Alternative splicing downstream of EMT enhances phenotypic plasticity and malignant behavior in colon cancer
Curation statements for this article:-
Curated by eLife
eLife assessment
This study provides a valuable analysis of the splicing landscape in colon cancer cells that have properties intermediate between those typically found in primary cancers ("epithelial") and those that are spreading by metastasis ("mesenchymal"). The strength of evidence provided is wide ranging and convincing, and supports current ideas that changes in the way that RNA from particular genes is processed plays a key role in cancer spread.
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
- Cancer Biology (eLife)
Abstract
Phenotypic plasticity allows carcinoma cells to transiently acquire the quasi-mesenchymal features necessary to detach from the primary mass and proceed along the invasion-metastasis cascade. A broad spectrum of epigenetic mechanisms is likely to cause the epithelial-to-mesenchymal (EMT) and mesenchymal-to-epithelial (MET) transitions necessary to allow local dissemination and distant metastasis. Here, we report on the role played by alternative splicing (AS) in eliciting phenotypic plasticity in epithelial malignancies with focus on colon cancer. By taking advantage of the coexistence of subpopulations of fully epithelial (EpCAM hi ) and quasi-mesenchymal and highly metastatic (EpCAM lo ) cells in conventional human cancer cell lines, we here show that the differential expression of ESRP1 and other RNA-binding proteins (RBPs) downstream of the EMT master regulator ZEB1 alters the AS pattern of a broad spectrum of targets including CD44 and NUMB , thus resulting in the generation of specific isoforms functionally associated with increased invasion and metastasis. Additional functional and clinical validation studies indicate that both the newly identified RBPs and the CD44s and NUMB2/4 splicing isoforms promote local invasion and distant metastasis and are associated with poor survival in colon cancer. The systematic elucidation of the spectrum of EMT-related RBPs and AS targets in epithelial cancers, apart from the insights in the mechanisms underlying phenotypic plasticity, will lead to the identification of novel and tumor-specific therapeutic targets.
Article activity feed
-
-
Author Response
Reviewer #1 (Public Review):
Switching between epithelial and mesenchymal populations is an important stage for cancer growth and metastasis but difficult to study as the cells in this transition are rare. In this study Xu et al investigate changes the splicing regulator environment and changes in specific splice events by monitoring colon cancer cell populations that have epithelial and mesenchymal properties (so are potentially in transition) compared their epithelial partners. Using these potentially transitioning cells should reveal new insights into the causative changes occurring during EMT, a key life threatening step in colon cancer progression, and other cancers too.
The authors were trying to establish if changes in the splicing environment occurred between epithelial and quasi-mesenchymal cells and to …
Author Response
Reviewer #1 (Public Review):
Switching between epithelial and mesenchymal populations is an important stage for cancer growth and metastasis but difficult to study as the cells in this transition are rare. In this study Xu et al investigate changes the splicing regulator environment and changes in specific splice events by monitoring colon cancer cell populations that have epithelial and mesenchymal properties (so are potentially in transition) compared their epithelial partners. Using these potentially transitioning cells should reveal new insights into the causative changes occurring during EMT, a key life threatening step in colon cancer progression, and other cancers too.
The authors were trying to establish if changes in the splicing environment occurred between epithelial and quasi-mesenchymal cells and to what extent this is important for colon cancer in establishing gene expression programs and cell behavior related to metastasis. The take home message is that these more "plastic" mesenchymal cells are expressing the mesenchymal transcription factor ZEB1 and reducing expression of the epithelial splicing factor ESRP1 (as well as some other RBPs). The FACS analysis showing that over-expression of ESRP1 alone can switch cell population ratios is very clear and indicates that reduction of this RBP plays a key role in making cells more metastatic. The lentiviral overexpression of CD44s and NUMB2/4 had very dramatic effects on increasing metastatic cellular properties. The clinical stratification analysis of splice isoforms and ZEB1/ESRP1 expression was very informative for understanding what is happening in actual tumors. The methods used and results from these studies are likely to have an impact on understanding the gene expression changes that take place during EMT.
Strengths: The authors have used cell lines that model switching cells between epithelial and quasimesenchymal, based on expression of the markers Epcam (epithelial cell adhesion molecule expressed in epithelial cells) and CD44. The study utilizes shRNA-mediated knockdown and lentiviral overexpression of
ESRP1 and splice isoforms, and monitors endogenous mRNA splice isoforms by RNAseq and qRTPCR, protein isoforms by western, cell surface expression of EpCAM and CD44 using FACS and metastatic potential using a mouse model, and patient gene expression data from TCGA.
Weaknesses: Some of the data here might be novel for colon cancer, but the roles of these RNA binding proteins and ESRP1 target exons are better known in other cancers. Both CD44 and NUMB are known ESRP1 targets already in cells undergoing plasticity (e.g. PMID: 30692202). RBM47 is already known to be downregulated in EMT and quaking upregulated (PMID: 28680090; PMID: 27044866). There is also a lot of literature on ESRP1 expression in cancer and EMT. This should be better discussed.Out of the 3 references mentioned, 2 are already discussed in the submitted manuscript, while the third (Rokavec et al.) has now been added to the Discussion. As specified above, we never claimed to be the first to report on these RBPs and downstream AS targets. Unfortunately, it is not clear how the reviewer wants us to improve on these aspects (“should be better discussed” is rather vague) but we have now tried to extend the discussion relative to these issues in the revised manuscript.
-
eLife assessment
This study provides a valuable analysis of the splicing landscape in colon cancer cells that have properties intermediate between those typically found in primary cancers ("epithelial") and those that are spreading by metastasis ("mesenchymal"). The strength of evidence provided is wide ranging and convincing, and supports current ideas that changes in the way that RNA from particular genes is processed plays a key role in cancer spread.
-
Reviewer #1 (Public Review):
Switching between epithelial and mesenchymal populations is an important stage for cancer growth and metastasis but difficult to study as the cells in this transition are rare. In this study Xu et al investigate changes the splicing regulator environment and changes in specific splice events by monitoring colon cancer cell populations that have epithelial and mesenchymal properties (so are potentially in transition) compared their epithelial partners. Using these potentially transitioning cells should reveal new insights into the causative changes occurring during EMT, a key life threatening step in colon cancer progression, and other cancers too.
The authors were trying to establish if changes in the splicing environment occurred between epithelial and quasi-mesenchymal cells and to what extent this is …
Reviewer #1 (Public Review):
Switching between epithelial and mesenchymal populations is an important stage for cancer growth and metastasis but difficult to study as the cells in this transition are rare. In this study Xu et al investigate changes the splicing regulator environment and changes in specific splice events by monitoring colon cancer cell populations that have epithelial and mesenchymal properties (so are potentially in transition) compared their epithelial partners. Using these potentially transitioning cells should reveal new insights into the causative changes occurring during EMT, a key life threatening step in colon cancer progression, and other cancers too.
The authors were trying to establish if changes in the splicing environment occurred between epithelial and quasi-mesenchymal cells and to what extent this is important for colon cancer in establishing gene expression programmes and cell behaviour related to metastasis. The take home message is that these more "plastic" mesenchymal cells are expressing the mesenchymal transcription factor ZEB1 and reducing expression of the epithelial splicing factor ESRP1 (as well as some other RBPs). The FACS analysis showing that over-expression of ESRP1 alone can switch cell population ratios is very clear and indicates that reduction of this RBP plays a key role in making cells more metastatic. The lentiviral overexpression of CD44s and NUMB2/4 had very dramatic effects on increasing metastatic cellular properties. The clinical stratification analysis of splice isoforms and ZEB1/ESRP1 expression was very informative for understanding what is happening in actual tumours. The methods used and results from these studies are likely to have an impact on understanding the gene expression changes that take place during EMT.
Strengths. The authors have used cell lines that model switching cells between epithelial and quasi-mesenchymal, based on expression of the markers Epcam (epithelial cell adhesion molecule expressed in epithelial cells) and CD44. The study utilises shRNA-mediated knockdown and lentiviral overexpression of ESRP1 and splice isoforms, and monitors endogenous mRNA splice isoforms by RNAseq and qRTPCR, protein isoforms by western, cell surface expression of EpCAM and CD44 using FACS and metastatic potential using a mouse model, and patient gene expression data from TCGA.
Weaknesses: Some of the data here might be novel for colon cancer, but the roles of these RNA binding proteins and ESRP1 target exons are better known in other cancers. Both CD44 and NUMB are known ESRP1 targets already in cells undergoing plasticity (e.g. PMID: 30692202). RBM47 is already known to be downregulated in EMT and quaking upregulated (PMID: 28680090; PMID: 27044866). There is also a lot of literature on ESRP1 expression in cancer and EMT. This should be better discussed.
-
Reviewer #2 (Public Review):
In the submitted article, Xu and co-workers have explored the alternative splicing of CD44 and NUMB isoforms responsible for promoting epithelial-to-mesenchymal transition in quasi-mesenchymal and highly metastatic subtype of colon cancer. In this regard, the authors have performed numerous RNA-seq and Gene Ontology analyses to identify differentially expressed RNA binding proteins and their associated pathways to understand the related alternative splicing events. CD44s and NUMB2/4 spliced isoforms have been identified as promoting the invasive and metastatic properties while negatively affecting the proliferation of the HCT116 and SW480 cells in Zeb1-ESPR1-dependent manner. Unfortunately, there exists discrepancy and inconsistency at a large extent in the experimental data, along with lack of novel …
Reviewer #2 (Public Review):
In the submitted article, Xu and co-workers have explored the alternative splicing of CD44 and NUMB isoforms responsible for promoting epithelial-to-mesenchymal transition in quasi-mesenchymal and highly metastatic subtype of colon cancer. In this regard, the authors have performed numerous RNA-seq and Gene Ontology analyses to identify differentially expressed RNA binding proteins and their associated pathways to understand the related alternative splicing events. CD44s and NUMB2/4 spliced isoforms have been identified as promoting the invasive and metastatic properties while negatively affecting the proliferation of the HCT116 and SW480 cells in Zeb1-ESPR1-dependent manner. Unfortunately, there exists discrepancy and inconsistency at a large extent in the experimental data, along with lack of novel findings as CD44 and NUMB alternative splicing is well investigated in other types of cancers.
-
-
-