erk/mapk signaling drives overexpression of the rac-gef ......nras-driven melanoma. in addition,...

Signal Transduction

ERK/MAPK Signaling Drives Overexpression ofthe Rac-GEF, PREX1, in BRAF- and NRAS-MutantMelanomaMeagan B. Ryan1, Alexander J. Finn2, Katherine H. Pedone3, Nancy E. Thomas2,Channing J. Der1,3, and Adrienne D. Cox1,3,4

Abstract

Recently, we identified that PREX1 overexpression is criticalfor metastatic but not tumorigenic growth in a mouse model ofNRAS-driven melanoma. In addition, a PREX1 gene signaturecorrelated with and was dependent on ERK MAPK activation inhuman melanoma cell lines. In the current study, the under-lying mechanism of PREX1 overexpression in human melano-ma was assessed. PREX1 protein levels were increased in mel-anoma tumor tissues and cell lines compared with benign neviand normal melanocytes, respectively. Suppression of PREX1by siRNA impaired invasion but not proliferation in vitro.PREX1-dependent invasion was attributable to PREX1-mediatedactivation of the small GTPase RAC1 but not the related smallGTPase CDC42. Pharmacologic inhibition of ERK signalingreduced PREX1 gene transcription and additionally regulated

PREX1 protein stability. This ERK-dependent upregulation ofPREX1 in melanoma, due to both increased gene transcriptionand protein stability, contrasts with the mechanisms identified inbreast and prostate cancers, in which PREX1 overexpression wasdriven by gene amplification andHDAC-mediated gene transcrip-tion, respectively. Thus, although PREX1 expression is aberrantlyupregulated and regulates RAC1 activity and invasion in thesethree different tumor types, the mechanisms of its upregulationare distinct and context dependent.

Implications: This study identifies an ERK-dependent mecha-nism that drives PREX1 upregulation and subsequent RAC1-dependent invasion in BRAF- and NRAS-mutant melanoma.Mol Cancer Res; 14(10); 1009–18. �2016 AACR.

IntroductionDriver roles in cancer have been identified for several mem-

bers of the Dbl family of Rho guanine nucleotide exchangefactors (RhoGEF), most prominently ECT2, TIAM1, VAV1/2/3,and PREX1/2 (1, 2). Increased expression and activation ofthese RhoGEFs result in enhanced activity of their Rho familysmall GTPase substrates in a context-dependent manner. Forexample, we recently identified overexpression of ECT2 proteinin ovarian cancer, mediated by gene amplification, that resultedin activation of RHOA in the cytosol and RAC1 in the nucleus(3). The critical importance of RAC1 in cancer cell migrationand invasion (4) has further focused attention on the mechan-

isms regulating activators of RAC1, such as the Dbl family ofRhoGEFs.

The highly related Dbl RhoGEFs, PREX1 and PREX2 (56%overall amino acid identity), which are GEFs for RAC1 and otherRho family small GTPases, such as CDC42 (5), have been impli-cated as cancer drivers in several tumor types. The first cancer-driving role for PREX1 was described in prostate cancer (6), inwhich limited analyses of tumor tissue revealed elevated levels ofPREX1 protein. PREX1 was also elevated in metastatic but notprimary prostate tumor cell lines. Suppression of PREX1 by RNAinterference in PC-3 human prostate cancer cells decreased thelevels of activated RAC and impaired tumor cell migration andinvasion in vitro. Conversely, ectopic expression of PREX1 stim-ulated RAC activation and promoted metastatic but not primarytumor growth of CWR22Rv1 prostate tumor cells. A follow-upstudy identified a histone deacetylase (HDAC)-mediated increasein PREX1 gene transcription as a basis for the increased levels ofPREX1 protein in prostate cancer (7).

PREX1 overexpression was also identified in estrogen receptor–positive luminal andHER2-positive breast cancers (8–10). PREX1protein was detected in approximately 60% of breast tumors butnot in normal breast tissue. PREX1 is located in a chromosomalregion frequently amplified in breast cancers, and PREX1 geneamplification was detected in breast cancer cell lines, supportinggene amplification as amechanism for PREX1 protein overexpres-sion in these tumor types. Silencing of PREX1 expression by RNAinterference reduced HER2-stimulated activation of RAC1 andimpaired tumor cell motility and invasion in vitro and tumori-genic growth in vivo (8–10).

A role in cancer for the related RhoGEF PREX2 has also beenidentified, but not by overexpression. Instead, missense

1Department of Pharmacology, The University of North Carolina atChapel Hill, Chapel Hill, North Carolina. 2Department of Dermatology,The University of North Carolina at Chapel Hill, Chapel Hill, NorthCarolina. 3Lineberger Comprehensive Cancer Center, The Universityof North Carolina at Chapel Hill, Chapel Hill, North Carolina. 4Depart-ment of Radiation Oncology, The University of North Carolina atChapel Hill, Chapel Hill, North Carolina.

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

Corresponding Authors: Channing J. Der, The University of North Carolinaat Chapel Hill, Lineberger Comprehensive Cancer Center, CB7295, ChapelHill, NC 27599-7295. Phone: 919-966-5634; Fax: 919-966-9673; E-mail:[email protected]; and Adrienne D. Cox, The University of North Carolinaat Chapel Hill, Lineberger Comprehensive Cancer Center, CB7295, ChapelHill, NC 27599-7295. Phone: 919-966-7712; Fax: 919-966-9673; E-mail:[email protected]

doi: 10.1158/1541-7786.MCR-16-0184

�2016 American Association for Cancer Research.

MolecularCancerResearch

www.aacrjournals.org 1009

on August 25, 2021. © 2016 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst July 14, 2016; DOI: 10.1158/1541-7786.MCR-16-0184

http://mcr.aacrjournals.org/

mutations in PREX2 have been identified in 25% of malignantmelanomas (11). Although no clear mutational hotspots havebeen seen in melanomas, experimental studies support a gain-of-function consequence of these mutations (11, 12). PREX2 mis-sense mutations have also been found in 38% of cutaneoussquamous cell carcinomas (13) and 17% of stomach adenocar-cinomas (14). To date, a similar level of activating missensemutations in PREX1 in cancer has not been reported. However,the occurrence of activatingmissensemutations (e.g., P29S) in thePREX1 substrate, RAC1, in approximately 11% of melanomas(15, 16) is also consistent with a driver role for overexpressedPREX1 in this disease.

Our previous studies revealed overexpression of PREX1 proteinin melanoma cell lines and tumor tissue (17). Furthermore, in amouse model of melanoma, we determined that Prex1-deficientmice were impaired in forming tumormetastases but not primarytumors. Here, extending our mouse model studies, we demon-strate that PREX1 protein is increased in humanmelanoma tumortissue and that PREX1 is required for human melanoma cellinvasion but not proliferation.

In a separate earlier study, we also identified PREX1 as a geneupregulated by the ERK MAPK in melanoma (18). The RAF–MEK–ERK protein kinase cascade is aberrantly activated in upto 80% of melanomas through BRAF or NRAS mutation andserves as a critical therapeutic target in this disease (19–22).These observations supported the possibility that PREX1 pro-tein overexpression in melanoma is driven by ERK activation.In addition, as PREX1 has a demonstrated driver role in othercancers, PREX1 overexpression may be a key driver of ERK-dependent melanoma growth. In this study, we determinedthat PREX1 protein overexpression is blocked by pharmaco-logic inhibitors of RAF–MEK–ERK signaling and that ERKregulates not only PREX1 gene transcription but also PREX1protein stability. Thus, there are significant cancer type distinctmechanisms that drive PREX1 overexpression in cancer.

Materials and MethodsHuman melanoma tissue and IHC

Following Institutional Review Board approval, primary andmetastatic melanomas were retrieved from a series of patientstreated at UNCHealth Care (ChapelHill, NC). Immunohistochem-ical staining was performed in the UNC Department of Dermatol-ogy Dermatopathology Laboratory as we have recently described(23). Briefly, freshly cut 4-mm thick sections of formalin-fixed andparaffin-embedded melanoma tissue blocks were stained using thefully automated Leica Bond III system. Sections were pretreatedusing on-board heat-induced epitope retrieval in EDTA buffer.Following incubation with PREX1 antibody (6F12; provided byMarcus Thelen, Institute for Research in Biomedicine, Bellinzona,Switzerland), chromogenic detectionwas performed using the LeicaRefined Red polymer detection system (Leica Microsystems). Somesections were also counterstained with hematoxylin and eosin.PREX1 antibody staining intensity was scored in a blinded mannerby a pathologist (A.J. Finn) as high, medium, low, or none.

Tissue cultureCutaneous melanoma, breast cancer, and prostate cancer cell

lines were obtained from ATCC. Cells weremaintained in DMEMsupplemented with 10% FBS and were not cultured for longerthan 6 months after receipt from cell banks.

siRNA transfection, proliferation, and invasion assaysFor siRNA knockdown, A375, WM2664, SK-MEL-119, and

Mel224 cells were plated in 6-well plates. Cells were transfectedwith 10 nmol/L siRNA against PREX1 (Thermo Fisher s33364,s33365, s33366; PREX1 #1, #2, #3, respectively), RAC1 (ThermoFisher s11711, s11712, s11713; RAC1 #1, #2, #3, respectively), ormismatch control (Dharmacon #D-001210-05), using Lipofecta-mine RNAimax (Life Technologies). Cells were serum starvedovernight for 18 hours and then seeded for invasion assays after48 hours of siRNA knockdown. For the proliferation assay,cells were seeded at 2–3 � 103 cells per well in 96-well platesand allowed to grow for 72 hours before incubation for 3 hours in3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide(MTT). MTT was solubilized in DMSO, and the absorbance wasread at A570. For the Boyden chamber invasion assay, 1–3� 104

cells were seeded into the upper chamber of Matrigel-coatedinvasion chambers in duplicate (Corning BioCoat) and allowedto invade toward 20% FBS in DMEM for 24 hours. Invasionchamberswerefixed and stained using aDiff-Quik Stain Kit (GE).Invasion chambers were imaged using a 10� objective lens on aNikon TS100 microscope at 5 fields per insert, and images wereanalyzed to calculate invaded cells per field using ImageJ soft-ware. For the collagen spheroid invasion assay, we slightlymodified a published protocol (24, 25). Briefly, 5–10 � 103

cells were seeded in ultra-low attachment round-bottomed 96-well plates (Corning) for 96hours, a sufficient time for the cells toorganize into spheroids. Spheroids were then transferred to 48-well plates and embedded in collagen (1mg/mL rat-tail collagen,BD). Spheroids were imaged at 0 hour and after 72 hours ofinvasion using a 5� objective on a Nikon TS100 microscope.Total spheroid area was calculated as fold change in area of 72-hour outgrowth versus 0-hour spheroid area, using ImageJ.

Pulldown assay to detect GTPase activityLevels of active,GTP-boundRAC1andCDC42were assessed by

an affinity pulldown assay as we described previously (26).Briefly, after 48 hours of siRNA-mediated PREX1 knockdown,whole-cell lysates were exposed to GST-PAK-PBD, which containsthe binding domain of the shared RAC1/CDC42 effector PAK1.After resolving pulldown samples on 15% SDS-PAGE gels andWestern blotting for RAC1 (clone 23A9, BD) and CDC42 (BD),levels of each GTP-bound GTPase were normalized to both totalprotein and the vinculin loading control (Sigma) by densitometryanalysis performed in ImageJ.

Drug treatment and Western blottingBRAF-mutant A375 andWM2664 cells were treated with BRAF

inhibitor vemurafenib (Selleckchem) or ERK inhibitorSCH772984 (kindly provided by Ahmed Samatar, DiscoveryOncology, Merck Research Laboratories, Boston, MA), andNRAS-mutant SK-MEL-119 and Mel224 cells were treated withMEK inhibitor trametinib (Selleckchem) or SCH772984 for 24and 48 hours before samples were collected in RIPA lysis buffer.For PREX1 protein stability experiments, cells were cotreated withcycloheximide (50 mg/mL) and SCH772984 for a 24-hour timecourse before samples were collected in RIPA. Whole-cell lysateswere resolved on 10% SDS-PAGE gels, and Western blotting wasperformed using antibodies to phospho-ERK1/2 (Thr202/Tyr204), total ERK1/2, phospho-RSK (Ser308), phospho-RSK(Thr359/Ser363), total RSK1/2/3, and c-myc (MYC;Cell SignalingTechnology); b-actin and vinculin (Sigma), and PREX1 (6F12;

Ryan et al.

Mol Cancer Res; 14(10) October 2016 Molecular Cancer Research1010




ref. 27). IRDye800-conjugated anti-mouse and anti-rabbit sec-ondary antibodies were from Rockland Immunochemicals.

Quantitative PCRTotal RNA was isolated using an RNeasy Kit (Qiagen), and

reverse transcription was performed using the High-CapacityRNA-to-cDNA Kit (Thermo Fisher). Real-time quantitative Taq-Man PCR was performed on the QuantStudio 6 Flex (ThermoFisher) with FAM/MGB–labeled probes against PREX1(Hs00368207_m1, Hs_001031512, Thermo Fisher) and endog-enous control VIC/TAMRA labeled b-actin (Thermo Fisher).

Statistical analysisData were analyzed using GraphPad Prism 6 software, and

statistical analyses were performed as indicated in the figurelegends.

ResultsPREX1 overexpression is correlated with elevated ERKactivation

We recently identified overexpression of PREX1 protein inmelanomas, determined that Prex1 deficiency impaired mousemelanoblast migration in vivo, and demonstrated that Prex1expression is required formetastasis in anNras-mutant geneticallyengineered mouse model of cutaneous melanoma (17). As ourprevious gene array analyses identified PREX1 as an ERK activa-tion–dependent gene (18), herewe assessed a relationship amongBRAF and NRAS mutation status, ERK activation, and PREX1proteinoverexpression inhumanmelanoma.Wefirst investigatedthe expression of PREX1 in a panel of humanmelanoma cell linesthat did or did not harbor BRAF orNRASmutations. Themajorityof BRAF- (3/4) or NRAS- (3/4) mutant cell lines exhibited sub-stantially higher PREX1 protein expression when compared with

normal melanocytes or with BRAF/NRAS wild-type cell lines(Fig. 1A). Generally, the level of activated, phosphorylated ERK(pERK) correlated with the level of PREX1. In our analyses,normal melanocytes may exhibit low pERK levels (18), or theymay also display low PREX1 protein levels even in the presence ofhigh pERK levels, as shown here.

We next utilized IHC to evaluate PREX1 protein expression andpERK levels in melanoma patient tissues. We first comparedPREX1 expression in benign melanocytic nevi (n ¼ 35) andhumanmelanoma tumors (n¼ 33; Fig. 1B). A range of expressionwas seen in nevi, with approximately 75% expressing low-to-medium levels of PREX1. As BRAF andNRASmutations are foundin a high percentage of nevi (28, 29), it is not surprising to findPREX1 in nevi as well as in melanoma tissue. However, high levelPREX1 expression was detected only in melanomas (�10%, Fig.1B).

ERK can phosphorylate numerous substrates present in boththe nucleus and the cytoplasm (30, 31), few of which have beenfirmly linked to specific outcomes of ERK-mediated signaling.Melanoma responses to pharmacologic inhibitors of the RAF–MEK–ERK pathway [e.g., clinically, to BRAF inhibition (32) andpreclinically, to inhibitors of ERK dimerization (33, 34)] corre-lated with suppression of cytoplasmic pERK. We therefore eval-uated the distribution of pERK in our human melanoma tissuesand found that levels of pERKwere correlatedwith those of PREX1both in the nucleus (Fig. 1C) and in the cytoplasm (Fig. 1D). Bothnuclear and cytoplasmic ERK activity may contribute to increasedexpression of PREX1.

PREX1 regulates invasion in a complex manner in both BRAF-and NRAS-mutant melanoma cell lines

An unexpected observation in our studies of PREX1 functionin a mouse model of melanoma was that Prex1 deficiency

B

DC

A

PREX1

1° M

elan

ocyt

es

PM

WK

Mel

505

Sk-

Mel

-23

Sk-

Mel

-119

Sk-

Mel

-28

A37

5

WM

2664

A20

58

VM

M39

Mel

224

Mel

147

WT NRAS BRAF

pERK

ERK1/2

β -Actin MelanomaNevi0

50

100

150

PREX

1 In

tens

ity (%

)

NoneLowMediumHigh

Medium/HighNone/Low0

50

100

150

Nuclear pERK vs. PREX1

Nuc

lear

pER

K in

tens

ity (%

)

NoneLowMediumHigh

PREX1 IntensityMedium/HighNone/Low

0

50

100

150

Cytoplasmic pERK vs. PREX1

Cyt

opla

smic

pER

K in

tens

ity (%

)

NoneLowMediumHigh

PREX1 Intensity

Figure 1.

PREX1 protein levels are elevated inmelanoma patient tumor tissues andcell lines, along with phospho-ERK. A,Western blot analysis of PREX1 protein,phospho-ERK (pERK), and total ERK1/2in a panel of wild-type (WT), BRAF-, orNRAS-mutant human melanoma tumorcell lines. B–D, human tissue samples ofbenign melanocytic nevi and malignantskin cutaneous melanoma weresubjected to IHC for PREX1 and pERK.Shown are the distribution of PREX1expression in nevi versus melanomasamples asmeasured by IHC (B); n¼ 35and 33, respectively. Samples were firstbinned according to no, low,medium, orhigh staining intensity for each protein,and then the distribution was graphedto show the relationship betweenPREX1 and the percentage of samplesthat stained positive for nuclear pERK(C) or cytoplasmic pERK (D).

ERK Regulates PREX1–RAC1 in Human Melanoma

www.aacrjournals.org Mol Cancer Res; 14(10) October 2016 1011




greatly impaired metastatic but not tumorigenic growth. Thisresult contrasts with studies evaluating the role of PREX1overexpression in human breast cancer cells, in which stableshRNA-mediated suppression of PREX1 reduced their tumori-genic growth (8–10). We therefore compared the effect ofPREX1 suppression in human melanoma cell lines on bothproliferation and invasion in vitro.

To evaluate the role of PREX1 overexpression in cell growth,we first used three independent siRNAs to knock down PREX1in two BRAF-mutant (A375 and WM2664) and two NRAS-mutant cell lines (SK-MEL-119 and Mel224; Fig. 2A, top). Wefound that transient (72 hours) suppression of PREX1 did notsignificantly reduce their proliferation in vitro (Fig. 2A, bottom),consistent with the lack of effect of Prex1 deficiency on thegrowth of primary melanomas in mice. Next, we evaluated therole of PREX1 in invasion by analysis of invasion throughMatrigel toward serum as a chemoattractant. We observed asurprisingly heterogeneous response to PREX1 knockdownthat was independent of BRAF or NRAS mutational status. Forexample, the NRAS-mutant line SK-MEL-119 exhibited a60% to 70% decrease in invasion upon knockdown of PREX1(P < 0.0001, Fig. 2B), whereas the BRAF-mutant cell lineA375 conversely exhibited an approximately 2-fold increase(P < 0.001). In contrast, the already very low degree of directedinvasion of the BRAF-mutant line WM2664 was unaffected byPREX1 knockdown. Similarly, the invasive NRAS-mutant cellline Mel224 was largely unaffected, despite efficient knock-down of PREX1. These data demonstrate that PREX1 plays acomplex and variable role in directed invasion toward anattractant.

Next, we investigated the role of PREX1 in a 3-dimensionalspheroid formation and collagen invasion assay, which mimicsthe in vivo tumor environment of human skin (35). Figure 2Cillustrates both spheroid formation and the subsequent invasionof cells from the spheroid into the surrounding collagenmatrix. Inthree of the four cell lines, knockdown of PREX1 impairedspheroid invasion into collagen, either trending (SK-MEL-119)or significantly so (Mel224, WM2664). In contrast, A375 spher-oidswere defective in formation and did not invade the surround-ing collagen matrix. The nearly doubled total spheroid area ofPREX1 knockdown A375 cells compared with mismatch controlcells observed after 4 days in culture was caused by a flattening ofthe 3-dimensional spheroid structure and not by increased inva-sion or by increased proliferation; no change in proliferationoccurred upon loss of PREX1 (Fig. 2A).

Collectively, our results suggest a complex and context-depen-dent role for PREX1 in driving both directed invasion and 3-dimensional spheroid collagen outgrowth of humanmelanomas,and one that is not dependent on BRAF or NRAS mutationalstatus.

PREX1 regulates active, GTP-bound RAC1 but not CDC42 inmelanoma cells

Although PREX1 is considered a RAC-selective GEF (36),PREX1 is also active on CDC42 (27). We therefore investigatedwhich Rho family small GTPases are activated downstream ofPREX1 in human melanoma cells. We found that knockdown ofPREX1 decreased the levels of activated RAC1, as measured byRAC1-GTP pulldown, in both BRAF-mutant A375 and NRAS-mutant SK-MEL-119 cells (Fig. 3). Despite the continued presenceof other Rac-GEFs capable of inducing nucleotide exchange on

RAC1, even incomplete loss of PREX1 was sufficient to cause asubstantial decrease inRAC1-GTP (Fig. 3). This effectwas selectivefor RAC1, as the levels of activated CDC42 did not decrease(Fig. 3). These results support a role for PREX1 in regulatingRAC1 activity and subsequent RAC1-driven invasive behavior ofmelanomas.

We next asked whether the loss of RAC1 was sufficient tophenocopy the impairment of invasion that we observed uponloss of PREX1. We found that knockdown of either PREX1 orRAC1 with three independent siRNAs for each (Fig. 4A) wassufficient to substantially impair spheroid formation of A375cells, as demonstrated by an increase in theflattened spheroid area(Fig. 4B and C). The degree of impairment upon knockdownof RAC1 was highly significant (P < 0.0001, Fig. 4C) and com-parable with the degree of impairment observed upon knock-down of PREX1 (P < 0.0001, Fig. 4C). Next, we observed that lossof RAC1 (Fig. 4D)was sufficient to prevent the vastmajority of SK-MEL-119 cell invasion in the Boyden chamber assay (Fig. 4E). Thisdecrease in invasion was similar to the decrease seen upon PREX1knockdown (P < 0.0001, Fig. 4F). The ability of RAC1 to pheno-copy PREX1 in impairing both spheroid formation and directedinvasion supports the idea that RAC1 is the most critical Rhofamily small GTPase downstream of PREX1 in regulating invasivemelanoma behavior. Of note, other Rho family small GTPases,such as RND3, have also been shown to be regulated by the RAF–MEK–ERK pathway and to contribute to melanoma invasion andspheroid outgrowth (37, 38). However, as a Rho-like rather than aRac-like GTPase, RND3 is unlikely to be a target of PREX1 in thiscontext (39).

PREX1 protein levels are positively regulated by ERK activity inmelanoma

Our evaluation of human melanoma cell lines and tumortissue found a correlation between phosphorylated ERK andPREX1 protein overexpression. To directly address whether ERKactivation is required for PREX1 overexpression, we evaluatedwhether pharmacologic inhibition of RAF–MEK–ERK signalingwould reduce PREX1 protein levels in BRAF- and NRAS-mutantmelanoma. We first treated NRAS-mutant SK-MEL-119 cellswith increasing concentrations of the MEK inhibitor trametinib.The more effective the inhibition of MEK, as measured bydecreasing levels of phosphorylated and activated ERK (pERK)and total MYC (an ERK substrate; ERK phosphorylation blocksdegradation), the greater the decrease in PREX1 protein (Fig. 5Aand B), suggesting a direct correlation between ERK activity andPREX1 protein levels. We next wanted to determine whetherthis dose-dependent decrease in PREX1 protein would alsooccur when the ERK MAPK cascade was inhibited at differentnodes, and whether such an effect is time dependent. Wetherefore treated SK-MEL-119 cells with two different concen-trations (1� EC50 and 5� EC50 for growth) of either trametinibor the ERK inhibitor SCH772984, for either 24 or 48 hours(Fig. 5C and D). PREX1 protein levels tracked closely with thelevel of pERK at each concentration of inhibitor, and this wassustained for 48 hours. Next, we investigated whether PREX1protein was similarly regulated downstream of the ERK–MAPKcascade in BRAF-mutant melanoma cells. We treated A375 cellssimilarly but with the BRAF inhibitor vemurafenib orSCH772984 for 24 or 48 hours (Fig. 5E and F). Similarly toSK-MEL-119 cells, levels of PREX1 in A375 cells tracked closelywith the levels of pERK and demonstrated a time-dependent

Ryan et al.





effect. In both SK-MEL-119 and A375 cells, phosphorylation ofthe ERK substrate RSK (pRSK) and total MYC served as effectivemarkers to demonstrate inhibition of ERK, as we have observed

in other settings (40). These results indicate that ERK MAPKactivity is an important contributor to the total amount ofPREX1 protein in melanoma cells.

A375 WM2664 SK-MEL-119 Mel224A

B

SARNFARB

PREX1Vinculin

siRNA MM 1 2 3PREX1

1.00 0.37 0.48 0.41

130

250MM 1 2 3

PREX1

1.00 0.56 0.59 0.56100

250MM 1 2 3

PREX1

1.00 0.50 0.51 0.49100130

250MM 1 2 3

PREX1

1.00 0.31 0.32 0.36100

250

A375 WM2664 SK-MEL-119 Mel224PREX1 #1

PREX1 #2 PREX1 #3

MM PREX1 #1

PREX1 #2 PREX1 #3

1# 1XERPMM

PREX1 #2 PREX1 #3

1# 1XERPMM

PREX1 #2 PREX1 #3

MM

MM 1 2 30

50

100

150

siRNA

Nor

mal

ized

gro

wth

(%)

PREX1MM 1 2 3

0

50

100

150

siRNA

Nor

mal

ized

gro

wth

(%)

PREX1MM 1 2 3

0

50

100

150

siRNA

Nor

mal

ized

gro

wth

(%)

PREX1MM 1 2 3

0

50

100

150

siRNA

Nor

mal

ized

gro

wth

(%)

PREX1

MM 1 2 30

100

200

300

siRNA

Inva

ded

cells

/fiel

d

***ns

***

PREX1MM 1 2 3

0

50

100

150

siRNA

Inva

ded

cells

/inse

rt ns

ns

ns

PREX1MM 1 2 3

0

50

100

150

siRNA

Inva

ded

cells

/fiel

d

******** ****

PREX1MM 1 2 3

0

50

100

150

siRNA

Inva

ded

cells

/fiel

d *

nsns

PREX1

A375 WM2664 SK-MEL-119 Mel224CPREX1 #1

PREX1 #2 PREX1 #3

1# 1XERPMM

PREX1 #2 PREX1 #3

1# 1XERPMM

PREX1 #2 PREX1 #3

1# 1XERPMM

PREX1 #2 PREX1 #3

MM

MM 1 2 30

50

100

150

200

250

siRNA

Sph

eroi

d ar

ea (%

of M

M)

********

****

PREX1MM 1 2 3

0

5

10

15

siRNA

Fold

cha

nge

in s

pher

oid

area

**

*

PREX1MM 1 2 3

0

5

10

15

siRNA

Fold

cha

nge

in s

pher

oid

area

ns nsns

PREX1MM 1 2 3

0

1

2

3

4

5

siRNA

Fold

cha

nge

in s

pher

oid

area

******

PREX1

Figure 2.

PREX1 regulates spheroid formation andinvasion, but not proliferation, of BRAF-and NRAS-mutant melanoma cells in acontext-dependent manner. A, BRAF-mutant A375 and WM2664 and NRAS-mutant SK-MEL-119 and Mel224 cells weretransfectedwith siRNA against PREX1 or amismatch control (MM) for 48 hours, andknockdown was confirmed by Westernblot analysis (top). Apparent molecularweights are indicated to the right of eachpanel; vinculin served as a loading control.Fold changes in protein expressioncompared with mismatch control areshown in numbers below each blot.Effects of PREX1 knockdown on growth inmonolayer culture were determined byMTT assay at 72 hours (bottom). B, todetermine the effects of PREX1knockdownon invasion, cellswere seededin the upper chamber of a Matrigel-coatedBoyden chamber and allowed to invadetoward serum for 24 hours, then stainedand imaged. ns, not significant. ImageJwas used to quantitate invaded cells perfield for 5 fields per insert in duplicateinserts (A375, SK-MEL-119, Mel224) orinvaded cells over both inserts (WM2664).C, for spheroid collagen invasion assays,spheroidswere allowed to form for 4 days.Total spheroid area was normalized tothat of mismatch control–treated cells;impaired spheroid formation is indicatedby increased area of the flattenedspheroid. WM2664, SK-MEL-119, andMel224 spheroids were embedded in acollagen matrix and imaged (day 0), andthe extent of cell outgrowth/invasion wasimaged 3 days later (thick gray lines). Foldchange in area from day 3 to day 0 wascalculated in ImageJ. Data, mean � SD.Statistical significance was evaluated byStudent t test, in which � , P < 0.05;�� , P < 0.01; �� , P < 0.001; and�� , P < 0.0001. Scale bar, 250 mm(invasion assays) and500mm(spheroids).Experiments shown are representative oftwo (WM2664, SK-MEL-119) or three(A375, SK-MEL-119) independentexperiments.






PREX1 levels are regulated by ERK both transcriptionally andposttranscriptionally in melanoma

Todeterminewhether the loss of PREX1protein uponblockadeof the ERK MAPK cascade was due to loss of PREX1 mRNA, wetreated two BRAF-mutant and two NRAS-mutant melanoma celllines with ERKMAPK cascade inhibitors. BRAF-mutant A375 andWM2664 cells were treated for 24 hours with vemurafenib orSCH772984 as above. TaqManquantitative PCR analysis revealedthat PREX1mRNA,measured by two independent probes, did notchange upon inhibition of BRAFor ERK inA375 cells (Fig. 6A) butdecreased dose dependently in WM2664 cells upon inhibitionof either BRAF or ERK (Fig. 6B). NRAS-mutant cells were treatedwith trametinib or SCH772984 as above.Weobserved that PREX1mRNA also decreased upon ERK inhibition in both SK-MEL-119(Fig. 6C) and Mel224 cells (Fig. 6D). Additional melanoma linesalso exhibited reduced PREX1 mRNA levels when treated withinhibitors of the ERKMAPK cascade, including BRAF-mutant SK-MEL-28 and NRAS-mutant SK-MEL-147 cells (SupplementaryFig. S3A and S3B). Thus, in the majority of melanoma cell lines,ERK MAPK regulates PREX1 protein levels both transcriptionallyand posttranscriptionally.

That ERK MAPK activity altered PREX1 protein levels in A375melanoma cells without changes at the transcriptional levelsuggested a posttranscriptional mechanism for ERKMAPK-medi-ated regulation of PREX1 protein in these cells. To investigate thispossibility, we treated A375 cells with vehicle or SCH772984 inthe presence of the protein synthesis inhibitor cycloheximide atvarious time points over 24 hours (Fig. 6E and F). Inhibition ofERK led to greater loss of PREX1 protein in the presence ofcycloheximide comparedwith vehicle-treated cells. Similar resultswere obtained upon treatment of SK-MEL-119 cells with trame-tinib in the presence of cycloheximide (Supplementary Fig. S3Cand S3D). These results indicate that ERK can regulate proteinstability as well as transcription of PREX1 in melanoma cell lines.

Finally, we tested the possibility that PREX1 is not only an ERKtarget but also an ERK activator. As it has been demonstrated thatPREX1 can regulate MEK–ERK signaling through RAC1 in breastcancer (8, 41), we also examined whether PREX1 can regulateERK1/2 phosphorylation in our melanoma lines. We found thatknockdown of PREX1 did not alter ERK1/2 phosphorylation inthe BRAF-mutant cell lines, A375 and WM2664, or the NRAS-mutant cell lines, SK-MEL-119 and Mel224 (SupplementaryFig. S4A and S4B and Supplementary Fig. S4C and S4D,respectively).

To determine the generality of ERK regulation of PREX1 expres-sion in nonmelanoma tumor types, we also tested whether thisheld true in breast and prostate cancer cell lines. PREX1 has beenshown to be overexpressed in these tumor types, but the role ofERK in its expression has not been explored. Unlike our observa-tions in melanoma cells, we observed that inhibition of the ERKMAPK cascade in T47D and MCF7 breast cancer cells did notreduce PREX1 protein (Supplementary Fig. S5A and S5C) ormRNA (Supplementary Fig. S5B and S5D). Conversely, in PC-3prostate cancer cells, inhibition of the ERKMAPK cascade reducedPREX1 mRNA levels Supplementary Fig. S5F) but had minimaleffects on PREX1 protein Supplementary Fig. S5E). These resultssupport distinct mechanisms of regulating PREX1 expression inmelanoma through ERK1/2 that do not apply in breast or prostatecancer, cancers in which BRAF and RAS mutation frequencies arelow. Overall, our results support that both ERK regulation ofPREX1 abundance andPREX1 regulationof ERKphosphorylationare context dependent andmay differ between breast and prostatecancers and cutaneous melanoma.

DiscussionWe determined that the ERKMAPK cascade plays an important

role in driving PREX1 protein overexpression in both BRAF- andNRAS-mutant melanomas. In contrast, ERK is not the key driverfor PREX1 overexpression in prostate (6, 7) or in breast carcino-mas (8–10, 42), where its abundance is associated with HDAC-dependent (7) PREX1 gene transcription and with HDAC- andmethylation-dependent PREX1 gene transcription (42) and geneamplification (8, 10, 42), respectively. Thus, there are strikingcancer-type differences inmechanisms driving PREX1 overexpres-sion. In support of this idea, PREX1 and PREX2 display distinctexpression and mutation patterns in breast cancer, prostate can-cer, and melanoma (Supplementary Fig. S1), and PREX1 inparticular is differentially amplified in breast cancer, prostatecancer, and cutaneous melanoma (Supplementary Fig. S2). Ourfindings also suggest that loss of PREX1–RAC1 signaling maycontribute to the clinical response of patients with BRAF-mutantmelanomas to BRAF and MEK inhibitors.

The effectiveness of these inhibitors provides compelling evi-dence that aberrant ERK signaling is a major driver of melanomagrowth.Despite this clear driver role, the ERK targets important formelanoma growth remain poorly characterized. The ERK1/2kinases can phosphorylate more than 200 known substrates

BA

RAC1

PREX1

Vinculin

Inpu

t

PREX1

RAC1-GTP

MM

siR

NA

CDC42-GTP

CDC4225

25

25

25

250

100

130

+ –+–

+ –+–

PREX1MMPREX1MM0.0

0.5

1.0

1.5

Rel

ativ

e pr

otei

n PREX1

siRNA

RAC1-GTP/total RAC1

SK-MEL-119A375

Figure 3.

PREX1 regulates active RAC1-GTP, butnot active CDC42-GTP, in melanomacells. A375 and SK-MEL-119 cells weretransfected with pooled siRNAs #1 to 3against PREX1 or mismatch (MM)control for 48 hours, then starvedovernight (18 hours). RAC1-GTP andCDC42-GTP were measured by GST-PAK-PBD pulldown (A). Apparentmolecular weights are indicated to theright of each panel; vinculin served as aloading control. Quantification (mean�SD) using ImageJ (B) is representativeof two independent experiments.

Ryan et al.





(30, 31) and serve as master regulators of numerous transcriptionfactors (43), both directly and indirectly, through both transcrip-tional and posttranslational mechanisms (40, 43–45). We havedetermined that ERK regulates PREX1 expression levels in part viaprotein stability, a mechanism also not observed in other cancers.ERK regulation of PREX1 protein stability presents a previouslyunknown mechanism of maintaining PREX1 protein expressionandmay explain the basis for the relatively high PREX1 expression

in malignant melanomas where the ERK MAPK cascade isupregulated.

Finally, the PREX1-related RhoGEF PREX2 is activated bymissense mutations in 25% of metastatic melanomas, especiallyby truncating mutations (11, 12), and mutational activation ofthe PREX1/2 target RAC1 has been observed in approximately11% of melanomas (15, 16). The rarity of PREX1 truncatingmutations and lack of apparent hotspots among the fewmissense

A

C

B

D

F

E

PREX1

Vinculin

siRNA 321MMPREX1

0.170.44 0.491.00100

250

Vinculin

RAC1

siRNA 321MMRAC1

0.060.160.151.00100

25

MM

RAC1PREX1

11

22

33

PREX1Vinculin

siRNA MM 321

PREX1

0.410.430.261.00 100

250

VinculinRAC1

siRNA MM 321RAC1

0.500.410.531.00100

25

MM

RAC1PREX1

11

22

33

A375

SK-MEL-119

A375

SK-MEL-119

321321MM0

50

100

150

200

250

siRNA

Sphe

roid

are

a (%

of M

M)

**** ********

**** *******

PREX1 RAC1

321321MM0

50

100

150

siRNA

Inva

ded

cells

/fiel

d

***

********

**** **** ****

RAC1PREX1

Figure 4.

RAC1 phenocopies PREX1 in regulatingspheroid formation in A375 andinvasion in SK-MEL-119. A375 andSK-MEL-119 cells were transfected withsiRNA against PREX1, RAC1, ormismatch (MM) control for 48 hoursbefore seeding into invasion chambers.Knockdown was confirmed by Westernblot analysis in A375 (A) and SK-MEL-119 (D). Apparent molecular weights areon the right of each panel; vinculinwas aloading control. Total spheroid area ofA375 cells was quantified after 4 daysusing ImageJ (white line); increasedflattened spheroid area indicatesimpaired spheroid formation (B and C).SK-MEL-119 cells were starvedovernight and seeded in a Boydenchamber assay with 20% serum as achemoattractant and allowed to invadefor 24 hours (10� magnification; E).Stained inserts were quantified forinvaded cells per field, 10 fields percondition, using ImageJ (F). Data,mean � SD. Student t test, where�� ,P<0.001; �� ,P<0.0001. Scale bar,250 mm (invasion assays) and 500 mm(spheroids).






mutations argue that this is not a significantmechanism of PREX1activation inmelanoma. Instead, our determination that PREX1 isoverexpressed and regulated at multiple levels in response to theERK MAPK cascade characterizes a third mechanism for drivingaberrant RAC1 signaling in melanoma. Our findings that loss ofRAC1 phenocopies loss of PREX1 with respect to invasive behav-ior regardless of BRAF or NRAS mutation status supports the

importance of the PREX1–RAC1 relationship as a promoter ofmelanoma cell invasion. Interestingly, downregulationof theRac-GEF TIAM1 by mutant BRAF was shown to enhance invasion ofhumanmelanoma cells (46). Although PREX1 was not examinedin that particular study, PREX1 has consistently demonstrated apositive role in invasion (6, 17, 47), whereas TIAM1 can be eithera positive or a negative regulator of this process (1, 46, 48–50).

0.0

0.5

1.0

1.5

Rel

ativ

e pr

otei

n pERKPREX1

Trametinib (nmol/L)

0.0

0.5

1.0

1.5

Rel

ativ

e pr

otei

n pERK

150300150300100 5000100 50000.0

0.5

1.0

1.5

Rel

ativ

e pr

otei

n PREX1

nmol/L

hSCH772984Vemurafenib

4824 24 48

DC

E F

BA

PREX1

pERK

Vinculin

ERK

pRSK

MYC

RSK

SCH772984Trametinib24 2448 48h

nmol/L 25050025050 025502550

100

250

35

3570

70

55

Vemurafenib SCH77298424 2448 48h

nmol/L 15030030 1500100 5000100 5000PREX1

pERK

Vinculin

ERK

pRSK

RSK

MYC

100

250

35

35

70

70

55

ERK1/2

PREX1

0 0.24

0.97

3.9

15.6

62.5

250

1,00

0

DMSO 0.24

0.97 3.9

15.6

62.5

250

1,00

0MYC

pERK

Vinculin

Trametinib (nmol/L)250

35

35

13055

A375

SK-MEL-119SK-MEL-119

SK-MEL-119SK-MEL-119

0.0

0.5

1.0

1.5

Rel

ativ

e pr

otei

n pERK

250500250500255025500.0

0.5

1.0

1.5

Rel

ativ

e pr

otei

n PREX1

nmol/L

hSCH772984Trametinib

24 48 4824

A375

Figure 5.

PREX1 protein levels are regulated by the ERK kinase cascade.NRAS-mutant SK-MEL-119 cells were first treatedwith the indicated concentrations of MEKi trametinibfor 48 hours, and lysates immunoblotted for PREX1, pERK, and MYC (A; quantified in B). SK-MEL-119 cells were next treated with trametinib or the ERKinhibitor SCH772984 for 24 or 48 hours, and lysates probed for pERK and PREX1 (C; quantified inD), and for pRSK and total MYC to monitor ERK pathway inhibition(C). Similarly, BRAF-mutant A375 cells were treated with the BRAF inhibitor vemurafenib or with SCH772984 for 24 or 48 hours and lysates probed asabove (E; quantified in F). Quantification is of n ¼ 3 experiments for SK-MEL-119 and n ¼ 4 for A375. Data, mean � SD.

Ryan et al.





Thus, the relative input from different upstream activators ofRAC1 can have a profound influence on melanoma invasion.

In summary, we have demonstrated that the ERK MAPK cascademediates overexpression of PREX1 in melanoma at multiple levels,andbymechanisms that aredistinct fromthose identifiedpreviouslyin other cancer types. Our results contribute to a better understand-ing of how Rac-GEFs are modulated in distinct cancer contexts.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: M.B. Ryan, A.J. Finn, C.J. Der, A.D. CoxDevelopment of methodology: M.B. Ryan, A.J. Finn, N.E. ThomasAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): M.B. Ryan, A.J. Finn, K.H. Pedone, N.E. Thomas,C.J. Der, A.D. CoxAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): M.B. Ryan, K.H. Pedone, C.J. Der, A.D. Cox

Writing, review, and/or revision of the manuscript: M.B. Ryan, C.J. Der,A.D. CoxAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): C.J. DerStudy supervision: N.E. Thomas, C.J. Der, A.D. Cox

Grant SupportC.J. Der and A.D. Cox are supported by grants from the NIH (CA42978,

CA179193, CA175747, and CA199235) and the Pancreatic Cancer ActionNetwork-AACR. C.J. Der is also supported by the Lustgarten Pancreatic CancerFoundation. M.B. Ryan has fellowship support from the NIH (Cancer CellBiology Training Grant T32 CA071341 and an NRSA predoctoral fellowship,F31 CA192829).

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received May 31, 2016; revised June 28, 2016; accepted July 1, 2016;published OnlineFirst July 14, 2016.

References1. Cook DR, Rossman KL, Der CJ. Rho guanine nucleotide exchange factors:

regulators of Rho GTPase activity in development and disease. Oncogene2014;33:4021–35.

2. Vigil D, Cherfils J, Rossman KL, Der CJ. Ras superfamily GEFs and GAPs:validated and tractable targets for cancer therapy? Nat Rev Cancer2010;10:842–57.

25050255DMSO0.0

0.5

1.0

1.5PR

EX1

mR

NA

nmol/L

SCH772984Trametinib

Probe AProbe B

2750550153DMSO0.0

0.5

1.0

1.5

PREX

1 m

RN

A

nmol/L

Trametinib SCH772984

Probe AProbe B

24168400.0

0.5

1.0

1.5

Rel

ativ

e PR

EX1 DMSO

SCH772984

h

BA

E

DC

F

WM2664A375

SK-MEL-119 Mel224

A375

VehicleSCH772984

h 1684016840

+ + + + ++ + + + +

2424PREX1pERK

ERK

MYCVinculin

CHX

250

35

35

55

100

A375

15030500100DMSO0.0

0.5

1.0

1.5

2.0

2.5

PREX

1 m

RN

A

nmol/L

Vemurafenib SCH772984

Probe AProbe B

100201250250DMSO0.0

0.5

1.0

1.5

PREX

1 m

RN

A

nmol/L

Vemurafenib SCH772984

Probe AProbe B

Figure 6.

PREX1 levels are regulated byERK both transcriptionally andposttranscriptionally. BRAF-mutantA375 and WM2664 cells were treatedwith vemurafenib or SCH772984 for 24hours and PREX1 mRNA levels weremeasured by TaqMan qPCR using twoindependent probes (A and B). NRAS-mutant SK-MEL-119 and Mel224 cellswere treated with trametinib orSCH772984 for 24 hours, and PREX1mRNA measured as above (C and D).TaqMan analyses indicate compiledresults of n ¼ 2 experiments forWM2664 and Mel224, and n ¼ 3experiments for A375 and SK-MEL-119.To test posttranscriptional regulation,A375 cells were treated with vehicle orSCH772984 in the presence of50 mg/mL cycloheximide, and lysateswere probed by Western blot analysisfor PREX1, pERK, and MYC (E).Quantification of PREX1 levels usingImageJ (F) is representativeof n ¼ 2 independent experiments.Data, mean � SD.






3. Huff LP, Decristo MJ, Trembath D, Kuan PF, YimM, Liu J, et al. The Role ofEct2 nuclear RhoGEF activity in ovarian cancer cell transformation. GenesCancer 2013;4:460–75.

4. Symons M, Segall JE. Rac and Rho driving tumor invasion: who's at thewheel? Genome Biol 2009;10:213.

5. Welch HC. Regulation and function of P-Rex family Rac-GEFs. SmallGTPases 2015;6:49–70.

6. Qin J, Xie Y, Wang B, Hoshino M, Wolff DW, Zhao J, et al. Upregulation ofPIP3-dependent Rac exchanger 1 (P-Rex1) promotes prostate cancermetas-tasis. Oncogene 2009;28:1853–63.

7. Wong CY, Wuriyanghan H, Xie Y, Lin MF, Abel PW, Tu Y. Epigeneticregulation of phosphatidylinositol 3,4,5-triphosphate-dependent Racexchanger 1 gene expression in prostate cancer cells. J Biol Chem 2011;286:25813–22.

8. Dillon LM, Bean JR, Yang W, Shee K, Symonds LK, Balko JM, et al. P-REX1creates a positive feedback loop to activate growth factor receptor, PI3K/AKT and MEK/ERK signaling in breast cancer. Oncogene 2015;34:3968–76.

9. Montero JC, Seoane S, Ocana A, Pandiella A. P-Rex1 participates inNeuregulin-ErbB signal transduction and its expression correlates withpatient outcome in breast cancer. Oncogene 2011;30:1059–71.

10. Sosa MS, Lopez-Haber C, Yang C, Wang H, LemmonMA, Busillo JM, et al.Identification of the Rac-GEF P-Rex1 as an essential mediator of ErbBsignaling in breast cancer. Mol Cell 2010;40:877–92.

11. BergerMF, Hodis E, Heffernan TP, Deribe YL, LawrenceMS, Protopopov A,et al. Melanoma genome sequencing reveals frequent PREX2 mutations.Nature 2012;485:502–6.

12. Lissanu Deribe Y, Shi Y, Rai K, Nezi L, Amin SB, Wu CC, et al. TruncatingPREX2 mutations activate its GEF activity and alter gene expressionregulation in NRAS-mutant melanoma. Proc Natl Acad Sci U S A 2016;113:E1296–305.

13. Li YY,HannaGJ, LagaAC,HaddadRI, Lorch JH,HammermanPS.Genomicanalysis ofmetastatic cutaneous squamous cell carcinoma. Clin Cancer Res2015;21:1447–56.

14. The Cancer Genome Atlas Research Network. Comprehensive molecularcharacterization of gastric adenocarcinoma. Nature 2014;513:202–9.

15. Hodis E,Watson IR, KryukovGV,Arold ST, ImielinskiM, Theurillat JP, et al.A landscape of driver mutations in melanoma. Cell 2012;150:251–63.

16. Krauthammer M, Kong Y, Ha BH, Evans P, Bacchiocchi A, McCusker JP,et al. Exome sequencing identifies recurrent somatic RAC1 mutations inmelanoma. Nat Genet 2012;44:1006–14.

17. Lindsay CR, Lawn S, Campbell AD, Faller WJ, Rambow F, Mort RL, et al.P-Rex1 is required for efficient melanoblast migration and melanomametastasis. Nat Commun 2011;2:555.

18. Shields JM, ThomasNE, CreggerM, Berger AJ, LeslieM, Torrice C, et al. Lackof extracellular signal-regulated kinase mitogen-activated protein kinasesignaling shows a new type of melanoma. Cancer Res 2007;67:1502–12.

19. Goetz EM, Ghandi M, Treacy DJ, Wagle N, Garraway LA. ERK mutationsconfer resistance to mitogen-activated protein kinase pathway inhibitors.Cancer Res 2014;74:7079–89.

20. Kwong LN, Boland GM, Frederick DT, Helms TL, Akid AT, Miller JP, et al.Co-clinical assessment identifies patterns of BRAF inhibitor resistance inmelanoma. J Clin Invest 2015;125:1459–70.

21. Ryan MB, Der CJ, Wang-Gillam A, Cox AD. Targeting RAS-mutant cancers:is ERK the key? Trends Cancer 2015;1:183–98.

22. Sullivan R, LoRusso P, Boerner S, Dummer R. Achievements and challengesofmolecular targeted therapy inmelanoma. AmSocClinOncol Educ Book2015;35:177–86.

23. PearlsteinMV, ZedekDC,Ollila DW, Treece A,GulleyML,Groben PA, et al.Validation of the VE1 immunostain for the BRAF V600E mutation inmelanoma. J Cutan Pathol 2014;41:724–32.

24. Chan KT, Asokan SB, King SJ, Bo T, Dubose ES, Liu W, et al. LKB1 loss inmelanoma disrupts directional migration toward extracellularmatrix cues.J Cell Biol 2014;207:299–315.

25. Vultur A, Villanueva J, Krepler C, Rajan G, Chen Q, Xiao M, et al. MEKinhibition affects STAT3 signaling and invasion in human melanoma celllines. Oncogene 2014;33:1850–61.

26. Fiordalisi JJ, Keller PJ, Cox AD. PRL tyrosine phosphatases regulate rhofamily GTPases to promote invasion and motility. Cancer Res 2006;66:3153–61.

27. Welch HC, Condliffe AM, Milne LJ, Ferguson GJ, Hill K, Webb LM, et al.P-Rex1 regulates neutrophil function. Curr Biol 2005;15:1867–73.

28. Pollock PM, Harper UL, Hansen KS, Yudt LM, Stark M, Robbins CM, et al.High frequency of BRAF mutations in nevi. Nat Genet 2003;33:19–20.

29. Poynter JN, Elder JT, Fullen DR, Nair RP, Soengas MS, Johnson TM, et al.BRAF andNRASmutations inmelanomaandmelanocytic nevi.MelanomaRes 2006;16:267–73.

30. Yoon S, Seger R. The extracellular signal-regulated kinase: multiple sub-strates regulate diverse cellular functions. Growth Factors 2006;24:21–44.

31. Roskoski R Jr. ERK1/2 MAP kinases: structure, function, and regulation.Pharmacol Res 2012;66:105–43.

32. Bollag G, Hirth P, Tsai J, Zhang J, Ibrahim PN, ChoH, et al. Clinical efficacyof aRAF inhibitor needsbroad target blockade inBRAF-mutantmelanoma.Nature 2010;467:596–9.

33. Casar B, Pinto A, Crespo P. Essential role of ERK dimers in the activation ofcytoplasmic but not nuclear substrates by ERK-scaffold complexes. MolCell 2008;31:708–21.

34. Herrero A, Pinto A, Colon-Bolea P, Casar B, JonesM, Agudo-Ibanez L, et al.Small molecule inhibition of ERK dimerization prevents tumorigenesis byRAS-ERK pathway oncogenes. Cancer Cell 2015;28:170–82.

35. Smalley KS, Haass NK, Brafford PA, Lioni M, Flaherty KT, Herlyn M.Multiple signaling pathways must be targeted to overcome drug resistancein cell lines derived from melanoma metastases. Mol Cancer Ther2006;5:1136–44.

36. Welch HC, Coadwell WJ, Ellson CD, FergusonGJ, Andrews SR, Erdjument-Bromage H, et al. P-Rex1, a PtdIns(3,4,5)P3- and Gbetagamma-regulatedguanine-nucleotide exchange factor for Rac. Cell 2002;108:809–21.

37. Klein RM, Spofford LS, Abel EV, Ortiz A, Aplin AE. B-RAF regulation ofRnd3 participates in actin cytoskeletal and focal adhesion organization.Mol Biol Cell 2008;19:498–508.

38. Klein RM, Aplin AE. Rnd3 regulation of the actin cytoskeleton promotesmelanomamigration and invasive outgrowth in three dimensions. CancerRes 2009;69:2224–33.

39. Damoulakis G, Gambardella L, Rossman KL, Lawson CD, Anderson KE,Fukui Y, et al. P-Rex1 directly activates RhoG to regulate GPCR-drivenRac signalling and actin polarity in neutrophils. J Cell Sci 2014;127:2589–600.

40. Hayes TK, Neel NF, Hu C, Gautam P, Chenard M, Long B, et al. Long-termERK inhibition in KRAS-mutant pancreatic cancer is associated with MYCdegradation and senescence-like growth suppression. Cancer Cell 2016;29:75–89.

41. EbiH,CostaC, Faber AC,NishtalaM,KotaniH, JuricD, et al. PI3K regulatesMEK/ERK signaling in breast cancer via the Rac-GEF, P-Rex1. Proc NatlAcad Sci U S A 2013;110:21124–9.

42. Barrio-Real L, Benedetti LG, EngelN, TuY, ChoS, Sukumar S, et al. Subtype-specific overexpression of the Rac-GEF P-REX1 in breast cancer is associatedwith promoter hypomethylation. Breast Cancer Res 2014;16:441.

43. Morrison DK. MAP kinase pathways. Cold Spring Harb Perspect Biol2012;4:pii:a011254.

44. Sears R, Nuckolls F, Haura E, Taya Y, Tamai K, Nevins JR. Multiple Ras-dependent phosphorylation pathways regulate Myc protein stability.Genes Dev 2000;14:2501–14.

45. Whitmarsh AJ. Regulation of gene transcription by mitogen-activatedprotein kinase signaling pathways. Biochim Biophys Acta 2007;1773:1285–98.

46. Monaghan-Benson E, Burridge K.Mutant B-RAF regulates a Rac-dependentcadherin switch in melanoma. Oncogene 2013;32:4836–44.

47. Campbell AD, Lawn S, McGarry LC, Welch HC, Ozanne BW, NormanJC. P-Rex1 cooperates with PDGFRbeta to drive cellular migration in 3Dmicroenvironments. PLoS One 2013;8:e53982.

48. Minard ME, Herynk MH, Collard JG, Gallick GE. The guanine nucle-otide exchange factor Tiam1 increases colon carcinoma growth atmetastatic sites in an orthotopic nude mouse model. Oncogene 2005;24:2568–73.

49. Uhlenbrock K, Eberth A, Herbrand U, Daryab N, Stege P, Meier F, et al. TheRacGEF Tiam1 inhibits migration and invasion of metastatic melanomavia a novel adhesive mechanism. J Cell Sci 2004;117:4863–71.

50. Xu K, Tian X, Oh SY, Movassaghi M, Naber SP, Kuperwasser C, et al. Thefibroblast Tiam1-osteopontin pathway modulates breast cancer invasionand metastasis. Breast Cancer Res 2016;18:14.


Ryan et al.




2016;14:1009-1018. Published OnlineFirst July 14, 2016.Mol Cancer Res Meagan B. Ryan, Alexander J. Finn, Katherine H. Pedone, et al. PREX1, in BRAF- and NRAS-Mutant MelanomaERK/MAPK Signaling Drives Overexpression of the Rac-GEF,

Updated version

10.1158/1541-7786.MCR-16-0184doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2016/10/04/1541-7786.MCR-16-0184.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/14/10/1009.full#ref-list-1

This article cites 50 articles, 14 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/14/10/1009.full#related-urls

This article has been cited by 5 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mcr.aacrjournals.org/content/14/10/1009To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-16-0184

http://mcr.aacrjournals.org/content/suppl/2016/10/04/1541-7786.MCR-16-0184.DC1

http://mcr.aacrjournals.org/content/14/10/1009.full#ref-list-1

http://mcr.aacrjournals.org/content/14/10/1009.full#related-urls

http://mcr.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mcr.aacrjournals.org/content/14/10/1009


erk/mapk signaling drives overexpression of the rac-gef ......nras-driven melanoma. in addition,...

Documents