Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance

Jose L. Orgaz; Eva Crosas-Molist; Amine Sadok; Anna Perdrix-Rosell; Oscar Maiques; Irene Rodriguez-Hernandez; Jo Monger; Silvia Mele; Mirella Georgouli; Victoria Bridgeman; Panagiotis Karagiannis; Rebecca Lee; Pahini Pandya; Lena Boehme; Fredrik Wallberg; Chris Tape; Sophia N. Karagiannis; Ilaria Malanchi; Victoria Sanz-Moreno

doi:10.1016/j.ccell.2019.12.003

Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance

Orgaz, Jose L.;Crosas-Molist, Eva;Sadok, Amine;Perdrix-Rosell, Anna;Maiques, Oscar;Rodriguez-Hernandez, Irene;Monger, Jo;Mele, Silvia;Georgouli, Mirella;Bridgeman, Victoria;Karagiannis, Panagiotis;Lee, Rebecca;Pandya, Pahini;Boehme, Lena;Wallberg, Fredrik;Tape, Chris;Karagiannis, Sophia N.;Malanchi, Ilaria;Sanz-Moreno, Victoria; 2020-01-01 00:00:00 Article Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance Graphical Abstract Authors Jose L. Orgaz, Eva Crosas-Molist, Amine Sadok, ..., actin-Myosin dynamics Sophia N. Karagiannis, Ilaria Malanchi, Victoria Sanz-Moreno transcriptional re-wiring Correspondence cytoskeleton ROCK-Myosin II- remodeling addicted tumor j.orgaz@qmul.ac.uk (J.L.O.), MAPKi v.sanz-moreno@qmul.ac.uk (V.S.-M.) anti-PD-1 MAPKi MAPK Myosin II Cross-resistant ROCK/Myosin II In Brief phenotype inhibition Orgaz et al. show that myosin II activity ROCK/Myosin II ROCK/Myosin II increases during melanoma adaptation to PD-L1 MAPK pathway inhibition. ROCK-myosin PD-L1 ROS ROS II signaling supports survival of resistant p-H2A.X melanoma cells and promotes immunosuppression. ROCK inhibitors FOXP3 + improve the efﬁcacy of MAPK inhibitors CD206 Treg CD206 CD206 FOXP3+ Mφ Mφ Mφ Treg and immunotherapies in melanoma FOXP3+ FOXP3+ CD206+ CD206+ Treg Treg Mφ CD206+ Mφ models. CD206+ FOXP3 Mφ Mφ Treg Highlights d Therapy-resistant melanoma cells restore myosin II activity to increase survival d High myosin II activity identiﬁes targeted and immunotherapy-resistant melanomas d ROCK-myosin II inhibition increases ROS-DNA damage and decreases PD-L1 and Tregs d ROCK inhibition enhances efﬁcacy of MAPK inhibitors and immunotherapies Orgaz et al., 2020, Cancer Cell 37, 85–103 January 13, 2020 ª 2019 The Authors. Published by Elsevier Inc. https://doi.org/10.1016/j.ccell.2019.12.003 Cancer Cell Article Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance 1,2, 1,2,10 3,10 1,2,4,10 1,2,10 Jose L. Orgaz, * Eva Crosas-Molist, Amine Sadok, Anna Perdrix-Rosell, Oscar Maiques, 1,2,10 1 5 2 4 Irene Rodriguez-Hernandez, Jo Monger, Silvia Mele, Mirella Georgouli, Victoria Bridgeman, 5,6 7 2 2 8 9 Panagiotis Karagiannis, Rebecca Lee, Pahini Pandya, Lena Boehme, Fredrik Wallberg, Chris Tape, 5 4 1,2,11, Sophia N. Karagiannis, Ilaria Malanchi, and Victoria Sanz-Moreno * Barts Cancer Institute, Queen Mary University of London, John Vane Science Building, Charterhouse Square, London EC1M 6BQ, UK Randall Division of Cell and Molecular Biophysics, King’s College London, New Hunt’s House, Guy’s Campus, London SE1 1UL, UK Translational Cancer Discovery Team, Cancer Research UK Cancer Therapeutics Unit, The Institute of Cancer Research, 15 Cotswold Road, Sutton, London SM2 5NG, UK Tumour Host Interaction, The Francis Crick Institute, 1 Midland Road, London NW1 1AT, UK St. John’s Institute of Dermatology, King’s College London & NIHR Biomedical Research Centre at Guy’s and St. Thomas’s Hospitals and King’s College London, London SE1 9RT, UK Department of Oncology, Haematology and Stem Cell Transplantation, University Hospital of Hamburg Eppendorf, Hamburg 20246, Germany Molecular Oncology Group, Cancer Research UK Manchester Institute, Manchester M20 4BX, UK The Institute of Cancer Research, Chester Beatty Laboratories, 237 Fulham Road, London SW3 6JB, UK Cell Communication Lab, UCL Cancer Institute, 72 Huntley Street, London WC1E 6DD, UK These authors contributed equally Lead Contact *Correspondence: j.orgaz@qmul.ac.uk (J.L.O.), v.sanz-moreno@qmul.ac.uk (V.S.-M.) https://doi.org/10.1016/j.ccell.2019.12.003 SUMMARY Despite substantial clinical beneﬁt of targeted and immune checkpoint blockade-based therapies in mela- noma, resistance inevitably develops. We show cytoskeletal remodeling and changes in expression and ac- tivity of ROCK-myosin II pathway during acquisition of resistance to MAPK inhibitors. MAPK regulates myosin II activity, but after initial therapy response, drug-resistant clones restore myosin II activity to increase survival. High ROCK-myosin II activity correlates with aggressiveness, identifying targeted therapy- and immunotherapy-resistant melanomas. Survival of resistant cells is myosin II dependent, regardless of the therapy. ROCK-myosin II ablation speciﬁcally kills resistant cells via intrinsic lethal reactive oxygen species and unresolved DNA damage and limits extrinsic myeloid and lymphoid immunosuppression. Efﬁcacy of tar- geted therapies and immunotherapies can be improved by combination with ROCK inhibitors. INTRODUCTION inhibitors (BRAFi) development (Chapman et al., 2011; Flaherty et al., 2010; Zhang, 2015). Unfortunately, most patients had par- Malignant melanoma has very poor survival rates (Balch et al., tial responses and disease progressed due to acquired resis- 2009) despite being at the forefront of personalized medicine tance (Larkin et al., 2014; Robert et al., 2015; Zhang, 2015). (Lau et al., 2016). Mutant BRAF (V600) is the most common Often, patients with resistance develop more metastases (Wagle oncogene in melanoma (Davies et al., 2002), driving proliferation, et al., 2011) and 20% of BRAF mutant melanoma patients never survival, and tumor progression by hyper-activating MEK and respond to BRAFi due to intrinsic resistance (Zhang, 2015). V600E ERK kinases (Gray-Schopfer et al., 2007). This led to BRAF Most resistance mechanisms involve MAPK reactivation Signiﬁcance Resistance to therapies is a persistent problem in melanoma management. Here, we identify an adaptation strategy in response to either targeted therapies or immunotherapies. Under treatment, melanoma cells undergo cytoskeletal remod- eling and consequent activation of ROCK-myosin II pathway. Such adaptation process renders resistant melanoma cells vulnerable to ROCK-myosin II inhibition, which can be exploited therapeutically. Cancer Cell 37, 85–103, January 13, 2020 ª 2019 The Authors. Published by Elsevier Inc. 85 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). A B C **** - BRAFi MEKi ROCKi **** **** Cytoskeleton remodeling/Rho GTPase signaling **** **** 1.0 **** Transcription Axon growth 0.8 silencing Translation 0.6 Notch-EMT 20 EIF2 activity 0.4 Development Development 0.2 Slit-Robo signaling Notch pathway Cell cycle NF-κB modulation 0 DNA replication by Notch - - F-actin Hoechst p-MLC2 D E WM983A WM983B WM983A WM983B **** **** ROCKi - BRAFi ROCKi -- + - - + **** **** 1.0 1.0 -+ - - + - BRAFi V600E BRAF p-MLC2 0.8 0.8 WM983A 0.6 0.6 MLC2 0.4 0.4 p-ERK1/2 0.2 0.2 ERK2 WM983B 0 0 GAPDH -- V600E Q61L 690cl2 BRAF /Pten null D04 NRAS ns F G 8 hr 24 hr 48 hr **** -+ - + - + BRAFi ** **** **** * *** 1.0 *** **** p-MLC2 1.0 2.0 **** ** 0.8 0.8 **** MLC2 1.5 0.6 0.6 0.4 0.4 p-MLC2/MLC2 1.0 0.2 p-ERK/ERK 0.2 p-ERK1/2 0 long exposure 0.5 - G - A MEKi ERK2 024 48 hr BRAFi GAPDH A375 WM983A H I J A375 sens PLX/R sens PLX/R A375/PLX/R - BRAFi -+ - + -+ - + BRAFi ** p-MLC2 ** MLC2 p-ERK1/2 ERK2 GAPDH WM983B WM88 sens PLX/R sens PLX/R p-MLC2 - +- + - + - + BRAFi DAPI F-actin GFP-MLC2 p-MLC2 250 250 250 MLC2 200 200 200 150 150 150 p-ERK1/2 100 100 100 50 50 50 GFP WT TASA TDSD ERK2 0 0 0 010 20 30 010 20 30 40 010 20 30 40 50 GAPDH Distance from edge (μm) A375/PLX/R M229-PLX/R M238-PLX/R Cytoskeleton remodeling Cytoskeleton remodeling Cytoskeleton remodeling/Rho GTPase signaling Chemotaxis - LPA signaling via GPCRs Reg. of actin Chemotaxis - LPA signaling via GPCRs Reg. of cytoskeleton Cell adhesion/Histamine in CFTR folding cytoskeleton in oligodendrocytes Cell adhesion/Histamine in VEGF signaling cell barrier integrity maturation cell barrier integrity H. pylori infection on Reg. of actin Signal transduction epithelial cells motility Oxidative stress Reg. of cytoskeleton STK3/4 Hippo, cytoskeleton mTORC2 (NOX) in remyelination YAP/TAZ Lipid metabolism HSF-1/chaperone VEGF signaling Cell cycle - Development - Notch Signaling Insulin signaling Huntington’s disease Non-genomic action chr.condensation Immune response_IL-4 retinoic acid Canonical Notch signaling in CRC Signal transduction PTMs in Gamma-secretase Gastrin in cell growth Lipid metabolism Signal transduction BAFF-induced signaling in osteogenesis WNT and Notch cardiac myogenesis and proliferation Insulin signaling mTORC2 Figure 1. MAPK Regulates Myosin II Activity in Melanoma (A) The 10 most enriched pathways in A375 cells after MEKi treatment compared with untreated cells from phospho-proteome data. (B) p-MLC2 and F-actin confocal images of A375M2 cells on collagen I after treatment (BRAFi PLX4720, MEKi trametinib, ROCKi GSK269962A). Scale bar, 25 mm. (legend continued on next page) 86 Cancer Cell 37, 85–103, January 13, 2020 MEKi BRAFi ERKi BRAFi MEKi ROCKi BRAFi BRAFi ROCKi MEKi ROCKi BRAFi ROCKi GFP WT TASA TDSD Cell morphology on collagen I % survival vs vehicle Cell morphology on collagen I Fluorescence int. p-MLC2 F-actin Hoechst Cell morphology on collagen I Relative levels BRAFi/vehicle p-MLC2/cell area (Konieczkowski et al., 2018). Therefore, combination of a BRAFi 2014b; Sanz-Moreno et al., 2008, 2011), and aggressive amoe- with a MEK inhibitor (MEKi) was approved (Flaherty et al., 2012; boid invasion (Cantelli et al., 2015; Medjkane et al., 2009; Orgaz Larkin et al., 2014; Long et al., 2014b). However, despite the et al., 2014b; Sanz-Moreno et al., 2008, 2011). improved responses, most patients still relapse (Flaherty et al., In vivo, ROCK inhibition diminishes tumor growth and meta- 2012; Konieczkowski et al., 2018). static spread (Itoh et al., 1999; Kumper et al., 2016; Sadok Improved survival in patients with melanoma was reported et al., 2015). However, the role of ROCK-myosin II during resis- after immune checkpoint inhibitor treatment (anti-PD-1 and tance to current cancer therapies has not been comprehensively anti-CTLA-4) (Hodi et al., 2010; Larkin et al., 2015; Sharma investigated. Intriguingly, PAK contributes to MAPKi resistance et al., 2017). However, there are patients who do not respond (Lu et al., 2017) and Cdc42-PAK2-myosin II regulates amoeboid or relapse due to resistance (Sharma et al., 2017). Therefore, invasion (Calvo et al., 2011; Gadea et al., 2008). drug resistance is a persistent problem in melanoma manage- Given the activation of pro-invasive/metastasis pathways ment. Better understanding of the biological/biochemical during melanoma cross-resistance (Hugo et al., 2015, 2016), changes in resistant cells will help develop improved treatments. we sought to investigate the role of cytoskeletal remodeling in Given the overlap between migration and pro-survival path- therapy resistance. ways, drivers of resistance have been linked to metastatic ability (Alexander and Friedl, 2012). Importantly, cross-resistance to RESULTS MAPK inhibitors (MAPKi) (Hugo et al., 2015) and immune check- point inhibitors (Hugo et al., 2016) has been described, involving MAPK Regulates Myosin II Activity in Melanoma transcriptomic alterations on genes key for epithelial-to-mesen- To gain unbiased insight into molecular changes in melanoma chymal transition (EMT), metastasis/invasion, extracellular ma- cells after MAPKi, we analyzed the phosphoproteome of V600E trix (ECM) remodeling, hypoxia and angiogenesis (Hugo et al., BRAF A375 melanoma cells early (24 h) on MEKi 2015, 2016). (GSK1120212 trametinib and PD184352) treatment (Figure 1A; ROCK-myosin II pathway is a key regulator of invasive and Table S1). Using MetaCore Pathway enrichment analysis, we metastatic behavior (Cantelli et al., 2015; Medjkane et al., found that cytoskeletal remodeling and Rho GTPase signaling 2009; Orgaz et al., 2014b; Sanz-Moreno et al., 2008, 2011). are top processes changing early on treatment (Figure 1A; Ta- Non-muscle myosin II has contractile properties and is regu- bles S1 and S2). lated by the phosphorylation of its light and heavy chains (Vice- Because Rho GTPase regulates invasion via ROCK-myosin II nte-Manzanares et al., 2009). Myosin II-driven contractility activity and amoeboid behavior (Jaffe and Hall, 2005; Olson, relies on multiple kinases. Rho-kinase (ROCK) inactivates the 2008; Sadok et al., 2015; Sahai and Marshall, 2002; Sanz-Mor- myosin light chain 2 (MLC2) phosphatase, which leads to eno et al., 2008), we studied how MAPK inhibition affected increased phosphorylation of MLC2 (p-MLC2) and myosin II melanoma phenotypes on collagen I-recapitulating dermal envi- activity (Ito et al., 2004; Olson, 2008). MLC2 is directly phos- ronments (Cantelli et al., 2015; Orgaz et al., 2014b; Sanz-Moreno phorylated by ROCK and myosin light chain kinase (MLCK) et al., 2008, 2011). Treatment of highly metastatic, amoeboid (Vicente-Manzanares et al., 2009). ZIP kinase can also phos- A375M2 melanoma cells with BRAFi PLX4720 and MEKi trame- phorylate MLC2 directly and indirectly (Haystead, 2005). How- tinib for 24 h led to loss of rounded-amoeboid behavior (Fig- ever, long-term depletion of ROCK1/2 cannot be substituted by ure 1B). Inhibition of myosin II with ROCKi GSK269962A induced any other kinase for generating actomyosin contractility loss of circularity and a collapsed cytoskeleton (Figure 1B). (Kumper et al., 2016). Myosin II activity drives contractile forces Reduced myosin II activity (p-MLC2) was observed after BRAF, required for migration (Clark et al., 2000; Lammermann and MEK, or ROCK inhibition (Figure 1C). Similar results were Sixt, 2009; Sahai and Marshall, 2002; Sanz-Moreno et al., observed in other human and mouse melanoma cells and other V600E 2008, 2011; Vicente-Manzanares et al., 2009; Wolf et al., MAPKi, including BRAF (WM983A, WM983B, 4599) (Fig- V600E 2003), metastatic colonization (Cantelli et al., 2015; Clark ures 1D, 1E, S1A, and S1B), BRAF /Pten-null 690cl2 (Figures Q61L/R et al., 2000; Hall, 2012; Herraiz et al., 2015; Orgaz et al., 1F and S1C) and NRAS (D04, MM485) (Figures 1F, S1D, (C) Quantiﬁcation of cell morphology and p-MLC2 by immunoﬂuorescence from (B). Left, boxplot (n > 200 cells pooled from 3 experiments); right, mean ± SEM (n = 90 cells [dots] pooled from 3 experiments). (D) Images and quantiﬁcation of cell morphology on collagen I after treatment (BRAFi PLX4720, ROCKi GSK269962A) (n > 346 cells pooled from 2 experiments). Arrows show collapsed phenotype. Scale bar, 100 mm. (E) p-MLC2 and p-ERK1/2 immunoblots from (D). (F) Cell morphology on collagen I after treatment (690cl2, MEKi PD184352, BRAFi PLX4032, ERKi SCH772984, n = 50 cells; D04, MEKi GSK1120212, AZD6244, n = 125–150 cells). (G) p-MLC2 and p-ERK1/2 levels after PLX4720 treatment (n = 5, mean ± SEM). (H) Survival of A375 cells stably overexpressing wild type (WT), constitutively inactive TASA, or constitutively active TDSD MLC2 a after 5-day treatment with 0.1 mM PLX4720 (n = 4). Confocal images of GFP-MLC2. Scale bar, 50 mm. (I) p-MLC2 and p-ERK1/2 immunoblots after PLX4720 treatment. (J) p-MLC2 and F-actin confocal images (BRAFi PLX4720). Scale bar, 25 mm. Representative ﬂuorescence intensity line scans (dashed lines in image) below. (K) The 10 most enriched pathways in BRAFi-resistant A375/PLX/R (Girotti et al., 2013), M229-PLX/R, and M238-PLX/R cells (Titz et al., 2016) compared with parental cell lines from phospho-proteome data. (A–F, I, and J) 24 h treatment. (C, D, F, and H) Boxplots show median (center line); interquartile range (box); min-max (whiskers). p values by Kruskal-Wallis with Dunn’s correction (C, D, and F), one-way ANOVA with Tukey’s (H) or Benjamini, Krieger, and Yekutieli correction (G), *p < 0.05, **p < 0.01, ****p < 0.0001. See also Figure S1 and Tables S1 and S2. Cancer Cell 37, 85–103, January 13, 2020 87 Figure 2. ROCK-Myosin II Pathway Is Tran- A B Cytoskeleton-related genes scriptionally Rewired during Development Log fold change expression vs Parental cell line of Resistance -0.5 0.5 MAPK blockade V600E Resistant BRAF (A) Cell lines used for gene expression (Obenauf time weeks months single-drug resistant (SDR) Group 1 Group 2 melanoma et al., 2015; Song et al., 2017). cell lines double-drug resistant (DDR) (B) Heatmap ofunsupervised hierarchicalclustering 48 hr DTP DTPP of 313 cytoskeleton-related genes in A375 48 h drug-tolerant drug-tolerant persister proliferating persister BRAFi (Obenauf et al., 2015); M229-, M238-, SKMEL28-, M395p2-, M395p1-, and M249-de- ACTA2 MYLK MYLK Transcriptomics (RNAseq) Cytoskeleton-related genes MYL9 rivatives (Songetal.,2017). Foldchangeexpression ITGA11 ARHGDIG MYL9 MYL2 MYL3 ARHGAP28 Rho GTPases, GEFs, GAPs, effectors MYH2 ACTC1 RHOBTB1 in resistant versus parental lines is shown. Myosin, actin proteins, cross-linkers, MYL7 FMNL1 ARHGAP29 Cytoskeleton-membrane linkers, integrins RHOU ANK1 ITGA8 MYOF Transcription factors CDC42EP1 (C) Percentage of upregulated/downregulated ARHGAP23 TLN2 ARHGEF17 MYL6 DAPK3 DAPK3 MYH9 VCL MYO7B MYH9 (1.5-fold) cytoskeleton-genes versus parental line. DOCK2 FMN2 PAK1 ITGB4 ITGA10 ITGA9 CITED1 POU3F2 BRN2 MYOZ2 (D) Percentage of upregulated genes. Boxplot: MYOZ1 ARHGEF10L ARHGAP24 C ANK3 RHOB MYOT SPTBN5 median (center line); interquartile range (box); ITGA1 RHOJ ARHGEF6 MYO3B LIMK2 ARHGAP15 LIMK2 ARHGDIB CDC42EP5 ITGA2 min-max (whiskers). p value by unpaired t test RAC2 Group 1 Group 2 ARHGAP22 ITGA3 ITGA5 MYOD1 ARHGAP6 ACTBL2 with Welch’s correction, ****p < 0.0001. 70 MYH15 ITGB3 ARHGAP25 PAK3 SEPT3 up in Resistant ARHGEF37 ARHGAP40 60 MYO10 (E) Left, schematic pathway. Right, percentage of ARHGAP31 down in Resistant DOCK10 ANK2 ARHGAP42 SPTBN1 ITGA4 50 PDK4 ARHGAP4 group 1 cell lines with upregulation of indicated MYH14 MYH10 MYO16 ARHGAP19 ARHGAP32 MYO5A ARHGAP12 ARHGAP5 genes. ACTN2 ITGAX DOCK4 FMN1 30 SPTBN2 ITGA6 ITGB8 MYL10 MYO5B See also Table S3. MYO1D ARHGEF35 ARHGEF5 20 SEPT4 DOCK3 ARHGEF40 MYL4 ARHGAP9 CDC42EP4 CDC42EP3 10 ITGB6 RHOV ACTR3C PAK6 PPP1R16B MYLK2 ARHGEF4 DIAPH3 ACTN1 SEPT11 ITGB1BP2 NEDD9 CDC42EP2 CIT LIMK1 DIAPH2 LIMK1 ARHGAP18 PDK3 ARHGEF11 ARPC1B WASL MYH16 MYOM3 WASH2P (Figures 1H and S1F). MLC2 overex- ARHGAP30 WASH1 WASH3P WASH7P 48 hr DTP DTPP DDR SDR SDR M249 DDR MYO1E MYOC RHOC ARHGEF2 FMNL3 ITGB5 pression did not affect p-ERK (Fig- DOCK6 DOCK7 SEPT9 ARHGAP21 LMNA MYO7A RAC3 SEPT14 DIAPH1 ure S1F). Moreover, high myosin II activ- WASF1 PAK4 PXN RHOBTB2 MYO19 ITGB3BP LMNB1 ECT2 ARHGAP11A ity A375M2 cells were more resistant to ARHGAP11B LMNB2 40 WASF3 **** MYH1 MYH8 SPTB RHOQ DOCK9 ITGA7 ARHGEF16 BRAFi and MEKi compared with low met- ARHGEF19 ITGB2 30 MYH6 RHOH ARHGAP26 CALM2 ACTB MYL12B ACTG1 MYL12B astatic, low myosin II activity A375 cells ACTR3 ACTR2 ACTR10 20 ARPC5 CDC42BPB ACTR1A SEPT2 CDC42BPA ARHGAP10 (Figures S1G and S1H). Similar results CDC42 ITGB1BP1 PDK1 HIF1A 10 DOCK11 RND3 ARHGAP44 DOCK5 ITGB1 TLN1 were observed using the pair WM983B MKL1 ACTA1 MKL1 SEPT5 MYL1 0 MYOG ARHGEF15 RHOBTB3 CFL2 1 2 Group ARHGAP33 (metastatic, high myosin II, and amoe- ARHGEF18 PPP1R12C Sep-08 ARHGAP1 ARHGAP17 PDK2 MKL2 ILK MKL2 SRF boid) versus WM983A (primary tumor, RHOD RND2 DOCK8 ARHGEF3 MYO15B MYO18B MYOM1 MYOM2 MYH7 ARHGEF33 low myosin II, and elongated) (Figures ARHGEF26 MYO1G ARHGEF9 ARHGAP20 MYO9A ROCK2 FMNL2 ROCK2 MYL5 MYOZ3 S1G and S1H). These data show that SEPT12 ITGAD E WAS ITGAM CDC42BPG ARHGEF25 MYL12A MYO6 MYL12A ARHGAP27 myosin II activity confers a survival ARHGAP39 PTK2B MYH13 MYH4 ROCK1/2 MYO9B MYO3A SPTBN4 RHOG ACTR1B 100 PAK2 advantage to BRAFi and could accel- RHOA WASF2 RAC1 ARHGEF1 MYL6B ARPC3 LIMK1/2 80 CFL1 P P EZR RND1 erate the onset of resistance. Accord- ITGA2B myosin ACTR3B CALM3 ZIPK 60 ARPC1A complex ARHGDIA ARPC5L ITGAE ACTR5 actin P ACTR6 ingly, restored or increased p-MLC2 MHC2 MLC2 ARPC2 40 CALM1 ARPC4 dynamics ITGB7 RHOF MYLK3 CALML4 MYO1A 20 ROCK1 ARHGEF38 was seen in several BRAFi-resistant Actomyosin MYO1H ROCK1 ACTN3 MYO1F MYO15A contractility MYLK4 0 MYH7B WASH5P MYO5C MRTF-A/B MYH3 compared with parental cell lines (Fig- MLCK PPP1R16A RHOT2 SEPT1 PPP1R12A MYO18A MYO1C ARHGEF10 PPP1R12B RDX ure 1I). MEKi did not affect p-MLC2 in SPTAN1 ARHGEF12 SEPT10 MYO1B ITGAV ACTN4 MLC2 MHC2 MLCK ZIPK MRTF MSN SEPT6 MYH11 ARHGEF7 resistant cells (Figure S1I), suggesting DOCK1 ARHGAP35 RHOT1 NF1 ACTR8 PTK2 FAK ITGAL SEPT7 STAT3 that MAPK-independent mechanisms may underlie p-MLC2 restoration. Impor- tantly, cortical p-MLC2 was delocalized after 24-h BRAFi treatment in A375 cells and restored in and S1E) cell lines. These data conﬁrm that myosin II is regulated by MAPK in melanoma. BRAFi-resistant cells (Figure 1J). Phosphoproteomic analysis of Restoration of ERK levels is observed during acquisition of several BRAFi-resistant melanoma cells compared with parental resistance to MAPKi (Konieczkowski et al., 2018; Lito et al., lines showed that cytoskeletal remodeling and Rho GTPase 2012; Obenauf et al., 2015). Twenty-four hours after BRAFi signaling were top enriched processes (Figure 1K; Table S2). reduced p-ERK was accompanied by reduced p-MLC2 (Fig- These data show that MAPK signaling regulates cytoskeletal ure 1G). However, 48 h after BRAFi treatment, p-MLC2 was myosin II and amoeboid behavior. During early responses to restored concomitantly with very modest increase in p-ERK (Fig- treatment, overexpression of myosin II allows melanoma cells ure 1G). These data show that, early after treatment, cells to survive, independently of MAPK activity. remodel their cytoskeleton to recover myosin II activity, resulting in uncoupling of MAPK signaling from actomyosin. ROCK-Myosin II Pathway Is Transcriptionally Rewired We next hypothesized that, under therapy, myosin II during Development of Resistance could play a role in survival of cells with reduced MAPK activity. Transcriptomic alterations drive resistance to MAPK-targeted Strikingly, overexpression of a phosphomimetic MLC2 (TDSD) therapy (Hugo et al., 2015). Transcriptomic data of melanoma (Takaki et al., 2017) increased survival of A375 cells under BRAFi cells at different stages of MAPKi resistance (Figure 2A): 48 h 88 Cancer Cell 37, 85–103, January 13, 2020 MYL9 MYL12A MYL12B MYH9 ROCK2 LIMK2 MYLK DAPK3 MKL1 MKL2 % cytoskeleton-related genes A375 M238 M229 M238 M229 % genes upregulated in Resistant M238 M229 M229 % cell lines with upregulation SKMEL28 M238 M229 SKMEL28 M263 M395p2 M395p1 M397 R4 R5 M238 DTPP M238 SDR SKMEL28 SDR SKMEL28 DDR M229 DDR M229 SDR M263 SDR A375 48 hr BRAFi M238 48 hr BRAFi M238 DTP M229 2d BRAFi M229 DTP M229 DTPP M395p2 SDR M395p1 SDR M397 SDR M249 DDR4 M249 DDR5 BRAFi BRAFi+MEKi A B C “High Myosin II activity” genes A375 A375/PLX/R Log fold change mRNA 2 (Sanz-Moreno 2011) sensitive resistant Resistant/sensitive -+ -+ BRAFi SKMEL28-SDR SKMEL28-DDR -1.0 1.0 - G - G - G - G ROCKi p<0.001 p=0.049 p-MLC2 MYL9 MYL12A MLC myosins MLC2 MYL12B MHC MYH9 p-ERK1/2 ROCK1 ROCKs ERK2 ROCK2 LIMK1 BRAFi-resistance MAPKi-resistance LIMKs GAPDH LIMK2 MKL1 100 33 59 59 100 59 97 35 % MLC2 activity MRTFs MKL2 100 100 12 14 100 100 70 70 % ERK activity vemurafenib 3 months D E F BRAFi sensitive BRAFi resistant (intrinsic) Patient #35 cell line A375 A375/PLX/R WM983A WM983B WM88 SKMEL5 LOX-IMVI WM793B - + - + - + - + - + - + - + BRAFi BRAFi - - G - G ROCKi - G - G - G - G - G - G - G - G - G - G - G - G ROCKi - - p-MLC2 p-MLC2 ROCKi MLC2 MLC2 p-ERK1/2 p-ERK1/2 BRAFi ERK2 ERK2 ROCKi + GAPDH GAPDH BRAFi % MLC2 activity 100 26 39 13 100 20 27 15 100 20 44 16 100 36 68 17 100 50 99 44 100 13 104 17 % MLC2 activity 100 28 99 29 100 100 8 9 100 100 7 9 100 99 10 12 100 100 25 26 100 100 58 60 100 100 30 25 % ERK activity % ERK activity 100 100 65 66 sensitive IC IC 50 50 Loewe BRAFi-Resistant 5.0 A375 0.5 μM WM983A 1.6 μM Antagonism Synergy 4.0 A375/PLX/R 0.06 μM WM983A/PLX/R 0.09 μM 2.0 1.5 1.0 0.5 25 0 0 0.05 0.05 0.1 0.1 0.0001 0.001 0.01 0.1 1 10 100 0.5 0.001 0.01 0.1 1 10 100 0.5 1 1 BRAFi (μM) 5 5 ROCKi (μM) ROCKi (μM) J **** K L A375/ **** **** Patient #35 PLX/R **** **** **** ** **** **** **** ** **** **** - ROCKi BRAFi ROCKi + BRAFi 100 **** 100 **** ROCKi - G - G - G - G - GG - ROCKi - B - B Myosin II inh. - + - + BRAFi - + BRAFi A375 A375/PLX/R BRAFi A375/PLX/R Patient #35 M N **** ns *** ** ** **** **** *** ** ** **** **** 2.0 ** ** 1.0 100 100 100 1.5 80 80 80 0.8 60 60 60 0.6 1.0 40 40 40 40 0.4 0.5 20 20 20 20 0.2 0 0 0 - ++ - rat MLC2 WT - ctrl ROCK1/2 ctrl siRNA ctrl ROCK1/2 ctrl siRNA WT TASA rat MLC2 ctrl MLC2 human siRNA MLC2 human siRNA ROCK1: 61% ROCK1: 66% % mRNA ROCK2: 69% 84% 88% 88% ROCK2: 69% 87% 85% 92% KD vs ctrl Figure 3. Survival of Targeted Therapy-Resistant Melanomas Is Dependent on ROCK-Driven Myosin II Activity (A) Fold change in mRNA levels of ROCK-myosin II pathway genes in A375/PLX/R, Colo829/PLX/R by qRT-PCR (n = 3); and from published RNA sequencing data (Song et al., 2017). (legend continued on next page) Cancer Cell 37, 85–103, January 13, 2020 89 MYL9 MYL12B MYH9 MYL9 MYL12B MYH9 A375/PLX/R A375 Relative melanoma A375/PLX/R Relative melanoma cell survival cell survival Colo829/PLX/R M229 SDR M238 SDR SKMEL28 SDR M229 DDR SKMEL28 DDR Relative melanoma cell survival IC ROCKi (μM) A375 Fold change WM983A in % dead cells (PI ) WM983B Relative melanoma cell survival WM88 % Control Relative melanoma cell survival (Obenauf et al., 2015; Song et al., 2017) or several weeks after 3C and S2A). P-MLC2 in resistant cells was ROCK dependent, treatment (drug-tolerant persisters [DTP], drug-tolerant prolifer- since several unrelated ROCKi (GSK269962A, H1152) (Feng ating persisters [DTPP]) (Song et al., 2017); and resistant cells et al., 2016) reduced p-MLC2 (Figures 3C and S2A). However, after months-years (single-drug resistant [SDR, BRAFi], dou- p-ERK was not affected by ROCKi. ble-drug resistant [DDR, BRAFi + MEKi]) (Song et al., 2017) Sensitive A375 cells lost circularity and became more spindle- were used to analyze changes in 313 manually curated cytoskel- shaped with long, thin protrusions after BRAF inhibition, with etal-related genes (Table S3). Unsupervised hierarchical clus- reduced p-MLC2 (Figures 1B–1G, 3D, and S2B). In contrast, tering classiﬁed melanoma cell lines into two groups (Figure 2B). A375/PLX/R cells did not change morphology after BRAFi treat- Group 1 clustered the majority of cell lines, including 48-h BRAFi ment, while ROCKi decreased their circularity and promoted a (when p-MLC2 was restored [Figure 1G]), DTP, DTPP, and SDR/ collapsed (Sadok et al., 2015) cytoskeleton (Figures 3D DDR stages, which had a signiﬁcant percentage of regulated and S2B). genes (1.5-fold up- or downregulated) compared with baseline/ We expanded these observations to PLX4720-resistant sensitive cell-speciﬁcally upregulated genes (Figures 2C and Colo829 (Figure S2C) and a panel of cell lines sensitive or intrin- 2D). Upregulated in group 1 were genes involved in generation/ sically resistant to BRAFi (Baenke et al., 2015; Konieczkowski maintenance of myosin II-driven contractility (Figure 2E), such et al., 2014)(Figures 3E, S2D, and S2E). Similar results were as myosin (MLC2 genes MYL9, MYL12A/B; and myosin heavy observed in A375 cells resistant to BRAFi dabrafenib + MEKi tra- chain 2 [MYH9]), ROCK2, MLCK (MYLK), ZIPK (DAPK3), metinib (Flaherty et al., 2012; Long et al., 2014b) (A375/D + T/R) LIMK2, and transcriptional co-activator MRTF (MKL1/2), which (Figures S2F and S2G); and in a resistant cell line established directly regulates MLC2 expression (Medjkane et al., 2009). Of from a patient with acquired resistance to BRAFi (patient no. note, myosin II activity promotes myosin II expression to self- 35) (Figures 3F and S2H). perpetuate (Calvo et al., 2013). Because therapy-resistant cells maintain high p-MLC2 These data show that group 1 melanomas adapt to therapy by (Figure 1I) and that myosin II increases survival under therapy rewiring their transcriptome to alter cytoskeletal gene expres- (Figure 1H), we assessed if myosin II could play a role in confer- sion, ultimately restoring myosin II activity. ring a survival advantage to therapy-resistant cells. Reduced p-MLC2 after ROCKi impaired survival of sensitive and BRAFi- Survival of Targeted Therapy-Resistant Melanomas Is resistant melanoma pairs (A375, WM983A, WM983B, WM88) Dependent on ROCK-Driven Myosin II Activity (Figures 3G, 3H, and S3A–S3C). BRAFi-resistant melanomas We next investigated if the ROCK-myosin II pathway could play a were 4- to 30-fold more sensitive to ROCKi GSK269962A (Fig- role in the survival of melanoma cells. Using qRT-PCR, we ures 3G, 3H, S3A, and S3C) and AT13148 (Figure S3A). Moder- conﬁrmed that MLC2 (MYL9, MYL12A/B) and other components ate synergistic effects between ROCKi and BRAFi were of the ROCK-MLC2 pathway (MYH9, ROCK1/2, LIMK, MKL1/2, observed in BRAFi-sensitive A375 cells (Figures 3I and S3D). MYLK) were increased at the mRNA level in BRAFi-resistant cell More pronounced synergy was observed by annexin V/propi- line pairs (A375 and Colo829 cells, Figure 3A). Similar results dium iodide (PI) cell death staining (Figure S3E). were obtained using publicly available data from M229, M238, Importantly, A375/PLX/R cells grown on collagen I had and SKMEL28 cells (Song et al., 2017)(Figure 3A). Gene set increased sensitivity to ROCKi (Figure 3J). We observed enrichment analysis (GSEA) showed that resistant cell lines dis- impaired survival after ROCKi treatment in several models of played similar transcriptomes to cells with high myosin II activity drug resistance: A375/PLX/R, A375/D + T/R, Colo829 and (Figure 3B). BRAFi-intrinsic resistant lines (Figure S3F); and patient no. 35 We compared the impact of MAPK inhibition on myosin II in cells (Figures 3K and S3G). Importantly, the survival advantage sensitive/resistant melanoma cells. P-MLC2 was decreased af- was provided by myosin II itself, since myosin II inhibitor blebbis- ter BRAFi treatment in sensitive but not in resistant A375/PLX/ tatin strongly suppressed survival (Figures 3L and S3H). More- R cells. P-ERK was reduced by BRAFi in sensitive cells (Figures over, siRNA targeting ROCK1/2, MYL9, MYL12B,or MYH9 (B) GSEA comparing high myosin II activity signature (Sanz-Moreno et al., 2011) to resistant cell lines (Song et al., 2017). Nominal p values shown, false discovery rate (FDR) < 0.2. (C) p-MLC2 and p-ERK1/2 immunoblots after 24 h treatment. (D) Images of cells from (C). Scale bar, 50 mm. (E and F) p-MLC2 and p-ERK1/2 immunoblots of sensitive and intrinsically resistant cells (E); and patient no. 35 cells (F) after 24 h treatment (8 h for WM88). Vertical line in diagram (F): cell line establishment. (G) Survival and half maximal inhibitory concentration (IC ) values after a 3-day treatment (n = 3). (H) IC values for GSK269962A. (I) Cell survival as synergy graph of A375 cells treated for 3 days (n = 4). (J) Images and quantiﬁcation of cell survival on collagen I for 9 days (n = 3). Scale bar, 100 mm. (K) Survival of patient no. 35 cells after 10 days (n = 3). (L) Survival after a 5- to 10-day blebbistatin and PLX4720 treatment (n = 3). (M) Survival 8 days after gene depletion by RNAi (n = 3; n = 4 A375/PLX/R myosin genes, patient no. 35 MYL12B, ROCK1/2). mRNA KD (percentage decrease versus control) by RT-PCR shown. (N) Cell death in A375/PLX/R cells 3 days after transient MLC2 KD and rescue with rat MLC2 WT or TASA (n = 3, left graph; n = 4, right graph). (C–K) ROCKi GSK269962A, BRAFi PLX4720. Graphs show mean ± SEM and individual data points (circle). p values by one-way ANOVA with Tukey’s (J, K, and N) or Dunnett’s correction (M, myosin genes); t test with Welch’s correction (L and M, ROCK), **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also Figures S2 and S3. 90 Cancer Cell 37, 85–103, January 13, 2020 A B 1.0 0.8 Patient #3 post 15 C/A 16 C/A Patient 2 Post-RAF/MEKi Pt17-DP avg Pt6-DP avg Patient #5 post 0.6 Patient 5 Post-RAF/MEKi Pt4-DP1 Pt19-DD-DP avg Pt7-DP1 Pt2-DP avg Pt20-DP avg 0.4 Pt8-DP avg Pt3-DP avg Pt15-DP avg Patient #2 post low expression Pt9-DP avg Patient 4 Post-RAF/MEKi 0.2 22 C/A high expression Pt10-DP avg Pt5-DP avg Pt22-DD-DP avg p=0.0226 Pt16-DP1 7 D/A Pt18-DD-DP1 Patient 1 Post-RAF/MEKi 24 D/A Pt24-DD-DP1 0 2 4 6 8 10 years Pt1-DP1 Pt23-DD-DP1 Patient 3 Post-RAFi Patient 6 Post-RAF/MEKi 25 C/A -0.3 0.3 Log fold change Resistant/baseline * D Pt30 Pt4 Pt18 Pt3 Pt44 Pt10 Pt77 Pt27 Pt28 Pt8 Pt62 100 Pt31 Pt5 Pt84 Pt94 Pt11 Pt89 Pt9 Resp NR -0.3 0.3 Log fold change On-anti-PD-1/baseline anti-PD-1 z-score Hugo 2016 Riaz 2017 Sanz-Moreno 2011 Waggle 2014 Hugo 2015 anti-PD-1 anti-PD-1 Myosin II MAPKi MAPKi -2.0 2.0 high low Resistant Baseline Resistant Baseline NR Resp On anti-PD-1 Baseline EMT/metastasis angiogenesis hypoxia wound healing TGF-β STAT3 NF-κB YAP pro-invasive/metastasis F G “High Myosin II activity” genes “High Myosin II activity” genes evading (Sanz-Moreno 2011) IL-8 (Sanz-Moreno 2011) intratumoural p=0.027 p<0.001 sustaining immunity cross- proliferative M2 macrophages resistance Met T cell markers signalling hallmarks T cell activation YAP evading apoptosis MAPKi-resistance anti-PD-1-resistance H IJ K Pre Post Pre Post Pre Post Pre Post Post vs Pre Post vs Pre Post vs Pre Post vs Pre p=0.0062 p=0.035 p=0.014 p=0.047 1.0 0.8 60 60 MAPKi immunotherapy (IT) 0.6 sequential MAPKi + IT 40 0.4 20 20 0.2 0 0 Pre Post Pre Post Pre Post Pre Post Figure 4. High Myosin II Levels Identify Therapy-Resistant Melanomas in Human Samples (A) Heatmap of fold change in expression of ROCK-myosin II pathway genes in MAPKi-resistant versus baseline patient samples from (Hugo et al., 2015; Kwong et al., 2015; Sun et al., 2014; Wagle et al., 2014). (B) Kaplan-Meier overall survival from The Cancer Genome Atlas according to expression of ROCK-myosin II genes (listed in A) (n = 389 melanoma patients). (legend continued on next page) Cancer Cell 37, 85–103, January 13, 2020 91 MYLK MYL9 LIMK2 LIMK1 MYH9 MYL12A MYL12B DAPK3 MKL1 ROCK1 ROCK2 MKL2 MYL9 MYLK LIMK1 MYL12A MYL12B LIMK2 DAPK3 MKL1 MYH9 ROCK1 ROCK2 MKL2 S100 % highest p-MLC2 staining MLC2 (MYL9) FPKM S100 (RNAseq) % Masson’s trichrome area + 2 CD206 cells/mm Overall Survival + + Ratio FOXP3 /CD4 reduced survival in A375/PLX/R and patient no. 35 cells We have previously generated a transcriptional signature for (Figure 3M). The decrease in survival after MLC2 knockdown amoeboid metastatic melanoma cells harboring high ROCK- (KD) was more pronounced in BRAFi-resistant cells (Figure S3I). driven myosin II activity (Cantelli et al., 2015; Sanz-Moreno Therefore, both MLC2 expression and phosphorylation by ROCK et al., 2011). We compared high myosin II signature; MAPK-tar- are required to promote survival of resistant cells. Importantly, geted therapy-resistant signatures (Hugo et al., 2015; Sun et al., RNAi-insensitive rat MLC2 (Calvo et al., 2013) overexpression 2014; Wagle et al., 2014); anti-PD-1/NR signature (Hugo et al., rescued the decreased survival observed after MLC2 depletion. 2016); and on-anti-PD-1-treatment signature (Riaz et al., 2017). This mechanism relied on MLC2 phosphorylation, since rescue Single sample GSEA (ssGSEA) showed that similar gene was impaired by TASA-MLC2 inactive phospho-mutant (Figures signatures are enriched in high myosin II amoeboid cells and 3N and S3J). therapy-resistant patient samples (Figure 4E), including EMT/ Overall, myosin II restoration confers a survival advantage to metastasis, angiogenesis, hypoxia, wound healing, transforming resistant melanomas. growth factor b (TGF-b)-, STAT3-, nuclear factor kB-, and YAP- signaling genes (Table S5). High Myosin II Levels Identify Cross-Resistant Global GSEA analysis showed a signiﬁcant overlap between Melanomas in Human Samples ‘‘high myosin II activity’’ melanoma cells (Cantelli et al., 2015; We next validated our ﬁndings in clinical samples from published Sanz-Moreno et al., 2011) and MAPKi-resistant melanomas datasets (Hugo et al., 2015; Kakavand et al., 2017; Kwong et al., with immunosuppressive macrophages, and pro-invasive and 2015; Long et al., 2014a; Rizos et al., 2014; Song et al., 2017; Sun pro-survival features (Hugo et al., 2015)(Figure 4F). There was signiﬁcant overlap between high myosin II and anti-PD-1/NR et al., 2014; Wagle et al., 2014)(Table S4). There was a subset of melanoma tumors (50%) with upregulation of ROCK-myosin II patient signatures (IPRES [Hugo et al., 2016]) (Figure 4G). pathway genes (Figures 4A, S4A, and S4B), in accordance with Myosin II-driven contractility is regulated by MLC2 gene data with resistant cell lines (Figure 2E). The Cancer Genome expression and phosphorylation/activity (Calvo et al., 2013; Atlas data showed that higher levels of ROCK-myosin II genes Medjkane et al., 2009; Olson, 2008). We assessed p-MLC2 levels in treatment-naive melanoma patients confer worse prognosis in paired patient melanoma sections before and after therapy (Figure 4B). MAPKi-resistant tumors quickly progress after (targeted therapy, immunotherapy [IT], or sequential targeted relapse (Wagle et al., 2011), indicative of aggressiveness. We and IT; Table S6). P-MLC2 levels were higher in all resistant tu- suggest that melanomas with intrinsically higher expression of mors after treatment (Figures 4H and S4D–S4G). Speciﬁcity of the ROCK-myosin II pathway are more aggressive and prone p-MLC2 antibody was validated by RNAi (Figure S4D). Collagen to develop resistance. density promotes myosin II activity (Laklai et al., 2016; Paszek Innately anti-PD-1-resistant (IPRES) tumors harbor a tran- et al., 2005), and ROCK-myosin II induces ECM stiffening scriptional signature of upregulated genes involved in the (Samuel et al., 2011). Increased ECM deposition was observed regulation of EMT, cell adhesion, ECM remodeling, angiogen- in resistant compared with pre-treatment samples (Figures 4I esis, and hypoxia (Hugo et al., 2016). MAPK-targeted and S4E–S4G). Melanoma cells with high ROCK-myosin II are therapies in melanoma induce similar signatures with immuno- highly secretory and polarize macrophages to tumor-promoting suppressive features (Hugo et al., 2015). These studies (CD206 ) phenotypes (Georgouli et al., 2019). Interestingly, suggest that non-genomic MAPKi resistance driven by tran- CD206 cells were increased in resistant compared with pre- scriptional upregulation of metastasis-related pathways medi- treatment samples (Figures 4J and S4E–S4G), correlating with ates cross-resistance to anti-PD-1 therapy. They also suggest higher p-MLC2 (Figure 4H). Immunosuppressive FOXP3 regula- that aggressive tumors resistant to one therapy (e.g., MAPKi) tory T cells (Tregs)/CD4 ratio was also increased in resistant will likely not respond to second therapy (anti-PD-1). There- samples (Figures 4K and S4F–S4G). These data suggest that fore, we next investigated if ROCK-myosin II could predict high MLC2 (MYL9) expression and/or activation (p-MLC2) in anti-PD-1 responses as part of a cross-resistance mechanism. melanoma cells together with immunosuppressive populations Samples before anti-PD-1 treatment (Hugo et al., 2016) and higher collagen densities identify therapy-resistant mela- showed higher MYL9 expression in non-responding (NR) nomas, suggesting their potential as biomarkers. than in responding (Resp) patients (Figure 4C). Increased Overall, resistant tumors and melanomas with high myosin II levels of ROCK-myosin II pathway genes were detected in a activity harbor a similar transcriptome. Importantly, ROCK- large subset of patients on anti-PD-1 treatment (Riaz et al., myosin II could be a key mediator of non-genomic cross- 2017)(Figures 4Dand S4C). resistance. (C) MYL9 mRNA in Resp (n = 15) and NR (n = 13) anti-PD-1 patients from (Hugo et al., 2016). Boxplot: median (center line); interquartile range (box); min-max (whiskers). (D) Heatmap of fold change in expression of ROCK-myosin II genes in on-anti-PD-1 versus baseline patient samples (Riaz et al., 2017). (E) Heatmaps show ssGSEA of cross-resistance gene signatures (NR, non-responder; Resp, responder). (F and G) GSEA comparing ‘‘high myosin II activity’’ signature (Sanz-Moreno et al., 2011) to a subset of MAPKi-resistant patient samples from (Hugo et al., 2015) (F) or anti-PD-1/NR samples (Hugo et al., 2016) (G). Chart pie in (F) with cross-resistance hallmarks from (Hugo et al., 2015). Nominal p values shown, FDR < 0.001 (F) and 0.145 (G). (H–K) Images (patient no. 17) and quantiﬁcation in 12 paired samples before and after therapies (including those in Figures S4E and S4F) of: p-MLC2 (% cells with + + highest score), melanoma marker S100 (inset) (H); Masson’s trichrome staining (percentage stained area/section) (I); CD206 cells (J); FOXP3 cells (K). Scale bars, 100 mm. p values by Mann-Whitney test (C, H–K). See also Figure S4 and Tables S4, S5, and S6. 92 Cancer Cell 37, 85–103, January 13, 2020 V600E anti-PD-1 BRAF V600E A B C 2x/week BRAF anti-PD-1/NR Patient #26 +ROCKi in vitro cell lines days 7 7 #26 #26/R x2 anti-PD-1 **** C57BL/6J **** 5555 4434 #26 #26/R **** **** anti-PD-1/ anti-PD-1/ before treatment anti-PD-1/Resistant **** **** Resp NR Resp NR * *** #26 #26/R **** **** **** 75 **** - G H - G H ROCKi 100 p-MLC2 80 80 MLC2 60 60 p-ERK1/2 40 40 ERK2 - G H - G H ROCKi 20 20 GAPDH 0 0 - G - G ROCKi - G - G ROCKi **** D E F V600E BRAF K601E 100 WT WT Patient #58 Ipi/R BRAF / NRAS BRAF vemurafenib ipilimumab pembrolizumab DTIC ipilimumab no therapy ipilimumab DTIC 3 months 3 months 1 month - ROCKi 1 month 4 months 1 month 1 month 3 months Patient #62T3 cell line Patient #33 cell line Patient #58 cell line 25 - + BRAFi - G - G ROCKi - G ROCKi - G ROCKi - ROCKi p-MLC2 p-MLC2 p-MLC2 **** Patient #33 Ipi/R MLC2 MLC2 MLC2 - ROCKi p-ERK1/2 p-ERK1/2 p-ERK1/2 ERK2 ERK2 ERK2 GAPDH GAPDH GAPDH 100 41 % MLC2 activity 100 33 % MLC2 activity % MLC2 activity 0 100 12 96 13 100 100 % ERK activity 100 100 % ERK activity - ROCKi 100 100 80 77 % ERK activity V600E BRAF G H vemurafenib ipilimumab I **** **** 5 months 2 months *** *** **** *** ns ns Patient #2 cell line **** *** 1.0 - ROCKi - + BRAFi 0.8 - G - G ROCKi 0.6 p-MLC2 0.4 MLC2 0.2 p-ERK1/2 ERK2 - GG - ROCKi - GG - ROCKi - + BRAFi - + BRAFi GAPDH 100 28 134 23 % MLC2 activity % ERK activity 100 100 82 86 J K L Patient #2 BRAFi+Ipi/R Patient #62T3 BRAFi+Ipi+Pembro/R Patient - ROCKi **** **** #2 #62T3 **** **** **** **** ** ** **** **** 100 100 100 100 80 80 80 80 60 60 60 40 40 40 **** 20 20 20 20 **** **** 100 0 0 0 0 ctrl ROCK1/2 ctrl siRNA ctrl ROCK1/2 ctrl siRNA - B - B Myosin II inh. 50 ROCK1: 83% ROCK1: 66% BRAFi % mRNA ROCK2: 77% 84% 97% 92% ROCK2: 59% 79% 89% 73% KD vs ctrl - G - G ROCKi BRAFi - + Figure 5. ROCK-Driven Myosin II Activity in Immunotherapy-Resistant Melanoma (A) Top, schematic of cell lines. Bottom, p-MLC2 immunoblots after treatment: n = 7 (G); n = 3 (H). (B) Survival of patient no. 26 cells treated for 10 days (n = 4). (legend continued on next page) Cancer Cell 37, 85–103, January 13, 2020 93 MYL9 MYL12B MYH9 MYL9 MYL12B MYH9 Relative spheroid- BRAFi - forming ability Relative melanoma cell survival Relative melanoma cell survival Relative melanoma cell survival Relative melanoma cell survival Relative melanoma cell survival Relative melanoma cell survival on collagen I BRAFi - Cell morphology on collagen I ROCK-Driven Myosin II Activity in Immunotherapy- no. 62T3, BRAFi did not affect p-MLC2, while ROCKi decreased Resistant Melanoma p-MLC2 in patient no. 2 cells (Figures 5H and S5I). Patient no. 2 Next we investigated whether survival of immunotherapy-resis- cells on collagen I displayed very rounded morphology even in tant melanomas could be dependent on ROCK-myosin II. To the presence of BRAFi, indicative of high p-MLC2 (Figure 5I). test this hypothesis in vitro, we used patient no. 26-derived cells ROCKi decreased circularity and induced very thin protrusions established pre- and post-anti-PD-1 resistance (Figure 5A). Both and a spindle-shaped morphology in patient no. 2 cells. A com- cell lines rely on ROCK to sustain p-MLC2 (Figures 5A and S5A). mon event during melanoma resistance is BRAFi/MEKi addic- Importantly, anti-PD-1/resistant cells were 2-fold more sensitive tion, which occurs when resistant melanomas become drug to ROCKi (Figure 5B). Increased sensitivity was further conﬁrmed dependent (Das Thakur et al., 2013; Hong et al., 2017; Kong in a resistant brain metastasis-derived cell line from patient no. et al., 2017; Moriceau et al., 2015; Sun et al., 2014). Patient no. V600E 26 (data not shown). We then grafted mouse Braf melanoma 2 cells displayed addiction to BRAFi on 2D cultures (Figure S5J), cell lines 5555 and 4434 cells (Dhomen et al., 2009) subcutane- but treatment with ROCKi impaired survival in the presence of ously onto fully immunocompetent C57BL/6J mice and treated BRAFi and further decreased survival upon BRAFi withdrawal with anti-PD-1, which led to variable responses. We isolated (Figure S5J). This agrees with data on BRAFi-resistant patient NR and Resp tumors and grew them ex vivo (Figures S5B and no. 35 and Colo829/PLX/R cells (Figures 3K, S3F, and S3G), S5C). Increased intrinsic sensitivity to ROCKi in vitro was found which also displayed varying degrees of BRAFi addiction. Inter- in anti-PD-1/NR-derived cells (Figure 5C), similar to the resistant estingly, patient no. 2 cells grew as compact spheroids on human cell lines (Figure 5B). As melanoma cells activate an im- collagen I under BRAFi treatment, but growth was abrogated mune-evasion program they also trigger cytoskeletal remodel- by ROCKi (Figures 5J and S5K), showing that myosin II drives ing, rendering them intrinsically vulnerable to ROCK-myosin II survival in BRAFi-addicted cells. Accordingly, myosin II inhibition inhibition. with blebbistatin or RNAi against ROCK or myosin II genes Using additional cell lines established from human melanomas impaired survival of patient no. 2 and no. 62T3 cells (Figures resistant to immunotherapy (patients no. 58 and no. 33), we 5K, 5L, and S5L). conﬁrmed that these melanomas harbored ROCK-dependent MRTF controls MLC2 expression (Medjkane et al., 2009) while p-MLC2 levels (Figures 5D and S5D). Cell survival was impaired MRTF activity is regulated by actin dynamics (Posern and Treis- after treatment with several ROCKi on 3D (Figures 5E and S5E) man, 2006). Expression of MRTF (MKL) was increased in resis- and 2D culture (Figure S5F). tant melanomas (Figures 2E and 4A) and its depletion impaired Our data predict that cells that do not respond to MAPKi––if BRAFi-resistant cell survival (Figure S5M). Accordingly, MYL9 they undergo cross-resistant transcriptional rewiring of their mRNA levels decreased after MRTF depletion (Figure S5M). cytoskeleton––they will not respond to immunotherapy either. Overall, melanomas with acquired and primary resistance to Such cross-resistance will be susceptible now to ROCKi. Patient targeted and immunotherapies rely on myosin II activity for their no. 62T3 cell line was established from a tumor with acquired survival. Consistently, p-MLC2 levels and cancer cell survival resistance to BRAFi and developed primary resistance to anti- were positively correlated in resistant lines (Figure S5N). CTLA-4 and anti-PD-1 (Figure 5F). After BRAF inhibition, p-MLC2 was not affected in these cells, while ROCK inhibition ROCK-Myosin II Inhibition Induces Lethal Reactive decreased p-MLC2 (Figures 5F and S5G). Similar to our previous Oxygen Species, DNA Damage, and Cell-Cycle Arrest data, survival of patient no. 62T3 cells was impaired with ROCKi We next investigated why resistant cells rely on myosin II for sur- (Figures 5G and S5H). vival. Resistant cells (Song et al., 2017) were enriched in oxida- Similarly, patient no. 2 cells were established from a tumor that tive stress and reactive oxygen species (ROS) metabolism never responded to targeted and immunotherapy (Figure 5H). gene signatures (Figure 6A) and had lower DNA damage repair The post-treatment-resistant biopsy had higher p-MLC2 gene expression (Figure 6B). Interestingly, ROCK-myosin II sup- compared with baseline tumor (Figure S4F). Similar to patient presses high ROS in migrating cells (Herraiz et al., 2015). ROCKi (C) Top, schematic of experiment. Bottom, survival of 4434 and 5555 anti-PD-1/non-responder (NR) lines versus responder (Resp) after a 3-day treatment (n = 3, 5555; n = 4, 4434). (D) p-MLC2 and p-ERK1/2 immunoblots of patient no. 58 (n = 4) and no. 33 (n = 3) cells after treatment. (E) Images and quantiﬁcation of cell survival on collagen I for 7 days (n = 3). (F) p-MLC2 and p-ERK1/2 immunoblots of patient no. 62T3 cells after treatment (n = 3). (G) Survival of patient no. 62T3 cells after a 10-day treatment (n = 3). (H) p-MLC2 and p-ERK1/2 immunoblots of patient no. 2 cells after treatment (n = 5). (I) Cell morphology of patient no. 2 cells on collagen I after treatment. n = 70 cells (dots) from 2 experiments. Scale bar, 50 mm. (J) Survival of patient no. 2 cells as spheroid-forming ability on collagen I for 16 days (n = 3); Scale bar, 100 mm. (K) Survival after a 10-day blebbistatin treatment (n = 3). (L) Survival 8 days after gene depletion by RNAi (n = 3; n = 4 patient no. 2 MYL12B-ROCK1/2, no. 62T3 MYL9; n = 5 no. 62T3 MYL12B). Average percentage mRNA KD (percentage decrease versus control) by qRT-PCR is shown. Vertical line in (D, F, and H): cell line establishment. (A–J) ROCKi GSK269962A, H1152; (F–K) BRAFi PLX4720. (A, D, F, H, and I) 24 h treatment. Graphs show mean ± SEM and individual data points (circle) except boxplot in I (median, center line; interquartile range, box; min-max, whiskers). p values by one-way ANOVA with Tukey’s (B, C, G, and J) or Dunnett’s correction (L, myosin genes); Kruskal-Wallis with Dunn’s correction (I), t test with Welch’s correction (E, K, and L, ROCK); **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also Figure S5. 94 Cancer Cell 37, 85–103, January 13, 2020 A B Role of BRCA1 in DNA Damage Response Cell Cycle Control of Chromosomal Replication GO_Positive_Regulation_Of_Reactive_Oxygen_Species_Metabolic_Process Aryl Hydrocarbon Receptor Signaling GO_Response_To_Reactive_Oxygen_Species DNA Double-Strand Break Repair by Homologous Recombination GO_Regulation_Of_Reactive_Oxygen_Species_Metabolic_Process DNA Double-Strand Break Repair by Non-Homologous End Joining ATM Signaling GO_Response_To_Oxidative_Stress NER Pathway Houstis_ROS BER pathway Response_To_Oxidative_Stress p53 Signaling GO_Positive_Regulation_Of_Reactive_Oxygen_Species_Biosynthetic_Process All-trans-decaprenyl Diphosphate Biosynthesis 02 468 10 -Log(p value) -Log(p value) C D **** ** * *** ** *** * ** *** **** 100 100 3 * multinucleated sR 75 G2/M - G - G ROCKi 2 50 p−H2A.X G1 dead GAPDH 25 25 0 0 0 0 sR sR s R sR sR sR sR sR - G - G - G - G ROCKi - 0.05 0.1 0.25 - 0.05 0.1 0.5 μM ROCKi sR sR Patient cells A375/ WM EF G PLX/R 793B #35 #2 #58 ** ** ns ns ** ** ** * **** **** 1.0 **** **** 1.0 **** ** **** ** 0.8 0.8 0.6 multinucleated 0.6 0.4 G2/M 0.2 0.4 0.2 G1 - G ROCKi dead Mcl-1 0 - G - G - GG - - G ROCKi - GG - - BB + GAPDH p-STAT3 - ++ - BRAFi STAT3 GAPDH H *** BRAFi ns ** 80 - G - G ROCKi ** *** 5 5 5 5 10 3.5% 4.6% 10 5.1% 16.9% 10 4.1% 3.2% 10 5.7% 21.3% 4 4 4 4 10 10 10 10 3 3 3 3 40 10 10 10 10 90.6% 1.3% 59.3% 18.8% 90.7% 2.1% 48.9% 23.9% 2 2 2 2 10 10 10 10 0 0 0 0 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 010 10 10 10 010 10 10 10 010 10 10 10 010 10 10 10 - GG - ROCKi Annexin V BRAFi -+ Patient #2 BRAFi+Ipi/R Patient #62T3 BRAFi+Ipi+Pembro/R ** **** ns ns *** 60 *** **** 20 20 -GG - ROCKi - GG - ROCKi -+ BRAFi - + BRAFi Figure 6. ROCK-Myosin II Inhibition Induces Lethal ROS, DNA Damage, and Cell-Cycle Arrest (A) GSEA of ROS/oxidative stress-related gene signatures in MAPKi-resistant versus sensitive cell lines (group 1) from (Song et al., 2017). Dashed line indicates statistical signiﬁcance. (B) The 10 most enriched canonical pathways in downregulated genes in MAPKi-resistant cell lines (group 1) from (Song et al., 2017). (C) Left, ROS levels in A375 (s) and A375/PLX/R (R) cells after treatment (n = 6). Right, quantiﬁcation of p-H2A.X immunoblots (n = 7). (legend continued on next page) Cancer Cell 37, 85–103, January 13, 2020 95 Relative ROS levels (mean fluorescence int.) % cells % dead cells % dead cells Relative p−H2AX levels Propidium iodide Relative p-STAT3/STAT3 levels % dead cells % cells Relative Mcl-1 protein levels induced higher levels of ROS (Figures 6C and S6A) and phos- High myosin II activity provides an advantage during early sur- phorylated H2A.X (p-H2A.X), indicative of DNA damage (Fig- vival in the lung, which is a limiting step in the metastatic process ure 6C), in BRAFi-resistant cells compared with sensitive cells. (Cantelli et al., 2015; Medjkane et al., 2009; Orgaz et al., 2014b; Resistant cells had lower expression of genes of the base exci- Sanz-Moreno et al., 2008, 2011). Many of the cross-resistance sion repair pathway (Figure S6B) that repairs ROS-mediated gene signatures were related to metastatic programs (Figure 4E). DNA damage (Krokan and Bjoras, 2013). Survival of patient no. 2 cells in the lung after tail vein injection Because BRAFi-resistant cells harbor higher ROS and have was improved after pre-treatment in vitro with BRAFi (Figure 7D). lost DNA damage repair machinery, ROCKi increases ROS However, when pre-treated with ROCKi, survival was impaired levels leading to unrepaired DNA damage. Unrepaired DNA (Figure 7D). Patient no. 35 BRAFi-addicted cell line (Figure 3K) damage can induce cell-cycle arrest that, if prolonged, can showed reduced growth and p-MLC2 levels in vivo after ROCKi lead to cell death (Shaltiel et al., 2015). Blocking myosin II ac- (Figures 7E and S7C). tivity using ROCKi resulted in a pronounced dose-dependent High myosin II activity cells (Cantelli et al., 2015; Sanz-Moreno cell-cycle arrest in BRAFi-resistant melanomas (Figures 6D et al., 2011) and MAPKi-resistant melanomas with immunosup- and S6C). Blebbistatin caused very similar results in resistant pressive features and pro-tumorigenic macrophages (Hugo cells (Figure 6E). As a result of ROS-DNA damage, resistant et al., 2015) display transcriptional overlap (Figure 4F). We as- cells suffer G2-M arrest and multinucleation. Accordingly, sessed myosin II activity and immunosuppressive populations time-lapse video microscopy showed that cells suffering cell- in A375/PLX/R xenografts (Figure 7A). ROCKi-treated tumors cycle arrest died after 72 h (Figure S6D). had reduced p-MLC2 (Figure 7F) and lower number of CD206 ROS production is counterbalanced by STAT3 (Poli and Cam- macrophages (Figure 7G), which could contribute to reduced poreale, 2015) and both high myosin II activity and resistant cells tumor growth. ROCKi decreased polarization to CD206 macro- harbor high STAT3 signaling (Figure 4E). ROCKi decreased phages as F4/80 content was not affected (Figure S7D), only + + p-STAT3 levels and its pro-survival target Mcl-1 in both targeted CD206 /F4/80 ratio (Figure 7G). ROCK-myosin II inhibition therapy- and immunotherapy-resistant cells (Figures 6F and 6G). could overcome cross-resistance to targeted/immunotherapies Moreover, we measured decreased survival in A375/PLX/R cells via intrinsic cell survival and extrinsic myeloid co-option. after 72 h of ROCKi treatment using 3-(4,5-dimethylthiazol-2-yl)- 2,5-diphenyl tetrazolium bromide assay (Figure S6E). Annexin ROCK-Myosin II Inhibition Improves Efﬁcacy of Immune V/PI staining (Figure S6F) showed increased cell death after Checkpoint Inhibitors ROCKi treatment in A375/PLX/R (Figures 6H and S6G), patient As high myosin II identiﬁes anti-PD-1/NR, we tested whether no. 2 (Figures 6I and S6H), no. 62T3 (Figure 6I), and no. 35 cells ROCKi could be given as combination therapy to improve (Figure S6I). response to anti-PD-1. We allografted treatment-naive 5555 Therefore, ROCK-driven myosin II protects tumor cells from cells into immunocompetent mice. Anti-PD-1 combined with toxic ROS levels, enabling correct cell-cycle progression and ROCKi (combo) induced signiﬁcantly more regressions of estab- providing pro-survival signals. Because resistant cells have lished tumors compared with single treatments (Figures 8A and altered ROS and loss of DNA damage repair genes, ROCK- S8A), and treatments were well tolerated based on weight (Fig- myosin II inhibition is particularly detrimental. ure S8B). ROCKi-treated tumors had reduced p-MLC2 after 5 days of treatment or at endpoint (Figures 8B and S8C). ROCKi Combining ROCK Inhibitors with BRAF Inhibitors In Vivo also decreased immunosuppressive cell populations at both To translate our ﬁndings to pre-clinical in vivo models, we com- 5 days and endpoint: CD206 macrophages (Figures 8C and + + bined BRAFi and ROCKi (low dose) GSK269962A in BRAFi- S8D) and FOXP3 Tregs (Figures 8D and S8D). F4/80 (Fig- + + + resistant A375/PLX/R xenografts in nude mice. Mice tolerated ure S8D) and other immune populations (CD3 , CD4 , CD8 drug treatments well (Figure S7A). The combination treatment cells) were not signiﬁcantly affected by ROCKi in tumors or was the most efﬁcient and induced regression of tumors and spleens (Figures S8E and S8F). ROCKi did not affect percentage + + improved mouse survival (Figures 7A and S7B). of CD4 and CD8 cells expressing PD-1 (data not shown). Patient no. 2 cells displayed PLX4720 addiction in vitro (Fig- CD206 polarization mainly occurred in tumors since polariza- ures 5J, S5J, and S5K) and also in vivo (Figure 7B), as seen by tion in the spleens was less than 1% (Figure S8F). increased growth in the presence of PLX4720. ROCKi reduced We analyzed inﬁltration in the tumor body (TB) and invasive + + growth and p-MLC2 levels of PLX4720-resistant patient no. 2 xe- front (IF) and found that TB were inﬁltrated with CD3 , CD4 , nografts (Figures 7B, 7C, and S7C). and CD8 cells––but mostly accumulated in the IF––while ROCKi (D and E) Cell-cycle analysis after treatment (n = 3–4). Sensitive (s)-resistant (R) pairs (D, left A375; right WM983A). A375/PLX/R (E); G, GSK269962A; B, bleb- bistatin. (F) p-STAT3 levels after treatment (n = 3 patient no. 35, WM793B; n = 4 A375/PLX/R; n = 5 patients no. 2 and 58). (G) Mcl-1 levels of A375/PLX/R cells after treatment (n = 3). (H and I) Percentage of dead cells by annexin V/PI staining of A375/PLX/R (H), patient no. 2 and no. 62T3 (I) cells after a 3-day treatment (n = 4 A375/PLX/R; n = 5 patient no. 2; n = 4 patient no. 62T3). (C, D, and F–I) ROCKi GSK269962A; (E, H, and I) BRAFi PLX4720. (C–G) 24 h treatment. (C and F–I) Mean ± SEM and individual data points (circle). Asterisks in (D and E) are statistical signiﬁcance in multinucleated cells. p values by one-way ANOVA with Tukey’s (D–F, H, and I) or Benjamini, Krieger, and Yekutieli correction (C); unpaired t test with Welch’s correction (G), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also Figure S6. 96 Cancer Cell 37, 85–103, January 13, 2020 ROCKi Figure 7. Combining ROCK Inhibitors with daily A375/PLX/R % regression A BRAFi BRAFi vs Combo p<0.0001 BRAF Inhibitors In Vivo nude days Veh vs Combo p<0.0001 *** 14 7 ROCKi vs Combo p<0.0001 x2 *** (A) Top, schematic of experiment. Left, growth of Vehicle vs Combo p<0.001 Vehicle ROCKi BRAFi Combo A375/PLX/R xenografts in nude mice after treat- 600 600 600 600 0% 0% 0% 40% regression 400 ment. Middle, Kaplan-Meier survival plot. Right, tumor volume at endpoint (n = 4–6 mice/group). 400 400 400 400 50 (B) Left, volume of patient no. 2 xenografts in NSG 200 200 200 200 0 mice after a 21-day treatment (n = 7 mice/group). 0 5 10 15 20 Days of treatment - GG - ROCKi Right, tumor growth at endpoint versus baseline. 0 0 0 5 10 15 05 10 15 0 5 10 15 05 10 15 - BRAFi (C) p-MLC2 staining in patient no. 2 xenografts. Days of treatment Scale bar, 100 mm. % regression (D) Survival of patient no. 2 cells in the mouse lung ROCKi BRAFi vs Combo p<0.0001 B C Patient #2 daily 24 h post-injection (n = 8–9 mice from 2 experi- BRAFi BRAFi vs Veh p<0.0001 BRAFi vs ROCKi p<0.0001 days ments). Scale bar, 100 mm. Veh vs ROCKi p=0.037 30 7 * NSG x3 ns Veh vs Combo ns (E) Left, volume of patient no. 35 xenografts in - ROCKi ROCKi vs Combo p=0.033 ** ** * *** 2 28% 43% 0% 28% regression NSG mice after a 10-day treatment (n = 6 mice/ **** ** group). Right, p-MLC2 staining. Scale bar, 1 200 100 mm. (F and G) Images and quantiﬁcation of p-MLC2 (F), CD206 (G) in A375/PLX/R xenografts from -1 + + 50 0 (A). Scale bars, 100 mm. Ratio of CD206 /F4/80 - GG - ROCKi shown. (F and G) Pooled data from 2 experiments. - GG - ROCKi -2 - BRAFi Vehicle ROCKi BRAFi Combo - BRAFi (A–G) ROCKi GSK269962A, BRAFi PLX4720. Boxplots show median (center line); interquartile ROCKi Patient #35 daily D E range (box); min-max (whiskers); and individual BRAFi Patient #2 Green CMFDA mice (circles). p values by ANOVA with Tukey’s: days NSG in vitro 24 hr 10-16 7 3 NSG (A) right graph, (B) left graph; Benjamini, Krieger, ROCKi *** BRAFi ** ns and Yekutieli (C, D, F, G, and E, right) or Dunnett’s **** - ROCKi *** * * *** *** 600 correction (E, left), Mantel-Cox (A, survival plot), **** **** - ROCKi chi-square test: percentage regressions in (A, left) and (B, right). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also 1 Figure S7. 0 0 - GG - ROCKi - GG - - GG - ROCKi ROCKi - BRAFi - BRAFi - BRAFi ROCKi ROCKi F G daily A375/PLX/R BRAFi daily A375/PLX/R BRAFi (Figures 8E, 8G, and S8H). After 7 days, nude days nude 14 7 days we treated with anti-PD-1, ROCKi, or 14 7 x2 x2 both. Tumors on anti-PD-1 grew rapidly *** * - ROCKi - ROCKi ** ** ns but combo therapy resulted in >40% 1500 1.5 *** ** **** * * **** **** 300 **** ** regression of established tumors and 1000 1.0 improved survival (Figures 8G and S8H). Treatments were tolerated (Figure S8I) 500 0.5 and ROCKi reduced p-MLC2 (Figure 8H, 0 left). Importantly, combo decreased 0 0 - GG - ROCKi - GG - - GG - ROCKi - BRAFi expression of immune checkpoint ligand - BRAFi - BRAFi PD-L1 on tumor cells (Figure 8H, right). Anti-PD-1/NR tumors polarized most macrophages into CD206 phenotype (Figure S8J) and combo did not alter distribution (Figure S8G). Moreover, ROCKi did not + + affect viability of CD8 T cells or tumor-killing ability in vitro (data decreased expression of PD-L1 on CD206 macrophages + + not shown). Therefore, ROCKi does not affect CD4 and CD8 (Figure 8I, left), while total macrophage content did not change cell functions tested. (Figure S8K). Finally, combo decreased Tregs (Figure 8I, right), We next analyzed anti-PD-1/NR and Resp tumors (Figures 8E while other immune populations did not change (Figure S8K). and 8F). NR had increased levels of p-MLC2 and CD206 cells ROCK-myosin II regulates TGF-b secretion from amoeboid compared with Resp while on anti-PD-1 treatment (Figure 8F, melanoma cells (Cantelli et al., 2015). TGF-b is a potent immuno- middle). FOXP3 Tregs did not change (Figure 8F, right). NR tu- suppressor that induces Tregs and myeloid-derived suppressor mors polarized most macrophages into CD206 compared with cells (Cantelli et al., 2017; Condamine et al., 2015; Nakamura less polarization in Resp (Figure 8F, right) and in parental 5555 et al., 2001). Therefore, ROCKi decreased TGF-b1 levels (Figure 8C). These data could in part explain the lack of response secreted by immunotherapy-resistant patient-derived cell lines to anti-PD-1 (Figure 8F, left). and by 5555 cells (Figure S8L). Interleukin-6, CCL2, TGF-b1, Then an anti-PD-1/NR (intrinsic resistance) was allografted and colony-stimulating factor 1/macrophage colony-stimulating into new recipient mice that were treated with anti-PD-1 factor immunomodulatory cytokines (Fisher et al., 2014; Manto- twice a week post-injection to maintain resistance in vivo vani et al., 2004; Qian et al., 2011; Roca et al., 2009) regulated by Cancer Cell 37, 85–103, January 13, 2020 97 BRAFi - 3 BRAFi Tumor volume (mm ) 3 Tumor volume (mm ) Survival in lung (fluorescence area/field) Log fold change in (arbitrary units) 2 tumor growth vs baseline p-MLC2 score Tumor volume (mm ) BRAFi - % survival BRAFi - BRAFi - + 2 CD206 cells/mm Tumor volume (mm ) p-MLC2 score + + Ratio CD206 / F4/80 p-MLC2 score ROCKi daily anti-PD-1 2x/week ** ROCKi - ns % regresssion C57BL/6J days 300 ** Anti-PD-1 vs Combo p=0.02 7-14 77 Veh+IgG vs Combo p<0.0001 randomize Veh+IgG vs Anti-PD-1 p=0.066 10% 5% 19% regression 33% -5 - GG - ROCKi -15 IgG anti-PD-1 Vehicle + IgG ROCKi anti-PD-1 Combo C D **** - ROCKi **** ** ns - ROCKi ** ns **** 150 1.0 ** ** ** * * 0.8 0.6 0.4 500 50 0.2 - GG - - GG - ROCKi ROCKi - GG - ROCKi IgG anti-PD-1 IgG anti-PD-1 IgG anti-PD-1 All on ROCKi daily 5555 anti-PD-1 anti-PD-1 2x/week 5555 anti-PD-1 2x/week anti-PD-1/NR allograft onto days analysis new recipient mice 7 7 7 77 x4 IHC days C57BL/6J randomize randomize IgG 800 * p=0.06 ns Anti-PD-1 p=0.057 3000 * 3000 1.0 80 0.8 200 2000 2000 NR (intrinsic) 0.6 0.4 100 1000 NR (acquired) 200 20 0.2 Resp (stable) Resp (regression, no tumor) 0 0 0 0 0 0 5 10 15 20 NR Resp NR Resp 0 NR Resp NR Resp NR Resp Days of treatment Anti-PD-1 Anti-PD-1 Anti-PD-1 Anti-PD-1 Anti-PD-1 % regression G H Anti-PD-1 vs Combo p<0.0001 ** Veh+IgG vs Combo p<0.0001 ns 300 ** 11% 14% 12.5% 43% regression 4 ** ** Vehicle + IgG ROCKi Anti-PD-1 200 Combo 75 200 -2 -10 010 20 30 40 -12 0 0 Days of treatment - GG - - GG - ROCKi Vehicle + IgG ROCKi anti-PD-1 Combo IgG anti-PD-1 IgG anti-PD-1 -ROCKi FOXP3 I CD206 Treg - ROCKi Mφ ** ** ns ** ** ** * PD-L1 - GG - ROCKi - GG - ROCKi PD-L1 / CD206 IgG anti-PD-1 IgG anti-PD-1 Figure 8. ROCK-Myosin II Inhibition Improves Efﬁcacy of Immune Checkpoint Inhibitors (A) Top, schematic of treatment. Bottom, growth of 5555 allografts in C57BL/6J mice after treatment. Pooled data from 3 experiments (n = 6–8 mice/group/ experiment). (legend continued on next page) 98 Cancer Cell 37, 85–103, January 13, 2020 Tumor volume (mm ) anti-PD-1 - anti-PD-1 - Log fold change tumor growth vs baseline Log fold change tumor growth vs baseline + 2 PD-L1 score on CD206 CD206 cells/mm p-MLC2 score + + Ratio CD206 / F4/80 + 2 FOXP3 Treg cells/mm % survival + 2 CD206 cells/mm anti-PD-1 - anti-PD-1 - + 2 F4/80 cells/mm p-MLC2 score + + Ratio CD206 / F4/80 anti-PD-1 - p-MLC2 score + 2 FOXP3 Treg cells/mm PD-L1 score (tumor cells) + 2 FOXP3 Treg cells/mm ROCK-myosin II (Georgouli et al., 2019; Le Dreau et al., 2010) etal features are observed in metastatic lesions compared with were upregulated in group 1 MAPKi-resistant melanomas primary tumors (Cantelli et al., 2015; Herraiz et al., 2015; Orgaz (Figure S8M). Therefore, blocking ROCK-myosin II reduces et al., 2014b; Sanz-Moreno et al., 2011), which suggests that immunosuppressive microenvironments, improving anti-PD-1 metastatic traits can be linked to drug resistance (Alexander action on pre-existing T cells (Mariathasan et al., 2018; Tauriello and Friedl, 2012). Pathways controlling invasion and metastasis et al., 2018). are aberrantly activated by non-mutational mechanisms––over- expression or signaling alteration (Alexander and Friedl, 2012; DISCUSSION Orgaz et al., 2014a)––in contrast with frequently mutated MAPK (Davies et al., 2002; Cancer Genome Atlas Network, Recurrent transcriptional alterations occur during development 2015). Rho GTPases are overexpressed in cancer (Orgaz of resistance to MAPKi (Song et al., 2017). In this study we ﬁnd et al., 2014a); particularly RhoC is a driver of melanoma that adaptation to therapy occurs early on treatment through metastasis by increased expression (Clark et al., 2000). Lower cytoskeletal remodeling leading to restoration/increase of frequency of mutations suggests that cancer cells are less ad- myosin II levels in resistant melanomas. Because targeted and dicted to these pathways and, upon inhibition, development of immunotherapy-resistant cells rely on ROCK-dependent myosin resistance could be less frequent. Although we have shown II for survival, this could be a key mediator of cross-resistance. that myosin II inhibition also impairs survival of therapy-sensi- Resistant melanomas increase either MLC2 expression and/or tive melanoma cells, therapy-resistant cells are more sensitive activity, which in turn increases and reinforces myosin II activity to ROCKi. This is due to resistant cells having gained certain survival traits, but acquired vulnerabilities in return, such as (Calvo et al., 2013; Medjkane et al., 2009). Cells under drug treatment upregulate myosin II as a pro-survival response to defective anti-oxidant and DNA damage repair responses. MAPK inhibition, resulting in uncoupling of ERK signals to the Inhibition of myosin II activity overcomes resistance in mela- cytoskeleton. noma through induction of lethal ROS, unresolved DNA damage, Although myosin II activity is controlled by BRN2-mediated and loss of pro-survival signaling, which leads to cell-cycle arrest downregulation of PDE5A and increased calcium signaling in and cell death. A recent study has described that HDAC BRAF mutant melanoma (Arozarena et al., 2011), our mechanism inhibitors (HDACi) also induce lethal ROS and DNA damage in seems operative in NRAS mutant melanoma. PDE5A expression MAPKi-resistant melanomas (Wang et al., 2018). It will be increases in MAPKi-resistant lines compared with parental important to investigate if/how HDACi regulate cytoskeletal (Song et al., 2017) in a similar fashion as MLC2 (MYL9) (data remodeling. not shown). Because p-MLC2 levels are restored/increased in The tumor microenvironment has a key role in resistance to resistant versus parental lines, there may be mechanisms block- therapies in melanoma (Almeida et al., 2019) and macrophages ing the inhibitory action of PDE5A on myosin II in resistant cells. can contribute to resistance to MAPKi through secretion of Moreover, myosin II levels are ROCK dependent in resistant pro-survival factors (Smith et al., 2014). Furthermore, TGF-b inhi- cells, so PDE5A may not regulate myosin II activity in this bition enhanced efﬁcacy of immune checkpoint inhibitors (Ma- context. riathasan et al., 2018; Tauriello et al., 2018). In addition to the MAPKi-resistant cells have been associated to bundled cell intrinsic effects we observe, we report how inhibition of collagen and pro-survival signals (Brighton et al., 2018). ROCK-myosin II reduces pro-tumorigenic CD206 macro- Increased ECM deposition found in resistant tumors could phages, which could contribute to reducing tumor growth. More- contribute to myosin II activity in vivo (Laklai et al., 2016; Paszek over, ROCK-myosin II inhibition decreases FOXP3 Tregs. et al., 2005). Likewise, ROCK-myosin II-driven contractility also Combination of ROCKi with anti-PD-1 also reduces PD-L1 induces ECM stiffening (Samuel et al., 2011), generating a feed- expression on both tumor cells and CD206 macrophages. back loop between myosin II and ECM. These effects could be due to lower STAT3 activity after ROCK Widely studied in cell migration (Jaffe and Hall, 2005; Olson, inhibition (Sanz-Moreno et al., 2011), since PD-L1 expression 2008; Sadok et al., 2015; Sahai and Marshall, 2002; Sanz-Mor- can be regulated by STAT3 (Marzec et al., 2008; Pardoll, eno et al., 2008), ROCK-myosin II is proposed here as a thera- 2012). Effects on T cells are likely due to ROCK-myosin II regu- peutic target that goes beyond this pro-migratory function. We lation of TGF-b in cancer cells (Cantelli et al., 2015). Decreased show how this machinery controls intrinsic survival and TGF-b production by melanoma induced by ROCKi can extrinsic immunosuppression. Importantly, contractile cytoskel- contribute to improved anti-PD-1 responses. + + + (B–D) Images and quantiﬁcation of p-MLC2 (B), CD206 (C), and FOXP3 (D) cells in 5555 tumors at endpoint (pooled data from 2 experiments). Ratio CD206 /F4/ 80 shown. Scale bars, 100 mm (p-MLC2, CD206) and 50 mm (FOXP3). (E) Schematic of experiment. + + + + (F) Left, growth of 5555 allografts after treatment. Right, quantiﬁcation of p-MLC2, CD206 , F4/80 , and ratio CD206 /F4/80 in anti-PD-1/NR or Resp tumors. (G) Left, growth of 5555 anti-PD-1/NR allografts in new recipient mice after treatment (n = 7-8 mice/group). Right, survival plot. (H) Images and quantiﬁcation of p-MLC2 and PD-L1 on tumor cells in tumors from (G). Scale bars, 25 mm. (I) Left, images and quantiﬁcation of PD-L1 on CD206 cells in tumors from (G). Images show merged pseudo-colors for each staining. Scale bar, 50 mm. Right, quantiﬁcation of FOXP3 Tregs in tumors from (G). (A–D and G–I) ROCKi GSK269962A. Boxplots show median (center line); interquartile range (box); min-max (whiskers); and individual mice (circles). p values by ANOVA with Benjamini, Krieger, and Yekutieli correction (B–D and H–I), t test (F), chi-square test: percentage regressions in (A) and (G). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also Figure S8. Cancer Cell 37, 85–103, January 13, 2020 99 B Phospho-proteomics ROCKi Fasudil has been used safely in Japan since 1995 to B Phospho-Peptide Enrichment Analysis treat subarachnoid hemorrhage (SAH) after a head trauma and B Quantitative Real Time One-Step PCR to prevent vasospasm associated with SAH (Feng et al., 2016; B Gene Expression Studies and Analysis Olson, 2008). ROCKi is given as a vasodilator to lower blood B Gene Enrichment Analyzes pressure (Olson, 2008), and Fasudil and other ROCKi are being B Tumor Xenografts tested in clinical trials for glaucoma and other vascular dis- B Immunotherapy Experiments eases, such as pulmonary hypertension and atherosclerosis B Survival in the Lung Assay (Olson, 2008). Optimal ROCKi could be tested in broader B Immunohistochemistry range of disease, as a strategy to extend clinical response to B Imaging and Scoring different cancer therapies or even as a single therapy in the d QUANTIFICATION AND STATISTICAL ANALYSIS case of drug-addicted tumors. Importantly, therapy-resistant d DATA AND CODE AVAILABILITY cells are more sensitive to ROCKi while its combination with current therapies seems to elicit a superior response. Lower doses of ROCKi and/or different schedule treatments could SUPPLEMENTAL INFORMATION be used in combination with current therapies to prolong their Supplemental Information can be found online at https://doi.org/10.1016/j. efﬁcacy and delay resistance. Alternatively, different delivery ccell.2019.12.003. strategies of ROCKi (local, antibody-drug conjugate) could be considered. In summary, we provide extensive evidence that targeting ACKNOWLEDGMENTS cytoskeletal regulators driving high myosin II activity The work was supported by Cancer Research UK (CRUK) C33043/A12065 (to overcomes resistance to targeted and immunotherapies in V.S.-M., J.L.O., and E.C.-M.), C33043/A24478 (to V.S.-M., E.C.-M., M.G., and melanoma. The cytoskeletal adaptations that occur very J.L.O.); C30122/A11527, C30122/A15774, C33043/A12065 (to S.N.K.); C107/ early on treatment provide not only a survival advantage but A10433, C107/A104339 (to A.S.); The Harry J. Lloyd Charitable Trust (to J.L.O. also a vulnerability, which can be later exploited. High and V.S.-M.); Barts Charity (to V.S.-M., I.R.-H., and J.M.); Royal Society myosin II activity identiﬁes therapy cross-resistant RG110591 (to V.S.-M.); Fundacio´ n Ramo´ n Areces (to E.C.-M.); Marie Sklo- dowska-Curie Action (H2020-MSCA-IF-2014-EF-ST) (to I.R.-H.); MR/ patients, suggesting its potential as a biomarker. Our work L023091/1 (to S.N.K. and S.M.); CRUK/NIHR in England/DoH for Scotland, opens the possibility that cytoskeletal remodeling could be a Wales and Northern Ireland ECMC (C10355/A15587) (to S.N.K.); Francis Crick conserved pro-survival mechanism of generating therapy- Institute core funding from CRUK (FC001112), MRC (FC001112), and Well- resistant cancer clones under the selection of other therapy come Trust (FC001112) (to I.M. and A.P.); NIHR Biomedical Research Centre regimes. (BRC) at Guy’s and St Thomas’ NHS Foundation Trust and King’s College Lon- don, IS-BRC-1215-20006 (to S.N.K.). Fluorescence-activated cell sorting was performed at BRC funded by NIHR. We are indebted to Richard Marais and his STAR+METHODS team (CRUK Manchester Institute) for kind provision of cell lines and Paul Lor- igan (University of Manchester) for providing patient samples (study approved Detailed methods are provided in the online version of this paper by Manchester Cancer Research Center Biobank Access Committee applica- and include the following: tion 13_RIMA_01; the role of the MCRC Biobank is to distribute samples and cannot endorse studies performed or interpretation of results). We thank d KEY RESOURCES TABLE Fran Balkwill, Colin Pegrum, and the Biological Services Unit at Barts Cancer Institute for help with mouse work; Romina Girotti and Jeremy Carlton for help- d LEAD CONTACT AND MATERIALS AVAILABILITY ful discussions; Fernando Calvo for MLC2 plasmids and advice on ssGSEA; d EXPERIMENTAL MODEL AND SUBJECT DETAILS Amaya Viros, Garry Ashton, Sandra Kumper, € and Michela Perani for technical B Patient-Derived Samples advice; Erik Sahai and Tohru Takaki for MLC2 lentivectors; Øystein Fodstad for B Cell Lines and Patient-Derived Cell Lines LOX-IMVI cells; and Wellcome Trust Functional Genomics Cell Bank for B Animals MM485 cells. d METHOD DETAILS B Chemicals AUTHOR CONTRIBUTIONS B Antibodies Conceptualization, V.S.-M. and J.L.O.; Methodology, J.L.O., A.P.-R., R.L., B Analysis of Cell Morphology I.M., V.S.-M., and F.W.; Investigation, J.L.O., E.C.-M., A.S., O.M., I.R.-H., B Long-Term Survival A.P.-R., J.M., V.B., M.G., P.P., L.B., S.M., P.K., C.T., and F.W.; Validation, B Long Term Survival on Collagen I Matrices J.L.O., E.C.-M., I.R.-H., A.P.-R., and O.M.; Writing – Original Draft, V.S.-M. B MTT Assay and J.L.O.; Writing – Review & Editing, V.S.-M., J.L.O., E.C.-M., and I.M.; B Cell Cycle Analysis Funding Acquisition, V.S.-M., J.L.O., I.M., and S.N.K.; Resources, I.M., B AnnexinV/Propidium Iodide FACS S.N.K., and R.L.; Supervision, V.S.-M. B ROS Detection B Time Lapse Microscopy DECLARATION OF INTERESTS B RNAi The authors declare no competing interests. B MLC2 Rescue Experiments B MLC2 Stable Overexpression Received: January 9, 2018 B Immunoﬂuorescence and Confocal Imaging Revised: September 4, 2019 B Immunoblotting Accepted: December 6, 2019 B TGF-b1 ELISA Published: January 13, 2020 100 Cancer Cell 37, 85–103, January 13, 2020 REFERENCES CCL2 and CXCL1 expression in astrocytes through beta1 and beta5 integrins. Glia 58, 1510–1521. Alexander, S., and Friedl, P. (2012). Cancer invasion and resistance: intercon- Feng, Y., LoGrasso, P.V., Defert, O., and Li, R. (2016). Rho kinase (ROCK) in- nected processes of disease progression and therapy failure. Trends Mol. hibitors and their therapeutic potential. J. Med. Chem. 59, 2269–2300. Med. 18, 13–26. Fisher, D.T., Appenheimer, M.M., and Evans, S.S. (2014). The two faces of IL-6 Almeida, F.V., Douglass, S.M., Fane, M.E., and Weeraratna, A.T. (2019). Bad in the tumor microenvironment. Semin. Immunol. 26, 38–47. company: microenvironmentally mediated resistance to targeted therapy in Flaherty, K.T., Puzanov, I., Kim, K.B., Ribas, A., McArthur, G.A., Sosman, J.A., melanoma. Pigment Cell Melanoma Res. 32, 237–247. O’Dwyer, P.J., Lee, R.J., Grippo, J.F., Nolop, K., and Chapman, P.B. (2010). Arozarena, I., Sanchez-Laorden, B., Packer, L., Hidalgo-Carcedo, C., Inhibition of mutated, activated BRAF in metastatic melanoma. N. Engl. J. Hayward, R., Viros, A., Sahai, E., and Marais, R. (2011). Oncogenic BRAF in- Med. 363, 809–819. duces melanoma cell invasion by downregulating the cGMP-speciﬁc phos- Flaherty, K.T., Infante, J.R., Daud, A., Gonzalez, R., Kefford, R.F., Sosman, J., phodiesterase PDE5A. Cancer Cell 19, 45–57. Hamid, O., Schuchter, L., Cebon, J., Ibrahim, N., et al. (2012). Combined BRAF and MEK inhibition in melanoma with BRAF V600 mutations. N. Engl. J. Med. Baenke, F., Chaneton, B., Smith, M., Van Den Broek, N., Hogan, K., Tang, H., 367, 1694–1703. Viros, A., Martin, M., Galbraith, L., Girotti, M.R., et al. (2015). Resistance to BRAF inhibitors induces glutamine dependency in melanoma cells. Mol. Gadea, G., Sanz-Moreno, V., Self, A., Godi, A., and Marshall, C.J. (2008). Oncol. 10, 73–84. DOCK10-mediated Cdc42 activation is necessary for amoeboid invasion of melanoma cells. Curr. Biol. 18, 1456–1465. Balch, C.M., Gershenwald, J.E., Soong, S.J., Thompson, J.F., Atkins, M.B., Byrd, D.R., Buzaid, A.C., Cochran, A.J., Coit, D.G., Ding, S., et al. (2009). Georgouli, M., Herraiz, C., Crosas-Molist, E., Fanshawe, B., Maiques, O., Final version of 2009 AJCC melanoma staging and classiﬁcation. J. Clin. Perdrix, A., Pandya, P., Rodriguez-Hernandez, I., Ilieva, K.M., Cantelli, G., Oncol. 27, 6199–6206. et al. (2019). Regional activation of myosin II in cancer cells drives tumor pro- gression via a secretory cross-talk with the immune microenvironment. Cell Bankhead, P., Loughrey, M.B., Fernandez, J.A., Dombrowski, Y., McArt, D.G., 176, 757–774.e23. Dunne, P.D., McQuaid, S., Gray, R.T., Murray, L.J., Coleman, H.G., et al. (2017). QuPath: open source software for digital pathology image analysis. Girotti, M.R., Pedersen, M., Sanchez-Laorden, B., Viros, A., Turajlic, S., Niculescu-Duvaz, D., Zambon, A., Sinclair, J., Hayes, A., Gore, M., et al. Sci. Rep. 7, 16878. (2013). Inhibiting EGF receptor or SRC family kinase signaling overcomes Brighton, H.E., Angus, S.P., Bo, T., Roques, J., Tagliatela, A.C., Darr, D.B., BRAF inhibitor resistance in melanoma. Cancer Discov. 3, 158–167. Karagoz, K., Sciaky, N., Gatza, M.L., Sharpless, N.E., et al. (2018). New mech- Gray-Schopfer, V., Wellbrock, C., and Marais, R. (2007). Melanoma biology anisms of resistance to MEK inhibitors in melanoma revealed by intravital im- and new targeted therapy. Nature 445, 851–857. aging. Cancer Res. 78, 542–557. Hall, A. (2012). Rho family GTPases. Biochem. Soc. Trans. 40, 1378–1382. Calvo, F., Sanz-Moreno, V., Agudo-Ibanez, L., Wallberg, F., Sahai, E., Marshall, C.J., and Crespo, P. (2011). RasGRF suppresses Cdc42-mediated Haystead, T.A. (2005). ZIP kinase, a key regulator of myosin protein phospha- tase 1. Cell Signal. 17, 1313–1322. tumour cell movement, cytoskeletal dynamics and transformation. Nat. Cell Biol. 13, 819–826. Herraiz, C., Calvo, F., Pandya, P., Cantelli, G., Rodriguez-Hernandez, I., Orgaz, J.L., Kang, N., Chu, T., Sahai, E., and Sanz-Moreno, V. (2015). Reactivation of Calvo, F., Ege, N., Grande-Garcia, A., Hooper, S., Jenkins, R.P., Chaudhry, p53 by a cytoskeletal sensor to control the balance between DNA damage and S.I., Harrington, K., Williamson, P., Moeendarbary, E., Charras, G., and tumor dissemination. J. Natl. Cancer Inst. 108, https://doi.org/10.1093/jnci/ Sahai, E. (2013). Mechanotransduction and YAP-dependent matrix remodel- djv289. ling is required for the generation and maintenance of cancer-associated ﬁbro- blasts. Nat. Cell Biol. 15, 637–646. Hirata, E., Girotti, M.R., Viros, A., Hooper, S., Spencer-Dene, B., Matsuda, M., Larkin, J., Marais, R., and Sahai, E. (2015). Intravital imaging reveals how BRAF Cancer Genome Atlas Network. (2015). Genomic classiﬁcation of cutaneous inhibition generates drug-tolerant microenvironments with high integrin beta1/ melanoma. Cell 161, 1681–1696. FAK signaling. Cancer Cell 27, 574–588. Cantelli, G., Orgaz, J.L., Rodriguez-Hernandez, I., Karagiannis, P., Maiques, Hodi, F.S., O’Day, S.J., McDermott, D.F., Weber, R.W., Sosman, J.A., Haanen, O., Matias-Guiu, X., Nestle, F.O., Marti, R.M., Karagiannis, S.N., and Sanz- J.B., Gonzalez, R., Robert, C., Schadendorf, D., Hassel, J.C., et al. (2010). Moreno, V. (2015). TGF-b-induced transcription sustains amoeboid melanoma Improved survival with ipilimumab in patients with metastatic melanoma. migration and dissemination. Curr. Biol. 25, 2899–2914. N. Engl. J. Med. 363, 711–723. Cantelli, G., Crosas-Molist, E., Georgouli, M., and Sanz-Moreno, V. (2017). Hong, A., Moriceau, G., Sun, L., Lomeli, S., Piva, M., Damoiseaux, R., Holmen, TGFBeta-induced transcription in cancer. Semin. Cancer Biol. 42, 60–69. S.L., Sharpless, N.E., Hugo, W., and Lo, R.S. (2017). Exploiting drug addiction Chapman, P.B., Hauschild, A., Robert, C., Haanen, J.B., Ascierto, P., Larkin, mechanisms to select against MAPKi-resistant melanoma. Cancer Discov. J., Dummer, R., Garbe, C., Testori, A., Maio, M., et al. (2011). Improved survival 8, 74–93. with vemurafenib in melanoma with BRAF V600E mutation. N. Engl. J. Med. Hugo, W., Shi, H., Sun, L., Piva, M., Song, C., Kong, X., Moriceau, G., Hong, A., 364, 2507–2516. Dahlman, K.B., Johnson, D.B., et al. (2015). Non-genomic and immune evolu- Clark, E.A., Golub, T.R., Lander, E.S., and Hynes, R.O. (2000). Genomic anal- tion of melanoma acquiring MAPKi resistance. Cell 162, 1271–1285. ysis of metastasis reveals an essential role for RhoC. Nature 406, 532–535. Hugo, W., Zaretsky, J.M., Sun, L., Song, C., Moreno, B.H., Hu-Lieskovan, S., Condamine, T., Ramachandran, I., Youn, J.I., and Gabrilovich, D.I. (2015). Berent-Maoz, B., Pang, J., Chmielowski, B., Cherry, G., et al. (2016). Genomic Regulation of tumor metastasis by myeloid-derived suppressor cells. Annu. and transcriptomic features of response to anti-PD-1 therapy in metastatic Rev. Med. 66, 97–110. melanoma. Cell 165, 35–44. Davies, H., Bignell, G.R., Cox, C., Stephens, P., Edkins, S., Clegg, S., Teague, Ito, M., Nakano, T., Erdodi, F., and Hartshorne, D.J. (2004). Myosin phospha- J., Woffendin, H., Garnett, M.J., Bottomley, W., et al. (2002). Mutations of the tase: structure, regulation and function. Mol. Cell. Biochem. 259, 197–209. BRAF gene in human cancer. Nature 417, 949–954. Itoh, K., Yoshioka, K., Akedo, H., Uehata, M., Ishizaki, T., and Narumiya, S. (1999). An essential part for Rho-associated kinase in the transcellular invasion Dhomen, N., Reis-Filho, J.S., da Rocha Dias, S., Hayward, R., Savage, K., of tumor cells. Nat. Med. 5, 221–225. Delmas, V., Larue, L., Pritchard, C., and Marais, R. (2009). Oncogenic Braf in- duces melanocyte senescence and melanoma in mice. Cancer Cell 15, Jaffe, A.B., and Hall, A. (2005). Rho GTPases: biochemistry and biology. Annu. 294–303. Rev. Cell Dev. Biol. 21, 247–269. Le Dreau, G., Kular, L., Nicot, A.B., Calmel, C., Melik-Parsadaniantz, S., Kakavand, H., Rawson, R.V., Pupo, G.M., Yang, J.Y.H., Menzies, A.M., Kitabgi, P., Laurent, M., and Martinerie, C. (2010). NOV/CCN3 upregulates Carlino, M.S., Kefford, R.F., Howle, J.R., Saw, R.P.M., Thompson, J.F., et al. Cancer Cell 37, 85–103, January 13, 2020 101 (2017). PD-L1 expression and immune escape in melanoma resistance to Marzec, M., Zhang, Q., Goradia, A., Raghunath, P.N., Liu, X., Paessler, M., MAPK inhibitors. Clin. Cancer Res. 23, 6054–6061. Wang, H.Y., Wysocka, M., Cheng, M., Ruggeri, B.A., and Wasik, M.A. (2008). Oncogenic kinase NPM/ALK induces through STAT3 expression of Kong, X., Kuilman, T., Shahrabi, A., Boshuizen, J., Kemper, K., Song, J.Y., immunosuppressive protein CD274 (PD-L1, B7-H1). Proc. Natl. Acad. Sci. U Niessen, H.W.M., Rozeman, E.A., Geukes Foppen, M.H., Blank, C.U., et al. SA 105, 20852–20857. (2017). Cancer drug addiction is relayed by an ERK2-dependent phenotype switch. Nature 550, 270–274. Medjkane, S., Perez-Sanchez, C., Gaggioli, C., Sahai, E., and Treisman, R. (2009). Myocardin-related transcription factors and SRF are required for cyto- Konieczkowski, D.J., Johannessen, C.M., Abudayyeh, O., Kim, J.W., Cooper, skeletal dynamics and experimental metastasis. Nat. Cell Biol. 11, 257–268. Z.A., Piris, A., Frederick, D.T., Barzily-Rokni, M., Straussman, R., Haq, R., et al. Mjelle, R., Hegre, S.A., Aas, P.A., Slupphaug, G., Drablos, F., Saetrom, P., and (2014). A melanoma cell state distinction inﬂuences sensitivity to MAPK pathway inhibitors. Cancer Discov. 4, 816–827. Krokan, H.E. (2015). Cell cycle regulation of human DNA repair and chromatin remodeling genes. DNA Repair (Amst.) 30, 53–67. Konieczkowski, D.J., Johannessen, C.M., and Garraway, L.A. (2018). A convergence-based framework for cancer drug resistance. Cancer Cell 33, Moriceau, G., Hugo, W., Hong, A., Shi, H., Kong, X., Yu, C.C., Koya, R.C., Samatar, A.A., Khanlou, N., Braun, J., et al. (2015). Tunable-combinatorial 801–815. mechanisms of acquired resistance limit the efﬁcacy of BRAF/MEK cotarget- Krokan, H.E., and Bjoras, M. (2013). Base excision repair. Cold Spring Harb. ing but result in melanoma drug addiction. Cancer Cell 27, 240–256. Perspect. Biol. 5, a012583. Nakamura, K., Kitani, A., and Strober, W. (2001). Cell contact-dependent Kumper, S., Mardakheh, F.K., McCarthy, A., Yeo, M., Stamp, G.W., Paul, A., immunosuppression by CD4(+)CD25(+) regulatory T cells is mediated by cell Worboys, J., Sadok, A., Jorgensen, C., Guichard, S., and Marshall, C.J. surface-bound transforming growth factor beta. J. Exp. Med. 194, 629–644. (2016). Rho-associated kinase (ROCK) function is essential for cell cycle pro- Obenauf, A.C., Zou, Y., Ji, A.L., Vanharanta, S., Shu, W., Shi, H., Kong, X., gression, senescence and tumorigenesis. Elife 5, e12994. Bosenberg, M.C., Wiesner, T., Rosen, N., et al. (2015). Therapy-induced Kwong, L.N., Boland, G.M., Frederick, D.T., Helms, T.L., Akid, A.T., Miller, J.P., tumour secretomes promote resistance and tumour progression. Nature Jiang, S., Cooper, Z.A., Song, X., Seth, S., et al. (2015). Co-clinical assessment 520, 368–372. identiﬁes patterns of BRAF inhibitor resistance in melanoma. J. Clin. Invest. Olson, M.F. (2008). Applications for ROCK kinase inhibition. Curr. Opin. Cell 125, 1459–1470. Biol. 20, 242–248. Laklai, H., Miroshnikova, Y.A., Pickup, M.W., Collisson, E.A., Kim, G.E., Orgaz, J.L., Herraiz, C., and Sanz-Moreno, V. (2014a). Rho GTPases modulate Barrett, A.S., Hill, R.C., Lakins, J.N., Schlaepfer, D.D., Mouw, J.K., et al. malignant transformation of tumor cells. Small GTPases 5, e29019. (2016). Genotype tunes pancreatic ductal adenocarcinoma tissue tension to induce matricellular ﬁbrosis and tumor progression. Nat. Med. 22, 497–505. Orgaz, J.L., Pandya, P., Dalmeida, R., Karagiannis, P., Sanchez-Laorden, B., Viros, A., Albrengues, J., Nestle, F.O., Ridley, A.J., Gaggioli, C., et al. Lammermann, T., and Sixt, M. (2009). Mechanical modes of ’amoeboid’ cell (2014b). Diverse matrix metalloproteinase functions regulate cancer amoeboid migration. Curr. Opin. Cell Biol. 21, 636–644. migration. Nat. Commun. 5, 4255. Larkin, J., Ascierto, P.A., Dreno, B., Atkinson, V., Liszkay, G., Maio, M., Pardoll, D.M. (2012). The blockade of immune checkpoints in cancer immuno- Mandala, M., Demidov, L., Stroyakovskiy, D., Thomas, L., et al. (2014). therapy. Nat. Rev. Cancer 12, 252–264. Combined vemurafenib and cobimetinib in BRAF-mutated melanoma. N. Engl. J. Med. 371, 1867–1876. Paszek, M.J., Zahir, N., Johnson, K.R., Lakins, J.N., Rozenberg, G.I., Gefen, A., Reinhart-King, C.A., Margulies, S.S., Dembo, M., Boettiger, D., et al. Larkin, J., Chiarion-Sileni, V., Gonzalez, R., Grob, J.J., Cowey, C.L., Lao, C.D., (2005). Tensional homeostasis and the malignant phenotype. Cancer Cell 8, Schadendorf, D., Dummer, R., Smylie, M., Rutkowski, P., et al. (2015). 241–254. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma. N. Engl. J. Med. 373, 23–34. Peng, G., Chun-Jen Lin, C., Mo, W., Dai, H., Park, Y.Y., Kim, S.M., Peng, Y., Mo, Q., Siwko, S., Hu, R., et al. (2014). Genome-wide transcriptome proﬁling Lau, P.K., Ascierto, P.A., and McArthur, G. (2016). Melanoma: the intersection of homologous recombination DNA repair. Nat. Commun. 5, 3361. of molecular targeted therapy and immune checkpoint inhibition. Curr. Opin. Immunol. 39, 30–38. Poli, V., and Camporeale, A. (2015). STAT3-mediated metabolic reprograming in cellular transformation and implications for drug resistance. Front. Oncol. Lito, P., Pratilas, C.A., Joseph, E.W., Tadi, M., Halilovic, E., Zubrowski, M., 5, 121. Huang, A., Wong, W.L., Callahan, M.K., Merghoub, T., et al. (2012). Relief of profound feedback inhibition of mitogenic signaling by RAF inhibitors attenu- Posern, G., and Treisman, R. (2006). Actin’ together: serum response factor, its ates their activity in BRAFV600E melanomas. Cancer Cell 22, 668–682. cofactors and the link to signal transduction. Trends Cell Biol. 16, 588–596. Long, G.V., Fung, C., Menzies, A.M., Pupo, G.M., Carlino, M.S., Hyman, J., Qian, B.Z., Li, J., Zhang, H., Kitamura, T., Zhang, J., Campion, L.R., Kaiser, Shahheydari, H., Tembe, V., Thompson, J.F., Saw, R.P., et al. (2014a). E.A., Snyder, L.A., and Pollard, J.W. (2011). CCL2 recruits inﬂammatory mono- Increased MAPK reactivation in early resistance to dabrafenib/trametinib com- cytes to facilitate breast-tumour metastasis. Nature 475, 222–225. bination therapy of BRAF-mutant metastatic melanoma. Nat. Commun. Riaz, N., Havel, J.J., Makarov, V., Desrichard, A., Urba, W.J., Sims, J.S., Hodi, 5, 5694. F.S., Martin-Algarra, S., Mandal, R., Sharfman, W.H., et al. (2017). Tumor and Long, G.V., Stroyakovskiy, D., Gogas, H., Levchenko, E., de Braud, F., Larkin, microenvironment evolution during immunotherapy with nivolumab. Cell 171, J., Garbe, C., Jouary, T., Hauschild, A., Grob, J.J., et al. (2014b). Combined 934–949.e16. BRAF and MEK inhibition versus BRAF inhibition alone in melanoma. Rizos, H., Menzies, A.M., Pupo, G.M., Carlino, M.S., Fung, C., Hyman, J., N. Engl. J. Med. 371, 1877–1888. Haydu, L.E., Mijatov, B., Becker, T.M., Boyd, S.C., et al. (2014). BRAF inhibitor Lu, H., Liu, S., Zhang, G., Bin, W., Zhu, Y., Frederick, D.T., Hu, Y., Zhong, W., resistance mechanisms in metastatic melanoma: spectrum and clinical Randell, S., Sadek, N., et al. (2017). PAK signalling drives acquired drug resis- impact. Clin. Cancer Res. 20, 1965–1977. tance to MAPK inhibitors in BRAF-mutant melanomas. Nature 550, 133–136. Robert, C., Karaszewska, B., Schachter, J., Rutkowski, P., Mackiewicz, A., Mantovani, A., Sica, A., Sozzani, S., Allavena, P., Vecchi, A., and Locati, M. Stroiakovski, D., Lichinitser, M., Dummer, R., Grange, F., Mortier, L., et al. (2004). The chemokine system in diverse forms of macrophage activation (2015). Improved overall survival in melanoma with combined dabrafenib and polarization. Trends Immunol. 25, 677–686. and trametinib. N. Engl. J. Med. 372, 30–39. Mariathasan, S., Turley, S.J., Nickles, D., Castiglioni, A., Yuen, K., Wang, Y., Roca, H., Varsos, Z.S., Sud, S., Craig, M.J., Ying, C., and Pienta, K.J. (2009). Kadel, E.E., III, Koeppen, H., Astarita, J.L., Cubas, R., et al. (2018). TGFbeta CCL2 and interleukin-6 promote survival of human CD11b+ peripheral blood attenuates tumour response to PD-L1 blockade by contributing to exclusion mononuclear cells and induce M2-type macrophage polarization. J. Biol. of T cells. Nature 554, 544–548. Chem. 284, 34342–34354. 102 Cancer Cell 37, 85–103, January 13, 2020 Sadok, A., McCarthy, A., Caldwell, J., Collins, I., Garrett, M.D., Yeo, M., (2017). Actomyosin drives cancer cell nuclear dysmorphia and threatens Hooper, S., Sahai, E., Kuemper, S., Mardakheh, F.K., and Marshall, C.J. genome stability. Nat. Commun. 8, 16013. (2015). Rho kinase inhibitors block melanoma cell migration and inhibit metas- Tauriello, D.V.F., Palomo-Ponce, S., Stork, D., Berenguer-Llergo, A., Badia- tasis. Cancer Res. 75, 2272–2284. Ramentol, J., Iglesias, M., Sevillano, M., Ibiza, S., Canellas, A., Hernando- Sahai, E., and Marshall, C.J. (2002). RHO-GTPases and cancer. Nat. Rev. Momblona, X., et al. (2018). TGFbeta drives immune evasion in genetically Cancer 2, 133–142. reconstituted colon cancer metastasis. Nature 554, 538–543. Samuel, M.S., Lopez, J.I., McGhee, E.J., Croft, D.R., Strachan, D., Timpson, Das Thakur, M., Salangsang, F., Landman, A.S., Sellers, W.R., Pryer, N.K., P., Munro, J., Schroder, E., Zhou, J., Brunton, V.G., et al. (2011). Levesque, M.P., Dummer, R., McMahon, M., and Stuart, D.D. (2013). Actomyosin-mediated cellular tension drives increased tissue stiffness and Modelling vemurafenib resistance in melanoma reveals a strategy to forestall beta-catenin activation to induce epidermal hyperplasia and tumor growth. drug resistance. Nature 494, 251–255. Cancer Cell 19, 776–791. Titz, B., Lomova, A., Le, A., Hugo, W., Kong, X., Ten Hoeve, J., Friedman, M., Sanz-Moreno, V., Gadea, G., Ahn, J., Paterson, H., Marra, P., Pinner, S., Sahai, Shi, H., Moriceau, G., Song, C., et al. (2016). JUN dependency in distinct early E., and Marshall, C.J. (2008). Rac activation and inactivation control plasticity and late BRAF inhibition adaptation states of melanoma. Cell Discov. 2, 16028. of tumor cell movement. Cell 135, 510–523. Di Veroli, G.Y., Fornari, C., Wang, D., Mollard, S., Bramhall, J.L., Richards, Sanz-Moreno, V., Gaggioli, C., Yeo, M., Albrengues, J., Wallberg, F., Viros, A., F.M., and Jodrell, D.I. (2016). Combeneﬁt: an interactive platform for the anal- Hooper, S., Mitter, R., Feral, C.C., Cook, M., et al. (2011). ROCK and JAK1 ysis and visualization of drug combinations. Bioinformatics 32, 2866–2868. signaling cooperate to control actomyosin contractility in tumor cells and Vicente-Manzanares, M., Ma, X., Adelstein, R.S., and Horwitz, A.R. (2009). stroma. Cancer Cell 20, 229–245. Non-muscle myosin II takes centre stage in cell adhesion and migration. Shaltiel, I.A., Krenning, L., Bruinsma, W., and Medema, R.H. (2015). The same, Nat. Rev. Mol. Cell Biol. 10, 778–790. only different––DNA damage checkpoints and their reversal throughout the Wagle, N., Emery, C., Berger, M.F., Davis, M.J., Sawyer, A., Pochanard, P., cell cycle. J. Cell Sci. 128, 607–620. Kehoe, S.M., Johannessen, C.M., Macconaill, L.E., Hahn, W.C., et al. (2011). Sharma, P., Hu-Lieskovan, S., Wargo, J.A., and Ribas, A. (2017). Primary, Dissecting therapeutic resistance to RAF inhibition in melanoma by tumor adaptive, and acquired resistance to cancer immunotherapy. Cell 168, genomic proﬁling. J. Clin. Oncol. 29, 3085–3096. 707–723. Wagle, N., Van Allen, E.M., Treacy, D.J., Frederick, D.T., Cooper, Z.A., Taylor- Smith, M.P., Sanchez-Laorden, B., O’Brien, K., Brunton, H., Ferguson, J., Weiner, A., Rosenberg, M., Goetz, E.M., Sullivan, R.J., Farlow, D.N., et al. Young, H., Dhomen, N., Flaherty, K.T., Frederick, D.T., Cooper, Z.A., et al. (2014). MAP kinase pathway alterations in BRAF-mutant melanoma patients (2014). The immune microenvironment confers resistance to MAPK pathway with acquired resistance to combined RAF/MEK inhibition. Cancer Discov. inhibitors through macrophage-derived TNFalpha. Cancer Discov. 4, 4, 61–68. 1214–1229. Wang, L., Leite de Oliveira, R., Huijberts, S., Bosdriesz, E., Pencheva, N., Song, C., Piva, M., Sun, L., Hong, A., Moriceau, G., Kong, X., Zhang, H., Brunen, D., Bosma, A., Song, J.Y., Zevenhoven, J., Los-de Vries, G.T., et al. Lomeli, S., Qian, J., Yu, C.C., et al. (2017). Recurrent tumor cell-intrinsic and (2018). An acquired vulnerability of drug-resistant melanoma with therapeutic -extrinsic alterations during MAPKi-induced melanoma regression and early potential. Cell 173, 1413–1425.e14. adaptation. Cancer Discov. 7, 1248–1265. Wolf, K., Muller, R., Borgmann, S., Brocker, E.B., and Friedl, P. (2003). Sun, C., Wang, L., Huang, S., Heynen, G.J., Prahallad, A., Robert, C., Haanen, Amoeboid shape change and contact guidance: T-lymphocyte crawling J., Blank, C., Wesseling, J., Willems, S.M., et al. (2014). Reversible and adap- through ﬁbrillar collagen is independent of matrix remodeling by MMPs and tive resistance to BRAF(V600E) inhibition in melanoma. Nature 508, 118–122. other proteases. Blood 102, 3262–3269. Takaki, T., Montagner, M., Serres, M.P., Le Berre, M., Russell, M., Collinson, Zhang, W. (2015). BRAF inhibitors: the current and the future. Curr. Opin. L., Szuhai, K., Howell, M., Boulton, S.J., Sahai, E., and Petronczki, M. Pharmacol. 23, 68–73. Cancer Cell 37, 85–103, January 13, 2020 103 STAR+METHODS KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies pT202/Y204-p44/42 (ERK1/2) Cell Signaling Technology Cat# 4370; RRID:AB_2315112 pThr18/Ser19-MLC2 Cell Signaling Technology Cat# 3674; RRID: AB_2147464 pSer19-MLC2 Cell Signaling Technology Cat# 3671; RRID: AB_330248 MLC2 Cell Signaling Technology Cat# 3672; RRID: AB_10692513 PD-L1 clone E1L3N Cell Signaling Technology Cat# 13684; RRID:AB_2687655 pY705-STAT3 Cell Signaling Technology Cat# 9145; RRID:AB_2491009 ERK2 Santa Cruz Biotechnology Cat# sc-154; RRID:AB_2141292 GAPDH Santa Cruz Biotechnology Cat# MAB374; RRID:AB_2107445 GFP Santa Cruz Biotechnology Cat# sc-8334; RRID:AB_641123 MCL-1 Santa Cruz Biotechnology Cat# sc-819; RRID:AB_2144105 STAT3 Santa Cruz Biotechnology Cat# sc-482; RRID:AB_632440 Rat IgG2a anti-PD-1 clone RMP1-14 BioXCell Cat# BE0146; RRID:AB_10949053 Rat IgG2a isotype control clone 2A3 BioXCell Cat# BE0089; RRID:AB_1107769 CD206 Abcam Cat# ab64693; RRID:AB_1523910 CD3 anti-mouse Abcam Cat# ab134096 CD4 anti-mouse, clone I3T4 Abcam Cat# ab183685; RRID:AB_2686917 FoxP3 anti-human, clone 236A/E7 Abcam Cat# ab20034; RRID:AB_445284 P-H2A.X (S139) Abcam Cat# ab2893; RRID:AB_303388 CD8a anti-mouse, clone Ly2 Invitrogen Cat# 14-0808-82; RRID:AB_2572861 F4/80 anti-mouse, clone BM8 Invitrogen Cat# MF48000; RRID:AB_10376289 FoxP3 anti-mouse, clone FJK-16s Invitrogen Cat# 14-5773-82; RRID:AB_467576 CD4 anti-human, clone 11E9 Novocastra Cat# NCL-L-CD4-368; RRID:AB_563559 Biological Samples Human melanoma pre-/post-therapy Paul Lorigan, Richard Marais N/A Chemicals, Peptides, and Recombinant Proteins PLX4720 Selleck #S1152 PLX4032 Selleck #S1267 GSK2118436 Dabrafenib ChemieTek #CT-DABRF GSK1120212 Trametinib Selleck #S2673 PD184352 Selleck #S1020 AZD6244 Selleck #S1008 SCH772984 Selleck #S7101 GSK269962A Axon MedChem # Axon 1167 H1152 Calbiochem #555550 AT13148 Selleck #S7563 (±)-Blebbistatin Calbiochem #203390 Critical Commercial Assays Human/Mouse TGF-b1 ELISA Biolegend #436707 Trichrome Stain (Masson) Kit Sigma #HT15-1KT Bouin’s solution Sigma #HT10132 Weigert’s iron hematoxylin solution Sigma # HT1079 (Continued on next page) e1 Cancer Cell 37, 85–103.e1–e9, January 13, 2020 Continued REAGENT or RESOURCE SOURCE IDENTIFIER Deposited Data Mass spectrometry A375 MEKi 24h This study ProteomeXchange via PRIDE repository Project # PXD002621 (https://www.ebi.ac. uk/pride/archive/projects/PXD002621) Experimental Models: Cell Lines Human: A375 ATCC ATCC Cat# CRL-7904; RRID:CVCL_0132 Human: Colo829 ATCC ATCC Cat# CRL-1974; RRID:CVCL_1137 Human: SKMEL5 ATCC ATCC Cat# HTB-70; RRID:CVCL_0527 Human: WM88 Coriell Institute Coriell Cat# WC00123; RRID:CVCL_6805 Human: WM983A Coriell Institute Coriell Cat# WC00048; RRID:CVCL_6808 Human: WM983B Coriell Institute Coriell Cat# WC00066; RRID:CVCL_6809 Human: WM793B Coriell Institute Coriell Cat# WC00062; RRID:CVCL_8787 Human: A375M2 Richard Hynes Clark et al., 2000 Human: LOX-IMVI Øystein Fodstad RRID:CVCL_1381 Human: D04 Kevin Harrington RRID:CVCL_H604 Human: MM485 Wellcome Trust Functional Genomics RRID:CVCL_2610 Cell Bank Human: A375/PLX/R Richard Marais RRID:CVCL_IW10 Baenke et al., 2015 Human: Colo829/PLX/R Richard Marais RRID:CVCL_IW11 Baenke et al., 2015 Human: A375/D+T/R Richard Marais N/A Human: Patient #2 Richard Marais N/A Human: Patient #35 Richard Marais N/A Human: Patient #62T3 Richard Marais N/A Human: Patient #58 Richard Marais N/A Human: Patient #33 Richard Marais N/A Mouse: 5555 Richard Marais Dhomen et al., 2009; Hirata et al., 2015 Mouse: 5555-anti-PD-1/NR This paper N/A Mouse: 4434 Richard Marais Dhomen et al., 2009; Hirata et al., 2015 Mouse: 4599 Richard Marais Dhomen et al., 2009; Hirata et al., 2015 Mouse: 690cl2 Richard Marais Dhomen et al., 2009 Human: HEK293T Jeremy Carlton ATCC Cat# CRL-3216; RRID:CVCL_0063 Experimental Models: Organisms/Strains Mouse: CD-1 nu/nu Charles River UK RRID:IMSR_CRL:086 Mouse: NOD/SCID/IL-2Rg-/- (NSG) Charles River UK RRID:IMSR_JAX:005557 Mouse: C57BL/6J Charles River UK RRID:IMSR_JAX:000664 Oligonucleotides See Table S7 for RNAi sequences Dharmacon Recombinant DNA pEGFP-N3-EGFP Fernando Calvo Takara, Clontech #U57609 pEGFP-N3-MLC2-EGFP (rat MLC2) Fernando Calvo Calvo et al., 2013 pEGFP-N3-MLC2-TASA (T18A Fernando Calvo Calvo et al., 2013 S19A) -EGFP pEGFP-N3-MLC2-TDSD (T18D Fernando Calvo Calvo et al., 2013 S19D) -EGFP pLVX-EGFP Erik Sahai, Tohru Takaki Takara, Clontech #632164 pLVX-MLC2-EGFP Erik Sahai, Tohru Takaki Takaki et al., 2017 (Continued on next page) Cancer Cell 37, 85–103.e1–e9, January 13, 2020 e2 Continued REAGENT or RESOURCE SOURCE IDENTIFIER pLVX-MLC2-TASA (T18A S19A)-EGFP Erik Sahai, Tohru Takaki Takaki et al., 2017 pLVX-MLC2-TDSD (T18A S19A)-EGFP Erik Sahai, Tohru Takaki Takaki et al., 2017 Software and Algorithms GSEA, ssGSEA Broad Institute http://www.broadinstitute. N/A org/gsea/index.jsp ImageJ https://imagej.nih.gov/ij/ N/A GraphPad Prism 8 GraphPad Software N/A SPSS IBM N/A LEAD CONTACT AND MATERIALS AVAILABILITY Further information and reasonable requests for resources and reagents should be directed to and will be fulﬁlled by the Lead Con- tact, Victoria Sanz-Moreno (v.sanz-moreno@qmul.ac.uk). All unique/stable reagents generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement. EXPERIMENTAL MODEL AND SUBJECT DETAILS Patient-Derived Samples Human melanoma samples were a kind gift from Paul Lorigan (University of Manchester). Tumor samples were collected under the Manchester Cancer Research Centre (MCRC) Biobank ethics application #07/H1003/161+5 with full informed consent from the pa- tients. The work presented in this manuscript was approved by MCRC Biobank Access Committee application 13_RIMA_01. Patient sample information is in Table S6. Cell Lines and Patient-Derived Cell Lines Cell lines used are listed in the Key Resources Table. Cell lines were cultured under standard conditions in complete medium (DMEM or RPMI medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (all from Gibco)). Cell lines were V600E tested to be free from mycoplasma contamination. All melanoma cell lines used were BRAF unless otherwise stated. A375, Colo829 and SKMEL5 cells were from ATCC. WM88, WM983A, WM983B, WM793B were purchased from Coriell Institute. A375M2 were from Dr Richard Hynes (HHMI, MIT, USA). LOX-IMVI cell line was a gift from Prof Øystein Fodstad (Oslo University Hospital). SKMEL5, WM983A, WM983B, WM793B, LOX-IMVI were grown in complete RPMI, WM88 was grown in complete DMEM. PLX4720-resistant WM983A, WM983B and WM88 cells were derived after exposure to PLX4720 for 2-3 months (1 mM PLX4720 for WM983A and WM983B; 0.5 mM PLX4720 for WM88), controls were treated with equivalent volume of DMSO. PLX4720-resistant (A375/PLX/R, Colo829/PLX/R) (Baenke et al., 2015) and dabrafenib+trametinib-resistant (A375/D+T/R) cell lines were a kind gift from Richard Marais (Cancer Research UK Manchester Institute). Resistant cells were generated after exposure of parental A375 and Colo829 to increasing concentrations of drugs (up to 1 mM PLX4720; 1 mM dabrafenib plus 10 nM trametinib) until cells resumed growth. Cells were grown in complete DMEM (A375-derivatives) or complete RPMI (Colo829-derivatives) supple- mented with 1 mM PLX4720 (A375/PLX/R and Colo829/PLX/R cells); 1 mM dabrafenib plus 10 nM trametinib (A375/D+T/R) or equiv- alent volume of DMSO (parental A375 and Colo829 cells). Patient-derived melanoma cell lines (#2, #35, #62T3, #58, #33) were a very kind gift from Richard Marais and were grown in RPMI. Patient #2, #35, #62T3 were grown in complete RPMI supplemented with 1 mM PLX4720. Patient #2 cell line was estab- lished from a patient with stage IV BRAF mutant melanoma with primary resistance to vemurafenib and ipilimumab. Patient #/35 cell line was established from a lymph node metastasis after treatment with vemurafenib for 3 months. Patient #62T3 cell line was established from a resected tumor upon disease progression following vemurafenib treatment (acquired resistance) and immunotherapy (refractory to ipilimumab and subsequent pembrolizumab). Patient #58 cell line (wild-type for BRAF/NRAS) was established from a metastasis from a patient that never responded to ipilimumab treatment (3 months). Patient #33 cell K601E line (BRAF ) was established from a metastasis from a patient that never responded to ipilimumab treatment (1 month). Patient #58 and #33 had also been treated with dacarbazine (DTIC) before ipilimumab. Patient #26 cell lines were established before and after nivolumab treatment. V600E Braf mouse melanoma cell lines 5555, 4434, 4599 and 690cl2 (from Richard Marais) were established from the following +/LSL-V600E +/o INK4a-/- +/LSL-V600E +/o C57BL/6 mouse models: Braf ;Tyr::CreERT2 ;p16 (5555, 4434); Braf ;Tyr::CreERT2 (4599); Pten-null +/LSL-V600E +/o INK4a-/- Braf ;Tyr::CreERT2 ;p16 ;Pten-/- (690cl2) (Dhomen et al., 2009; Hirata et al., 2015). NRAS mutant cell lines used: D04 was from Kevin Harrington (The Institute of Cancer Research); MM485 was obtained from the Wellcome Trust Functional Ge- nomics Cell Bank (UK). HEK293T cells were from Jeremy Carlton (The Francis Crick Institute). e3 Cancer Cell 37, 85–103.e1–e9, January 13, 2020 A375, A375/PLX/R, Colo829, Colo829/PLX/R, SKMEL5 cells and Patient-derived cell lines were conﬁrmed by STR proﬁling at CRUK Manchester Institute; A375M2, WM983A, WM983B at King’s College London; WM88 and WM793B cells were purchased from Coriell Institute in June 2014. Animals All animals were maintained under speciﬁc pathogen-free conditions and handled in accordance with the Institutional Committees on Animal Welfare of the UK Home Ofﬁce (The Home Ofﬁce Animals Scientiﬁc Procedures Act, 1986). All animal experiments were approved by the Ethical Review Process Committees at Barts Cancer Institute, King’s College London and The Francis Crick Institute, in accordance with the Animals (Scientiﬁc Procedures) Act 1986 and according to the guidelines of the Committee of the National Cancer Research Institute. Animals used in this study were from Charles River UK: 5-week-old female nude CD-1 nu/nu mice; 5-8-week old NOD/SCID/ IL-2Rg-/- (NSG) mice (male and female); 5-7-week-old female C57BL/6J mice. Tumors were allowed to establish, sizes (average 60-100 mm ) were matched and then mice were randomly allocated to groups of 6-8 animals. No blinding was used in the treatment schedules for these experiments since the different treatments were identiﬁed by ear notching/mark on tail. Based on previous studies in the literature (Hong et al., 2017; Kong et al., 2017) and our own experience, groups of 6-8 animals were used to have sufﬁcient animals per group to provide statistically signiﬁcant data while keeping the number of animals used to a minimum. Tumor size was determined by caliper measurements of tumor length, width and depth and tumor volume was calcu- lated as volume = 0.5236 x length x width x depth (mm). Note that this formula calculates smaller tumors (approximately 2-fold smaller) compared to those calculated using the formula volume = 0.5236 x length x width (mm). METHOD DETAILS Chemicals Chemicals used in this study (stocks resuspended in DMSO unless otherwise stated): BRAFi PLX4720 and PLX4032 (Selleck), BRAFi Dabrafenib (GSK2118436, ChemieTek), MEKi Trametinib (GSK1120212, Selleck), MEKi PD184352 (Selleck), MEKi AZD6244 (Selleck), ERKi SCH772984 (Selleck), ROCKi GSK269962A (Axon Medchem), ROCKi H1152 (resuspended in water; Calbiochem), AGC kinase inhibitor and ROCKi AT13148 (Selleck), myosin II inhibitor blebbistatin (in 95% DMSO; Calbiochem). Concentrations used unless otherwise stated in other STAR Methods sections: 5 mM ROCKi GSK269962A, 5 mM ROCKi H1152, 5 mM ROCKi AT13148, 25 mM myosin II inhibitor blebbistatin, 5 mM BRAFi PLX4720. ‘‘Analysis of cell morphology’’ section lists the inhibitors and concentrations used for those experiments. Antibodies Antibodies and concentrations used: pThr18/Ser19-MLC2 (#3674; 1:750, immunoblot), pSer19-MLC2 (#3671; 1:50, immunohisto- chemistry; 1:200, immunoﬂuorescence), MLC2 (#3672; 1:750), pT202/Y204-p44/42 (ERK1/2) (#4370; 1:1,000), pY705-STAT3 (#9145; 1:750), PD-L1 (clone E1L3N, #13684, 1:200) from Cell Signaling Technology; STAT3 (sc-482; 1:500), ERK2 (sc-154; 1:1,000), MCL-1 (sc-819; 1:1,000), GFP (sc-8334; 1:1,000) from Santa Cruz Biotechnology; GAPDH (MAB374; 1:10,000) from Milli- pore; P-H2A.X (S139) (ab2893;1:1000), CD206 (ab64693; 1:1,000), CD3 (anti-mouse, ab134096; 1:500), CD4 (anti-mouse, clone I3T4, ab183685; 1:300), FoxP3 (anti-human, clone 236A/E7, ab20034; 1:200) from Abcam; F4/80 (anti-mouse, clone BM8, MF48000, 1:1000), CD8a (anti-mouse, clone Ly2, 14-0808-82; 1:200), FoxP3 (anti-mouse, clone FJK-16s, 14-5773-82; 1:200) from Invitrogen; CD4 (anti-human, clone 11E9, NCL-L-CD4-368; 1:300) from Novocastra. Analysis of Cell Morphology Cell morphology was analyzed on still phase-contrast images (cells on plastic or on collagen I) using ImageJ software (http://rsb.info. nih.gov/ij/). In order to quantify cell morphology on 2D and on collagen matrices, the morphology descriptor Circularity was used after manually drawing around the cell. Values closer to 1 represent rounded morphology; values closer to 0 represent more spread and/or spindle-shaped cells with multiple protrusions. Treatments were for 24 hr as follows: A375M2 cells with 50 nM BRAFi PLX4720, 0.1 nM MEKi GSK1120212, 1 mM ROCKi GSK269962A (Figure 1B); WM983A/B cells with 5 mM ROCKi GSK269962A, 5 mM BRAFi PLX4720 (Figures 1D and 1E); 690cl2 cells with 200 nM MEKi PD184352, 200 nM BRAFi PLX4032, 500 nM ERKi SCH772984 (Figures 1F and S1C); D04, MM485 cells with 50 nM MEKi GSK1120212, 50 nM AZD6244 (Figures 1F, S1D, and S1E); 4599 cells with 500 nM MEKi GSK1120212, 1 nM MEKi AZD6244 (Figure S1B). A375 and A375/PLX/R on plastic (Figure 3D); and Patient #2 cells on collagen I (Figure 5I) were treated with 5 mM ROCKi GSK269962A, 5 mM BRAFi PLX4720 or both. Long-Term Survival Long-term survival was performed on tissue culture plastic dishes unless otherwise speciﬁed. Cells were seeded in 6-well plates (10,000 cells/well) and treated for 5-14 days, re-adding drugs in fresh media every 2-3 days (daily for blebbistatin). Then cells were ﬁxed with 1% formaldehyde and stained with 0.25% crystal violet. Plates were scanned and images analyzed using ImageJ software. For experiments with inhibitors, percentage of the well covered by crystal violet-stained cells was calculated and shown relative to control cells. For dose-response experiments, cells were seeded in 12-well or 96-well plates and survival was analyzed Cancer Cell 37, 85–103.e1–e9, January 13, 2020 e4 after 3-5 days treatment with indicated drugs using crystal violet. Crystal violet was solubilized with 10% acetic acid and absorbance was measured at 590 nm. In dose-response experiments, BRAFi-resistant cells were cultured in the presence of BRAFi throughout the experiment unless otherwise stated. In Figure 5C, 4434- and 5555-derivatives were treated with 0.1 mM ROCKi. For synergy experiments, 1,000 A375 cells were seeded in 96-well plates, cultured overnight and next day treated in quintuplicates with ROCKi GSK269962A or BRAFi PLX4720, either alone or in several combinations in complete medium. Three days later, plates were ﬁxed, stained with crystal violet and solubilized and quantiﬁed as above. Values were normalized to vehicle controls and analyzed with Combeneﬁt software (Loewe model) (Di Veroli et al., 2016). Average of 4 independent experiments is shown. Long Term Survival on Collagen I Matrices Cells were grown on collagen I matrices as described (Orgaz et al., 2014b). Brieﬂy, bovine collagen I (PureCol, #5005-B; Advanced BioMatrix) thick gels were polymerized at 1.7 mg/ml in 24-well plates. Cells were seeded at 10,000 cell/well and treatments started 16 hr later for 5-14 days. In experiments using A375-derivatives, cells were treated with 1 mM ROCKi, 1 mM BRAFi or both. Patient- derived cell lines were treated with 5 mM ROCKi. Fresh complete media with drugs was added every 2-3 days. At the end of the exper- iment collagen I gels were ﬁxed with 4% formaldehyde and phase-contrast images were taken. Percentage of area covered by cells was quantiﬁed using QuPath software Version 0.1.2 and a SLIC superpixel image segmentation (Gaussian sigma value 5 pixels, superpixel spacing 20 pixels) (Bankhead et al., 2017). Software was trained to identify cells and background (surrounding collagen). Detection measurements were then exported to Excel and values for area/pixel were normalized to each untreated control as per- centage of area covered by cells. For Patient #2 cells, spheroid-forming ability was quantiﬁed as the sum of areas occupied by spher- oids from phase-contrast images using ImageJ. MTT Assay Cells were seeded in 96-well plates (2,000 cells/well). Drugs were added every 2 days. Three days after seeding, plates were incu- bated with MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; Millipore) following the manufacturer’s instructions and absorbance measured at 572 nm. Background at 630 nm was subtracted and data represented as relative viability. Cell Cycle Analysis For DNA cell cycle analysis, ﬂoating and adherent cells were ﬁxed in 70% ethanol at -20 C, washed in phosphate-buffered saline and treated with 40 mg/ml propidium iodide (PI) (Biolegend) and 100 mg/ml ribonuclease (Sigma) for 25 min at 37 C. Staining was detected using a FACS BD Canto II (BD Biosciences) and analyzed and plotted using FlowJo (FlowJo LLC). The starting gating of the whole cell population, excluding any debris, was performed with FSC-A/SSC-A. Using this as a parental gate, doublets were excluded using PerCP-Cy5.5-A/ PerCP-Cy5.5-W (PI). The gated singlets were represented as histograms for PerCP-Cy5.5-A to show the peaks for the cell cycle phases. AnnexinV/Propidium Iodide FACS Floating and adherent cells were collected, spun down, and labelled with FITC Annexin V Apoptosis Detection Kit with PI (#640914, Biolegend UK Ltd), following the manufacturer’s instructions. Staining was detected using a FACS BD Canto II and analyzed and plotted using FlowJo. The starting gating of the whole cell population, excluding any debris, was performed with FSC-A/SSC-A. This was followed by a double exclusion of doublets using ﬁrst FSC-H/FSC-W and then SSC-H/SSC-W. The gated singlets were then gated as ‘quad gates’ using FITC-A (AnnexinV) versus PerCP-Cy5.5-A (PI) and represented as FACS dot plots. Graphs show high low percentage of dead cells as the sum of percentage of early apoptotic (annexin V , propidium iodide ) and percentage of late high high apoptotic/necrotic cells (annexin V , propidium iodide ). ROS Detection Cells were treated with 1 mM ROCKi for 24 hr (A375 pair) or 48 hr (WM983A pair). Then cells were collected and ROS levels were detected using CellROX Green Flow Cytometry Assay Kit (C10492, Life Technologies), according to the manufacturer’s instructions. FACS and gating strategy were as described in Cell Cycle section. Time Lapse Microscopy Multi-site bright-ﬁeld microscopy of cells in 24-well plates was performed in a humidiﬁed chamber at 37 C and 5% CO using a 10X/0.3 NA Plan Fluor ELWD objective lens on a fully motorized (Prior Scientiﬁc) multi-ﬁeld Nikon TE2000 microscope with an ORCA camera (Hamamatsu) controlled by Micro-Manager (https://micro-manager.org/) and ImageJ. Sixteen hr after seeding, cells were treated for 72 hr with ROCKi, BRAFi or both in the presence of 1.5 mM PI to identify dead cells. Total number of cells, per- centage of multinucleated (alive, dead) and total dead cells were quantiﬁed for 72 hr. RNAi One hundred thousand cells were plated per 35-mm dish and transfected the next day with 26 nM siRNA oligonucleotides, using Optimem-I and Lipofectamine 2000 (Invitrogen). Forty eight hr after transfection cells were harvested and equal numbers re-seeded on 35-mm wells. Cells were transfected again 2 days later and plates were ﬁxed and stained with crystal violet 2-4 days after the second transfection. Crystal violet was solubilized and absorbance at 590 nm measured as above. Cells were grown in the presence e5 Cancer Cell 37, 85–103.e1–e9, January 13, 2020 of 1 mM PLX4720 during the whole experiment. All siRNA sequences were On-Targetplus (OT) from Dharmacon (Lafayette, USA) and are listed in Table S7. MLC2 Rescue Experiments One hundred thousand cells were plated per 35-mm dish and transfected the next day with Lipofectamine 2000 and 1 mg plasmid encoding GFP (as control), wild-type rat MLC2 (MYL12B) fused with GFP or inactive phospho-mutant TASA-MLC2 fused with GFP (T18A, S19A) (Calvo et al., 2013) (plasmids were a gift from Fernando Calvo). Next day cells were transfected with 26 nM siRNA oligonucleotides against MYL12B. Cell death was assessed 2-3 days after siRNA transfection by PI (1.5 mM) incorporation by + + FACS. Percentage of dead cells (PI ) was quantiﬁed within transfected (GFP ) cells. MLC2 Stable Overexpression Lentivectors encoding EGFP-fused rat MLC2-derivatives (wild-type, phospho-mimetic TDSD (T18D, S19D) and inactive phospho- mutant TASA (T18A, S19A)) (Takaki et al., 2017) were a kind gift from Erik Sahai and Tohru Takaki (The Francis Crick Institute). HEK293T cells were transfected with MLC2-lentivectors and packaging plasmids using standard procedures, and after 48 hr super- natants were collected and spun down to remove debris. A375 cells were transduced with lentiviral supernatants for 8 hr, and 48 hr later cells were selected with 1 mg/ml puromycin for 5 days, then cells were used for subsequent experiments. Immunoﬂuorescence and Confocal Imaging Cells were ﬁxed with 4% formaldehyde, permeabilised with 0.2% Triton X-100 for 5 min, blocked with 5% BSA-PBS for 1 hr at room temperature, and incubated with anti-p-MLC2 (p-MLC2S19, 1:200 in 5% BSA-PBS) overnight at 4 C. Alexa-488 anti-rabbit second- ary antibody (Life Technologies) was used at 1:500 for 1 hr at room temperature. F-actin was detected with Phalloidin (1 hr RT) and nuclei were stained with Hoechst 33258 (Life Technologies). Imaging was carried out on a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss) with C-Apochromat 3 40/1.2 NA (water) or a Plan Apochromat 3 63/1.4 NA (oil) objective lenses and Zen software (Carl Zeiss). Line scan analysis was performed in ImageJ. Immunoblotting Cells were lysed in Laemmli buffer and snap frozen. Lysates were then boiled, sonicated for 15 s and spun down. Cell lysates were fractionated using sodium dodecyl sulfate-polyacrylamide (SDS-PAGE) gels in non-reducing conditions, and transferred subse- quently to PVDF ﬁlters. Membranes were blocked in 5% BSA in 0.1% Tween 20-TBS. Primary antibodies were incubated overnight at 4 C. For detection, ECL or Prime ECL detection systems coupled to HRP-conjugated secondary antibodies (GE Healthcare) with X-ray ﬁlms and an Amersham Imager 600 were used. Bands were quantiﬁed using ImageJ. Levels of phospho-proteins were calcu- lated after correction to total levels of the relevant protein. TGF-b1 ELISA Cells were seeded on T6-well plates (150,000 cells/well), next day cells were washed 3 times and then grown in serum-free media with or without ROCKi GSK269962A (5 mM). Forty-eight hr later supernatants were collected, spun down and assayed fresh or frozen TM at -80 C. TGF-b1 levels were detected by ELISA using Total TGF-b1 Legend Max ELISA Kit with Pre-coated plate (#436707, Bio- legend) on neat samples diluted 1/5 following the manufacturer’s instructions. Phospho-proteomics Preparation of tandem mass tagged (TMT)-multivariate phosphoproteomic samples. Cells treated with MEKi (200 nM GSK1120212 trametinib or 200 nM PD184352) or vehicle (DMSO) for 24 hr were lysed in 6 M urea, sonicated, centrifuged to clear cell debris and protein concentration was determined by BCA (Pierce 23225). 100 mg of each condition was individually digested by FASP (PMID: 19377485) using 1:100 Lys-C (Wako 125-05061), 1:100 Trypsin (Worthington), and amine-TMT-10 plex labeled (Pierce 90111) on membrane (iFASP) (PMID: 23692318). TMT channel assignment: 126 = Control (Bio. Rep. 1); 127N = Control (Bio. Rep. 2), 127C = Control (Bio. Rep. 3); 128N = Control (Bio. Rep. 4); 128C = MEKi A (Bio. Rep. 1); 129N = MEKi A (Bio. Rep. 2); 129C = MEKi A (Bio. Rep. 3); 130N = MEKi B (Bio. Rep. 1); 130C = MEKi B (Bio. Rep. 2); 131 = MEKi B (Bio. Rep. 3) (A= GSK1120212, B= PD184352). Peptides were then eluted, pooled, lyophilized and subjected to automated phosphopeptide enrichment (APE) (PMID: 25233145). Phosphopeptides were desalted using OLIGO R3 resin (Life Technologies 1-1339-03) and lyophilised prior to LC-MS/MS analysis (see below). Data-dependent acquisition LC-MS/MS. Phosphopeptide samples were resuspended in 0.1% formic acid and analyzed on a Q-Exactive Plus mass spectrometer (Thermo Scientiﬁc) coupled to a Dionex Ultimate 3000 RSLCnano System (Thermo Scientiﬁc). Reversed-phase chromatographic separation was performed on a C18 PepMap 300 A trap cartridge (0.3 mm i.d. x 5 mm, 5 mm bead size; loaded in a bi-directional manner), a 75 mm i.d. x 50 cm column (5 mm bead size) using a 120 min linear gradient of 0-50% solvent B (MeCN 100% + 0.1% formic acid (FA)) against solvent A (H2O 100% + 0.1% FA) with a ﬂow rate of 300 nL/min. The mass spec- trometer was operated in the data-dependent mode to automatically switch between dual Orbitrap MS and MS/MS acquisition. Sur- vey full scan MS spectra (from m/z 400-2000) were acquired in the Orbitrap with a resolution of 70,000 at m/z 400 and FT target value of 1 x 106 ions. The 20 most abundant ions were selected for fragmentation using higher-energy collisional dissociation (HCD) and dynamically excluded for 30 s. Fragmented ions were scanned in the Orbitrap at a resolution 35,000 at m/z 400. The isolation window Cancer Cell 37, 85–103.e1–e9, January 13, 2020 e6 was reduced to 1.2 m/z (to reduce ion co-isolation) and a MS/MS ﬁxed ﬁrst mass of 120 m/z was used (to ensure consistent TMT reporter ion coverage). For accurate mass measurement, the lock mass option was enabled using the polydimethylcyclosiloxane ion (m/z 445.120025) as an internal calibrant. For peptide identiﬁcation, raw data ﬁles produced in Xcalibur 2.1 (Thermo Scientiﬁc) were processed in Proteome Discoverer 1.4 (Thermo Scientiﬁc) and searched against Human Unitprot database using Mascot (v2.2). Searches were performed with a precursor mass tolerance set to 10 ppm, fragment mass tolerance set to 0.05 Da and a maximum number of missed cleavages set to 2. Static modiﬁcations were limited to carbamidomethylation of cysteine, and variable modiﬁcations used were oxidation of methionine, deamidation of asparagine/glutamine, and phosphorylation of serine, threonine and tyrosine residues. Peptides were further ﬁltered using a mascot signiﬁcance threshold <0.05, a peptide ion Score >20 and a FDR <0.01 (evaluated by Percolator (PMID: 17952086)). Phospho-site localization probabilities were calculated with phosphoRS 3.1 (>75%) (PMID: 22073976). For relative phosphopeptide quantiﬁcation, MEKi/vehicle ratios were calculated by Proteome Discoverer 1.4. See Data and Code Availability section below for further details. Phosphoproteomic data analysis. Phosphopeptides from Proteome Discoverer 1.4 were normalised against total protein levels (from SILAC in-gel digest experiments), and protein-level phospho-site locations (phosphoRS 3.1 score >75%, maximum 4-PTM/ peptide) were manually annotated using PhosphoSitePlus. Precursor ion spectra, extracted ion chromatograms, and product ion spectra were manually inspected for each regulated phosphopeptide. Empirical parent kinases were manually identiﬁed by refer- enced Uniprot annotation and putative parent kinases were manually assigned using ScanSite (PMID: 12824383) 3 (top 1 percentile of all sites, lowest score). Phospho-sites that did not meet these conditions were not annotated. Regulated phospho-peptides in Table S1 were those which were signiﬁcant across both MEKi (GSK1120212 and PD184352) compared to vehicle-treated cells. Phospho-Peptide Enrichment Analysis Pathway enrichment analyzes of the list of phospho-peptides increased in MEKi-treated A375 compared to vehicle-treated A375 cells (this study, see Phospho-proteomics section; Table S1); A375/PLX/R compared to A375 cells (data from (Girotti et al., 2013)) and M229- and M238-vemurafenib-resistant vs parental cells from (Titz et al., 2016) were performed using MetaCore from GeneGo Inc. (https://portal.genego.com/). Quantitative Real Time One-Step PCR RNA was isolated using TriZol (Life technologies). For experiments comparing expression in parental vs BRAFi-resistant cells (A375- and Colo829-derivatives), resistant cells were cultured with 1 mM PLX4720 and sensitive cells with equivalent volume of DMSO for 24 hr. QuantiTect Primer Assays (Qiagen) and Brilliant II SYBR Green QRT-PCR 1-step system (Agilent Technologies) with 100 ng RNA were used following the manufacturer’s instructions. GAPDH was used as loading control. The following QuantiTect Primers were used (Qiagen): GAPDH (QT00079247), LIMK1 (QT00008680), LIMK2 (QT00084357), MKL1 (QT00067921), MKL2 (QT00010115), MYH9 (QT00073101), MYL9 (QT00072268), MYL12A (QT01665741), MYL12B (QT00075264), ROCK1 (QT00034972), ROCK2 (QT00011165). Primer sequences are not provided by Qiagen, as stated in their website: ‘Sequences of the QuantiTect Primer Assays are not provided. Approximate location of primers within a speciﬁc gene can be viewed on the Product Detail pages retrieved via our GeneGlobe data base.’ Gene Expression Studies and Analysis Normalized gene expression microarray and RNAseq (FPKM, fragments per kilobase of transcripts per million mapped reads) data from published studies were downloaded from Gene Expression Omnibus (GEO) unless otherwise stated: Hugo 2015 (GSE65185 and GSE65184) (Hugo et al., 2015); Hugo 2016 (GSE78220) (Hugo et al., 2016); Kakavand 2017 (GSE99898) (Kakavand et al., 2017); Kwong 2015 (European Genome-phenome Archive (EGA S00001000992)) (Kwong et al., 2015); Long 2014 (GSE61992) (Long et al., 2014a); Obenauf 2015 (GSE64741) (Obenauf et al., 2015); Rizos 2014 (GSE50509) (Rizos et al., 2014); Riaz 2017 (Ipi-naive cohort; GSE91061) (Riaz et al., 2017); Song 2017 (GSE75299, GSE103630) (Song et al., 2017); Sun 2014 (GSE50535) (Sun et al., 2014); Wagle 2014 (GSE77940) (Wagle et al., 2014). In patients with several biopsies, their average is shown (see Table S4). RSEM-normalized expression data and clinical information of human melanoma samples (70 primary and 319 metastatic mela- nomas) from The Cancer Genome Atlas (TCGA) database were downloaded from Firehose (https://gdac.broadinstitute.org/). Only TCGA samples with no neo-adjuvant treatment prior to tumor resection were considered. The ROCK-myosin II pathway expression signature (MYL9, MYL12A, MYL12B, MYH9, ROCK1, ROCK2, LIMK1, LIMK2, MKL1, MKL2, MYLK, DAPK3) was generated by the sum of normalized expression values of signature genes for each TCGA patient. ROCK-myosin II pathway signature was categorized as low or high using the mean expression. Heatmaps and unsupervised hierarchical clustering analyzes were generated using Multiexperiment Viewer (http://www.tm4.org/ mev.html). Distance metric used for the clustering was Euclidean distance. In patients with several biopsies, their average is shown. Gene Enrichment Analyzes Gene sets for cross-resistance processes (EMT, metastasis, angiogenesis, hypoxia, wound healing, TGF-b, STAT3, NF-kB, YAP) were downloaded and analyzed using Gene Set Enrichment Analysis (GSEA) software (http://www.broadinstitute.org/gsea/index. jsp) with the settings: permutations-1,000, permutation type-gene set, metric for ranking genes-t-test. Signiﬁcantly enriched gene sets in resistant vs baseline samples were considered according to p value <0.05 and FDR <0.25 in at least 2 of the 5 comparisons e7 Cancer Cell 37, 85–103.e1–e9, January 13, 2020 performed. To calculate the gene-signature score in each sample, we used single-sample Gene Set Enrichment Analysis (ssGSEA) Projection Software from GenePattern platform (https://www.broadinstitute.org/cancer/software/genepattern). For the transcriptional signature of melanoma cells with high myosin II activity, genes upregulated in high myosin II activity compared to low myosin II activity melanoma cells (cells treated with ROCKi and blebbistatin) (Cantelli et al., 2015; Sanz-Moreno et al., 2011) were selected using a fold changeR 1.5 and a p value <0.01. GSEA analysis was performed as described above. Enrich- ment plot (green line) show upregulation of gene signature in indicated samples (resistant, non-responders or on-treatment). Nominal p values are shown along plot, false discovery rate (FDR) in ﬁgure legend. For analysis of ROS-related gene signatures, all available ROS/oxidative stress gene sets were downloaded from GSEA Broad Institute (http://www.broadinstitute.org/gsea/index.jsp). Graph shows (-Log ) p value. For analysis of expression of DNA repair genes, we compiled a DNA repair gene signature from the list in (Mjelle et al., 2015) and the homologous recombination defect signature (Peng et al., 2014). Network enrichment analysis of genes commonly downregulated (<0.65-fold) in at least 4 of 7 cell lines from Group 1 (Figure 2B) was performed using Ingenuity Pathway Analysis (Qiagen). Tumor Xenografts A375/PLX/R cells (1 x 10 ) were injected subcutaneously into the right ﬂank of 5-week-old female nude CD-1 mice (Charles River). 6 6 Patient #2 cells (4 x 10 ) or Patient #35 cells (6 x 10 ) were injected into 5-8-week old NOD/SCID/ IL-2Rg-/- (NSG, Charles River) mice (male and female). Tumors were allowed to establish, sizes (average 60-100 mm ) were matched and then mice were randomly allo- cated to groups of 7-8 animals. Treatment was by orogastric gavage with 45 mg/kg PLX4720, 10-25 mg/kg GSK269962A or both drugs together. GSK269962A was used at 25 mg/kg for A375/PLX/R and 10 mg/kg for Patient #2, #35. Drugs were dissolved in 5% DMSO or in 6% DMSO+50% PEG300+ 9% Tween 80. All the drugs were administered daily, 7 days a week. Tumor size was determined by caliper measurements of tumor length, width and depth and tumor volume was calculated as volume = 0.5236 x length x width x depth (mm). Immunotherapy Experiments 5555 cells (100,000, 250,000 or 1 million) were subcutaneously injected into the right ﬂank of 5-7-week-old female C57BL/6J mice. After 7-14 days, mice with tumors (50-80 mm ) were randomly allocated into groups of 6-7 animals and treated daily with ROCKi GSK269962A (10 mg/kg, oral gavage) or vehicle and every 3 days with anti-PD-1 monoclonal antibody (InVivoPlus clone RMP1-14, BioXCell #BE0146) (10 mg/kg, intraperitoneally (i.p.)) or rat IgG2a isotype control (clone 2A3 BioXCell # BE0089). Vehicle for ROCKi was 5% DMSO or 5% DMSO, 10% Tween 80, 6.5% ethanol. Tumor volume was determined as above. Anti-PD-1-non- responder (NR) lines were established in culture by digesting tumors with a mixture of Liberases (TH and TM, 75 mg/ml each, Roche Diagnostics) and 1 mg/ml DNase I (Sigma) in HBSS for 1 hr at 37 C with shaking, and then passed through 100 mM strainers. For ex- periments using 5555-anti-PD-1/NR cells, 1 million cells were injected subcutaneously into 7-week old C57BL/6J mice. Next day, all mice were given 1 dose of anti-PD-1 (10 mg/kg) i.p., and then again 3 days later. At day 7, mice were randomized into 4 treatment groups (ROCKi, anti-PD-1, combo or control) as above. Survival in the Lung Assay Patient #2 cells were pre-treated for 24 hr with 5 mM PLX4720, 5 mM GSK269962A or both (control had DMSO), then cells were labelled with 10 mM CMFDA-Green in OptiMem (Life Technologies) for 10 min, trypsinized and equal numbers were injected into the tail vein of NSG mice in 100 ml PBS along with drugs (same concentrations as pre-treatment). At the time of injection, mice (male and female) were 6-10 weeks old and weighed around 20-22 g; mice were age and sex-matched between the groups. Mice were sacriﬁced 30 min (to conﬁrm that equal numbers arrived at the lung) and 24 hr after tail vein injection. The lungs were ex- tracted, washed twice with PBS, ﬁxed (4% formaldehyde for 16 hr at 4 C) and examined for ﬂuorescently-labelled cells under a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss) with a 20X objective. Lung retention is represented as ﬂuorescence area (CMFDA- Green from melanoma cells) per ﬁeld, and approximately 20 ﬁelds per mouse lung were analyzed. Each experiment had 4-5 mice/ condition, and experiments were replicated twice and data pooled together. Quantiﬁcation of survival in the lung 24 hr after injection is shown as mean ﬂuorescence area/ﬁeld. Immunohistochemistry Tumors and spleens were formalin-ﬁxed and parafﬁn-embedded using standard protocols. For cell pellets, transfected cells were harvested 48 hr after transfection using a cell scraper, spun down, ﬁxed with 4% formalin for 30 min and washed with PBS. Cell pellet was resuspended in 2% agarose and then embedded in parafﬁn. Four mm thick sections were incubated at 60 C for 20 min and then subjected to antigen retrieval using Access Super Tris pH 9 buffer (A.Menarini Diagnostics) at 110 C for 6 min in a Decloaking Cham- ber NxGen (Biocare Medical). Samples were blocked with Dual Endogenous Enzyme-Blocking Reagent (Dako) for 10 min and then were incubated with primary antibodies for 40 min at RT, washed and then incubated with biotinylated secondary antibodies (rabbit, mouse or rat; 1:200; Vector-Labs) for 30 min at RT. Signal was then ampliﬁed using VECTASTAIN ABC HRP kit (PK-4000) for 20 min at RT and the reaction was developed using VIP substrate (SK-4600, Vector-Labs) for 10 min at RT. Stainings were counter- stained with Hematoxylin. Positive and negative controls were included in each experiment, including staining of melanoma markers Cancer Cell 37, 85–103.e1–e9, January 13, 2020 e8 HMB45/Melan-A or S100. For ECM staining, samples were ﬁxed in Bouin’s solution (HT10132, Sigma) for 1 hr at 60 C, then stained with Weigert’s iron hematoxylin solution (HT1079, Sigma) for 5 min at RT and with Trichrome Stain (Masson) Kit (HT15-1KT, Sigma) following the manufacturer’s instructions. Imaging and Scoring Sections from tumor xenograft experiments and from paired melanoma samples from 12 patients (tumor tissue before and after treatment) were imaged using NanoZoomer S210 slide scanner (Hamamatsu, Japan). Staining quantiﬁcation was performed using QuPath 0.1.2 (Bankhead et al., 2017). For p-MLC2 stainings, whole sections were scanned and images were analyzed performing positive cell detection, and three different thresholds were applied according to the intensity scores (0, 1, 2 and 3). Next, the software was trained by creating random trees classiﬁcation algorithm combined with the intensity information, in order to differentiate tumor from stroma, necrosis and immune cells. Values used in the analysis correspond to the quantiﬁcation of p-MLC2 in the invasive front (mouse tumors) or highest score in the whole section (human samples). To characterize the immune inﬁltrate (CD206, F4/80, CD3, CD4, CD8 and FOXP3) a similar approach was performed using QuPath. First, positive cell detection was applied, using only a single value to differentiate negative (blue) from positive (red). Data are repre- sented as cellular density (cells/mm ). For PD-L1 analysis, CD206 cells were identiﬁed and both PD-L1 and CD206 stainings were aligned using QuPath 2.03m. From + + CD206 staining, positive detections (CD206 ) were transferred to PD-L1 in order to quantify the actual score for PD-L1 in CD206 cells. The negative detection for CD206 was used to quantify PD-L1 on tumor cells, these were identiﬁed as CD206 after discarding stromal/immune cells. Image composition was performed artiﬁcially attributing a color code, and images were overlaid using ImageJ (trackEM2). For PD-L1 and CD206, merge images in Figure 8I were generated with QuPath by overlaying pseudo-color images for each staining. For ECM analysis with Masson’s Trichrome staining, whole section images were quantiﬁed with QuPath applying a SLIC algorithm for segmentation of sections according to pixel density. Next, colors were deconvoluted and the green channel was used to quantify the percentage of the area occupied by collagen. QUANTIFICATION AND STATISTICAL ANALYSIS GraphPad Prism (GraphPad Software) was used to perform unpaired two-tailed t-test, Mann-Whitney test, Wilcoxon test, one-way or two-way ANOVA with post hoc tests (Tukey’s, Dunnet’s, Benjamini, Krieger and Yekutieli correction), Kruskal-Wallis, Deming linear regression, Spearman correlation and Chi-square test. Survival curves were estimated by the Kaplan-Meier method and the log-rank test using SPSS (IBM). Details of statistical analysis performed are in the ﬁgure legends. Bar graphs report mean ± SEM with indi- vidual data points as explained in ﬁgure legends. Box plots show median (center line); interquartile range (box); min-max values (whis- kers). In Figure legends, ‘‘n’’ means number of independent experiments unless otherwise stated. Signiﬁcance was deﬁned as p<0.05. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns not signiﬁcant. DATA AND CODE AVAILABILITY The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://www.proteomexchange. org) via the PRIDE partner repository (PMID: 23203882) with the dataset identiﬁer PXD002621 (https://www.ebi.ac.uk/pride/archive/ projects/PXD002621). e9 Cancer Cell 37, 85–103.e1–e9, January 13, 2020 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Cancer Cell Unpaywall http://www.deepdyve.com/lp/unpaywall/myosin-ii-reactivation-and-cytoskeletal-remodeling-as-a-hallmark-and-a-1fFnKO7z8G

Loading next page...

References (103)

S. Mariathasan, S. Turley, D. Nickles, A. Castiglioni, Kobe Yuen, Yulei Wang, E. Kadel, H. Koeppen, Jillian Astarita, R. Cubas, S. Jhunjhunwala, R. Banchereau, Yagai Yang, Y. Guan, Cecile Chalouni, James Ziai, Y. Şenbabaoğlu, Stephen Santoro, D. Sheinson, Jeffrey Hung, J. Giltnane, A. Pierce, Kathryn Mesh, Steve Lianoglou, J. Riegler, R. Carano, P. Eriksson, M. Hoglund, Loan Somarriba, D. Halligan, M. Heijden, Y. Loriot, J. Rosenberg, L. Fong, I. Mellman, Daniel Chen, Marjorie Green, C. Derleth, G. Fine, P. Hegde, R. Bourgon, T. Powles (2018)
TGF-β attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells
Nature, 554
C. Balch, J. Gershenwald, S. Soong, J. Thompson, M. Atkins, D. Byrd, A. Buzaid, A. Cochran, D. Coit, Shouluan Ding, A. Eggermont, K. Flaherty, P. Gimotty, J. Kirkwood, K. McMasters, M. Mihm, D. Morton, M. Ross, A. Sober, V. Sondak (2009)
Final version of 2009 AJCC melanoma staging and classification.
Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 27 36
Willy Hugo, J. Zaretsky, Lu Sun, Chunying Song, B. Moreno, S. Hu-Lieskovan, B. Berent-Maoz, Jia Pang, B. Chmielowski, Grace Cherry, E. Seja, Shirley Lomeli, X. Kong, M. Kelley, J. Sosman, Douglas Johnson, A. Ribas, R. Lo (2016)
Genomic and Transcriptomic Features of Response to Anti-PD-1 Therapy in Metastatic Melanoma
Cell, 165
C. Robert, B. Karaszewska, J. Schachter, P. Rutkowski, A. Mackiewicz, D. Stroiakovski, M. Lichinitser, R. Dummer, F. Grange, L. Mortier, V. Chiarion-Sileni, K. Drucis, I. Krajsová, A. Hauschild, P. Lorigan, P. Wolter, G. Long, K. Flaherty, P. Nathan, A. Ribas, Anne-Marie Martin, P. Sun, W. Crist, J. Legos, S. Rubin, S. Little, D. Schadendorf (2015)
Improved overall survival in melanoma with combined dabrafenib and trametinib.
The New England journal of medicine, 372 1
P. Lito, C. Pratilas, E. Joseph, M. Tadi, Ensar Halilovic, Matthew Zubrowski, A. Huang, W. Wong, M. Callahan, T. Merghoub, J. Wolchok, E. Stanchina, S. Chandarlapaty, Poulikos Poulikakos, J. Fagin, N. Rosen (2012)
Relief of profound feedback inhibition of mitogenic signaling by RAF inhibitors attenuates their activity in BRAFV600E melanomas.
Cancer cell, 22 5
K. Flaherty, I. Puzanov, Kevin Kim, A. Ribas, G. McArthur, J. Sosman, P. O'dwyer, Richard Lee, J. Grippo, K. Nolop, P. Chapman (2010)
Inhibition of mutated, activated BRAF in metastatic melanoma.
The New England journal of medicine, 363 9
A. Mantovani, A. Sica, S. Sozzani, P. Allavena, A. Vecchi, M. Locati (2004)
The chemokine system in diverse forms of macrophage activation and polarization.
Trends in immunology, 25 12
T. Takaki, Marco Montagner, Murielle Serres, M. Berre, Matthew Russell, L. Collinson, K. Szuhai, M. Howell, S. Boulton, E. Sahai, M. Petronczki (2017)
Actomyosin drives cancer cell nuclear dysmorphia and threatens genome stability
Nature Communications, 8
Herraiz (2015)
Reactivation of p53 by a cytoskeletal sensor to control the balance between DNA damage and tumor dissemination
J. Natl. Cancer Inst., 108
X. Kong, T. Kuilman, A. Shahrabi, J. Boshuizen, K. Kemper, Ji-Ying Song, H. Niessen, E. Rozeman, M. Foppen, C. Blank, D. Peeper (2017)
Cancer Drug Addiction is Relayed by an ERK2-Dependent Phenotype Switch
Nature, 550
V. Poli, A. Camporeale (2015)
STAT3-Mediated Metabolic Reprograming in Cellular Transformation and Implications for Drug Resistance
Frontiers in Oncology, 5
V. Sanz-Moreno, C. Gaggioli, Maggie Yeo, Jean Albrengues, Fredrik Wållberg, A. Virós, S. Hooper, R. Mitter, C. Féral, Martin Cook, J. Larkin, R. Marais, G. Meneguzzi, E. Sahai, C. Marshall (2011)
ROCK and JAK1 signaling cooperate to control actomyosin contractility in tumor cells and stroma.
Cancer cell, 20 2
(2013)
pEGFP-N3-MLC2-TASA (T18A S19A) -EGFP
G. Long, Carina Fung, A. Menzies, G. Pupo, M. Carlino, J. Hyman, Hamideh Shahheydari, V. Tembe, J. Thompson, R. Saw, J. Howle, N. Hayward, P. Johansson, R. Scolyer, R. Kefford, H. Rizos (2014)
Increased MAPK reactivation in early resistance to dabrafenib/trametinib combination therapy of BRAF-mutant metastatic melanoma
Nature Communications, 5
M. Marzec, Qian Zhang, Ami Goradia, P. Raghunath, Xiaobin Liu, M. Paessler, Hong Wang, M. Wysocka, M. Cheng, B. Ruggeri, M. Wasik (2008)
Oncogenic kinase NPM/ALK induces through STAT3 expression of immunosuppressive protein CD274 (PD-L1, B7-H1)
Proceedings of the National Academy of Sciences, 105
Yangbo Feng, P. Lograsso, O. Defert, Rongshi Li (2016)
Rho Kinase (ROCK) Inhibitors and Their Therapeutic Potential.
Journal of medicinal chemistry, 59 6
Hezhe Lu, Shujing Liu, Gao Zhang, Bin Wu, Yueyao Zhu, Dennie Frederick, Yi Hu, Wenqun Zhong, S. Randell, Norah Sadek, Wei Zhang, Gang Chen, Chaoran Cheng, Jingwen Zeng, Lawrence Wu, Jie Zhang, Xiaoming Liu, Wei Xu, C. Krepler, Katrin Sproesser, M. Xiao, B. Miao, Jianglan Liu, Claire Song, Jephrey Liu, G. Karakousis, L. Schuchter, Yiling Lu, G. Mills, Yusheng Cong, J. Chernoff, Jun Guo, G. Boland, R. Sullivan, Zhi Wei, J. Field, R. Amaravadi, K. Flaherty, M. Herlyn, Xiaowei Xu, Wei Guo (2017)
PAK Signaling Drives Acquired Drug Resistance to MAPK Inhibitors in BRAF-mutant Melanomas
Nature, 550
E. Hirata, M. Girotti, A. Virós, S. Hooper, B. Spencer‐Dene, M. Matsuda, J. Larkin, R. Marais, E. Sahai (2015)
Intravital Imaging Reveals How BRAF Inhibition Generates Drug-Tolerant Microenvironments with High Integrin β1/FAK Signaling
Cancer Cell, 27
Gilles Gadea, V. Sanz-Moreno, A. Self, A. Godi, C. Marshall (2008)
DOCK10-Mediated Cdc42 Activation Is Necessary for Amoeboid Invasion of Melanoma Cells
Current Biology, 18
K. Itoh, K. Yoshioka, H. Akedo, M. Uehata, T. Ishizaki, S. Narumiya (1999)
An essential part for Rho–associated kinase in the transcellular invasion of tumor cells
Nature Medicine, 5
G. Long, D. Stroyakovskiy, H. Gogas, E. Levchenko, F. Braud, J. Larkin, C. Garbe, T. Jouary, A. Hauschild, J. Grob, V. Sileni, C. Lebbé, M. Mandalà, M. Millward, A. Arance, I. Bondarenko, J. Haanen, J. Hansson, J. Utikal, V. Ferraresi, N. Kovalenko, P. Mohr, V. Probachai, D. Schadendorf, P. Nathan, C. Robert, A. Ribas, D. Demarini, J. Irani, M. Casey, D. Ouellet, Anne-Marie Martin, N. Le, K. Patel, K. Flaherty (2014)
Combined BRAF and MEK inhibition versus BRAF inhibition alone in melanoma.
The New England journal of medicine, 371 20
M. Girotti, M. Pedersen, B. Sanchez-Laorden, A. Virós, S. Turajlic, D. Niculescu-Duvaz, A. Zambon, J. Sinclair, A. Hayes, M. Gore, P. Lorigan, C. Springer, J. Larkin, C. Jørgensen, R. Marais (2013)
Inhibiting EGF receptor or SRC family kinase signaling overcomes BRAF inhibitor resistance in melanoma.
Cancer discovery, 3 2
G. Cantelli, E. Crosas-Molist, M. Georgouli, V. Sanz-Moreno (2017)
TGFΒ-induced transcription in cancer
Seminars in Cancer Biology, 42
R. Mjelle, S. Hegre, P. Aas, G. Slupphaug, F. Drabløs, P. Saetrom, H. Krokan (2015)
Cell cycle regulation of human DNA repair and chromatin remodeling genes.
DNA repair, 30
Masaaki Ito, T. Nakano, F. Erdődi, D. Hartshorne (2004)
Myosin phosphatase: Structure, regulation and function
Molecular and Cellular Biochemistry, 259
Stephanie Alexander, P. Friedl (2012)
Cancer invasion and resistance: interconnected processes of disease progression and therapy failure.
Trends in molecular medicine, 18 1
S. Kümper, F. Mardakheh, Afshan Mccarthy, Maggie Yeo, G. Stamp, A. Paul, Jonathan Worboys, Amine Sadok, C. Jørgensen, S. Guichard, C. Marshall (2016)
Rho-associated kinase (ROCK) function is essential for cell cycle progression, senescence and tumorigenesis
eLife, 5
J. Orgaz, C. Herraiz, V. Sanz-Moreno (2014)
Rho GTPases modulate malignant transformation of tumor cells
Small GTPases, 5
A. Jaffe, Alan Hall (2005)
Rho GTPases: biochemistry and biology.
Annual review of cell and developmental biology, 21
A. Obenauf, Y. Zou, A. Ji, Sakari Vanharanta, Weiping Shu, Hubing Shi, X. Kong, M. Bosenberg, T. Wiesner, N. Rosen, R. Lo, J. Massagué (2015)
Therapy-induced tumour secretomes promote resistance and tumour progression
Nature, 520
T. Lämmermann, M. Sixt (2009)
Mechanical modes of 'amoeboid' cell migration.
Current opinion in cell biology, 21 5
I. Shaltiel, L. Krenning, W. Bruinsma, R. Medema (2015)
The same, only different – DNA damage checkpoints and their reversal throughout the cell cycle
Journal of Cell Science, 128
G. Posern, R. Treisman (2006)
Actin' together: serum response factor, its cofactors and the link to signal transduction.
Trends in cell biology, 16 11
Katarina Wolf, Regina Müller, S. Borgmann, E. Bröcker, P. Friedl (2003)
Amoeboid shape change and contact guidance: T-lymphocyte crawling through fibrillar collagen is independent of matrix remodeling by MMPs and other proteases.
Blood, 102 9
J. Larkin, V. Chiarion-Sileni, R. Gonzalez, J. Grob, C. Cowey, C. Lao, D. Schadendorf, R. Dummer, M. Smylie, P. Rutkowski, P. Ferrucci, A. Hill, J. Wagstaff, M. Carlino, J. Haanen, M. Maio, I. Márquez-Rodas, G. McArthur, P. Ascierto, G. Long, M. Callahan, M. Postow, M. Postow, K. Grossmann, M. Sznol, B. Dréno, L. Bastholt, A. Yang, L. Rollin, C. Horak, F. Hodi, J. Wolchok, J. Wolchok (2015)
Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma.
The New England journal of medicine, 373 1
D. Fisher, M. Appenheimer, S. Evans (2014)
The two faces of IL-6 in the tumor microenvironment.
Seminars in immunology, 26 1
E. Clark, T. Golub, E. Lander, R. Hynes (2000)
Genomic analysis of metastasis reveals an essential role for RhoC
Nature, 406
Michael Smith, B. Sanchez-Laorden, K. O’Brien, Holly Brunton, J. Ferguson, H. Young, N. Dhomen, K. Flaherty, Dennie Frederick, Z. Cooper, J. Wargo, R. Marais, C. Wellbrock (2014)
The immune microenvironment confers resistance to MAPK pathway inhibitors through macrophage-derived TNFα.
Cancer discovery, 4 10
J. Larkin, P. Ascierto, B. Dréno, V. Atkinson, G. Liszkay, M. Maio, M. Mandalà, L. Demidov, D. Stroyakovskiy, L. Thomas, L. Cruz-Merino, C. Dutriaux, C. Garbe, M. Sovak, I. Chang, N. Choong, S. Hack, G. McArthur, A. Ribas (2014)
Combined vemurafenib and cobimetinib in BRAF-mutated melanoma.
The New England journal of medicine, 371 20
Chunying Song, M. Piva, Lu Sun, Aayoung Hong, G. Moriceau, X. Kong, Hong Zhang, Shirley Lomeli, J. Qian, C. Yu, R. Damoiseaux, M. Kelley, Kimberley Dahlman, P. Scumpia, J. Sosman, Douglas Johnson, A. Ribas, Willy Hugo, R. Lo (2017)
Recurrent Tumor Cell-Intrinsic and -Extrinsic Alterations during MAPKi-Induced Melanoma Regression and Early Adaptation.
Cancer discovery, 7 11
R. Akbani, K. Akdemir, B. Aksoy, Monique Albert, Adrian Ally, S. Amin, H. Arachchi, Arshi Arora, J. Auman, Brenda Ayala, J. Baboud, M. Balasundaram, S. Balu, N. Barnabas, J. Bartlett, Pamela Bartlett, B. Bastian, S. Baylin, M. Behera, D. Belyaev, C. Benz, Brady Bernard, R. Beroukhim, Natalie Bir, Aaron Black, T. Bodenheimer, L. Boice, G. Boland, R. Bono, M. Bootwalla, M. Bosenberg, Jay Bowen, R. Bowlby, C. Bristow, Laura Brockway-Lunardi, Denise Brooks, Jakub Brzeziński, W. Bshara, Elizabeth Buda, W. Burns, Y. Butterfield, Michael Button, T. Calderone, G. Cappellini, Candace Carter, S. Carter, L. Cherney, A. Cherniack, A. Chevalier, L. Chin, Juok Cho, R. Cho, Yoon-La Choi, Andy Chu, Sudha Chudamani, K. Cibulskis, G. Ciriello, A. Clarke, S. Coons, L. Cope, D. Crain, Erin Curley, Ludmila Danilova, S. D’Atri, Tanja Davidsen, M. Davies, K. Delman, John Demchok, Qixia Deng, Y. Deribe, Noreen Dhalla, R. Dhir, D. Dicara, M. Dinikin, M. Dubina, J. Ebrom, Sophie Egea, G. Eley, J. Engel, J. Eschbacher, K. Fedosenko, Ina Felau, T. Fennell, M. Ferguson, Sheila Fisher, K. Flaherty, S. Frazer, J. Frick, Victoria Fulidou, S. Gabriel, Jianjiong Gao, J. Gardner, L. Garraway, J. Gastier-Foster, Carmelo Gaudioso, N. Gehlenborg, G. Genovese, M. Gerken, J. Gershenwald, G. Getz, C. Gomez-Fernandez, T. Gribbin, J. Grimsby, Benjamin Gross, R. Guin, Tony Gutschner, Angela Hadjipanayis, R. Halaban, B. Hanf, D. Haussler, L. Haydu, D. Hayes, N. Hayward, David Heiman, Lynn Herbert, J. Herman, P. Hersey, K. Hoadley, Eran Hodis, R. Holt, D. Hoon, Susan Hoppough, A. Hoyle, F. Huang, Meijuan Huang, S. Huang, C. Hutter, Matthew Ibbs, L. Iype, A. Jacobsen, V. Jakrot, Alyssa Janning, W. Jeck, S. Jefferys, M. Jensen, Corbin Jones, Steven Jones, Z. Ju, H. Kakavand, Hyojin Kang, R. Kefford, F. Khuri, Jaegil Kim, J. Kirkwood, J. Klode, Anil Korkut, Konstanty Korski, M. Krauthammer, R. Kucherlapati, L. Kwong, W. Kycler, M. Ladanyi, Phillip Lai, P. Laird, E. Lander, M. Lawrence, A. Lazar, Radoslaw Łaźniak, Darlene Lee, Jeffrey Lee, Junehawk Lee, Kenneth Lee, Semin Lee, William Lee, E. Leporowska, K. Leraas, H. Li, T. Lichtenberg, Lee Lichtenstein, Pei Lin, Shiyun Ling, Jia Liu, Ouida Liu, Wenbin Liu, G. Long, Yiling Lu, Singer Ma, Yussanne Ma, A. Mackiewicz, Harshad Mahadeshwar, Jared Malke, D. Mallery, G. Manikhas, G. Mann, M. Marra, Brenna Matejka, Michael Mayo, Sousan Mehrabi, S. Meng, M. Meyerson, P. Mieczkowski, J. Miller, Martin Miller, G. Mills, F. Moiseenko, Richard Moore, S. Morris, C. Morrison, D. Morton, S. Moschos, Lisle Mose, F. Muller, A. Mungall, D. Murawa, P. Murawa, Bradley Murray, L. Nezi, S. Ng, D. Nicholson, M. Noble, A. Osunkoya, T. Owonikoko, B. Ozenberger, E. Pagani, O. Paklina, A. Pantazi, Michael Parfenov, J. Parfitt, P. Park, W. Park, J. Parker, F. Passarelli, R. Penny, C. Perou, T. Pihl, O. Potapova, V. Prieto, A. Protopopov, M. Quinn, Amie Radenbaugh, Kunal Rai, S. Ramalingam, Ayush Raman, N. Ramirez, Ricardo Ramirez, U. Rao, W. Rathmell, X. Ren, Sheila Reynolds, J. Roach, A. Robertson, M. Ross, J. Roszik, G. Russo, G. Saksena, Charles Saller, Yardena Samuels, C. Sander, C. Sander, G. Sandusky, Netty Santoso, M. Saul, R. Saw, D. Schadendorf, J. Schein, N. Schultz, S. Schumacher, Charles Schwallier, R. Scolyer, J. Seidman, P. Sekhar, H. Sekhon, Y. Şenbabaoğlu, S. Seth, K. Shannon, S. Sharpe, N. Sharpless, K. Shaw, C. Shelton, T. Shelton, R. Shen, Margi Sheth, Yan Shi, C. Shiau, I. Shmulevich, G. Sica, J. Simons, Rileen Sinha, Payal Sipahimalani, H. Sofia, Matthew Soloway, Xingzhi Song, C. Sougnez, A. Spillane, A. Spychała, J. Stretch, Joshua Stuart, W. Suchorska, A. Sucker, S. Sumer, Yichao Sun, M. Synott, B. Tabak, Teresa Tabler, Angela Tam, Donghui Tan, Jiabin Tang, R. Tarnuzzer, Katherine Tarvin, Honorata Tatka, B. Taylor, M. Teresiak, N. Thiessen, J. Thompson, L. Thorne, V. Thorsson, J. Trent, T. Triche, K. Tsai, Peiling Tsou, D. Berg, E. Allen, Umadevi Veluvolu, R. Verhaak, Douglas Voet, Olga Voronina, Vonn Walter, Jessica Walton, Yunhu Wan, Yuling Wang, Zhining Wang, S. Waring, I. Watson, Nils Weinhold, J. Weinstein, D. Weisenberger, P. White, M. Wilkerson, J. Wilmott, L. Wise, M. Wiznerowicz, S. Woodman, Chang-Jiun Wu, Chia-Chin Wu, Junyuan Wu, Ye Wu, Ruibin Xi, A. Xu, D. Yang, Liming Yang, Lixing Yang, Travis Zack, J. Zenklusen, Hailei Zhang, Jianhua Zhang, Wei Zhang, Xiao-fei Zhao, Jingchun Zhu, Kelsey Zhu, L. Zimmer, E. Zmuda, L. Zou (2015)
Genomic Classification of Cutaneous Melanoma
Cell, 161
Bo Hang (2015)
Base Excision Repair
Definitions
J. Orgaz, Pahini Pandya, Rimple Dalmeida, P. Karagiannis, B. Sanchez-Laorden, A. Virós, Jean Albrengues, F. Nestle, A. Ridley, C. Gaggioli, R. Marais, S. Karagiannis, V. Sanz-Moreno (2014)
Diverse matrix metalloproteinase functions regulate cancer amoeboid migration
Nature communications, 5
F. Calvo, Nil Ege, A. Grande-García, S. Hooper, Robert Jenkins, S. Chaudhry, K. Harrington, P. Williamson, E. Moeendarbary, G. Charras, E. Sahai (2013)
Mechano-transduction and YAP-dependent matrix remodelling is required for the generation and maintenance of cancer associated fibroblasts
Nature cell biology, 15
H. Rizos, A. Menzies, G. Pupo, M. Carlino, Carina Fung, J. Hyman, L. Haydu, Branka Mijatov, T. Becker, Suzanah Boyd, J. Howle, R. Saw, J. Thompson, R. Kefford, R. Scolyer, G. Long (2014)
BRAF Inhibitor Resistance Mechanisms in Metastatic Melanoma: Spectrum and Clinical Impact
Clinical Cancer Research, 20
K. Flaherty, J. Infante, A. Daud, R. Gonzalez, R. Kefford, J. Sosman, O. Hamid, L. Schuchter, J. Cebon, N. Ibrahim, R. Kudchadkar, H. Burris, G. Falchook, A. Algazi, K. Lewis, G. Long, I. Puzanov, P. Lebowitz, A. Singh, S. Little, P. Sun, A. Allred, D. Ouellet, Kevin Kim, K. Patel, J. Weber (2012)
Combined BRAF and MEK inhibition in melanoma with BRAF V600 mutations.
The New England journal of medicine, 367 18
B. Titz, A. Lomova, A. Le, Willy Hugo, X. Kong, J. Hoeve, Michael Friedman, Hubing Shi, G. Moriceau, Chunying Song, Aayoung Hong, M. Atefi, Richard Li, Evangelia Komisopoulou, A. Ribas, R. Lo, T. Graeber (2016)
JUN dependency in distinct early and late BRAF inhibition adaptation states of melanoma
Cell Discovery, 2
G. Moriceau, Willy Hugo, Aayoung Hong, Hubing Shi, X. Kong, C. Yu, R. Koya, A. Samatar, N. Khanlou, J. Braun, K. Ruchalski, H. Seifert, J. Larkin, K. Dahlman, Douglas Johnson, A. Algazi, J. Sosman, A. Ribas, R. Lo (2015)
Tunable-combinatorial mechanisms of acquired resistance limit the efficacy of BRAF/MEK cotargeting but result in melanoma drug addiction.
Cancer cell, 27 2
N. Wagle, Caroline Emery, M. Berger, Matthew Davis, Allison Sawyer, Panisa Pochanard, Sarah Kehoe, Cory Johannessen, L. Macconaill, W. Hahn, M. Meyerson, L. Garraway (2011)
Dissecting therapeutic resistance to RAF inhibition in melanoma by tumor genomic profiling.
Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 29 22
Daniele Tauriello, Sergio Palomo-Ponce, Diana Stork, A. Berenguer-Llergo, Jordi Badia-Ramentol, M. Iglesias, Marta Sevillano, S. Ibiza, A. Cañellas, Xavier Hernando-Momblona, Daniel Byrom, Joan Matarin, A. Calon, Elisa Rivas, A. Nebreda, A. Riera, Camille Attolini, E. Batlle (2018)
TGFβ drives immune evasion in genetically reconstituted colon cancer metastasis
Nature, 554
Weijiang Zhang (2015)
BRAF inhibitors: the current and the future.
Current opinion in pharmacology, 23
N. Dhomen, J. Reis-Filho, S. Dias, R. Hayward, K. Savage, V. Delmas, L. Larue, C. Pritchard, R. Marais (2009)
Oncogenic Braf induces melanocyte senescence and melanoma in mice.
Cancer cell, 15 4
Filipe Almeida, Stephen Douglass, Mitchell Fane, A. Weeraratna (2018)
Bad company: Microenvironmentally mediated resistance to targeted therapy in melanoma
Pigment Cell & Melanoma Research, 32
T. Condamine, I. Ramachandran, Je-in Youn, D. Gabrilovich (2015)
Regulation of tumor metastasis by myeloid-derived suppressor cells.
Annual review of medicine, 66
F. Calvo, V. Sanz-Moreno, Lorena Agudo-Ibáñez, Fredrik Wållberg, E. Sahai, C. Marshall, P. Crespo (2011)
RasGRF suppresses Cdc42-mediated tumour cell movement, cytoskeletal dynamics and transformation
Nature Cell Biology, 13
S. Medjkane, C. Perez-Sanchez, C. Gaggioli, E. Sahai, R. Treisman (2009)
Myocardin-Related Transcription Factors and SRF are required for cytoskeletal dynamics, invasion and experimental metastasis
Nature cell biology, 11
D. Konieczkowski, Cory Johannessen, L. Garraway (2018)
A Convergence-Based Framework for Cancer Drug Resistance.
Cancer cell, 33 5
Eduardo Vasquez
to record in KAR
D. Konieczkowski, Cory Johannessen, Omar Abudayyeh, Jong Kim, Z. Cooper, A. Piris, Dennie Frederick, Michal Barzily-Rokni, R. Straussman, R. Haq, D. Fisher, J. Mesirov, W. Hahn, K. Flaherty, J. Wargo, P. Tamayo, L. Garraway (2014)
A melanoma cell state distinction influences sensitivity to MAPK pathway inhibitors.
Cancer discovery, 4 7
Liqin Wang, R. Oliveira, S. Huijberts, Evert Bosdriesz, N. Pencheva, D. Brunen, A. Bosma, Ji-Ying Song, J. Zevenhoven, G. Vries, H. Horlings, B. Nuijen, J. Beijnen, J. Schellens, R. Bernards (2018)
An Acquired Vulnerability of Drug-Resistant Melanoma with Therapeutic Potential
Cell, 173
H. Kakavand, R. Rawson, G. Pupo, J. Yang, A. Menzies, M. Carlino, R. Kefford, J. Howle, R. Saw, J. Thompson, J. Wilmott, G. Long, R. Scolyer, H. Rizos (2017)
PD-L1 Expression and Immune Escape in Melanoma Resistance to MAPK Inhibitors
Clinical Cancer Research, 23
V. Sanz-Moreno, Gilles Gadea, Jessica Ahn, H. Paterson, P. Marra, Sophie Pinner, E. Sahai, C. Marshall (2008)
Rac Activation and Inactivation Control Plasticity of Tumor Cell Movement
Cell, 135
P. Sharma, S. Hu-Lieskovan, J. Wargo, A. Ribas (2017)
Primary, Adaptive, and Acquired Resistance to Cancer Immunotherapy
Cell, 168
M. Vicente-Manzanares, Xuefei Ma, R. Adelstein, A. Horwitz (2009)
Non-muscle myosin II takes centre stage in cell adhesion and migration
Nature Reviews Molecular Cell Biology, 10
T. Haystead (2005)
ZIP kinase, a key regulator of myosin protein phosphatase 1.
Cellular signalling, 17 11
A. Hall (2012)
Rho family GTPases.
Biochemical Society transactions, 40 6
Aayoung Hong, G. Moriceau, Lu Sun, Shirley Lomeli, M. Piva, R. Damoiseaux, S. Holmen, N. Sharpless, Willy Hugo, R. Lo (2018)
Exploiting Drug Addiction Mechanisms to Select against MAPKi-Resistant Melanoma.
Cancer discovery, 8 1
F. Baenke, B. Chaneton, Matthew Smith, Niels Broek, Kate Hogan, Haoran Tang, A. Virós, Matthew Martin, L. Galbraith, M. Girotti, N. Dhomen, E. Gottlieb, R. Marais (2015)
Resistance to BRAF inhibitors induces glutamine dependency in melanoma cells
Molecular Oncology, 10
Binzhi Qian, Jiu-feng Li, Hui Zhang, T. Kitamura, Jinghang Zhang, Liam Campion, E. Kaiser, L. Snyder, J. Pollard (2011)
CCL2 recruits inflammatory monocytes to facilitate breast tumor metastasis
Nature, 475
P. Lau, P. Ascierto, G. McArthur (2016)
Melanoma: the intersection of molecular targeted therapy and immune checkpoint inhibition.
Current opinion in immunology, 39
L. Kwong, G. Boland, Dennie Frederick, Timothy Helms, Ahmad Akid, J. Miller, Shan Jiang, Z. Cooper, Xingzhi Song, S. Seth, Jennifer Kamara, A. Protopopov, G. Mills, K. Flaherty, J. Wargo, L. Chin (2015)
Co-clinical assessment identifies patterns of BRAF inhibitor resistance in melanoma.
The Journal of clinical investigation, 125 4
Chong Sun, Liqin Wang, Sidong Huang, Guus Heynen, A. Prahallad, C. Robert, J. Haanen, C. Blank, J. Wesseling, Stefan Willems, D. Zecchin, S. Hobor, P. Bajpe, C. Lieftink, C. Mateus, S. Vagner, Wipawadee Grernrum, I. Hofland, A. Schlicker, L. Wessels, R. Beijersbergen, A. Bardelli, F. Nicolantonio, A. Eggermont, R. Bernards (2014)
Reversible and adaptive resistance to BRAF(V600E) inhibition in melanoma
Nature, 508
I. Arozarena, B. Sanchez-Laorden, Leisl Packer, Cristina Hidalgo-Carcedo, R. Hayward, A. Virós, E. Sahai, R. Marais (2011)
Oncogenic BRAF induces melanoma cell invasion by downregulating the cGMP-specific phosphodiesterase PDE5A.
Cancer cell, 19 1
G. Peng, Curtis Lin, W. Mo, H. Dai, Yun‐Yong Park, S. Kim, Yang Peng, Qianxing Mo, S. Siwko, Ruozhen Hu, Ju-Seog Lee, B. Hennessy, S. Hanash, G. Mills, Shiaw-Yih Lin (2014)
Genome-wide Transcriptome Profiling of Homologous Recombination DNA Repair
Nature communications, 5
V. Gray-Schopfer, C. Wellbrock, R. Marais (2007)
Melanoma biology and new targeted therapy
Nature, 445
Tumor Xenografts A375/PLX/R cells (1 x 10 6 ) were injected subcutaneously into the right flank of 5-week-old female nude CD-1 mice
G. Dréau, L. Kular, A. Nicot, C. Calmel, S. Mélik-Parsadaniantz, P. Kitabgi, M. Laurent, C. Martinerie (2010)
NOV/CCN3 upregulates CCL2 and CXCL1 expression in astrocytes through β1 and β5 integrins
Glia, 58
N. Wagle, E. Allen, D. Treacy, Dennie Frederick, Z. Cooper, A. Taylor-Weiner, Mara Rosenberg, E. Goetz, R. Sullivan, Deborah Farlow, Dennis Friedrich, K. Anderka, Danielle Perrin, Cory Johannessen, A. McKenna, K. Cibulskis, G. Kryukov, Eran Hodis, D. Lawrence, Sheila Fisher, G. Getz, S. Gabriel, S. Carter, K. Flaherty, J. Wargo, L. Garraway (2014)
MAP kinase pathway alterations in BRAF-mutant melanoma patients with acquired resistance to combined RAF/MEK inhibition.
Cancer discovery, 4 1
(2014)
In patients with several biopsies, their average is shown (see Table S4)
Kazuhiko Nakamura, A. Kitani, W. Strober (2001)
Cell Contact–Dependent Immunosuppression by Cd4+Cd25+Regulatory T Cells Is Mediated by Cell Surface–Bound Transforming Growth Factor β
The Journal of Experimental Medicine, 194
All siRNA sequences were On-Targetplus (OT) from Dharmacon (Lafayette, USA) and are listed in Table S7
P. Bankhead, M. Loughrey, José Fernández, Yvonne Dombrowski, D. McArt, P. Dunne, S. McQuaid, R. Gray, L. Murray, H. Coleman, J. James, M. Salto‐Tellez, P. Hamilton (2017)
QuPath: Open source software for digital pathology image analysis
Scientific Reports, 7
C. Herraiz, F. Calvo, Pahini Pandya, G. Cantelli, I. Rodriguez-Hernandez, J. Orgaz, Nara Kang, Tinghine Chu, E. Sahai, V. Sanz-Moreno (2015)
Reactivation of p53 by a Cytoskeletal Sensor to Control the Balance Between DNA Damage and Tumor Dissemination
JNCI Journal of the National Cancer Institute, 108
Hailey Brighton, Steven Angus, Tao Bo, J. Roques, Alicia Tagliatela, David Darr, Kubra Karagoz, N. Sciaky, Michael Gatza, N. Sharpless, G. Johnson, J. Bear (2018)
New Mechanisms of Resistance to MEK Inhibitors in Melanoma Revealed by Intravital Imaging.
Cancer research, 78 2
N. Riaz, J. Havel, V. Makarov, A. Desrichard, John-William Sidhom, T. Chan (2017)
Tumor and Microenvironment Evolution during Immunotherapy with Nivolumab
Cell, 171
Hanane Laklai, Y. Miroshnikova, Michael Pickup, E. Collisson, Grace Kim, Alexander Barrett, Ryan Hill, J. Lakins, D. Schlaepfer, J. Mouw, Valerie LeBleu, Nilotpal Roy, S. Novitskiy, J. Johansen, V. Poli, R. Kalluri, C. Iacobuzio-Donahue, L. Wood, M. Hebrok, K. Hansen, H. Moses, V. Weaver (2016)
Genotype tunes pancreatic ductal adenocarcinoma tissue tension to induce matricellular-fibrosis and tumor progression
Nature medicine, 22
G. Veroli, Chiara Fornari, Dennis Wang, S. Mollard, J. Bramhall, F. Richards, D. Jodrell (2016)
Combenefit: an interactive platform for the analysis and visualization of drug combinations
Bioinformatics, 32
Sadok (2015)
Rho kinase inhibitors block melanoma cell migration and inhibit metastasis
Cancer Res., 75
M. Olson (2008)
Applications for ROCK kinase inhibition.
Current opinion in cell biology, 20 2
F. Hodi, S. O'Day, D. McDermott, R. Weber, J. Sosman, J. Haanen, R. Gonzalez, C. Robert, D. Schadendorf, Jessica Hassel, W. Akerley, A. Eertwegh, J. Lutzky, P. Lorigan, Julia Vaubel, G. Linette, D. Hogg, C. Ottensmeier, C. Lebbé, C. Peschel, I. Quirt, Joseph Clark, J. Wolchok, J. Weber, Jason Tian, M. Yellin, G. Nichol, A. Hoos, W. Urba (2010)
Improved survival with ipilimumab in patients with metastatic melanoma.
The New England journal of medicine, 363 8
H. Davies, G. Bignell, Charles Cox, P. Stephens, S. Edkins, S. Clegg, J. Teague, H. Woffendin, Mathew Garnett, W. Bottomley, Neil Davis, E. Dicks, Rebecca Ewing, Yvonne Floyd, K. Gray, S. Hall, R. Hawes, Jaime Hughes, Vivian Kosmidou, A. Menzies, Catherine Mould, A. Parker, C. Stevens, S. Watt, S. Hooper, Rebecca Wilson, H. Jayatilake, B. Gusterson, C. Cooper, J. Shipley, D. Hargrave, K. Pritchard-Jones, N. Maitland, G. Chenevix-Trench, G. Riggins, D. Bigner, G. Palmieri, A. Cossu, A. Flanagan, A. Nicholson, Judy Ho, Suet Leung, S. Yuen, Barbara Weber, H. Seigler, T. Darrow, H. Paterson, R. Marais, Christopher Marshall, R. Wooster, Michael Stratton, P. Futreal (2002)
Mutations of the BRAF gene in human cancer
Nature, 417
E. Sahai, C. Marshall (2002)
RHO–GTPases and cancer
Nature Reviews Cancer, 2
G. Cantelli, J. Orgaz, I. Rodriguez-Hernandez, P. Karagiannis, O. Maiques, X. Matías-Guiu, F. Nestle, R. Martí, S. Karagiannis, V. Sanz-Moreno (2015)
TGF-β-Induced Transcription Sustains Amoeboid Melanoma Migration and Dissemination
Current Biology, 25
M. Samuel, Jose Lopez, E. McGhee, D. Croft, D. Strachan, P. Timpson, J. Munro, E. Schröder, Jing Zhou, V. Brunton, N. Barker, H. Clevers, O. Sansom, K. Anderson, V. Weaver, M. Olson (2011)
Actomyosin-mediated cellular tension drives increased tissue stiffness and β-catenin activation to induce epidermal hyperplasia and tumor growth.
Cancer cell, 19 6
P. Chapman, A. Hauschild, C. Robert, J. Haanen, P. Ascierto, J. Larkin, R. Dummer, C. Garbe, A. Testori, M. Maio, D. Hogg, P. Lorigan, C. Lebbé, T. Jouary, D. Schadendorf, A. Ribas, S. O'Day, J. Sosman, J. Kirkwood, A. Eggermont, B. Dréno, K. Nolop, Jiang Li, B. Nelson, J. Hou, Richard Lee, K. Flaherty, G. McArthur (2011)
Improved survival with vemurafenib in melanoma with BRAF V600E mutation.
The New England journal of medicine, 364 26
M. Georgouli, C. Herraiz, E. Crosas-Molist, Bruce Fanshawe, O. Maiques, Anna Perdrix, Pahini Pandya, I. Rodriguez-Hernandez, Kristina Ilieva, G. Cantelli, P. Karagiannis, S. Mele, Hoyin Lam, D. Josephs, X. Matías-Guiu, R. Martí, F. Nestle, J. Orgaz, I. Malanchi, G. Fruhwirth, S. Karagiannis, V. Sanz-Moreno (2019)
Regional Activation of Myosin II in Cancer Cells Drives Tumor Progression via a Secretory Cross-Talk with the Immune Microenvironment
Cell, 176
D. Pardoll (2012)
The blockade of immune checkpoints in cancer immunotherapy
Nature Reviews Cancer, 12
M. Paszek, Nastaran Zahir, Kandice Johnson, J. Lakins, G. Rozenberg, A. Gefen, C. Reinhart-King, S. Margulies, M. Dembo, D. Boettiger, D. Hammer, V. Weaver (2005)
Tensional homeostasis and the malignant phenotype.
Cancer cell, 8 3
(2015)
Resistant samples vs baseline (right graph) (Hugo et al., 2015; Kakavand et al., 2017
Krokan (2013)
Base excision repair
Cold Spring Harb. Perspect. Biol., 5
M. Thakur, Fernando Salangsang, Allison Landman, W. Sellers, N. Pryer, M. Levesque, R. Dummer, M. McMahon, D. Stuart (2013)
Modelling vemurafenib resistance in melanoma reveals a strategy to forestall drug resistance
Nature, 494
Willy Hugo, Hubing Shi, Lu Sun, M. Piva, Chunying Song, X. Kong, G. Moriceau, Aayoung Hong, K. Dahlman, Douglas Johnson, J. Sosman, A. Ribas, R. Lo (2015)
Non-genomic and Immune Evolution of Melanoma Acquiring MAPKi Resistance
Cell, 162
H. Roca, Z. Varsos, S. Sud, Matthew Craig, C. Ying, K. Pienta (2009)
CCL2 and Interleukin-6 Promote Survival of Human CD11b+ Peripheral Blood Mononuclear Cells and Induce M2-type Macrophage Polarization*
The Journal of Biological Chemistry, 284

Publisher: Unpaywall
ISSN: 1535-6108
DOI: 10.1016/j.ccell.2019.12.003
Publisher site: See Article on Publisher Site

Abstract

Article Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance Graphical Abstract Authors Jose L. Orgaz, Eva Crosas-Molist, Amine Sadok, ..., actin-Myosin dynamics Sophia N. Karagiannis, Ilaria Malanchi, Victoria Sanz-Moreno transcriptional re-wiring Correspondence cytoskeleton ROCK-Myosin II- remodeling addicted tumor j.orgaz@qmul.ac.uk (J.L.O.), MAPKi v.sanz-moreno@qmul.ac.uk (V.S.-M.) anti-PD-1 MAPKi MAPK Myosin II Cross-resistant ROCK/Myosin II In Brief phenotype inhibition Orgaz et al. show that myosin II activity ROCK/Myosin II ROCK/Myosin II increases during melanoma adaptation to PD-L1 MAPK pathway inhibition. ROCK-myosin PD-L1 ROS ROS II signaling supports survival of resistant p-H2A.X melanoma cells and promotes immunosuppression. ROCK inhibitors FOXP3 + improve the efﬁcacy of MAPK inhibitors CD206 Treg CD206 CD206 FOXP3+ Mφ Mφ Mφ Treg and immunotherapies in melanoma FOXP3+ FOXP3+ CD206+ CD206+ Treg Treg Mφ CD206+ Mφ models. CD206+ FOXP3 Mφ Mφ Treg Highlights d Therapy-resistant melanoma cells restore myosin II activity to increase survival d High myosin II activity identiﬁes targeted and immunotherapy-resistant melanomas d ROCK-myosin II inhibition increases ROS-DNA damage and decreases PD-L1 and Tregs d ROCK inhibition enhances efﬁcacy of MAPK inhibitors and immunotherapies Orgaz et al., 2020, Cancer Cell 37, 85–103 January 13, 2020 ª 2019 The Authors. Published by Elsevier Inc. https://doi.org/10.1016/j.ccell.2019.12.003 Cancer Cell Article Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance 1,2, 1,2,10 3,10 1,2,4,10 1,2,10 Jose L. Orgaz, * Eva Crosas-Molist, Amine Sadok, Anna Perdrix-Rosell, Oscar Maiques, 1,2,10 1 5 2 4 Irene Rodriguez-Hernandez, Jo Monger, Silvia Mele, Mirella Georgouli, Victoria Bridgeman, 5,6 7 2 2 8 9 Panagiotis Karagiannis, Rebecca Lee, Pahini Pandya, Lena Boehme, Fredrik Wallberg, Chris Tape, 5 4 1,2,11, Sophia N. Karagiannis, Ilaria Malanchi, and Victoria Sanz-Moreno * Barts Cancer Institute, Queen Mary University of London, John Vane Science Building, Charterhouse Square, London EC1M 6BQ, UK Randall Division of Cell and Molecular Biophysics, King’s College London, New Hunt’s House, Guy’s Campus, London SE1 1UL, UK Translational Cancer Discovery Team, Cancer Research UK Cancer Therapeutics Unit, The Institute of Cancer Research, 15 Cotswold Road, Sutton, London SM2 5NG, UK Tumour Host Interaction, The Francis Crick Institute, 1 Midland Road, London NW1 1AT, UK St. John’s Institute of Dermatology, King’s College London & NIHR Biomedical Research Centre at Guy’s and St. Thomas’s Hospitals and King’s College London, London SE1 9RT, UK Department of Oncology, Haematology and Stem Cell Transplantation, University Hospital of Hamburg Eppendorf, Hamburg 20246, Germany Molecular Oncology Group, Cancer Research UK Manchester Institute, Manchester M20 4BX, UK The Institute of Cancer Research, Chester Beatty Laboratories, 237 Fulham Road, London SW3 6JB, UK Cell Communication Lab, UCL Cancer Institute, 72 Huntley Street, London WC1E 6DD, UK These authors contributed equally Lead Contact *Correspondence: j.orgaz@qmul.ac.uk (J.L.O.), v.sanz-moreno@qmul.ac.uk (V.S.-M.) https://doi.org/10.1016/j.ccell.2019.12.003 SUMMARY Despite substantial clinical beneﬁt of targeted and immune checkpoint blockade-based therapies in mela- noma, resistance inevitably develops. We show cytoskeletal remodeling and changes in expression and ac- tivity of ROCK-myosin II pathway during acquisition of resistance to MAPK inhibitors. MAPK regulates myosin II activity, but after initial therapy response, drug-resistant clones restore myosin II activity to increase survival. High ROCK-myosin II activity correlates with aggressiveness, identifying targeted therapy- and immunotherapy-resistant melanomas. Survival of resistant cells is myosin II dependent, regardless of the therapy. ROCK-myosin II ablation speciﬁcally kills resistant cells via intrinsic lethal reactive oxygen species and unresolved DNA damage and limits extrinsic myeloid and lymphoid immunosuppression. Efﬁcacy of tar- geted therapies and immunotherapies can be improved by combination with ROCK inhibitors. INTRODUCTION inhibitors (BRAFi) development (Chapman et al., 2011; Flaherty et al., 2010; Zhang, 2015). Unfortunately, most patients had par- Malignant melanoma has very poor survival rates (Balch et al., tial responses and disease progressed due to acquired resis- 2009) despite being at the forefront of personalized medicine tance (Larkin et al., 2014; Robert et al., 2015; Zhang, 2015). (Lau et al., 2016). Mutant BRAF (V600) is the most common Often, patients with resistance develop more metastases (Wagle oncogene in melanoma (Davies et al., 2002), driving proliferation, et al., 2011) and 20% of BRAF mutant melanoma patients never survival, and tumor progression by hyper-activating MEK and respond to BRAFi due to intrinsic resistance (Zhang, 2015). V600E ERK kinases (Gray-Schopfer et al., 2007). This led to BRAF Most resistance mechanisms involve MAPK reactivation Signiﬁcance Resistance to therapies is a persistent problem in melanoma management. Here, we identify an adaptation strategy in response to either targeted therapies or immunotherapies. Under treatment, melanoma cells undergo cytoskeletal remod- eling and consequent activation of ROCK-myosin II pathway. Such adaptation process renders resistant melanoma cells vulnerable to ROCK-myosin II inhibition, which can be exploited therapeutically. Cancer Cell 37, 85–103, January 13, 2020 ª 2019 The Authors. Published by Elsevier Inc. 85 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). A B C **** - BRAFi MEKi ROCKi **** **** Cytoskeleton remodeling/Rho GTPase signaling **** **** 1.0 **** Transcription Axon growth 0.8 silencing Translation 0.6 Notch-EMT 20 EIF2 activity 0.4 Development Development 0.2 Slit-Robo signaling Notch pathway Cell cycle NF-κB modulation 0 DNA replication by Notch - - F-actin Hoechst p-MLC2 D E WM983A WM983B WM983A WM983B **** **** ROCKi - BRAFi ROCKi -- + - - + **** **** 1.0 1.0 -+ - - + - BRAFi V600E BRAF p-MLC2 0.8 0.8 WM983A 0.6 0.6 MLC2 0.4 0.4 p-ERK1/2 0.2 0.2 ERK2 WM983B 0 0 GAPDH -- V600E Q61L 690cl2 BRAF /Pten null D04 NRAS ns F G 8 hr 24 hr 48 hr **** -+ - + - + BRAFi ** **** **** * *** 1.0 *** **** p-MLC2 1.0 2.0 **** ** 0.8 0.8 **** MLC2 1.5 0.6 0.6 0.4 0.4 p-MLC2/MLC2 1.0 0.2 p-ERK/ERK 0.2 p-ERK1/2 0 long exposure 0.5 - G - A MEKi ERK2 024 48 hr BRAFi GAPDH A375 WM983A H I J A375 sens PLX/R sens PLX/R A375/PLX/R - BRAFi -+ - + -+ - + BRAFi ** p-MLC2 ** MLC2 p-ERK1/2 ERK2 GAPDH WM983B WM88 sens PLX/R sens PLX/R p-MLC2 - +- + - + - + BRAFi DAPI F-actin GFP-MLC2 p-MLC2 250 250 250 MLC2 200 200 200 150 150 150 p-ERK1/2 100 100 100 50 50 50 GFP WT TASA TDSD ERK2 0 0 0 010 20 30 010 20 30 40 010 20 30 40 50 GAPDH Distance from edge (μm) A375/PLX/R M229-PLX/R M238-PLX/R Cytoskeleton remodeling Cytoskeleton remodeling Cytoskeleton remodeling/Rho GTPase signaling Chemotaxis - LPA signaling via GPCRs Reg. of actin Chemotaxis - LPA signaling via GPCRs Reg. of cytoskeleton Cell adhesion/Histamine in CFTR folding cytoskeleton in oligodendrocytes Cell adhesion/Histamine in VEGF signaling cell barrier integrity maturation cell barrier integrity H. pylori infection on Reg. of actin Signal transduction epithelial cells motility Oxidative stress Reg. of cytoskeleton STK3/4 Hippo, cytoskeleton mTORC2 (NOX) in remyelination YAP/TAZ Lipid metabolism HSF-1/chaperone VEGF signaling Cell cycle - Development - Notch Signaling Insulin signaling Huntington’s disease Non-genomic action chr.condensation Immune response_IL-4 retinoic acid Canonical Notch signaling in CRC Signal transduction PTMs in Gamma-secretase Gastrin in cell growth Lipid metabolism Signal transduction BAFF-induced signaling in osteogenesis WNT and Notch cardiac myogenesis and proliferation Insulin signaling mTORC2 Figure 1. MAPK Regulates Myosin II Activity in Melanoma (A) The 10 most enriched pathways in A375 cells after MEKi treatment compared with untreated cells from phospho-proteome data. (B) p-MLC2 and F-actin confocal images of A375M2 cells on collagen I after treatment (BRAFi PLX4720, MEKi trametinib, ROCKi GSK269962A). Scale bar, 25 mm. (legend continued on next page) 86 Cancer Cell 37, 85–103, January 13, 2020 MEKi BRAFi ERKi BRAFi MEKi ROCKi BRAFi BRAFi ROCKi MEKi ROCKi BRAFi ROCKi GFP WT TASA TDSD Cell morphology on collagen I % survival vs vehicle Cell morphology on collagen I Fluorescence int. p-MLC2 F-actin Hoechst Cell morphology on collagen I Relative levels BRAFi/vehicle p-MLC2/cell area (Konieczkowski et al., 2018). Therefore, combination of a BRAFi 2014b; Sanz-Moreno et al., 2008, 2011), and aggressive amoe- with a MEK inhibitor (MEKi) was approved (Flaherty et al., 2012; boid invasion (Cantelli et al., 2015; Medjkane et al., 2009; Orgaz Larkin et al., 2014; Long et al., 2014b). However, despite the et al., 2014b; Sanz-Moreno et al., 2008, 2011). improved responses, most patients still relapse (Flaherty et al., In vivo, ROCK inhibition diminishes tumor growth and meta- 2012; Konieczkowski et al., 2018). static spread (Itoh et al., 1999; Kumper et al., 2016; Sadok Improved survival in patients with melanoma was reported et al., 2015). However, the role of ROCK-myosin II during resis- after immune checkpoint inhibitor treatment (anti-PD-1 and tance to current cancer therapies has not been comprehensively anti-CTLA-4) (Hodi et al., 2010; Larkin et al., 2015; Sharma investigated. Intriguingly, PAK contributes to MAPKi resistance et al., 2017). However, there are patients who do not respond (Lu et al., 2017) and Cdc42-PAK2-myosin II regulates amoeboid or relapse due to resistance (Sharma et al., 2017). Therefore, invasion (Calvo et al., 2011; Gadea et al., 2008). drug resistance is a persistent problem in melanoma manage- Given the activation of pro-invasive/metastasis pathways ment. Better understanding of the biological/biochemical during melanoma cross-resistance (Hugo et al., 2015, 2016), changes in resistant cells will help develop improved treatments. we sought to investigate the role of cytoskeletal remodeling in Given the overlap between migration and pro-survival path- therapy resistance. ways, drivers of resistance have been linked to metastatic ability (Alexander and Friedl, 2012). Importantly, cross-resistance to RESULTS MAPK inhibitors (MAPKi) (Hugo et al., 2015) and immune check- point inhibitors (Hugo et al., 2016) has been described, involving MAPK Regulates Myosin II Activity in Melanoma transcriptomic alterations on genes key for epithelial-to-mesen- To gain unbiased insight into molecular changes in melanoma chymal transition (EMT), metastasis/invasion, extracellular ma- cells after MAPKi, we analyzed the phosphoproteome of V600E trix (ECM) remodeling, hypoxia and angiogenesis (Hugo et al., BRAF A375 melanoma cells early (24 h) on MEKi 2015, 2016). (GSK1120212 trametinib and PD184352) treatment (Figure 1A; ROCK-myosin II pathway is a key regulator of invasive and Table S1). Using MetaCore Pathway enrichment analysis, we metastatic behavior (Cantelli et al., 2015; Medjkane et al., found that cytoskeletal remodeling and Rho GTPase signaling 2009; Orgaz et al., 2014b; Sanz-Moreno et al., 2008, 2011). are top processes changing early on treatment (Figure 1A; Ta- Non-muscle myosin II has contractile properties and is regu- bles S1 and S2). lated by the phosphorylation of its light and heavy chains (Vice- Because Rho GTPase regulates invasion via ROCK-myosin II nte-Manzanares et al., 2009). Myosin II-driven contractility activity and amoeboid behavior (Jaffe and Hall, 2005; Olson, relies on multiple kinases. Rho-kinase (ROCK) inactivates the 2008; Sadok et al., 2015; Sahai and Marshall, 2002; Sanz-Mor- myosin light chain 2 (MLC2) phosphatase, which leads to eno et al., 2008), we studied how MAPK inhibition affected increased phosphorylation of MLC2 (p-MLC2) and myosin II melanoma phenotypes on collagen I-recapitulating dermal envi- activity (Ito et al., 2004; Olson, 2008). MLC2 is directly phos- ronments (Cantelli et al., 2015; Orgaz et al., 2014b; Sanz-Moreno phorylated by ROCK and myosin light chain kinase (MLCK) et al., 2008, 2011). Treatment of highly metastatic, amoeboid (Vicente-Manzanares et al., 2009). ZIP kinase can also phos- A375M2 melanoma cells with BRAFi PLX4720 and MEKi trame- phorylate MLC2 directly and indirectly (Haystead, 2005). How- tinib for 24 h led to loss of rounded-amoeboid behavior (Fig- ever, long-term depletion of ROCK1/2 cannot be substituted by ure 1B). Inhibition of myosin II with ROCKi GSK269962A induced any other kinase for generating actomyosin contractility loss of circularity and a collapsed cytoskeleton (Figure 1B). (Kumper et al., 2016). Myosin II activity drives contractile forces Reduced myosin II activity (p-MLC2) was observed after BRAF, required for migration (Clark et al., 2000; Lammermann and MEK, or ROCK inhibition (Figure 1C). Similar results were Sixt, 2009; Sahai and Marshall, 2002; Sanz-Moreno et al., observed in other human and mouse melanoma cells and other V600E 2008, 2011; Vicente-Manzanares et al., 2009; Wolf et al., MAPKi, including BRAF (WM983A, WM983B, 4599) (Fig- V600E 2003), metastatic colonization (Cantelli et al., 2015; Clark ures 1D, 1E, S1A, and S1B), BRAF /Pten-null 690cl2 (Figures Q61L/R et al., 2000; Hall, 2012; Herraiz et al., 2015; Orgaz et al., 1F and S1C) and NRAS (D04, MM485) (Figures 1F, S1D, (C) Quantiﬁcation of cell morphology and p-MLC2 by immunoﬂuorescence from (B). Left, boxplot (n > 200 cells pooled from 3 experiments); right, mean ± SEM (n = 90 cells [dots] pooled from 3 experiments). (D) Images and quantiﬁcation of cell morphology on collagen I after treatment (BRAFi PLX4720, ROCKi GSK269962A) (n > 346 cells pooled from 2 experiments). Arrows show collapsed phenotype. Scale bar, 100 mm. (E) p-MLC2 and p-ERK1/2 immunoblots from (D). (F) Cell morphology on collagen I after treatment (690cl2, MEKi PD184352, BRAFi PLX4032, ERKi SCH772984, n = 50 cells; D04, MEKi GSK1120212, AZD6244, n = 125–150 cells). (G) p-MLC2 and p-ERK1/2 levels after PLX4720 treatment (n = 5, mean ± SEM). (H) Survival of A375 cells stably overexpressing wild type (WT), constitutively inactive TASA, or constitutively active TDSD MLC2 a after 5-day treatment with 0.1 mM PLX4720 (n = 4). Confocal images of GFP-MLC2. Scale bar, 50 mm. (I) p-MLC2 and p-ERK1/2 immunoblots after PLX4720 treatment. (J) p-MLC2 and F-actin confocal images (BRAFi PLX4720). Scale bar, 25 mm. Representative ﬂuorescence intensity line scans (dashed lines in image) below. (K) The 10 most enriched pathways in BRAFi-resistant A375/PLX/R (Girotti et al., 2013), M229-PLX/R, and M238-PLX/R cells (Titz et al., 2016) compared with parental cell lines from phospho-proteome data. (A–F, I, and J) 24 h treatment. (C, D, F, and H) Boxplots show median (center line); interquartile range (box); min-max (whiskers). p values by Kruskal-Wallis with Dunn’s correction (C, D, and F), one-way ANOVA with Tukey’s (H) or Benjamini, Krieger, and Yekutieli correction (G), *p < 0.05, **p < 0.01, ****p < 0.0001. See also Figure S1 and Tables S1 and S2. Cancer Cell 37, 85–103, January 13, 2020 87 Figure 2. ROCK-Myosin II Pathway Is Tran- A B Cytoskeleton-related genes scriptionally Rewired during Development Log fold change expression vs Parental cell line of Resistance -0.5 0.5 MAPK blockade V600E Resistant BRAF (A) Cell lines used for gene expression (Obenauf time weeks months single-drug resistant (SDR) Group 1 Group 2 melanoma et al., 2015; Song et al., 2017). cell lines double-drug resistant (DDR) (B) Heatmap ofunsupervised hierarchicalclustering 48 hr DTP DTPP of 313 cytoskeleton-related genes in A375 48 h drug-tolerant drug-tolerant persister proliferating persister BRAFi (Obenauf et al., 2015); M229-, M238-, SKMEL28-, M395p2-, M395p1-, and M249-de- ACTA2 MYLK MYLK Transcriptomics (RNAseq) Cytoskeleton-related genes MYL9 rivatives (Songetal.,2017). Foldchangeexpression ITGA11 ARHGDIG MYL9 MYL2 MYL3 ARHGAP28 Rho GTPases, GEFs, GAPs, effectors MYH2 ACTC1 RHOBTB1 in resistant versus parental lines is shown. Myosin, actin proteins, cross-linkers, MYL7 FMNL1 ARHGAP29 Cytoskeleton-membrane linkers, integrins RHOU ANK1 ITGA8 MYOF Transcription factors CDC42EP1 (C) Percentage of upregulated/downregulated ARHGAP23 TLN2 ARHGEF17 MYL6 DAPK3 DAPK3 MYH9 VCL MYO7B MYH9 (1.5-fold) cytoskeleton-genes versus parental line. DOCK2 FMN2 PAK1 ITGB4 ITGA10 ITGA9 CITED1 POU3F2 BRN2 MYOZ2 (D) Percentage of upregulated genes. Boxplot: MYOZ1 ARHGEF10L ARHGAP24 C ANK3 RHOB MYOT SPTBN5 median (center line); interquartile range (box); ITGA1 RHOJ ARHGEF6 MYO3B LIMK2 ARHGAP15 LIMK2 ARHGDIB CDC42EP5 ITGA2 min-max (whiskers). p value by unpaired t test RAC2 Group 1 Group 2 ARHGAP22 ITGA3 ITGA5 MYOD1 ARHGAP6 ACTBL2 with Welch’s correction, ****p < 0.0001. 70 MYH15 ITGB3 ARHGAP25 PAK3 SEPT3 up in Resistant ARHGEF37 ARHGAP40 60 MYO10 (E) Left, schematic pathway. Right, percentage of ARHGAP31 down in Resistant DOCK10 ANK2 ARHGAP42 SPTBN1 ITGA4 50 PDK4 ARHGAP4 group 1 cell lines with upregulation of indicated MYH14 MYH10 MYO16 ARHGAP19 ARHGAP32 MYO5A ARHGAP12 ARHGAP5 genes. ACTN2 ITGAX DOCK4 FMN1 30 SPTBN2 ITGA6 ITGB8 MYL10 MYO5B See also Table S3. MYO1D ARHGEF35 ARHGEF5 20 SEPT4 DOCK3 ARHGEF40 MYL4 ARHGAP9 CDC42EP4 CDC42EP3 10 ITGB6 RHOV ACTR3C PAK6 PPP1R16B MYLK2 ARHGEF4 DIAPH3 ACTN1 SEPT11 ITGB1BP2 NEDD9 CDC42EP2 CIT LIMK1 DIAPH2 LIMK1 ARHGAP18 PDK3 ARHGEF11 ARPC1B WASL MYH16 MYOM3 WASH2P (Figures 1H and S1F). MLC2 overex- ARHGAP30 WASH1 WASH3P WASH7P 48 hr DTP DTPP DDR SDR SDR M249 DDR MYO1E MYOC RHOC ARHGEF2 FMNL3 ITGB5 pression did not affect p-ERK (Fig- DOCK6 DOCK7 SEPT9 ARHGAP21 LMNA MYO7A RAC3 SEPT14 DIAPH1 ure S1F). Moreover, high myosin II activ- WASF1 PAK4 PXN RHOBTB2 MYO19 ITGB3BP LMNB1 ECT2 ARHGAP11A ity A375M2 cells were more resistant to ARHGAP11B LMNB2 40 WASF3 **** MYH1 MYH8 SPTB RHOQ DOCK9 ITGA7 ARHGEF16 BRAFi and MEKi compared with low met- ARHGEF19 ITGB2 30 MYH6 RHOH ARHGAP26 CALM2 ACTB MYL12B ACTG1 MYL12B astatic, low myosin II activity A375 cells ACTR3 ACTR2 ACTR10 20 ARPC5 CDC42BPB ACTR1A SEPT2 CDC42BPA ARHGAP10 (Figures S1G and S1H). Similar results CDC42 ITGB1BP1 PDK1 HIF1A 10 DOCK11 RND3 ARHGAP44 DOCK5 ITGB1 TLN1 were observed using the pair WM983B MKL1 ACTA1 MKL1 SEPT5 MYL1 0 MYOG ARHGEF15 RHOBTB3 CFL2 1 2 Group ARHGAP33 (metastatic, high myosin II, and amoe- ARHGEF18 PPP1R12C Sep-08 ARHGAP1 ARHGAP17 PDK2 MKL2 ILK MKL2 SRF boid) versus WM983A (primary tumor, RHOD RND2 DOCK8 ARHGEF3 MYO15B MYO18B MYOM1 MYOM2 MYH7 ARHGEF33 low myosin II, and elongated) (Figures ARHGEF26 MYO1G ARHGEF9 ARHGAP20 MYO9A ROCK2 FMNL2 ROCK2 MYL5 MYOZ3 S1G and S1H). These data show that SEPT12 ITGAD E WAS ITGAM CDC42BPG ARHGEF25 MYL12A MYO6 MYL12A ARHGAP27 myosin II activity confers a survival ARHGAP39 PTK2B MYH13 MYH4 ROCK1/2 MYO9B MYO3A SPTBN4 RHOG ACTR1B 100 PAK2 advantage to BRAFi and could accel- RHOA WASF2 RAC1 ARHGEF1 MYL6B ARPC3 LIMK1/2 80 CFL1 P P EZR RND1 erate the onset of resistance. Accord- ITGA2B myosin ACTR3B CALM3 ZIPK 60 ARPC1A complex ARHGDIA ARPC5L ITGAE ACTR5 actin P ACTR6 ingly, restored or increased p-MLC2 MHC2 MLC2 ARPC2 40 CALM1 ARPC4 dynamics ITGB7 RHOF MYLK3 CALML4 MYO1A 20 ROCK1 ARHGEF38 was seen in several BRAFi-resistant Actomyosin MYO1H ROCK1 ACTN3 MYO1F MYO15A contractility MYLK4 0 MYH7B WASH5P MYO5C MRTF-A/B MYH3 compared with parental cell lines (Fig- MLCK PPP1R16A RHOT2 SEPT1 PPP1R12A MYO18A MYO1C ARHGEF10 PPP1R12B RDX ure 1I). MEKi did not affect p-MLC2 in SPTAN1 ARHGEF12 SEPT10 MYO1B ITGAV ACTN4 MLC2 MHC2 MLCK ZIPK MRTF MSN SEPT6 MYH11 ARHGEF7 resistant cells (Figure S1I), suggesting DOCK1 ARHGAP35 RHOT1 NF1 ACTR8 PTK2 FAK ITGAL SEPT7 STAT3 that MAPK-independent mechanisms may underlie p-MLC2 restoration. Impor- tantly, cortical p-MLC2 was delocalized after 24-h BRAFi treatment in A375 cells and restored in and S1E) cell lines. These data conﬁrm that myosin II is regulated by MAPK in melanoma. BRAFi-resistant cells (Figure 1J). Phosphoproteomic analysis of Restoration of ERK levels is observed during acquisition of several BRAFi-resistant melanoma cells compared with parental resistance to MAPKi (Konieczkowski et al., 2018; Lito et al., lines showed that cytoskeletal remodeling and Rho GTPase 2012; Obenauf et al., 2015). Twenty-four hours after BRAFi signaling were top enriched processes (Figure 1K; Table S2). reduced p-ERK was accompanied by reduced p-MLC2 (Fig- These data show that MAPK signaling regulates cytoskeletal ure 1G). However, 48 h after BRAFi treatment, p-MLC2 was myosin II and amoeboid behavior. During early responses to restored concomitantly with very modest increase in p-ERK (Fig- treatment, overexpression of myosin II allows melanoma cells ure 1G). These data show that, early after treatment, cells to survive, independently of MAPK activity. remodel their cytoskeleton to recover myosin II activity, resulting in uncoupling of MAPK signaling from actomyosin. ROCK-Myosin II Pathway Is Transcriptionally Rewired We next hypothesized that, under therapy, myosin II during Development of Resistance could play a role in survival of cells with reduced MAPK activity. Transcriptomic alterations drive resistance to MAPK-targeted Strikingly, overexpression of a phosphomimetic MLC2 (TDSD) therapy (Hugo et al., 2015). Transcriptomic data of melanoma (Takaki et al., 2017) increased survival of A375 cells under BRAFi cells at different stages of MAPKi resistance (Figure 2A): 48 h 88 Cancer Cell 37, 85–103, January 13, 2020 MYL9 MYL12A MYL12B MYH9 ROCK2 LIMK2 MYLK DAPK3 MKL1 MKL2 % cytoskeleton-related genes A375 M238 M229 M238 M229 % genes upregulated in Resistant M238 M229 M229 % cell lines with upregulation SKMEL28 M238 M229 SKMEL28 M263 M395p2 M395p1 M397 R4 R5 M238 DTPP M238 SDR SKMEL28 SDR SKMEL28 DDR M229 DDR M229 SDR M263 SDR A375 48 hr BRAFi M238 48 hr BRAFi M238 DTP M229 2d BRAFi M229 DTP M229 DTPP M395p2 SDR M395p1 SDR M397 SDR M249 DDR4 M249 DDR5 BRAFi BRAFi+MEKi A B C “High Myosin II activity” genes A375 A375/PLX/R Log fold change mRNA 2 (Sanz-Moreno 2011) sensitive resistant Resistant/sensitive -+ -+ BRAFi SKMEL28-SDR SKMEL28-DDR -1.0 1.0 - G - G - G - G ROCKi p<0.001 p=0.049 p-MLC2 MYL9 MYL12A MLC myosins MLC2 MYL12B MHC MYH9 p-ERK1/2 ROCK1 ROCKs ERK2 ROCK2 LIMK1 BRAFi-resistance MAPKi-resistance LIMKs GAPDH LIMK2 MKL1 100 33 59 59 100 59 97 35 % MLC2 activity MRTFs MKL2 100 100 12 14 100 100 70 70 % ERK activity vemurafenib 3 months D E F BRAFi sensitive BRAFi resistant (intrinsic) Patient #35 cell line A375 A375/PLX/R WM983A WM983B WM88 SKMEL5 LOX-IMVI WM793B - + - + - + - + - + - + - + BRAFi BRAFi - - G - G ROCKi - G - G - G - G - G - G - G - G - G - G - G - G ROCKi - - p-MLC2 p-MLC2 ROCKi MLC2 MLC2 p-ERK1/2 p-ERK1/2 BRAFi ERK2 ERK2 ROCKi + GAPDH GAPDH BRAFi % MLC2 activity 100 26 39 13 100 20 27 15 100 20 44 16 100 36 68 17 100 50 99 44 100 13 104 17 % MLC2 activity 100 28 99 29 100 100 8 9 100 100 7 9 100 99 10 12 100 100 25 26 100 100 58 60 100 100 30 25 % ERK activity % ERK activity 100 100 65 66 sensitive IC IC 50 50 Loewe BRAFi-Resistant 5.0 A375 0.5 μM WM983A 1.6 μM Antagonism Synergy 4.0 A375/PLX/R 0.06 μM WM983A/PLX/R 0.09 μM 2.0 1.5 1.0 0.5 25 0 0 0.05 0.05 0.1 0.1 0.0001 0.001 0.01 0.1 1 10 100 0.5 0.001 0.01 0.1 1 10 100 0.5 1 1 BRAFi (μM) 5 5 ROCKi (μM) ROCKi (μM) J **** K L A375/ **** **** Patient #35 PLX/R **** **** **** ** **** **** **** ** **** **** - ROCKi BRAFi ROCKi + BRAFi 100 **** 100 **** ROCKi - G - G - G - G - GG - ROCKi - B - B Myosin II inh. - + - + BRAFi - + BRAFi A375 A375/PLX/R BRAFi A375/PLX/R Patient #35 M N **** ns *** ** ** **** **** *** ** ** **** **** 2.0 ** ** 1.0 100 100 100 1.5 80 80 80 0.8 60 60 60 0.6 1.0 40 40 40 40 0.4 0.5 20 20 20 20 0.2 0 0 0 - ++ - rat MLC2 WT - ctrl ROCK1/2 ctrl siRNA ctrl ROCK1/2 ctrl siRNA WT TASA rat MLC2 ctrl MLC2 human siRNA MLC2 human siRNA ROCK1: 61% ROCK1: 66% % mRNA ROCK2: 69% 84% 88% 88% ROCK2: 69% 87% 85% 92% KD vs ctrl Figure 3. Survival of Targeted Therapy-Resistant Melanomas Is Dependent on ROCK-Driven Myosin II Activity (A) Fold change in mRNA levels of ROCK-myosin II pathway genes in A375/PLX/R, Colo829/PLX/R by qRT-PCR (n = 3); and from published RNA sequencing data (Song et al., 2017). (legend continued on next page) Cancer Cell 37, 85–103, January 13, 2020 89 MYL9 MYL12B MYH9 MYL9 MYL12B MYH9 A375/PLX/R A375 Relative melanoma A375/PLX/R Relative melanoma cell survival cell survival Colo829/PLX/R M229 SDR M238 SDR SKMEL28 SDR M229 DDR SKMEL28 DDR Relative melanoma cell survival IC ROCKi (μM) A375 Fold change WM983A in % dead cells (PI ) WM983B Relative melanoma cell survival WM88 % Control Relative melanoma cell survival (Obenauf et al., 2015; Song et al., 2017) or several weeks after 3C and S2A). P-MLC2 in resistant cells was ROCK dependent, treatment (drug-tolerant persisters [DTP], drug-tolerant prolifer- since several unrelated ROCKi (GSK269962A, H1152) (Feng ating persisters [DTPP]) (Song et al., 2017); and resistant cells et al., 2016) reduced p-MLC2 (Figures 3C and S2A). However, after months-years (single-drug resistant [SDR, BRAFi], dou- p-ERK was not affected by ROCKi. ble-drug resistant [DDR, BRAFi + MEKi]) (Song et al., 2017) Sensitive A375 cells lost circularity and became more spindle- were used to analyze changes in 313 manually curated cytoskel- shaped with long, thin protrusions after BRAF inhibition, with etal-related genes (Table S3). Unsupervised hierarchical clus- reduced p-MLC2 (Figures 1B–1G, 3D, and S2B). In contrast, tering classiﬁed melanoma cell lines into two groups (Figure 2B). A375/PLX/R cells did not change morphology after BRAFi treat- Group 1 clustered the majority of cell lines, including 48-h BRAFi ment, while ROCKi decreased their circularity and promoted a (when p-MLC2 was restored [Figure 1G]), DTP, DTPP, and SDR/ collapsed (Sadok et al., 2015) cytoskeleton (Figures 3D DDR stages, which had a signiﬁcant percentage of regulated and S2B). genes (1.5-fold up- or downregulated) compared with baseline/ We expanded these observations to PLX4720-resistant sensitive cell-speciﬁcally upregulated genes (Figures 2C and Colo829 (Figure S2C) and a panel of cell lines sensitive or intrin- 2D). Upregulated in group 1 were genes involved in generation/ sically resistant to BRAFi (Baenke et al., 2015; Konieczkowski maintenance of myosin II-driven contractility (Figure 2E), such et al., 2014)(Figures 3E, S2D, and S2E). Similar results were as myosin (MLC2 genes MYL9, MYL12A/B; and myosin heavy observed in A375 cells resistant to BRAFi dabrafenib + MEKi tra- chain 2 [MYH9]), ROCK2, MLCK (MYLK), ZIPK (DAPK3), metinib (Flaherty et al., 2012; Long et al., 2014b) (A375/D + T/R) LIMK2, and transcriptional co-activator MRTF (MKL1/2), which (Figures S2F and S2G); and in a resistant cell line established directly regulates MLC2 expression (Medjkane et al., 2009). Of from a patient with acquired resistance to BRAFi (patient no. note, myosin II activity promotes myosin II expression to self- 35) (Figures 3F and S2H). perpetuate (Calvo et al., 2013). Because therapy-resistant cells maintain high p-MLC2 These data show that group 1 melanomas adapt to therapy by (Figure 1I) and that myosin II increases survival under therapy rewiring their transcriptome to alter cytoskeletal gene expres- (Figure 1H), we assessed if myosin II could play a role in confer- sion, ultimately restoring myosin II activity. ring a survival advantage to therapy-resistant cells. Reduced p-MLC2 after ROCKi impaired survival of sensitive and BRAFi- Survival of Targeted Therapy-Resistant Melanomas Is resistant melanoma pairs (A375, WM983A, WM983B, WM88) Dependent on ROCK-Driven Myosin II Activity (Figures 3G, 3H, and S3A–S3C). BRAFi-resistant melanomas We next investigated if the ROCK-myosin II pathway could play a were 4- to 30-fold more sensitive to ROCKi GSK269962A (Fig- role in the survival of melanoma cells. Using qRT-PCR, we ures 3G, 3H, S3A, and S3C) and AT13148 (Figure S3A). Moder- conﬁrmed that MLC2 (MYL9, MYL12A/B) and other components ate synergistic effects between ROCKi and BRAFi were of the ROCK-MLC2 pathway (MYH9, ROCK1/2, LIMK, MKL1/2, observed in BRAFi-sensitive A375 cells (Figures 3I and S3D). MYLK) were increased at the mRNA level in BRAFi-resistant cell More pronounced synergy was observed by annexin V/propi- line pairs (A375 and Colo829 cells, Figure 3A). Similar results dium iodide (PI) cell death staining (Figure S3E). were obtained using publicly available data from M229, M238, Importantly, A375/PLX/R cells grown on collagen I had and SKMEL28 cells (Song et al., 2017)(Figure 3A). Gene set increased sensitivity to ROCKi (Figure 3J). We observed enrichment analysis (GSEA) showed that resistant cell lines dis- impaired survival after ROCKi treatment in several models of played similar transcriptomes to cells with high myosin II activity drug resistance: A375/PLX/R, A375/D + T/R, Colo829 and (Figure 3B). BRAFi-intrinsic resistant lines (Figure S3F); and patient no. 35 We compared the impact of MAPK inhibition on myosin II in cells (Figures 3K and S3G). Importantly, the survival advantage sensitive/resistant melanoma cells. P-MLC2 was decreased af- was provided by myosin II itself, since myosin II inhibitor blebbis- ter BRAFi treatment in sensitive but not in resistant A375/PLX/ tatin strongly suppressed survival (Figures 3L and S3H). More- R cells. P-ERK was reduced by BRAFi in sensitive cells (Figures over, siRNA targeting ROCK1/2, MYL9, MYL12B,or MYH9 (B) GSEA comparing high myosin II activity signature (Sanz-Moreno et al., 2011) to resistant cell lines (Song et al., 2017). Nominal p values shown, false discovery rate (FDR) < 0.2. (C) p-MLC2 and p-ERK1/2 immunoblots after 24 h treatment. (D) Images of cells from (C). Scale bar, 50 mm. (E and F) p-MLC2 and p-ERK1/2 immunoblots of sensitive and intrinsically resistant cells (E); and patient no. 35 cells (F) after 24 h treatment (8 h for WM88). Vertical line in diagram (F): cell line establishment. (G) Survival and half maximal inhibitory concentration (IC ) values after a 3-day treatment (n = 3). (H) IC values for GSK269962A. (I) Cell survival as synergy graph of A375 cells treated for 3 days (n = 4). (J) Images and quantiﬁcation of cell survival on collagen I for 9 days (n = 3). Scale bar, 100 mm. (K) Survival of patient no. 35 cells after 10 days (n = 3). (L) Survival after a 5- to 10-day blebbistatin and PLX4720 treatment (n = 3). (M) Survival 8 days after gene depletion by RNAi (n = 3; n = 4 A375/PLX/R myosin genes, patient no. 35 MYL12B, ROCK1/2). mRNA KD (percentage decrease versus control) by RT-PCR shown. (N) Cell death in A375/PLX/R cells 3 days after transient MLC2 KD and rescue with rat MLC2 WT or TASA (n = 3, left graph; n = 4, right graph). (C–K) ROCKi GSK269962A, BRAFi PLX4720. Graphs show mean ± SEM and individual data points (circle). p values by one-way ANOVA with Tukey’s (J, K, and N) or Dunnett’s correction (M, myosin genes); t test with Welch’s correction (L and M, ROCK), **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also Figures S2 and S3. 90 Cancer Cell 37, 85–103, January 13, 2020 A B 1.0 0.8 Patient #3 post 15 C/A 16 C/A Patient 2 Post-RAF/MEKi Pt17-DP avg Pt6-DP avg Patient #5 post 0.6 Patient 5 Post-RAF/MEKi Pt4-DP1 Pt19-DD-DP avg Pt7-DP1 Pt2-DP avg Pt20-DP avg 0.4 Pt8-DP avg Pt3-DP avg Pt15-DP avg Patient #2 post low expression Pt9-DP avg Patient 4 Post-RAF/MEKi 0.2 22 C/A high expression Pt10-DP avg Pt5-DP avg Pt22-DD-DP avg p=0.0226 Pt16-DP1 7 D/A Pt18-DD-DP1 Patient 1 Post-RAF/MEKi 24 D/A Pt24-DD-DP1 0 2 4 6 8 10 years Pt1-DP1 Pt23-DD-DP1 Patient 3 Post-RAFi Patient 6 Post-RAF/MEKi 25 C/A -0.3 0.3 Log fold change Resistant/baseline * D Pt30 Pt4 Pt18 Pt3 Pt44 Pt10 Pt77 Pt27 Pt28 Pt8 Pt62 100 Pt31 Pt5 Pt84 Pt94 Pt11 Pt89 Pt9 Resp NR -0.3 0.3 Log fold change On-anti-PD-1/baseline anti-PD-1 z-score Hugo 2016 Riaz 2017 Sanz-Moreno 2011 Waggle 2014 Hugo 2015 anti-PD-1 anti-PD-1 Myosin II MAPKi MAPKi -2.0 2.0 high low Resistant Baseline Resistant Baseline NR Resp On anti-PD-1 Baseline EMT/metastasis angiogenesis hypoxia wound healing TGF-β STAT3 NF-κB YAP pro-invasive/metastasis F G “High Myosin II activity” genes “High Myosin II activity” genes evading (Sanz-Moreno 2011) IL-8 (Sanz-Moreno 2011) intratumoural p=0.027 p<0.001 sustaining immunity cross- proliferative M2 macrophages resistance Met T cell markers signalling hallmarks T cell activation YAP evading apoptosis MAPKi-resistance anti-PD-1-resistance H IJ K Pre Post Pre Post Pre Post Pre Post Post vs Pre Post vs Pre Post vs Pre Post vs Pre p=0.0062 p=0.035 p=0.014 p=0.047 1.0 0.8 60 60 MAPKi immunotherapy (IT) 0.6 sequential MAPKi + IT 40 0.4 20 20 0.2 0 0 Pre Post Pre Post Pre Post Pre Post Figure 4. High Myosin II Levels Identify Therapy-Resistant Melanomas in Human Samples (A) Heatmap of fold change in expression of ROCK-myosin II pathway genes in MAPKi-resistant versus baseline patient samples from (Hugo et al., 2015; Kwong et al., 2015; Sun et al., 2014; Wagle et al., 2014). (B) Kaplan-Meier overall survival from The Cancer Genome Atlas according to expression of ROCK-myosin II genes (listed in A) (n = 389 melanoma patients). (legend continued on next page) Cancer Cell 37, 85–103, January 13, 2020 91 MYLK MYL9 LIMK2 LIMK1 MYH9 MYL12A MYL12B DAPK3 MKL1 ROCK1 ROCK2 MKL2 MYL9 MYLK LIMK1 MYL12A MYL12B LIMK2 DAPK3 MKL1 MYH9 ROCK1 ROCK2 MKL2 S100 % highest p-MLC2 staining MLC2 (MYL9) FPKM S100 (RNAseq) % Masson’s trichrome area + 2 CD206 cells/mm Overall Survival + + Ratio FOXP3 /CD4 reduced survival in A375/PLX/R and patient no. 35 cells We have previously generated a transcriptional signature for (Figure 3M). The decrease in survival after MLC2 knockdown amoeboid metastatic melanoma cells harboring high ROCK- (KD) was more pronounced in BRAFi-resistant cells (Figure S3I). driven myosin II activity (Cantelli et al., 2015; Sanz-Moreno Therefore, both MLC2 expression and phosphorylation by ROCK et al., 2011). We compared high myosin II signature; MAPK-tar- are required to promote survival of resistant cells. Importantly, geted therapy-resistant signatures (Hugo et al., 2015; Sun et al., RNAi-insensitive rat MLC2 (Calvo et al., 2013) overexpression 2014; Wagle et al., 2014); anti-PD-1/NR signature (Hugo et al., rescued the decreased survival observed after MLC2 depletion. 2016); and on-anti-PD-1-treatment signature (Riaz et al., 2017). This mechanism relied on MLC2 phosphorylation, since rescue Single sample GSEA (ssGSEA) showed that similar gene was impaired by TASA-MLC2 inactive phospho-mutant (Figures signatures are enriched in high myosin II amoeboid cells and 3N and S3J). therapy-resistant patient samples (Figure 4E), including EMT/ Overall, myosin II restoration confers a survival advantage to metastasis, angiogenesis, hypoxia, wound healing, transforming resistant melanomas. growth factor b (TGF-b)-, STAT3-, nuclear factor kB-, and YAP- signaling genes (Table S5). High Myosin II Levels Identify Cross-Resistant Global GSEA analysis showed a signiﬁcant overlap between Melanomas in Human Samples ‘‘high myosin II activity’’ melanoma cells (Cantelli et al., 2015; We next validated our ﬁndings in clinical samples from published Sanz-Moreno et al., 2011) and MAPKi-resistant melanomas datasets (Hugo et al., 2015; Kakavand et al., 2017; Kwong et al., with immunosuppressive macrophages, and pro-invasive and 2015; Long et al., 2014a; Rizos et al., 2014; Song et al., 2017; Sun pro-survival features (Hugo et al., 2015)(Figure 4F). There was signiﬁcant overlap between high myosin II and anti-PD-1/NR et al., 2014; Wagle et al., 2014)(Table S4). There was a subset of melanoma tumors (50%) with upregulation of ROCK-myosin II patient signatures (IPRES [Hugo et al., 2016]) (Figure 4G). pathway genes (Figures 4A, S4A, and S4B), in accordance with Myosin II-driven contractility is regulated by MLC2 gene data with resistant cell lines (Figure 2E). The Cancer Genome expression and phosphorylation/activity (Calvo et al., 2013; Atlas data showed that higher levels of ROCK-myosin II genes Medjkane et al., 2009; Olson, 2008). We assessed p-MLC2 levels in treatment-naive melanoma patients confer worse prognosis in paired patient melanoma sections before and after therapy (Figure 4B). MAPKi-resistant tumors quickly progress after (targeted therapy, immunotherapy [IT], or sequential targeted relapse (Wagle et al., 2011), indicative of aggressiveness. We and IT; Table S6). P-MLC2 levels were higher in all resistant tu- suggest that melanomas with intrinsically higher expression of mors after treatment (Figures 4H and S4D–S4G). Speciﬁcity of the ROCK-myosin II pathway are more aggressive and prone p-MLC2 antibody was validated by RNAi (Figure S4D). Collagen to develop resistance. density promotes myosin II activity (Laklai et al., 2016; Paszek Innately anti-PD-1-resistant (IPRES) tumors harbor a tran- et al., 2005), and ROCK-myosin II induces ECM stiffening scriptional signature of upregulated genes involved in the (Samuel et al., 2011). Increased ECM deposition was observed regulation of EMT, cell adhesion, ECM remodeling, angiogen- in resistant compared with pre-treatment samples (Figures 4I esis, and hypoxia (Hugo et al., 2016). MAPK-targeted and S4E–S4G). Melanoma cells with high ROCK-myosin II are therapies in melanoma induce similar signatures with immuno- highly secretory and polarize macrophages to tumor-promoting suppressive features (Hugo et al., 2015). These studies (CD206 ) phenotypes (Georgouli et al., 2019). Interestingly, suggest that non-genomic MAPKi resistance driven by tran- CD206 cells were increased in resistant compared with pre- scriptional upregulation of metastasis-related pathways medi- treatment samples (Figures 4J and S4E–S4G), correlating with ates cross-resistance to anti-PD-1 therapy. They also suggest higher p-MLC2 (Figure 4H). Immunosuppressive FOXP3 regula- that aggressive tumors resistant to one therapy (e.g., MAPKi) tory T cells (Tregs)/CD4 ratio was also increased in resistant will likely not respond to second therapy (anti-PD-1). There- samples (Figures 4K and S4F–S4G). These data suggest that fore, we next investigated if ROCK-myosin II could predict high MLC2 (MYL9) expression and/or activation (p-MLC2) in anti-PD-1 responses as part of a cross-resistance mechanism. melanoma cells together with immunosuppressive populations Samples before anti-PD-1 treatment (Hugo et al., 2016) and higher collagen densities identify therapy-resistant mela- showed higher MYL9 expression in non-responding (NR) nomas, suggesting their potential as biomarkers. than in responding (Resp) patients (Figure 4C). Increased Overall, resistant tumors and melanomas with high myosin II levels of ROCK-myosin II pathway genes were detected in a activity harbor a similar transcriptome. Importantly, ROCK- large subset of patients on anti-PD-1 treatment (Riaz et al., myosin II could be a key mediator of non-genomic cross- 2017)(Figures 4Dand S4C). resistance. (C) MYL9 mRNA in Resp (n = 15) and NR (n = 13) anti-PD-1 patients from (Hugo et al., 2016). Boxplot: median (center line); interquartile range (box); min-max (whiskers). (D) Heatmap of fold change in expression of ROCK-myosin II genes in on-anti-PD-1 versus baseline patient samples (Riaz et al., 2017). (E) Heatmaps show ssGSEA of cross-resistance gene signatures (NR, non-responder; Resp, responder). (F and G) GSEA comparing ‘‘high myosin II activity’’ signature (Sanz-Moreno et al., 2011) to a subset of MAPKi-resistant patient samples from (Hugo et al., 2015) (F) or anti-PD-1/NR samples (Hugo et al., 2016) (G). Chart pie in (F) with cross-resistance hallmarks from (Hugo et al., 2015). Nominal p values shown, FDR < 0.001 (F) and 0.145 (G). (H–K) Images (patient no. 17) and quantiﬁcation in 12 paired samples before and after therapies (including those in Figures S4E and S4F) of: p-MLC2 (% cells with + + highest score), melanoma marker S100 (inset) (H); Masson’s trichrome staining (percentage stained area/section) (I); CD206 cells (J); FOXP3 cells (K). Scale bars, 100 mm. p values by Mann-Whitney test (C, H–K). See also Figure S4 and Tables S4, S5, and S6. 92 Cancer Cell 37, 85–103, January 13, 2020 V600E anti-PD-1 BRAF V600E A B C 2x/week BRAF anti-PD-1/NR Patient #26 +ROCKi in vitro cell lines days 7 7 #26 #26/R x2 anti-PD-1 **** C57BL/6J **** 5555 4434 #26 #26/R **** **** anti-PD-1/ anti-PD-1/ before treatment anti-PD-1/Resistant **** **** Resp NR Resp NR * *** #26 #26/R **** **** **** 75 **** - G H - G H ROCKi 100 p-MLC2 80 80 MLC2 60 60 p-ERK1/2 40 40 ERK2 - G H - G H ROCKi 20 20 GAPDH 0 0 - G - G ROCKi - G - G ROCKi **** D E F V600E BRAF K601E 100 WT WT Patient #58 Ipi/R BRAF / NRAS BRAF vemurafenib ipilimumab pembrolizumab DTIC ipilimumab no therapy ipilimumab DTIC 3 months 3 months 1 month - ROCKi 1 month 4 months 1 month 1 month 3 months Patient #62T3 cell line Patient #33 cell line Patient #58 cell line 25 - + BRAFi - G - G ROCKi - G ROCKi - G ROCKi - ROCKi p-MLC2 p-MLC2 p-MLC2 **** Patient #33 Ipi/R MLC2 MLC2 MLC2 - ROCKi p-ERK1/2 p-ERK1/2 p-ERK1/2 ERK2 ERK2 ERK2 GAPDH GAPDH GAPDH 100 41 % MLC2 activity 100 33 % MLC2 activity % MLC2 activity 0 100 12 96 13 100 100 % ERK activity 100 100 % ERK activity - ROCKi 100 100 80 77 % ERK activity V600E BRAF G H vemurafenib ipilimumab I **** **** 5 months 2 months *** *** **** *** ns ns Patient #2 cell line **** *** 1.0 - ROCKi - + BRAFi 0.8 - G - G ROCKi 0.6 p-MLC2 0.4 MLC2 0.2 p-ERK1/2 ERK2 - GG - ROCKi - GG - ROCKi - + BRAFi - + BRAFi GAPDH 100 28 134 23 % MLC2 activity % ERK activity 100 100 82 86 J K L Patient #2 BRAFi+Ipi/R Patient #62T3 BRAFi+Ipi+Pembro/R Patient - ROCKi **** **** #2 #62T3 **** **** **** **** ** ** **** **** 100 100 100 100 80 80 80 80 60 60 60 40 40 40 **** 20 20 20 20 **** **** 100 0 0 0 0 ctrl ROCK1/2 ctrl siRNA ctrl ROCK1/2 ctrl siRNA - B - B Myosin II inh. 50 ROCK1: 83% ROCK1: 66% BRAFi % mRNA ROCK2: 77% 84% 97% 92% ROCK2: 59% 79% 89% 73% KD vs ctrl - G - G ROCKi BRAFi - + Figure 5. ROCK-Driven Myosin II Activity in Immunotherapy-Resistant Melanoma (A) Top, schematic of cell lines. Bottom, p-MLC2 immunoblots after treatment: n = 7 (G); n = 3 (H). (B) Survival of patient no. 26 cells treated for 10 days (n = 4). (legend continued on next page) Cancer Cell 37, 85–103, January 13, 2020 93 MYL9 MYL12B MYH9 MYL9 MYL12B MYH9 Relative spheroid- BRAFi - forming ability Relative melanoma cell survival Relative melanoma cell survival Relative melanoma cell survival Relative melanoma cell survival Relative melanoma cell survival Relative melanoma cell survival on collagen I BRAFi - Cell morphology on collagen I ROCK-Driven Myosin II Activity in Immunotherapy- no. 62T3, BRAFi did not affect p-MLC2, while ROCKi decreased Resistant Melanoma p-MLC2 in patient no. 2 cells (Figures 5H and S5I). Patient no. 2 Next we investigated whether survival of immunotherapy-resis- cells on collagen I displayed very rounded morphology even in tant melanomas could be dependent on ROCK-myosin II. To the presence of BRAFi, indicative of high p-MLC2 (Figure 5I). test this hypothesis in vitro, we used patient no. 26-derived cells ROCKi decreased circularity and induced very thin protrusions established pre- and post-anti-PD-1 resistance (Figure 5A). Both and a spindle-shaped morphology in patient no. 2 cells. A com- cell lines rely on ROCK to sustain p-MLC2 (Figures 5A and S5A). mon event during melanoma resistance is BRAFi/MEKi addic- Importantly, anti-PD-1/resistant cells were 2-fold more sensitive tion, which occurs when resistant melanomas become drug to ROCKi (Figure 5B). Increased sensitivity was further conﬁrmed dependent (Das Thakur et al., 2013; Hong et al., 2017; Kong in a resistant brain metastasis-derived cell line from patient no. et al., 2017; Moriceau et al., 2015; Sun et al., 2014). Patient no. V600E 26 (data not shown). We then grafted mouse Braf melanoma 2 cells displayed addiction to BRAFi on 2D cultures (Figure S5J), cell lines 5555 and 4434 cells (Dhomen et al., 2009) subcutane- but treatment with ROCKi impaired survival in the presence of ously onto fully immunocompetent C57BL/6J mice and treated BRAFi and further decreased survival upon BRAFi withdrawal with anti-PD-1, which led to variable responses. We isolated (Figure S5J). This agrees with data on BRAFi-resistant patient NR and Resp tumors and grew them ex vivo (Figures S5B and no. 35 and Colo829/PLX/R cells (Figures 3K, S3F, and S3G), S5C). Increased intrinsic sensitivity to ROCKi in vitro was found which also displayed varying degrees of BRAFi addiction. Inter- in anti-PD-1/NR-derived cells (Figure 5C), similar to the resistant estingly, patient no. 2 cells grew as compact spheroids on human cell lines (Figure 5B). As melanoma cells activate an im- collagen I under BRAFi treatment, but growth was abrogated mune-evasion program they also trigger cytoskeletal remodel- by ROCKi (Figures 5J and S5K), showing that myosin II drives ing, rendering them intrinsically vulnerable to ROCK-myosin II survival in BRAFi-addicted cells. Accordingly, myosin II inhibition inhibition. with blebbistatin or RNAi against ROCK or myosin II genes Using additional cell lines established from human melanomas impaired survival of patient no. 2 and no. 62T3 cells (Figures resistant to immunotherapy (patients no. 58 and no. 33), we 5K, 5L, and S5L). conﬁrmed that these melanomas harbored ROCK-dependent MRTF controls MLC2 expression (Medjkane et al., 2009) while p-MLC2 levels (Figures 5D and S5D). Cell survival was impaired MRTF activity is regulated by actin dynamics (Posern and Treis- after treatment with several ROCKi on 3D (Figures 5E and S5E) man, 2006). Expression of MRTF (MKL) was increased in resis- and 2D culture (Figure S5F). tant melanomas (Figures 2E and 4A) and its depletion impaired Our data predict that cells that do not respond to MAPKi––if BRAFi-resistant cell survival (Figure S5M). Accordingly, MYL9 they undergo cross-resistant transcriptional rewiring of their mRNA levels decreased after MRTF depletion (Figure S5M). cytoskeleton––they will not respond to immunotherapy either. Overall, melanomas with acquired and primary resistance to Such cross-resistance will be susceptible now to ROCKi. Patient targeted and immunotherapies rely on myosin II activity for their no. 62T3 cell line was established from a tumor with acquired survival. Consistently, p-MLC2 levels and cancer cell survival resistance to BRAFi and developed primary resistance to anti- were positively correlated in resistant lines (Figure S5N). CTLA-4 and anti-PD-1 (Figure 5F). After BRAF inhibition, p-MLC2 was not affected in these cells, while ROCK inhibition ROCK-Myosin II Inhibition Induces Lethal Reactive decreased p-MLC2 (Figures 5F and S5G). Similar to our previous Oxygen Species, DNA Damage, and Cell-Cycle Arrest data, survival of patient no. 62T3 cells was impaired with ROCKi We next investigated why resistant cells rely on myosin II for sur- (Figures 5G and S5H). vival. Resistant cells (Song et al., 2017) were enriched in oxida- Similarly, patient no. 2 cells were established from a tumor that tive stress and reactive oxygen species (ROS) metabolism never responded to targeted and immunotherapy (Figure 5H). gene signatures (Figure 6A) and had lower DNA damage repair The post-treatment-resistant biopsy had higher p-MLC2 gene expression (Figure 6B). Interestingly, ROCK-myosin II sup- compared with baseline tumor (Figure S4F). Similar to patient presses high ROS in migrating cells (Herraiz et al., 2015). ROCKi (C) Top, schematic of experiment. Bottom, survival of 4434 and 5555 anti-PD-1/non-responder (NR) lines versus responder (Resp) after a 3-day treatment (n = 3, 5555; n = 4, 4434). (D) p-MLC2 and p-ERK1/2 immunoblots of patient no. 58 (n = 4) and no. 33 (n = 3) cells after treatment. (E) Images and quantiﬁcation of cell survival on collagen I for 7 days (n = 3). (F) p-MLC2 and p-ERK1/2 immunoblots of patient no. 62T3 cells after treatment (n = 3). (G) Survival of patient no. 62T3 cells after a 10-day treatment (n = 3). (H) p-MLC2 and p-ERK1/2 immunoblots of patient no. 2 cells after treatment (n = 5). (I) Cell morphology of patient no. 2 cells on collagen I after treatment. n = 70 cells (dots) from 2 experiments. Scale bar, 50 mm. (J) Survival of patient no. 2 cells as spheroid-forming ability on collagen I for 16 days (n = 3); Scale bar, 100 mm. (K) Survival after a 10-day blebbistatin treatment (n = 3). (L) Survival 8 days after gene depletion by RNAi (n = 3; n = 4 patient no. 2 MYL12B-ROCK1/2, no. 62T3 MYL9; n = 5 no. 62T3 MYL12B). Average percentage mRNA KD (percentage decrease versus control) by qRT-PCR is shown. Vertical line in (D, F, and H): cell line establishment. (A–J) ROCKi GSK269962A, H1152; (F–K) BRAFi PLX4720. (A, D, F, H, and I) 24 h treatment. Graphs show mean ± SEM and individual data points (circle) except boxplot in I (median, center line; interquartile range, box; min-max, whiskers). p values by one-way ANOVA with Tukey’s (B, C, G, and J) or Dunnett’s correction (L, myosin genes); Kruskal-Wallis with Dunn’s correction (I), t test with Welch’s correction (E, K, and L, ROCK); **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also Figure S5. 94 Cancer Cell 37, 85–103, January 13, 2020 A B Role of BRCA1 in DNA Damage Response Cell Cycle Control of Chromosomal Replication GO_Positive_Regulation_Of_Reactive_Oxygen_Species_Metabolic_Process Aryl Hydrocarbon Receptor Signaling GO_Response_To_Reactive_Oxygen_Species DNA Double-Strand Break Repair by Homologous Recombination GO_Regulation_Of_Reactive_Oxygen_Species_Metabolic_Process DNA Double-Strand Break Repair by Non-Homologous End Joining ATM Signaling GO_Response_To_Oxidative_Stress NER Pathway Houstis_ROS BER pathway Response_To_Oxidative_Stress p53 Signaling GO_Positive_Regulation_Of_Reactive_Oxygen_Species_Biosynthetic_Process All-trans-decaprenyl Diphosphate Biosynthesis 02 468 10 -Log(p value) -Log(p value) C D **** ** * *** ** *** * ** *** **** 100 100 3 * multinucleated sR 75 G2/M - G - G ROCKi 2 50 p−H2A.X G1 dead GAPDH 25 25 0 0 0 0 sR sR s R sR sR sR sR sR - G - G - G - G ROCKi - 0.05 0.1 0.25 - 0.05 0.1 0.5 μM ROCKi sR sR Patient cells A375/ WM EF G PLX/R 793B #35 #2 #58 ** ** ns ns ** ** ** * **** **** 1.0 **** **** 1.0 **** ** **** ** 0.8 0.8 0.6 multinucleated 0.6 0.4 G2/M 0.2 0.4 0.2 G1 - G ROCKi dead Mcl-1 0 - G - G - GG - - G ROCKi - GG - - BB + GAPDH p-STAT3 - ++ - BRAFi STAT3 GAPDH H *** BRAFi ns ** 80 - G - G ROCKi ** *** 5 5 5 5 10 3.5% 4.6% 10 5.1% 16.9% 10 4.1% 3.2% 10 5.7% 21.3% 4 4 4 4 10 10 10 10 3 3 3 3 40 10 10 10 10 90.6% 1.3% 59.3% 18.8% 90.7% 2.1% 48.9% 23.9% 2 2 2 2 10 10 10 10 0 0 0 0 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 010 10 10 10 010 10 10 10 010 10 10 10 010 10 10 10 - GG - ROCKi Annexin V BRAFi -+ Patient #2 BRAFi+Ipi/R Patient #62T3 BRAFi+Ipi+Pembro/R ** **** ns ns *** 60 *** **** 20 20 -GG - ROCKi - GG - ROCKi -+ BRAFi - + BRAFi Figure 6. ROCK-Myosin II Inhibition Induces Lethal ROS, DNA Damage, and Cell-Cycle Arrest (A) GSEA of ROS/oxidative stress-related gene signatures in MAPKi-resistant versus sensitive cell lines (group 1) from (Song et al., 2017). Dashed line indicates statistical signiﬁcance. (B) The 10 most enriched canonical pathways in downregulated genes in MAPKi-resistant cell lines (group 1) from (Song et al., 2017). (C) Left, ROS levels in A375 (s) and A375/PLX/R (R) cells after treatment (n = 6). Right, quantiﬁcation of p-H2A.X immunoblots (n = 7). (legend continued on next page) Cancer Cell 37, 85–103, January 13, 2020 95 Relative ROS levels (mean fluorescence int.) % cells % dead cells % dead cells Relative p−H2AX levels Propidium iodide Relative p-STAT3/STAT3 levels % dead cells % cells Relative Mcl-1 protein levels induced higher levels of ROS (Figures 6C and S6A) and phos- High myosin II activity provides an advantage during early sur- phorylated H2A.X (p-H2A.X), indicative of DNA damage (Fig- vival in the lung, which is a limiting step in the metastatic process ure 6C), in BRAFi-resistant cells compared with sensitive cells. (Cantelli et al., 2015; Medjkane et al., 2009; Orgaz et al., 2014b; Resistant cells had lower expression of genes of the base exci- Sanz-Moreno et al., 2008, 2011). Many of the cross-resistance sion repair pathway (Figure S6B) that repairs ROS-mediated gene signatures were related to metastatic programs (Figure 4E). DNA damage (Krokan and Bjoras, 2013). Survival of patient no. 2 cells in the lung after tail vein injection Because BRAFi-resistant cells harbor higher ROS and have was improved after pre-treatment in vitro with BRAFi (Figure 7D). lost DNA damage repair machinery, ROCKi increases ROS However, when pre-treated with ROCKi, survival was impaired levels leading to unrepaired DNA damage. Unrepaired DNA (Figure 7D). Patient no. 35 BRAFi-addicted cell line (Figure 3K) damage can induce cell-cycle arrest that, if prolonged, can showed reduced growth and p-MLC2 levels in vivo after ROCKi lead to cell death (Shaltiel et al., 2015). Blocking myosin II ac- (Figures 7E and S7C). tivity using ROCKi resulted in a pronounced dose-dependent High myosin II activity cells (Cantelli et al., 2015; Sanz-Moreno cell-cycle arrest in BRAFi-resistant melanomas (Figures 6D et al., 2011) and MAPKi-resistant melanomas with immunosup- and S6C). Blebbistatin caused very similar results in resistant pressive features and pro-tumorigenic macrophages (Hugo cells (Figure 6E). As a result of ROS-DNA damage, resistant et al., 2015) display transcriptional overlap (Figure 4F). We as- cells suffer G2-M arrest and multinucleation. Accordingly, sessed myosin II activity and immunosuppressive populations time-lapse video microscopy showed that cells suffering cell- in A375/PLX/R xenografts (Figure 7A). ROCKi-treated tumors cycle arrest died after 72 h (Figure S6D). had reduced p-MLC2 (Figure 7F) and lower number of CD206 ROS production is counterbalanced by STAT3 (Poli and Cam- macrophages (Figure 7G), which could contribute to reduced poreale, 2015) and both high myosin II activity and resistant cells tumor growth. ROCKi decreased polarization to CD206 macro- harbor high STAT3 signaling (Figure 4E). ROCKi decreased phages as F4/80 content was not affected (Figure S7D), only + + p-STAT3 levels and its pro-survival target Mcl-1 in both targeted CD206 /F4/80 ratio (Figure 7G). ROCK-myosin II inhibition therapy- and immunotherapy-resistant cells (Figures 6F and 6G). could overcome cross-resistance to targeted/immunotherapies Moreover, we measured decreased survival in A375/PLX/R cells via intrinsic cell survival and extrinsic myeloid co-option. after 72 h of ROCKi treatment using 3-(4,5-dimethylthiazol-2-yl)- 2,5-diphenyl tetrazolium bromide assay (Figure S6E). Annexin ROCK-Myosin II Inhibition Improves Efﬁcacy of Immune V/PI staining (Figure S6F) showed increased cell death after Checkpoint Inhibitors ROCKi treatment in A375/PLX/R (Figures 6H and S6G), patient As high myosin II identiﬁes anti-PD-1/NR, we tested whether no. 2 (Figures 6I and S6H), no. 62T3 (Figure 6I), and no. 35 cells ROCKi could be given as combination therapy to improve (Figure S6I). response to anti-PD-1. We allografted treatment-naive 5555 Therefore, ROCK-driven myosin II protects tumor cells from cells into immunocompetent mice. Anti-PD-1 combined with toxic ROS levels, enabling correct cell-cycle progression and ROCKi (combo) induced signiﬁcantly more regressions of estab- providing pro-survival signals. Because resistant cells have lished tumors compared with single treatments (Figures 8A and altered ROS and loss of DNA damage repair genes, ROCK- S8A), and treatments were well tolerated based on weight (Fig- myosin II inhibition is particularly detrimental. ure S8B). ROCKi-treated tumors had reduced p-MLC2 after 5 days of treatment or at endpoint (Figures 8B and S8C). ROCKi Combining ROCK Inhibitors with BRAF Inhibitors In Vivo also decreased immunosuppressive cell populations at both To translate our ﬁndings to pre-clinical in vivo models, we com- 5 days and endpoint: CD206 macrophages (Figures 8C and + + bined BRAFi and ROCKi (low dose) GSK269962A in BRAFi- S8D) and FOXP3 Tregs (Figures 8D and S8D). F4/80 (Fig- + + + resistant A375/PLX/R xenografts in nude mice. Mice tolerated ure S8D) and other immune populations (CD3 , CD4 , CD8 drug treatments well (Figure S7A). The combination treatment cells) were not signiﬁcantly affected by ROCKi in tumors or was the most efﬁcient and induced regression of tumors and spleens (Figures S8E and S8F). ROCKi did not affect percentage + + improved mouse survival (Figures 7A and S7B). of CD4 and CD8 cells expressing PD-1 (data not shown). Patient no. 2 cells displayed PLX4720 addiction in vitro (Fig- CD206 polarization mainly occurred in tumors since polariza- ures 5J, S5J, and S5K) and also in vivo (Figure 7B), as seen by tion in the spleens was less than 1% (Figure S8F). increased growth in the presence of PLX4720. ROCKi reduced We analyzed inﬁltration in the tumor body (TB) and invasive + + growth and p-MLC2 levels of PLX4720-resistant patient no. 2 xe- front (IF) and found that TB were inﬁltrated with CD3 , CD4 , nografts (Figures 7B, 7C, and S7C). and CD8 cells––but mostly accumulated in the IF––while ROCKi (D and E) Cell-cycle analysis after treatment (n = 3–4). Sensitive (s)-resistant (R) pairs (D, left A375; right WM983A). A375/PLX/R (E); G, GSK269962A; B, bleb- bistatin. (F) p-STAT3 levels after treatment (n = 3 patient no. 35, WM793B; n = 4 A375/PLX/R; n = 5 patients no. 2 and 58). (G) Mcl-1 levels of A375/PLX/R cells after treatment (n = 3). (H and I) Percentage of dead cells by annexin V/PI staining of A375/PLX/R (H), patient no. 2 and no. 62T3 (I) cells after a 3-day treatment (n = 4 A375/PLX/R; n = 5 patient no. 2; n = 4 patient no. 62T3). (C, D, and F–I) ROCKi GSK269962A; (E, H, and I) BRAFi PLX4720. (C–G) 24 h treatment. (C and F–I) Mean ± SEM and individual data points (circle). Asterisks in (D and E) are statistical signiﬁcance in multinucleated cells. p values by one-way ANOVA with Tukey’s (D–F, H, and I) or Benjamini, Krieger, and Yekutieli correction (C); unpaired t test with Welch’s correction (G), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also Figure S6. 96 Cancer Cell 37, 85–103, January 13, 2020 ROCKi Figure 7. Combining ROCK Inhibitors with daily A375/PLX/R % regression A BRAFi BRAFi vs Combo p<0.0001 BRAF Inhibitors In Vivo nude days Veh vs Combo p<0.0001 *** 14 7 ROCKi vs Combo p<0.0001 x2 *** (A) Top, schematic of experiment. Left, growth of Vehicle vs Combo p<0.001 Vehicle ROCKi BRAFi Combo A375/PLX/R xenografts in nude mice after treat- 600 600 600 600 0% 0% 0% 40% regression 400 ment. Middle, Kaplan-Meier survival plot. Right, tumor volume at endpoint (n = 4–6 mice/group). 400 400 400 400 50 (B) Left, volume of patient no. 2 xenografts in NSG 200 200 200 200 0 mice after a 21-day treatment (n = 7 mice/group). 0 5 10 15 20 Days of treatment - GG - ROCKi Right, tumor growth at endpoint versus baseline. 0 0 0 5 10 15 05 10 15 0 5 10 15 05 10 15 - BRAFi (C) p-MLC2 staining in patient no. 2 xenografts. Days of treatment Scale bar, 100 mm. % regression (D) Survival of patient no. 2 cells in the mouse lung ROCKi BRAFi vs Combo p<0.0001 B C Patient #2 daily 24 h post-injection (n = 8–9 mice from 2 experi- BRAFi BRAFi vs Veh p<0.0001 BRAFi vs ROCKi p<0.0001 days ments). Scale bar, 100 mm. Veh vs ROCKi p=0.037 30 7 * NSG x3 ns Veh vs Combo ns (E) Left, volume of patient no. 35 xenografts in - ROCKi ROCKi vs Combo p=0.033 ** ** * *** 2 28% 43% 0% 28% regression NSG mice after a 10-day treatment (n = 6 mice/ **** ** group). Right, p-MLC2 staining. Scale bar, 1 200 100 mm. (F and G) Images and quantiﬁcation of p-MLC2 (F), CD206 (G) in A375/PLX/R xenografts from -1 + + 50 0 (A). Scale bars, 100 mm. Ratio of CD206 /F4/80 - GG - ROCKi shown. (F and G) Pooled data from 2 experiments. - GG - ROCKi -2 - BRAFi Vehicle ROCKi BRAFi Combo - BRAFi (A–G) ROCKi GSK269962A, BRAFi PLX4720. Boxplots show median (center line); interquartile ROCKi Patient #35 daily D E range (box); min-max (whiskers); and individual BRAFi Patient #2 Green CMFDA mice (circles). p values by ANOVA with Tukey’s: days NSG in vitro 24 hr 10-16 7 3 NSG (A) right graph, (B) left graph; Benjamini, Krieger, ROCKi *** BRAFi ** ns and Yekutieli (C, D, F, G, and E, right) or Dunnett’s **** - ROCKi *** * * *** *** 600 correction (E, left), Mantel-Cox (A, survival plot), **** **** - ROCKi chi-square test: percentage regressions in (A, left) and (B, right). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also 1 Figure S7. 0 0 - GG - ROCKi - GG - - GG - ROCKi ROCKi - BRAFi - BRAFi - BRAFi ROCKi ROCKi F G daily A375/PLX/R BRAFi daily A375/PLX/R BRAFi (Figures 8E, 8G, and S8H). After 7 days, nude days nude 14 7 days we treated with anti-PD-1, ROCKi, or 14 7 x2 x2 both. Tumors on anti-PD-1 grew rapidly *** * - ROCKi - ROCKi ** ** ns but combo therapy resulted in >40% 1500 1.5 *** ** **** * * **** **** 300 **** ** regression of established tumors and 1000 1.0 improved survival (Figures 8G and S8H). Treatments were tolerated (Figure S8I) 500 0.5 and ROCKi reduced p-MLC2 (Figure 8H, 0 left). Importantly, combo decreased 0 0 - GG - ROCKi - GG - - GG - ROCKi - BRAFi expression of immune checkpoint ligand - BRAFi - BRAFi PD-L1 on tumor cells (Figure 8H, right). Anti-PD-1/NR tumors polarized most macrophages into CD206 phenotype (Figure S8J) and combo did not alter distribution (Figure S8G). Moreover, ROCKi did not + + affect viability of CD8 T cells or tumor-killing ability in vitro (data decreased expression of PD-L1 on CD206 macrophages + + not shown). Therefore, ROCKi does not affect CD4 and CD8 (Figure 8I, left), while total macrophage content did not change cell functions tested. (Figure S8K). Finally, combo decreased Tregs (Figure 8I, right), We next analyzed anti-PD-1/NR and Resp tumors (Figures 8E while other immune populations did not change (Figure S8K). and 8F). NR had increased levels of p-MLC2 and CD206 cells ROCK-myosin II regulates TGF-b secretion from amoeboid compared with Resp while on anti-PD-1 treatment (Figure 8F, melanoma cells (Cantelli et al., 2015). TGF-b is a potent immuno- middle). FOXP3 Tregs did not change (Figure 8F, right). NR tu- suppressor that induces Tregs and myeloid-derived suppressor mors polarized most macrophages into CD206 compared with cells (Cantelli et al., 2017; Condamine et al., 2015; Nakamura less polarization in Resp (Figure 8F, right) and in parental 5555 et al., 2001). Therefore, ROCKi decreased TGF-b1 levels (Figure 8C). These data could in part explain the lack of response secreted by immunotherapy-resistant patient-derived cell lines to anti-PD-1 (Figure 8F, left). and by 5555 cells (Figure S8L). Interleukin-6, CCL2, TGF-b1, Then an anti-PD-1/NR (intrinsic resistance) was allografted and colony-stimulating factor 1/macrophage colony-stimulating into new recipient mice that were treated with anti-PD-1 factor immunomodulatory cytokines (Fisher et al., 2014; Manto- twice a week post-injection to maintain resistance in vivo vani et al., 2004; Qian et al., 2011; Roca et al., 2009) regulated by Cancer Cell 37, 85–103, January 13, 2020 97 BRAFi - 3 BRAFi Tumor volume (mm ) 3 Tumor volume (mm ) Survival in lung (fluorescence area/field) Log fold change in (arbitrary units) 2 tumor growth vs baseline p-MLC2 score Tumor volume (mm ) BRAFi - % survival BRAFi - BRAFi - + 2 CD206 cells/mm Tumor volume (mm ) p-MLC2 score + + Ratio CD206 / F4/80 p-MLC2 score ROCKi daily anti-PD-1 2x/week ** ROCKi - ns % regresssion C57BL/6J days 300 ** Anti-PD-1 vs Combo p=0.02 7-14 77 Veh+IgG vs Combo p<0.0001 randomize Veh+IgG vs Anti-PD-1 p=0.066 10% 5% 19% regression 33% -5 - GG - ROCKi -15 IgG anti-PD-1 Vehicle + IgG ROCKi anti-PD-1 Combo C D **** - ROCKi **** ** ns - ROCKi ** ns **** 150 1.0 ** ** ** * * 0.8 0.6 0.4 500 50 0.2 - GG - - GG - ROCKi ROCKi - GG - ROCKi IgG anti-PD-1 IgG anti-PD-1 IgG anti-PD-1 All on ROCKi daily 5555 anti-PD-1 anti-PD-1 2x/week 5555 anti-PD-1 2x/week anti-PD-1/NR allograft onto days analysis new recipient mice 7 7 7 77 x4 IHC days C57BL/6J randomize randomize IgG 800 * p=0.06 ns Anti-PD-1 p=0.057 3000 * 3000 1.0 80 0.8 200 2000 2000 NR (intrinsic) 0.6 0.4 100 1000 NR (acquired) 200 20 0.2 Resp (stable) Resp (regression, no tumor) 0 0 0 0 0 0 5 10 15 20 NR Resp NR Resp 0 NR Resp NR Resp NR Resp Days of treatment Anti-PD-1 Anti-PD-1 Anti-PD-1 Anti-PD-1 Anti-PD-1 % regression G H Anti-PD-1 vs Combo p<0.0001 ** Veh+IgG vs Combo p<0.0001 ns 300 ** 11% 14% 12.5% 43% regression 4 ** ** Vehicle + IgG ROCKi Anti-PD-1 200 Combo 75 200 -2 -10 010 20 30 40 -12 0 0 Days of treatment - GG - - GG - ROCKi Vehicle + IgG ROCKi anti-PD-1 Combo IgG anti-PD-1 IgG anti-PD-1 -ROCKi FOXP3 I CD206 Treg - ROCKi Mφ ** ** ns ** ** ** * PD-L1 - GG - ROCKi - GG - ROCKi PD-L1 / CD206 IgG anti-PD-1 IgG anti-PD-1 Figure 8. ROCK-Myosin II Inhibition Improves Efﬁcacy of Immune Checkpoint Inhibitors (A) Top, schematic of treatment. Bottom, growth of 5555 allografts in C57BL/6J mice after treatment. Pooled data from 3 experiments (n = 6–8 mice/group/ experiment). (legend continued on next page) 98 Cancer Cell 37, 85–103, January 13, 2020 Tumor volume (mm ) anti-PD-1 - anti-PD-1 - Log fold change tumor growth vs baseline Log fold change tumor growth vs baseline + 2 PD-L1 score on CD206 CD206 cells/mm p-MLC2 score + + Ratio CD206 / F4/80 + 2 FOXP3 Treg cells/mm % survival + 2 CD206 cells/mm anti-PD-1 - anti-PD-1 - + 2 F4/80 cells/mm p-MLC2 score + + Ratio CD206 / F4/80 anti-PD-1 - p-MLC2 score + 2 FOXP3 Treg cells/mm PD-L1 score (tumor cells) + 2 FOXP3 Treg cells/mm ROCK-myosin II (Georgouli et al., 2019; Le Dreau et al., 2010) etal features are observed in metastatic lesions compared with were upregulated in group 1 MAPKi-resistant melanomas primary tumors (Cantelli et al., 2015; Herraiz et al., 2015; Orgaz (Figure S8M). Therefore, blocking ROCK-myosin II reduces et al., 2014b; Sanz-Moreno et al., 2011), which suggests that immunosuppressive microenvironments, improving anti-PD-1 metastatic traits can be linked to drug resistance (Alexander action on pre-existing T cells (Mariathasan et al., 2018; Tauriello and Friedl, 2012). Pathways controlling invasion and metastasis et al., 2018). are aberrantly activated by non-mutational mechanisms––over- expression or signaling alteration (Alexander and Friedl, 2012; DISCUSSION Orgaz et al., 2014a)––in contrast with frequently mutated MAPK (Davies et al., 2002; Cancer Genome Atlas Network, Recurrent transcriptional alterations occur during development 2015). Rho GTPases are overexpressed in cancer (Orgaz of resistance to MAPKi (Song et al., 2017). In this study we ﬁnd et al., 2014a); particularly RhoC is a driver of melanoma that adaptation to therapy occurs early on treatment through metastasis by increased expression (Clark et al., 2000). Lower cytoskeletal remodeling leading to restoration/increase of frequency of mutations suggests that cancer cells are less ad- myosin II levels in resistant melanomas. Because targeted and dicted to these pathways and, upon inhibition, development of immunotherapy-resistant cells rely on ROCK-dependent myosin resistance could be less frequent. Although we have shown II for survival, this could be a key mediator of cross-resistance. that myosin II inhibition also impairs survival of therapy-sensi- Resistant melanomas increase either MLC2 expression and/or tive melanoma cells, therapy-resistant cells are more sensitive activity, which in turn increases and reinforces myosin II activity to ROCKi. This is due to resistant cells having gained certain survival traits, but acquired vulnerabilities in return, such as (Calvo et al., 2013; Medjkane et al., 2009). Cells under drug treatment upregulate myosin II as a pro-survival response to defective anti-oxidant and DNA damage repair responses. MAPK inhibition, resulting in uncoupling of ERK signals to the Inhibition of myosin II activity overcomes resistance in mela- cytoskeleton. noma through induction of lethal ROS, unresolved DNA damage, Although myosin II activity is controlled by BRN2-mediated and loss of pro-survival signaling, which leads to cell-cycle arrest downregulation of PDE5A and increased calcium signaling in and cell death. A recent study has described that HDAC BRAF mutant melanoma (Arozarena et al., 2011), our mechanism inhibitors (HDACi) also induce lethal ROS and DNA damage in seems operative in NRAS mutant melanoma. PDE5A expression MAPKi-resistant melanomas (Wang et al., 2018). It will be increases in MAPKi-resistant lines compared with parental important to investigate if/how HDACi regulate cytoskeletal (Song et al., 2017) in a similar fashion as MLC2 (MYL9) (data remodeling. not shown). Because p-MLC2 levels are restored/increased in The tumor microenvironment has a key role in resistance to resistant versus parental lines, there may be mechanisms block- therapies in melanoma (Almeida et al., 2019) and macrophages ing the inhibitory action of PDE5A on myosin II in resistant cells. can contribute to resistance to MAPKi through secretion of Moreover, myosin II levels are ROCK dependent in resistant pro-survival factors (Smith et al., 2014). Furthermore, TGF-b inhi- cells, so PDE5A may not regulate myosin II activity in this bition enhanced efﬁcacy of immune checkpoint inhibitors (Ma- context. riathasan et al., 2018; Tauriello et al., 2018). In addition to the MAPKi-resistant cells have been associated to bundled cell intrinsic effects we observe, we report how inhibition of collagen and pro-survival signals (Brighton et al., 2018). ROCK-myosin II reduces pro-tumorigenic CD206 macro- Increased ECM deposition found in resistant tumors could phages, which could contribute to reducing tumor growth. More- contribute to myosin II activity in vivo (Laklai et al., 2016; Paszek over, ROCK-myosin II inhibition decreases FOXP3 Tregs. et al., 2005). Likewise, ROCK-myosin II-driven contractility also Combination of ROCKi with anti-PD-1 also reduces PD-L1 induces ECM stiffening (Samuel et al., 2011), generating a feed- expression on both tumor cells and CD206 macrophages. back loop between myosin II and ECM. These effects could be due to lower STAT3 activity after ROCK Widely studied in cell migration (Jaffe and Hall, 2005; Olson, inhibition (Sanz-Moreno et al., 2011), since PD-L1 expression 2008; Sadok et al., 2015; Sahai and Marshall, 2002; Sanz-Mor- can be regulated by STAT3 (Marzec et al., 2008; Pardoll, eno et al., 2008), ROCK-myosin II is proposed here as a thera- 2012). Effects on T cells are likely due to ROCK-myosin II regu- peutic target that goes beyond this pro-migratory function. We lation of TGF-b in cancer cells (Cantelli et al., 2015). Decreased show how this machinery controls intrinsic survival and TGF-b production by melanoma induced by ROCKi can extrinsic immunosuppression. Importantly, contractile cytoskel- contribute to improved anti-PD-1 responses. + + + (B–D) Images and quantiﬁcation of p-MLC2 (B), CD206 (C), and FOXP3 (D) cells in 5555 tumors at endpoint (pooled data from 2 experiments). Ratio CD206 /F4/ 80 shown. Scale bars, 100 mm (p-MLC2, CD206) and 50 mm (FOXP3). (E) Schematic of experiment. + + + + (F) Left, growth of 5555 allografts after treatment. Right, quantiﬁcation of p-MLC2, CD206 , F4/80 , and ratio CD206 /F4/80 in anti-PD-1/NR or Resp tumors. (G) Left, growth of 5555 anti-PD-1/NR allografts in new recipient mice after treatment (n = 7-8 mice/group). Right, survival plot. (H) Images and quantiﬁcation of p-MLC2 and PD-L1 on tumor cells in tumors from (G). Scale bars, 25 mm. (I) Left, images and quantiﬁcation of PD-L1 on CD206 cells in tumors from (G). Images show merged pseudo-colors for each staining. Scale bar, 50 mm. Right, quantiﬁcation of FOXP3 Tregs in tumors from (G). (A–D and G–I) ROCKi GSK269962A. Boxplots show median (center line); interquartile range (box); min-max (whiskers); and individual mice (circles). p values by ANOVA with Benjamini, Krieger, and Yekutieli correction (B–D and H–I), t test (F), chi-square test: percentage regressions in (A) and (G). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not signiﬁcant. See also Figure S8. Cancer Cell 37, 85–103, January 13, 2020 99 B Phospho-proteomics ROCKi Fasudil has been used safely in Japan since 1995 to B Phospho-Peptide Enrichment Analysis treat subarachnoid hemorrhage (SAH) after a head trauma and B Quantitative Real Time One-Step PCR to prevent vasospasm associated with SAH (Feng et al., 2016; B Gene Expression Studies and Analysis Olson, 2008). ROCKi is given as a vasodilator to lower blood B Gene Enrichment Analyzes pressure (Olson, 2008), and Fasudil and other ROCKi are being B Tumor Xenografts tested in clinical trials for glaucoma and other vascular dis- B Immunotherapy Experiments eases, such as pulmonary hypertension and atherosclerosis B Survival in the Lung Assay (Olson, 2008). Optimal ROCKi could be tested in broader B Immunohistochemistry range of disease, as a strategy to extend clinical response to B Imaging and Scoring different cancer therapies or even as a single therapy in the d QUANTIFICATION AND STATISTICAL ANALYSIS case of drug-addicted tumors. Importantly, therapy-resistant d DATA AND CODE AVAILABILITY cells are more sensitive to ROCKi while its combination with current therapies seems to elicit a superior response. Lower doses of ROCKi and/or different schedule treatments could SUPPLEMENTAL INFORMATION be used in combination with current therapies to prolong their Supplemental Information can be found online at https://doi.org/10.1016/j. efﬁcacy and delay resistance. Alternatively, different delivery ccell.2019.12.003. strategies of ROCKi (local, antibody-drug conjugate) could be considered. In summary, we provide extensive evidence that targeting ACKNOWLEDGMENTS cytoskeletal regulators driving high myosin II activity The work was supported by Cancer Research UK (CRUK) C33043/A12065 (to overcomes resistance to targeted and immunotherapies in V.S.-M., J.L.O., and E.C.-M.), C33043/A24478 (to V.S.-M., E.C.-M., M.G., and melanoma. The cytoskeletal adaptations that occur very J.L.O.); C30122/A11527, C30122/A15774, C33043/A12065 (to S.N.K.); C107/ early on treatment provide not only a survival advantage but A10433, C107/A104339 (to A.S.); The Harry J. Lloyd Charitable Trust (to J.L.O. also a vulnerability, which can be later exploited. High and V.S.-M.); Barts Charity (to V.S.-M., I.R.-H., and J.M.); Royal Society myosin II activity identiﬁes therapy cross-resistant RG110591 (to V.S.-M.); Fundacio´ n Ramo´ n Areces (to E.C.-M.); Marie Sklo- dowska-Curie Action (H2020-MSCA-IF-2014-EF-ST) (to I.R.-H.); MR/ patients, suggesting its potential as a biomarker. Our work L023091/1 (to S.N.K. and S.M.); CRUK/NIHR in England/DoH for Scotland, opens the possibility that cytoskeletal remodeling could be a Wales and Northern Ireland ECMC (C10355/A15587) (to S.N.K.); Francis Crick conserved pro-survival mechanism of generating therapy- Institute core funding from CRUK (FC001112), MRC (FC001112), and Well- resistant cancer clones under the selection of other therapy come Trust (FC001112) (to I.M. and A.P.); NIHR Biomedical Research Centre regimes. (BRC) at Guy’s and St Thomas’ NHS Foundation Trust and King’s College Lon- don, IS-BRC-1215-20006 (to S.N.K.). Fluorescence-activated cell sorting was performed at BRC funded by NIHR. We are indebted to Richard Marais and his STAR+METHODS team (CRUK Manchester Institute) for kind provision of cell lines and Paul Lor- igan (University of Manchester) for providing patient samples (study approved Detailed methods are provided in the online version of this paper by Manchester Cancer Research Center Biobank Access Committee applica- and include the following: tion 13_RIMA_01; the role of the MCRC Biobank is to distribute samples and cannot endorse studies performed or interpretation of results). We thank d KEY RESOURCES TABLE Fran Balkwill, Colin Pegrum, and the Biological Services Unit at Barts Cancer Institute for help with mouse work; Romina Girotti and Jeremy Carlton for help- d LEAD CONTACT AND MATERIALS AVAILABILITY ful discussions; Fernando Calvo for MLC2 plasmids and advice on ssGSEA; d EXPERIMENTAL MODEL AND SUBJECT DETAILS Amaya Viros, Garry Ashton, Sandra Kumper, € and Michela Perani for technical B Patient-Derived Samples advice; Erik Sahai and Tohru Takaki for MLC2 lentivectors; Øystein Fodstad for B Cell Lines and Patient-Derived Cell Lines LOX-IMVI cells; and Wellcome Trust Functional Genomics Cell Bank for B Animals MM485 cells. d METHOD DETAILS B Chemicals AUTHOR CONTRIBUTIONS B Antibodies Conceptualization, V.S.-M. and J.L.O.; Methodology, J.L.O., A.P.-R., R.L., B Analysis of Cell Morphology I.M., V.S.-M., and F.W.; Investigation, J.L.O., E.C.-M., A.S., O.M., I.R.-H., B Long-Term Survival A.P.-R., J.M., V.B., M.G., P.P., L.B., S.M., P.K., C.T., and F.W.; Validation, B Long Term Survival on Collagen I Matrices J.L.O., E.C.-M., I.R.-H., A.P.-R., and O.M.; Writing – Original Draft, V.S.-M. B MTT Assay and J.L.O.; Writing – Review & Editing, V.S.-M., J.L.O., E.C.-M., and I.M.; B Cell Cycle Analysis Funding Acquisition, V.S.-M., J.L.O., I.M., and S.N.K.; Resources, I.M., B AnnexinV/Propidium Iodide FACS S.N.K., and R.L.; Supervision, V.S.-M. B ROS Detection B Time Lapse Microscopy DECLARATION OF INTERESTS B RNAi The authors declare no competing interests. B MLC2 Rescue Experiments B MLC2 Stable Overexpression Received: January 9, 2018 B Immunoﬂuorescence and Confocal Imaging Revised: September 4, 2019 B Immunoblotting Accepted: December 6, 2019 B TGF-b1 ELISA Published: January 13, 2020 100 Cancer Cell 37, 85–103, January 13, 2020 REFERENCES CCL2 and CXCL1 expression in astrocytes through beta1 and beta5 integrins. Glia 58, 1510–1521. Alexander, S., and Friedl, P. (2012). Cancer invasion and resistance: intercon- Feng, Y., LoGrasso, P.V., Defert, O., and Li, R. (2016). Rho kinase (ROCK) in- nected processes of disease progression and therapy failure. Trends Mol. hibitors and their therapeutic potential. J. Med. Chem. 59, 2269–2300. Med. 18, 13–26. Fisher, D.T., Appenheimer, M.M., and Evans, S.S. (2014). The two faces of IL-6 Almeida, F.V., Douglass, S.M., Fane, M.E., and Weeraratna, A.T. (2019). Bad in the tumor microenvironment. Semin. Immunol. 26, 38–47. company: microenvironmentally mediated resistance to targeted therapy in Flaherty, K.T., Puzanov, I., Kim, K.B., Ribas, A., McArthur, G.A., Sosman, J.A., melanoma. Pigment Cell Melanoma Res. 32, 237–247. O’Dwyer, P.J., Lee, R.J., Grippo, J.F., Nolop, K., and Chapman, P.B. (2010). Arozarena, I., Sanchez-Laorden, B., Packer, L., Hidalgo-Carcedo, C., Inhibition of mutated, activated BRAF in metastatic melanoma. N. Engl. J. Hayward, R., Viros, A., Sahai, E., and Marais, R. (2011). Oncogenic BRAF in- Med. 363, 809–819. duces melanoma cell invasion by downregulating the cGMP-speciﬁc phos- Flaherty, K.T., Infante, J.R., Daud, A., Gonzalez, R., Kefford, R.F., Sosman, J., phodiesterase PDE5A. Cancer Cell 19, 45–57. Hamid, O., Schuchter, L., Cebon, J., Ibrahim, N., et al. (2012). Combined BRAF and MEK inhibition in melanoma with BRAF V600 mutations. N. Engl. J. Med. Baenke, F., Chaneton, B., Smith, M., Van Den Broek, N., Hogan, K., Tang, H., 367, 1694–1703. Viros, A., Martin, M., Galbraith, L., Girotti, M.R., et al. (2015). Resistance to BRAF inhibitors induces glutamine dependency in melanoma cells. Mol. Gadea, G., Sanz-Moreno, V., Self, A., Godi, A., and Marshall, C.J. (2008). Oncol. 10, 73–84. DOCK10-mediated Cdc42 activation is necessary for amoeboid invasion of melanoma cells. Curr. Biol. 18, 1456–1465. Balch, C.M., Gershenwald, J.E., Soong, S.J., Thompson, J.F., Atkins, M.B., Byrd, D.R., Buzaid, A.C., Cochran, A.J., Coit, D.G., Ding, S., et al. (2009). Georgouli, M., Herraiz, C., Crosas-Molist, E., Fanshawe, B., Maiques, O., Final version of 2009 AJCC melanoma staging and classiﬁcation. J. Clin. Perdrix, A., Pandya, P., Rodriguez-Hernandez, I., Ilieva, K.M., Cantelli, G., Oncol. 27, 6199–6206. et al. (2019). Regional activation of myosin II in cancer cells drives tumor pro- gression via a secretory cross-talk with the immune microenvironment. Cell Bankhead, P., Loughrey, M.B., Fernandez, J.A., Dombrowski, Y., McArt, D.G., 176, 757–774.e23. Dunne, P.D., McQuaid, S., Gray, R.T., Murray, L.J., Coleman, H.G., et al. (2017). QuPath: open source software for digital pathology image analysis. Girotti, M.R., Pedersen, M., Sanchez-Laorden, B., Viros, A., Turajlic, S., Niculescu-Duvaz, D., Zambon, A., Sinclair, J., Hayes, A., Gore, M., et al. Sci. Rep. 7, 16878. (2013). Inhibiting EGF receptor or SRC family kinase signaling overcomes Brighton, H.E., Angus, S.P., Bo, T., Roques, J., Tagliatela, A.C., Darr, D.B., BRAF inhibitor resistance in melanoma. Cancer Discov. 3, 158–167. Karagoz, K., Sciaky, N., Gatza, M.L., Sharpless, N.E., et al. (2018). New mech- Gray-Schopfer, V., Wellbrock, C., and Marais, R. (2007). Melanoma biology anisms of resistance to MEK inhibitors in melanoma revealed by intravital im- and new targeted therapy. Nature 445, 851–857. aging. Cancer Res. 78, 542–557. Hall, A. (2012). Rho family GTPases. Biochem. Soc. Trans. 40, 1378–1382. Calvo, F., Sanz-Moreno, V., Agudo-Ibanez, L., Wallberg, F., Sahai, E., Marshall, C.J., and Crespo, P. (2011). RasGRF suppresses Cdc42-mediated Haystead, T.A. (2005). ZIP kinase, a key regulator of myosin protein phospha- tase 1. Cell Signal. 17, 1313–1322. tumour cell movement, cytoskeletal dynamics and transformation. Nat. Cell Biol. 13, 819–826. Herraiz, C., Calvo, F., Pandya, P., Cantelli, G., Rodriguez-Hernandez, I., Orgaz, J.L., Kang, N., Chu, T., Sahai, E., and Sanz-Moreno, V. (2015). Reactivation of Calvo, F., Ege, N., Grande-Garcia, A., Hooper, S., Jenkins, R.P., Chaudhry, p53 by a cytoskeletal sensor to control the balance between DNA damage and S.I., Harrington, K., Williamson, P., Moeendarbary, E., Charras, G., and tumor dissemination. J. Natl. Cancer Inst. 108, https://doi.org/10.1093/jnci/ Sahai, E. (2013). Mechanotransduction and YAP-dependent matrix remodel- djv289. ling is required for the generation and maintenance of cancer-associated ﬁbro- blasts. Nat. Cell Biol. 15, 637–646. Hirata, E., Girotti, M.R., Viros, A., Hooper, S., Spencer-Dene, B., Matsuda, M., Larkin, J., Marais, R., and Sahai, E. (2015). Intravital imaging reveals how BRAF Cancer Genome Atlas Network. (2015). Genomic classiﬁcation of cutaneous inhibition generates drug-tolerant microenvironments with high integrin beta1/ melanoma. Cell 161, 1681–1696. FAK signaling. Cancer Cell 27, 574–588. Cantelli, G., Orgaz, J.L., Rodriguez-Hernandez, I., Karagiannis, P., Maiques, Hodi, F.S., O’Day, S.J., McDermott, D.F., Weber, R.W., Sosman, J.A., Haanen, O., Matias-Guiu, X., Nestle, F.O., Marti, R.M., Karagiannis, S.N., and Sanz- J.B., Gonzalez, R., Robert, C., Schadendorf, D., Hassel, J.C., et al. (2010). Moreno, V. (2015). TGF-b-induced transcription sustains amoeboid melanoma Improved survival with ipilimumab in patients with metastatic melanoma. migration and dissemination. Curr. Biol. 25, 2899–2914. N. Engl. J. Med. 363, 711–723. Cantelli, G., Crosas-Molist, E., Georgouli, M., and Sanz-Moreno, V. (2017). Hong, A., Moriceau, G., Sun, L., Lomeli, S., Piva, M., Damoiseaux, R., Holmen, TGFBeta-induced transcription in cancer. Semin. Cancer Biol. 42, 60–69. S.L., Sharpless, N.E., Hugo, W., and Lo, R.S. (2017). Exploiting drug addiction Chapman, P.B., Hauschild, A., Robert, C., Haanen, J.B., Ascierto, P., Larkin, mechanisms to select against MAPKi-resistant melanoma. Cancer Discov. J., Dummer, R., Garbe, C., Testori, A., Maio, M., et al. (2011). Improved survival 8, 74–93. with vemurafenib in melanoma with BRAF V600E mutation. N. Engl. J. Med. Hugo, W., Shi, H., Sun, L., Piva, M., Song, C., Kong, X., Moriceau, G., Hong, A., 364, 2507–2516. Dahlman, K.B., Johnson, D.B., et al. (2015). Non-genomic and immune evolu- Clark, E.A., Golub, T.R., Lander, E.S., and Hynes, R.O. (2000). Genomic anal- tion of melanoma acquiring MAPKi resistance. Cell 162, 1271–1285. ysis of metastasis reveals an essential role for RhoC. Nature 406, 532–535. Hugo, W., Zaretsky, J.M., Sun, L., Song, C., Moreno, B.H., Hu-Lieskovan, S., Condamine, T., Ramachandran, I., Youn, J.I., and Gabrilovich, D.I. (2015). Berent-Maoz, B., Pang, J., Chmielowski, B., Cherry, G., et al. (2016). Genomic Regulation of tumor metastasis by myeloid-derived suppressor cells. Annu. and transcriptomic features of response to anti-PD-1 therapy in metastatic Rev. Med. 66, 97–110. melanoma. Cell 165, 35–44. Davies, H., Bignell, G.R., Cox, C., Stephens, P., Edkins, S., Clegg, S., Teague, Ito, M., Nakano, T., Erdodi, F., and Hartshorne, D.J. (2004). Myosin phospha- J., Woffendin, H., Garnett, M.J., Bottomley, W., et al. (2002). Mutations of the tase: structure, regulation and function. Mol. Cell. Biochem. 259, 197–209. BRAF gene in human cancer. Nature 417, 949–954. Itoh, K., Yoshioka, K., Akedo, H., Uehata, M., Ishizaki, T., and Narumiya, S. (1999). An essential part for Rho-associated kinase in the transcellular invasion Dhomen, N., Reis-Filho, J.S., da Rocha Dias, S., Hayward, R., Savage, K., of tumor cells. Nat. Med. 5, 221–225. Delmas, V., Larue, L., Pritchard, C., and Marais, R. (2009). Oncogenic Braf in- duces melanocyte senescence and melanoma in mice. Cancer Cell 15, Jaffe, A.B., and Hall, A. (2005). Rho GTPases: biochemistry and biology. Annu. 294–303. Rev. Cell Dev. Biol. 21, 247–269. Le Dreau, G., Kular, L., Nicot, A.B., Calmel, C., Melik-Parsadaniantz, S., Kakavand, H., Rawson, R.V., Pupo, G.M., Yang, J.Y.H., Menzies, A.M., Kitabgi, P., Laurent, M., and Martinerie, C. (2010). NOV/CCN3 upregulates Carlino, M.S., Kefford, R.F., Howle, J.R., Saw, R.P.M., Thompson, J.F., et al. Cancer Cell 37, 85–103, January 13, 2020 101 (2017). PD-L1 expression and immune escape in melanoma resistance to Marzec, M., Zhang, Q., Goradia, A., Raghunath, P.N., Liu, X., Paessler, M., MAPK inhibitors. Clin. Cancer Res. 23, 6054–6061. Wang, H.Y., Wysocka, M., Cheng, M., Ruggeri, B.A., and Wasik, M.A. (2008). Oncogenic kinase NPM/ALK induces through STAT3 expression of Kong, X., Kuilman, T., Shahrabi, A., Boshuizen, J., Kemper, K., Song, J.Y., immunosuppressive protein CD274 (PD-L1, B7-H1). Proc. Natl. Acad. Sci. U Niessen, H.W.M., Rozeman, E.A., Geukes Foppen, M.H., Blank, C.U., et al. SA 105, 20852–20857. (2017). Cancer drug addiction is relayed by an ERK2-dependent phenotype switch. Nature 550, 270–274. Medjkane, S., Perez-Sanchez, C., Gaggioli, C., Sahai, E., and Treisman, R. (2009). Myocardin-related transcription factors and SRF are required for cyto- Konieczkowski, D.J., Johannessen, C.M., Abudayyeh, O., Kim, J.W., Cooper, skeletal dynamics and experimental metastasis. Nat. Cell Biol. 11, 257–268. Z.A., Piris, A., Frederick, D.T., Barzily-Rokni, M., Straussman, R., Haq, R., et al. Mjelle, R., Hegre, S.A., Aas, P.A., Slupphaug, G., Drablos, F., Saetrom, P., and (2014). A melanoma cell state distinction inﬂuences sensitivity to MAPK pathway inhibitors. Cancer Discov. 4, 816–827. Krokan, H.E. (2015). Cell cycle regulation of human DNA repair and chromatin remodeling genes. DNA Repair (Amst.) 30, 53–67. Konieczkowski, D.J., Johannessen, C.M., and Garraway, L.A. (2018). A convergence-based framework for cancer drug resistance. Cancer Cell 33, Moriceau, G., Hugo, W., Hong, A., Shi, H., Kong, X., Yu, C.C., Koya, R.C., Samatar, A.A., Khanlou, N., Braun, J., et al. (2015). Tunable-combinatorial 801–815. mechanisms of acquired resistance limit the efﬁcacy of BRAF/MEK cotarget- Krokan, H.E., and Bjoras, M. (2013). Base excision repair. Cold Spring Harb. ing but result in melanoma drug addiction. Cancer Cell 27, 240–256. Perspect. Biol. 5, a012583. Nakamura, K., Kitani, A., and Strober, W. (2001). Cell contact-dependent Kumper, S., Mardakheh, F.K., McCarthy, A., Yeo, M., Stamp, G.W., Paul, A., immunosuppression by CD4(+)CD25(+) regulatory T cells is mediated by cell Worboys, J., Sadok, A., Jorgensen, C., Guichard, S., and Marshall, C.J. surface-bound transforming growth factor beta. J. Exp. Med. 194, 629–644. (2016). Rho-associated kinase (ROCK) function is essential for cell cycle pro- Obenauf, A.C., Zou, Y., Ji, A.L., Vanharanta, S., Shu, W., Shi, H., Kong, X., gression, senescence and tumorigenesis. Elife 5, e12994. Bosenberg, M.C., Wiesner, T., Rosen, N., et al. (2015). Therapy-induced Kwong, L.N., Boland, G.M., Frederick, D.T., Helms, T.L., Akid, A.T., Miller, J.P., tumour secretomes promote resistance and tumour progression. Nature Jiang, S., Cooper, Z.A., Song, X., Seth, S., et al. (2015). Co-clinical assessment 520, 368–372. identiﬁes patterns of BRAF inhibitor resistance in melanoma. J. Clin. Invest. Olson, M.F. (2008). Applications for ROCK kinase inhibition. Curr. Opin. Cell 125, 1459–1470. Biol. 20, 242–248. Laklai, H., Miroshnikova, Y.A., Pickup, M.W., Collisson, E.A., Kim, G.E., Orgaz, J.L., Herraiz, C., and Sanz-Moreno, V. (2014a). Rho GTPases modulate Barrett, A.S., Hill, R.C., Lakins, J.N., Schlaepfer, D.D., Mouw, J.K., et al. malignant transformation of tumor cells. Small GTPases 5, e29019. (2016). Genotype tunes pancreatic ductal adenocarcinoma tissue tension to induce matricellular ﬁbrosis and tumor progression. Nat. Med. 22, 497–505. Orgaz, J.L., Pandya, P., Dalmeida, R., Karagiannis, P., Sanchez-Laorden, B., Viros, A., Albrengues, J., Nestle, F.O., Ridley, A.J., Gaggioli, C., et al. Lammermann, T., and Sixt, M. (2009). Mechanical modes of ’amoeboid’ cell (2014b). Diverse matrix metalloproteinase functions regulate cancer amoeboid migration. Curr. Opin. Cell Biol. 21, 636–644. migration. Nat. Commun. 5, 4255. Larkin, J., Ascierto, P.A., Dreno, B., Atkinson, V., Liszkay, G., Maio, M., Pardoll, D.M. (2012). The blockade of immune checkpoints in cancer immuno- Mandala, M., Demidov, L., Stroyakovskiy, D., Thomas, L., et al. (2014). therapy. Nat. Rev. Cancer 12, 252–264. Combined vemurafenib and cobimetinib in BRAF-mutated melanoma. N. Engl. J. Med. 371, 1867–1876. Paszek, M.J., Zahir, N., Johnson, K.R., Lakins, J.N., Rozenberg, G.I., Gefen, A., Reinhart-King, C.A., Margulies, S.S., Dembo, M., Boettiger, D., et al. Larkin, J., Chiarion-Sileni, V., Gonzalez, R., Grob, J.J., Cowey, C.L., Lao, C.D., (2005). Tensional homeostasis and the malignant phenotype. Cancer Cell 8, Schadendorf, D., Dummer, R., Smylie, M., Rutkowski, P., et al. (2015). 241–254. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma. N. Engl. J. Med. 373, 23–34. Peng, G., Chun-Jen Lin, C., Mo, W., Dai, H., Park, Y.Y., Kim, S.M., Peng, Y., Mo, Q., Siwko, S., Hu, R., et al. (2014). Genome-wide transcriptome proﬁling Lau, P.K., Ascierto, P.A., and McArthur, G. (2016). Melanoma: the intersection of homologous recombination DNA repair. Nat. Commun. 5, 3361. of molecular targeted therapy and immune checkpoint inhibition. Curr. Opin. Immunol. 39, 30–38. Poli, V., and Camporeale, A. (2015). STAT3-mediated metabolic reprograming in cellular transformation and implications for drug resistance. Front. Oncol. Lito, P., Pratilas, C.A., Joseph, E.W., Tadi, M., Halilovic, E., Zubrowski, M., 5, 121. Huang, A., Wong, W.L., Callahan, M.K., Merghoub, T., et al. (2012). Relief of profound feedback inhibition of mitogenic signaling by RAF inhibitors attenu- Posern, G., and Treisman, R. (2006). Actin’ together: serum response factor, its ates their activity in BRAFV600E melanomas. Cancer Cell 22, 668–682. cofactors and the link to signal transduction. Trends Cell Biol. 16, 588–596. Long, G.V., Fung, C., Menzies, A.M., Pupo, G.M., Carlino, M.S., Hyman, J., Qian, B.Z., Li, J., Zhang, H., Kitamura, T., Zhang, J., Campion, L.R., Kaiser, Shahheydari, H., Tembe, V., Thompson, J.F., Saw, R.P., et al. (2014a). E.A., Snyder, L.A., and Pollard, J.W. (2011). CCL2 recruits inﬂammatory mono- Increased MAPK reactivation in early resistance to dabrafenib/trametinib com- cytes to facilitate breast-tumour metastasis. Nature 475, 222–225. bination therapy of BRAF-mutant metastatic melanoma. Nat. Commun. Riaz, N., Havel, J.J., Makarov, V., Desrichard, A., Urba, W.J., Sims, J.S., Hodi, 5, 5694. F.S., Martin-Algarra, S., Mandal, R., Sharfman, W.H., et al. (2017). Tumor and Long, G.V., Stroyakovskiy, D., Gogas, H., Levchenko, E., de Braud, F., Larkin, microenvironment evolution during immunotherapy with nivolumab. Cell 171, J., Garbe, C., Jouary, T., Hauschild, A., Grob, J.J., et al. (2014b). Combined 934–949.e16. BRAF and MEK inhibition versus BRAF inhibition alone in melanoma. Rizos, H., Menzies, A.M., Pupo, G.M., Carlino, M.S., Fung, C., Hyman, J., N. Engl. J. Med. 371, 1877–1888. Haydu, L.E., Mijatov, B., Becker, T.M., Boyd, S.C., et al. (2014). BRAF inhibitor Lu, H., Liu, S., Zhang, G., Bin, W., Zhu, Y., Frederick, D.T., Hu, Y., Zhong, W., resistance mechanisms in metastatic melanoma: spectrum and clinical Randell, S., Sadek, N., et al. (2017). PAK signalling drives acquired drug resis- impact. Clin. Cancer Res. 20, 1965–1977. tance to MAPK inhibitors in BRAF-mutant melanomas. Nature 550, 133–136. Robert, C., Karaszewska, B., Schachter, J., Rutkowski, P., Mackiewicz, A., Mantovani, A., Sica, A., Sozzani, S., Allavena, P., Vecchi, A., and Locati, M. Stroiakovski, D., Lichinitser, M., Dummer, R., Grange, F., Mortier, L., et al. (2004). The chemokine system in diverse forms of macrophage activation (2015). Improved overall survival in melanoma with combined dabrafenib and polarization. Trends Immunol. 25, 677–686. and trametinib. N. Engl. J. Med. 372, 30–39. Mariathasan, S., Turley, S.J., Nickles, D., Castiglioni, A., Yuen, K., Wang, Y., Roca, H., Varsos, Z.S., Sud, S., Craig, M.J., Ying, C., and Pienta, K.J. (2009). Kadel, E.E., III, Koeppen, H., Astarita, J.L., Cubas, R., et al. (2018). TGFbeta CCL2 and interleukin-6 promote survival of human CD11b+ peripheral blood attenuates tumour response to PD-L1 blockade by contributing to exclusion mononuclear cells and induce M2-type macrophage polarization. J. Biol. of T cells. Nature 554, 544–548. Chem. 284, 34342–34354. 102 Cancer Cell 37, 85–103, January 13, 2020 Sadok, A., McCarthy, A., Caldwell, J., Collins, I., Garrett, M.D., Yeo, M., (2017). Actomyosin drives cancer cell nuclear dysmorphia and threatens Hooper, S., Sahai, E., Kuemper, S., Mardakheh, F.K., and Marshall, C.J. genome stability. Nat. Commun. 8, 16013. (2015). Rho kinase inhibitors block melanoma cell migration and inhibit metas- Tauriello, D.V.F., Palomo-Ponce, S., Stork, D., Berenguer-Llergo, A., Badia- tasis. Cancer Res. 75, 2272–2284. Ramentol, J., Iglesias, M., Sevillano, M., Ibiza, S., Canellas, A., Hernando- Sahai, E., and Marshall, C.J. (2002). RHO-GTPases and cancer. Nat. Rev. Momblona, X., et al. (2018). TGFbeta drives immune evasion in genetically Cancer 2, 133–142. reconstituted colon cancer metastasis. Nature 554, 538–543. Samuel, M.S., Lopez, J.I., McGhee, E.J., Croft, D.R., Strachan, D., Timpson, Das Thakur, M., Salangsang, F., Landman, A.S., Sellers, W.R., Pryer, N.K., P., Munro, J., Schroder, E., Zhou, J., Brunton, V.G., et al. (2011). Levesque, M.P., Dummer, R., McMahon, M., and Stuart, D.D. (2013). Actomyosin-mediated cellular tension drives increased tissue stiffness and Modelling vemurafenib resistance in melanoma reveals a strategy to forestall beta-catenin activation to induce epidermal hyperplasia and tumor growth. drug resistance. Nature 494, 251–255. Cancer Cell 19, 776–791. Titz, B., Lomova, A., Le, A., Hugo, W., Kong, X., Ten Hoeve, J., Friedman, M., Sanz-Moreno, V., Gadea, G., Ahn, J., Paterson, H., Marra, P., Pinner, S., Sahai, Shi, H., Moriceau, G., Song, C., et al. (2016). JUN dependency in distinct early E., and Marshall, C.J. (2008). Rac activation and inactivation control plasticity and late BRAF inhibition adaptation states of melanoma. Cell Discov. 2, 16028. of tumor cell movement. Cell 135, 510–523. Di Veroli, G.Y., Fornari, C., Wang, D., Mollard, S., Bramhall, J.L., Richards, Sanz-Moreno, V., Gaggioli, C., Yeo, M., Albrengues, J., Wallberg, F., Viros, A., F.M., and Jodrell, D.I. (2016). Combeneﬁt: an interactive platform for the anal- Hooper, S., Mitter, R., Feral, C.C., Cook, M., et al. (2011). ROCK and JAK1 ysis and visualization of drug combinations. Bioinformatics 32, 2866–2868. signaling cooperate to control actomyosin contractility in tumor cells and Vicente-Manzanares, M., Ma, X., Adelstein, R.S., and Horwitz, A.R. (2009). stroma. Cancer Cell 20, 229–245. Non-muscle myosin II takes centre stage in cell adhesion and migration. Shaltiel, I.A., Krenning, L., Bruinsma, W., and Medema, R.H. (2015). The same, Nat. Rev. Mol. Cell Biol. 10, 778–790. only different––DNA damage checkpoints and their reversal throughout the Wagle, N., Emery, C., Berger, M.F., Davis, M.J., Sawyer, A., Pochanard, P., cell cycle. J. Cell Sci. 128, 607–620. Kehoe, S.M., Johannessen, C.M., Macconaill, L.E., Hahn, W.C., et al. (2011). Sharma, P., Hu-Lieskovan, S., Wargo, J.A., and Ribas, A. (2017). Primary, Dissecting therapeutic resistance to RAF inhibition in melanoma by tumor adaptive, and acquired resistance to cancer immunotherapy. Cell 168, genomic proﬁling. J. Clin. Oncol. 29, 3085–3096. 707–723. Wagle, N., Van Allen, E.M., Treacy, D.J., Frederick, D.T., Cooper, Z.A., Taylor- Smith, M.P., Sanchez-Laorden, B., O’Brien, K., Brunton, H., Ferguson, J., Weiner, A., Rosenberg, M., Goetz, E.M., Sullivan, R.J., Farlow, D.N., et al. Young, H., Dhomen, N., Flaherty, K.T., Frederick, D.T., Cooper, Z.A., et al. (2014). MAP kinase pathway alterations in BRAF-mutant melanoma patients (2014). The immune microenvironment confers resistance to MAPK pathway with acquired resistance to combined RAF/MEK inhibition. Cancer Discov. inhibitors through macrophage-derived TNFalpha. Cancer Discov. 4, 4, 61–68. 1214–1229. Wang, L., Leite de Oliveira, R., Huijberts, S., Bosdriesz, E., Pencheva, N., Song, C., Piva, M., Sun, L., Hong, A., Moriceau, G., Kong, X., Zhang, H., Brunen, D., Bosma, A., Song, J.Y., Zevenhoven, J., Los-de Vries, G.T., et al. Lomeli, S., Qian, J., Yu, C.C., et al. (2017). Recurrent tumor cell-intrinsic and (2018). An acquired vulnerability of drug-resistant melanoma with therapeutic -extrinsic alterations during MAPKi-induced melanoma regression and early potential. Cell 173, 1413–1425.e14. adaptation. Cancer Discov. 7, 1248–1265. Wolf, K., Muller, R., Borgmann, S., Brocker, E.B., and Friedl, P. (2003). Sun, C., Wang, L., Huang, S., Heynen, G.J., Prahallad, A., Robert, C., Haanen, Amoeboid shape change and contact guidance: T-lymphocyte crawling J., Blank, C., Wesseling, J., Willems, S.M., et al. (2014). Reversible and adap- through ﬁbrillar collagen is independent of matrix remodeling by MMPs and tive resistance to BRAF(V600E) inhibition in melanoma. Nature 508, 118–122. other proteases. Blood 102, 3262–3269. Takaki, T., Montagner, M., Serres, M.P., Le Berre, M., Russell, M., Collinson, Zhang, W. (2015). BRAF inhibitors: the current and the future. Curr. Opin. L., Szuhai, K., Howell, M., Boulton, S.J., Sahai, E., and Petronczki, M. Pharmacol. 23, 68–73. Cancer Cell 37, 85–103, January 13, 2020 103 STAR+METHODS KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies pT202/Y204-p44/42 (ERK1/2) Cell Signaling Technology Cat# 4370; RRID:AB_2315112 pThr18/Ser19-MLC2 Cell Signaling Technology Cat# 3674; RRID: AB_2147464 pSer19-MLC2 Cell Signaling Technology Cat# 3671; RRID: AB_330248 MLC2 Cell Signaling Technology Cat# 3672; RRID: AB_10692513 PD-L1 clone E1L3N Cell Signaling Technology Cat# 13684; RRID:AB_2687655 pY705-STAT3 Cell Signaling Technology Cat# 9145; RRID:AB_2491009 ERK2 Santa Cruz Biotechnology Cat# sc-154; RRID:AB_2141292 GAPDH Santa Cruz Biotechnology Cat# MAB374; RRID:AB_2107445 GFP Santa Cruz Biotechnology Cat# sc-8334; RRID:AB_641123 MCL-1 Santa Cruz Biotechnology Cat# sc-819; RRID:AB_2144105 STAT3 Santa Cruz Biotechnology Cat# sc-482; RRID:AB_632440 Rat IgG2a anti-PD-1 clone RMP1-14 BioXCell Cat# BE0146; RRID:AB_10949053 Rat IgG2a isotype control clone 2A3 BioXCell Cat# BE0089; RRID:AB_1107769 CD206 Abcam Cat# ab64693; RRID:AB_1523910 CD3 anti-mouse Abcam Cat# ab134096 CD4 anti-mouse, clone I3T4 Abcam Cat# ab183685; RRID:AB_2686917 FoxP3 anti-human, clone 236A/E7 Abcam Cat# ab20034; RRID:AB_445284 P-H2A.X (S139) Abcam Cat# ab2893; RRID:AB_303388 CD8a anti-mouse, clone Ly2 Invitrogen Cat# 14-0808-82; RRID:AB_2572861 F4/80 anti-mouse, clone BM8 Invitrogen Cat# MF48000; RRID:AB_10376289 FoxP3 anti-mouse, clone FJK-16s Invitrogen Cat# 14-5773-82; RRID:AB_467576 CD4 anti-human, clone 11E9 Novocastra Cat# NCL-L-CD4-368; RRID:AB_563559 Biological Samples Human melanoma pre-/post-therapy Paul Lorigan, Richard Marais N/A Chemicals, Peptides, and Recombinant Proteins PLX4720 Selleck #S1152 PLX4032 Selleck #S1267 GSK2118436 Dabrafenib ChemieTek #CT-DABRF GSK1120212 Trametinib Selleck #S2673 PD184352 Selleck #S1020 AZD6244 Selleck #S1008 SCH772984 Selleck #S7101 GSK269962A Axon MedChem # Axon 1167 H1152 Calbiochem #555550 AT13148 Selleck #S7563 (±)-Blebbistatin Calbiochem #203390 Critical Commercial Assays Human/Mouse TGF-b1 ELISA Biolegend #436707 Trichrome Stain (Masson) Kit Sigma #HT15-1KT Bouin’s solution Sigma #HT10132 Weigert’s iron hematoxylin solution Sigma # HT1079 (Continued on next page) e1 Cancer Cell 37, 85–103.e1–e9, January 13, 2020 Continued REAGENT or RESOURCE SOURCE IDENTIFIER Deposited Data Mass spectrometry A375 MEKi 24h This study ProteomeXchange via PRIDE repository Project # PXD002621 (https://www.ebi.ac. uk/pride/archive/projects/PXD002621) Experimental Models: Cell Lines Human: A375 ATCC ATCC Cat# CRL-7904; RRID:CVCL_0132 Human: Colo829 ATCC ATCC Cat# CRL-1974; RRID:CVCL_1137 Human: SKMEL5 ATCC ATCC Cat# HTB-70; RRID:CVCL_0527 Human: WM88 Coriell Institute Coriell Cat# WC00123; RRID:CVCL_6805 Human: WM983A Coriell Institute Coriell Cat# WC00048; RRID:CVCL_6808 Human: WM983B Coriell Institute Coriell Cat# WC00066; RRID:CVCL_6809 Human: WM793B Coriell Institute Coriell Cat# WC00062; RRID:CVCL_8787 Human: A375M2 Richard Hynes Clark et al., 2000 Human: LOX-IMVI Øystein Fodstad RRID:CVCL_1381 Human: D04 Kevin Harrington RRID:CVCL_H604 Human: MM485 Wellcome Trust Functional Genomics RRID:CVCL_2610 Cell Bank Human: A375/PLX/R Richard Marais RRID:CVCL_IW10 Baenke et al., 2015 Human: Colo829/PLX/R Richard Marais RRID:CVCL_IW11 Baenke et al., 2015 Human: A375/D+T/R Richard Marais N/A Human: Patient #2 Richard Marais N/A Human: Patient #35 Richard Marais N/A Human: Patient #62T3 Richard Marais N/A Human: Patient #58 Richard Marais N/A Human: Patient #33 Richard Marais N/A Mouse: 5555 Richard Marais Dhomen et al., 2009; Hirata et al., 2015 Mouse: 5555-anti-PD-1/NR This paper N/A Mouse: 4434 Richard Marais Dhomen et al., 2009; Hirata et al., 2015 Mouse: 4599 Richard Marais Dhomen et al., 2009; Hirata et al., 2015 Mouse: 690cl2 Richard Marais Dhomen et al., 2009 Human: HEK293T Jeremy Carlton ATCC Cat# CRL-3216; RRID:CVCL_0063 Experimental Models: Organisms/Strains Mouse: CD-1 nu/nu Charles River UK RRID:IMSR_CRL:086 Mouse: NOD/SCID/IL-2Rg-/- (NSG) Charles River UK RRID:IMSR_JAX:005557 Mouse: C57BL/6J Charles River UK RRID:IMSR_JAX:000664 Oligonucleotides See Table S7 for RNAi sequences Dharmacon Recombinant DNA pEGFP-N3-EGFP Fernando Calvo Takara, Clontech #U57609 pEGFP-N3-MLC2-EGFP (rat MLC2) Fernando Calvo Calvo et al., 2013 pEGFP-N3-MLC2-TASA (T18A Fernando Calvo Calvo et al., 2013 S19A) -EGFP pEGFP-N3-MLC2-TDSD (T18D Fernando Calvo Calvo et al., 2013 S19D) -EGFP pLVX-EGFP Erik Sahai, Tohru Takaki Takara, Clontech #632164 pLVX-MLC2-EGFP Erik Sahai, Tohru Takaki Takaki et al., 2017 (Continued on next page) Cancer Cell 37, 85–103.e1–e9, January 13, 2020 e2 Continued REAGENT or RESOURCE SOURCE IDENTIFIER pLVX-MLC2-TASA (T18A S19A)-EGFP Erik Sahai, Tohru Takaki Takaki et al., 2017 pLVX-MLC2-TDSD (T18A S19A)-EGFP Erik Sahai, Tohru Takaki Takaki et al., 2017 Software and Algorithms GSEA, ssGSEA Broad Institute http://www.broadinstitute. N/A org/gsea/index.jsp ImageJ https://imagej.nih.gov/ij/ N/A GraphPad Prism 8 GraphPad Software N/A SPSS IBM N/A LEAD CONTACT AND MATERIALS AVAILABILITY Further information and reasonable requests for resources and reagents should be directed to and will be fulﬁlled by the Lead Con- tact, Victoria Sanz-Moreno (v.sanz-moreno@qmul.ac.uk). All unique/stable reagents generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement. EXPERIMENTAL MODEL AND SUBJECT DETAILS Patient-Derived Samples Human melanoma samples were a kind gift from Paul Lorigan (University of Manchester). Tumor samples were collected under the Manchester Cancer Research Centre (MCRC) Biobank ethics application #07/H1003/161+5 with full informed consent from the pa- tients. The work presented in this manuscript was approved by MCRC Biobank Access Committee application 13_RIMA_01. Patient sample information is in Table S6. Cell Lines and Patient-Derived Cell Lines Cell lines used are listed in the Key Resources Table. Cell lines were cultured under standard conditions in complete medium (DMEM or RPMI medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (all from Gibco)). Cell lines were V600E tested to be free from mycoplasma contamination. All melanoma cell lines used were BRAF unless otherwise stated. A375, Colo829 and SKMEL5 cells were from ATCC. WM88, WM983A, WM983B, WM793B were purchased from Coriell Institute. A375M2 were from Dr Richard Hynes (HHMI, MIT, USA). LOX-IMVI cell line was a gift from Prof Øystein Fodstad (Oslo University Hospital). SKMEL5, WM983A, WM983B, WM793B, LOX-IMVI were grown in complete RPMI, WM88 was grown in complete DMEM. PLX4720-resistant WM983A, WM983B and WM88 cells were derived after exposure to PLX4720 for 2-3 months (1 mM PLX4720 for WM983A and WM983B; 0.5 mM PLX4720 for WM88), controls were treated with equivalent volume of DMSO. PLX4720-resistant (A375/PLX/R, Colo829/PLX/R) (Baenke et al., 2015) and dabrafenib+trametinib-resistant (A375/D+T/R) cell lines were a kind gift from Richard Marais (Cancer Research UK Manchester Institute). Resistant cells were generated after exposure of parental A375 and Colo829 to increasing concentrations of drugs (up to 1 mM PLX4720; 1 mM dabrafenib plus 10 nM trametinib) until cells resumed growth. Cells were grown in complete DMEM (A375-derivatives) or complete RPMI (Colo829-derivatives) supple- mented with 1 mM PLX4720 (A375/PLX/R and Colo829/PLX/R cells); 1 mM dabrafenib plus 10 nM trametinib (A375/D+T/R) or equiv- alent volume of DMSO (parental A375 and Colo829 cells). Patient-derived melanoma cell lines (#2, #35, #62T3, #58, #33) were a very kind gift from Richard Marais and were grown in RPMI. Patient #2, #35, #62T3 were grown in complete RPMI supplemented with 1 mM PLX4720. Patient #2 cell line was estab- lished from a patient with stage IV BRAF mutant melanoma with primary resistance to vemurafenib and ipilimumab. Patient #/35 cell line was established from a lymph node metastasis after treatment with vemurafenib for 3 months. Patient #62T3 cell line was established from a resected tumor upon disease progression following vemurafenib treatment (acquired resistance) and immunotherapy (refractory to ipilimumab and subsequent pembrolizumab). Patient #58 cell line (wild-type for BRAF/NRAS) was established from a metastasis from a patient that never responded to ipilimumab treatment (3 months). Patient #33 cell K601E line (BRAF ) was established from a metastasis from a patient that never responded to ipilimumab treatment (1 month). Patient #58 and #33 had also been treated with dacarbazine (DTIC) before ipilimumab. Patient #26 cell lines were established before and after nivolumab treatment. V600E Braf mouse melanoma cell lines 5555, 4434, 4599 and 690cl2 (from Richard Marais) were established from the following +/LSL-V600E +/o INK4a-/- +/LSL-V600E +/o C57BL/6 mouse models: Braf ;Tyr::CreERT2 ;p16 (5555, 4434); Braf ;Tyr::CreERT2 (4599); Pten-null +/LSL-V600E +/o INK4a-/- Braf ;Tyr::CreERT2 ;p16 ;Pten-/- (690cl2) (Dhomen et al., 2009; Hirata et al., 2015). NRAS mutant cell lines used: D04 was from Kevin Harrington (The Institute of Cancer Research); MM485 was obtained from the Wellcome Trust Functional Ge- nomics Cell Bank (UK). HEK293T cells were from Jeremy Carlton (The Francis Crick Institute). e3 Cancer Cell 37, 85–103.e1–e9, January 13, 2020 A375, A375/PLX/R, Colo829, Colo829/PLX/R, SKMEL5 cells and Patient-derived cell lines were conﬁrmed by STR proﬁling at CRUK Manchester Institute; A375M2, WM983A, WM983B at King’s College London; WM88 and WM793B cells were purchased from Coriell Institute in June 2014. Animals All animals were maintained under speciﬁc pathogen-free conditions and handled in accordance with the Institutional Committees on Animal Welfare of the UK Home Ofﬁce (The Home Ofﬁce Animals Scientiﬁc Procedures Act, 1986). All animal experiments were approved by the Ethical Review Process Committees at Barts Cancer Institute, King’s College London and The Francis Crick Institute, in accordance with the Animals (Scientiﬁc Procedures) Act 1986 and according to the guidelines of the Committee of the National Cancer Research Institute. Animals used in this study were from Charles River UK: 5-week-old female nude CD-1 nu/nu mice; 5-8-week old NOD/SCID/ IL-2Rg-/- (NSG) mice (male and female); 5-7-week-old female C57BL/6J mice. Tumors were allowed to establish, sizes (average 60-100 mm ) were matched and then mice were randomly allocated to groups of 6-8 animals. No blinding was used in the treatment schedules for these experiments since the different treatments were identiﬁed by ear notching/mark on tail. Based on previous studies in the literature (Hong et al., 2017; Kong et al., 2017) and our own experience, groups of 6-8 animals were used to have sufﬁcient animals per group to provide statistically signiﬁcant data while keeping the number of animals used to a minimum. Tumor size was determined by caliper measurements of tumor length, width and depth and tumor volume was calcu- lated as volume = 0.5236 x length x width x depth (mm). Note that this formula calculates smaller tumors (approximately 2-fold smaller) compared to those calculated using the formula volume = 0.5236 x length x width (mm). METHOD DETAILS Chemicals Chemicals used in this study (stocks resuspended in DMSO unless otherwise stated): BRAFi PLX4720 and PLX4032 (Selleck), BRAFi Dabrafenib (GSK2118436, ChemieTek), MEKi Trametinib (GSK1120212, Selleck), MEKi PD184352 (Selleck), MEKi AZD6244 (Selleck), ERKi SCH772984 (Selleck), ROCKi GSK269962A (Axon Medchem), ROCKi H1152 (resuspended in water; Calbiochem), AGC kinase inhibitor and ROCKi AT13148 (Selleck), myosin II inhibitor blebbistatin (in 95% DMSO; Calbiochem). Concentrations used unless otherwise stated in other STAR Methods sections: 5 mM ROCKi GSK269962A, 5 mM ROCKi H1152, 5 mM ROCKi AT13148, 25 mM myosin II inhibitor blebbistatin, 5 mM BRAFi PLX4720. ‘‘Analysis of cell morphology’’ section lists the inhibitors and concentrations used for those experiments. Antibodies Antibodies and concentrations used: pThr18/Ser19-MLC2 (#3674; 1:750, immunoblot), pSer19-MLC2 (#3671; 1:50, immunohisto- chemistry; 1:200, immunoﬂuorescence), MLC2 (#3672; 1:750), pT202/Y204-p44/42 (ERK1/2) (#4370; 1:1,000), pY705-STAT3 (#9145; 1:750), PD-L1 (clone E1L3N, #13684, 1:200) from Cell Signaling Technology; STAT3 (sc-482; 1:500), ERK2 (sc-154; 1:1,000), MCL-1 (sc-819; 1:1,000), GFP (sc-8334; 1:1,000) from Santa Cruz Biotechnology; GAPDH (MAB374; 1:10,000) from Milli- pore; P-H2A.X (S139) (ab2893;1:1000), CD206 (ab64693; 1:1,000), CD3 (anti-mouse, ab134096; 1:500), CD4 (anti-mouse, clone I3T4, ab183685; 1:300), FoxP3 (anti-human, clone 236A/E7, ab20034; 1:200) from Abcam; F4/80 (anti-mouse, clone BM8, MF48000, 1:1000), CD8a (anti-mouse, clone Ly2, 14-0808-82; 1:200), FoxP3 (anti-mouse, clone FJK-16s, 14-5773-82; 1:200) from Invitrogen; CD4 (anti-human, clone 11E9, NCL-L-CD4-368; 1:300) from Novocastra. Analysis of Cell Morphology Cell morphology was analyzed on still phase-contrast images (cells on plastic or on collagen I) using ImageJ software (http://rsb.info. nih.gov/ij/). In order to quantify cell morphology on 2D and on collagen matrices, the morphology descriptor Circularity was used after manually drawing around the cell. Values closer to 1 represent rounded morphology; values closer to 0 represent more spread and/or spindle-shaped cells with multiple protrusions. Treatments were for 24 hr as follows: A375M2 cells with 50 nM BRAFi PLX4720, 0.1 nM MEKi GSK1120212, 1 mM ROCKi GSK269962A (Figure 1B); WM983A/B cells with 5 mM ROCKi GSK269962A, 5 mM BRAFi PLX4720 (Figures 1D and 1E); 690cl2 cells with 200 nM MEKi PD184352, 200 nM BRAFi PLX4032, 500 nM ERKi SCH772984 (Figures 1F and S1C); D04, MM485 cells with 50 nM MEKi GSK1120212, 50 nM AZD6244 (Figures 1F, S1D, and S1E); 4599 cells with 500 nM MEKi GSK1120212, 1 nM MEKi AZD6244 (Figure S1B). A375 and A375/PLX/R on plastic (Figure 3D); and Patient #2 cells on collagen I (Figure 5I) were treated with 5 mM ROCKi GSK269962A, 5 mM BRAFi PLX4720 or both. Long-Term Survival Long-term survival was performed on tissue culture plastic dishes unless otherwise speciﬁed. Cells were seeded in 6-well plates (10,000 cells/well) and treated for 5-14 days, re-adding drugs in fresh media every 2-3 days (daily for blebbistatin). Then cells were ﬁxed with 1% formaldehyde and stained with 0.25% crystal violet. Plates were scanned and images analyzed using ImageJ software. For experiments with inhibitors, percentage of the well covered by crystal violet-stained cells was calculated and shown relative to control cells. For dose-response experiments, cells were seeded in 12-well or 96-well plates and survival was analyzed Cancer Cell 37, 85–103.e1–e9, January 13, 2020 e4 after 3-5 days treatment with indicated drugs using crystal violet. Crystal violet was solubilized with 10% acetic acid and absorbance was measured at 590 nm. In dose-response experiments, BRAFi-resistant cells were cultured in the presence of BRAFi throughout the experiment unless otherwise stated. In Figure 5C, 4434- and 5555-derivatives were treated with 0.1 mM ROCKi. For synergy experiments, 1,000 A375 cells were seeded in 96-well plates, cultured overnight and next day treated in quintuplicates with ROCKi GSK269962A or BRAFi PLX4720, either alone or in several combinations in complete medium. Three days later, plates were ﬁxed, stained with crystal violet and solubilized and quantiﬁed as above. Values were normalized to vehicle controls and analyzed with Combeneﬁt software (Loewe model) (Di Veroli et al., 2016). Average of 4 independent experiments is shown. Long Term Survival on Collagen I Matrices Cells were grown on collagen I matrices as described (Orgaz et al., 2014b). Brieﬂy, bovine collagen I (PureCol, #5005-B; Advanced BioMatrix) thick gels were polymerized at 1.7 mg/ml in 24-well plates. Cells were seeded at 10,000 cell/well and treatments started 16 hr later for 5-14 days. In experiments using A375-derivatives, cells were treated with 1 mM ROCKi, 1 mM BRAFi or both. Patient- derived cell lines were treated with 5 mM ROCKi. Fresh complete media with drugs was added every 2-3 days. At the end of the exper- iment collagen I gels were ﬁxed with 4% formaldehyde and phase-contrast images were taken. Percentage of area covered by cells was quantiﬁed using QuPath software Version 0.1.2 and a SLIC superpixel image segmentation (Gaussian sigma value 5 pixels, superpixel spacing 20 pixels) (Bankhead et al., 2017). Software was trained to identify cells and background (surrounding collagen). Detection measurements were then exported to Excel and values for area/pixel were normalized to each untreated control as per- centage of area covered by cells. For Patient #2 cells, spheroid-forming ability was quantiﬁed as the sum of areas occupied by spher- oids from phase-contrast images using ImageJ. MTT Assay Cells were seeded in 96-well plates (2,000 cells/well). Drugs were added every 2 days. Three days after seeding, plates were incu- bated with MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; Millipore) following the manufacturer’s instructions and absorbance measured at 572 nm. Background at 630 nm was subtracted and data represented as relative viability. Cell Cycle Analysis For DNA cell cycle analysis, ﬂoating and adherent cells were ﬁxed in 70% ethanol at -20 C, washed in phosphate-buffered saline and treated with 40 mg/ml propidium iodide (PI) (Biolegend) and 100 mg/ml ribonuclease (Sigma) for 25 min at 37 C. Staining was detected using a FACS BD Canto II (BD Biosciences) and analyzed and plotted using FlowJo (FlowJo LLC). The starting gating of the whole cell population, excluding any debris, was performed with FSC-A/SSC-A. Using this as a parental gate, doublets were excluded using PerCP-Cy5.5-A/ PerCP-Cy5.5-W (PI). The gated singlets were represented as histograms for PerCP-Cy5.5-A to show the peaks for the cell cycle phases. AnnexinV/Propidium Iodide FACS Floating and adherent cells were collected, spun down, and labelled with FITC Annexin V Apoptosis Detection Kit with PI (#640914, Biolegend UK Ltd), following the manufacturer’s instructions. Staining was detected using a FACS BD Canto II and analyzed and plotted using FlowJo. The starting gating of the whole cell population, excluding any debris, was performed with FSC-A/SSC-A. This was followed by a double exclusion of doublets using ﬁrst FSC-H/FSC-W and then SSC-H/SSC-W. The gated singlets were then gated as ‘quad gates’ using FITC-A (AnnexinV) versus PerCP-Cy5.5-A (PI) and represented as FACS dot plots. Graphs show high low percentage of dead cells as the sum of percentage of early apoptotic (annexin V , propidium iodide ) and percentage of late high high apoptotic/necrotic cells (annexin V , propidium iodide ). ROS Detection Cells were treated with 1 mM ROCKi for 24 hr (A375 pair) or 48 hr (WM983A pair). Then cells were collected and ROS levels were detected using CellROX Green Flow Cytometry Assay Kit (C10492, Life Technologies), according to the manufacturer’s instructions. FACS and gating strategy were as described in Cell Cycle section. Time Lapse Microscopy Multi-site bright-ﬁeld microscopy of cells in 24-well plates was performed in a humidiﬁed chamber at 37 C and 5% CO using a 10X/0.3 NA Plan Fluor ELWD objective lens on a fully motorized (Prior Scientiﬁc) multi-ﬁeld Nikon TE2000 microscope with an ORCA camera (Hamamatsu) controlled by Micro-Manager (https://micro-manager.org/) and ImageJ. Sixteen hr after seeding, cells were treated for 72 hr with ROCKi, BRAFi or both in the presence of 1.5 mM PI to identify dead cells. Total number of cells, per- centage of multinucleated (alive, dead) and total dead cells were quantiﬁed for 72 hr. RNAi One hundred thousand cells were plated per 35-mm dish and transfected the next day with 26 nM siRNA oligonucleotides, using Optimem-I and Lipofectamine 2000 (Invitrogen). Forty eight hr after transfection cells were harvested and equal numbers re-seeded on 35-mm wells. Cells were transfected again 2 days later and plates were ﬁxed and stained with crystal violet 2-4 days after the second transfection. Crystal violet was solubilized and absorbance at 590 nm measured as above. Cells were grown in the presence e5 Cancer Cell 37, 85–103.e1–e9, January 13, 2020 of 1 mM PLX4720 during the whole experiment. All siRNA sequences were On-Targetplus (OT) from Dharmacon (Lafayette, USA) and are listed in Table S7. MLC2 Rescue Experiments One hundred thousand cells were plated per 35-mm dish and transfected the next day with Lipofectamine 2000 and 1 mg plasmid encoding GFP (as control), wild-type rat MLC2 (MYL12B) fused with GFP or inactive phospho-mutant TASA-MLC2 fused with GFP (T18A, S19A) (Calvo et al., 2013) (plasmids were a gift from Fernando Calvo). Next day cells were transfected with 26 nM siRNA oligonucleotides against MYL12B. Cell death was assessed 2-3 days after siRNA transfection by PI (1.5 mM) incorporation by + + FACS. Percentage of dead cells (PI ) was quantiﬁed within transfected (GFP ) cells. MLC2 Stable Overexpression Lentivectors encoding EGFP-fused rat MLC2-derivatives (wild-type, phospho-mimetic TDSD (T18D, S19D) and inactive phospho- mutant TASA (T18A, S19A)) (Takaki et al., 2017) were a kind gift from Erik Sahai and Tohru Takaki (The Francis Crick Institute). HEK293T cells were transfected with MLC2-lentivectors and packaging plasmids using standard procedures, and after 48 hr super- natants were collected and spun down to remove debris. A375 cells were transduced with lentiviral supernatants for 8 hr, and 48 hr later cells were selected with 1 mg/ml puromycin for 5 days, then cells were used for subsequent experiments. Immunoﬂuorescence and Confocal Imaging Cells were ﬁxed with 4% formaldehyde, permeabilised with 0.2% Triton X-100 for 5 min, blocked with 5% BSA-PBS for 1 hr at room temperature, and incubated with anti-p-MLC2 (p-MLC2S19, 1:200 in 5% BSA-PBS) overnight at 4 C. Alexa-488 anti-rabbit second- ary antibody (Life Technologies) was used at 1:500 for 1 hr at room temperature. F-actin was detected with Phalloidin (1 hr RT) and nuclei were stained with Hoechst 33258 (Life Technologies). Imaging was carried out on a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss) with C-Apochromat 3 40/1.2 NA (water) or a Plan Apochromat 3 63/1.4 NA (oil) objective lenses and Zen software (Carl Zeiss). Line scan analysis was performed in ImageJ. Immunoblotting Cells were lysed in Laemmli buffer and snap frozen. Lysates were then boiled, sonicated for 15 s and spun down. Cell lysates were fractionated using sodium dodecyl sulfate-polyacrylamide (SDS-PAGE) gels in non-reducing conditions, and transferred subse- quently to PVDF ﬁlters. Membranes were blocked in 5% BSA in 0.1% Tween 20-TBS. Primary antibodies were incubated overnight at 4 C. For detection, ECL or Prime ECL detection systems coupled to HRP-conjugated secondary antibodies (GE Healthcare) with X-ray ﬁlms and an Amersham Imager 600 were used. Bands were quantiﬁed using ImageJ. Levels of phospho-proteins were calcu- lated after correction to total levels of the relevant protein. TGF-b1 ELISA Cells were seeded on T6-well plates (150,000 cells/well), next day cells were washed 3 times and then grown in serum-free media with or without ROCKi GSK269962A (5 mM). Forty-eight hr later supernatants were collected, spun down and assayed fresh or frozen TM at -80 C. TGF-b1 levels were detected by ELISA using Total TGF-b1 Legend Max ELISA Kit with Pre-coated plate (#436707, Bio- legend) on neat samples diluted 1/5 following the manufacturer’s instructions. Phospho-proteomics Preparation of tandem mass tagged (TMT)-multivariate phosphoproteomic samples. Cells treated with MEKi (200 nM GSK1120212 trametinib or 200 nM PD184352) or vehicle (DMSO) for 24 hr were lysed in 6 M urea, sonicated, centrifuged to clear cell debris and protein concentration was determined by BCA (Pierce 23225). 100 mg of each condition was individually digested by FASP (PMID: 19377485) using 1:100 Lys-C (Wako 125-05061), 1:100 Trypsin (Worthington), and amine-TMT-10 plex labeled (Pierce 90111) on membrane (iFASP) (PMID: 23692318). TMT channel assignment: 126 = Control (Bio. Rep. 1); 127N = Control (Bio. Rep. 2), 127C = Control (Bio. Rep. 3); 128N = Control (Bio. Rep. 4); 128C = MEKi A (Bio. Rep. 1); 129N = MEKi A (Bio. Rep. 2); 129C = MEKi A (Bio. Rep. 3); 130N = MEKi B (Bio. Rep. 1); 130C = MEKi B (Bio. Rep. 2); 131 = MEKi B (Bio. Rep. 3) (A= GSK1120212, B= PD184352). Peptides were then eluted, pooled, lyophilized and subjected to automated phosphopeptide enrichment (APE) (PMID: 25233145). Phosphopeptides were desalted using OLIGO R3 resin (Life Technologies 1-1339-03) and lyophilised prior to LC-MS/MS analysis (see below). Data-dependent acquisition LC-MS/MS. Phosphopeptide samples were resuspended in 0.1% formic acid and analyzed on a Q-Exactive Plus mass spectrometer (Thermo Scientiﬁc) coupled to a Dionex Ultimate 3000 RSLCnano System (Thermo Scientiﬁc). Reversed-phase chromatographic separation was performed on a C18 PepMap 300 A trap cartridge (0.3 mm i.d. x 5 mm, 5 mm bead size; loaded in a bi-directional manner), a 75 mm i.d. x 50 cm column (5 mm bead size) using a 120 min linear gradient of 0-50% solvent B (MeCN 100% + 0.1% formic acid (FA)) against solvent A (H2O 100% + 0.1% FA) with a ﬂow rate of 300 nL/min. The mass spec- trometer was operated in the data-dependent mode to automatically switch between dual Orbitrap MS and MS/MS acquisition. Sur- vey full scan MS spectra (from m/z 400-2000) were acquired in the Orbitrap with a resolution of 70,000 at m/z 400 and FT target value of 1 x 106 ions. The 20 most abundant ions were selected for fragmentation using higher-energy collisional dissociation (HCD) and dynamically excluded for 30 s. Fragmented ions were scanned in the Orbitrap at a resolution 35,000 at m/z 400. The isolation window Cancer Cell 37, 85–103.e1–e9, January 13, 2020 e6 was reduced to 1.2 m/z (to reduce ion co-isolation) and a MS/MS ﬁxed ﬁrst mass of 120 m/z was used (to ensure consistent TMT reporter ion coverage). For accurate mass measurement, the lock mass option was enabled using the polydimethylcyclosiloxane ion (m/z 445.120025) as an internal calibrant. For peptide identiﬁcation, raw data ﬁles produced in Xcalibur 2.1 (Thermo Scientiﬁc) were processed in Proteome Discoverer 1.4 (Thermo Scientiﬁc) and searched against Human Unitprot database using Mascot (v2.2). Searches were performed with a precursor mass tolerance set to 10 ppm, fragment mass tolerance set to 0.05 Da and a maximum number of missed cleavages set to 2. Static modiﬁcations were limited to carbamidomethylation of cysteine, and variable modiﬁcations used were oxidation of methionine, deamidation of asparagine/glutamine, and phosphorylation of serine, threonine and tyrosine residues. Peptides were further ﬁltered using a mascot signiﬁcance threshold <0.05, a peptide ion Score >20 and a FDR <0.01 (evaluated by Percolator (PMID: 17952086)). Phospho-site localization probabilities were calculated with phosphoRS 3.1 (>75%) (PMID: 22073976). For relative phosphopeptide quantiﬁcation, MEKi/vehicle ratios were calculated by Proteome Discoverer 1.4. See Data and Code Availability section below for further details. Phosphoproteomic data analysis. Phosphopeptides from Proteome Discoverer 1.4 were normalised against total protein levels (from SILAC in-gel digest experiments), and protein-level phospho-site locations (phosphoRS 3.1 score >75%, maximum 4-PTM/ peptide) were manually annotated using PhosphoSitePlus. Precursor ion spectra, extracted ion chromatograms, and product ion spectra were manually inspected for each regulated phosphopeptide. Empirical parent kinases were manually identiﬁed by refer- enced Uniprot annotation and putative parent kinases were manually assigned using ScanSite (PMID: 12824383) 3 (top 1 percentile of all sites, lowest score). Phospho-sites that did not meet these conditions were not annotated. Regulated phospho-peptides in Table S1 were those which were signiﬁcant across both MEKi (GSK1120212 and PD184352) compared to vehicle-treated cells. Phospho-Peptide Enrichment Analysis Pathway enrichment analyzes of the list of phospho-peptides increased in MEKi-treated A375 compared to vehicle-treated A375 cells (this study, see Phospho-proteomics section; Table S1); A375/PLX/R compared to A375 cells (data from (Girotti et al., 2013)) and M229- and M238-vemurafenib-resistant vs parental cells from (Titz et al., 2016) were performed using MetaCore from GeneGo Inc. (https://portal.genego.com/). Quantitative Real Time One-Step PCR RNA was isolated using TriZol (Life technologies). For experiments comparing expression in parental vs BRAFi-resistant cells (A375- and Colo829-derivatives), resistant cells were cultured with 1 mM PLX4720 and sensitive cells with equivalent volume of DMSO for 24 hr. QuantiTect Primer Assays (Qiagen) and Brilliant II SYBR Green QRT-PCR 1-step system (Agilent Technologies) with 100 ng RNA were used following the manufacturer’s instructions. GAPDH was used as loading control. The following QuantiTect Primers were used (Qiagen): GAPDH (QT00079247), LIMK1 (QT00008680), LIMK2 (QT00084357), MKL1 (QT00067921), MKL2 (QT00010115), MYH9 (QT00073101), MYL9 (QT00072268), MYL12A (QT01665741), MYL12B (QT00075264), ROCK1 (QT00034972), ROCK2 (QT00011165). Primer sequences are not provided by Qiagen, as stated in their website: ‘Sequences of the QuantiTect Primer Assays are not provided. Approximate location of primers within a speciﬁc gene can be viewed on the Product Detail pages retrieved via our GeneGlobe data base.’ Gene Expression Studies and Analysis Normalized gene expression microarray and RNAseq (FPKM, fragments per kilobase of transcripts per million mapped reads) data from published studies were downloaded from Gene Expression Omnibus (GEO) unless otherwise stated: Hugo 2015 (GSE65185 and GSE65184) (Hugo et al., 2015); Hugo 2016 (GSE78220) (Hugo et al., 2016); Kakavand 2017 (GSE99898) (Kakavand et al., 2017); Kwong 2015 (European Genome-phenome Archive (EGA S00001000992)) (Kwong et al., 2015); Long 2014 (GSE61992) (Long et al., 2014a); Obenauf 2015 (GSE64741) (Obenauf et al., 2015); Rizos 2014 (GSE50509) (Rizos et al., 2014); Riaz 2017 (Ipi-naive cohort; GSE91061) (Riaz et al., 2017); Song 2017 (GSE75299, GSE103630) (Song et al., 2017); Sun 2014 (GSE50535) (Sun et al., 2014); Wagle 2014 (GSE77940) (Wagle et al., 2014). In patients with several biopsies, their average is shown (see Table S4). RSEM-normalized expression data and clinical information of human melanoma samples (70 primary and 319 metastatic mela- nomas) from The Cancer Genome Atlas (TCGA) database were downloaded from Firehose (https://gdac.broadinstitute.org/). Only TCGA samples with no neo-adjuvant treatment prior to tumor resection were considered. The ROCK-myosin II pathway expression signature (MYL9, MYL12A, MYL12B, MYH9, ROCK1, ROCK2, LIMK1, LIMK2, MKL1, MKL2, MYLK, DAPK3) was generated by the sum of normalized expression values of signature genes for each TCGA patient. ROCK-myosin II pathway signature was categorized as low or high using the mean expression. Heatmaps and unsupervised hierarchical clustering analyzes were generated using Multiexperiment Viewer (http://www.tm4.org/ mev.html). Distance metric used for the clustering was Euclidean distance. In patients with several biopsies, their average is shown. Gene Enrichment Analyzes Gene sets for cross-resistance processes (EMT, metastasis, angiogenesis, hypoxia, wound healing, TGF-b, STAT3, NF-kB, YAP) were downloaded and analyzed using Gene Set Enrichment Analysis (GSEA) software (http://www.broadinstitute.org/gsea/index. jsp) with the settings: permutations-1,000, permutation type-gene set, metric for ranking genes-t-test. Signiﬁcantly enriched gene sets in resistant vs baseline samples were considered according to p value <0.05 and FDR <0.25 in at least 2 of the 5 comparisons e7 Cancer Cell 37, 85–103.e1–e9, January 13, 2020 performed. To calculate the gene-signature score in each sample, we used single-sample Gene Set Enrichment Analysis (ssGSEA) Projection Software from GenePattern platform (https://www.broadinstitute.org/cancer/software/genepattern). For the transcriptional signature of melanoma cells with high myosin II activity, genes upregulated in high myosin II activity compared to low myosin II activity melanoma cells (cells treated with ROCKi and blebbistatin) (Cantelli et al., 2015; Sanz-Moreno et al., 2011) were selected using a fold changeR 1.5 and a p value <0.01. GSEA analysis was performed as described above. Enrich- ment plot (green line) show upregulation of gene signature in indicated samples (resistant, non-responders or on-treatment). Nominal p values are shown along plot, false discovery rate (FDR) in ﬁgure legend. For analysis of ROS-related gene signatures, all available ROS/oxidative stress gene sets were downloaded from GSEA Broad Institute (http://www.broadinstitute.org/gsea/index.jsp). Graph shows (-Log ) p value. For analysis of expression of DNA repair genes, we compiled a DNA repair gene signature from the list in (Mjelle et al., 2015) and the homologous recombination defect signature (Peng et al., 2014). Network enrichment analysis of genes commonly downregulated (<0.65-fold) in at least 4 of 7 cell lines from Group 1 (Figure 2B) was performed using Ingenuity Pathway Analysis (Qiagen). Tumor Xenografts A375/PLX/R cells (1 x 10 ) were injected subcutaneously into the right ﬂank of 5-week-old female nude CD-1 mice (Charles River). 6 6 Patient #2 cells (4 x 10 ) or Patient #35 cells (6 x 10 ) were injected into 5-8-week old NOD/SCID/ IL-2Rg-/- (NSG, Charles River) mice (male and female). Tumors were allowed to establish, sizes (average 60-100 mm ) were matched and then mice were randomly allo- cated to groups of 7-8 animals. Treatment was by orogastric gavage with 45 mg/kg PLX4720, 10-25 mg/kg GSK269962A or both drugs together. GSK269962A was used at 25 mg/kg for A375/PLX/R and 10 mg/kg for Patient #2, #35. Drugs were dissolved in 5% DMSO or in 6% DMSO+50% PEG300+ 9% Tween 80. All the drugs were administered daily, 7 days a week. Tumor size was determined by caliper measurements of tumor length, width and depth and tumor volume was calculated as volume = 0.5236 x length x width x depth (mm). Immunotherapy Experiments 5555 cells (100,000, 250,000 or 1 million) were subcutaneously injected into the right ﬂank of 5-7-week-old female C57BL/6J mice. After 7-14 days, mice with tumors (50-80 mm ) were randomly allocated into groups of 6-7 animals and treated daily with ROCKi GSK269962A (10 mg/kg, oral gavage) or vehicle and every 3 days with anti-PD-1 monoclonal antibody (InVivoPlus clone RMP1-14, BioXCell #BE0146) (10 mg/kg, intraperitoneally (i.p.)) or rat IgG2a isotype control (clone 2A3 BioXCell # BE0089). Vehicle for ROCKi was 5% DMSO or 5% DMSO, 10% Tween 80, 6.5% ethanol. Tumor volume was determined as above. Anti-PD-1-non- responder (NR) lines were established in culture by digesting tumors with a mixture of Liberases (TH and TM, 75 mg/ml each, Roche Diagnostics) and 1 mg/ml DNase I (Sigma) in HBSS for 1 hr at 37 C with shaking, and then passed through 100 mM strainers. For ex- periments using 5555-anti-PD-1/NR cells, 1 million cells were injected subcutaneously into 7-week old C57BL/6J mice. Next day, all mice were given 1 dose of anti-PD-1 (10 mg/kg) i.p., and then again 3 days later. At day 7, mice were randomized into 4 treatment groups (ROCKi, anti-PD-1, combo or control) as above. Survival in the Lung Assay Patient #2 cells were pre-treated for 24 hr with 5 mM PLX4720, 5 mM GSK269962A or both (control had DMSO), then cells were labelled with 10 mM CMFDA-Green in OptiMem (Life Technologies) for 10 min, trypsinized and equal numbers were injected into the tail vein of NSG mice in 100 ml PBS along with drugs (same concentrations as pre-treatment). At the time of injection, mice (male and female) were 6-10 weeks old and weighed around 20-22 g; mice were age and sex-matched between the groups. Mice were sacriﬁced 30 min (to conﬁrm that equal numbers arrived at the lung) and 24 hr after tail vein injection. The lungs were ex- tracted, washed twice with PBS, ﬁxed (4% formaldehyde for 16 hr at 4 C) and examined for ﬂuorescently-labelled cells under a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss) with a 20X objective. Lung retention is represented as ﬂuorescence area (CMFDA- Green from melanoma cells) per ﬁeld, and approximately 20 ﬁelds per mouse lung were analyzed. Each experiment had 4-5 mice/ condition, and experiments were replicated twice and data pooled together. Quantiﬁcation of survival in the lung 24 hr after injection is shown as mean ﬂuorescence area/ﬁeld. Immunohistochemistry Tumors and spleens were formalin-ﬁxed and parafﬁn-embedded using standard protocols. For cell pellets, transfected cells were harvested 48 hr after transfection using a cell scraper, spun down, ﬁxed with 4% formalin for 30 min and washed with PBS. Cell pellet was resuspended in 2% agarose and then embedded in parafﬁn. Four mm thick sections were incubated at 60 C for 20 min and then subjected to antigen retrieval using Access Super Tris pH 9 buffer (A.Menarini Diagnostics) at 110 C for 6 min in a Decloaking Cham- ber NxGen (Biocare Medical). Samples were blocked with Dual Endogenous Enzyme-Blocking Reagent (Dako) for 10 min and then were incubated with primary antibodies for 40 min at RT, washed and then incubated with biotinylated secondary antibodies (rabbit, mouse or rat; 1:200; Vector-Labs) for 30 min at RT. Signal was then ampliﬁed using VECTASTAIN ABC HRP kit (PK-4000) for 20 min at RT and the reaction was developed using VIP substrate (SK-4600, Vector-Labs) for 10 min at RT. Stainings were counter- stained with Hematoxylin. Positive and negative controls were included in each experiment, including staining of melanoma markers Cancer Cell 37, 85–103.e1–e9, January 13, 2020 e8 HMB45/Melan-A or S100. For ECM staining, samples were ﬁxed in Bouin’s solution (HT10132, Sigma) for 1 hr at 60 C, then stained with Weigert’s iron hematoxylin solution (HT1079, Sigma) for 5 min at RT and with Trichrome Stain (Masson) Kit (HT15-1KT, Sigma) following the manufacturer’s instructions. Imaging and Scoring Sections from tumor xenograft experiments and from paired melanoma samples from 12 patients (tumor tissue before and after treatment) were imaged using NanoZoomer S210 slide scanner (Hamamatsu, Japan). Staining quantiﬁcation was performed using QuPath 0.1.2 (Bankhead et al., 2017). For p-MLC2 stainings, whole sections were scanned and images were analyzed performing positive cell detection, and three different thresholds were applied according to the intensity scores (0, 1, 2 and 3). Next, the software was trained by creating random trees classiﬁcation algorithm combined with the intensity information, in order to differentiate tumor from stroma, necrosis and immune cells. Values used in the analysis correspond to the quantiﬁcation of p-MLC2 in the invasive front (mouse tumors) or highest score in the whole section (human samples). To characterize the immune inﬁltrate (CD206, F4/80, CD3, CD4, CD8 and FOXP3) a similar approach was performed using QuPath. First, positive cell detection was applied, using only a single value to differentiate negative (blue) from positive (red). Data are repre- sented as cellular density (cells/mm ). For PD-L1 analysis, CD206 cells were identiﬁed and both PD-L1 and CD206 stainings were aligned using QuPath 2.03m. From + + CD206 staining, positive detections (CD206 ) were transferred to PD-L1 in order to quantify the actual score for PD-L1 in CD206 cells. The negative detection for CD206 was used to quantify PD-L1 on tumor cells, these were identiﬁed as CD206 after discarding stromal/immune cells. Image composition was performed artiﬁcially attributing a color code, and images were overlaid using ImageJ (trackEM2). For PD-L1 and CD206, merge images in Figure 8I were generated with QuPath by overlaying pseudo-color images for each staining. For ECM analysis with Masson’s Trichrome staining, whole section images were quantiﬁed with QuPath applying a SLIC algorithm for segmentation of sections according to pixel density. Next, colors were deconvoluted and the green channel was used to quantify the percentage of the area occupied by collagen. QUANTIFICATION AND STATISTICAL ANALYSIS GraphPad Prism (GraphPad Software) was used to perform unpaired two-tailed t-test, Mann-Whitney test, Wilcoxon test, one-way or two-way ANOVA with post hoc tests (Tukey’s, Dunnet’s, Benjamini, Krieger and Yekutieli correction), Kruskal-Wallis, Deming linear regression, Spearman correlation and Chi-square test. Survival curves were estimated by the Kaplan-Meier method and the log-rank test using SPSS (IBM). Details of statistical analysis performed are in the ﬁgure legends. Bar graphs report mean ± SEM with indi- vidual data points as explained in ﬁgure legends. Box plots show median (center line); interquartile range (box); min-max values (whis- kers). In Figure legends, ‘‘n’’ means number of independent experiments unless otherwise stated. Signiﬁcance was deﬁned as p<0.05. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns not signiﬁcant. DATA AND CODE AVAILABILITY The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://www.proteomexchange. org) via the PRIDE partner repository (PMID: 23203882) with the dataset identiﬁer PXD002621 (https://www.ebi.ac.uk/pride/archive/ projects/PXD002621). e9 Cancer Cell 37, 85–103.e1–e9, January 13, 2020

Journal

Cancer Cell – Unpaywall

Published: Jan 1, 2020

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance

Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance

Myosin II Reactivation and Cytoskeletal Remodeling as a Hallmark and a Vulnerability in Melanoma Therapy Resistance

References (103)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies