Author Manuscript Published OnlineFirst on March 4, 2019; DOI: 10.1158/1541-7786.MCR-18-0407 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Somatic hypermutation of the YAP oncogene in a human cutaneous melanoma Xiaomeng Zhang1,*, Jian Zhong Tang1,2,*, Ismael A. Vergara1,*, Youfang Zhang1,3,4, Pacman Szeto1,3,4, Lie Yang1, Christopher Mintoff1, Andrew Colebatch1, Lachlan McIntosh1,5,11, Katrina A. Mitchell1, Evangeline Shaw1, Helen Rizos6,7, Georgina V. Long7, Nicholas Hayward7,8, Grant A. McArthur1,9, Anthony T. Papenfuss1,5,9,11, Kieran F. Harvey1,9,10,† and Mark Shackleton1,3,4,7,8,† 1Peter MacCallum Cancer Centre, Melbourne, Victoria, Australia 2Olivia Newton-John Cancer Research Institute, Melbourne, Victoria, Australia 3Central Clinical School, Monash University, Melbourne, Victoria, Australia 4Alfred Health, Melbourne, Victoria, Australia 5The Walter and Eliza Hall Institute, Melbourne, Victoria, Australia 6Faculty of Medicine and Health Sciences, Macquarie University, Sydney, NSW, Australia 7Melanoma Institute of Australia, Sydney, NSW, Australia 8QIMR Berghofer Medical Research Institute, Brisbane, QLD, Australia 9Sir Peter MacCallum Department of Oncology, 10Department of Pathology and 11Department of Mathematics and Statistics, University of Melbourne, Melbourne, Victoria, Australia *†These authors contributed equally Correspondence: Mark Shackleton, Alfred Health and Monash University, 55 Commercial Road, Melbourne, Vic 3004, Australia; e: [email protected], tel: +61 3 9076 2000 Running Title: YAP and melanoma Financial Support: The Cancer Council of Victoria, Australia; Grant #1080255 Conflict of Interest: None 1 Downloaded from mcr.aacrjournals.org on September 26, 2021. © 2019 American Association for Cancer Research. Author Manuscript Published OnlineFirst on March 4, 2019; DOI: 10.1158/1541-7786.MCR-18-0407 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. ABSTRACT Melanoma is usually driven by mutations in BRAF or NRAS that trigger hyperactivation of mitogen-activated protein kinase (MAPK) signalling. However, MAPK-targeted therapies are not sustainably effective in most patients. Accordingly, characterizing mechanisms that co-operatively drive melanoma progression is key to improving patient outcomes. One possible mechanism is the Hippo signalling pathway, which regulates cancer progression via its central oncoproteins YAP and TAZ, although is thought to be only rarely affected by direct mutation. As YAP hyperactivation occurs in uveal melanoma, we investigated this oncogene in cutaneous melanoma. YAP protein expression was elevated in most benign nevi and primary cutaneous melanomas but present at only very low levels in normal melanocytes. In patient-derived xenografts and melanoma cell lines, we observed variable reliance of cell viability on Hippo pathway signalling that was independent of TAZ activity and also of classical melanoma driver mutations such as BRAF and NRAS. Finally, in genotyping studies of melanoma, we observed the first ever hyperactivating YAP mutations in a human cancer, manifest as seven distinct missense point mutations that caused serine to alanine transpositions. Strikingly, these mutate four serine residues known to be targeted by the Hippo pathway and we show that they lead to hyperactivation of YAP. Implications: Our studies highlight the YAP oncoprotein as a potential therapeutic target in select sub-groups of melanoma patients, although successful treatment with anti-YAP therapies will depend on identification of biomarkers additional to YAP protein expression. KEYWORDS Melanoma, Hippo pathway, YAP 2 Downloaded from mcr.aacrjournals.org on September 26, 2021. © 2019 American Association for Cancer Research. Author Manuscript Published OnlineFirst on March 4, 2019; DOI: 10.1158/1541-7786.MCR-18-0407 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. INTRODUCTION Recent spectacular advances in treatment of melanoma, a common and deadly form of skin cancer, have leveraged discoveries of mechanisms of disease initiation and progression. A paradigm- changing example was the observation that most oncogenic melanoma-initiating mutations, classically in BRAF or NRAS, increase mitogen-activated protein kinase (MAPK) signalling (1). This led to BRAF- and MAPK kinase (MEK)-targeted therapies that dramatically changed treatment (2, 3). Unfortunately, although most BRAF-mutant melanomas respond to combination BRAF/MEK inhibition, resistance usually develops (4). For patients with non-BRAF-mutant disease, BRAF targeting is usually ineffective, although some respond to MEK targeting (5). Even T-cell checkpoint targeting with immunotherapy, increasingly and successfully applied to a wide variety of cancers including melanoma (6, 7), is not efficacious or sustained in many patients, highlighting the importance of identifying novel drivers of disease progression and therapy resistance. For instance, the Hippo signalling pathway was found to mediate resistance to MAPK pathway targeting in cancers such as melanoma (8-10), spurring efforts to develop Hippo-targeted therapies (11) that might overcome such resistance. The Hippo pathway, which is a critical regulator of organ size, was discovered in Drosophila screens for regulators of tissue growth (12). Subsequently shown to regulate growth in mice (13, 14), Hippo pathway proteins regulate organ size by modulating nuclear access of the transcription co-activators YAP and TAZ (also known as WWTR1). Hippo signalling is regulated by properties such as cell polarity and adhesion and has been linked to G-protein coupled receptor signalling. These signals are typically conveyed by a core kinase cassette, which is comprised of Sterile 20-like kinases (e.g. MST1 and MST2), NDR kinases (e.g. LATS1 and LATS2) and two sets of adaptor proteins (Salvador and Mob family proteins). LATS1/2 repress YAP and TAZ by phosphorylation of serine residues (at least 5 in YAP), which promotes binding to 14-3-3 proteins and cytoplasmic sequestration. YAP and TAZ regulate gene expression with transcription factors, particularly TEAD family proteins (15). Hippo proteins regulate many hallmarks of cancer and pathway activity is deregulated in many cancers (16). Despite this, mutations directly linked to alterations in Hippo pathway activity appear surprisingly uncommon in human cancer. Exceptions are the upstream pathway gene neurofibromin 2 (NF2), which is frequently mutated in mesothelioma (17) and meningioma (18), and the G protein genes GNAQ and GNA11, which collectively are mutated in 85% of uveal melanomas and hyperactivate YAP (19-21). Putative gain-of-function mutations of YAP are apparently very rare, with only one reported, a YAP1-TFE3 fusion in epithelioid hemangioendothelioma (22). In 3 Downloaded from mcr.aacrjournals.org on September 26, 2021. © 2019 American Association for Cancer Research. Author Manuscript Published OnlineFirst on March 4, 2019; DOI: 10.1158/1541-7786.MCR-18-0407 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. melanoma, Menzel and colleagues found copy number gains directly affecting YAP in 4 – 10 % of patients and, overall, 62% of melanomas had copy number alterations affecting known Hippo pathway genes (23). Unlike uveal melanoma, the role of Hippo signalling in the far more prevalent cutaneous melanoma is poorly defined, particularly in therapy-naïve contexts. Expression of YAP in cutaneous melanoma was described, with one study (23) but not another (24) reporting elevated expression with increasing disease stage. Nallet-Staub and colleagues noted that TAZ expression followed YAP expression, and that benign nevi displayed YAP/TAZ levels similar to melanomas (24) (although this was not reproduced (23)). In functional studies and consistent with other data (16), YAP, TAZ or TEAD were found to promote malignant behaviours in melanoma cell lines (23, 24) . Interestingly, whereas YAP and TAZ inhibited invasion in 1205Lu and SKMEL28 cells (24), heterogeneity in clonogenic growth amongst lines was noted upon TAZ knockdown. This raises the possibility that these oncoproteins might regulate malignant behaviors in only some cancers. Further, it is unknown whether YAP and TAZ, which are paralogous proteins, can compensate for each other’s inhibition. Both these possibilities have implications for development of inhibitors of YAP- or TAZ-mediated transcription. To address these issues, we used melanoma patient samples, patient-derived xenografts (PDX) and cell lines to investigate systematically the Hippo pathway in melanoma. We found Hippo pathway deregulation across a broad spectrum of melanocytic neoplasia, from benign nevi to melanomas. In molecular studies, we identified a patient melanoma with multiple YAP mutations that led to its hyperactivation. In PDX and cell lines, we found that melanoma cells rely on YAP for survival only variably, that this is independent of mutations in BRAF and NRAS, and that TAZ does not consistently compensate for YAP depletion in maintaining melanoma cell viability. We thus report the first obvious gain of function missense mutations in the YAP oncogene in human cancer and implicate the Hippo pathway as a potential therapeutic target in a subset of melanomas. MATERIALS AND METHODS Human tissues and cell lines. Human tissues were obtained from the Melbourne Melanoma Project, the Victorian Cancer