Pharmacological Inhibition of Myocardin-related Transcription Factor Pathway Blocks Lung Metastases of RhoC-Overexpressing Melanoma

Rho GTPases play significant roles in human cancer (1). Although RhoA and RhoC share considerable homology, they may coordinate different aspects of cancer progression. RhoC appears to be most important for cellular invasion and metastasis (2) while RhoA plays a role in transformation and tumor proliferation (3). In a screen designed to identify genes important for melanoma metastasis, RhoC was found to be upregulated and it contributed to melanoma lung metastasis (4). Rho GTPases are well-known for regulating the actin cytoskeleton (5). RhoA/C activation causes the formation of myosin-rich F-actin bundles called stress fibers through their downstream effectors Rho-associated kinase (ROCK) and Diaphanous-related formin-1 (mDia1; refs. 6, 7). By modulation of the actin cytoskeleton, Rho GTPases also regulate gene expression though myocardin-related transcription cofactors (MRTF-A/B) also known as megakaryoblastic leukemia proteins (MKL1/2). In the nucleus, MRTF cooperates with serum response factor (SRF) to induce transcription of genes involved in proliferation, migration, invasion, and metastasis (8, 9). In the N-terminal region of MRTF are two G-actin–binding RPEL motifs which sequester MRTF in the cytoplasm. Rho activity reduces cytosolic G-actin, allowing MRTF to translocate into the nucleus and activate gene transcription. Stable knockdown of MRTF or SRF in B16M2 melanoma or human MDA-MB-231 breast cancer cells with high RhoC expression results in a dramatic reduction of in vivo lung metastasis and in vitro cellular migration (10). This evidence suggests that gene regulation through MRTF mediates RhoC-induced cancer metastasis.

To assess the potential for targeting MRTF-regulated gene transcription in cancer metastasis, we used a previously identified small-molecule inhibitor of the MRTF pathway (11). Our initial compound, CCG-1423, selectively blocked Rho GTPase–regulated prostate cancer cell invasion at low micromolar concentrations (11). Further structure–activity relationship optimization of this compound series to reduce toxicity (12) and to enhance potency resulted in a new analogue (13), CCG-203971, that also blocks MRTF-induced gene expression and prostate cancer migration. The Rho/MRTF pathway has also recently been highlighted as a major signaling crossroad in the development of pathologic fibrosis in diseases including idiopathic pulmonary fibrosis, scleroderma, and Crohn disease (14, 15). We have recently shown efficacy of CCG-203971 in a murine model of skin fibrosis (16) but a close examination of MRTF pathway inhibitors in models of cancer progression has not been performed.

Here we investigate the Rho/MRTF pathway in two metastatic melanoma cell lines, SK-Mel-19 and SK-Mel-147. There is a clear dichotomy between the two lines where the high level of RhoC, constitutive MRTF-A nuclear localization, and function in SK-Mel-147 is associated with in vivo and in vitro correlates of cancer metastasis including clonogenicity, migration, and invasion. As a novel approach to antimetastasis therapeutics, our MRTF pathway inhibitor, CCG-203971, blocks cellular migration and invasion and profoundly suppresses SK-Mel-147 lung metastasis in vivo.

Human cutaneous melanoma cell lines, SK-Mel-19 and SK-Mel-147, obtained from Dr. Maria Soengas at the University of Michigan, were cultured in DMEM (Life Technologies) supplemented with 10% FBS (Life Technologies) including penicillin/streptomycin (Life Technologies). Short tandem repeat profiles were conducted on the melanoma cell lines (Genewiz). The profiles obtained for the melanoma cell lines do not match any established published profiles, and we are unable to identify published profiles for SK-Mel-19 and SK-Mel-147. Cells were expanded and frozen immediately prior to authentication and then resuscitated only two to three months before experiments. We verified the driver mutations in our laboratory using MSU Sanger sequencing service expected from previous reports (i.e., BRAF^V600E for SK-Mel-19 and NRAS^Q61L for SK-Mel-147; refs. 17, 18). For proliferation experiments, 1 × 10⁴ cells were plated into 24-well plates in DMEM containing 10% FBS. Cells were allowed to attach overnight and then medium was replaced with DMEM containing 0.5% FBS. After 72 hours of growth, cells were stained with 0.4% Trypan blue stain (Invitrogen). Cell counting was performed using the Countess Automated Cell Counter (Invitrogen) according to the manufacturer's instructions.

Cells (1 × 10⁵) were plated onto fibronectin-coated coverslips in DMEM with 10% FBS and allowed to attach overnight. Medium was changed to DMEM with 0.5% FBS for 24 hours plus the indicated concentration of CCG-203971 or 0.1% DMSO. Cells were then fixed in 3.7% formaldehyde for 10 minutes at room temperature and permeabilized with 0.25% Triton X-100 for 10 minutes at room temperature. For immunocytochemistry studies, primary antibody for MRTF-A (Santa Cruz Biotechnology, cat# SC-21558) was diluted 1:100 and incubated for 2 hours at room temperature. Fluorophore-conjugated secondary (Alexa Fluor 594 Donkey Anti-Goat IgG, Invitrogen) was diluted 1:1,000 and added for 1 hour. For stress fiber assays, a final concentration of 100 nmol/L Acti-Stain phalloidin (Cytoskeleton) in 100 μL 1× PBS/1% BSA was added to each coverslip for 30 minutes at room temperature. Cells were mounted (Prolong Gold antifade-reagent with DAPI, Invitrogen) and imaged on an upright fluorescence microscope (Nikon E-800) at 60× magnification.

Cells (1 × 10⁶) were plated into 60-mm dishes and starved overnight in DMEM containing 0.5% FBS. Cells were lysed and RNA was isolated using the RNeasy Kit (Qiagen) following the manufacturer's instructions. DNAse-treated RNA (1 μg) was used as a template for synthesizing cDNA utilizing the TaqMan Reverse-Transcription Reagents Kit (Invitrogen). SYBR green qPCR (SABiosciences) was performed using a Stratagene Mx3000P (Agilent Technologies) and C_t values were analyzed relative to GAPDH expression. Primer sequences used were as follows: GAPDH: 5′-GGAAGGGCTCATGACCACAG-3′, 3′ ACAGTCTTCTGGGTGGCAGTG-5′; CTGF: 5′-CAGAGTGGAGCGCCTGTT-3′, 3′-CTGCAGGAGGCGTTGTCA-5′; CYR61: 5′-CGCGCTGCTGTAAGGTCT-3′, 3′–TTTTGCTGCAGTCCTCGTTG-5′; MYL9: 5′-CATCCATGAGGACCACCTCCG-3′, 3′–CTGGGGTGGCCTAGTCGTC-5′.

Cells (4 × 10⁵) were plated in DMEM containing 10% FBS in a 12-well plate. After 24 hours, a scratch was made using a 200-μL pipette tip. Medium was replaced with DMEM containing 2.0% FBS and images of the wound were taken at time = 0 and 24 hours using a bright field microscope (Olympus CKX41). Area quantification of the scratch was determined using ImageJ software as described previously (13, 19). Briefly, the binary (white and black) threshold for each image was manually adjusted to leave only the open area of the scratch black. The area of this scratch was determined and the percent closure was calculated by comparing the difference in area from t = 0 hour to t = 24 hours.

Cells (5 × 10⁴) were added to the top of an 8-μm pore diameter Transwell chamber (Millipore) in DMEM containing 0.5% FBS with 10 μmol/L CCG-203971 or 0.1% DMSO control. Medium containing 0.5% FBS and compound or DMSO was also added to the bottom chamber. Cells were allowed to migrate for 6 hours, and after wiping the remaining cells off the top of the Transwell chamber, cells which migrated were fixed and stained in 4% formaldehyde/0.5% crystal violet. The number of migrated cells for each well was determined by examining four random nonoverlapping fields of view at 20× magnification using a bright field microscope (Olympus CKX41). For knockdown experiments, cells transfected for 48 hours with siRNA pools were plated, allowed to migrate, and analyzed as stated above.

ON-TARGET plus SMART pools (GE Dharmacon) targeting human MRTF-A (GE cat# L-015434-00) or nontargeting pool control (GE cat# D-001810-10-05) were used according to the manufacturer's recommendations. Briefly, 25 nmol/L siRNA smart pools were transfected using DharmaFECT (GE cat# T-2001-01). Western blot validation of MRTF-A expression was conducted using a fraction of the same transfected population used for the Transwell assays.

Real-time cell analysis (RTCA) of cell invasion was monitored using a CIM-plate 16 and xCELLigence DP System (Acea Bioscience, Inc). Matrigel Matrix (BD Biosciences) 5% (v/v) was diluted 1:20 in basal RPMI media and was added in the top chamber and allowed to equilibrate for 4 hours in an incubator at 37°C in 5% CO₂. Cells (3 × 10⁵/well) were added at t = 5 hours. For compound effects in invasion assays, cells were treated with 10 μmol/L CCG-203971 upon placement into the top chamber. Cell index was measured every 30 minutes.

Cells (3 × 10⁴) were seeded into 96-well plates in DMEM containing 10% FBS overnight. Cells were then cotransfected with 56 ng SRE.L- Luciferase reporter selective for Rho/MRTF described in ref. 20, and 2.3 ng RhoC plasmid diluted with Lipofectamine 2000 (Invitrogen) in Opti-MEM (Life Technologies). After 6 hours, compounds were added for overnight treatment. The next day, WST-1 (10 μL/well) reagent (Promega) was added to measure viability. After 1 hour, absorbance was read (450 nm) before lysing the cells for luciferase measurement (Promega Luciferase Assay System).

Cells were grown for 6 days in DMEM with 2.0% FBS. Two hundred live cells, determined by Trypan blue exclusion assay, were seeded into a 6-well plate and allowed to form colonies in DMEM containing 2.0% FBS for 10 days. Colonies were fixed and stained in 3.7% formaldehyde/0.5% crystal violet for 10 minutes at room temperature. Only cell clusters consisting of at least 50 cells were counted as colonies. For CCG-203971 treatment, cells (2 × 10⁵) were seeded into 6-well dishes and allowed to adhere overnight. The following day, the medium was replaced with DMEM containing 0.5% FBS and 0.1% DMSO or 10 μmol/L CCG-203971 was added. Seventy-two hours later, 200 or 1 × 10³ cells were plated per group in DMEM with 10% FBS in 6-well dishes and allowed to form colonies for 7 days. Colonies were fixed and stained as stated above. Colony counts and mean area quantification were determined using ImageJ software. Briefly, the image threshold was adjusted to black and white and further processed to smooth colony pixilation. The images were then analyzed for colony size (pixel²) fixed at 50–infinity and for circularity defined at 0.2–1.0.

Cells (5 × 10⁵) were plated in 60-mm dishes in DMEM with 10% FBS and allowed to adhere overnight. Media were replaced with DMEM with 0.5% FBS, and 0.1% DMSO or 10 μmol/L CCG-203971 was added. Cells were allowed to grow for 72 hours and were then collected using 1× versene (Gibco), permeabilized, and fixed using 70% ethanol for 30 minutes on ice. Cells were centrifuged (200 × g, 5 minutes), and resuspended in DNA extraction buffer (192 mL 0.2 mol/L Na₂HPO₄, 8 mL 0.1 mol/L citric acid, pH 7.8, filter sterilized) for 5 minutes. Cells were centrifuged (300 × g, 5 minutes) to remove DNA extraction buffer and resuspended in 250 μL of freshly prepared DNA staining solution (2 μg/mL propidium iodide, 0.2 mg/mL RNase A in 1× PBS) and incubated at room temperature for 30 minutes in the dark. Fluorescence was detected using a C6 BD Accuri flow cytometer (BD Accuri). Fluorescence was quantified using CFlow software (BD Accuri). Cell-cycle analysis was conducted using FlowJo v10.0.8 software (Tree Star, Inc.). A total of 1 × 10⁴ events were detected in experimental triplicate for untreated cells, 0.1% DMSO, and 10 μmol/L CCG-203971. Three independent experiments were conducted.

Cells were starved overnight in DMEM containing 0.5% FBS. Cells were lysed [10 mmol/L Tris-Cl, 1 mmol/L EDTA, 1% Triton X-100, 0.1% SDS, 140 mmol/L NaCl, protease inhibitor cocktail (Roche cat# 11873580001)], sonicated two times for 15 seconds, and centrifuged to quantitate soluble proteins using a BCA protein assay as recommended by the manufacturer (Pierce BCA Protein Assay, Thermo Scientific). Protein lysate (20 μg) were resolved using 15% SDS-PAGE gels, transferred to PVDF membranes, and blocked in 5% dried milk in Tween Tris-buffered saline. Membranes were incubated in 1:1,000 diluted primary antibodies RhoC (Cell Signaling Technology, cat# 3430), RhoA (Santa Cruz Biotechnology, cat# SC-418), MRTF-A (Santa Cruz Biotechnology, cat# SC-21558), or GAPDH (Cell Signaling Technology, cat# 14C10) overnight at 4°C. HRP-conjugated goat anti-mouse (Santa Cruz Biotechnology, sc-2060), HRP-conjugated goat anti-rabbit (Sigma-Aldrich, A0545), or HRP-conjugated bovine anti-goat (Santa Cruz Biotechnology, cat# SC-2352) antibodies were diluted 1:10,000. Blots were developed using SuperSignal Pico chemiluminescent substrate (Thermo Scientific). Bands were visualized and quantified using a LI-COR Odyssey Fc or visualized using Kodak X-OMAT Film Developer and quantified using ImageJ software.

Ten-week old, female, NOD.CB17-Prkdc^scid/J (NOD SCID) mice were purchased from Jackson Laboratory (stock number: 001303). The mice were separated into three groups of 6. One group received tail vein injection of GFP-expressing SK-Mel-19 (2 × 10⁶ cells). The other two groups were injected with GFP-expressing SK-Mel-147 (2 × 10⁶ cells). Following tail vein injection, the SK-Mel-147 mice began twice daily intraperitoneal injection of 100 mg/kg CCG-203971 or 50-μL DMSO (vehicle control). After 18 days, the mice were euthanized and lungs were collected for histologic analysis. All studies were performed according to protocols approved by the University of Michigan, University Committee on the Use and Care of Animals (UCUCA).

For the quantification and imaging of lung tissue from the mouse metastasis model, paraffin-embedded sections were dewaxed, unmasked, and treated as described previously (21). To visualize GFP-expressing melanoma cells in the lung section, the Vector Rabbit-on-Mouse basic kit (Vector Laboratories) was utilized according to the manufacturer's instructions. For each experiment, two tissue sections from the same animal were developed: one with primary antibody [rabbit anti-GFP (1:250 Invitrogen A-6455)] included and one with the primary antibody left out of the reaction. Visualization of GFP was done using the Vectastain Elite ABC kit (Vector Laboratories), used according to the manufacturer's instructions. All sections were developed using 3,3′-diaminobenzidine (Vector Laboratories) and counterstained with hematoxylin. Images were taken at 20× magnification using a Nikon Eclipse Ti-S microscope with MMI Cell Tools software, Version 3.47. (MMI). Treatment and control slides were imaged using the same exposure time and LUT settings. The slides were blindly quantified for the number of metastatic nodules on each section followed by the use of ImageJ software to determine the cross-sectional area (μm²) of each nodule. The area was represented as an average size for each mouse and the total tumor burden was calculated by adding the total cross-sectional areas for all tumors in each mouse then averaged for each group.

Statistical analysis was performed by ANOVA comparison followed by Bonferonni posttest for multiple comparisons where *P > 0.05, **P > 0.01, ***P > 0.001, and ****P > 0.0001. Comparison of survival curves was performed by log-rank method where *P > 0.05 and **P > 0.01.

Previously published results showed that SK-Mel-147 cells were highly metastatic in immunocompromised mice while SK-Mel-19 cells were not (22), so to confirm these observations, we compared phenotypes of SK-Mel-147 and SK-Mel-19 using in vitro models of cancer aggressiveness. SK-Mel-147 readily forms colonies while SK-Mel-19 is unable to form quantifiable colonies under these conditions (Fig. 1A). We next examined cellular migration and invasion using the scratch-wound migration assay (Fig. 1B) as well as Matrigel invasion assay (Fig. 1C), respectively. Consistent with previous results, SK-Mel-147 cells were much more migratory and invasive than SK-Mel-19. Rho GTPase signaling has been shown to promote cellular migration and invasion (23). More specifically in melanoma, RhoC expression was elevated in cell lines selected in mice for increased metastatic behavior (5). To determine the relative Rho signaling potential within SK-Mel-147 and SK-Mel-19, we quantified RhoA and RhoC protein expression (Fig. 1D and E). RhoC protein is overexpressed in the more aggressive melanoma cell line, SK-Mel-147, while RhoA levels are not significantly different between the cell lines (Fig. 1E). RhoA plays an important role in proliferation and cytokinesis, whereas RhoC seems to be more involved in migration, invasion, and metastasis (3, 4). In addition to increased RhoC expression, SK-Mel-147 also shows modestly increased MRTF-A expression levels compared with SK-Mel-19, but it is not statistically significant (Fig. 1D and E). To understand the potential contribution of RhoC and MRTF in clinical melanoma, we explored survival data for cutaneous melanoma datasets in The Cancer Genome Atlas (TCGA; ref. 24). Stratification of high top quartile versus low bottom quartile MRTF-A or RhoC expression levels shows a significant reduction of percent survival in patients with high MRTF-A or RhoC expression (Fig. 1F). Interestingly, stratification of RhoA, RhoB, and MRTF-B high versus low expression does not result in significant differences in melanoma patient percent survival (Supplementary Fig. S1). Of note, the yes-associated protein (YAP) and its paralog (TAZ), which have been reported to regulate similar genes as does MRTF-A, such as CYR61, and CTGF (25), did not show significant differences in patient survival based on high versus low expression levels in cutaneous melanoma (Supplementary Fig. S1).

To evaluate MRTF activity, we started by detecting stress fiber formation in the two melanoma cell lines. Stress fiber formation leads to G-actin depletion and MRTF translocation into the nucleus (8). SK-Mel-147 cells exhibit strong F-actin staining and stress fiber formation in the absence of serum stimulation (Fig. 2A). SK-Mel-19 showed reduced F-actin staining and very little stress fiber formation. Consistent with the actin-based mechanism governing MRTF localization and activity, SK-Mel-147 also exhibited spontaneously nuclear MRTF-A under these conditions (Fig. 2A). To show that MRTF-A localization was dependent on F-actin dynamics, we stimulated SK-Mel-19 cells with 20% FBS to induce stress fiber formation which induced MRTF-A nuclear localization (Supplementary Fig. S2). We also treated SK-Mel-147 cells with latrunculin B which inhibits F-actin formation and MRTF-A relocated to the cytosol (Supplementary Fig. S2). MRTF activity is also regulated by the relative levels of G-actin (26). To assess MRTF activity, we determined the mRNA expression of three known Rho/MRTF target genes, myosin regulatory light chain 9 (MYL9), cysteine-rich angiogenic inducer 61 (CYR61), and connective tissue growth factor (CTGF). Consistent with the predominantly nuclear MRTF-A, SK-Mel-147 shows dramatically increased expression of these three Rho/MRTF-regulated genes compared with SK-Mel-19 (Fig. 2B). Along with being MRTF target genes, MYL9, CYR61, and CTGF play important roles in cellular migration and metastasis (10, 27). SK-Mel-147 cells clearly have increased MRTF activation as compared with SK-Mel-19 cells, suggesting that this may contribute to the metastatic potential of these cells.

We next assessed the potential for using a novel small-molecule inhibitor of the Rho/MRTF pathway to block in vitro and in vivo models of melanoma metastasis. To evaluate transcriptional activity of MRTF, we performed luciferase assays with a modified serum response element promoter (SRE.L) that is selective for MRTF/SRF but does not respond to TCF/SRF complexes (11, 28). CCG-203971 inhibits RhoC-induced SRE-Luciferase expression in SK-Mel-147 cells with an IC₅₀ of approximately 6 μmol/L. CCG-203971 does not cause cellular toxicity up to 100 μmol/L using the WST-1 viability assay in a 24-hour time period (Fig. 3A). In addition, CCG-203971 reduces the nuclear localization of MRTF-A in SK-Mel-147 cells in a concentration-dependent manner (Fig. 3B). By inhibiting the Rho/MRTF pathway, treating SK-Mel-147 cells with CCG-203971 also reduces the expression of endogenous MYL9, CYR61, and CTGF (Fig. 3C).

To evaluate the effect of CCG-203971 on cellular growth in the aggressive melanoma cell line (SK-Mel-147) compared with the less aggressive cell line (SK-Mel-19), we determined the effect of Rho/MRTF pathway inhibition on cellular proliferation. While it had no effect on cell numbers at 24 hours, CCG-203971 caused a significant reduction in the number of SK-Mel-147 cells at 72 hours. CCG-203971 is more potent in reducing cell numbers in our cell proliferation assay in the SK-Mel-147 cell line compared with SK-Mel-19 (Fig. 4A). To understand whether CCG-203971 affected cell proliferation or induced apoptosis, we conducted cell-cycle analysis using flow cytometry in SK-Mel-147 cells following 10 μmol/L CCG-203971 treatment for 72 hours. CCG-203971 caused G₀–G₁ arrest as evidenced by a significant increase in G₀–G₁ phase and a decrease in S-phase compared with DMSO control treatment (Fig. 4B and C). CCG-1423, the precursor compound to CCG-203971, has been shown to downregulate genes that are involved in cell cycle, specifically G₁–S transition, including targets of the transcription factor E2F (GEO Accession number: GSE30188; ref. 29). There was no significant increase in sub-G₁ in CCG-203971–treated SK-Mel-147 cells (Fig. 4B), suggesting that CCG-203971 does not induce apoptosis. Consistent with the effects of CCG-203971 to induce G₀–G₁ cell-cycle arrest, it also inhibited the clonogenicity of the SK-Mel-147 cells. Following a 72-hour treatment with DMSO or CCG-203971, equal numbers of viable cells were plated for colony formation in the absence of our Rho/MRTF pathway inhibitor. The CCG-203971–treated cells formed roughly half as many colonies compared with control treatment and the colonies which formed were approximately 50% reduced in size (Fig. 4D).

To determine whether inhibition of MRTF localization and activity affects phenotypes of the aggressive melanoma line, we investigated the influence of CCG-203971 at the cellular level. Using Transwell migration and label-free Matrigel invasion assays, we found that 10 μmol/L CCG-203971 inhibited cellular migration (Fig. 5A) and invasion (Fig. 5B) of the RhoC-overexpressing melanoma line SK-Mel-147. This migratory phenotype is consistent with an action on MRTF-A as siRNA knockdown of MRTF-A expression (Fig. 5C) inhibits SK-Mel-147 Transwell migration (Fig. 5D).

To compare the metastatic potential of the high versus low RhoC-expressing cell lines, we injected mice with GFP-expressing SK-Mel-19 and SK-Mel-147 cells. The RhoC-overexpressing SK-Mel-147 cells readily formed lung metastases following intravenous injection (Fig. 6A). The SK-Mel-147–injected mice not only displayed more metastatic nodules (∼3×) than did SK-Mel-19–injected mice, but the tumors were much larger (∼5× greater area) than those in the SK-Mel-19–injected mice (Fig. 6B). Thus, the total tumor burden was greater by approximately 15 times. To determine the potential for MRTF pathway inhibitors as metastasis-targeting therapeutics, we treated mice with CCG-203971 (100 mg/kg twice daily) or vehicle control for 18 days following tail vein injection of GFP-expressing SK-Mel-147 cells. The CCG-203971–treated group had about half as many tumors which were also about 1/4 the size. This resulted in a final tumor burden around 7 times lower than in the vehicle-treated group (Fig. 6B).

Mutations in BRAF and NRAS, leading to constitutive activation of the MAPK pathway, are found in nearly approximately 60%–70% of primary human cutaneous melanomas (30). However, these driver mutations do not appear to govern the propensity for metastasis; BRAF vs. NRAS mutation status does not predict patient overall survival (31). In addition, current targeted therapeutics, such as the BRAF^V600E inhibitor vemurafenib, demonstrate profound initial results in a majority of BRAF^V600E-expressing melanoma tumors (32). However, these responses are limited in duration and resistance typically develops within months. Consequently, it is important to better understand factors that drive aggressiveness and metastasis of human melanomas. On the basis of the role of RhoC and MRTF-regulated gene transcription in genetic studies in mouse B16F2 melanoma and breast cancer models (2, 4, 11) and in the TCGA dataset (24), we wanted to explore the role of those mechanisms in human metastatic melanoma utilizing our novel Rho/MRTF pathway inhibitor, CCG-203971.

While mutations in Rho proteins are rare, review of cancer genomic information through cBioPortal, (33) showed that 31% of cutaneous melanomas had amplification or mutation of RhoA/C, MKL1/2 (gene names for MRTF-A/B) or upstream activators (GNA12, GNA13, GNAQ, GNA11, or ARHGEF11). Furthermore, an outlier analysis of melanoma in Oncomine, (34) showed that 5%–20% of melanomas across several studies had marked upregulation of RhoC at the transcriptional level. Mechanisms driving this are not known but could include regulation by the ETS-1 transcription factor (35, 36). Loss of E-cadherin in melanoma leads to ETS-1 activation and enhanced RhoC expression (37). RhoC induces c-Jun expression (likely through MRTF/SRF mechanisms) which contributes to cell survival (37). This could partially explain the increased clonogenicity observed in SK-Mel-147, the RhoC-overexpressing melanoma cell line, as well as the increased lung colonization.

In light of the multiple genes and pathways that can lead to activation of Rho and MRTF-regulated transcription (38, 39), inhibition at a downstream step in the pathway would be beneficial to disrupt signals from the numerous activation mechanisms (40, 37). Interestingly, CCG-203971 treatment appears to influence SK-Mel-147 cellular morphology in such a manner that SK-Mel-147 cells more closely resemble SK-Mel-19 cellular morphology. Further studies will need to be conducted to better understand this observation. Our compound, CCG-203971 prevents the nuclear accumulation of MRTF but the molecular mechanism remains uncertain. Two proposed models include direct binding to MRTF (41) or modulation of the intranuclear actin regulation by MICAL2 (42). Genetic validation of our pharmacologic approach to suppress metastasis has already been shown for both melanoma and breast cancer using MRTF-A/B double knockdown or silencing SRF (11). We also confirm that in SK-Mel-147 cells, siRNA knockdown of MRTF-A alone is sufficient to markedly reduce migration in vitro. The majority of recent literature strongly implicates MRTF and SRF as being critical factors in the regulation cell motility (43–45). Despite the commonality between MRTF/SRF- and YAP/TEAD–regulated target genes, as well as cell type–specific mutual coactivation of gene expression (46), only MRTF-A expression levels were correlated with patient survival (Fig. 3; Supplementary Fig. S2).

The potential for an antimetastatic small-molecule therapeutic that targets this gene transcription mechanism is illustrated by our result showing marked suppression of SK-Mel-147 lung metastasis by in vivo treatment with CCG-203971. While it has been argued that antimetastatic therapies are not useful or are difficult to test clinically, there is increasing recognition that this approach has value (47). CCG-203971 also effectively causes G₁ cell-cycle arrest which is a phenotype shared with current MAPK pathway–targeted therapeutics (48, 49). As G₁-arrested melanoma cells have increased sensitivity to MAPK pathway inhibitors (49) this suggests the potential benefit of combination treatment with Rho/MRTF pathway inhibitors and MAPK pathway inhibitors. Continued efforts to optimize MRTF transcription inhibitors like CCG-203971 will be needed. In particular, enhancing potency along with reduction of acute toxicity (12, 13) is required for chronic treatments that would be required for antimetastasis therapies. Furthermore, improved bioavailability and pharmacokinetics would facilitate both preclinical and potentially clinical studies.

In addition to melanoma, spontaneous nuclear localization of MRTF-A has been reported in MDA-MB-231 breast cancer cells (11). The MDA-MB-231 cells are often used as a model of aggressive breast cancer cells. Intriguingly, they also harbor a Ras mutation (KRAS^G13D; ref. 50) similar to the NRAS^Q61N in SK-Mel-147. Additional work is needed to assess the generality of these observations.

In this report, we provide evidence that Rho-regulated, MRTF-mediated gene transcription signaling plays an important role in migration, invasion, and metastasis of human cutaneous melanoma. Thus, expression levels of RhoC and MRTF target genes may provide useful biomarkers to identify cancers with propensity for metastasis as well as for targeted therapies by this approach. Small-molecule inhibitors of this pathway, such as enhanced CCG-203971 analogues, represent an exciting class of pharmacologic probes and potential future therapeutics.

A.J. Haak has ownership interest (including patents) in a Patent Application (20160145251). S.D. Larsen has ownership interest (including patents) in a United States Patent Application (20160145251). R.R. Neubig has ownership interest (including patents) in a United States Patent Application 20160145251. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A.J. Haak, K.M. Appleton, S.D. Larsen, M.E. Verhaegen, R.R. Neubig, E.M. Lisabeth

Development of methodology: A.J. Haak, K.M. Appleton, S.M. Wade, C.E. Rockwell

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.J. Haak, K.M. Appleton, S. Misek, Y. Ji, S.M. Wade, C.E. Rockwell, M. Airik, M.A. Krook, E.R. Lawlor

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.J. Haak, K.M. Appleton, S. Misek, Y. Ji, S.M. Wade, C.E. Rockwell, M. Airik, M.A. Krook, S.D. Larsen, E.R. Lawlor, R.R. Neubig

Writing, review, and/or revision of the manuscript: A.J. Haak, K.M. Appleton, S.M. Wade, J.L. Bell, C.E. Rockwell, M. Airik, M.A. Krook, S.D. Larsen, M.E. Verhaegen, E.R. Lawlor, R.R. Neubig, E.M. Lisabeth

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K.M. Appleton, M.A. Krook, M.E. Verhaegen

Study supervision: E.R. Lawlor, R.R. Neubig

Other (chemical synthesis): J.L. Bell

We would like to thank Riya Malhotra, Alexandra Turley, and Joseph Zagorski for technical assistance with the flow cytometry experiments. We would also like to thank Raelene Van Noord for her assistance with the mouse metastasis study and Nadia Ayala-Lopez, Humphrey Petersen-Jones, and Dr. Stephanie Watts for assistance with the immunohistochemical analysis of melanoma metastasis.

This work was supported in part by a Pharmacological Sciences Training Program grant GM007767 from NIGMS (to A.J. Haak), NCIT32CA009676 (to M.A. Krook), and SPOREU54CA168512 (to E.R. Lawlor).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.