ERK/MAPK Signaling Drives Overexpression of the Rac-GEF, PREX1

Author Manuscript Published OnlineFirst on July 14, 2016; DOI: 10.1158/1541-7786.MCR-16-0184 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

1 ERK/MAPK Signaling Drives Overexpression of the Rac-GEF, PREX1,

2 in BRAF- and NRAS-mutant Melanoma

4 Meagan B. Ryan1, Alexander J. Finn2, Katherine H. Pedone4, Nancy E. Thomas2,

5 Channing J. Der1,4*, Adrienne D. Cox1,3,4*

7 Departments of 1Pharmacology, 2Dermatology and 3Radiation Oncology, and

8 4Lineberger Comprehensive Cancer Center, The University of North Carolina at Chapel

9 Hill, Chapel Hill, North Carolina

11 Running Title: ERK Regulates PREX1-RAC1 in Human Melanoma

12 Keywords: PREX1, RhoGEF, Rac-GEF, RAC1, ERK, melanoma

13 Financial Support: CJD and ADC are supported by grants from the NIH (CA42978,

14 CA179193, CA175747, CA199235) and the Pancreatic Cancer Action Network-AACR.

15 CJD is also supported by the Lustgarten Pancreatic Cancer Foundation. MBR has

16 fellowship support from the NIH (Cancer Cell Biology Training Grant T32 CA071341 and

17 an NRSA predoctoral fellowship, F31 CA192829).

18 *Correspondence:

19 Channing J. Der, Lineberger Comprehensive Cancer Center, University of North

20 Carolina at Chapel Hill, Chapel Hill, NC 27599-7295, USA. Phone: 919-966-5634; Fax:

21 919-966-9673, E-mail: [email protected]

22 Adrienne D. Cox, Lineberger Comprehensive Cancer Center, University of North

23 Carolina at Chapel Hill, Chapel Hill, NC 27599-7295, USA. Phone: 919-966-7712; Fax:

Downloaded from mcr.aacrjournals.org on October 5, 2021. © 2016 American Association for Cancer Research. Author Manuscript Published OnlineFirst on July 14, 2016; DOI: 10.1158/1541-7786.MCR-16-0184 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

ERK Regulates PREX1-RAC1 in Human Melanoma

24 919-966-9673, E-mail: [email protected]

25 Conflicts of Interest: The authors declare no conflict of interest.

26 Word Count: (5,248)

27 Number of Figures: 6 (plus 5 supplemental)

28 Number of Tables: 0

ERK Regulates PREX1-RAC1 in Human Melanoma

30 ABSTRACT

31 Recently we identified that PREX1 overexpression is critical for metastatic but not

32 tumorigenic growth in a mouse model of NRAS-driven melanoma. In addition, a PREX1

33 gene signature correlated with and was dependent on ERK mitogen-activated protein

34 kinase (MAPK) activation in human melanoma cell lines. In the current study, the

35 underlying mechanism of PREX1 overexpression in human melanoma was assessed.

36 PREX1 protein levels were increased in melanoma tumor tissues and cell lines

37 compared with benign nevi and normal melanocytes, respectively. Suppression of

38 PREX1 by siRNA impaired invasion but not proliferation in vitro. PREX1-dependent

39 invasion was attributable to PREX1-mediated activation of the small GTPase RAC1 but

40 not the related small GTPase CDC42. Pharmacologic inhibition of ERK signaling

41 reduced PREX1 gene transcription and additionally regulated PREX1 protein stability.

42 This ERK-dependent upregulation of PREX1 in melanoma, due to both increased gene

43 transcription and protein stability, contrasts with the mechanisms identified in breast and

44 prostate cancers, where PREX1 overexpression was driven by gene amplification and

45 HDAC-mediated gene transcription, respectively. Thus, although PREX1 expression is

46 aberrantly upregulated and regulates RAC1 activity and invasion in these three different

47 tumor types, the mechanisms of its upregulation are distinct and context-dependent.

48 Implications:

49 This study identifies a ERK-dependent mechanism that drives PREX1 upregulation and

50 subsequent RAC1-dependent invasion in BRAF- and NRAS-mutant melanoma.

ERK Regulates PREX1-RAC1 in Human Melanoma

51 INTRODUCTION

52 Driver roles in cancer have been identified for several members of the Dbl family

53 of Rho guanine nucleotide exchange factors (RhoGEFs), most prominently ECT2,

54 TIAM1, VAV1/2/3 and PREX1/2 (1,2). Increased expression and activation of these

55 RhoGEFs result in enhanced activity of their Rho family small GTPase substrates in a

56 context-dependent manner. For example, we recently identified overexpression of

57 ECT2 protein in ovarian cancer, mediated by gene amplification, that resulted in

58 activation of RHOA in the cytosol and RAC1 in the nucleus (3). The critical importance

59 of RAC1 in cancer cell migration and invasion (4) has further focused attention on the

60 mechanisms regulating activators of RAC1, such as the Dbl family of RhoGEFs.

61 The highly related Dbl RhoGEFs PREX1 and PREX2 (56% overall amino acid

62 identity), which are GEFs for RAC1 and other Rho family small GTPases such as

63 CDC42 (5), have been implicated as cancer drivers in several tumor types. The first

64 cancer-driving role for PREX1 was described in prostate cancer (6), where limited

65 analyses of tumor tissue revealed elevated levels of PREX1 protein. PREX1 was also

66 elevated in metastatic but not primary prostate tumor cell lines. Suppression of PREX1

67 by RNA interference in PC-3 human prostate cancer cells decreased the levels of

68 activated RAC and impaired tumor cell migration and invasion in vitro. Conversely,

69 ectopic expression of PREX1 stimulated RAC activation, and promoted metastatic but

70 not primary tumor growth of CWR22Rv1 prostate tumor cells. A follow-up study

71 identified a histone deacetylase (HDAC)-mediated increase in PREX1 gene

72 transcription as a basis for the increased levels of PREX1 protein in prostate cancer (7).

ERK Regulates PREX1-RAC1 in Human Melanoma

73 PREX1 overexpression was also identified in estrogen receptor-positive luminal

74 and HER2-positive breast cancers (8-10). PREX1 protein was detected in ~60% of

75 breast tumors but not in normal breast tissue. PREX1 is located in a chromosomal

76 region frequently amplified in breast cancers and PREX1 gene amplification was

77 detected in breast cancer cell lines, supporting gene amplification as a mechanism for

78 PREX1 protein overexpression in these tumor types. Silencing of PREX1 expression by

79 RNA interference reduced HER2-stimulated activation of RAC1, and impaired tumor cell

80 motility and invasion in vitro and tumorigenic growth in vivo (8-10).

81 A role in cancer for the related RhoGEF PREX2 has also been identified, but not

82 by overexpression. Instead, missense mutations in PREX2 have been identified in 25%

83 of malignant melanomas (11). Although no clear mutational hotspots have been seen in

84 melanomas, experimental studies support a gain-of-function consequence of these

85 mutations (11,12). PREX2 missense mutations have also been found in 38% of

86 cutaneous squamous cell carcinomas (13) and 17% of stomach adenocarcinomas (14).

87 To date, a similar level of activating missense mutations in PREX1 in cancer has not

88 been reported. However, the occurrence of activating missense mutations (e.g., P29S)

89 in the PREX1 substrate, RAC1, in ~11% of melanomas (15,16) is also consistent with a

90 driver role for overexpressed PREX1 in this disease.

91 Our previous studies revealed overexpression of PREX1 protein in melanoma

92 cell lines and tumor tissue (17). Further, in a mouse model of melanoma, we

93 determined that Prex1-deficient mice were impaired in forming tumor metastases but

94 not primary tumors. Here, extending our mouse model studies, we demonstrate that

ERK Regulates PREX1-RAC1 in Human Melanoma

95 PREX1 protein is increased in human melanoma tumor tissue and that PREX1 is

96 required for human melanoma cell invasion but not proliferation.

97 In a separate earlier study, we also identified PREX1 as a gene upregulated by

98 the ERK mitogen-activated protein kinase (MAPK) in melanoma (18). The RAF-MEK-

99 ERK protein kinase cascade is aberrantly activated in up to 80% of melanomas through

100 BRAF or NRAS mutation, and serves as a critical therapeutic target in this disease (19-

101 22). These observations supported the possibility that PREX1 protein overexpression in

102 melanoma is driven by ERK activation. Additionally, since PREX1 has a demonstrated

103 driver role in other cancers, PREX1 overexpression may be a key driver of ERK-

104 dependent melanoma growth. In the present study, we determined that PREX1 protein

105 overexpression is blocked by pharmacologic inhibitors of RAF-MEK-ERK signaling, and

106 that ERK regulates not only PREX1 gene transcription but also PREX1 protein stability.

107 Thus, there are significant cancer type-distinct mechanisms that drive PREX1

108 overexpression in cancer.

109

110 MATERIALS AND METHODS 111 112 Human melanoma tissue and immunohistochemistry (IHC)

113 Following institutional review board approval, primary and metastatic melanomas

114 were retrieved from a series of patients treated at UNC Healthcare.

115 Immunohistochemical staining was performed in the UNC Department of Dermatology

116 Dermatopathology Laboratory as we have recently described (23). Briefly, freshly cut 4-

117 µm thick sections of formalin-fixed and paraffin-embedded melanoma tissue blocks

118 were stained using the fully automated Leica Bond III system. Sections were pretreated

ERK Regulates PREX1-RAC1 in Human Melanoma

119 using an onboard heat-induced epitope retrieval in EDTA buffer. Following incubation

120 with PREX1 antibody (6F12; provided by Marcus Thelen, IRB, Switzerland),

121 chromogenic detection was performed using the Leica Refined Red polymer detection

122 system (Leica Microsystems). Some sections were also counterstained with

123 hematoxylin and eosin (H&E). PREX1 antibody staining intensity was scored in a

124 blinded manner by a pathologist (AJ Finn) as high, medium, low or none.

125 Tissue culture

126 Cutaneous melanoma, breast cancer and prostate cancer cell lines were

127 obtained from ATCC. Cells were maintained in Dulbecco’s Modified Eagle Medium

128 (DMEM) supplemented with 10% fetal bovine serum (FBS), and were not cultured for

129 longer than 6 months after receipt from cell banks.

130 siRNA transfection, proliferation, and invasion assays

131 For siRNA knockdown, A375, WM2664, SK-MEL-119 and Mel224 cells were

132 plated in 6-well plates. Cells were transfected with 10 nM siRNA against PREX1

133 (Thermo Fisher s33364, s33365, s33366; PREX1 #1, #2, #3, respectively), RAC1

134 (Thermo Fisher s11711, s11712, s11713; RAC1 #1, #2, #3, respectively) or mismatch

135 control (Dharmacon #D-001210-05), using Lipofectamine RNAimax (Life Technologies).

136 Cells were serum-starved overnight for 18 h and then seeded for invasion assays after

137 48 h of siRNA knockdown. For the proliferation assay, cells were seeded at 2-3 x 103

138 cells/well in 96-well plates and allowed to grow for 72 h before incubation for 3 h in 3-

139 (4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT). MTT was solubilized

140 in DMSO and the absorbance was read at A570. For the Boyden chamber invasion

141 assay, 1-3 x 104 cells were seeded into the upper chamber of Matrigel-coated invasion

ERK Regulates PREX1-RAC1 in Human Melanoma

142 chambers in duplicate (Corning BioCoat) and allowed to invade towards 20% FBS in

143 DMEM for 24 h. Invasion chambers were fixed and stained using a Diff-Quik staining kit

144 (GE). Invasion chambers were imaged using a 10x objective lens on a Nikon TS100

145 microscope at 5 fields per insert, and images were analyzed to calculate invaded cells

146 per field using ImageJ software. For the collagen spheroid invasion assay, we slightly

147 modified a published protocol (24,25). Briefly, 5-10 x 103 cells were seeded in ultra-low

148 attachment round-bottomed 96-well plates (Corning) for 96 h, a sufficient time for the

149 cells to organize into spheroids. Spheroids were then transferred to 48-well plates and

150 embedded in collagen (1 mg/ml rat-tail collagen, BD). Spheroids were imaged at 0 h

151 and after 72 h of invasion using a 5x objective on a Nikon TS100 microscope. Total

152 spheroid area was calculated as fold-change in area of 72 h outgrowth versus 0 h

153 spheroid area, using ImageJ.

154 Pulldown assay to detect GTPase activity

155 Levels of active, GTP-bound RAC1 and CDC42 were assessed by an affinity

156 pulldown assay as we described previously (26). Briefly, after 48 h of siRNA-mediated

157 PREX1 knockdown, whole cell lysates were exposed to GST-PAK-PBD, which contains

158 the binding domain of the shared RAC1/CDC42 effector PAK1. After resolving pulldown

159 samples on 15% SDS-PAGE gels and western blotting for RAC1 (clone 23A9, BD) and

160 CDC42 (BD), levels of each GTP-bound GTPase were normalized to both total protein

161 and the vinculin loading control (Sigma) by densitometry analysis performed in ImageJ.

162 Drug treatment and western blotting

163 BRAF-mutant A375 and WM2664 cells were treated with BRAF inhibitor

164 vemurafenib (Selleckchem) or ERK inhibitor SCH772984 (Merck, kindly provided by

ERK Regulates PREX1-RAC1 in Human Melanoma

165 Ahmed Samatar), and NRAS-mutant SK-MEL-119 and Mel224 cells were treated with

166 MEK inhibitor trametinib (Selleckchem) or SCH772984 for 24 and 48 h before samples

167 were collected in RIPA lysis buffer. For PREX1 protein stability experiments, cells were

168 co-treated with cycloheximide (50 µg/ml) and SCH772984 for a 24 h timecourse before

169 samples were collected in RIPA. Whole cell lysates were resolved on 10% SDS-PAGE

170 gels and western blotting was performed using antibodies to phospho-ERK1/2

171 (Thr202/Tyr204), total ERK1/2, phospho-RSK (Ser308), phospho-RSK

172 (Thr359/Ser363), total RSK1/2/3, and c-myc (MYC) (Cell Signaling); β-actin and vinculin

173 (Sigma), and PREX1 (6F12) (27). IRDye800-conjugated anti-mouse and anti-rabbit

174 secondary antibodies were from Rockland Immunochemicals.

175 Quantitative PCR

176 Total RNA was isolated using an RNeasy kit (Qiagen) and reverse transcription

177 was performed using the High Capacity RNA-to-cDNA kit (Thermo Fisher). Real time

178 quantitative Taqman PCR was performed on the QuantStudio 6 Flex (Thermo Fisher)

179 with FAM/MGB labeled probes against PREX1 (Hs00368207_m1, Hs_001031512,

180 Thermo Fisher) and endogenous control VIC/TAMRA labeled β-actin (Thermo Fisher).

181 Statistical analysis

182 Data were analyzed using GraphPad Prism 6 software and statistical analyses

183 were performed as indicated in the Figure Legends.

184

185 RESULTS

186 PREX1 overexpression is correlated with elevated ERK activation

ERK Regulates PREX1-RAC1 in Human Melanoma

187 We recently identified overexpression of PREX1 protein in melanomas,

188 determined that Prex1 deficiency impaired mouse melanoblast migration in vivo, and

189 demonstrated that Prex1 expression is required for metastasis in an Nras-mutant

190 genetically engineered mouse model of cutaneous melanoma (17). Since our previous

191 gene array analyses identified PREX1 as an ERK activation-dependent gene (18), here

192 we assessed a relationship among BRAF and NRAS mutation status, ERK activation

193 and PREX1 protein overexpression in human melanoma. We first investigated the

194 expression of PREX1 in a panel of human melanoma cell lines that did or did not harbor

195 BRAF or NRAS mutations. The majority of BRAF- (3 of 4) or NRAS- (3 of 4) mutant cell

196 lines exhibited substantially higher PREX1 protein expression when compared with

197 normal melanocytes or with BRAF/NRAS wild type cell lines (Figure 1A). Generally, the

198 level of activated, phosphorylated ERK (pERK) correlated with the level of PREX1. In

199 our analyses, normal melanocytes may exhibit low pERK levels (18), or they may also

200 display low PREX1 protein levels even in the presence of high pERK levels, as shown

201 here.

202 We next utilized immunohistochemistry (IHC) to evaluate PREX1 protein

203 expression and pERK levels in melanoma patient tissues. We first compared PREX1

204 expression in benign melanocytic nevi (n=35) and human melanoma tumors (n=33)

205 (Figure 1B). A range of expression was seen in nevi, with ~75% expressing low-to-

206 medium levels of PREX1. Since BRAF and NRAS mutations are found in a high

207 percentage of nevi (28,29), it is not surprising to find PREX1 in nevi as well as in

208 melanoma tissue. However, high level PREX1 expression was detected only in

209 melanomas (~10%, Figure 1B).

ERK Regulates PREX1-RAC1 in Human Melanoma

210 ERK can phosphorylate numerous substrates present in both the nucleus and

211 the cytoplasm (30,31), few of which have been firmly linked to specific outcomes of

212 ERK-mediated signaling. Melanoma responses to pharmacological inhibitors of the

213 RAF-MEK-ERK pathway (e.g., clinically, to BRAF inhibition (32) and preclinically, to

214 inhibitors of ERK dimerization (33,34)) correlated with suppression of cytoplasmic

215 pERK. We therefore evaluated the distribution of pERK in our human melanoma

216 tissues, and found that levels of pERK were correlated with those of PREX1 both in the

217 nucleus (Figure 1C) and in the cytoplasm (Figure 1D). Both nuclear and cytoplasmic

218 ERK activity may contribute to increased expression of PREX1.

219

220 PREX1 regulates invasion in a complex manner in both BRAF- and NRAS-mutant

221 melanoma cell lines

222 An unexpected observation in our studies of PREX1 function in a mouse model

223 of melanoma was that Prex1 deficiency greatly impaired metastatic but not tumorigenic

224 growth. This result contrasts with studies evaluating the role of PREX1 overexpression

225 in human breast cancer cells, where stable shRNA-mediated suppression of PREX1

226 reduced their tumorigenic growth (8-10). We therefore compared the effect of PREX1

227 suppression in human melanoma cell lines on both proliferation and invasion in vitro.

228 To evaluate the role of PREX1 overexpression in cell growth, we first used three

229 independent siRNAs to knock down PREX1 in two BRAF-mutant (A375 and WM2664)

230 and two NRAS-mutant cell lines (SK-MEL-119 and Mel224) (Figure 2A, upper panels).

231 We found that transient (72 h) suppression of PREX1 did not significantly reduce their

232 proliferation in vitro (Figure 2A, lower panels), consistent with the lack of effect of Prex1

ERK Regulates PREX1-RAC1 in Human Melanoma

233 deficiency on the growth of primary melanomas in mice. Next, we evaluated the role of

234 PREX1 in invasion by analysis of invasion through Matrigel towards serum as a

235 chemoattractant. We observed a surprisingly heterogeneous response to PREX1

236 knockdown that was independent of BRAF or NRAS mutational status. For example,

237 the NRAS-mutant line SK-MEL-119 exhibited a 60-70% decrease in invasion upon

238 knockdown of PREX1 (p<0.0001, Figure 2B) whereas the BRAF-mutant cell line A375

239 conversely exhibited a ~2-fold increase (p<0.001). In contrast, the already very low

240 degree of directed invasion of the BRAF-mutant line WM2664 was unaffected by

241 PREX1 knockdown. Similarly, the invasive NRAS-mutant cell line Mel224 was largely

242 unaffected, despite efficient knockdown of PREX1. These data demonstrate that

243 PREX1 plays a complex and variable role in directed invasion towards an attractant.

244 Next, we investigated the role of PREX1 in a three-dimensional spheroid

245 formation and collagen invasion assay, which mimics the in vivo tumor environment of

246 human skin (35). Figure 2C illustrates both spheroid formation and the subsequent

247 invasion of cells from the spheroid into the surrounding collagen matrix. In three of the

248 four cell lines, knockdown of PREX1 impaired spheroid invasion into collagen, either

249 trending (SK-MEL-119) or significantly so (Mel224, WM2664). In contrast, A375

250 spheroids were defective in formation and did not invade the surrounding collagen

251 matrix. The nearly doubled total spheroid area of PREX1-knockdown A375 cells

252 compared to mismatch control cells observed after 4 days in culture was caused by a

253 flattening of the three-dimensional spheroid structure and not by increased invasion or

254 by increased proliferation; no change in proliferation occurred upon loss of PREX1 (Fig.

255 2A).

ERK Regulates PREX1-RAC1 in Human Melanoma

256 Collectively, our results suggest a complex and context-dependent role for

257 PREX1 in driving both directed invasion and three-dimensional spheroid collagen

258 outgrowth of human melanomas, and one that is not dependent on BRAF or NRAS

259 mutational status.

260 PREX1 regulates active, GTP-bound RAC1 but not CDC42 in melanoma cells

261 Although PREX1 is considered a RAC-selective GEF (36), PREX1 is also active

262 on CDC42 (27). We therefore investigated which Rho family small GTPases are

263 activated downstream of PREX1 in human melanoma cells. We found that knockdown

264 of PREX1 decreased the levels of activated RAC1, as measured by RAC1-GTP

265 pulldown, in both BRAF-mutant A375 and NRAS-mutant SK-MEL-119 cells (Figure 3).

266 Despite the continued presence of other RacGEFs capable of inducing nucleotide

267 exchange on RAC1, even incomplete loss of PREX1 was sufficient to cause a

268 substantial decrease in RAC1-GTP (Figure 3). This effect was selective for RAC1, as

269 the levels of activated CDC42 did not decrease (Figure 3). These results support a role

270 for PREX1 in regulating RAC1 activity and subsequent RAC1-driven invasive behavior

271 of melanomas.

272 We next asked if the loss of RAC1 was sufficient to phenocopy the impairment of

273 invasion that we observed upon loss of PREX1. We found that knockdown of either

274 PREX1 or RAC1 with three independent siRNAs for each (Figure 4A) was sufficient to

275 substantially impair spheroid formation of A375 cells, as demonstrated by an increase in

276 the flattened spheroid area (Figure 4B,C). The degree of impairment upon knockdown

277 of RAC1 was highly significant (p<0.0001, Figure 4C) and comparable to the degree of

278 impairment observed upon knockdown of PREX1 (p<0.0001, Figure 4C). Next, we

ERK Regulates PREX1-RAC1 in Human Melanoma

279 observed that loss of RAC1 (Figure 4D) was sufficient to prevent the vast majority of

280 SK-MEL-119 cell invasion in the Boyden chamber assay (Figure 4E). This decrease in

281 invasion was similar to the decrease seen upon PREX1 knockdown (p<0.0001, Figure

282 4F). The ability of RAC1 to phenocopy PREX1 in impairing both spheroid formation and

283 directed invasion supports the idea that RAC1 is the most critical Rho family small

284 GTPase downstream of PREX1 in regulating invasive melanoma behavior. Of note,

285 other Rho family small GTPases such as RND3 have also been shown to be regulated

286 by the RAF-MEK-ERK pathway and to contribute to melanoma invasion and spheroid

287 outgrowth (37,38). However, as a Rho-like rather than a Rac-like GTPase, RND3 is

288 unlikely to be a target of PREX1 in this context (39).

289 PREX1 protein levels are positively regulated by ERK activity in melanoma

290 Our evaluation of human melanoma cell lines and tumor tissue found a

291 correlation between phosphorylated ERK and PREX1 protein overexpression. To

292 directly address whether ERK activation is required for PREX1 overexpression, we

293 evaluated whether pharmacologic inhibition of RAF-MEK-ERK signaling would reduce

294 PREX1 protein levels in BRAF- and NRAS-mutant melanoma. We first treated NRAS-

295 mutant SK-MEL-119 cells with increasing concentrations of the MEK inhibitor trametinib.

296 The more effective the inhibition of MEK, as measured by decreasing levels of

297 phosphorylated and activated ERK (pERK) and total MYC (an ERK substrate; ERK

298 phosphorylation blocks degradation), the greater the decrease in PREX1 protein

299 (Figures 5A,B), suggesting a direct correlation between ERK activity and PREX1 protein

300 levels. We next wanted to determine whether this dose-dependent decrease in PREX1

301 protein would also occur when the ERK MAPK cascade was inhibited at different nodes,

ERK Regulates PREX1-RAC1 in Human Melanoma

302 and whether such an effect is time-dependent. We therefore treated SK-MEL-119 cells

303 with two different concentrations (1 x EC50 and 5 x EC50 for growth) of either trametinib

304 or the ERK inhibitor SCH772984, for either 24 or 48 h (Figures 5C,D). PREX1 protein

305 levels tracked closely with the level of pERK at each concentration of inhibitor, and this

306 was sustained for 48 h. Next, we investigated whether PREX1 protein was similarly

307 regulated downstream of the ERK-MAPK cascade in BRAF-mutant melanoma cells. We

308 treated A375 cells similarly but with the BRAF inhibitor vemurafenib or SCH772984 for

309 24 or 48 h (Figures 5E,F). Similarly to SK-MEL-119 cells, levels of PREX1 in A375 cells

310 tracked closely with the levels of pERK and demonstrated a time-dependent effect. In

311 both SK-MEL-119 and A375 cells, phosphorylation of the ERK substrate RSK (pRSK)

312 and total MYC served as effective markers to demonstrate inhibition of ERK, as we

313 have observed in other settings (40). These results indicate that ERK MAPK activity is

314 an important contributor to the total amount of PREX1 protein in melanoma cells.

315 PREX1 levels are regulated by ERK both transcriptionally and post-

316 transcriptionally in melanoma

317 To determine if the loss of PREX1 protein upon blockade of the ERK MAPK

318 cascade was due to loss of PREX1 mRNA, we treated two BRAF-mutant and two

319 NRAS-mutant melanoma cell lines with ERK MAPK cascade inhibitors. BRAF-mutant

320 A375 and WM2664 cells were treated for 24 h with vemurafenib or SCH772984 as

321 above. Taqman quantitative PCR analysis revealed that PREX1 mRNA, measured by

322 two independent probes, did not change upon inhibition of BRAF or ERK in A375 cells

323 (Figure 6A) but decreased dose-dependently in WM2664 cells upon inhibition of either

324 BRAF or ERK (Figure 6B). NRAS-mutant cells were treated with trametinib or

ERK Regulates PREX1-RAC1 in Human Melanoma

325 SCH772984 as above. We observed that PREX1 mRNA also decreased upon ERK

326 inhibition in both SK-MEL-119 (Figure 6C) and Mel224 cells (Figure 6D). Additional

327 melanoma lines also exhibited reduced PREX1 mRNA levels when treated with

328 inhibitors of the ERK MAPK cascade, including BRAF-mutant SK-MEL-28 and NRAS-

329 mutant SK-MEL-147 cells (Figures S3A,B). Thus, in the majority of melanoma cell

330 lines, ERK MAPK regulates PREX1 protein levels both transcriptionally and post-

331 transcriptionally.

332 That ERK MAPK activity altered PREX1 protein levels in A375 melanoma cells

333 without changes at the transcriptional level suggested a post-transcriptional mechanism

334 for ERK MAPK-mediated regulation of PREX1 protein in these cells. To investigate this

335 possibility, we treated A375 cells with vehicle or SCH772984 in the presence of the

336 protein synthesis inhibitor cycloheximide at various time points over 24 h (Figures 6E,F).

337 Inhibition of ERK led to greater loss of PREX1 protein in the presence of cycloheximide

338 compared to vehicle-treated cells. Similar results were obtained upon treatment of SK-

339 MEL-119 cells with trametinib in the presence of cycloheximide (Figures S3C,D). These

340 results indicate that ERK can regulate protein stability as well as transcription of PREX1

341 in melanoma cell lines.

342 Finally, we tested the possibility that PREX1 is not only an ERK target but also

343 an ERK activator. Since it has been demonstrated that PREX1 can regulate MEK-ERK

344 signaling through RAC1 in breast cancer (8,41), we also examined whether PREX1 can

345 regulate ERK1/2 phosphorylation in our melanoma lines. We found that knockdown of

346 PREX1 did not alter ERK1/2 phosphorylation in the BRAF-mutant cell lines, A375 and

ERK Regulates PREX1-RAC1 in Human Melanoma

347 WM2664, or the NRAS-mutant cell lines, SK-MEL-119 and Mel224 (Figures S4A,B and

348 S4C,D, respectively).

349 To determine the generality of ERK regulation of PREX1 expression in non-

350 melanoma tumor types, we also tested whether they held true in breast and prostate

351 cancer cell lines. PREX1 has been shown to be overexpressed in these tumor types,

352 but the role of ERK in its expression has not been explored. Unlike our observations in

353 melanoma cells, we observed that inhibition of the ERK MAPK cascade in T47D and

354 MCF7 breast cancer cells did not reduce PREX1 protein (Figures S5A,C) or mRNA

355 (Figures S5B,D). Conversely, in PC-3 prostate cancer cells, inhibition of the ERK MAPK

356 cascade reduced PREX1 mRNA levels (Figure S5F) but had minimal effects on PREX1

357 protein (Figure S5E). These results support distinct mechanisms of regulating PREX1

358 expression in melanoma through ERK1/2 that do not apply in breast or prostate cancer,

359 cancers in which BRAF and RAS mutation frequencies are low. Overall, our results

360 support that both ERK regulation of PREX1 abundance and PREX1 regulation of ERK

361 phosphorylation are context-dependent, and may differ between breast and prostate

362 cancers and cutaneous melanoma.

363

364 DISCUSSION

365 We determined that the ERK MAPK cascade plays an important role in driving

366 PREX1 protein overexpression in both BRAF- and NRAS-mutant melanomas. In

367 contrast, ERK is not the key driver for PREX1 overexpression in prostate (6,7) or in

368 breast carcinomas (8-10,42), where its abundance is associated with HDAC-dependent

369 (7) PREX1 gene transcription and with HDAC- and methylation-dependent PREX1 gene

ERK Regulates PREX1-RAC1 in Human Melanoma

370 transcription (42) and gene amplification (8,10,42), respectively. Thus, there are striking

371 cancer-type differences in mechanisms driving PREX1 overexpression. In support of

372 this idea, PREX1 and PREX2 display distinct expression and mutation patterns in

373 breast cancer, prostate cancer, and melanoma (Figure S1), and PREX1 in particular is

374 differentially amplified in breast cancer, prostate cancer and cutaneous melanoma

375 (Figure S2). Our findings also suggest that loss of PREX1-RAC1 signaling may

376 contribute to the clinical response of patients with BRAF-mutant melanomas to BRAF

377 and MEK inhibitors.

378 The effectiveness of these inhibitors provides compelling evidence that aberrant

379 ERK signaling is a major driver of melanoma growth. Despite this clear driver role, the

380 ERK targets important for melanoma growth remain poorly characterized. The ERK1/2

381 kinases can phosphorylate more than 200 known substrates (30,31) and serve as

382 master regulators of numerous transcription factors (43), both directly and indirectly,

383 through both transcriptional and post-translational mechanisms (40,43-45) We have

384 determined that ERK regulates PREX1 expression levels in part via protein stability, a

385 mechanism also not observed in other cancers. ERK regulation of PREX1 protein

386 stability presents a previously unknown mechanism of maintaining PREX1 protein

387 expression and may explain the basis for the relatively high PREX1 expression in

388 malignant melanomas where the ERK MAPK cascade is upregulated.

389 Finally, the PREX1-related RhoGEF PREX2 is activated by missense mutations

390 in 25% of metastatic melanomas, especially by truncating mutations mutations (11,12),

391 and mutational activation of the PREX1/2 target RAC1 has been observed in ~11% of

392 melanomas (15,16). The rarity of PREX1 truncating mutations and lack of apparent

ERK Regulates PREX1-RAC1 in Human Melanoma

393 hotspots among the few missense mutations argue that this is not a significant

394 mechanism of PREX1 activation in melanoma. Instead, our determination that PREX1 is

395 overexpressed and regulated at multiple levels in response to the ERK MAPK cascade

396 characterizes a third mechanism for driving aberrant RAC1 signaling in melanoma. Our

397 findings that loss of RAC1 phenocopies loss of PREX1 with respect to invasive behavior

398 regardless of BRAF or NRAS mutation status supports the importance of the PREX1-

399 RAC1 relationship as a promoter of melanoma cell invasion. Interestingly,

400 downregulation of the RacGEF TIAM1 by mutant BRAF was shown to enhance invasion

401 of human melanoma cells (46). Although PREX1 was not examined in that particular

402 study, PREX1 has consistently demonstrated a positive role in invasion (6,17,47),

403 whereas TIAM1 can be either a positive or a negative regulator of this process (1,46,48-

404 50). Thus, the relative input from different upstream activators of RAC1 can have a

405 profound influence on melanoma invasion.

406 In summary, we have demonstrated that the ERK MAPK cascade mediates

407 overexpression of PREX1 in melanoma at multiple levels, and by mechanisms that are

408 distinct from those identified previously in other cancer types. Our results contribute to

409 a better understanding of how RacGEFs are modulated in distinct cancer contexts.

410

411 AUTHORS' CONTRIBUTIONS

412 MB Ryan, CJ Der and AD Cox conceived and designed the study. MB Ryan performed

413 all experiments except those shown in Figures 1A and 1B. KH Pedone performed the

414 experiment shown in Figure 1A. AJ Finn performed the experiment shown in Figure 1B.

ERK Regulates PREX1-RAC1 in Human Melanoma

415 NE Thomas provided melanoma tumor samples. MB Ryan, CJ Der and AD Cox wrote

416 the manuscript, and all authors reviewed it.

417

418

ERK Regulates PREX1-RAC1 in Human Melanoma

419 REFERENCES

420 1. Cook DR, Rossman KL, Der CJ. Rho guanine nucleotide exchange factors:

421 regulators of Rho GTPase activity in development and disease. Oncogene

422 2014;33(31):4021-35.

423 2. Vigil D, Cherfils J, Rossman KL, Der CJ. Ras superfamily GEFs and GAPs:

424 validated and tractable targets for cancer therapy? Nat Rev Cancer

425 2010;10(12):842-57.

426 3. Huff LP, Decristo MJ, Trembath D, Kuan PF, Yim M, Liu J, et al. The Role of Ect2

427 Nuclear RhoGEF Activity in Ovarian Cancer Cell Transformation. Genes Cancer

428 2013;4(11-12):460-75.

429 4. Symons M, Segall JE. Rac and Rho driving tumor invasion: who's at the wheel?

430 Genome Biol 2009;10(3):213.

431 5. Welch HC. Regulation and function of P-Rex family Rac-GEFs. Small GTPases

432 2015;6(2):49-70.

433 6. Qin J, Xie Y, Wang B, Hoshino M, Wolff DW, Zhao J, et al. Upregulation of PIP3-

434 dependent Rac exchanger 1 (P-Rex1) promotes prostate cancer metastasis.

435 Oncogene 2009;28(16):1853-63.

436 7. Wong CY, Wuriyanghan H, Xie Y, Lin MF, Abel PW, Tu Y. Epigenetic regulation

437 of phosphatidylinositol 3,4,5-triphosphate-dependent Rac exchanger 1 gene

438 expression in prostate cancer cells. J Biol Chem 2011;286(29):25813-22.

439 8. Dillon LM, Bean JR, Yang W, Shee K, Symonds LK, Balko JM, et al. P-REX1

440 creates a positive feedback loop to activate growth factor receptor, PI3K/AKT

441 and MEK/ERK signaling in breast cancer. Oncogene 2015;34(30):3968-76.

ERK Regulates PREX1-RAC1 in Human Melanoma

442 9. Montero JC, Seoane S, Ocana A, Pandiella A. P-Rex1 participates in Neuregulin-

443 ErbB signal transduction and its expression correlates with patient outcome in

444 breast cancer. Oncogene 2011;30(9):1059-71.

445 10. Sosa MS, Lopez-Haber C, Yang C, Wang H, Lemmon MA, Busillo JM, et al.

446 Identification of the Rac-GEF P-Rex1 as an essential mediator of ErbB signaling

447 in breast cancer. Mol Cell 2010;40(6):877-92.

448 11. Berger MF, Hodis E, Heffernan TP, Deribe YL, Lawrence MS, Protopopov A, et

449 al. Melanoma genome sequencing reveals frequent PREX2 mutations. Nature

450 2012;485(7399):502-6.

451 12. Lissanu Deribe Y, Shi Y, Rai K, Nezi L, Amin SB, Wu CC, et al. Truncating

452 PREX2 mutations activate its GEF activity and alter gene expression regulation

453 in NRAS-mutant melanoma. Proc Natl Acad Sci U S A 2016;113(9):E1296-305.

454 13. Li YY, Hanna GJ, Laga AC, Haddad RI, Lorch JH, Hammerman PS. Genomic

455 analysis of metastatic cutaneous squamous cell carcinoma. Clin Cancer Res

456 2015;21(6):1447-56.

457 14. Comprehensive molecular characterization of gastric adenocarcinoma. Nature

458 2014;513(7517):202-9.

459 15. Hodis E, Watson IR, Kryukov GV, Arold ST, Imielinski M, Theurillat JP, et al. A

460 landscape of driver mutations in melanoma. Cell 2012;150(2):251-63.

461 16. Krauthammer M, Kong Y, Ha BH, Evans P, Bacchiocchi A, McCusker JP, et al.

462 Exome sequencing identifies recurrent somatic RAC1 mutations in melanoma.

463 Nat Genet 2012;44(9):1006-14.

ERK Regulates PREX1-RAC1 in Human Melanoma

464 17. Lindsay CR, Lawn S, Campbell AD, Faller WJ, Rambow F, Mort RL, et al. P-

465 Rex1 is required for efficient melanoblast migration and melanoma metastasis.

466 Nat Commun 2011;2:555.

467 18. Shields JM, Thomas NE, Cregger M, Berger AJ, Leslie M, Torrice C, et al. Lack

468 of extracellular signal-regulated kinase mitogen-activated protein kinase signaling

469 shows a new type of melanoma. Cancer Res 2007;67(4):1502-12.

470 19. Goetz EM, Ghandi M, Treacy DJ, Wagle N, Garraway LA. ERK mutations confer

471 resistance to mitogen-activated protein kinase pathway inhibitors. Cancer Res

472 2014;74(23):7079-89.

473 20. Kwong LN, Boland GM, Frederick DT, Helms TL, Akid AT, Miller JP, et al. Co-

474 clinical assessment identifies patterns of BRAF inhibitor resistance in melanoma.

475 J Clin Invest 2015;125(4):1459-70.

476 21. Ryan MB, Der CJ, Wang-Gillam A, Cox AD. Targeting RAS-mutant cancers: is

477 ERK the key? Trends Cancer 2015;1(3):183-98.

478 22. Sullivan R, LoRusso P, Boerner S, Dummer R. Achievements and challenges of

479 molecular targeted therapy in melanoma. Am Soc Clin Oncol Educ Book

480 2015:177-86.

481 23. Pearlstein MV, Zedek DC, Ollila DW, Treece A, Gulley ML, Groben PA, et al.

482 Validation of the VE1 immunostain for the BRAF V600E mutation in melanoma. J

483 Cutan Pathol 2014;41(9):724-32.

484 24. Chan KT, Asokan SB, King SJ, Bo T, Dubose ES, Liu W, et al. LKB1 loss in

485 melanoma disrupts directional migration toward extracellular matrix cues. J Cell

486 Biol 2014;207(2):299-315.

ERK Regulates PREX1-RAC1 in Human Melanoma

487 25. Vultur A, Villanueva J, Krepler C, Rajan G, Chen Q, Xiao M, et al. MEK inhibition

488 affects STAT3 signaling and invasion in human melanoma cell lines. Oncogene

489 2014;33(14):1850-61.

490 26. Fiordalisi JJ, Keller PJ, Cox AD. PRL tyrosine phosphatases regulate rho family

491 GTPases to promote invasion and motility. Cancer Res 2006;66(6):3153-61.

492 27. Welch HC, Condliffe AM, Milne LJ, Ferguson GJ, Hill K, Webb LM, et al. P-Rex1

493 regulates neutrophil function. Curr Biol 2005;15(20):1867-73.

494 28. Pollock PM, Harper UL, Hansen KS, Yudt LM, Stark M, Robbins CM, et al. High

495 frequency of BRAF mutations in nevi. Nat Genet 2003;33(1):19-20.

496 29. Poynter JN, Elder JT, Fullen DR, Nair RP, Soengas MS, Johnson TM, et al.

497 BRAF and NRAS mutations in melanoma and melanocytic nevi. Melanoma Res

498 2006;16(4):267-73.

499 30. Yoon S, Seger R. The extracellular signal-regulated kinase: multiple substrates

500 regulate diverse cellular functions. Growth Factors 2006;24(1):21-44.

501 31. Roskoski R, Jr. ERK1/2 MAP kinases: structure, function, and regulation.

502 Pharmacol Res 2012;66(2):105-43.

503 32. Bollag G, Hirth P, Tsai J, Zhang J, Ibrahim PN, Cho H, et al. Clinical efficacy of a

504 RAF inhibitor needs broad target blockade in BRAF-mutant melanoma. Nature

505 2010;467(7315):596-9.

506 33. Casar B, Pinto A, Crespo P. Essential role of ERK dimers in the activation of

507 cytoplasmic but not nuclear substrates by ERK-scaffold complexes. Mol Cell

508 2008;31(5):708-21.

ERK Regulates PREX1-RAC1 in Human Melanoma

509 34. Herrero A, Pinto A, Colon-Bolea P, Casar B, Jones M, Agudo-Ibanez L, et al.

510 Small Molecule Inhibition of ERK Dimerization Prevents Tumorigenesis by RAS-

511 ERK Pathway Oncogenes. Cancer Cell 2015;28(2):170-82.

512 35. Smalley KS, Haass NK, Brafford PA, Lioni M, Flaherty KT, Herlyn M. Multiple

513 signaling pathways must be targeted to overcome drug resistance in cell lines

514 derived from melanoma metastases. Mol Cancer Ther 2006;5(5):1136-44.

515 36. Welch HC, Coadwell WJ, Ellson CD, Ferguson GJ, Andrews SR, Erdjument-

516 Bromage H, et al. P-Rex1, a PtdIns(3,4,5)P3- and Gbetagamma-regulated

517 guanine-nucleotide exchange factor for Rac. Cell 2002;108(6):809-21.

518 37. Klein RM, Spofford LS, Abel EV, Ortiz A, Aplin AE. B-RAF regulation of Rnd3

519 participates in actin cytoskeletal and focal adhesion organization. Mol Biol Cell

520 2008;19(2):498-508.

521 38. Klein RM, Aplin AE. Rnd3 regulation of the actin cytoskeleton promotes

522 melanoma migration and invasive outgrowth in three dimensions. Cancer Res

523 2009;69(6):2224-33.

524 39. Damoulakis G, Gambardella L, Rossman KL, Lawson CD, Anderson KE, Fukui

525 Y, et al. P-Rex1 directly activates RhoG to regulate GPCR-driven Rac signalling

526 and actin polarity in neutrophils. J Cell Sci 2014;127(Pt 11):2589-600.

527 40. Hayes TK, Neel NF, Hu C, Gautam P, Chenard M, Long B, et al. Long-Term ERK

528 Inhibition in KRAS-Mutant Pancreatic Cancer Is Associated with MYC

529 Degradation and Senescence-like Growth Suppression. Cancer Cell

530 2016;29(1):75-89.

ERK Regulates PREX1-RAC1 in Human Melanoma

531 41. Ebi H, Costa C, Faber AC, Nishtala M, Kotani H, Juric D, et al. PI3K regulates

532 MEK/ERK signaling in breast cancer via the Rac-GEF, P-Rex1. Proc Natl Acad

533 Sci U S A 2013;110(52):21124-9.

534 42. Barrio-Real L, Benedetti LG, Engel N, Tu Y, Cho S, Sukumar S, et al. Subtype-

535 specific overexpression of the Rac-GEF P-REX1 in breast cancer is associated

536 with promoter hypomethylation. Breast Cancer Res 2014;16(5):441.

537 43. Morrison DK. MAP kinase pathways. Cold Spring Harb Perspect Biol 2012;4(11).

538 44. Sears R, Nuckolls F, Haura E, Taya Y, Tamai K, Nevins JR. Multiple Ras-

539 dependent phosphorylation pathways regulate Myc protein stability. Genes Dev

540 2000;14(19):2501-14.

541 45. Whitmarsh AJ. Regulation of gene transcription by mitogen-activated protein

542 kinase signaling pathways. Biochim Biophys Acta 2007;1773(8):1285-98.

543 46. Monaghan-Benson E, Burridge K. Mutant B-RAF regulates a Rac-dependent

544 cadherin switch in melanoma. Oncogene 2013;32(40):4836-44.

545 47. Campbell AD, Lawn S, McGarry LC, Welch HC, Ozanne BW, Norman JC. P-

546 Rex1 cooperates with PDGFRbeta to drive cellular migration in 3D

547 microenvironments. PLoS One 2013;8(1):e53982.

548 48. Minard ME, Herynk MH, Collard JG, Gallick GE. The guanine nucleotide

549 exchange factor Tiam1 increases colon carcinoma growth at metastatic sites in

550 an orthotopic nude mouse model. Oncogene 2005;24(15):2568-73.

551 49. Uhlenbrock K, Eberth A, Herbrand U, Daryab N, Stege P, Meier F, et al. The

552 RacGEF Tiam1 inhibits migration and invasion of metastatic melanoma via a

553 novel adhesive mechanism. J Cell Sci 2004;117(Pt 20):4863-71.

ERK Regulates PREX1-RAC1 in Human Melanoma

554 50. Xu K, Tian X, Oh SY, Movassaghi M, Naber SP, Kuperwasser C, et al. The

555 fibroblast Tiam1-osteopontin pathway modulates breast cancer invasion and

556 metastasis. Breast Cancer Res 2016;18(1):14.

557

558 FIGURE LEGENDS

559 Figure 1. PREX1 protein levels are elevated in melanoma patient tumor tissues

560 and cell lines, along with phospho-ERK. (A) Western blot analysis of PREX1 protein,

561 phospho-ERK (pERK) and total ERK1/2 in a panel of WT, BRAF- or NRAS-mutant

562 human melanoma tumor cell lines. (B-D) Human tissue samples of benign melanocytic

563 nevi and malignant skin cutaneous melanoma were subjected to IHC for PREX1 and

564 pERK. Shown are (B) the distribution of PREX1 expression in nevi versus melanoma

565 samples as measured by IHC; n=35 and 33, respectively. Samples were first binned

566 according to no, low, medium or high staining intensity for each protein, and then the

567 distribution was graphed to show the relationship between PREX1 and the percent of

568 samples that stained positive for (C) nuclear pERK or (D) cytoplasmic pERK.

569

570 Figure 2. PREX1 regulates spheroid formation and invasion, but not proliferation,

571 of BRAF- and NRAS-mutant melanoma cells in a context-dependent manner. (A)

572 BRAF-mutant A375 and WM2664 and NRAS-mutant SK-MEL-119 and Mel224 cells

573 were transfected with siRNA against PREX1 or a mismatch control (MM) for 48 h, and

574 knockdown was confirmed by western blot (upper panels). Apparent molecular weights

575 are indicated to the right of each panel; vinculin served as a loading control. Fold

ERK Regulates PREX1-RAC1 in Human Melanoma

576 changes in protein expression compared to MM control are shown in numbers below

577 each blot. Effects of PREX1 knockdown on growth in monolayer culture were

578 determined by MTT assay at 72 hr (lower panels). (B) To determine the effects of

579 PREX1 knockdown on invasion, cells were seeded in the upper chamber of a Matrigel-

580 coated Boyden chamber, and allowed to invade towards serum for 24 h, then stained

581 and imaged. ImageJ was used to quantitate invaded cells per field for 5 fields per insert

582 in duplicate inserts (A375, SK-MEL-119, Mel224) or invaded cells over both inserts

583 (WM2664). (C) For spheroid collagen invasion assays, spheroids were allowed to form

584 for 4 days. Total spheroid area was normalized to that of mismatch control-treated

585 cells; impaired spheroid formation is indicated by increased area of the flattened

586 spheroid. WM2664, SK-MEL-119, and Mel224 spheroids were embedded in a collagen

587 matrix and imaged (day 0) and the extent of cell outgrowth/invasion was imaged 3 days

588 later (thick gray lines). Fold change in area from day 3 to day 0 was calculated in

589 ImageJ. Data are represented as mean ± SD and statistical significance was evaluated

590 by Student’s t-test, where *: p<0.05, **: p<0.01, ***: p<0.001, ****: p<0.0001. Scale bar

591 represents 250 µm for invasion assays and 500 µm for spheroids. Experiments shown

592 are representative of two (WM2664, SK-MEL-119) or three (A375, SK-MEL-119)

593 independent experiments.

594 Figure 3. PREX1 regulates active RAC1-GTP, but not active CDC42-GTP, in

595 melanoma cells. A375 and SK-MEL-119 cells were transfected with pooled siRNAs

596 #1-3 against PREX1 or MM control for 48 h, then starved overnight (18 h). RAC1-GTP

597 and CDC42-GTP were measured by GST-PAK-PBD pulldown (A). Apparent molecular

598 weights are indicated to the right of each panel; vinculin served as a loading control.

ERK Regulates PREX1-RAC1 in Human Melanoma

599 Quantification (mean ± SD) using ImageJ (B) is representative of two independent

600 experiments.

601 Figure 4. RAC1 phenocopies PREX1 in regulating spheroid formation in A375 and

602 invasion in SK-MEL-119. A375 and SK-MEL-119 cells were transfected with siRNA

603 against PREX1, RAC1, or MM control for 48 h before seeding into invasion chambers.

604 Knockdown was confirmed by western blot in A375 (panel A) and SK-MEL-119 (panel

605 D). Apparent molecular weights are on the right of each panel; vinculin was a loading

606 control. Total spheroid area of A375 cells was quantified after 4 days using ImageJ

607 (white line); increased flattened spheroid area indicates impaired spheroid formation

608 (panels B,C). SK-MEL-119 cells were starved overnight and seeded in a Boyden

609 chamber assay with 20% serum as a chemoattractant and allowed to invade for 24h

610 (10x magnification) (E). Stained inserts were quantified for invaded cells/field, 10 fields

611 per condition, using ImageJ (F). Data are represented as mean ± SD. Student’s t-test,

612 where *: p<0.05, **: p<0.01, ***: p<0.001, ****: p<0.0001). Scale bar represents 250 µm

613 for invasion assays and 500 µm for spheroids.

614 Figure 5. PREX1 protein levels are regulated by the ERK kinase cascade. NRAS-

615 mutant SK-MEL-119 cells were first treated with the indicated concentrations of MEKi

616 trametinib for 48 h, and lysates immunoblotted for PREX1, pERK and MYC (A;

617 quantified in B). SK-MEL-119 cells were next treated with trametinib or the ERK

618 inhibitor SCH772984 for 24 or 48 h, and lysates probed for pERK and PREX1 (C;

619 quantified in D), and for pRSK and total MYC to monitor ERK pathway inhibition (C).

620 Similarly, BRAF-mutant A375 cells were treated with the BRAF inhibitor vemurafenib or

621 with SCH772984 for 24 or 48 h and lysates probed as above (E; quantified in F).

ERK Regulates PREX1-RAC1 in Human Melanoma

622 Quantification is of n=3 experiments for SK-MEL-119 and n=4 for A375. Data are

623 represented as mean ± SD.

624 Figure 6. PREX1 levels are regulated by ERK both transcriptionally and post-

625 transcriptionally. BRAF-mutant A375 and WM2664 cells were treated with

626 vemurafenib or SCH772984 for 24 h and PREX1 mRNA levels were measured by

627 Taqman qPCR using two independent probes (A,B). NRAS-mutant SK-MEL-119 and

628 Mel224 cells were treated with trametinib or SCH772984 for 24 h, and PREX1 mRNA

629 measured as above (C,D). Taqman analyses indicate compiled results of n=2

630 experiments for WM2664 and Mel224, and n=3 experiments for A375 and SK-MEL-119.

631 To test posttranscriptional regulation, A375 cells were treated with vehicle or

632 SCH772984 in the presence of 50 µg/ml cycloheximide, and lysates were probed by

633 western blot for PREX1, pERK and MYC (E). Quantification of PREX1 levels using

634 ImageJ (F) is representative of n=2 independent experiments. Data are represented as

635 mean ± SD.

A B

) 150

% High

(

y Medium t

PREX1 i

s 100 Low n

e None

pERK t

1 50 X

ERK1/2 E R

P 0 -actin Nevi Melanoma

C D )

) Nuclear pERK vs PREX1

% Cytoplasmic pERK vs PREX1

(

( t 150

i 150

y s

t High

i High

n s

Medium e

n Medium

e I t Low Low 100

n 100 I None K

None

E 50 50

l c

0 p 0

o t

N None/Low Medium/High None/Low Medium/High y

PREX1 Intensity C PREX1 Intensity

A BRAF NRAS A375 WM2664 SK-MEL-119 Mel224 PREX1 PREX1 PREX1 PREX1 siRNA MM 1 2 3 MM 1 2 3 MM 1 2 3 MM 1 2 3 250 250 250 250 PREX1 130 130 Vinculin 100 100 100

1.00 0.37 0.48 0.41 1.00 0.56 0.59 0.56 1.00 0.50 0.51 0.49 1.00 0.31 0.32 0.36

) ) ) )

% % % %

( ( (

( 150 150 150 150

h h h h

t t t t

w w w w

o o o o

r r r

r 100 100 100 100

G G G G

d d d d

e e e e

z z z z

i i i

i 50 50 50 50

l l l l

a a a a

m m m m

r r r r

o o o

o 0 0 0 0

N N N N MM 1 2 3 MM 1 2 3 MM 1 2 3 MM 1 2 3 PREX1 PREX1 PREX1 PREX1 siRNA siRNA siRNA siRNA

B A375 WM2664 SK-MEL-119 Mel224 MM PREX1 #1 MM PREX1 #1 MM PREX1 #1 MM PREX1 #1

PREX1 #2 PREX1 #3 PREX1 #2 PREX1 #3 PREX1 #2 PREX1 #3 PREX1 #2 PREX1 #3

d d d

l l

l 300 150 ns 150 150

e e

e ns *

i i

i *** ns

F F F

/ / /

s s s

l l l

*** l

l l l ns ns

200 l 100 100 100

e e e

ns e

C C C

d d d ****

d ****

e e e

100 e 50 50 50

d d d

d ****

a a a

v v v

n n n

I I I 0 I 0 0 0 MM 1 2 3 MM 1 2 3 MM 1 2 3 MM 1 2 3 PREX1 PREX1 PREX1 PREX1 siRNA siRNA siRNA siRNA

C A375 WM2664 SK-MEL-119 Mel224 MM PREX1 #1 MM PREX1 #1 MM PREX1 #1 MM PREX1 #1

PREX1 #2 PREX1 #3 PREX1 #2 PREX1 #3 PREX1 #2 PREX1 #3 PREX1 #2 PREX1 #3

a a a

e e e

)

r r r

A A A

d d d

i i 250 i

15 15 5

o o o

r r

**** r o

****

e e

200 e 4

h h h

% ****

(

p p p

10 * 10 ns *

S S S

a ns

150 * 3

e *** **

n n n

i i

i ns

e e 100 e

g g

g 5 5

n n n

a a a

r 50 1

h h h

C C C

p 0 0 0

d d d

l l l

S MM 1 2 3 MM 1 2 3 MM 1 2 3 MM 1 2 3

o o o

F F PREX1 F PREX1 PREX1 PREX1 siRNA siRNA siRNA siRNA

A B

MM + - + - PREX1 - - + 1.5 siRNA siRNA + PREX1

25 n i

e RAC1-GTP/total RAC1

RAC1-GTP t o

25 r 1.0 P

CDC42-GTP

e v

25 i t

a 0.5

RAC1 l e

25 R CDC42 0.0

Input 250 MM PREX1 MM PREX1 PREX1 siRNA 130 Vinculin A375 SK-MEL-119 100

A PREX1 B MM siRNA MM 1 2 3 250 PREX1

Vinculin 100 1.00 0.44 0.49 0.17 RAC1 siRNA MM 1 2 3 PREX1 RAC1 25 RAC1 Vinculin 1 1 100

1.00 0.15 0.16 0.06 )

C M

M 250

f ****

o **** 200 **** 2 2

% **** **** ( ***

a 150

e r

A 100

i o

r 50 3 3

e h

p 0

S MM 1 2 3 1 2 3 PREX1 RAC1 siRNA D SK-MEL-119 E SK-MEL-119 PREX1 MM siRNA MM 1 2 3 250 PREX1 Vinculin 100 1.00 0.26 0.43 0.41 RAC1 PREX1 RAC1 siRNA MM 1 2 3 25 RAC1 Vinculin 1 1 100 1.00 0.53 0.41 0.50

F d

l 150

i F

/ 2 2

s l

l 100 e

C ***

e 50 d **** ****

a ****

v **** **** 3 3 n I 0 MM 1 2 3 1 2 3 PREX1 RAC1 siRNA Downloaded from mcr.aacrjournals.org on October 5, 2021. © 2016 American Association for Cancer Research. Author Manuscript Published OnlineFirst on July 14, 2016; DOI: 10.1158/1541-7786.MCR-16-0184 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Figure 5

SK-MEL-119 1.5 SK-MEL-119 n

A B i pERK e

t PREX1 o

Trametinib (nM) r 1.0 P 250

PREX1 e

t a

pERK l 0.5 35 e ERK1/2 R 35 0.0 MYC 4 7 9 6 5 0 0 55 O 2 9 . . . 5 0 S . . 3 5 2 2 0 Vinculin 130 M 0 0 1 6 1 D Trametinib (nM) C SK-MEL-119 D SK-MEL-119

n 1.5 i pERK

Trametinib SCH772984 e

t o r 1.0

h 24 48 24 48 P

v i

nM 0 5 25 0 5 25 0 50 250 0 50 250 t 0.5

a l PREX1 250 e R 0.0

pERK n 1.5 0 5 25 0 5 25 0 50 250 0 50 250

i PREX1 e

35 t o

ERK r 1.0

35 e v

pRSK i t 0.5

70 a l RSK e 70 R 0.0 MYC 0 5 25 0 5 25 0 50 250 0 50 250 55 nM Vinculin Trametinib SCH772984 100 h 24 48 24 48 A375 E F A375

n 1.5 pERK

Vemurafenib SCH772984 i

t o

h 24 48 24 48 r 1.0

nM 0 100 500 0 100 500 0 30 150 0 30 150 e

v i

250 t 0.5 a

PREX1 l e

R 0.0 pERK 0 100500 0 100500 0 30 150 0 30 150 35Treatment (nM) n 1.5 PREX1

i Vemurafenib SCH772984 e ERK Time (h) t o 24 48 r 1.0 24 48

35 P

pRSK e

v i

70 t 0.5

a l

RSK e

70 R MYC 0.0 55 nM 0 100500 0 100500 0 30 150 0 30 150 Vinculin Vemurafenib SCH772984 100 h 24 48 24 48

A B A375 WM2664 2.5 1.5

Probe A Probe A A

2.0 A N

Probe B N Probe B R

R 1.0 m

1.5 m

1 X

1.0 X E

E 0.5

R P 0.5 P 0.0 0.0 DMSO 100 500 30 150 nM DMSO 250 1250 20 100 nM Vemurafenib SCH772984 Vemurafenib SCH772984

C D 1.5 SK-MEL-119 1.5 Mel224

Probe A Probe A

A A N

N Probe B Probe B R

R 1.0 1.0

m m

1 1

X X E

E 0.5 0.5

R R

P P

0.0 0.0 DMSO 5 25 50 250 nM DMSO 3 15 550 2750 nM Trametinib SCH772984 Trametinib SCH772984

E F A375 A375 CHX Vehicle

SCH772984 1 1.5 DMSO

h 0 4 8 16 24 0 4 8 16 24 X SCH772984 250 E

PREX1 R 1.0

pERK e

v i

35 t 0.5

a l

ERK e

35 R 0.0 MYC 55 0 4 8 16 24 h Vinculin 100

ERK/MAPK Signaling Drives Overexpression of the Rac-GEF, PREX1, in BRAF- and NRAS-mutant Melanoma

Meagan B. Ryan, Alexander J. Finn, Katherine H. Pedone, et al.

Mol Cancer Res Published OnlineFirst July 14, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-16-0184

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2016/10/04/1541-7786.MCR-16-0184.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet been Manuscript edited.

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/early/2016/07/14/1541-7786.MCR-16-0184. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.