Oncogenic PKA signaling increases c-MYC protein expression through multiple targetable mechanisms

  1. Gary KL Chan
  2. Samantha Maisel
  3. Yeonjoo C Hwang
  4. Bryan C Pascual
  5. Rebecca RB Wolber
  6. Phuong Vu
  7. Krushna C Patra
  8. Mehdi Bouhaddou
  9. Heidi L Kenerson
  10. Huat C Lim
  11. Donald Long
  12. Raymond S Yeung
  13. Praveen Sethupathy
  14. Danielle L Swaney
  15. Nevan J Krogan
  16. Rigney E Turnham
  17. Kimberly J Riehle
  18. John D Scott
  19. Nabeel Bardeesy
  20. John D Gordan

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    In this article, global kinome profiling using fibrolamellar carcinoma and melanoma cell line models was employed to identify key effectors of protein kinase A (PKA) oncogenic signaling, which is hyperactivated in these cancer types. Based on use of molecular and cellular biology assays, authors proposed a model whereby the oncogenic effects of PKA are at least in part mediated by Aurora Kinase A (AURKA)- and PIM2-dependent regulation of MYC family members, and provide evidence that cancers with constitutive activation of PKA may be sensitive to AURKA inhibitors. Overall, it was thought that this study is of broad interest inasmuch as it provides new insights into the molecular underpinnings of oncogenic PKA signaling, and suggests the potential of using AURKA inhibitors to target malignancies characterized by aberrant PKA activation. With stronger mechanistic data linking constitutive PKA signaling to activation of AURKA and PIM2 and MYC regulation and in vivo experiments to support the conclusions, this manuscript will be of interest to researchers in the fields of cancer research, therapeutics, signal transduction and molecular and cell biology.

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

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Genetic alterations that activate protein kinase A (PKA) are found in many tumor types. Yet, their downstream oncogenic signaling mechanisms are poorly understood. We used global phosphoproteomics and kinase activity profiling to map conserved signaling outputs driven by a range of genetic changes that activate PKA in human cancer. Two signaling networks were identified downstream of PKA: RAS/MAPK components and an Aurora Kinase A (AURKA)/glycogen synthase kinase (GSK3) sub-network with activity toward MYC oncoproteins. Findings were validated in two PKA-dependent cancer models: a novel, patient-derived fibrolamellar carcinoma (FLC) line that expresses a DNAJ-PKAc fusion and a PKA-addicted melanoma model with a mutant type I PKA regulatory subunit. We identify PKA signals that can influence both de novo translation and stability of the proto-oncogene c-MYC. However, the primary mechanism of PKA effects on MYC in our cell models was translation and could be blocked with the eIF4A inhibitor zotatifin. This compound dramatically reduced c-MYC expression and inhibited FLC cell line growth in vitro. Thus, targeting PKA effects on translation is a potential treatment strategy for FLC and other PKA-driven cancers.

Article activity feed

  1. Version published to 10.7554/elife.69521 on eLife
  2. eLife

    Author Response

    Reviewer #1 (Public Review):

    This study provides evidence for previously unknown relationship between oncogenic protein kinase A (PKA) signaling and MYC family members. Specifically, the authors have employed a combination of systems biology and biochemical assays to capture mediators of oncogenic PKA signaling in a fibrolamellar carcinoma and melanoma cell line. This lead to identification of Aurora A and PIM kinases as potential effectors of constitutively active PKA. Aurora A and PIM kinases have been previously shown to stabilize MYC proteins. Accordingly, evidence is provided that the effects of PKA/Aurora A and PKA/PIM axis are mediated via MYC. Collectively, these findings suggest a model whereby the effects of aberrant PKA signaling are mediated via Aurora A and PIM kinases and related feedback mechanisms that ultimately result in stabilization of MYC proteins. Importantly, PKA-driven cancer cell lines exhibited high sensitivity to Aurora A kinase inhibitors in cell culture-based assays. These findings not only provide pioneering insights into oncogenic PKA signaling, but may also have implications for developing therapeutic approaches for neoplasia that harbor constitutively active PKA.

    Strengths:

    This study addresses the role of aberrant PKA signaling in cancer, which represents a major gap in knowledge in cancer biology. Systems biology approaches and dissection of signaling networks downstream of constitutively active PKA are found to be exciting in the context of this study and likely to provide a wealth of information for future studies. Results from samples obtained from fibrolamellar carcinoma patients partially confirmed correlations observed in cell lines, which was seen as an advantage. Notwithstanding that, it was thought that orthogonal genetic validation may in some cases be warranted, pharmacological approaches using e.g. Aurora A inhibitors hold a promise for accelerated translation of observed findings into the clinic.

    We appreciate this positive assessment of our work and are hopeful that we have solidified the significance and potential impact of our findings through additional analysis.

    Weaknesses:

    The major drawback of the study is the lack of in vivo models to validate observations garnered from the cell lines. This is particularly important considering that experiments carried out in samples from fibrolamellar carcinoma patients suggested additional Aurora A and PIM kinase-independent mechanisms of PKA-driven increase in MYC levels and likely in neoplastic growth may be implicated in vivo. In addition, it was thought that more mechanistic evidence is required for linking PKA to PIM kinase, especially because different PIM kinases were implicated in stabilization of MYC in fibrolamellar carcinoma vs. melanoma cell lines. Finally, although pharmacological approaches were appreciated, due to potential issues with the specificity of the inhibitors, it was thought that orthogonal genetic approaches are warranted to further corroborate the proposed model.

    We acknowledge the lack of in vivo treatment modeling in this manuscript. The work presented here provides motivation for these important experiments, but they remain outside the scope of this manuscript. The expansion of the manuscript in revision with new investigations into protein translation and several additional data sets creates a more complete systems biology analysis of PKA signaling and PKA-induced signaling dependencies. This expanded scope makes in vivo validation of specific treatments and treatment combinations an even larger undertaking. The text has been modified to emphasize this point. We further acknowledge the accuracy of the reviewer’s assessment of our findings on PIM2. The limited reagents to study PIM kinases made this relatively difficult to expand. We shifted the focus of the work to include assessment of PKA effects on mRNA translation as a mechanism of c-MYC regulation. We have strengthened our assessments with loss- and gain-of-function genetic and pharmacological models, which we believe will more completely answer the reviewer’s concerns.

    Reviewer #2 (Public Review):

    Protein kinase A (PKA) is often stimulated and contributes to cancer growth, yet the downstream kinase signaling cascades remain unclear. Here the authors use a global phosphoproteomics and kinome activity profile to show that not only is the RAS/MAPK pathway activated, as expected, but the authors also suggest Aurora kinase A (AURKA) and PIM kinases are activated to stabilize the expression of MYC expression; a potent oncoprotein associated with poor prognosis and aggressive disease. The authors use a number of different cell lines in this study, but focus on fibrolamellar carcinoma as PKA is known to contribute to this disease.

    Strengths: It has been notoriously difficult to map kinases and their substrates as these protein-protein interactions are not always amenable to traditional biochemical techniques due to their labile nature, and kinase substrate consensus sites are often overlapping and not highly specific. Thus, the authors' pipeline to delineate such kinase cascades is quite novel and useful. They apply it here to determine PKA signaling in cancer using sophisticated computational strategies and then validate with classic molecular techniques.

    We appreciate this positive assessment of our analytical tools and the importance of understanding oncogenic PKA signaling.

    Weaknesses: The lack of mechanistic evidence linking aberrant PKA activation with regulation of MYC family members was considered to be a major weakness of the study. As it stands, it is hard to delineate whether observed changes in the levels of MYC family members are indeed a consequence of aberrant PKA signaling. It also remains unclear which MYC phosphorylation sites are implicated in the context of neoplastic PKA function and whether MYC family members are regulated at the level of protein stability or mRNA translation. Moreover, some methodological issues (e.g. using single siRNAs) were also observed. Collectively it was thought that these weaknesses should be addressed to corroborate author's conclusions.

    We acknowledge these concerns about our initially submitted manuscript and present extensive data that advances the manuscript in answering the key questions posed by the reviewer. We note that with the development of data showing PKA-induced phosphorylation of translation initiation components and sensitivity of c-MYC levels to eIF4A inhibition, some detailed evaluations of c-MYC phosphorylation were not undertaken, although key c-MYC mutants were tested in the course of our study and are included for reviewer interest.

  3. eLife

    Evaluation Summary:

    In this article, global kinome profiling using fibrolamellar carcinoma and melanoma cell line models was employed to identify key effectors of protein kinase A (PKA) oncogenic signaling, which is hyperactivated in these cancer types. Based on use of molecular and cellular biology assays, authors proposed a model whereby the oncogenic effects of PKA are at least in part mediated by Aurora Kinase A (AURKA)- and PIM2-dependent regulation of MYC family members, and provide evidence that cancers with constitutive activation of PKA may be sensitive to AURKA inhibitors. Overall, it was thought that this study is of broad interest inasmuch as it provides new insights into the molecular underpinnings of oncogenic PKA signaling, and suggests the potential of using AURKA inhibitors to target malignancies characterized by aberrant PKA activation. With stronger mechanistic data linking constitutive PKA signaling to activation of AURKA and PIM2 and MYC regulation and in vivo experiments to support the conclusions, this manuscript will be of interest to researchers in the fields of cancer research, therapeutics, signal transduction and molecular and cell biology.

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

  4. eLife

    Reviewer #1 (Public Review):

    This study provides evidence for previously unknown relationship between oncogenic protein kinase A (PKA) signaling and MYC family members. Specifically, the authors have employed a combination of systems biology and biochemical assays to capture mediators of oncogenic PKA signaling in a fibrolamellar carcinoma and melanoma cell line. This lead to identification of Aurora A and PIM kinases as potential effectors of constitutively active PKA. Aurora A and PIM kinases have been previously shown to stabilize MYC proteins. Accordingly, evidence is provided that the effects of PKA/Aurora A and PKA/PIM axis are mediated via MYC. Collectively, these findings suggest a model whereby the effects of aberrant PKA signaling are mediated via Aurora A and PIM kinases and related feedback mechanisms that ultimately result in stabilization of MYC proteins. Importantly, PKA-driven cancer cell lines exhibited high sensitivity to Aurora A kinase inhibitors in cell culture-based assays. These findings not only provide pioneering insights into oncogenic PKA signaling, but may also have implications for developing therapeutic approaches for neoplasia that harbor constitutively active PKA.

    Strengths:

    This study addresses the role of aberrant PKA signaling in cancer, which represents a major gap in knowledge in cancer biology. Systems biology approaches and dissection of signaling networks downstream of constitutively active PKA are found to be exciting in the context of this study and likely to provide a wealth of information for future studies. Results from samples obtained from fibrolamellar carcinoma patients partially confirmed correlations observed in cell lines, which was seen as an advantage. Notwithstanding that, it was thought that orthogonal genetic validation may in some cases be warranted, pharmacological approaches using e.g. Aurora A inhibitors hold a promise for accelerated translation of observed findings into the clinic.

    Weaknesses:

    The major drawback of the study is the lack of in vivo models to validate observations garnered from the cell lines. This is particularly important considering that experiments carried out in samples from fibrolamellar carcinoma patients suggested additional Aurora A and PIM kinase-independent mechanisms of PKA-driven increase in MYC levels and likely in neoplastic growth may be implicated in vivo. In addition, it was thought that more mechanistic evidence is required for linking PKA to PIM kinase, especially because different PIM kinases were implicated in stabilization of MYC in fibrolamellar carcinoma vs. melanoma cell lines. Finally, although pharmacological approaches were appreciated, due to potential issues with the specificity of the inhibitors, it was thought that orthogonal genetic approaches are warranted to further corroborate the proposed model.

  5. eLife

    Reviewer #2 (Public Review):

    Protein kinase A (PKA) is often stimulated and contributes to cancer growth, yet the downstream kinase signaling cascades remain unclear. Here the authors use a global phsophoproteomics and kinome activity profile to show that not only is the RAS/MAPK pathway activated, as expected, but the authors also suggest Aurora kinase A (AURKA) and PIM kinases are activated to stabilize the expression of MYC expression; a potent oncoprotein associated with poor prognosis and aggressive disease. The authors use a number of different cell lines in this study, but focus on fibrolamellar carcinoma as PKA is known to contribute to this disease.

    Strengths: It has been notoriously difficult to map kinases and their substrates as these protein-protein interactions are not always amenable to traditional biochemical techniques due to their labile nature, and kinase substrate consensus sites are often overlapping and not highly specific. Thus, the authors' pipeline to delineate such kinase cascades is quite novel and useful. They apply it here to determine PKA signaling in cancer using sophisticated computational strategies and then validate with classic molecular techniques.

    Weaknesses: The lack of mechanistic evidence linking aberrant PKA activation with regulation of MYC family members was considered to be a major weakness of the study. As it stands, it is hard to delineate whether observed changes in the levels of MYC family members are indeed a consequence of aberrant PKA signaling. It also remains unclear which MYC phosphorylation sites are implicated in the context of neoplastic PKA function and whether MYC family members are regulated at the level of protein stability or mRNA translation. Moreover, some methodological issues (e.g. using single siRNAs) were also observed. Collectively it was thought that these weaknesses should be addressed to corroborate author's conclusions.

  6. eLife

    Reviewer #3 (Public Review):

    The study by Chan, Gordan et al, utilizes state of the art global kinome profiling to map the shared signaling networks driven by diverse genetic changes resulting in PKA activation in human cancer. Among the many kinases whose activity is modulated by active PKA, they authors centered their study on Aurora Kinase A (AURKA), and its ability to regulate c-MYC and n-MYC protein levels. They propose an AURKA-MYC regulatory network, and a possible positive feedback loop mediated by the kinase PIM2, which can be disrupted by AURKA inhibition. The study has many elements of novelty, which could be of translational importance. The strength of the study is the use of global kinome profiling and the identification of multiple candidate signaling nodes downstream from PKA. The weakness is the limited mechanistic information provided on possible direct regulatory processes intervening in kinase activation, and the need to enhance the rigor of the studies and to provide quantitative analysis of the data to increase the confidence regarding the proposed novel mechanisms at this stage.

  7. Version published to 10.1101/2021.04.16.438110v1 on bioRxiv