Recurrent disruption of tumour suppressor genes in cancer by somatic mutations in cleavage and polyadenylation signals
Curation statements for this article:-
Curated by eLife
eLife assessment
To test if somatic mutations in cancer genomes are enriched with mutations in polyadenylation signal regions, the authors observed an increased enrichment of somatic mutations that may affect the function of polyA signals and confirmed that these mutations may influence gene expression through a minigene expression experiment. This important study advances our understanding of noncoding somatic mutations by identifying a novel class of mutations that affect 3'UTR polyadenylation signals enriched in tumor suppressor genes in cancer. The evidence supporting the conclusions is convincing, with rigorous statistical analyses and experimental validation.
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
The expression of eukaryotic genes relies on the precise 3’-terminal cleavage and polyadenylation of newly synthesized pre-mRNA transcripts. Defects in these processes have been associated with various diseases, including cancer. While cancer-focused sequencing studies have identified numerous driver mutations in protein-coding sequences, noncoding drivers – particularly those affecting the cis-elements required for pre-mRNA cleavage and polyadenylation – have received less attention. Here, we systematically analysed cancer somatic mutations affecting 3’UTR polyadenylation signals using the Pan-Cancer Analysis of Whole Genomes (PCAWG) dataset. We found a striking enrichment of cancer-specific somatic mutations that disrupt strong and evolutionarily conserved cleavage and polyadenylation signals within tumour suppressor genes. Further bioinformatics and experimental analyses conducted as a part of our study suggest that these mutations have a profound capacity to downregulate the expression of tumour suppressor genes. Thus, this work uncovers a novel class of noncoding somatic mutations with significant potential to drive cancer progression.
Article activity feed
-
eLife assessment
To test if somatic mutations in cancer genomes are enriched with mutations in polyadenylation signal regions, the authors observed an increased enrichment of somatic mutations that may affect the function of polyA signals and confirmed that these mutations may influence gene expression through a minigene expression experiment. This important study advances our understanding of noncoding somatic mutations by identifying a novel class of mutations that affect 3'UTR polyadenylation signals enriched in tumor suppressor genes in cancer. The evidence supporting the conclusions is convincing, with rigorous statistical analyses and experimental validation.
-
Reviewer #1 (Public Review):
Kainov et al investigated the prevalence of mutations in 3'UTR that affect gene expression in cancer to identify noncoding cancer drivers.
The authors used data from normal controls (1000 genome data) and compared it to cancer data (PCAWG). They found that in cancer 3'UTR mutations had a stronger effect on cleavage than the normal population. These mutations are negatively selected in the normal population and positively selected in cancers. The authors used PCAWG data set to identify such mutations and found that the mutations that lead to a reduction of gene expression are enriched in tumor suppressor genes and those that are increased in gene expression are enriched for oncogenes. 3'UTR mutations that reduce gene expression or occur in TSGs co-occur with non-synonymous mutations. The authors then validate …
Reviewer #1 (Public Review):
Kainov et al investigated the prevalence of mutations in 3'UTR that affect gene expression in cancer to identify noncoding cancer drivers.
The authors used data from normal controls (1000 genome data) and compared it to cancer data (PCAWG). They found that in cancer 3'UTR mutations had a stronger effect on cleavage than the normal population. These mutations are negatively selected in the normal population and positively selected in cancers. The authors used PCAWG data set to identify such mutations and found that the mutations that lead to a reduction of gene expression are enriched in tumor suppressor genes and those that are increased in gene expression are enriched for oncogenes. 3'UTR mutations that reduce gene expression or occur in TSGs co-occur with non-synonymous mutations. The authors then validate the effect of 3'UTR mutations experimentally using a luciferase reporter assay. These data identify a novel class of noncoding driver genes with mutations in 3'UTR that impact polyadenylation and thus gene expression.
This is an elegant study with fundamental insight into identifying cancer driver genes. The conclusions of this paper are mostly well supported by data, but some aspects of data analysis need to be extended.
(1) It would be important for the authors to show if the findings of this study hold for metastatic cancers since most deaths occur due to metastasis and tumor heterogeneity changes when cancer progresses to metastasis. The authors should use the Hartwig data and show if metastatic cancers are enriched for 3'UTR mutations.
(2) Figure 2 should show the distribution of 3'UTR mutations by cancer type especially since authors go on to use colorectal cancer only for validations. It would be helpful to bring Figures S3A and S3C to this panel since these findings make the connections to cancer biology. Are any molecular functions enriched in addition to biological processes? Are kinases, phosphatases, etc more or less affected by 3'UTR mutations?
(3) Figure 3 looks at the co-occurrence of 3'UTR mutations with non-synonymous mutations but what about copy number change? You would expect the loss of the other allele to be enriched. Along the same line, are these data phased? Do you know that the non-synonymous mutations are in the other allele or in the same allele that shows 3'UTR mutation?
-
Reviewer #2 (Public Review):
Summary:
To evaluate whether somatic mutations in cancer genomes are enriched with mutations in polyadenylation signal regions, the authors analyzed 1000 genomes data and PCAWG data as a control and experimental set, respectively. They observed increased enrichment of somatic mutations that may affect the function of polyA signals and confirmed that these mutations may influence the expression of the gene through a minigene expression experiment.
Strengths:
This study provides a systematic evaluation of polyA signal, which makes it valuable. Overall, the analytic approach and results are solid and supported by experimental validation.
Weaknesses:
(1) This study uses APARENT2 as a tool to evaluate functional alteration in polyA signal sequences. Based on the original paper and the results shown in this paper, …
Reviewer #2 (Public Review):
Summary:
To evaluate whether somatic mutations in cancer genomes are enriched with mutations in polyadenylation signal regions, the authors analyzed 1000 genomes data and PCAWG data as a control and experimental set, respectively. They observed increased enrichment of somatic mutations that may affect the function of polyA signals and confirmed that these mutations may influence the expression of the gene through a minigene expression experiment.
Strengths:
This study provides a systematic evaluation of polyA signal, which makes it valuable. Overall, the analytic approach and results are solid and supported by experimental validation.
Weaknesses:
(1) This study uses APARENT2 as a tool to evaluate functional alteration in polyA signal sequences. Based on the original paper and the results shown in this paper, the algorithm appears to be of high quality. However, the whole study is dependent on the output of APARENT2. Therefore, it would be nice to
(a) run and show a positive control run, which can show that the algorithm works well, and
(b) describe the rationale for selecting this algorithm in the main text.(2) Are there recurrent somatic mutation calls (= exactly the same mutation across different tumor samples) in the poly(A) region of certain genes?
(3) The authors nicely showed that the minigene with A>G mutation altered gene expression. Maybe one can reach a similar conclusion by analyzing a cancer dataset that has mutation and gene expression data? That is, genes with or without polyA mutations show different expression levels.
-
Author response:
We thank both reviewers for their constructive comments. We will do our best incorporating the requested analyses and answering reviewers’ questions in the revision
-
-
-
