PREreview of shRNA drop-out screen identifies BRD4 targeting transcription from RNA polymerase II system to activate β-catenin to promote soft-tissue tumor proliferations

This review reflects comments and contributions from Arpita Ghosh, Teena Bajaj, Rebecca A. Shelley, Marina Schernthanner, Sourav Mukherjee, Emma Phillips & Femi Arogundade. Review synthesized by Arpita Ghosh & Garima Jain.

Brief summary of the study 

The authors present a screen carried out in soft-tissue tumour cells, in which the chromatin reader BRD4 was silenced in the presence of PDGF-BB, which was included to promote proliferation. BRD4 was chosen as it was identified in their lab in a previous screen to be important for soft-tissue tumour cell growth. They explore the effects of BRD4 silencing and inhibition on soft-tissue tumour cell morphology, proliferation and cell cycle. Some settings relate effects of BRD4 silencing to stimulation of the cells using PDGF-BB, a cytokine which, amongst other things, is known to activate beta-catenin signalling. 

Major comments  

• This preprint was extremely difficult to read as the authors did not clearly explain why or how they performed the experiments and over-state their findings throughout the manuscript. There are major omissions of the reporting of the methodologies behind key experiments - most notably how the screen was performed and what kind of cells were used. In order to improve the community value of this preprint, the authors should carefully revisit it and reframe their data in terms of what it actually concretely shows, removing all statements which are not backed up by their experiments. Any conclusions about molecular interactions, transcriptional activity, tumorigenesis, prognosis etc are not supported by the data in this paper. The authors very often make bold statements, using language such as “strongly suggest,” “key regulator” and “driver,” throughout. These should be toned down or removed. The authors should also take care to cite any previous findings they are referring to - such as the ChipSeq data they refer to here.

• The title of this manuscript clearly emphasises the shRNA screen, which is only shown and mentioned in the last figure. In addition, the ordering of the figures is not very intuitive - it might make more sense to start with the findings from figure 6 (i.e. introduce the shRNA screen, which also highlights key proteins) and then delve into mechanistic details as done in figures 1-5

• A very thorough proofreading of the entire manuscript is required to make sentences more conducive, comprehensible and error free (tense and grammar). Many sentences are incomplete in their meaning. For example, wherever there is an “effect” mentioned, on what the effect is observed is usually missing, or explanation is missing for what a “tumor D” cells are, etc.

• Looking at pRB levels is not enough to draw conclusions about apoptosis. The authors should consider exploring caspase activity or annexinV staining to investigate apoptosis. Further, an increase in subG1 DNA levels in the cell cycle analysis assay is expected if cells are dying. Additionally, it would be good to check the total GSK levels.

• More details are required for the shRNA screen mentioning controls. Top hits should be plotted and lists should be provided for the same as supplementary information.

• To comment on cellular proliferation as shown on Figure 1A, a Ki67 or BrdU assay would be required.

• The cell density plays a major role in governing cell shape and area. From the representative images in Figure 2(a), it is evident that cell density of shNTC is higher than BRD4 sh#1 and BRD4 sh#2. To strongly claim the observation from the Figure, better representative images should be chosen and details for cell number stated in the methods.

• Vinculin expression is decreased upon BRD4 knock down. Why Vinculin was chosen with pRb is unclear. While in the results section, the changes in the Vinculin expression is not mentioned. If Vinculin was chosen as a control, then a better control must be chosen as Vinculin expression depends upon changes in cellular cytoskeleton i.e., cell shape or cell area.

• In Figure 4, the authors did not distinguish between nuclear and cytoplasmic b-catenin. Given that they look at b-catenin target genes, they should have probably focused on nuclear protein fractions/lysates and then normalised the levels to a nuclear housekeeping gene.

• In Figure 5, the actin levels at 800 nM are reduced by approximately the same amount as the reduction seen for B-catenin and c-Abl. It would be recommended that the authors repeat the experiment and show quantification and statistics or remove these claims.

• In the discussion, the tone needs to be dialled down for the overhyped statements and claims which were not tested experimentally. For example, no interaction studies were carried out. Explanation for what “sorafenib” is and why it is relevant to the discussion of the data presented here is missing. More data would be required to make claims about the effect of BRD4 knockdown on the differentiation status of the cells used in this study (eg, levels of differentiation markers). The data do not allow for any conclusions about prognosis. The claim that "transcription from RNA polymerase II" activity is regulated cannot be made by simply showing that some of the candidate genes from the screen are part of this signature. The authors would need to perform additional experiments to test the involvement of RNA polymerase II. The authors showed that BRD4 silencing reduced HIF-1a expression but with that they cannot claim that BRD4 "activates" HIF-1a in these cells.

These claims should be removed or should be stated as hypotheses and should definitely not appear in the title. 

Minor comments 

• All the acronyms used should be first introduced with their full forms and also a line stating why and how that is related to this study would be useful for readers and reviewers.

• A Figure explaining the methodology for shRNA screen need not be a primary figure as it is widely known in the field. Although this can be incorporated as a supplementary figure.

• Western blots seem edited with very high contrast, should be kept neutral to observe actual effects, or else raw figures for the blots can be put in the supplementary file. 

• Microscopy images provided for cells have very low visibility. Better images should be provided.

• Missing figure legends.

Comments on reporting 

• Cell lines used should be clearly stated.

• Methods used are very vague without stating detail like concentrations used for inhibitors or in general in explaining the protocols. 

• Mutations done are very unclear; what mutations, where they have been introduced and also the rationale behind them needs to be stated.

•  All observations reported should be re-checked carefully like “significantly upregulated 72” does not state what exactly 72 is.

• It is not clear exactly how any statistical measurements were made for the experiments. The figure legend reports n=8. The authors should clarify whether they mean 8 cells were measured or 8 independent experiments. Statistical significance is missing from many figures.

• All the pathways and upstream or downstream targets mentioned in the manuscript body should be re-visited to elaborate and clearly state what target is upstream and downstream of what other target.

Suggestions for future work:

• Re-organizing the entire manuscript to make the story more interesting and also adding value to the title of the paper.

• Compiling a few figure panels together, or rearranging their order might be necessary to frame the story better.

Competing interests

The author declares that they have no competing interests.