Article activity feed

  1. Evaluation Summary:

    This paper provides insight into a potentially new genetically defined subset of prostate tumors driven by concurrent loss of two tumor suppressor genes. This study both validates previous findings and provides new data that is compelling overall. With some additional statistical and biochemical evidence to support the conclusions, the work would be of interest to cancer biologists studying molecular mechanisms of prostate cancer.

    (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. Reviewer #3 agreed to share their name with the authors.)

    Was this evaluation helpful?
  2. Reviewer #1 (Public Review):

    Gupta et al. investigate a new molecular mechanism whereby the ETS transcription factor, ETV1, is upregulated in prostate cancer. Through a series of experiments in prostate epithelial and prostate cancer cell lines, including gene knockdown, knockout and reconstitution, they demonstrated that the concomitant loss of ERF and CIC enhance malignant phenotypes such as cell viability, invasiveness and migratory capacity. Their in vitro results were supported by in vivo subcutaneous tumour xenograft assays in immunodeficient mice. Additional analyses of publicly available data and multiple in-house assays indicated that ERF and CIC target ETV1, acting as transcriptional repressors and modulating ETV1-mediated transcriptional pathways. Finally, the authors show that ETV1 chemical and genetic inhibition moderately decrease cell viability and significantly decrease invasiveness in ERF and CIC deficient prostate cancer cells.

    A major strength of this paper is the range and number of analyses performed to test their hypothesis that CIC and ERF cooperate to suppress ETS target genes in prostate cancer. The authors combine both publicly available and in-house data to answer their research questions, which are logically set out in the results section. However, there are also limitations specific to these data that slightly diminish the quality of the paper and make interpretation of their results difficult for the reader.

    The premise of the molecular work is based on data from the cBioPortal but it is difficult to fully grasp the results presented due to study and assay numbers being omitted and figures being hard to interpret. The significance (or lack thereof) is also not specified in the text for a number of the subsequent cell line analyses and could be made clearer, especially when the authors are describing a trend rather than significant results. A key analysis method, single-sample Gene Set Enrichment Analysis, used to answer a question central to the paper's conclusions (whether ERF and CIC regulate ETV1 transcription), is poorly explained and presented in the methods and results sections. Furthermore, the methods section does not align with the results section, there is a missing methodology (e.g., how was the PNT2 gene expression data generated?), there are instances of figures being misnumbered and/or insufficiently described/labelled, and missing supplementary data. Finally, while the authors present what appears to be very clinically relevant data showing sensitivity to ETV1 inhibition was enhanced in cells with both ERF and CIC loss, they only present experiments in a single prostate cancer cell line. Given the potential clinical relevance of these data, further in vitro and in vivo assays in the other available cell lines would have provided further evidence for their conclusions, especially given the higher metastatic potential of one of these (PC-3 cells).

    Despite the limitations described above, the interpretation and overall conclusions the authors draw from their analyses are generally sound. The study represents an advance in our understanding of how ETS family transcription factors are dysregulated in prostate cancer and suggests a new sub-class of prostate cancer patients based on somatic tumour alterations. Significantly, these patients could one day benefit from targeted ETV1 inhibitors, which are currently being assessed in clinical trials for other cancers.

    Was this evaluation helpful?
  3. Reviewer #2 (Public Review):

    Gupta et al test a model proposing that a subset of prostate tumors are driven by loss of the tumor suppressor transcription factors CIC and ERF, which results in higher expression of the oncogenic transcription factor ETV1. Importantly, this would provide an oncogenic role for ETV1 in ETS-fusion negative prostate cancers. This is significant because we have an incomplete understanding of all of the molecular mechanisms that can promote prostate cancer. The authors use publicly available prostate cancer patient data to show that concurrent loss of CIC and ERF occurs in about 10% of prostate tumors and that these tend to be outside of the well-known molecular subclasses of ETS-fusion, SPOP, and FOXA1. They are not, however, mutually exclusive with these subclasses. The authors then utilize a normal prostate cell line and two prostate cancer cell lines to demonstrate that CIC and ERF can inhibit cancer-related phenotypes and growth in xenograft models. The authors confirm previous findings that CIC acts to repress ETV1, ETV4, and ETV5 and discover that ERF represses ETV1, but not ETV4 or ETV5. Finally, the authors describe an intriguing but confusing finding that ETV1 inhibition reduces cell growth only when CIC is deleted. Overall, the manuscript does a good job validating previous findings in other systems in prostate cells (CIC inhibition of PEA3 factors) and adding new findings (ERF inhibition of ETV1). Together these data provide a compelling story regarding a novel subset of prostate tumors. However, there are several key claims that are based on data that are not statistically significant or properly controlled.

    1. The data do not support the conclusion that CIC and ERF cooperate to limit prostate cancer progression. Further, there is marginal data to support the conclusion that CIC and ERF have an additive effect on this process. This is because there is no statistical significance reported for comparisons between single and dual inhibition of CIC and ERF in Figure 2b, 2d, 2f, 2g, or in any of the xenograft experiments. The data in Figure 2e is not convincing on this point, and the discussion of Figure 2e in the results section is over-interpreted. As mentioned above there is a small amount of statistically significant data to support an additive effect in the DU145 system - Figures 2b and 2d.

    2. It is not shown that CIC and ERF loss correlate with an increase in ETV1 in patient data.

    3. A large number of the findings supporting the conclusions are based on exogenous over-expression of ERF. It is unclear how high this over-expression is (is it higher than ERF levels in PNT2 cells), and if this exogenous expression is generally toxic to all cells. Further, all of the phenotypic changes measured (clonogenic growth, invasion, xenograft growth) could be the result of altered cell proliferation. It is unclear what the data in Figures 3c and 3h really represent. These could be endpoints to a growth curve, but after 7 days of growth, these cells should be over-confluent or were so sparsely plated that this is essentially a clonogenic growth assay.

    4. ChIP-PCR experiments in Figure 4e-g are not properly controlled. Crosslinking immunoprecipitation with an antibody to a nuclear protein often increases pull-down of all genomic DNA non-specifically. Further, it is not clear why the text states that the authors were unable to immunoprecipitate DNA with the anti-ERF antibody (with regard to ChIP-seq), but can for ChIP-PCR.

    5. Findings that clearly do not reach statistical significance (Such as Fig 4l) should not be reported as showing an effect in the results section.

    6. The results in Figure 5 showing the necessity of ETV1 function only when CIC is depleted are intriguing. However, the size of the error bars makes these findings difficult to interpret. In particular, it is unclear if there is an effect of siCIC or sgCIC1 alone.

    Was this evaluation helpful?
  4. Reviewer #3 (Public Review):

    This study highlights the functional consequences of combined genomic losses of CIC and ERF which results in the activation of ETV1, in the absence of the canonical fusion event involving TMPRSS2 in a subset of prostate cancer. ETV1 is an oncogenic driver of cell growth and metastatic behaviour in many cancer types including prostate cancer. The experiments performed provided tantalizing evidence on the biological and functional consequences of combined losses of CIC and ERF and appeared to support the findings of the mined publicly available cancer genomic datasets.

    The manuscript could be improved by providing evidence of proteomic interactions between CIC and ERF proteins in the form of immune-precipitation and Western protein blots. The authors had provided predominantly genomic, transcriptomic, and functional data. In most parts, the manuscript is logical and thorough and leveraged available genomic data. This is followed by genomic-functional experimentations. Given the postulate of co-operativity between CIC and ERF, it would be logical to investigate their potential proteomic interactions.

    Was this evaluation helpful?