that could explain cancer predisposition in humans related to epimutations an epigenetic hereditary abnormality in gene expression transmitted from mother to child [ 7 ].

Clinical relevance of mutant-allele tumor heterogeneity and lung adenocarcinoma


  • Mutational Heterogeneity in Human Cancers: Origin and Consequences The impact of DNA sequencing on future cancer research and. Midsdiferos1966 - 18.05.2021 at 15:03