In cancer, the epithelial-to-mesenchymal transition (EMT) is associated with tumour stemness, metastasis and resistance to therapy. It has recently been proposed that, rather than being a binary process, EMT occurs through distinct intermediate states. However, there is no direct in vivo evidence for this idea. Here we screen a large panel of cell surface markers in skin and mammary primary tumours, and identify the existence of multiple tumour subpopulations associated with different EMT stages: from epithelial to completely mesenchymal states, passing through intermediate hybrid states. Although all EMT subpopulations presented similar tumour-propagating cell capacity, they displayed differences in cellular plasticity, invasiveness and metastatic potential. Their transcriptional and epigenetic landscapes identify the underlying gene regulatory networks, transcription factors and signalling pathways that control these different EMT transition states. Finally, these tumour subpopulations are localized in different niches that differentially regulate EMT transition states.
This is a preview of subscription content, access via your institution
Open Access articles citing this article.
Military Medical Research Open Access 02 December 2022
Journal of Hematology & Oncology Open Access 08 September 2022
Cell Regeneration Open Access 01 September 2022
Subscribe to Nature+
Get immediate online access to Nature and 55 other Nature journal
Subscribe to Journal
Get full journal access for 1 year
only $3.90 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.
Nieto, M. A., Huang, R. Y., Jackson, R. A. & Thiery, J. P. EMT: 2016. Cell 166, 21–45 (2016).
Puisieux, A. & Brabletz, T. & Caramel, J. Oncogenic roles of EMT-inducing transcription factors. Nat. Cell Biol. 16, 488–494 (2014).
Huang, R. Y. et al. An EMT spectrum defines an anoikis-resistant and spheroidogenic intermediate mesenchymal state that is sensitive to e-cadherin restoration by a src-kinase inhibitor, saracatinib (AZD0530). Cell Death Dis. 4, e915 (2013).
Zhang, J. et al. TGF-β-induced epithelial-to-mesenchymal transition proceeds through stepwise activation of multiple feedback loops. Sci. Signal. 7, ra91 (2014).
Hong, T. et al. An Ovol2-Zeb1 mutual inhibitory circuit governs bidirectional and multi-step transition between epithelial and mesenchymal states. PLoS Comput. Biol. 11, e1004569 (2015).
Jolly, M. K. et al. Stability of the hybrid epithelial/mesenchymal phenotype. Oncotarget 7, 27067–27084 (2016).
Bierie, B. et al. Integrin-β4 identifies cancer stem cell-enriched populations of partially mesenchymal carcinoma cells. Proc. Natl Acad. Sci. USA 114, E2337–E2346 (2017).
Latil, M. et al. Cell-type-specific chromatin states differentially prime squamous cell carcinoma tumor-initiating cells for epithelial to mesenchymal transition. Cell Stem Cell 20, 191–204 (2017).
Chen, Q., Zhang, X. H. & Massagué, J. Macrophage binding to receptor VCAM-1 transmits survival signals in breast cancer cells that invade the lungs. Cancer Cell 20, 538–549 (2011).
Seguin, L. et al. An integrin β3–KRAS–RalB complex drives tumour stemness and resistance to EGFR inhibition. Nat. Cell Biol. 16, 457–468 (2014).
Slack-Davis, J. K., Atkins, K. A., Harrer, C., Hershey, E. D. & Conaway, M. Vascular cell adhesion molecule-1 is a regulator of ovarian cancer peritoneal metastasis. Cancer Res. 69, 1469–1476 (2009).
Wang, J. et al. CD51 correlates with the TGF-beta pathway and is a functional marker for colorectal cancer stem cells. Oncogene 36, 1351–1363 (2017).
Van Keymeulen, A. et al. Reactivation of multipotency by oncogenic PIK3CA induces breast tumour heterogeneity. Nature 525, 119–123 (2015).
Ye, X. et al. Upholding a role for EMT in breast cancer metastasis. Nature 547, E1–E3 (2017).
Ye, X. et al. Distinct EMT programs control normal mammary stem cells and tumour-initiating cells. Nature 525, 256–260 (2015).
DelPozo Martin, Y. et al. Mesenchymal cancer cell-stroma crosstalk promotes niche activation, epithelial reversion, and metastatic colonization. Cell Reports 13, 2456–2469 (2015).
Hidalgo, M. et al. Patient-derived xenograft models: an emerging platform for translational cancer research. Cancer Discov. 4, 998–1013 (2014).
Puram, S. V. et al. Single-cell transcriptomic analysis of primary and metastatic tumor ecosystems in head and neck cancer. Cell 171, 1611–1624 (2017).
Mani, S. A. et al. The epithelial–mesenchymal transition generates cells with properties of stem cells. Cell 133, 704–715 (2008).
Ye, X. & Weinberg, R. A. Epithelial-mesenchymal plasticity: a central regulator of cancer progression. Trends Cell Biol. 25, 675–686 (2015).
Meacham, C. E. & Morrison, S. J. Tumour heterogeneity and cancer cell plasticity. Nature 501, 328–337 (2013).
Diepenbruck, M. & Christofori, G. Epithelial–mesenchymal transition (EMT) and metastasis: yes, no, maybe? Curr. Opin. Cell Biol. 43, 7–13 (2016).
Zheng, X. et al. Epithelial-to-mesenchymal transition is dispensable for metastasis but induces chemoresistance in pancreatic cancer. Nature 527, 525–530 (2015).
Fischer, K. R. et al. Epithelial-to-mesenchymal transition is not required for lung metastasis but contributes to chemoresistance. Nature 527, 472–476 (2015).
Aiello, N. M. et al. Upholding a role for EMT in pancreatic cancer metastasis. Nature 547, E7–E8 (2017).
Tsai, J. H., Donaher, J. L., Murphy, D. A., Chau, S. & Yang, J. Spatiotemporal regulation of epithelial–mesenchymal transition is essential for squamous cell carcinoma metastasis. Cancer Cell 22, 725–736 (2012).
Ocaña, O. H. et al. Metastatic colonization requires the repression of the epithelial–mesenchymal transition inducer Prrx1. Cancer Cell 22, 709–724 (2012).
Siegel, P. M. & Massagué, J. Cytostatic and apoptotic actions of TGF-β in homeostasis and cancer. Nat. Rev. Cancer 3, 807–821 (2003).
Srinivas, S. et al. Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev. Biol. 1, 4 (2001).
Barker, N. et al. Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature 449, 1003–1007 (2007).
Tuveson, D. A. et al. Endogenous oncogenic K-ras(G12D) stimulates proliferation and widespread neoplastic and developmental defects. Cancer Cell 5, 375–387 (2004).
Jonkers, J. et al. Synergistic tumor suppressor activity of BRCA2 and p53 in a conditional mouse model for breast cancer. Nat. Genet. 29, 418–425 (2001).
Van Keymeulen, A. et al. Distinct stem cells contribute to mammary gland development and maintenance. Nature 479, 189–193 (2011).
Lapouge, G. et al. Skin squamous cell carcinoma propagating cells increase with tumour progression and invasiveness. EMBO J. 31, 4563–4575 (2012).
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Kiselev, V. Y. et al. SC3: consensus clustering of single-cell RNA-seq data. Nat. Methods 14, 483–486 (2017).
Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386 (2014).
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).
Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008).
Heinz, S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38, 576–589 (2010).
McLean, C. Y. et al. GREAT improves functional interpretation of cis-regulatory regions. Nat. Biotechnol. 28, 495–501 (2010).
Anders, S., Pyl, P. T. & Huber, W. HTSeq—a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169 (2015).
Hu, Y. & Smyth, G. K. ELDA: extreme limiting dilution analysis for comparing depleted and enriched populations in stem cells and other assays. J. Immunol. Methods 15, (70–78 (2009).
Medina-Rivera, A. et al. RSAT 2015: regulatory sequence analysis tools. Nucleic Acids Res. 43, W50–W56 (2015).
We thank the ULB animal facility, the ULB genomic core facility (F. Libert and A. Lefort), G. Berx (Ghent University) and M. Mazzone (VIB-KUL) for MMTV-PyMT mice. I.P. is supported by TELEVIE. P.A.S. is supported by FNRS. A.S., D.B. and T.V. are supported by KULeuven SymBioSys, Stiching Tegen Kanker, FWO postdoctoral fellowship #12W7318N and Marie Sklodowska-Curie fellowship #12O5617N. The Department of Pathology acknowledges Fonds Yvonne Boel. The PDX platform is supported by Fonds Erasme. C.Bla. is supported by WELBIO, FNRS, Fondation Contre le Cancer, ULB Foundation, European Research Council, Worldwide Cancer Research and the Foundation Baillet Latour.
Nature thanks B. E. Bernstein, E. Thompson and the other anonymous reviewer(s) for their contribution to the peer review of this work.
The authors declare no competing interests.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
Extended Data Fig. 1 Screening of cell surface markers homogeneously and heterogeneously expressed in Epcam+ and Epcam− TCs in skin SCCs.
a, b, FACS profile of the isotype control and homogeneously or heterogeneously expressed markers in Epcam+ (a) and Epcam− (b) TCs using BD lyoplate panel. c, Screening of all markers found to be heterogeneously expressed in at least one tumour in the initial BD lyoplate screen on a much larger cohort of tumours. Epcam− TCs were more heterogeneous than Epcam+ TCs and CD51, CD61 and CD106 were heterogeneously expressed in at least 75% of the tumours analysed and therefore were selected and used in combination for further experiments.
Extended Data Fig. 2 Gating strategy for FACS analysis and cell sorting of the different tumour subpopulations.
a, FACS plots showing the gating strategy used to FACS-isolate Epcam+ and Epcam− cells from YFP+ tumours (Lgr5CreER/KrasG12D/p53KO and K8CreER/Pik3ca/p53KO). b, FACS plots showing the six different subpopulations of Epcam− TCs: Epcam−CD106−CD51−CD61− (triple-negative), Epcam−CD106+CD51−CD61−, Epcam−CD106−CD51+CD61−, Epcam−CD106+CD51+CD61−, Epcam−CD106−CD51+CD61+ and Epcam−CD106+CD51+CD61+ (triple-positive) populations. c–h, Post-sort purity plots for triple-negative (c), CD106+ (d), CD51+ (e), CD106+CD51+ (f), CD51+CD61+ (g) and triple-positive (h) subpopulations (n = 1). i, FACS plots showing the gating strategy used to FACS-isolate Epcam+ and Epcam− cells from MMTV-PyMT mammary tumours.
a–e, Dimensionality reduction of scRNA-seq data using PCA colouring for pure epithelial genes Esrp1 and Ovol1 (a), hybrid epithelial genes Krt5 and Trp63 (b), hybrid mesenchymal genes Prrx1, Snai1, Zeb1, Zeb2, Pdgfra and Col3a1 (c, d) and pure mesenchymal genes Aspn and Mmp19 (e). Every dot represents a single cell and its shape represents the FACS sorting criteria (Epcam+ or Epcam−). The colour scale represents the normalized expression of the respective genes. Pure epithelial genes are expressed only in Epcam+ TCs, hybrid epithelial genes in both Epcam+ and hybrid Epcam− TCs, hybrid mesenchymal genes in hybrid Epcam− TCs and highly mesenchymal Epcam− TCs, and pure mesenchymal genes only in highly mesenchymal Epcam− TCs. a–e, n = 66 Epcam+ and n = 277 Epcam− cells from one tumour. f, Immunostaining showing the co-expression of Epcam and E-cadherin in the majority of epithelial TCs. Rare Epcam + cells are E-cadherin− (arrowheads). Scale bars, 50 μm.
a, Percentage of Epcam+ cells in the two groups of tumours that differ by the frequency of the different CD106, CD51, CD61 subpopulations in metaplastic-like breast tumours resulting from activation of the oncogenic mutation Pik3caH1047R and p53 deletion in luminal cells of the mammary gland (n = 11 biological replicates, mean ± s.e.m.). b, FACS profile showing the six different CD106, CD51, CD61 subpopulations in the most frequent group 1 mixed tumours containing Epcam+ and Epcam− TCs. c, d, Distribution of the different tumour subpopulations in group 1 (c; mixed tumours with Epcam+ and Epcam− TCs) and in group 2 (d; mesenchymal Epcam− TCs) (n = 10 biological replicates, mean ± s.e.m.). e, Immunostaining of K14 and vimentin in cystospin of FACS-isolated YFP+Epcam+, YFP+Epcam− triple-negative and YFP+Epcam− triple-positive populations. Scale bars, 20 μm; n = 3 biological replicates. f, Proportions of cells expressing K14 and vimentin based on cytospin counts; n = 3 biological replicates. g, Number of lung metastases arising from the injection of 2,000 YFP+ TCs from different EMT subpopulations. Dots represent the number of metastases per condition and per mouse (n = 2). h, qRT–PCR showing the downregulation of epithelial markers and the upregulation of EMT markers in the different Epcam− tumour subpopulations. Histogram represents mean ( + s.e.m.) fold-change in gene expression compared to Epcam+ tumour cells for mesenchymal genes and compared to triple-positive tumour cells for epithelial genes (n = 7 Epcam+, n = 3 triple-negative, n = 5 CD106+, n = 3 CD51+, n = 6 CD106+CD51+, n = 3 CD51+CD61+, n = 6 triple-positive, all are biological replicates). i–l, Immunostaining showing the expression of K14 and vimentin, YFP and K14, YFP and vimentin, or YFP and CD61 in area 1 (K14+Vim− CD61−; i), area 2 (K14+Vim+CD61−; j), area 3 (K14+Vim+CD61+; k) and area 4 (K14−Vim+CD61+; l). Scale bars, 50 μm. m–p, Immunostaining showing the expression of YFP and CD45, CD68, CD31 or Lyve1 in area 1 (m), area 2 (n), area 3 (o) and area 4 (p), showing that immune infiltration and the density of vascular and lymphatic vessels are associated with late stages of EMT. Scale bars, 50 μm, n = 3 biological replicates. In hair follicle EMT-derived tumours, the chromatin and transcriptional landscape of the cancer cell of origin prime the tumour to undergo EMT during tumorigenesis; in Pik3ca/p53KO mammary metaplastic tumours this is not always the case. Metaplastic EMT tumours are more frequent in luminal-derived tumours than in basal/myoepithelial cells, which under physiological conditions present much more EMT-like features do than luminal cells.
a, Percentage of Epcam+ cells in two groups of tumours that differ by the frequency of the different CD106, CD51, CD61 subpopulations in MMTV-PyMT luminal-like mammary tumours (n = 5 mice, mean ± s.e.m.). b, Distribution of the different tumour subpopulations in group 2 (mixed tumours with Epcam+ and Epcam− TCs) (n = 9 tumours, n = 5 mice, mean ± s.e.m.). c, FACS profile showing the six different CD106, CD51, CD61 subpopulations in the group containing mixed tumours with Epcam+ and Epcam− TCs. d, Immunostaining of K8 and vimentin in cystospin of FACS-isolated YFP+Epcam+, YFP+Epcam− triple-negative and YFP+Epcam− triple-positive populations. Scale bars, 20 μm, n = 2 biological replicates. e, Proportion of cells expressing K8 and vimentin based on cytospin counts (n = 2 tumours). f, Number of lung metastases arising from the injection of 10,000 TCs from the different EMT subpopulations (Epcam+, n = 5; triple-negative, n = 5; CD106+, n = 5; triple-positive, n = 4). All are biological replicates, mean + s.e.m. Two-sided Mann–Whitney test: Epcam+ versus triple-negative, P = 0.0079; Epcam+ versus CD106+, P = 0.0079; triple-negative versus triple-positive, P = 0.0238; CD106+ versus triple-positive, P = 0.0476; mean + s.e.m. g, qRT–PCR showing the downregulation of epithelial markers and the upregulation of EMT markers in the different Epcam− tumour subpopulations. Histogram represents mean ± s.e.m. fold-change in gene expression compared to Epcam+ tumour cells for mesenchymal genes and compared to triple-positive tumour cells for epithelial genes (Epcam+, n = 5; triple-negative, n = 3; CD106+, n = 5; triple-positive, n = 4; all are biological replicates). h–j, Immunostaining of pancytokeratin (PanKRT) and vimentin, PyMT and pancytokeratin or PyMT and vimentin in area 1 (pancytokeratin+Vim−; h), area 2 (pancytokeratin+Vim+; i) and area 3 (pancytokeratin−Vim+; j). Scale bars, 50 μm. k–m, Immunostaining of PyMT and CD68 or Lyve1 in area 1 (k), area 2 (l) and area 3 (m), showing that immune infiltration and the density of lymphatic vessels are associated with late-stage EMT. Scale bars, 50 μm, n = 2 biological replicates.
a–d, Immunostaining of pancytokeratin and vimentin, human-specific marker Ku80 and pancytokeratin or Ku80 and vimentin in poorly differentiated human breast (a), lung (b) or oesophagus (c, d) PDX. In these examples, three different areas can be identified: area 1 (Ku80+pancytokeratin+vimentin−), area 2 (Ku80+pancytokeratin+vimentin+) and area 3 (Ku80+pancytokeratin−vimentin+), showing that TCs with different degrees of EMT can be identified in a subset of poorly differentiated human cancers. Scale bars, 50 μm. e, PDX derived from aggressive and highly metastatic pleomorphic lung carcinoma, showing the hybrid EMT state as defined by co-expression of pancytokeratin and vimentin by the majority of TCs.
Extended Data Fig. 7 Transcription factor binding motif enrichment in opened chromatin regions in the different EMT populations.
a, Heat map representing unsupervised clustering of RNA-seq data from the different tumour subpopulations, showing that Epcam+ epithelial TCs, Epcam− hybrid TCs and Epcam− mesenchymal TCs fall into different clusters (n = 3 biological replicates). b, Representation of chromatin remodelling and associated transcription factors (TFs) during EMT based on the transcription factor motifs that were statistically enriched in ATAC-seq peaks that differed between tumour subpopulations as determined by Homer analysis. c–f, Transcription factor motifs enriched in the ATAC-seq peaks that were upregulated between the indicated subpopulations as determined by Homer analysis using known or de novo motif search (asterisk) or using JASPAR motif matrix (hash symbol). g–j, Transcription factor motifs enriched in the ATAC-seq peaks that were upregulated between the indicated tumour subpopulations as determined by Homer analysis using known or de novo motif search (asterisk). Green, core transcription factors; pink, epithelial transcription factors; yellow, mesenchymal transcription factors. c–j, Analysis based on ATAC-seq results from different subpopulations derived from one tumour.
a, Heat map of normalized scRNA-seq expression of markers. Markers are obtained from the scRNA-seq data using SC3 with clustering parameter k = 5 and selecting genes with an area under the curve (AUC) > 0.8 and P < 0.01. Columns represent cells and rows represent genes. Colouring represents the normalized expression in the scRNA-seq data for each cell. b, Dimensionality reduction of scRNA-seq data using PCA colouring for clusters identified by SC3 clustering. Dots represent single cells; shape represents FACS sorting criteria. c, Dimensionality reduction of scRNA-seq data using PCA colouring for clusters identified by Monocle using only the top 500 genes with the highest covariability with CD51, CD61 and CD106 expression. Dots represent single cells; colour represents Monocle cluster; shape represents FACS sorting criteria. d, Pseudotime-ordering trajectory of scRNA-seq data using Monocle. The top 500 differentially expressed genes across clusters identified in a semi-supervised approach using Monocle were used to order cells in pseudotime using the DDRTree method. Dots represent single cells; colours represent assigned Monocle clusters. e, Heat map representing normalized scRNA-seq data after branched expression analysis modelling (BEAM) using Monocle for the top 100 differentially expressed genes across the branch-point. The pre-branch here is set to be the state corresponding to the Epcam+ FACS-sorted cells. Cell fates 1 and 2 represent the two branches of the pseudotime trajectory corresponding to more hybrid and mesenchymal EMT phases, respectively. Heat map colours represent gene-wise normalized expression across pseudotime. Genes are hierarchically clustered according to their expression pattern. a–e, n = 66 Epcam+ and n = 277 Epcam− cells from one tumour.
Extended Data Fig. 9 Effect of constitutive overexpression of p63 and inhibition of TGF-β signalling in vivo on the different EMT subpopulations.
a, Immunostaining of K14 and vimentin, YFP and K14 or YFP and vimentin in hair follicle-derived skin SCCs with constitutive overexpression of p63, showing that EMT is blocked at an early stage and only area 1 epithelial TCs (K14+vimentin−) and area 2 early hybrid TCs (K14+vimentin+) can be detected. Scale bars, 50 μm, n = 5 biological replicates. b, Experimental strategy to inhibit TGF-β signalling during hair follicle-derived SCC tumorigenesis. c, Immunostaining of K14 and vimentin or YFP K14, vimentin, CD61, CD45, CD68, CD31 or Lyve1 in area 1 (K14+vimentin−), area 2 (K14+vimentin+), area 3 (K14+vimentin+CD61+) and area 4 (K14−vimentin+CD61+) in mice treated with control antibodies. d, Immunostaining showing the expression of K14 and vimentin, or YFP and K14, vimentin, CD61, CD45, CD68, CD31 or Lyve1 in area 1 and area 2 of mice treated with anti-TGFβ antibodies, showing that inhibition of TGF-β or Smad2 signalling blocks EMT progression at the early stages of EMT and no areas corresponding to late EMT states (area 3 and 4) could be detected. c, d, Scale bars, 50 μm, n = 3 biological replicates.
a, Experimental strategy to deplete macrophages by administrating anti-Csfr1 and anti-Ccl2 antibodies. b, Immunostaining of K14 and vimentin or YFP and K14, vimentin, YFP, CD45, CD68, CD31 or Lyve1 in area 1 (K14+vimentin−), area 2 (K14+vimentin+), area 3 (K14+vimentin+CD61+) and area 4 (K14−vimentin+CD61+) in mice treated with control antibodies. Scale bars, 50 μm. c, Immunostaining of K14 and vimentin or YFP and K14, vimentin, CD61, CD45, CD68, CD31 or Lyve1 in area 1 and area 2 of mice treated with anti-Csf1r and anti-Ccl2 antibodies, showing that EMT is blocked at its early stages and no areas corresponding to late EMT stages (areas 3 and 4) could be detected. Scale bars, 50 μm, n = 3 biological replicates.
About this article
Cite this article
Pastushenko, I., Brisebarre, A., Sifrim, A. et al. Identification of the tumour transition states occurring during EMT. Nature 556, 463–468 (2018). https://doi.org/10.1038/s41586-018-0040-3
This article is cited by
Military Medical Research (2022)
Which patients are prone to suffer liver metastasis? A review of risk factors of metachronous liver metastasis of colorectal cancer
European Journal of Medical Research (2022)
BMC Biology (2022)
Cell Regeneration (2022)
Stem Cell Research & Therapy (2022)