The G protein-coupled receptor TBXA2R activates ERMs to control cell motility and invasion of triple-negative breast cancer cells.

  1. Kevin Leguay
  2. Omaima Naffati
  3. Yu Yan He
  4. Mireille Rogue
  5. Chloe Tesniere
  6. Melania Gombos
  7. Hellen Kuasne
  8. Louis Gaboury
  9. Christian Le Gouill
  10. Sylvain Meloche
  11. Michel Bouvier
  12. Sebastien Carreno

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Cell migration and invasion are critical processes for cancer cell metastasis, relying on the ability of cells to adapt their morphology. Proteins of the ezrin, radixin, and moesin (ERM) family are key regulators of cell morphogenesis and essential determinants of cancer cell metastasis. However, the mechanisms by which ERMs are activated in metastatic cells remain poorly understood. Here, we identify the thromboxane A2 receptor (TBXA2R), a G protein-coupled receptor overexpressed in multiple cancers, as a critical activator of ERMs, enhancing the motility and invasion of triple-negative breast cancer (TNBC) cells. We found that TBXA2R activates ERMs by engaging the Gαq/11 and Gα12/13 subfamilies, the small GTPase RhoA, and its Ser/Thr kinase effectors SLK and LOK. Furthermore, we demonstrate that TBXA2R promotes TNBC cell motility and invasion in vitro and metastatic colonization in vivo, dependent on ERM function. These findings reveal a novel signaling axis by which a member of the largest class of receptors activates key metastatic determinants, thereby controlling various aspects of metastasis. This discovery opens new avenues for developing targeted therapies against cancer metastasis.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    The authors do not wish to provide a response at this time.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #4

    Evidence, reproducibility and clarity

    Overall, the authors show an interesting and conclusive work on the activation of ERM proteins upon TBXA2R signaling. The use of the ebBRET biosensor to assess ERM-protein activation enables elegant investigation of activation modalities. The Thromboxane A2 analogue U46619 robustly shows activation of ERM proteins in ebBRET assays as well as an increase in ERM-protein phosphorylation status. The functional effects of this signaling pathway are shown convincingly for moesin, where moesin mediates an TBXA2R mediated increase in cell motility, invasion and metastasis of triple-negative breast cancer Hs578 cells in vitro and in vivo. Nonetheless, some points need to be clarified.

    Significance

    Comment 1: In the title the authors state, that ERM-activation via TBXA2R is controlling invasion and motility of triple-negative breast cancer cells. In the manuscript, there is only data supporting this assumption for moesin (MSN). Therefore, the authors need to change the title accordingly or support additional experiments for the other two ERM-proteins radixin and ezrin. Throughout the experiments, the p-ERM antibody is used to measure ERM-protein activation. Since the effects on invasion and motility observed in Hs578 cells are mainly mediated through moesin, it would be necessary to see, at least for one experiment per cell line (HEK293T, Hs578) the detailed phosphorylation status of ezrin, radixin and moesin separately. As there are specific, phospho-detecting antibodies for this case, this could be done rather easy. Furthermore, showing specific increase of phosphorylated moesin would support the functional data shown in Figure 5 and 6. To investigate the functional effect of TBXA2R mediated activation of ezrin and radixin on cell motility and invasion, similar experiments could be done in e.g. HMC-1-8 breast cancer cells (high ezrin expression) and HCC1187 (high radixin expression).

    Comment 2: Figure 1A, C, D: The concentration of staurosporine is with 100 nM relatively high for kinase inhibition. It would be informative to see the assay with increasing staurosporine concentrations, e.g. from 1 nM to 50 nM. In general, a concentration of 1-10 nM should be sufficient for kinase inhibition, preventing unspecific effects of the drug.

    Comment 3: The citation for the p-ERM antibody is confusing, as there is only p-Moe used in the cited paper (Roubinet, 2011). There is a p-ERM antibody commercially available (Cell Signaling, Phospho-ezrin (Thr567)/radixin (Thr564)/moesin (Thr558) Antibody #3141). Could you clarify which antibody you are using?

    Comment 4: From the inhibitor experiments using C3 transferase toxin (Figure 2), the authors conclude that RhoA plays a role in TBXA2R mediated ERM activation. As mentioned in the manufacturer's description, C3 toxin is inhibiting RhoA, RhoB and RhoC. Therefore, it would be necessary to repeat those experiments under RhoA knockdown conditions (e.g. using an siRNA-based approach) to state that specifically RhoA is involved.

    Comment 5: To assess, if the findings in Figure 5 and 6 are due to the higher moesin expression in Hs578 cells or are linked to a specific function of moesin, a re-expression experiment would be informative. To achieve this, the 2D and 3D migration experiments could be redone after re-expression of moesin, ezrin and radixin separately in moesin knockdown conditions.

    Minor comments:

    • Even though U46619 is a known Thromboxane A2 analogue, including negative and positive controls would strengthen the results. In detail, this could be done by showing a known protein which gets phosphorylated downstream of TBXA2R signaling and a protein which is not affected by this signaling pathway alongside the shown effects on ERM-proteins.
    • Figure 1 J: There are no statistics comparing the conditions of SQ-29548 treated cells in presence/absence of U46619, that should be added.
    • Figure 1 G, H: How was the quantification for cell periphery performed? In detail, how were the thresholds set for cell periphery / not cell periphery?
    • Figure 3 H:
      • The labelling indicating presence of U46619 is missing.
      • Also, what is the rationale behind normalizing MB-453 for 3 cell lines and comparing the BT-549 to MB-157?
    • Suppl. Fig 4 D: Define y-axis better. Absorbance at what wave length?
    • Define FERM and ERMAD abbreviations in introduction.
  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary

    The Ezrin, radixin, and moesin (ERM) family of proteins orchestrate morphological changes that potentiate metastatic invasion in cancer cells. In this study, Leguay et al. identify the GPCR, TBXA2R, as a key activator of the ERM proteins which promotes motility and invasion in triple-negative breast cancer (TNBC) cells. Using BRET-based sensors developed by them previously for monitoring the activation of ERM proteins and building upon their previous findings on the role of the small GTPase RhoA in the activation of ERM proteins, the authors carefully dissect the molecular pathway leading to the activation of ERM proteins upon stimulation of the TBX2AR. The authors also establish the pathological relevance of the pathway in TNBC using in vitro and in vivo models, opening up possibilities for targeting this pathway in cancer cells. Overall, the study is well-conceived and executed, and the results are clearly described and presented in the manuscript. However, the following comments must be addressed before publication.

    Major comments

    Fig 1C - Why p-ERM was normalized over Ezrin and not ERM? It would be more appropriate and consistent to normalize against the ERM signal as done in other experiments in the manuscript.

    Fig 1E and S3C - The levels of total ERM also seem to change with increasing treatment times. This must be clarified and discussed in the manuscript.

    Fig 1F - Why is the mean of all three independent experiments not presented here as in S3C?

    Fig 2E - Though SLK seems to play a dominant role in the phosphorylation of ERM in HEK293T cells, the depletion of LOK also substantially reduces the phosphorylation of ERM in the representative figure (Fig 2E), which is not reflected in the quantification (Fig 2F). Indeed, both SLK and LOK seem to be equally crucial in Hs578T cells (Fig 4I), unlike the conclusion here. The authors must check if the quantifications were affected by any white spots in the blot for total ERM as seen in the representative figure. If necessary, the authors must include additional replicates, and the model in Fig 2G should be updated accordingly. If the contributions of LOK are indeed quite minimal in HEK293T cells, then the difference in Hs578T cells must be adequately highlighted and discussed rather than broadly mentioning similar results were observed in both cell lines. The discussion mentions that SLK kinases are the only kinases needed for ERM activation, which conflicts with findings from Hs578T cells, where both SLK and LOK contribute to ERM phosphorylation (Fig 4I). The authors should revise this to reflect their data accurately.

    Minor comments

    FigS3B should cite the source dataset and not just the database. Also, details of how the extracted data was processed (if any) should be described clearly.

    When multiple treatments are involved (for, e.g. U46619 and staurosporine), the exact sequence of treatments and the overlap in timings of different treatments must be clearly mentioned. E.g. fig 1A and 1C. There are a few grammatical errors which need to be fixed. E.g. Paragraph 2 in the second section of results - We next aimed to identify (not identifying) which kinase(s) acts downstream of TBX2AR

    Significance

    Triple-negative breast cancer, which is characterized by a lack of estrogen, progesterone or HER2 receptors, is a highly metastatic and aggressive form of breast cancer with poor prognosis. Currently, there are fewer treatment options than other types of invasive breast cancer. The current study opens up the possibility of targeting the TBXA2R or the downstream signalling components in TNBC, which are still expressed in TNBC cells. However, certain TNBC sub-types express low levels of p-ERM and TBX2AR (Fig 3E, 3F), indicating a minor role for TBX2AR pathway and targeting this pathway in these subtypes may be inefficient. In addition, certain subtypes express high p-ERM and low TBX2AR indicating alternative pathways for ERM activation. Currently, it is not clear which other GPCRs can contribute to ERM activation by engaging similar downstream effectors. A comprehensive screening of different GPCR antagonists could identify alternative strategies to target the ERM-mediated metastasis in TNBC cells that show low expression of TBX2AR.

    Audience The manuscript is relevant to a broad audience, especially to cell biologists, cancer biologists and clinical scientists.

    The reviewer's field of expertise includes cell signaling, gene expression, and RNA biology in mammalian systems. Moderate expertise in cancer biology. Limited knowledge of histopathological analysis.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Leguay et al present an interesting and logical series studies that investigate the activity and signaling of the GPCR TBXA2R in TNBC cells. The premise of the overall study is that metastasis is often associated with a more invasive/motile cancer cell phenotype. The investigators have an interest in ERM (Ezrin, Radixin, Moesin) proteins, which have been implicated in cell motility. The authors link stimulation of TBXAR2, a GPCR, to activation of ERM proteins and also show that TBXAR2 is associated with worse outcome in TNBC patients. Through the use of genetic and pharmacologic tools the authors provide convincing biochemical and cell based data to support their model that stimulation of TBXAR2 activates Gα11 & Gα12/13 which subsequently stimulate RhoA and SLK/LOK which then phosphorylate ERMs. The authors show relevant biologic consequences of the pathway. Data include orthogonal assays with similar results and the manuscript is written clearly and the data are displayed well. Overall it is a solid story that is largely well done. There are a few comments that should be addressed.

    Comments:

    1. All the biochemical/cell based in vitro data exploit the use of small molecule agonists of TBXAR2, not the natural ligand. A comment on this and why use of TXA2 is not feasible would be helpful to the reader.
    2. The data in figures 1-5 are solid and clear. However, I suggest adding a higher magnification inset for the IHC images shown in Fig 3E. It would be useful to be able to distinguish cells in the IHC, a higher mag shot should suffice.
    3. A) The use of Hs578t cells for the in vivo modeling is unfortunate. Additionally, the use of iv injection to in a study focused on cell invasion is also unfortunate. The metastatic propensity of Hs578t is not clear, in fact a recent report comparing metastasis in breast cancer cell lines shows that Hs578t perform poorly in terms of metastasis after orthotopic injection (see PMID 38468326). I searched the literature a bit to try and find other examples of iv injection of Hs578t cells, I found 1 (PMID:27654855, I did not search exhaustively), this paper shows significant lung metastasis and does not mention liver metastases. Were other breast cancer cells investigated for the in vivo studies?

    B) Why I was interested is because the typical organ that is seeded post iv injection is the lungs (as seen in the above ref), liver metastases post iv injection are not common, especially with breast cancer cells. What did the lungs look like in your experiments?

    C) Further while the data presented in figure 6 are supportive of the overall conclusions, the data is modest at best in terms of metastatic burden. Repetition of the experiment using a breast cancer cell line injected orthotopically would likely be more useful in highlighting the importance of the pathway to metastasis.
    I understand performing an orthotopic assay may be outside the scope of the study, but it would provide greater impact given the focus of the paper on cell invasion.

    Cross-commenting

    I think reviewer comments are generally aligned. I was least critical but appreciate the concerns of the other reviewers, especially rev #1 who requested additional validation and controls. In my opinion in vivo studies are not robust, I expect that is due to cell line choice. Repetition of the in vivo study with a breast cancer cell line that is capable of metastasis (from a primary tumor) would be more effective.

    Significance

    The manuscript presents a solid, logical flow and the biochemical/cell based in vitro data are clean. Clear differences between groups, appropriate controls, and displayed effectively.

    The challenge is the in vivo study. IV injection of cancer cells is a valid model for seeding and growing in a target organ BUT it does not reflect cell invasion, which is typically thought of as a step that occurs earlier in the metastatic cascade. That said, the data are supportive with conclusions but not necessarily consistent with expected results based on iv injection of this cell line. A caveat is that the cell line used is characterized as having metastatic characteristics in vitro but is not a consistent metastatic line in vivo. The recommendation is the perform a new in vivo experiment. An orthotopic injection of a strongly metastatic cell line, such as MDA MB 231 or other (see paper ref aboved) would be a more stringent and accurate test of the importance of the pathway to cell invasion in vivo.

  5. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Summary

    This manuscript investigates the role of the thromboxane A2 receptor (TBXA2R) in activating ERM (ezrin, radixin, and moesin) proteins to promote cell motility and invasion in triple-negative breast cancer (TNBC) cells. Using TBXA2R stimulation and a series of in vitro and in vivo experiments, the authors report that ERM activation is mediated through a TBXA2R signaling pathway involving Gαq/11 and Gα12/13 subunits, RhoA, and SLK/LOK kinases. They propose that this pathway enhances cell migration, invasion, and metastatic potential in TNBC.

    General criticisms

    Experimental design and analyses are adequate, even though certain experiments lack appropriate controls or employ the wrong statistical tests. However, the study primarily relies on a single TNBC cell line and heavy use of overexpression systems and/or small molecule inhibitors, raising concerns about the generalizability and specificity of the findings. Furthermore, several conclusions appear premature and unsupported by the current data. Critical controls and additional validation experiments are necessary to support the claims about the role of TBXA2R in metastasis and to justify the strong mechanistic conclusions drawn.

    Specific criticisms

    Figure 1

    TBXA2R expression should be shown to understand whether different ebBRET signals are dependent on the overexpression levels of TBXA2R.

    E-F: As ERM levels change over time, one would like to understand whether this is due to misloading or whether there is an underlying biological event going on in the stimulated cells. Are total ERM levels really changing over time? Please add a blot for 1-2 housekeeping proteins as loading controls. This is also crucial to clarify the kinetics of ERM activation; such notable intensity variations make quantifications of non-linear WB signals not fully reliable. In F, mean and SD should be plotted.

    G: The authors need to use a PM marker if they want to claim that pERM increases at the cell cortex. TBXA2R localization should also be shown.

    Figure 2

    A: This reviewer cannot see the purported partial inhibition in Ga12/13 KO cells. Are differences between the two KOs significant? Furthermore, there are reports indicating that YM-254890 may not be specific for Gaq. Experiments on double KO cells are needed to assess the possible redundancy between the two Ga subfamilies. C-D: it is important to add a positive control for the activity of Y-27632 in these experiments. Please show that a ROCK-dependent effect is inhibited in the treated cells. G: The working model is premature as it is unknown whether ROCKi was active. While asking for ROCK1/2 KO cells would be too much, this claim is far-fetched.

    Figure 3

    B: In the legend, it is not clear what grey and light read colours mark. E-F: This reviewer finds it difficult to believe that p-ERM and TBXA2R signal intensities at the cell cortex could be reliably quantified using IHC images. The representative samples would indicate that p-ERM and TBXA2R positivity are not correlated. It would be crucial to show examples for each of the TNBC subgroups the existence of which is inferred based on p-ERM and TBXA2R staining. The conclusion that "no TNBC samples exhibited high TBXA2R expression and low levels of p-ERMs, further supporting a role for TBXA2R signalling in ERM activation in TNBC" is an overstatement.

    Figure 4

    The authors wrote that "We focused on the Hs578T cell line, which showed a median level of TBXA2R mRNA expression among the six TNBC cell lines tested". I do not understand the rationale for it as anti-TBXA2R antibodies detecting endogenous TBXA2R are available and thus why not use the median protein levels?

    Figure 5

    Effects of the knockouts are subtle, and rescue experiments would be needed to corroborate these results. The employed statistical analysis is prone to overestimating differences. The authors should use the superplots instead. The authors might also decide to use other TNBC cell lines to explore the functional relevance of this pathway in BC progression. This is particularly important because Hs578T are poorly tumorigenic, and they often do not form palpable tumours in mice.

    Figure 6

    The fact that Hs578T are poorly tumorigenic in mice is likely the reason why the authors used the experimental metastasis model. However, it is puzzling that metastases were studied in the liver but not in the lungs. Furthermore, the whole approach is rather artefactual as the TBXA2R agonist was administered for the entire duration of these experiments. What is the pathological relevance of such a study? Including a spontaneous metastasis model or alternative TNBC lines that mimic human disease more closely would help strengthen the functional relevance of this pathway in BC progression and study's translational relevance.

    Figure S2

    B-M: the pERM signal appears to be perinuclear in some of the tested cell lines. Please use a PM marker.

    Figure S3

    The authors should use the superplots to analyse the cell migration data.

    Discussion

    The claim that "our findings demonstrated that kinases of the SLK family are the only kinases needed for ERM activation by TBXA2R" should be tuned down as only 2 cell lines were tested. In this section, the authors should also discuss the proposed pro-metastatic functions of TXA2 and TXA2R in more detail, including vascular permeability. The sweeping conclusion that "TBXA2R expression correlates with phosphorylation and activation of ERMs in TNBC patient samples" clashes with the authors' own results; please stick to the data.

    Concluding remarks

    This study investigates a signaling pathway whereby TBXA2R thorugh ERM activation enhances the migratory and invasive potential of TNBC cells. However, several improvements are needed to support the main claims. The dependence on a single TNBC cell line, reliance on pharmacological inhibitors with potential off-target effects, and limited in vivo relevance detract from the generalizability of the findings. Additional TNBC models, adeguate controls, and a broader focus on natural metastasis patterns would make the conclusions more compelling. Moderating certain overstated claims would be needed to align the interpretations with the actual data.

    Cross-commenting

    I found comments in the other reviewers' reports that align with my criticisms on the mouse experiments as well as with those pertaining to the tissue culture work.

    Significance

    General comments

    The manuscript investigates the role of TBXA2R in the regulation of ERM in the context of TNBC metastasis. Much of this TBXA2R signalling axis is already known, as well as that SLK and LOK can phosphorylate ERM in other cell systems. Similarly, the positive role of ERM in cell migration/invasion and cancer progression has long been reported. The somewhat unexpected finding that ERM phosphorylation is independent of ROCK remains not fully convincing. The BC-related part is problematic as the continuous administration a TBXA2R agonist is required for key tumour metrics to show some differences in vivo. This calls into question the main conclusion of the work, namely that the TBXA2R/ERM-dependent pathway is activated during BC progression in TNBC cells.

    Audience

    Specialists interested in GPCRs and signal transduction or in the cytoskeleton.

    Expertise

    Rev: cancer cell biology, signal transduction, cytoskeleton, actin biochemistry, multiplexed imaging, mouse model of human diseases.

    Co-rev: nanoparticles, cell biology.

  6. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    The authors do not wish to provide a response at this time.

  7. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #4

    Evidence, reproducibility and clarity

    Overall, the authors show an interesting and conclusive work on the activation of ERM proteins upon TBXA2R signaling. The use of the ebBRET biosensor to assess ERM-protein activation enables elegant investigation of activation modalities. The Thromboxane A2 analogue U46619 robustly shows activation of ERM proteins in ebBRET assays as well as an increase in ERM-protein phosphorylation status. The functional effects of this signaling pathway are shown convincingly for moesin, where moesin mediates an TBXA2R mediated increase in cell motility, invasion and metastasis of triple-negative breast cancer Hs578 cells in vitro and in vivo. Nonetheless, some points need to be clarified.

    Significance

    Comment 1: In the title the authors state, that ERM-activation via TBXA2R is controlling invasion and motility of triple-negative breast cancer cells. In the manuscript, there is only data supporting this assumption for moesin (MSN). Therefore, the authors need to change the title accordingly or support additional experiments for the other two ERM-proteins radixin and ezrin. Throughout the experiments, the p-ERM antibody is used to measure ERM-protein activation. Since the effects on invasion and motility observed in Hs578 cells are mainly mediated through moesin, it would be necessary to see, at least for one experiment per cell line (HEK293T, Hs578) the detailed phosphorylation status of ezrin, radixin and moesin separately. As there are specific, phospho-detecting antibodies for this case, this could be done rather easy. Furthermore, showing specific increase of phosphorylated moesin would support the functional data shown in Figure 5 and 6. To investigate the functional effect of TBXA2R mediated activation of ezrin and radixin on cell motility and invasion, similar experiments could be done in e.g. HMC-1-8 breast cancer cells (high ezrin expression) and HCC1187 (high radixin expression).

    Comment 2: Figure 1A, C, D: The concentration of staurosporine is with 100 nM relatively high for kinase inhibition. It would be informative to see the assay with increasing staurosporine concentrations, e.g. from 1 nM to 50 nM. In general, a concentration of 1-10 nM should be sufficient for kinase inhibition, preventing unspecific effects of the drug.

    Comment 3: The citation for the p-ERM antibody is confusing, as there is only p-Moe used in the cited paper (Roubinet, 2011). There is a p-ERM antibody commercially available (Cell Signaling, Phospho-ezrin (Thr567)/radixin (Thr564)/moesin (Thr558) Antibody #3141). Could you clarify which antibody you are using?

    Comment 4: From the inhibitor experiments using C3 transferase toxin (Figure 2), the authors conclude that RhoA plays a role in TBXA2R mediated ERM activation. As mentioned in the manufacturer's description, C3 toxin is inhibiting RhoA, RhoB and RhoC. Therefore, it would be necessary to repeat those experiments under RhoA knockdown conditions (e.g. using an siRNA-based approach) to state that specifically RhoA is involved.

    Comment 5: To assess, if the findings in Figure 5 and 6 are due to the higher moesin expression in Hs578 cells or are linked to a specific function of moesin, a re-expression experiment would be informative. To achieve this, the 2D and 3D migration experiments could be redone after re-expression of moesin, ezrin and radixin separately in moesin knockdown conditions.

    Minor comments:

    • Even though U46619 is a known Thromboxane A2 analogue, including negative and positive controls would strengthen the results. In detail, this could be done by showing a known protein which gets phosphorylated downstream of TBXA2R signaling and a protein which is not affected by this signaling pathway alongside the shown effects on ERM-proteins.
    • Figure 1 J: There are no statistics comparing the conditions of SQ-29548 treated cells in presence/absence of U46619, that should be added.
    • Figure 1 G, H: How was the quantification for cell periphery performed? In detail, how were the thresholds set for cell periphery / not cell periphery?
    • Figure 3 H:
      • The labelling indicating presence of U46619 is missing.
      • Also, what is the rationale behind normalizing MB-453 for 3 cell lines and comparing the BT-549 to MB-157?
    • Suppl. Fig 4 D: Define y-axis better. Absorbance at what wave length?
    • Define FERM and ERMAD abbreviations in introduction.
  8. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary

    The Ezrin, radixin, and moesin (ERM) family of proteins orchestrate morphological changes that potentiate metastatic invasion in cancer cells. In this study, Leguay et al. identify the GPCR, TBXA2R, as a key activator of the ERM proteins which promotes motility and invasion in triple-negative breast cancer (TNBC) cells. Using BRET-based sensors developed by them previously for monitoring the activation of ERM proteins and building upon their previous findings on the role of the small GTPase RhoA in the activation of ERM proteins, the authors carefully dissect the molecular pathway leading to the activation of ERM proteins upon stimulation of the TBX2AR. The authors also establish the pathological relevance of the pathway in TNBC using in vitro and in vivo models, opening up possibilities for targeting this pathway in cancer cells. Overall, the study is well-conceived and executed, and the results are clearly described and presented in the manuscript. However, the following comments must be addressed before publication.

    Major comments

    Fig 1C - Why p-ERM was normalized over Ezrin and not ERM? It would be more appropriate and consistent to normalize against the ERM signal as done in other experiments in the manuscript.

    Fig 1E and S3C - The levels of total ERM also seem to change with increasing treatment times. This must be clarified and discussed in the manuscript.

    Fig 1F - Why is the mean of all three independent experiments not presented here as in S3C?

    Fig 2E - Though SLK seems to play a dominant role in the phosphorylation of ERM in HEK293T cells, the depletion of LOK also substantially reduces the phosphorylation of ERM in the representative figure (Fig 2E), which is not reflected in the quantification (Fig 2F). Indeed, both SLK and LOK seem to be equally crucial in Hs578T cells (Fig 4I), unlike the conclusion here. The authors must check if the quantifications were affected by any white spots in the blot for total ERM as seen in the representative figure. If necessary, the authors must include additional replicates, and the model in Fig 2G should be updated accordingly. If the contributions of LOK are indeed quite minimal in HEK293T cells, then the difference in Hs578T cells must be adequately highlighted and discussed rather than broadly mentioning similar results were observed in both cell lines. The discussion mentions that SLK kinases are the only kinases needed for ERM activation, which conflicts with findings from Hs578T cells, where both SLK and LOK contribute to ERM phosphorylation (Fig 4I). The authors should revise this to reflect their data accurately.

    Minor comments

    FigS3B should cite the source dataset and not just the database. Also, details of how the extracted data was processed (if any) should be described clearly.

    When multiple treatments are involved (for, e.g. U46619 and staurosporine), the exact sequence of treatments and the overlap in timings of different treatments must be clearly mentioned. E.g. fig 1A and 1C. There are a few grammatical errors which need to be fixed. E.g. Paragraph 2 in the second section of results - We next aimed to identify (not identifying) which kinase(s) acts downstream of TBX2AR

    Significance

    Triple-negative breast cancer, which is characterized by a lack of estrogen, progesterone or HER2 receptors, is a highly metastatic and aggressive form of breast cancer with poor prognosis. Currently, there are fewer treatment options than other types of invasive breast cancer. The current study opens up the possibility of targeting the TBXA2R or the downstream signalling components in TNBC, which are still expressed in TNBC cells. However, certain TNBC sub-types express low levels of p-ERM and TBX2AR (Fig 3E, 3F), indicating a minor role for TBX2AR pathway and targeting this pathway in these subtypes may be inefficient. In addition, certain subtypes express high p-ERM and low TBX2AR indicating alternative pathways for ERM activation. Currently, it is not clear which other GPCRs can contribute to ERM activation by engaging similar downstream effectors. A comprehensive screening of different GPCR antagonists could identify alternative strategies to target the ERM-mediated metastasis in TNBC cells that show low expression of TBX2AR.

    Audience The manuscript is relevant to a broad audience, especially to cell biologists, cancer biologists and clinical scientists.

    The reviewer's field of expertise includes cell signaling, gene expression, and RNA biology in mammalian systems. Moderate expertise in cancer biology. Limited knowledge of histopathological analysis.

  9. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Leguay et al present an interesting and logical series studies that investigate the activity and signaling of the GPCR TBXA2R in TNBC cells. The premise of the overall study is that metastasis is often associated with a more invasive/motile cancer cell phenotype. The investigators have an interest in ERM (Ezrin, Radixin, Moesin) proteins, which have been implicated in cell motility. The authors link stimulation of TBXAR2, a GPCR, to activation of ERM proteins and also show that TBXAR2 is associated with worse outcome in TNBC patients. Through the use of genetic and pharmacologic tools the authors provide convincing biochemical and cell based data to support their model that stimulation of TBXAR2 activates Gα11 & Gα12/13 which subsequently stimulate RhoA and SLK/LOK which then phosphorylate ERMs. The authors show relevant biologic consequences of the pathway. Data include orthogonal assays with similar results and the manuscript is written clearly and the data are displayed well. Overall it is a solid story that is largely well done. There are a few comments that should be addressed.

    Comments:

    1. All the biochemical/cell based in vitro data exploit the use of small molecule agonists of TBXAR2, not the natural ligand. A comment on this and why use of TXA2 is not feasible would be helpful to the reader.
    2. The data in figures 1-5 are solid and clear. However, I suggest adding a higher magnification inset for the IHC images shown in Fig 3E. It would be useful to be able to distinguish cells in the IHC, a higher mag shot should suffice.
    3. A) The use of Hs578t cells for the in vivo modeling is unfortunate. Additionally, the use of iv injection to in a study focused on cell invasion is also unfortunate. The metastatic propensity of Hs578t is not clear, in fact a recent report comparing metastasis in breast cancer cell lines shows that Hs578t perform poorly in terms of metastasis after orthotopic injection (see PMID 38468326). I searched the literature a bit to try and find other examples of iv injection of Hs578t cells, I found 1 (PMID:27654855, I did not search exhaustively), this paper shows significant lung metastasis and does not mention liver metastases. Were other breast cancer cells investigated for the in vivo studies?

    B) Why I was interested is because the typical organ that is seeded post iv injection is the lungs (as seen in the above ref), liver metastases post iv injection are not common, especially with breast cancer cells. What did the lungs look like in your experiments?

    C) Further while the data presented in figure 6 are supportive of the overall conclusions, the data is modest at best in terms of metastatic burden. Repetition of the experiment using a breast cancer cell line injected orthotopically would likely be more useful in highlighting the importance of the pathway to metastasis.
    I understand performing an orthotopic assay may be outside the scope of the study, but it would provide greater impact given the focus of the paper on cell invasion.

    Cross-commenting

    I think reviewer comments are generally aligned. I was least critical but appreciate the concerns of the other reviewers, especially rev #1 who requested additional validation and controls. In my opinion in vivo studies are not robust, I expect that is due to cell line choice. Repetition of the in vivo study with a breast cancer cell line that is capable of metastasis (from a primary tumor) would be more effective.

    Significance

    The manuscript presents a solid, logical flow and the biochemical/cell based in vitro data are clean. Clear differences between groups, appropriate controls, and displayed effectively.

    The challenge is the in vivo study. IV injection of cancer cells is a valid model for seeding and growing in a target organ BUT it does not reflect cell invasion, which is typically thought of as a step that occurs earlier in the metastatic cascade. That said, the data are supportive with conclusions but not necessarily consistent with expected results based on iv injection of this cell line. A caveat is that the cell line used is characterized as having metastatic characteristics in vitro but is not a consistent metastatic line in vivo. The recommendation is the perform a new in vivo experiment. An orthotopic injection of a strongly metastatic cell line, such as MDA MB 231 or other (see paper ref aboved) would be a more stringent and accurate test of the importance of the pathway to cell invasion in vivo.

  10. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Summary

    This manuscript investigates the role of the thromboxane A2 receptor (TBXA2R) in activating ERM (ezrin, radixin, and moesin) proteins to promote cell motility and invasion in triple-negative breast cancer (TNBC) cells. Using TBXA2R stimulation and a series of in vitro and in vivo experiments, the authors report that ERM activation is mediated through a TBXA2R signaling pathway involving Gαq/11 and Gα12/13 subunits, RhoA, and SLK/LOK kinases. They propose that this pathway enhances cell migration, invasion, and metastatic potential in TNBC.

    General criticisms

    Experimental design and analyses are adequate, even though certain experiments lack appropriate controls or employ the wrong statistical tests. However, the study primarily relies on a single TNBC cell line and heavy use of overexpression systems and/or small molecule inhibitors, raising concerns about the generalizability and specificity of the findings. Furthermore, several conclusions appear premature and unsupported by the current data. Critical controls and additional validation experiments are necessary to support the claims about the role of TBXA2R in metastasis and to justify the strong mechanistic conclusions drawn.

    Specific criticisms

    Figure 1

    TBXA2R expression should be shown to understand whether different ebBRET signals are dependent on the overexpression levels of TBXA2R.

    E-F: As ERM levels change over time, one would like to understand whether this is due to misloading or whether there is an underlying biological event going on in the stimulated cells. Are total ERM levels really changing over time? Please add a blot for 1-2 housekeeping proteins as loading controls. This is also crucial to clarify the kinetics of ERM activation; such notable intensity variations make quantifications of non-linear WB signals not fully reliable. In F, mean and SD should be plotted.

    G: The authors need to use a PM marker if they want to claim that pERM increases at the cell cortex. TBXA2R localization should also be shown.

    Figure 2

    A: This reviewer cannot see the purported partial inhibition in Ga12/13 KO cells. Are differences between the two KOs significant? Furthermore, there are reports indicating that YM-254890 may not be specific for Gaq. Experiments on double KO cells are needed to assess the possible redundancy between the two Ga subfamilies. C-D: it is important to add a positive control for the activity of Y-27632 in these experiments. Please show that a ROCK-dependent effect is inhibited in the treated cells. G: The working model is premature as it is unknown whether ROCKi was active. While asking for ROCK1/2 KO cells would be too much, this claim is far-fetched.

    Figure 3

    B: In the legend, it is not clear what grey and light read colours mark. E-F: This reviewer finds it difficult to believe that p-ERM and TBXA2R signal intensities at the cell cortex could be reliably quantified using IHC images. The representative samples would indicate that p-ERM and TBXA2R positivity are not correlated. It would be crucial to show examples for each of the TNBC subgroups the existence of which is inferred based on p-ERM and TBXA2R staining. The conclusion that "no TNBC samples exhibited high TBXA2R expression and low levels of p-ERMs, further supporting a role for TBXA2R signalling in ERM activation in TNBC" is an overstatement.

    Figure 4

    The authors wrote that "We focused on the Hs578T cell line, which showed a median level of TBXA2R mRNA expression among the six TNBC cell lines tested". I do not understand the rationale for it as anti-TBXA2R antibodies detecting endogenous TBXA2R are available and thus why not use the median protein levels?

    Figure 5

    Effects of the knockouts are subtle, and rescue experiments would be needed to corroborate these results. The employed statistical analysis is prone to overestimating differences. The authors should use the superplots instead. The authors might also decide to use other TNBC cell lines to explore the functional relevance of this pathway in BC progression. This is particularly important because Hs578T are poorly tumorigenic, and they often do not form palpable tumours in mice.

    Figure 6

    The fact that Hs578T are poorly tumorigenic in mice is likely the reason why the authors used the experimental metastasis model. However, it is puzzling that metastases were studied in the liver but not in the lungs. Furthermore, the whole approach is rather artefactual as the TBXA2R agonist was administered for the entire duration of these experiments. What is the pathological relevance of such a study? Including a spontaneous metastasis model or alternative TNBC lines that mimic human disease more closely would help strengthen the functional relevance of this pathway in BC progression and study's translational relevance.

    Figure S2

    B-M: the pERM signal appears to be perinuclear in some of the tested cell lines. Please use a PM marker.

    Figure S3

    The authors should use the superplots to analyse the cell migration data.

    Discussion

    The claim that "our findings demonstrated that kinases of the SLK family are the only kinases needed for ERM activation by TBXA2R" should be tuned down as only 2 cell lines were tested. In this section, the authors should also discuss the proposed pro-metastatic functions of TXA2 and TXA2R in more detail, including vascular permeability. The sweeping conclusion that "TBXA2R expression correlates with phosphorylation and activation of ERMs in TNBC patient samples" clashes with the authors' own results; please stick to the data.

    Concluding remarks

    This study investigates a signaling pathway whereby TBXA2R thorugh ERM activation enhances the migratory and invasive potential of TNBC cells. However, several improvements are needed to support the main claims. The dependence on a single TNBC cell line, reliance on pharmacological inhibitors with potential off-target effects, and limited in vivo relevance detract from the generalizability of the findings. Additional TNBC models, adeguate controls, and a broader focus on natural metastasis patterns would make the conclusions more compelling. Moderating certain overstated claims would be needed to align the interpretations with the actual data.

    Cross-commenting

    I found comments in the other reviewers' reports that align with my criticisms on the mouse experiments as well as with those pertaining to the tissue culture work.

    Significance

    General comments

    The manuscript investigates the role of TBXA2R in the regulation of ERM in the context of TNBC metastasis. Much of this TBXA2R signalling axis is already known, as well as that SLK and LOK can phosphorylate ERM in other cell systems. Similarly, the positive role of ERM in cell migration/invasion and cancer progression has long been reported. The somewhat unexpected finding that ERM phosphorylation is independent of ROCK remains not fully convincing. The BC-related part is problematic as the continuous administration a TBXA2R agonist is required for key tumour metrics to show some differences in vivo. This calls into question the main conclusion of the work, namely that the TBXA2R/ERM-dependent pathway is activated during BC progression in TNBC cells.

    Audience

    Specialists interested in GPCRs and signal transduction or in the cytoskeleton.

    Expertise

    Rev: cancer cell biology, signal transduction, cytoskeleton, actin biochemistry, multiplexed imaging, mouse model of human diseases.

    Co-rev: nanoparticles, cell biology.

  11. Version published to 10.1101/2023.03.28.534587 on bioRxiv