Recent molecular classifications of colorectal cancer (CRC) based on global gene expression profiles have defined subtypes displaying resistance to therapy and poor prognosis. Upon evaluation of these classification systems, we discovered that their predictive power arises from genes expressed by stromal cells rather than epithelial tumor cells. Bioinformatic and immunohistochemical analyses identify stromal markers that associate robustly with disease relapse across the various classifications. Functional studies indicate that cancer-associated fibroblasts (CAFs) increase the frequency of tumor-initiating cells, an effect that is dramatically enhanced by transforming growth factor (TGF)-β signaling. Likewise, we find that all poor-prognosis CRC subtypes share a gene program induced by TGF-β in tumor stromal cells. Using patient-derived tumor organoids and xenografts, we show that the use of TGF-β signaling inhibitors to block the cross-talk between cancer cells and the microenvironment halts disease progression.
This is a preview of subscription content, access via your institution
Open Access articles citing this article.
Pseudogenes and the associated ceRNA network as potential prognostic biomarkers for colorectal cancer
Scientific Reports Open Access 22 October 2022
Chemotherapy-induced complement signaling modulates immunosuppression and metastatic relapse in breast cancer
Nature Communications Open Access 02 October 2022
KIT promotes tumor stroma formation and counteracts tumor-suppressive TGFβ signaling in colorectal cancer
Cell Death & Disease Open Access 16 July 2022
Subscribe to Journal
Get full journal access for 1 year
only $6.58 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Tax calculation will be finalised during checkout.
Get time limited or full article access on ReadCube.
All prices are NET prices.
De Sousa E Melo, F. et al. Poor-prognosis colon cancer is defined by a molecularly distinct subtype and develops from serrated precursor lesions. Nat. Med. 19, 614–618 (2013).
Sadanandam, A. et al. A colorectal cancer classification system that associates cellular phenotype and responses to therapy. Nat. Med. 19, 619–625 (2013).
Marisa, L. et al. Gene expression classification of colon cancer into molecular subtypes: characterization, validation, and prognostic value. PLoS Med. 10, e1001453 (2013).
Sadanandam, A. et al. Reconciliation of classification systems defining molecular subtypes of colorectal cancer: interrelationships and clinical implications. Cell Cycle 13, 353–357 (2014).
Nishida, N. et al. Microarray analysis of colorectal cancer stromal tissue reveals upregulation of two oncogenic miRNA clusters. Clin. Cancer Res. 18, 3054–3070 (2012).
Calon, A. et al. Dependency of colorectal cancer on a TGF-β–driven program in stromal cells for metastasis initiation. Cancer Cell 22, 571–584 (2012).
Uhlen, M. et al. Towards a knowledge-based Human Protein Atlas. Nat. Biotechnol. 28, 1248–1250 (2010).
Mouradov, D. et al. Colorectal cancer cell lines are representative models of the main molecular subtypes of primary cancer. Cancer Res. 74, 3238–3247 (2014).
O'Brien, C.A., Pollett, A., Gallinger, S. & Dick, J.E. A human colon cancer cell capable of initiating tumour growth in immunodeficient mice. Nature 445, 106–110 (2007).
Ricci-Vitiani, L. et al. Identification and expansion of human colon-cancer-initiating cells. Nature 445, 111–115 (2007).
Oskarsson, T., Batlle, E. & Massague, J. Metastatic stem cells: sources, niches, and vital pathways. Cell Stem Cell 14, 306–321 (2014).
Thiery, J.P., Acloque, H., Huang, R.Y. & Nieto, M.A. Epithelial-mesenchymal transitions in development and disease. Cell 139, 871–890 (2009).
Markowitz, S.D. & Bertagnolli, M.M. Molecular origins of cancer: molecular basis of colorectal cancer. N. Engl. J. Med. 361, 2449–2460 (2009).
Grady, W.M. & Markowitz, S.D. Genetic and epigenetic alterations in colon cancer. Annu. Rev. Genomics Hum. Genet. 3, 101–128 (2002).
Markowitz, S. et al. Inactivation of the type II TGF-β receptor in colon cancer cells with microsatellite instability. Science 268, 1336–1338 (1995).
Jung, P. et al. Isolation and in vitro expansion of human colonic stem cells. Nat. Med. 17, 1225–1227 (2011).
Sato, T. et al. Long-term expansion of epithelial organoids from human colon, adenoma, adenocarcinoma, and Barrett's epithelium. Gastroenterology 141, 1762–1772 (2011).
Bueno, L. et al. Semi-mechanistic modelling of the tumour growth inhibitory effects of LY2157299, a new type I receptor TGF-β kinase antagonist, in mice. Eur. J. Cancer 44, 142–150 (2008).
Seoane, J., Le, H.V. & Massague, J. Myc suppression of the p21Cip1 Cdk inhibitor influences the outcome of the p53 response to DNA damage. Nature 419, 729–734 (2002).
Seoane, J. et al. TGFβ influences Myc, Miz-1 and Smad to control the CDK inhibitor p15INK4b. Nat. Cell Biol. 3, 400–408 (2001).
Giannelli, G., Villa, E. & Lahn, M. Transforming growth factor-β as a therapeutic target in hepatocellular carcinoma. Cancer Res. 74, 1890–1894 (2014).
Kosinski, C. et al. Gene expression patterns of human colon tops and basal crypts and BMP antagonists as intestinal stem cell niche factors. Proc. Natl. Acad. Sci. USA 104, 15418–15423 (2007).
Berdiel-Acer, M. et al. A 5-gene classifier from the carcinoma-associated fibroblast transcriptomic profile and clinical outcome in colorectal cancer. Oncotarget. 5, 6437–6452 (2014).
O'Shannessy, D.J. et al. Influence of tumor microenvironment on prognosis in colorectal cancer: tissue architecture–dependent signature of endosialin (TEM-1) and associated proteins. Oncotarget. 5, 3983–3995 (2014).
Francí, C. et al. Snail1 protein in the stroma as a new putative prognosis marker for colon tumours. PLoS ONE 4, e5595 (2009).
Ngan, C.Y. et al. Quantitative evaluation of vimentin expression in tumour stroma of colorectal cancer. Br. J. Cancer 96, 986–992 (2007).
Calon, A., Tauriello, D.V. & Batlle, E. TGF-β in CAF-mediated tumor growth and metastasis. Semin. Cancer Biol. 25, 15–22 (2014).
Barrett, T. & Edgar, R. Gene expression omnibus: microarray data storage, submission, retrieval, and analysis. Methods Enzymol. 411, 352–369 (2006).
Jorissen, R.N. et al. Metastasis-associated gene expression changes predict poor outcomes in patients with Dukes stage B and C colorectal cancer. Clin. Cancer Res. 15, 7642–7651 (2009).
De Sousa E Melo, F. et al. Methylation of cancer-stem-cell-associated Wnt target genes predicts poor prognosis in colorectal cancer patients. Cell Stem Cell 9, 476–485 (2011).
Benjamini, Y., Krieger, A.M. & Yekutieli, D. Adaptive linear step-up procedures that control the false discovery rate. Biometrika 93, 491–507 (2006).
Tibshirani, R., Hastie, T., Narasimhan, B. & Chu, G. Diagnosis of multiple cancer types by shrunken centroids of gene expression. Proc. Natl. Acad. Sci. USA 99, 6567–6572 (2002).
Merlos-Suárez, A. et al. The intestinal stem cell signature identifies colorectal cancer stem cells and predicts disease relapse. Cell Stem Cell 8, 511–524 (2011).
Lombardo, Y. et al. Bone morphogenetic protein 4 induces differentiation of colorectal cancer stem cells and increases their response to chemotherapy in mice. Gastroenterology 140, 297–309 (2011).
Morikawa, K., Walker, S.M., Jessup, J.M. & Fidler, I.J. In vivo selection of highly metastatic cells from surgical specimens of different primary human colon carcinomas implanted into nude mice. Cancer Res. 48, 1943–1948 (1988).
Morikawa, K. et al. Influence of organ environment on the growth, selection, and metastasis of human colon carcinoma cells in nude mice. Cancer Res. 48, 6863–6871 (1988).
Böttinger, E.P., Jakubczak, J.L., Haines, D.C., Bagnall, K. & Wakefield, L.M. Transgenic mice overexpressing a dominant-negative mutant type II transforming growth factor β receptor show enhanced tumorigenesis in the mammary gland and lung in response to the carcinogen 7,12-dimethylbenz-[a]-anthracene. Cancer Res. 57, 5564–5570 (1997).
Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA 102, 15545–15550 (2005).
Witten, I.H. et al. Weka: practical machine learning tools and techniques with Java implementations. Proceedings of the ICONIP/ANZIIS/ANNES'99 Workshop on Emerging Knowledge Engineering and Connectionist-Based Information Systems 192–196 (1999).
Warren, R.S., Yuan, H., Matli, M.R., Gillett, N.A. & Ferrara, N. Regulation by vascular endothelial growth factor of human colon cancer tumorigenesis in a mouse model of experimental liver metastasis. J. Clin. Invest. 95, 1789–1797 (1995).
Hu, Y. & Smyth, G.K. ELDA: extreme limiting dilution analysis for comparing depleted and enriched populations in stem cell and other assays. J. Immunol. Methods 347, 70–78 (2009).
Lander, E.S. et al. Initial sequencing and analysis of the human genome. Nature 409, 860–921 (2001).
Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).
O'Brien, C.A. et al. ID1 and ID3 regulate the self-renewal capacity of human colon cancer–initiating cells through p21. Cancer Cell 21, 777–792 (2012).
McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 20, 1297–1303 (2010).
DePristo, M.A. et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat. Genet. 43, 491–498 (2011).
Pathmanathan, S., Hamilton, E., Atcheson, E. & Timson, D.J. The interaction of IQGAPs with calmodulin-like proteins. Biochem. Soc. Trans. 39, 694–699 (2011).
Clevers, H. & Batlle, E. EphB/EphrinB receptors and Wnt signaling in colorectal cancer. Cancer Res. 66, 2–5 (2006).
Cingolani, P. et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly (Austin) 6, 80–92 (2012).
R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2014).
Kent, W.J. et al. The human genome browser at UCSC. Genome Res. 12, 996–1006 (2002).
Lawrence, M., Gentleman, R. & Carey, V. rtracklayer: an R package for interfacing with genome browsers. Bioinformatics 25, 1841–1842 (2009).
Wood, S.N. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J. Royal Statist. Soc. (B) 73, 3–36 (2011).
Venkatraman, E.S. & Olshen, A.B. A faster circular binary segmentation algorithm for the analysis of array CGH data. Bioinformatics 23, 657–663 (2007).
We thank G. Stassi (University of Palermo) for providing PDO7 and PDO8, L. Wakefield (US National Cancer Institute) for providing the plasmid encoding DNR, I. Joval for assistance in mounting the figures, M. Virtudes Cespedes and R. Mangues (IIB Sant Pau) for logistic support with CRC samples, and all members of the Batlle laboratory for support and discussions We are grateful for the excellent assistance of the IRB Barcelona core facilities for Histology, Functional Genomics and Advanced Digital Micropscopy. D.V.F.T. holds a Juan de la Cierva postdoctoral fellowship, from the Spanish Ministry of Economy and Competitiveness, and E.L. holds a fellowship from Fundación Olga Torres and Asociación Española contra el Cáncer (AECC). This work has been supported by grants from the Doctor Josef Steiner Foundation, AECC, Red Temática de Investigación Cooperativa en Cáncer, Instituto de Salud Carlos III (RTICC:RD12/0036/0024) and grant SAF2011-27068, the latter two from the Spanish Ministry of Economy and Competitiveness, and by ‘Xarxa de Bancs de Tumors’ sponsored by Pla Director d'Oncologia de Catalunya (XBTC).
The authors declare no competing financial interests.
Integrated supplementary information
Supplementary Figure 1 High levels of genes characteristic of CAFs identify poor-prognosis patients across multiple subtypes in the De Sousa E Melo et al. molecular classification.
(a) Clustering analysis of the 146-probeset signature used to classify CRC patients into subtypes by De Sousa E Melo et al. in the GSE14333 cohort. We allowed unsupervised hierarchical clustering of the probesets using correlation distance and the Ward agglomeration method, whereas we enforced the classification of patients into subtypes. Data show normalized, centered and scaled Affymetrix probeset intensities on a log2 scale. The HR lane represents the hazard ratio for the corresponding genes on a log2 scale. Genes significantly upregulated in the microdissected epithelial or stromal compartment are depicted in the epithelial (LCM) and stroma (LCM) lanes, respectively. Genes specifically upregulated in epithelial, endothelial, leukocyte or FAP cell populations are represented in the EpCAM, CD31, CD45 and FAP lanes, respectively. Patients with high average expression (z scores >0) of the CAF-enriched gene cluster are marked in green (GP-HiC). (b) The smooth estimate of HR (+1 s.d.) (increase in recurrence risk for one standard deviation) shows higher risk of relapse for patients within the good-prognosis subtypes (CCS1 + CCS2) presenting higher average expression of the CAF cluster (GP-HiC group). Red dashed lines indicate 95% confidence bands. HR values for the linear continuous model and corresponding P values are indicated. (c) The Kaplan-Meier curve shows the recurrence-free survival of patients within good-prognosis cancer subtypes and presenting low expression levels of the CAF cluster gene set (blue; GP-LoC) or patients within good-prognosis cancer subtypes that present high expression levels of the CAF cluster gene set (yellow; GP-HiC), both compared to the stem cell–like poor-prognosis subtype (red; CCS3). HR and P values are indicated.
Supplementary Figure 2 High levels of genes characteristic of CAFs in C4 poor-prognosis patients in the Marisa et al. molecular classification.
(a) Clustering analysis of the 1,459-probeset signature used to classify CRC patients into subtypes by Marisa et al. in the GSE39582 cohort. We allowed unsupervised hierarchical clustering of the probesets using correlation distance and the Ward agglomeration method, whereas we enforced the classification of patients into subtypes. Data show normalized, centered and scaled Affymetrix probeset intensities on a log2 scale. The HR lane represents the hazard ratio for the corresponding genes on a log2 scale. Genes significantly upregulated in the microdissected epithelial or stromal compartment are depicted in the epithelial (LCM) and stroma (LCM) lanes, respectively. Genes specifically upregulated in epithelial, endothelial, leukocyte or FAP cell populations are represented in the EpCAM, CD31, CD45 and FAP lanes, respectively. Patients with high average expression (z scores >0) of the CAF-enriched gene cluster are marked in green (GP-HiC). (b) The smooth estimate of HR (+1 s.d.) (increase in recurrence risk for one standard deviation) shows higher risk of relapse for patients within the good-prognosis subtypes (C1 + C2 + C3 + C5 + C6) presenting higher average expression of the CAF cluster (GP-HiC group). Red dashed lines indicate 95% confidence bands. HR values for the linear continuous model and corresponding P values are indicated. (c) The Kaplan-Meier curve shows the recurrence-free survival of patients within good-prognosis cancer subtypes and presenting low expression levels of the CAF cluster gene set (blue; GP-LoC) or patients within good-prognosis cancer subtypes that present high expression levels of the CAF cluster gene set (yellow; GP-HiC), both compared to the C4 poor-prognosis subtype (red). HR and P values are indicated.
(a) The pie chart shows the expression pattern distribution (percentage) of clinically relevant antibodies in individuals with CRC within the Human Protein Atlas database. Stromal-specific antibodies account for approximately 31% of this distribution, including the newly identified CALD1 and POSTN proteins (red). Antibodies to FAP and IGFBP7 belong to the 62% of antibodies that stain both the stromal and epithelial compartments (yellow). We found that 3% of antibodies are epithelial specific (blue), whereas the rest were not detectable in either fraction (4%; gray slice). (b) z-score mean of CALD1, FAP, POSTN and IGFBP7 in the cell populations from disaggregated primary CRC samples. ***P < 0.001.
Supplementary Figure 4 Poor-prognosis CRC subtypes are characterized by high levels of TGF-β and TBRS expression.
(a) z-score means of TGFB1, TGFB2 and TGFB3 mRNA expression levels in the three GEO data sets in each molecular subtype. The poor-prognosis group (red box) for each data set contains the highest levels of TGFB1 and TGFB3. (b,c) z-score means of TBRSs (TGF-β–responsive signatures) in the three data sets in each molecular subtype. The TBRSs are extracted from (b) macrophages (Ma-TBRS) or (c) lymphocytes (T-TBRS) treated with recombinant TGF-β1. The poor-prognosis groups for each data set are depicted in red. P values for pairwise comparisons are in Supplementary Table 7.
Colony formation assay (a) and qRT-PCR analysis of cell cycle genes (b) in HT29-M6 and KM12L4a cells treated for 7 d with recombinant TGF-β1 (red bar) compared to untreated (blue bar) CRC cells. qRT-PCR data are normalized to PP1A levels and are presented as fold change versus the control (Con) (n ≥ 3 independent experiments). (c) The Kaplan-Meier curves show the disease-free survival (DFS) of mice after subcutaneous injections of 6 × 103 KM12L4aTGFβ cells alone (n = 8 tumor cell inoculations; black line) or coinoculated with FIB-shCon (n = 8 tumor cell inoculations; red line) or FIB-shTBRI (n = 8 tumor cell inoculations; blue line). FIB, fibroblasts (5 × 104 cells). *P < 0.05. (d) Growth kinetics of HT29-M6 (blue; n = 4 tumor cell inoculations) and HT29-M6TGFβ (red; n = 4 tumor cell inoculations) cells inoculated subcutaneously alone or in combination with FIB (green and yellow lines, respectively). Day 1, day of first detection. Values are means ± s.e.m.
Supplementary Figure 6 Coinjection of CRC cells with wild-type or TGF-β pathway–defective fibroblasts did not modify xenograft growth rates.
(a) Growth kinetics of HT29-M6TGFβ cells coinoculated subcutaneously with FIB-shCon (n = 8 tumor cell inoculations; blue) or FIB-shTBRI (n = 5 tumor cell inoculations; red). (b) Growth kinetics of HT29-M6TGFβ cells coinoculated subcutaneously with FIB-Con (n = 7 tumor cell inoculations; blue) or FIB-DNR (n = 5 tumor cell inoculations; red). Day 1, day of first detection. Values are means ± s.e.m. (c,d) Histology of xenografts with HT29-M6 and HT29-M6TGFβ cells inoculated subcutaneously alone (Con) or in combination with FIB, FIB-shTBRI or FIB-DNR. (Scale bars, 200 μm). (c) We use trichrome Masson staining to mark epithelial cells in red and collagen fibers in blue. (d) EpCAM immunostaining marks epithelial tumor cells. The histological patterns were equivalent in all conditions. (e) We measured the surfaces occupied by stromal and epithelial cells using trichrome Masson staining (Online Methods) in multiple tumors formed after subcutaneous inoculation of HT29-M6TGFβ cells alone or in combination with FIB or FIB-DNR cells. This analysis confirmed no significant differences between the three experimental conditions.
Supplementary Figure 7 Fibroblasts lacking a TGF-β response showed minor capacity to support tumor formation.
(a) Kaplan-Meier curves showing the disease-free survival (DFS) of mice after subcutaneous inoculation of HT29-M6 cells alone (black line; n = 6 tumor cell inoculations) or coinjection with FIB-shCon (red line; n = 6 tumor cell inoculations) or FIB-shTBRI (blue line; n = 6 tumo cells inoculations) cells. FIB, fibroblasts (5 × 104 cells). (b) Growth kinetics of FIB cultured in the presence (red line) or absence (blue line) of recombinant TGF-β1. Cells were treated with TGF-β1 from day 1. Growth was followed for 6 d, and the number of live cells was monitored at the indicated time points. Values are means ± s.d. **P < 0.01. (c) The table shows raw data for the subcutaneous inoculation experiments in the indicated conditions. The columns tumors/n indicate the number of tumor initiation events per number of inoculations. The percentage (%) of successful tumor initiation events and DFS (days ± s.d.) are indicated for each condition. (d) Estimated frequency of tumor-initiating cells (TICs) calculated using ELDA (Online Methods) from the above data (***P < 0.001 for comparisons against TGF-β + FIB); 95% confidence intervals are indicated.
Supplementary Figure 8 Heterogeneity of patient-derived tumor organoids (PDO) growth under TGF-β treatment.
Growth kinetics of tumor organoids cultured in the presence or absence of recombinant TGF-β1 (red line) or LY2157299 (yellow line). The blue line represents untreated cells. Growth was followed for 7 days (D), and tumor organoids were counted at the indicated time points. Values are means ± s.d.
(a) Kaplan-Meier curves show disease-free survival (DFS) for mice injected subcutaneously with cells from PDO1, 2, 3, 4, 5, 6, 7 or 8, treated (red line; n = 12 tumor cell inoculations) or untreated (blue line; n = 12 tumor cell inoculations) with LY2157299. *P < 0.05; ***P < 0.001. (b) Tumor volume over time for the indicated tumor organoids (from a), measured from the day of appearance, treated (red) or untreated (blue) with LY2157299. Values are means ± s.e.m. (c) Kaplan-Meier curves and growth curves for mice injected subcutaneously with cells from PDO5 induced (ind.) with TGF-β1, treated (yellow) or untreated (blue) with doxycycline (4 × 105 cells; n = 18 tumor cell inoculations). Values are means ± s.e.m.
Supplementary Figures 1–9 and Supplementary Note. (PDF 1934 kb)
Genes positively associated to relapse (HR > 1, P < 0.01) and upregulated in poor-prognosis subtypes in each cohort (probe level data). (XLS 1455 kb)
HR and stromal enrichment of poor-prognosis gene sets across multiple thresholds of gene expression and cutoffs of significance. (XLS 39 kb)
Reclassification of CRC patients according to the EpCAM+ cluster or excluding the stromal cluster. (XLS 45 kb)
Expression patterns of genes commonly associated with poor prognosis in different patient cohorts according to the Human Protein Atlas database. (XLS 41 kb)
Cox proportional hazards multivariate analysis of stromal protein intensities, adjusted by AJCC stage, sex, age, tumor site and treatment. (XLS 30 kb)
Statistical analysis of IHC data (adjusted for TMA, stage, sex, age, tumor site and treatment). (XLS 21 kb)
Statistics for group-to-group comparisons for TGF-β and FTBRS average levels in different CRC patient data sets. (XLS 35 kb)
TGF-β response signatures (TBRSs) in normal stromal cells. (XLS 278 kb)
TBRS predicts disease relapse in the three CRC patient cohorts independently of the main clinical parameters. (XLS 28 kb)
Clinicopathological features and summary of genetic alterations in major pathways according to exome sequencing in tumor organoids. (XLS 29 kb)
About this article
Cite this article
Calon, A., Lonardo, E., Berenguer-Llergo, A. et al. Stromal gene expression defines poor-prognosis subtypes in colorectal cancer. Nat Genet 47, 320–329 (2015). https://doi.org/10.1038/ng.3225
This article is cited by
Serum-based measurements of stromal activation through ADAM12 associate with poor prognosis in colorectal cancer
BMC Cancer (2022)
Cancer Cell International (2022)
Cancer-associated fibroblasts and resistance to anticancer therapies: status, mechanisms, and countermeasures
Cancer Cell International (2022)
Pseudogenes and the associated ceRNA network as potential prognostic biomarkers for colorectal cancer
Scientific Reports (2022)
Nature Cancer (2022)