Extracellular signal-regulated kinase mediates chromatin rewiring and lineage transformation in lung cancer
Curation statements for this article:-
Curated by eLife
Evaluation Summary:
This manuscript will be of interest to cancer biologists studying cell fate transitions, particularly adenocarcinoma-to-small cell transitions that occur in prostate and lung cancer, which is a timely topic. While there is not a single linear mechanism identified that fully explains Kras-induced neuroendocrine cell fate suppression in all contexts, multiple new findings will likely be built upon by the field. Overall, the data are properly controlled and the key claims are supported.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. All reviewers agreed to share their names with the authors.)
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
- Cancer Biology (eLife)
Abstract
Lineage transformation between lung cancer subtypes is a poorly understood phenomenon associated with resistance to treatment and poor patient outcomes. Here, we aimed to model this transition to define underlying biological mechanisms and identify potential avenues for therapeutic intervention. Small cell lung cancer (SCLC) is neuroendocrine in identity and, in contrast to non-SCLC (NSCLC), rarely contains mutations that drive the MAPK pathway. Likewise, NSCLCs that transform to SCLC concomitantly with development of therapy resistance downregulate MAPK signaling, suggesting an inverse relationship between pathway activation and lineage state. To test this, we activated MAPK in SCLC through conditional expression of mutant KRAS or EGFR, which revealed suppression of the neuroendocrine differentiation program via ERK. We found that ERK induces the expression of ETS factors that mediate transformation into a NSCLC-like state. ATAC-seq demonstrated ERK-driven changes in chromatin accessibility at putative regulatory regions and global chromatin rewiring at neuroendocrine and ETS transcriptional targets. Further, ERK-mediated induction of ETS factors as well as suppression of neuroendocrine differentiation were dependent on histone acetyltransferase activities of CBP/p300. Overall, we describe how the ERK-CBP/p300-ETS axis promotes a lineage shift between neuroendocrine and non-neuroendocrine lung cancer phenotypes and provide rationale for the disruption of this program during transformation-driven resistance to targeted therapy.
Article activity feed
-
-
Reviewer #3 (Public Review):
The authors have studied preclinical models of human small cell lung cancer (SCLC) using characterized SCLC cell lines that have been manipulated to conditionally express mutant EGFR (L858R) or KRAS (G12V) alleles and then assessing their morphology in cell culture, expression of neuroendocrine differentiation markers and transcription factors, and main signaling pathways such as the MAPK pathway. They focus on this because activation of ERK and the MAPK pathways are seen in nearly all non-small cell lung cancers (NSCLCs) including those with EGFR or KRAS mutations but mutations in these driver oncogenes or active ERK and MAPK pathway are essentially never found in SCLCs. In addition, chromatin modifications are assessed after manipulations and functional genomics targeting and pharmacologic inhibition of …
Reviewer #3 (Public Review):
The authors have studied preclinical models of human small cell lung cancer (SCLC) using characterized SCLC cell lines that have been manipulated to conditionally express mutant EGFR (L858R) or KRAS (G12V) alleles and then assessing their morphology in cell culture, expression of neuroendocrine differentiation markers and transcription factors, and main signaling pathways such as the MAPK pathway. They focus on this because activation of ERK and the MAPK pathways are seen in nearly all non-small cell lung cancers (NSCLCs) including those with EGFR or KRAS mutations but mutations in these driver oncogenes or active ERK and MAPK pathway are essentially never found in SCLCs. In addition, chromatin modifications are assessed after manipulations and functional genomics targeting and pharmacologic inhibition of various components of the MAPK pathway are tested to see their effect on NE expression. Because of the known clinical phenomenon of transformation to SCLC like tumors by lung adenocarcinomas with EGFR mutations that become resistant to EGFR tyrosine kinase inhibitors, findings from the SCLC studies were applied to try to experimentally generate such LUAD to SCLC transformation. Overall, they found that activation of ERK/MAPK pathway by oncogenic mutations led to loss of NE differentiation and that the "ERK-CBP/p300-ETS axis promotes a lineage shift between neuroendocrine and non-neuroendocrine lung cancer phenotypes". They conclude: "In summary, we provide the first reported biological rationale for why alterations in MAPK pathway are rarely found in SCLC and describe the molecular underpinnings of how the central node in this pathway, ERK2, suppresses the NE differentiation program. " The authors conclusions and claims are justified by the experiments and data they present and they provide a mechanistic basis of what happens with MAPK/ERK activation in SCLC, why one does not find MAPK/ERK activation in SCLC, or the presence of related oncogenic driver mutations such as mutant KRAS or EGFR.
-
Reviewer #2 (Public Review):
Cell fate transitions (such as adenocarcinoma converting to small cell neuroendocrine fate) are an increasing phenomenon observed during therapeutic resistance in lung cancer, prostate cancer, and possibly other cancer types. It is important to understand these mechanisms if we ultimately seek to tailor treatment to a patient's disease and/or to control the pathways that lead to treatment resistance. However, the mechanisms that underly these cell fate changes are not well understood. It has been previously observed (Calbo et al, Cancer Cell, 2011) that activated mutant Kras (commonly associated with adenocarcinoma fate) can promote a non-neuroendocrine fate in SCLC, but the mechanisms are unknown.
Predominantly using three human small cell lung cancer (SCLC) cell lines, Inoue and colleagues use genetic and …
Reviewer #2 (Public Review):
Cell fate transitions (such as adenocarcinoma converting to small cell neuroendocrine fate) are an increasing phenomenon observed during therapeutic resistance in lung cancer, prostate cancer, and possibly other cancer types. It is important to understand these mechanisms if we ultimately seek to tailor treatment to a patient's disease and/or to control the pathways that lead to treatment resistance. However, the mechanisms that underly these cell fate changes are not well understood. It has been previously observed (Calbo et al, Cancer Cell, 2011) that activated mutant Kras (commonly associated with adenocarcinoma fate) can promote a non-neuroendocrine fate in SCLC, but the mechanisms are unknown.
Predominantly using three human small cell lung cancer (SCLC) cell lines, Inoue and colleagues use genetic and pharmacological approaches to focus on potential mechanisms by which Egfr/Kras/Mapk signaling can repress neuroendocrine fate. They make a number of interesting observations that extend our understanding of neuroendocrine cell fate regulation including:
Kras-induced NE suppression appears to depend mostly on ERK2, and not ERK1 or PI3K signaling.
Kras activation induces chromatin changes including increased H3K27Ac in 2/3 cell lines; increased H3K27Ac in response to HDAC inhibition is associated with NE suppression. Pharmacological inhibition of CBP/p300 (a HAT that promotes H3K27Ac) reduces H3K27Ac and restores NE suppression. Altogether, these findings are consistent with the notion that SCLC cannot tolerate high levels of H3K27Ac.
Kras induces the MSK/RSK pathway consistently in cell lines but appears to be functionally-relevant to NE fate only in H82 cells.
Kras activation induces chromatin occupancy at ERG and ETS family transcription factor (Etv1, 4, 5) binding sites in 2/3 cell lines, and induces ETV4 (2/3 lines) and ETV5 protein levels (3/3 lines). ETV1 and ETV5 overexpression are sufficient to inhibit NE fate markers in context-dependent manner. Ets family induction appears to occur in a CIC-independent manner.
In addition, some interesting negative data is presented, for example, SOX9 is induced upon Kras activation in 3/3 cell lines but it was not functionally relevant for NE suppression; Notch1, Notch2, and HES1 (known NE fate suppressors) are induced by Kras activation in a cell context-specific manner, but they did not appear functionally-relevant to NE suppression based on HES1 knockout and a pharmacological inhibitor of Notch signaling; Rb1 loss was not sufficient to promote NE fate in EGFR/p53 mutant cell lines, despite its known association with adeno-to-SCLC conversion. Overall, the conclusions in the manuscript are well justified. These findings will be of interest to those especially in lung and prostate cancer studying cell fate conversions in the context of EGFR and AR inhibitor resistance, respectively. These observations will be built upon by these fields.
Weaknesses:
- One recurring issue in the manuscript is that the observations are often not consistent across the three cell lines and are context-specific effects, and the potential reasons could be explained better. The cell lines chosen unfortunately (but interestingly) represent some of the major cell states of SCLC. H2107 represents the ASCL1+ NE-high subset of SCLC (and has some MYCL). H82 and H524 represent the C-Myc (MYC)-high subset of SCLC, with H82 having a MYC amplification, and both representing the NEUROD1 subtype (which tend to be associated with more MYC). Assessment of NE score using a common approach in the field (Zhang et al, TLCR) shows that H82 cells are already considerably NE-low, with H524 as NE-intermediate/high, and H2017 as NE-high. So, this may be related to why H82 seemed to be the most permissive cell line to change NE fate in multiple assays.
In addition, H2107 and H524 appear to have EP300 mutations, which may contribute to their NE-high nature and contribute to the refractory response to A485 treatment based on the author's model. It's known that MYCL and MYC-driven cell lines differ in numerous aspects from transcriptional signatures, super enhancer usage, metabolic regulation, therapeutic response, etc. This information could be mentioned in the results and discussed when mentioned as a factor near line 540.
- Related to Figure 4, the authors show that p300 pharmacological inhibition can restore NE fate in presence of Kras. Given that drugs can have off-target effects, it would be helpful to know if genetic knockdown/knockout of p300 phenocopies these effects. Given that CREBBP (CBP) or EP300 (p300) mutations are common in SCLC, it is also relevant whether any of these cell lines have CREBBP (CBP) or EP300 (p300) mutations. It appears H2107 and H524 may have EP300 mutations, and it would be good to know whether the authors have tried to restore EP300 function.
-
Reviewer #1 (Public Review):
The paper is investigating the mechanism of lineage switch in lung cancer. In about 10-15% of lung cancers treated with inhibitors of oncogenic receptors such as EGFR or KRAS, cancer cells emerge over time with newly acquired features of neuroendocrine differentiation. The authors proposed that it is a direct result of inhibition of MAPK pathway signaling so that reduced MAPK activity activates previously silent genes regulating neural crest differentiation. While this theory is of interest, the experiments presented herein are construed on the opposite sequence by way of introducing activated MAPK via oncogenic KRAS or EGFR to 3 neuroendocrine cell lines resulting in lower expression of neuronal transcription factors. The authors propose MAPK-activated ETS family TFs are responsible for the repression of NE …
Reviewer #1 (Public Review):
The paper is investigating the mechanism of lineage switch in lung cancer. In about 10-15% of lung cancers treated with inhibitors of oncogenic receptors such as EGFR or KRAS, cancer cells emerge over time with newly acquired features of neuroendocrine differentiation. The authors proposed that it is a direct result of inhibition of MAPK pathway signaling so that reduced MAPK activity activates previously silent genes regulating neural crest differentiation. While this theory is of interest, the experiments presented herein are construed on the opposite sequence by way of introducing activated MAPK via oncogenic KRAS or EGFR to 3 neuroendocrine cell lines resulting in lower expression of neuronal transcription factors. The authors propose MAPK-activated ETS family TFs are responsible for the repression of NE lineage.
Several principal issues presented by the authors raise some concerns:
Despite presenting some evidence to the effect of suppression of NE transcription factors by overactivating MAPK signaling, the conversion of adenocarcinoma to NE (the opposite transition) is not being addressed in the paper. Therefore, it is rather illogical to investigate the process of transition that is not taking place in the real world.
The authors do not consider a possibility of multi-clonality of human cancers and clonal competition as a mechanism leading to acquired resistance and the emergence of NE clones that are not suppressible by the inhibitors of MAPK pathway (e.g. EGFR inhibitors, or KRAS/RAF/MEK inhibitors). Starting the experiments with clonal populations of long-term cultured cell lines may be an insurmountable difficulty to switch these cells between the epithelial and NE phenotypes which proved to be frustratingly non-productive in the hands of the authors. Taken out of context of tumor microenvironment, these phenotypic transitions may be co-regulated by a combination of cell-intrinsic and extrinsic factors.
Despite zeroing in on ETVs downstream of ERK1/2, the paper does not go as far as showing the direct effect of these TFs as repressors of NE differentiation (ASCL1, BRN2, NEUROD1 etc.).
The line of evidence that Dox-activated MAPK signaling via massive over expression of KRAS or EGFR induces dramatic increase in marks of transcriptionally active chromatin (such H3K27ac and others) is to be expected in this entirely artificial system. Indeed, the addition of doxycycline results in massive burst of proliferation and overexpression of ETV1 and ETV4, the canonical MAPK targets. Again, this switch appears unrelated with the opposite of epithelial-to-NE de-differentiation.
-
Evaluation Summary:
This manuscript will be of interest to cancer biologists studying cell fate transitions, particularly adenocarcinoma-to-small cell transitions that occur in prostate and lung cancer, which is a timely topic. While there is not a single linear mechanism identified that fully explains Kras-induced neuroendocrine cell fate suppression in all contexts, multiple new findings will likely be built upon by the field. Overall, the data are properly controlled and the key claims are supported.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. All reviewers agreed to share their names with the authors.)
-
