CoChat
Sign Up
PreprintCited by 277RNA Research and SplicingCancer-related molecular mechanisms researchCancer Genomics and Diagnostics

The DOI matches its Crossref record; 61 of the 61 references it cites were found in a registry or an index.

Large-scale analysis of genome and transcriptome alterations in multiple tumors unveils novel cancer-relevant splicing networks

Open access
Sebestyén, Endre, Singh, Babita, Miñana, Belén, Pagès, Amadís, Mateo, Francesca, Pujana, Miguel Angel, Valcárcel, Juan, Eyras, Eduardo·2016·Genome Research
DOI 10.1101/gr.199935.115PMC PMC4889968

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Alternative splicing is regulated by multiple RNA-binding proteins and influences the expression of most eukaryotic genes. However, the role of this process in human disease, and particularly in cancer, is only starting to be unveiled. We systematically analyzed mutation, copy number, and gene expression patterns of 1348 RNA-binding protein (RBP) genes in 11 solid tumor types, together with alternative splicing changes in these tumors and the enrichment of binding motifs in the alternatively spliced sequences. Our comprehensive study reveals widespread alterations in the expression of RBP genes, as well as novel mutations and copy number variations in association with multiple alternative splicing changes in cancer drivers and oncogenic pathways. Remarkably, the altered splicing patterns in several tumor types recapitulate those of undifferentiated cells. These patterns are predicted to be mainly controlled by MBNL1 and involve multiple cancer drivers, including the mitotic gene NUMA1 . We show that NUMA1 alternative splicing induces enhanced cell proliferation and centrosome amplification in nontumorigenic mammary epithelial cells. Our study uncovers novel splicing networks that potentially contribute to cancer development and progression.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1101/gr.199935.115

Cites

61 of 61 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1. Cancer alterations in splicing factors. (A) Up- (red) and down-regulation (blue) patterns splicing factors (x-axis) in the different tumor types (y-axis) compared to normal samples. Only the 132 SFs (out of 162 tested) with differential expression are shown. The color intensity indicates the log2-fold change (log2 FC). The bar plot above indicates the frequency of tumor types with up- (red) or down- (blue) regulation for each factor. Factors are clustered into three groups according to whether they show frequent down-regulation (Downregulated) or up-regulation (Upregulated) in tumors, or whether they tend to show an opposite pattern between the three kidney tumors (KICH, KIRC, KIRP) and the rest of the tumor types (Opposing). Factors previously described to have oncogenic or tumor-suppressing activities (Supplemental Table S3) are labeled in red. SF3B1, SRPK1, SRPK2, and SRPK3 were included for comparison. (B) Number of samples in log10 scale (y-axis) in which RBPs (green), driver genes (red) (Supplemental Tables S5, S6), and the rest of the genes (blue) show mutations in each tumor type (x-axis). Distributions are represented as box plots, with outliers represented as dots. All comparisons of drivers vs. RBPs or drivers vs. others are significant (one-sided Wilcoxon test P-values < 1.7×10−5). Comparisons of others vs. RBPs are significant (one-sided Wilcoxon test P-values < 0.05), except for LUAD, LUSC, and PRAD. Drivers were extracted from the literature (Methods). (C) Copy number variation (CNV) gains (left panel) and losses (right panel) of the tested splicing factors. The size of the circle corresponds to the proportion of samples with CNVs and darker colors represent cases where >50% of the CNVs are focal. Only those CNVs with a frequency of amplification > 5% or deletion > 1% are shown.
Figure 1. Cancer alterations in splicing factors. (A) Up- (red) and down-regulation (blue) patterns splicing factors (x-axis) in the different tumor types (y-axis) compared to normal samples. Only the 132 SFs (out of 162 tested) with differential expression are shown. The color intensity indicates the log2-fold change (log2 FC). The bar plot above indicates the frequency of tumor types with up- (red) or down- (blue) regulation for each factor. Factors are clustered into three groups according to whether they show frequent down-regulation (Downregulated) or up-regulation (Upregulated) in tumors, or whether they tend to show an opposite pattern between the three kidney tumors (KICH, KIRC, KIRP) and the rest of the tumor types (Opposing). Factors previously described to have oncogenic or tumor-suppressing activities (Supplemental Table S3) are labeled in red. SF3B1, SRPK1, SRPK2, and SRPK3 were included for comparison. (B) Number of samples in log10 scale (y-axis) in which RBPs (green), driver genes (red) (Supplemental Tables S5, S6), and the rest of the genes (blue) show mutations in each tumor type (x-axis). Distributions are represented as box plots, with outliers represented as dots. All comparisons of drivers vs. RBPs or drivers vs. others are significant (one-sided Wilcoxon test P-values < 1.7×10−5). Comparisons of others vs. RBPs are significant (one-sided Wilcoxon test P-values < 0.05), except for LUAD, LUSC, and PRAD. Drivers were extracted from the literature (Methods). (C) Copy number variation (CNV) gains (left panel) and losses (right panel) of the tested splicing factors. The size of the circle corresponds to the proportion of samples with CNVs and darker colors represent cases where >50% of the CNVs are focal. Only those CNVs with a frequency of amplification > 5% or deletion > 1% are shown.CC BY-NC · PubMed Central
Figure 2. Differentially spliced events in tumors. (A, upper panel) Number of paired-samples used per tumor type. (Lower panel) Number of differentially spliced events per tumor type compared to normal samples, split according to type of event: alternative 3′ splice-site (A3), alternative 5′ splice-site (A5), mutually exclusive exon (MX), retained intron (RI), and skipping exon (SE) (Supplemental Table S6). (B) Proportion of driver and nondriver genes with differentially spliced events. We indicated in red those tumors for which the enrichment is significant. (C) Cancer hallmarks (x-axis) that are enriched (Fisher's exact test P-value < 0.05) in differentially spliced events in each tumor type (y-axis). The color indicates the odds ratio of the enrichment. Hallmarks that are also enriched according to gene expression are indicated with a black dot. (D) Number of differentially spliced events related to protein-affecting mutations in RBP genes color-labeled by tumor type. Only cases with at least 10 associated differentially spliced events are shown. SF3B1 is included for comparison. (E) Proportion of samples per tumor type with protein-affecting mutations in RBP genes with at least 10 associated differentially spliced events. (F) Number of protein-affecting mutations (y-axis) along the HNRNPL protein (x-axis), color-labeled according to whether they are substitutions, insertions, or deletions. Protein domains are indicated in light red. (G) Distribution of PSI values for the A5 event in CASP8 associated with the mutations of HNRNPL in COAD, separated into normal samples (Norm), tumor samples without protein-affecting mutations (T–NM), and tumor samples with protein-affecting mutations (T–M). (H) Enrichment or depletion of specific event types in association with mutations in HNRNPL (red bars) compared to the overall proportions of events (black bars). Significant differences (P < 0.05, Fisher's exact test) are labeled in red. Contingency tables are provided in Supplemental Table S12.
Figure 2. Differentially spliced events in tumors. (A, upper panel) Number of paired-samples used per tumor type. (Lower panel) Number of differentially spliced events per tumor type compared to normal samples, split according to type of event: alternative 3′ splice-site (A3), alternative 5′ splice-site (A5), mutually exclusive exon (MX), retained intron (RI), and skipping exon (SE) (Supplemental Table S6). (B) Proportion of driver and nondriver genes with differentially spliced events. We indicated in red those tumors for which the enrichment is significant. (C) Cancer hallmarks (x-axis) that are enriched (Fisher's exact test P-value < 0.05) in differentially spliced events in each tumor type (y-axis). The color indicates the odds ratio of the enrichment. Hallmarks that are also enriched according to gene expression are indicated with a black dot. (D) Number of differentially spliced events related to protein-affecting mutations in RBP genes color-labeled by tumor type. Only cases with at least 10 associated differentially spliced events are shown. SF3B1 is included for comparison. (E) Proportion of samples per tumor type with protein-affecting mutations in RBP genes with at least 10 associated differentially spliced events. (F) Number of protein-affecting mutations (y-axis) along the HNRNPL protein (x-axis), color-labeled according to whether they are substitutions, insertions, or deletions. Protein domains are indicated in light red. (G) Distribution of PSI values for the A5 event in CASP8 associated with the mutations of HNRNPL in COAD, separated into normal samples (Norm), tumor samples without protein-affecting mutations (T–NM), and tumor samples with protein-affecting mutations (T–M). (H) Enrichment or depletion of specific event types in association with mutations in HNRNPL (red bars) compared to the overall proportions of events (black bars). Significant differences (P < 0.05, Fisher's exact test) are labeled in red. Contingency tables are provided in Supplemental Table S12.CC BY-NC · PubMed Central
Figure 3. Common and specific events in tumors. (A) Common events and splicing factors between pairs of tumor types. For each pair of tumor types and for each splicing factor differentially expressed in both tumor types, we plot the correlation of ΔPSI values for events that have a correlation of |R| > 0.5 (Spearman) with these splicing factors in both tumor types. Only factors with more than 50 associated events in both tumor types are shown. Each event is only plotted once, and the color of the plot corresponds to the most common correlating splicing factor. Correlations between ΔPSI values are indicated. In red or green, we highlight those higher than 0.8 or lower than −0.8, respectively. (B) ΔPSI correlations for the pairs LUAD–BRCA, PRAD–KIRC, KIRC–HNSC, and LUSC–HNSC, for the common events separated according to their potential splicing factor regulators. Points representing the same event associated with more than one factor are plotted with a slight displacement to avoid overlap. (C) Principal Component Analysis (PCA) plot of 380 tumor-specific alternative splicing events colored by tumor type.
Figure 3. Common and specific events in tumors. (A) Common events and splicing factors between pairs of tumor types. For each pair of tumor types and for each splicing factor differentially expressed in both tumor types, we plot the correlation of ΔPSI values for events that have a correlation of |R| > 0.5 (Spearman) with these splicing factors in both tumor types. Only factors with more than 50 associated events in both tumor types are shown. Each event is only plotted once, and the color of the plot corresponds to the most common correlating splicing factor. Correlations between ΔPSI values are indicated. In red or green, we highlight those higher than 0.8 or lower than −0.8, respectively. (B) ΔPSI correlations for the pairs LUAD–BRCA, PRAD–KIRC, KIRC–HNSC, and LUSC–HNSC, for the common events separated according to their potential splicing factor regulators. Points representing the same event associated with more than one factor are plotted with a slight displacement to avoid overlap. (C) Principal Component Analysis (PCA) plot of 380 tumor-specific alternative splicing events colored by tumor type.CC BY-NC · PubMed Central
Figure 4. Enriched RNA binding motifs in differentially spliced events. (A) Enriched RNA binding motifs in differentially spliced skipping exon events in each tumor type, separated by inclusion (upper panels) or skipping (lower panels) events, and by upstream (left), exonic (middle), or downstream (right) regions. Only enriched motifs for splicing factors that are differentially expressed are indicated in each tumor type. RBPs (y-axis) are grouped according to Figure 1A. Gene up- and down-regulation is indicated in red and blue, respectively. The color intensity indicates the Z-score of the motif enrichment. Similar plots for the other event types are given in Supplemental Figures S14–S18. (B) Proportion of enriched motifs in inclusion (ΔPSI > 0.1) (left panel) and skipping (ΔPSI < 0.1) (right panel) events, in each of the event regions (x-axis): upstream (U), exon (E), and downstream (D). Proportions are separated according to whether the RBP gene is up- (red) or down- (blue) regulated. (C) Total proportion (y-axis) of differentially spliced events in each tumor type (x-axis) that are predicted as targets of one or more differentially expressed RBPs with significance Z-score > 1.96. (D) Proportion of differentially spliced events (marked in green) that are predicted as targets of each RBP (x-axis) in each tumor type (y-axis), with significance Z-score > 1.96. RBPs are grouped according to Figure 1A. Upper panels indicate the number of unique cancer drivers with differentially spliced events predicted as targets of each RBP across all tumors. RBPs are ordered according to this value within each group. (E) Correlation (Pearson R) of ΔPSI values (x-axes) in breast tumors (BRCA) and prostate tumors (PRAD) with the ΔPSI of events (y-axes) from the knockdown of MBNL1 and MBNL2 in HeLa cells (lower panels) and from the comparison of stem cells (ESCs) with differentiated cells (CL) (upper panels). Events with a predicted MBNL binding motif are indicated in blue.
Figure 4. Enriched RNA binding motifs in differentially spliced events. (A) Enriched RNA binding motifs in differentially spliced skipping exon events in each tumor type, separated by inclusion (upper panels) or skipping (lower panels) events, and by upstream (left), exonic (middle), or downstream (right) regions. Only enriched motifs for splicing factors that are differentially expressed are indicated in each tumor type. RBPs (y-axis) are grouped according to Figure 1A. Gene up- and down-regulation is indicated in red and blue, respectively. The color intensity indicates the Z-score of the motif enrichment. Similar plots for the other event types are given in Supplemental Figures S14–S18. (B) Proportion of enriched motifs in inclusion (ΔPSI > 0.1) (left panel) and skipping (ΔPSI < 0.1) (right panel) events, in each of the event regions (x-axis): upstream (U), exon (E), and downstream (D). Proportions are separated according to whether the RBP gene is up- (red) or down- (blue) regulated. (C) Total proportion (y-axis) of differentially spliced events in each tumor type (x-axis) that are predicted as targets of one or more differentially expressed RBPs with significance Z-score > 1.96. (D) Proportion of differentially spliced events (marked in green) that are predicted as targets of each RBP (x-axis) in each tumor type (y-axis), with significance Z-score > 1.96. RBPs are grouped according to Figure 1A. Upper panels indicate the number of unique cancer drivers with differentially spliced events predicted as targets of each RBP across all tumors. RBPs are ordered according to this value within each group. (E) Correlation (Pearson R) of ΔPSI values (x-axes) in breast tumors (BRCA) and prostate tumors (PRAD) with the ΔPSI of events (y-axes) from the knockdown of MBNL1 and MBNL2 in HeLa cells (lower panels) and from the comparison of stem cells (ESCs) with differentiated cells (CL) (upper panels). Events with a predicted MBNL binding motif are indicated in blue.CC BY-NC · PubMed Central
Figure 5. Networks of splicing regulation. Modules of alternative splicing regulation according to cancer hallmarks in breast (A) and colon (B) tumors. For each cluster of RBPs (x-axis), we indicate in gray the total number of gene targets linked to them in each hallmark (y-axis). Only enriched hallmarks are shown. We indicate in red the number of cancer drivers associated with each RBP, and in each cluster RBPs are ordered according to the total number of genes they are associated with. Representation of the regulatory modules for G2M checkpoint (C) and WNT/beta-catenin (D) hallmarks in breast tumors, and for the angiogenesis hallmark (E) in colon tumors. RBPs are indicated as square boxes in red or blue depending on whether they are up- or down-regulated. Target genes are presented as white diamonds for cancer drivers and white boxes for the rest. Connections indicate predicted splicing regulation by an RBP.
Figure 5. Networks of splicing regulation. Modules of alternative splicing regulation according to cancer hallmarks in breast (A) and colon (B) tumors. For each cluster of RBPs (x-axis), we indicate in gray the total number of gene targets linked to them in each hallmark (y-axis). Only enriched hallmarks are shown. We indicate in red the number of cancer drivers associated with each RBP, and in each cluster RBPs are ordered according to the total number of genes they are associated with. Representation of the regulatory modules for G2M checkpoint (C) and WNT/beta-catenin (D) hallmarks in breast tumors, and for the angiogenesis hallmark (E) in colon tumors. RBPs are indicated as square boxes in red or blue depending on whether they are up- or down-regulated. Target genes are presented as white diamonds for cancer drivers and white boxes for the rest. Connections indicate predicted splicing regulation by an RBP.CC BY-NC · PubMed Central
Figure 6. Regulation of NUMA1 alternative splicing by MBNL1 in breast luminal tumors. (A) PSI value distributions in tumor and paired normal sample for luminal A (LA) and luminal B (LB) breast tumors for the events in NUMA1 (LA: ΔPSI = −0.22, P-value = 7.81 × 10−7, LB: ΔPSI = −0.23, P-value = 0.037) and NUMB (LA: ΔPSI = 0.28, P-value = 0.0001, LB: ΔPSI = 0.28, P-value = 0.016). All P-values given are corrected for multiple testing. (B) Semiquantitative RT-PCR isoform analysis upon knockdowns of MBNL1 (lanes 2–10) and QKI (lanes 12–14) at different siRNA concentrations and the controls with scrambled siRNAs (lanes 1,11). The diagrams to the right indicate the position of the alternatively spliced exons. The bottom lanes correspond to the RT-PCR amplification of RNA from the PUM gene in the same samples, which are used as a control for RNA expression. (C) Resazurin-based assays of cell viability/proliferation. Measurements were performed in triplicate at 72, 96, and 120 h. The plot shows measurements upon knockdowns of MBNL1 (siMBNL1) and QKI (siQKI), upon transfection of AONs targeting the 3′ and 5′ splice-sites independently and both together, and the corresponding controls (scrambled siRNA and random AON). (D, left panel) Graph showing the results of the evaluation of centrosome amplification upon knockdown of MBNL1 (siMBNL1) or upon transfection of AONs targeting 5′ splice-sites (5′ ss AON), compared to the corresponding controls siScrambled (P = 0.4271) and random AON (P = 0.04356), respectively (one-sided Mann-Whitney U test). (Right panels) Representative merged (TUBG1 and DAPI) images of immunofluorescence assays. (E) Correlation of NUMA1 event PSI (x-axis) with the CIN25 signature of aneuploidy (y-axis) across the tumor (red) and normal (blue) samples for luminal A (upper panel) (R = −0.4 Spearman) and B (lower panel) (R = −0.33 Spearman).
Figure 6. Regulation of NUMA1 alternative splicing by MBNL1 in breast luminal tumors. (A) PSI value distributions in tumor and paired normal sample for luminal A (LA) and luminal B (LB) breast tumors for the events in NUMA1 (LA: ΔPSI = −0.22, P-value = 7.81 × 10−7, LB: ΔPSI = −0.23, P-value = 0.037) and NUMB (LA: ΔPSI = 0.28, P-value = 0.0001, LB: ΔPSI = 0.28, P-value = 0.016). All P-values given are corrected for multiple testing. (B) Semiquantitative RT-PCR isoform analysis upon knockdowns of MBNL1 (lanes 2–10) and QKI (lanes 12–14) at different siRNA concentrations and the controls with scrambled siRNAs (lanes 1,11). The diagrams to the right indicate the position of the alternatively spliced exons. The bottom lanes correspond to the RT-PCR amplification of RNA from the PUM gene in the same samples, which are used as a control for RNA expression. (C) Resazurin-based assays of cell viability/proliferation. Measurements were performed in triplicate at 72, 96, and 120 h. The plot shows measurements upon knockdowns of MBNL1 (siMBNL1) and QKI (siQKI), upon transfection of AONs targeting the 3′ and 5′ splice-sites independently and both together, and the corresponding controls (scrambled siRNA and random AON). (D, left panel) Graph showing the results of the evaluation of centrosome amplification upon knockdown of MBNL1 (siMBNL1) or upon transfection of AONs targeting 5′ splice-sites (5′ ss AON), compared to the corresponding controls siScrambled (P = 0.4271) and random AON (P = 0.04356), respectively (one-sided Mann-Whitney U test). (Right panels) Representative merged (TUBG1 and DAPI) images of immunofluorescence assays. (E) Correlation of NUMA1 event PSI (x-axis) with the CIN25 signature of aneuploidy (y-axis) across the tumor (red) and normal (blue) samples for luminal A (upper panel) (R = −0.4 Spearman) and B (lower panel) (R = −0.33 Spearman).CC BY-NC · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 61 references

Cited by

Sorted by citations

See all 277

Related

Metadata

Authors
Sebestyén, E., Singh, B., Miñana, B., Pagès, A., Mateo, F., Pujana, M. A., Valcárcel, J., & Eyras, E.
Affiliations
  • Universitat Pompeu Fabra (ES)
  • Institut d'Investigació Biomédica de Bellvitge (ES)
  • Institució Catalana de Recerca i Estudis Avançats (ES)
ORCID
Venue
Genome Research
Publisher
Cold Spring Harbor Laboratory
Published
2016-04-13
Volume
26
Issue
6
Pages
732-744
ISSN
  • 1088-9051 (ISSN-L)
  • 1549-5469
DOI
10.1101/gr.199935.115
PMC
PMC4889968
OpenAlex
W2338046461
Cited by
276
References
61
Citations by year
2026: 12 · 2025: 18 · 2024: 25 · 2023: 21 · 2022: 38 · 2021: 45 · 2020: 43 · 2019: 27 · 2018: 22 · 2017: 22 · 2016: 3
Type
posted content
Language
en
Open access
Locations
  • Genome Research (publishedVersion)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc)
  • PubMed Central (publishedVersion, cc-by-nc)
  • PubMed Central (publishedVersion)
  • PubMed (publishedVersion)
  • Repositori digital de la UPF (Universitat Pompeu Fabra) (publishedVersion, cc-by-nc)
  • HighWire Press Open Archive (submittedVersion)
  • RECERCAT (Consorci de Serveis Universitaris de Catalunya) (submittedVersion, cc-by-nc)
Funders
  • Ministerio de Economía y Competitividad
  • European Commission
  • FEDER (BIO2014-52566-R, PIE13/00022-ONCOPROFILE, PI15/00854, RD12/0036/0008)
  • Consolider RNAREG (CSD2009-00080)
  • Agència de Gestió d'Ajuts Universitaris i de Recerca (SGR2014-1121)
  • Sandra Ibarra Foundation for Cancer (FSI2013)
  • Fundación Botín
  • Santander Universities Global Division
  • MINECO
  • AGAUR (SGR2014-364)
  • AECC
  • Instituto de Salud Carlos III
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › RNA Research and Splicing
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Cancer-related molecular mechanisms research
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Cancer Genomics and Diagnostics
Keywords
alternative splicing, RNA-binding proteins, cancer, solid tumors, MBNL1, NUMA1, copy number variation, gene expression changes, binding motifs, centrosome amplification, cell proliferation, mammary epithelial cells, undifferentiated cells
MeSH terms
RNA Splicing Factors, Binding Sites, Humans, Mutation, Neoplasms, Genome, Human, Gene Expression Regulation, Neoplastic, Alternative Splicing, Amino Acid Motifs, Gene Expression Profiling, Cell Proliferation, Transcriptome

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review