CoChat
Sign Up
Cited by 162Breast Cancer Treatment StudiesAdvanced Proteomics Techniques and ApplicationsBioinformatics and Genomic Networks

The DOI matches its Crossref record; 37 of the 37 references it cites were found in a registry or an index.

Integrated proteotranscriptomics of breast cancer reveals globally increased protein-mRNA concordance associated with subtypes and survival

Open access
Tang, Wei, Zhou, Ming, Dorsey, Tiffany H., Prieto, DaRue A., Wang, Xin W., Ruppin, Eytan, Veenstra, Timothy D., Ambs, Stefan·2018·Genome Medicine
DOI 10.1186/s13073-018-0602-xPMC PMC6276229
Open this paper in CoChat — free accountOpen PDF

Abstract

Background: Transcriptome analysis of breast cancer discovered distinct disease subtypes of clinical significance. However, it remains a challenge to define disease biology solely based on gene expression because tumor biology is often the result of protein function. Here, we measured global proteome and transcriptome expression in human breast tumors and adjacent non-cancerous tissue and performed an integrated proteotranscriptomic analysis. Methods: We applied a quantitative liquid chromatography/mass spectrometry-based proteome analysis using an untargeted approach and analyzed protein extracts from 65 breast tumors and 53 adjacent non-cancerous tissues. Additional gene expression data from Affymetrix Gene Chip Human Gene ST Arrays were available for 59 tumors and 38 non-cancerous tissues in our study. We then applied an integrated analysis of the proteomic and transcriptomic data to examine relationships between them, disease characteristics, and patient survival. Findings were validated in a second dataset using proteome and transcriptome data from “The Cancer Genome Atlas” and the Clinical Proteomic Tumor Analysis Consortium. Results: We found that the proteome describes differences between cancerous and non-cancerous tissues that are not revealed by the transcriptome. The proteome, but not the transcriptome, revealed an activation of infection-related signal pathways in basal-like and triple-negative tumors. We also observed that proteins rather than mRNAs are increased in tumors and show that this observation could be related to shortening of the 3′ untranslated region of mRNAs in tumors. The integrated analysis of the two technologies further revealed a global increase in protein-mRNA concordance in tumors. Highly correlated protein-gene pairs were enriched in protein processing and disease metabolic pathways. The increased concordance between transcript and protein levels was additionally associated with aggressive disease, including basal-like/triple-negative tumors, and decreased patient survival. We also uncovered a strong positive association between protein-mRNA concordance and proliferation of tumors. Finally, we observed that protein expression profiles co-segregate with a Myc activation signature and separate breast tumors into two subgroups with different survival outcomes. Conclusions: Our study provides new insights into the relationship between protein and mRNA expression in breast cancer and shows that an integrated analysis of the proteome and transcriptome has the potential of uncovering novel disease characteristics.

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 · 10.1186/s13073-018-0602-x

Cites

37 of 37 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Quantification of proteins in the breast tissues. a Mass spectrometry-based proteomics workflow. b Dynamic range of protein expression in our study. c Overlap in the identified proteins between our study (Tang et al.) and Mertins et al. [13]. d Abundance of individual proteins correlates between our study and Mertins et al. (Spearman’s rank correlation rho = 0.56). Correlation is better for high than low abundance proteins. Scaled total spectral counts (z-scaled log converted PSMs) were plotted on the x- and y-axes. e A principal component analysis (PCA) based on the abundance of 7141 proteins in 118 tissues (breast tumors n = 65; adjacent non-cancerous tissues n = 53). The two-dimension PCA plot shows distinct clustering of the tumor (red) and non-cancerous (blue) tissues
Fig. 1 Quantification of proteins in the breast tissues. a Mass spectrometry-based proteomics workflow. b Dynamic range of protein expression in our study. c Overlap in the identified proteins between our study (Tang et al.) and Mertins et al. [13]. d Abundance of individual proteins correlates between our study and Mertins et al. (Spearman’s rank correlation rho = 0.56). Correlation is better for high than low abundance proteins. Scaled total spectral counts (z-scaled log converted PSMs) were plotted on the x- and y-axes. e A principal component analysis (PCA) based on the abundance of 7141 proteins in 118 tissues (breast tumors n = 65; adjacent non-cancerous tissues n = 53). The two-dimension PCA plot shows distinct clustering of the tumor (red) and non-cancerous (blue) tissuesCC BY · PubMed Central
Fig. 2 Pathway association of differentially expressed proteins and transcripts. a Enrichment of differentially expressed proteins between tumors and adjacent non-cancerous tissue (n = 2643; 2165 upregulated and 478 downregulated tumor proteins) in KEGG pathways. Enriched pathways at ≤ 10% FDR are shown for two categories, proteins upregulated (red) and downregulated (green) in tumors. The red and green bars highlight the pathways with enrichment for proteins without a similar enrichment for differentially expressed mRNAs. The gray bars indicate those pathways that were enriched for both proteins and mRNA in tumors. Several metabolism pathways, such as the pentose phosphate pathway, starch and sucrose metabolism, gluconeogenesis and glycolysis, galactose metabolism, and glutathione metabolism, were only enriched for upregulated proteins in tumors but not upregulated mRNAs. Four other pathways, TCA cycle, gap junction, pyruvate metabolism, and long-term depression, showed enrichment for proteins upregulated in tumors whereas pathway-associated mRNAs tended to be downregulated in these same tumors (green-bordered bars). For analysis, all proteins were ranked using Wald statistic and imported into the GSEA pre-ranked module. The KEGG gene sets of MSigDB were selected as the reference database. b Heatmap showing subtype-specific enrichment of proteins in KEGG BRITE categories and pathways. Hierarchical clustering using the pathway enrichment scores [FDR-based] yielded three branches with differential enrichment of proteins in KEGG pathways by breast cancer molecular subtype (luminal A, HER2-positive, ER-negative, triple-negative, and basal-like). Shown are unique and common features between subtypes. Red indicates upregulated proteins that are significantly enriched in pathways, whereas blue indicates downregulated proteins with pathway enrichment
Fig. 2 Pathway association of differentially expressed proteins and transcripts. a Enrichment of differentially expressed proteins between tumors and adjacent non-cancerous tissue (n = 2643; 2165 upregulated and 478 downregulated tumor proteins) in KEGG pathways. Enriched pathways at ≤ 10% FDR are shown for two categories, proteins upregulated (red) and downregulated (green) in tumors. The red and green bars highlight the pathways with enrichment for proteins without a similar enrichment for differentially expressed mRNAs. The gray bars indicate those pathways that were enriched for both proteins and mRNA in tumors. Several metabolism pathways, such as the pentose phosphate pathway, starch and sucrose metabolism, gluconeogenesis and glycolysis, galactose metabolism, and glutathione metabolism, were only enriched for upregulated proteins in tumors but not upregulated mRNAs. Four other pathways, TCA cycle, gap junction, pyruvate metabolism, and long-term depression, showed enrichment for proteins upregulated in tumors whereas pathway-associated mRNAs tended to be downregulated in these same tumors (green-bordered bars). For analysis, all proteins were ranked using Wald statistic and imported into the GSEA pre-ranked module. The KEGG gene sets of MSigDB were selected as the reference database. b Heatmap showing subtype-specific enrichment of proteins in KEGG BRITE categories and pathways. Hierarchical clustering using the pathway enrichment scores [FDR-based] yielded three branches with differential enrichment of proteins in KEGG pathways by breast cancer molecular subtype (luminal A, HER2-positive, ER-negative, triple-negative, and basal-like). Shown are unique and common features between subtypes. Red indicates upregulated proteins that are significantly enriched in pathways, whereas blue indicates downregulated proteins with pathway enrichmentCC BY · PubMed Central
Fig. 3 Correlation between steady-state protein and mRNA abundance in breast tumors and association with subtypes and survival. a Density plot showing the global Spearman correlation for protein-mRNA pairs within breast tumors (n = 59; 5677 protein-mRNA pairs) and adjacent non-cancerous tissue (n = 38; 3316 pairs). Protein and mRNA abundance was positively correlated in all tissues samples with a mean Spearman’s correlation coefficient (rho) of 0.31 in tumors, which was significantly higher (P = 1.5 × 10−14, Wilcoxon rank sum test) than the rho of 0.19 in adjacent non-cancerous tissues. b Protein-mRNA pairs have considerably different global correlations among breast cancer subtypes (triple-negative, basal-like, HER2, and luminal A). Triple-negative tumors (TNBC) and tumors of the basal-like subtype (Basal) had the highest mean correlation while luminal A tumors (LumA) had the lowest correlation. c Concordance between protein and transcript levels varies by tumor grade. Poorly differentiated grade III tumors had the highest mean correlation between protein and mRNA pairs. b, c P values were calculated using the Kruskal Wallis test. d, e Estrogen receptor (ER)-negative tumors and tumors from African-American patients show an increased concordance between protein and mRNA abundance. f Global protein-mRNA concordance values closely correlate with the proliferation score in the 59 tumors (r = 0.7; P = 8 × 10−10, Pearson correlation) but not in the 38 adjacent non-cancerous tissues (g). The dashed line shows regression. h High concordance between protein-mRNA pairs in tumors is associated with decreased breast cancer survival. Tumors were stratified into three groups according to their global correlation coefficient across all protein-mRNA pairs [< 25%, 25–75%, and > 75% (tumors with highest global correlation)]. Shown is a Kaplan-Meier plot for these three groups. Tumors with the highest correlations between protein and mRNA pairs had the worst outcome [Cox regression hazard ratio (HR) comparing 75% vs. 25% tumor group = 6.91, 95% CI 1.5–31.7 adjusted for disease subtypes; Ptrend across the three groups = 0.004]
Fig. 3 Correlation between steady-state protein and mRNA abundance in breast tumors and association with subtypes and survival. a Density plot showing the global Spearman correlation for protein-mRNA pairs within breast tumors (n = 59; 5677 protein-mRNA pairs) and adjacent non-cancerous tissue (n = 38; 3316 pairs). Protein and mRNA abundance was positively correlated in all tissues samples with a mean Spearman’s correlation coefficient (rho) of 0.31 in tumors, which was significantly higher (P = 1.5 × 10−14, Wilcoxon rank sum test) than the rho of 0.19 in adjacent non-cancerous tissues. b Protein-mRNA pairs have considerably different global correlations among breast cancer subtypes (triple-negative, basal-like, HER2, and luminal A). Triple-negative tumors (TNBC) and tumors of the basal-like subtype (Basal) had the highest mean correlation while luminal A tumors (LumA) had the lowest correlation. c Concordance between protein and transcript levels varies by tumor grade. Poorly differentiated grade III tumors had the highest mean correlation between protein and mRNA pairs. b, c P values were calculated using the Kruskal Wallis test. d, e Estrogen receptor (ER)-negative tumors and tumors from African-American patients show an increased concordance between protein and mRNA abundance. f Global protein-mRNA concordance values closely correlate with the proliferation score in the 59 tumors (r = 0.7; P = 8 × 10−10, Pearson correlation) but not in the 38 adjacent non-cancerous tissues (g). The dashed line shows regression. h High concordance between protein-mRNA pairs in tumors is associated with decreased breast cancer survival. Tumors were stratified into three groups according to their global correlation coefficient across all protein-mRNA pairs [< 25%, 25–75%, and > 75% (tumors with highest global correlation)]. Shown is a Kaplan-Meier plot for these three groups. Tumors with the highest correlations between protein and mRNA pairs had the worst outcome [Cox regression hazard ratio (HR) comparing 75% vs. 25% tumor group = 6.91, 95% CI 1.5–31.7 adjusted for disease subtypes; Ptrend across the three groups = 0.004]CC BY · PubMed Central
Fig. 4 Global concordance between protein-mRNA pairs in TCGA breast tumors and association with molecular subtypes and survival. a Density plot showing global Spearman correlation for 70 protein-mRNA pairs within 404 TCGA breast tumors. b Significantly different global correlations for protein-mRNA pairs among the molecular breast cancer subtypes (P = 3 × 10−8, Kruskal Wallis test), with basal-like tumors having the highest mean correlation. Shown are the PAM50-defined subtypes for TCGA tumors. c Global concordance between protein-mRNA pairs in TCGA tumors is associated with survival. Stratification of tumors into three groups [< 25%, 25–75%, and > 75% (tumors with highest global correlation)]. Tumors with the highest mean correlation between protein and mRNA pairs had the worst outcome [HR 75% vs. 25% tumor group = 2.6, 95% CI 1.01–6.65; Ptrend across the three groups = 0.043]. d Increased global concordance between protein-mRNA pairs is associated with reduced survival in patients with luminal A tumors. Ptrend = 0.0083
Fig. 4 Global concordance between protein-mRNA pairs in TCGA breast tumors and association with molecular subtypes and survival. a Density plot showing global Spearman correlation for 70 protein-mRNA pairs within 404 TCGA breast tumors. b Significantly different global correlations for protein-mRNA pairs among the molecular breast cancer subtypes (P = 3 × 10−8, Kruskal Wallis test), with basal-like tumors having the highest mean correlation. Shown are the PAM50-defined subtypes for TCGA tumors. c Global concordance between protein-mRNA pairs in TCGA tumors is associated with survival. Stratification of tumors into three groups [< 25%, 25–75%, and > 75% (tumors with highest global correlation)]. Tumors with the highest mean correlation between protein and mRNA pairs had the worst outcome [HR 75% vs. 25% tumor group = 2.6, 95% CI 1.01–6.65; Ptrend across the three groups = 0.043]. d Increased global concordance between protein-mRNA pairs is associated with reduced survival in patients with luminal A tumors. Ptrend = 0.0083CC BY · PubMed Central
Fig. 5 Global protein-mRNA correlations across breast tissues and their association with tumor characteristics. a Density plot for global Spearman correlation between protein and mRNA pairs across tumors (n = 59; 5677 protein-mRNA pairs) and adjacent non-cancerous tissue (n = 38; 3316 pairs). Mean global correlation coefficients are significantly different between tumor and adjacent non-cancerous tissue (P < 2.2 × 10−16, Wilcoxon rank sum test). b Concordance between protein and mRNA pairs is associated with discrete pathways in KEGG, e.g., protein processing and metabolism-related pathways. Protein-mRNA pairs for these pathways tend to show significantly increased concordances. Shown are the 10 highest ranked KEGG pathways that are enriched for high concordance protein-mRNA pairs. Multiple test-adjusted P values from the Kolmogorov-Smirnov test. c Protein-mRNA correlation for all proteins in tumors (“proteins,” 5677 protein-mRNA pairs), for tumor proteins with significant differences in abundance between tumor and adjacent non-cancerous tissue (“tumor signature,” n = 2258 pairs), and for basal-like tumor proteins with significant differences in abundance between tumor and adjacent non-cancerous tissue (“basal-like signature,” 159 pairs). Basal-like signature protein-mRNA pairs have the highest concordance. Definition of the signatures is described in the “Methods” section. d Concordance between protein and mRNA levels in breast tumors is associated with the predicted stability of proteins and mRNAs. Protein-mRNA pairs consisting of a protein and mRNA that are both stable have the highest mean concordance
Fig. 5 Global protein-mRNA correlations across breast tissues and their association with tumor characteristics. a Density plot for global Spearman correlation between protein and mRNA pairs across tumors (n = 59; 5677 protein-mRNA pairs) and adjacent non-cancerous tissue (n = 38; 3316 pairs). Mean global correlation coefficients are significantly different between tumor and adjacent non-cancerous tissue (P < 2.2 × 10−16, Wilcoxon rank sum test). b Concordance between protein and mRNA pairs is associated with discrete pathways in KEGG, e.g., protein processing and metabolism-related pathways. Protein-mRNA pairs for these pathways tend to show significantly increased concordances. Shown are the 10 highest ranked KEGG pathways that are enriched for high concordance protein-mRNA pairs. Multiple test-adjusted P values from the Kolmogorov-Smirnov test. c Protein-mRNA correlation for all proteins in tumors (“proteins,” 5677 protein-mRNA pairs), for tumor proteins with significant differences in abundance between tumor and adjacent non-cancerous tissue (“tumor signature,” n = 2258 pairs), and for basal-like tumor proteins with significant differences in abundance between tumor and adjacent non-cancerous tissue (“basal-like signature,” 159 pairs). Basal-like signature protein-mRNA pairs have the highest concordance. Definition of the signatures is described in the “Methods” section. d Concordance between protein and mRNA levels in breast tumors is associated with the predicted stability of proteins and mRNAs. Protein-mRNA pairs consisting of a protein and mRNA that are both stable have the highest mean concordanceCC BY · PubMed Central
Fig. 6 Protein abundance profiles separate breast tumors into two groups with greatly different patient survival. a Consensus matrix plot of NMF clustering for 59 breast tumors based on protein abundance levels in these tumors. Two groups of tumors emerged. Group 1 represents tumors enriched for the basal-like subtype (P < 0.01, Fisher’s exact test) while group 2 represents tumors enriched for the luminal A subtype (P = 0.03). Myc signaling was the strongest classifier for these two tumor groups (P < 1 × 10−8, Fisher’s exact test); most tumors in group 1 contained a Myc activation signature. For analysis, protein counts were normalized and log transformed, and consensus matrices were computed at K = 2–7. NMF class assignment for K = 2 was the most robust. The consensus index for each pair of samples is represented by a color gradient from white (0%) to red (100%) in the consensus matrix. b Enrichment pattern for transcription factor binding sites in genes that encode differently expressed proteins between tumor groups 1 and 2. Group 1—upregulated proteins are commonly encoded by genes with a predicted Myc binding motif, highlighted in red bars. Different bars represent different Myc binding motifs. GSEA enrichment score is captured by the blue dots. c Kaplan-Meier survival analysis comparing tumors in group 1 with tumors in group 2. Survival of patients in group 1 was significantly shorter than the survival of patients in group 2 (log-rank test, P = 0.027; HR = 2.65, 95% CI 1.08–6.51, using Cox regression)
Fig. 6 Protein abundance profiles separate breast tumors into two groups with greatly different patient survival. a Consensus matrix plot of NMF clustering for 59 breast tumors based on protein abundance levels in these tumors. Two groups of tumors emerged. Group 1 represents tumors enriched for the basal-like subtype (P < 0.01, Fisher’s exact test) while group 2 represents tumors enriched for the luminal A subtype (P = 0.03). Myc signaling was the strongest classifier for these two tumor groups (P < 1 × 10−8, Fisher’s exact test); most tumors in group 1 contained a Myc activation signature. For analysis, protein counts were normalized and log transformed, and consensus matrices were computed at K = 2–7. NMF class assignment for K = 2 was the most robust. The consensus index for each pair of samples is represented by a color gradient from white (0%) to red (100%) in the consensus matrix. b Enrichment pattern for transcription factor binding sites in genes that encode differently expressed proteins between tumor groups 1 and 2. Group 1—upregulated proteins are commonly encoded by genes with a predicted Myc binding motif, highlighted in red bars. Different bars represent different Myc binding motifs. GSEA enrichment score is captured by the blue dots. c Kaplan-Meier survival analysis comparing tumors in group 1 with tumors in group 2. Survival of patients in group 1 was significantly shorter than the survival of patients in group 2 (log-rank test, P = 0.027; HR = 2.65, 95% CI 1.08–6.51, using Cox regression)CC BY · 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

Cited by

Sorted by citations

See all 162

Metadata

Authors
Tang, W., Zhou, M., Dorsey, T. H., Prieto, D. A., Wang, X. W., Ruppin, E., Veenstra, T. D., & Ambs, S.
Affiliations
  • National Institutes of Health (US)
  • Frederick National Laboratory for Cancer Research (US)
  • National Cancer Institute (US)
Corresponding author
Ambs, Stefan
ORCID
Venue
Genome Medicine
Publisher
Springer Science and Business Media LLC
Published
2018-12-03
Volume
10
Issue
1
Pages
94-94
ISSN
1756-994X (ISSN-L)
DOI
10.1186/s13073-018-0602-x
PMC
PMC6276229
OpenAlex
W2902072367
Cited by
162
References
37
Citations by year
2026: 7 · 2025: 16 · 2024: 20 · 2023: 44 · 2022: 25 · 2021: 35 · 2020: 9 · 2019: 6
FWCI
3.78
Type
journal article
Language
en
Update notices
correction (2025-06-24)
Open access
Licence
Locations
  • Genome Medicine (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, other-oa)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DigitalCommons-Cedarville (Cedarville University) (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
National Cancer Institute (ZIA BC 010887)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Breast Cancer Treatment Studies
  • Physical Sciences › Chemistry › Spectroscopy › Advanced Proteomics Techniques and Applications
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
Keywords
breast cancer, proteotranscriptomics, breast cancer subtypes, basal-like breast cancer, triple-negative breast cancer, tumor cell proliferation, patient survival
MeSH terms
Breast Neoplasms, Chromatography, Liquid, Female, Humans, Mass Spectrometry, Signal Transduction, Gene Expression Regulation, Neoplastic, Oligonucleotide Array Sequence Analysis, Gene Expression Profiling, Proteomics

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