phosphoproteomic screen identifies potential therapeutic … · signaling and regulation...

Signaling and Regulation

Phosphoproteomic Screen Identifies Potential TherapeuticTargets in Melanoma

Kathryn Tworkoski1, Garima Singhal1, Sebastian Szpakowski2, Christina Ivins Zito1, Antonella Bacchiocchi3,Viswanathan Muthusamy3, Marcus Bosenberg3, Michael Krauthammer1, Ruth Halaban3, and David F. Stern1

AbstractTherapies directed against receptor tyrosine kinases are effective in many cancer subtypes, including lung and

breast cancer. We used a phosphoproteomic platform to identify active receptor tyrosine kinases that mightrepresent therapeutic targets in a panel of 25 melanoma cell strains. We detected activated receptors includingTYRO3, AXL, MERTK, EPHB2, MET, IGF1R, EGFR, KIT, HER3, and HER4. Statistical analysis of receptortyrosine kinase activation as well as ligand and receptor expression indicates that some receptors, such as FGFR3,may be activated via autocrine circuits. Short hairpin RNA knockdown targeting three of the active kinasesidentified in the screen, AXL, HER3, and IGF1R, inhibited the proliferation of melanoma cells and knockdown ofactive AXL also reduced melanoma cell migration. The changes in cellular phenotype observed on AXLknockdown seem to be modulated via the STAT3 signaling pathway, whereas the IGF1R-dependent alterationsseem to be regulated by the AKT signaling pathway. Ultimately, this study identifies several novel targets fortherapeutic intervention in melanoma. Mol Cancer Res; 9(6); 801–12. �2011 AACR.

Introduction

Next generation cancer therapies that target receptortyrosine kinases (RTK) have a major impact on the dis-ease-free progression and survival of patients with breastcancer and non–small cell lung carcinoma (NSCLC). RTKinhibitors in clinical use include the antibodies trastuzumabforHER2-positive breast cancer and cetuximab in epidermalgrowth factor receptor (EGFR)-positive colorectal cancer.Small molecule kinase inhibitors also target EGFR inNSCLC (Erlotinib), and KIT or platelet-derived growthfactor receptor (PDGFR) in gastrointestinal stromal tumors(Imatinib).The incidence of melanoma has steadily risen in the

United States over the past 30 years, yet individuals withlate-stage melanoma have a median survival time of only9 months (1–3). Imatinib has been used to treat melanomaswith mutationally activated KIT, whereas experimentalstudies show that targeting RTKs such as MET and insu-lin-like growth factor 1 receptor (IGF1R) may inhibit

melanoma cell proliferation and survival (4–9). The useof RTK-directed therapeutics in melanoma has, however,been limited by the lack of available information aboutactive RTKs in this disease.To identify potential RTK therapeutic targets in mela-

noma, we surveyed the expression of all 58 human RTKsand their agonists in a panel of 24 low-passage melanomacell strains and 1 commercially available melanoma cell line.The functional activation of 42 human RTKs was alsoexamined and used to identify which RTKs are mostfrequently and intensely activated in melanoma. Shorthairpin RNA (shRNA)-mediated knockdown of AXL,HER3, and IGF1R, 3 of the RTKs identified in the screen,resulted in decreased melanoma cell proliferation. AXLknockdown also reduced melanoma cell migration andthese cellular responses seem to be regulated by the STAT3signaling pathway. These data identify new candidates fortherapeutic intervention in melanoma.

Materials and Methods

Cell cultureYale University (YU)-designated andWW165 melanoma

cell strains were derived from primary and metastatic lesionsas described (10). Melanoma tumors were excised as part ofpatient clinical care and were collected with the partici-pants’ informed written consent according to HIPAA(Health Insurance Portability and Accountability Act) reg-ulations with the approval of the Yale Human InvestigationCommittee. Primary and metastatic melanoma cell strainswere used prior to passage 20. Most melanoma cells werecultured in OptiMEM (Invitrogen) supplemented with 5%

Authors' Affiliations: 1Department of Pathology, 2Graduate Program inComputational Biology and Bioinformatics, and 3Department of Derma-tology, Yale University School of Medicine, New Haven, Connecticut

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

G. Singhal, S. Szpakowski, and C.I. Zito contributed equally to the study.

Corresponding Author: David F. Stern, BML 348 310 Cedar Street, NewHaven, CT 06510. Phone: 203-785-4832; Fax: 203-785-7467; E-mail:[email protected]

doi: 10.1158/1541-7786.MCR-10-0512

�2011 American Association for Cancer Research.

MolecularCancer

Research

www.aacrjournals.org 801

on May 25, 2021. © 2011 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst April 26, 2011; DOI: 10.1158/1541-7786.MCR-10-0512

http://mcr.aacrjournals.org/

FBS and 1% penicillin/streptomycin (basal medium). TheWW165 primary melanoma and YUHEIK mucosal mel-anoma cells were grown in basal medium supplementedwith 0.1 mmol/L IBMX (3-isobutyl-1-methylxanthine;Sigma-Aldrich). Normal human melanocytes from new-born foreskins (NBMEL) and discarded adult skin(RMP32F) were grown in basal medium supplementedwith 16 nmol/L 12-O-tetradecanoyl phorbol-13-acetate,0.1 mmol/L 3-isobutyl-1-methylxanthine, 2.5 nmol/L cho-lera toxin, 1 mmol/L Na3VO4, and 0.1 mmol/L N6,20-O-dibutyryladenosine 30,50-cyclic monophosphate (Sigma-Aldrich), termed TICVA (11). HEK293T cells from theAmerican Type Culture Collection (ATCC) were grown inDulbecco's modified Eagle's medium containing 10% FBS,1% penicillin/streptomycin, and 1% HEPES. MDA-MB-231 cells, MDA-MB-453 cells, and BT474 cells (ATCC)were grown in RPMI containing 10% FBS and 1%penicillin/streptomycin.

Gene expression analysisWhole genome gene expression analysis was based on

data generated by the Yale SPORE in Skin Cancer (12).Briefly, NimbleGen human whole genome expressionmicroarrays (array 2005-04-20_Human_60mer and array2006-08-03_HG18_60mer) were used for hybridizationof cDNA from normal melanocytes and melanoma cellsat NimbleGen Systems Iceland LLC, Vínlandsleið 2-4,113 Reykjavik, Iceland (currently Roche Applied Science,Basel, Switzerland) and by the Yale W.M. Keck Founda-tion Biotechnology Resource as described (12). Themicroarray data from melanoma cell lines (YUCAS,YUCOT, YUDOSO, YUGOE, YUHEF, YUHUY,YUKIM, YUKSI, YULAC, YUPLA, YUROB, YUROL,YUSIV, YUTICA, YUZOR, WW165, and YUHEIK)and 4 independent normal melanocytes cell cultures(NBMEL1-3 and RMP32F) were used to determine agene expression call (expressed/nonexpressed) and a meanRTK expression value. An expression call threshold(761.0) was determined by Stat4 library in R. Briefly,parameters of a bimodal normal distribution were esti-mated separating the gene expression intensity data intonoise (mnoise ¼ 279.9) and signal (msignal ¼ 2,360.6).Genes were considered expressed if their observed inten-sity value (x) was greater than the msignal � 1 SDsignal (i.e.,x > 761.0). To assess correlation between RTK activationand ligand expression, we obtained a list of 473 knownreceptors ligands pairs from the Database of Ligand–Receptor Partners (13). The list was manually augmentedwith additional literature-derived receptor-ligand pairs.Within our dataset 118 receptor-ligands pairs were usedto correlate ligand expression (determined by NimbleGenarray) with RTK phosphorylation (determined phospho-RTK Array).RNA from snap-frozen melanoma tumors was extracted

directly from frozen tumor blocks (YUCOT, YUSIV, andYUSTE) or after microdissection of snap-frozen tumors(YUDOSO and YULAC). RNA was isolated by the RNeasyMini Kit from Qiagen and reverse transcribed with the

iScript cDNA Synthesis Kit from BioRad by using 0.8 mg ofRNA per reaction. Universal TaqMan Master Mix (AppliedBiosystems) was used to conduct quantitative real-timePCR (qRT-PCR) by using a 1 to 10 dilution of the resultingcDNA. The following primers were used following themanufacturer's protocols: AXL (Hs0024357_m1), HER3(Hs00951455_m1), and glyceraldehyde 3 phosphate dehy-drogenase (GAPDH; Hs99999905_m1; Applied Biosys-tems). Relative mRNA expression was determined with theDCt method by using GAPDH as the reference gene. Geneexpression values for tumor RNAs were normalized to anegative control. A single RNA preparation was analyzed foreach microdissected tumor, and duplicates for each tumorblock.

Cell lysis, RTK arrays, immunoblotting, andhierarchical clusteringWhere indicated, cells were starved in 0.1% serum prior to

lysis, and pervanadate (50 mmol/L final) was added for 20minutes before lysis. Cells were lysed in NP-40 lysis buffer[1%NP-40, 150mmol/LNaCl, 50mmol/LTris (pH 7.4), 5mmol/L EDTA, 10% glycerol with complete EDTA-freeprotease inhibitor tablets (Roche) and phosphatase inhibitorcocktails 1 and 2 (Sigma-Aldrich) added immediately beforecell lysis]. Cell lysates (250 mg)were analyzedwith theHumanPhospho-RTK Array Kit (R&D Systems); array maps athttp://www.rndsystems.com/pdf/ary001.pdf. Protein sam-ples were prepared for electrophoresis by addition of 5�Laemmli sample buffer, and immunoblotting was carriedout on nitrocellulose membranes blocked in 5% milk inTris-buffered saline Tween-20 (1.5 mol/L NaCl, 0.2 mol/LTris-HCl, 0.05% Tween 20). Membranes were incubatedwith antibodies in 5% milk or 5% bovine serum albuminovernight at 4�C: anti-pHER3, anti-pIGF1R, anti-IGF1R,anti-pKIT, anti-MERTK, anti-pMET, anti-MET, anti-pSTAT3, anti-STAT3, anti-pAKT, anti-AKT (Cell SignalingTechnology), anti-HER3, anti-GAPDH (Santa Cruz Bio-technology), anti-TYRO3 (Abcam), and anti-KIT (Dako).Secondary antibodies conjugated to horseradish peroxidasewere used at a 1:10,000 dilution for 1 hour at room tem-perature prior to development (Thermo Scientific).Nonsupervised hierarchical clustering was done by manu-

ally determined phospho-array data values ranging from 0(background) to 5 (maximum phosphorylation; Tables 1and 2). Clustering of RTK data was done by Pearson'scorrelation with complete linkage along sample and proteindimensions.

ImmunoprecipitationHER3 immunoprecipitation (IP) was done with 1 mg

protein and 800 ng HER3 antibody sc-7390 (Santa Cruz);KIT IP with 300 mg protein and 1 mg A4502 (Dako); METIP with 550 mg protein and 1mgMET antibody sc-10 (SantaCruz); IGF1R IP with 500 mg protein and 100 ng antibody3027 (Cell Signaling Technology) at 4�C overnight. Then,12 mL of a 50% slurry of protein A/G beads (ThermoScientific) was added to each IP for 1 hour. Beads werewashed twice with NP-40 lysis buffer and once with

Tworkoski et al.

Mol Cancer Res; 9(6) June 2011 Molecular Cancer Research802




salt-Tris buffer (0.1 mol/L NaCl; 10 mmol/L Tris-HCl, pH7.4), then incubated at 100�C for 8 minutes in 2.5�Laemmli sample buffer.

shRNA and lentiviral infectionshRNA targeting AXL [RHS3979-9568950, AXL1 (14) and

RHS3979-9568949, AXL 2), HER3 (RHS3979-9630819;

Table 1. Survey of tyrosine phosphorylated receptors by RTK capture array analysis in untreatedmelanoma cell lines

Without PV

Q61

R

Q61

R

Q61

K

V60

0E

V60

0E

V60

0E

V60

0E

V60

0E

V60

0E

V60

0K

V60

0K

V60

0K

BT

474

MB

231

MB

453

NB

ME

L

YU

GO

E

YU

HE

F

YU

HE

IK

YU

HO

IN

YU

KIM

YU

NIG

E

YU

PLA

YU

RO

B

YU

RO

L

YU

SIV

YU

ZO

R

YU

CA

S

YU

TIC

A

YU

DO

SO

YU

CO

T

YU

GE

N8

YU

HU

Y

YU

SA

C2

YU

SIK

YU

ST

E

YU

KS

I

YU

LAC

YU

MA

C

EGFR 5 0 5 0 1 5 3 0 2 1 1 3 0 0 1 2 0 2 0 3 2 0 1 1 5 3 1HER2 4 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0HER3 3 0 0 0 0 3 0 0 0 1 0 0 4 0 0 2 2 2 0 1 0 0 2 0 1 0 0HER4 0 0 0 0 0 1 0 0 3 0 0 0 1 0 0 0 0 0 0 1 0 0 0 1 5 0 1

FGFR1 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0FGFR2a 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0FGFR3 0 0 0 1 1 1 3 0 5 0 0 1 0 0 0 2 2 3 3 4 3 0 2 1 0 0 0FGFR4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0InsulinR 0 0 0 0 3 0 3 1 5 0 0 0 0 0 0 4 3 3 0 3 3 3 4 3 0 0 0IGF1R 0 0 0 3 5 5 3 1 5 2 3 5 4 5 5 5 5 5 0 5 5 3 5 5 3 5 5AXL 4 3 3 0 0 0 0 3 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0

TYRO3 1 0 0 5 3 0 3 4 4 2 1 5 5 0 5 4 1 0 0 4 3 3 4 5 3 5 5MERTK 1 0 0 1 0 1 0 0 3 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0

MET 0 0 1 1 0 1 0 0 4 0 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0MST1R 0 0 0 0 1 1 3 0 0 0 0 1 1 0 0 0 2 1 0 1 2 0 0 0 0 1 0

PDGFRa 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0PDGFRb 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

KIT 0 0 0 0 0 1 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0FLT-3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

M-CSFR 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0RET 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 4 0 0 0

ROR1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 1 0 0 0 0 0 1 0ROR2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0TIE-1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0TIE-2 1 0 0 0 0 0 3 0 2 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0TRKA 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0TRKB 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0TRKC 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

VEGFR1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0VEGFR2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0VEGFR3 0 0 0 0 0 1 0 0 3 0 0 0 0 0 0 0 3 3 0 0 0 0 0 0 0 0 0MUSK 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHA1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHA2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHA3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHA4 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHA6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHA7 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHB1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHB2 0 0 0 0 1 1 2 1 2 0 0 1 1 0 1 2 3 3 2 3 3 0 0 1 0 3 1EPHB4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHB6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Breast Cancer Melanomas and Normal Newborn Melanocytes

WT BRAF /NRAS

NRAS BRAF

ND* V60

0K

V60

0K

YU

RIF

WW

165

1 30 00 01 00 01 00 00 01 35 50 05 50 03 00 10 01 01 00 00 03 00 00 00 00 00 00 00 00 00 03 00 00 00 00 00 00 00 01 00 00 00 0

NOTE: Summary of receptor activations in breast cancer cell lines, melanoma cell lines, and newborn melanocytes. Signal intensitywas visually scored on a scale from 0 (background) to 5 (most intense). Melanoma cells are grouped by NRAS/BRAF status.aBRAF and NRAS not determined (ND) for breast cancer cell lines.Abbreviation: PV, Pervanadate.

Receptor Tyrosine Kinases in Melanoma

www.aacrjournals.org Mol Cancer Res; 9(6) June 2011 803




ref. 15), IGF1R (TRCN0000121300; ref. 16), empty vectorpLKO control (RHS4080; Open Biosystems), and ascrambled shRNA control (plasmid 1864; Addgene) wereused. Lentivirus stocks were produced by cotransfecting 293Tcells with pMD2.G and psPAX2 (Addgene) by usingFuGENE6 (Roche), with supernatants collected daily for3 days and pooled (Addgene protocol). Experimental infec-tions were conducted at a multiplicity of infection of 5.

Cell proliferationCells were plated in duplicate clear-bottom 96-well plates

at a density of 1,000 cells per well in 100 mL of OptiMEM.Approximately 6 hours after plating, cells were incubatedwith lentivirus in OptiMEM containing 4 mg/mL polybrenefor 20 hours. Immediately following infection, a T0 timepoint was taken by incubating one plate with CellTiter-Glo,per the manufacturer's instructions (Promega). Relative cell

Table 2. Survey of tyrosine phosphorylated receptors by RTK capture array analysis in pervanadate-treated melanoma cell lines

With PV

Q61

R

Q61

R

Q61

K

V60

0E

V60

0E

V60

0E

V60

0E

V60

0E

V60

0E

V60

0K

V60

0K

NB

ME

L

YU

GO

E

YU

HE

F

YU

HE

IK

YU

HO

IN

YU

KIM

YU

NIG

E

YU

PLA

YU

RO

B

YU

RO

L

YU

SIV

YU

ZO

R

YU

CA

S

YU

TIC

A

YU

DO

SO

YU

CO

T

YU

GE

N8

YU

HU

Y

YU

SA

C2

YU

SIK

YU

ST

E

YU

KS

I

YU

LAC

EGFR 0 0 5 3 5 0 0 5 3 0 5 5 2 0 0 0 0 0 0 0 0 0 0HER2 1 1 0 3 5 0 4 1 1 1 0 0 1 4 4 1 3 4 0 1 0 0 0HER3 1 5 5 5 0 5 5 0 5 5 0 5 5 5 5 5 5 5 5 5 5 5 5HER4 3 1 4 0 5 4 1 0 3 3 4 0 1 1 3 1 4 3 1 1 4 5 0

FGFR1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 0 0 0 0 0 0 0FGFR2a 3 0 0 0 0 0 0 0 0 0 0 0 0 2 2 0 0 0 0 0 0 0 0FGFR3 4 1 3 0 4 4 1 1 4 3 0 1 4 5 3 5 5 5 1 3 3 1 1FGFR4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0InsulinR 4 5 2 5 0 3 0 1 1 1 0 3 4 3 3 4 4 3 5 3 3 0 0IGF1R 5 5 5 5 4 3 3 5 4 4 3 5 5 5 5 4 5 5 5 5 5 5 5AXL 0 0 3 0 4 0 0 5 0 0 1 1 1 3 3 0 1 0 1 0 1 1 0

TYRO3 0 0 0 0 0 0 1 0 0 0 0 1 1 3 2 0 3 5 3 1 1 3 3MERTK 1 0 3 0 0 3 2 0 0 1 0 0 1 1 0 2 3 3 1 1 0 0 3

MET 4 5 5 5 5 2 3 3 0 3 3 0 5 3 0 2 5 0 0 1 3 0 3MST1R 5 1 5 5 3 4 0 0 3 3 0 1 1 3 5 4 3 5 1 3 1 0 3

PDGFRa 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0PDGFRb 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0

KIT 5 0 3 5 0 1 0 0 0 0 0 0 0 0 0 0 0 5 0 1 0 0 3FLT-3 3 1 3 0 1 3 0 1 1 3 0 1 3 3 3 2 3 3 3 1 1 1 1

M-CSFR 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0RET 0 1 2 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0

ROR1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0ROR2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0TIE-1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0TIE-2 0 0 3 0 0 2 0 0 0 1 0 0 0 1 3 2 1 0 0 0 0 0 0TRKA 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0TRKB 0 0 2 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0TRKC 0 0 2 0 0 0 0 0 0 0 0 0 0 0 3 1 0 0 1 0 0 0 0

VEGFR1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0VEGFR2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0VEGFR3 3 1 3 0 1 3 0 1 1 3 0 1 3 3 5 2 3 3 1 1 1 1 1

MUSK 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0EPHA1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHA2 0 0 0 0 3 2 0 1 0 0 0 3 0 0 0 0 0 0 0 0 1 0 0EPHA3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHA4 1 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 2 3 3 1 0 0 0EPHA6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHA7 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0EPHB1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHB2 1 1 2 0 1 0 0 0 0 0 0 0 2 0 0 0 1 0 1 0 0 0 3EPHB4 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0EPHB6 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Melanomas and Normal Newborn Melanocytes

WT BRAF /NRAS

NRAS BRAF

V60

0K

V60

0K

V60

0K

YU

MA

C

YU

RIF

WW

165

0 0 00 1 35 0 51 1 00 0 00 0 00 3 10 0 01 0 15 3 31 0 11 0 11 0 31 5 33 1 30 0 00 0 05 5 33 1 10 0 01 0 10 0 00 0 00 0 00 0 10 0 10 0 00 0 00 0 00 0 03 1 10 0 00 0 00 0 00 0 01 0 10 0 00 0 10 0 01 0 10 0 00 0 0

NOTE: Summary of RTK activation in melanoma cell lines and newborn melanocytes scored as described in Table 1.

Tworkoski et al.





accumulation was calculated by dividing the value for theT72 time point by the value for the T0 time point. Assayswere repeated 4 times, with a representative plot shown foreach.

Cell migrationCell lines were stably infected with scrambled shRNA or

AXL shRNA and were seeded at 30,000 cells/well in 24-wellplates with 8-mm filter inserts (BD Biosciences) in Opti-MEM with 0.1% FBS. OptiMEM with 5% FBS was usedas a chemoattractant. Cell lines were analyzed in triplicate,with a representative graph chosen for each.

Results

Transcriptional profiling of melanoma cell linesWe used transcription profiling to identify RTKs com-

monly expressed in melanoma cell lines and in normalmelanocytes. Of the 58 human RTK genes, 43 RTKs weretranscribed above a conservatively chosen expression thresh-old in at least one melanocyte culture, and 53 RTKs wereabove the threshold in at least one melanoma cell line(Supplementary Tables S1 and S2). Some RTKs, includingIGF1R, MET, and TYRO3, were highly expressed in mostmelanomas and normal melanocytes, whereas fibroblastgrowth receptor 2 (FGFR2), EGFR, and ROS1 were rarely,if ever, detected above the expression threshold in eithermelanoma cells or melanocytes. AXL and HER3 wereexpressed at higher levels in melanoma cells than in mel-anocytes, but KIT and MERTK expression was lower inmelanoma cells compared with melanocytes (Supplemen-tary Tables S1 and S2).

Survey of active RTKs in melanomaGenerally, the abundance of RTK mRNA does not

predict functional activity; although RTKs such asHER2 and MET are frequently activated via gene ampli-fication and overexpression, mutationally activated RTKscan be potent oncogenes when expressed at low levels (17–18). Because tyrosine (Tyr) phosphorylation is indicativeof the signaling activity of RTKs (19), we surveyedreceptor activation by using phospho-arrays that capture42 RTKs from cell lysates and are probed with anti-phosphotyrosine (anti-PTyr). To enhance the endogenousPTyr signals, cells were also analyzed after treatment withpervanadate, a tyrosine phosphatase inhibitor. Pervana-date treatment significantly enhanced the cell lysate signalsseen on the phospho-RTK arrays, but did not greatly alterthe pattern of RTK activation (Fig. 1A). Phospho-arraycontrols included breast cancer BT474 and MDA-MB-453 cells, which are known to express active HER2, andMDA-MB-231 cells, which express little or no activeHER3, but have substantial amounts of AXL (Table 1;refs. 20–22).As expected, RTK mRNA levels were only loosely

correlated with receptor activation. For instance, colony-stimulating factor 1 receptor (CSF1R) mRNA levels werehigh in both melanoma cell lines and normal melanocytes,

yet CSF1R was rarely phosphorylated (Tables 1 and 2;Supplementary Tables S1 and S2). The same trend wasalso observed for EPHA2 in melanoma cell lines andFGFR1 in normal melanocytes. Of the 42 RTKs analyzedvia phospho-arrays, IGF1R, HER3, and MET were fre-quently activated and transcribed above the expressionthreshold in melanoma cell lines. Phosphorylation of otherRTKs was more variable in comparison with RNA expres-sion, suggesting the importance of other modulators suchas growth factors or active coreceptors. Pervanadate treat-ment generally increased the number of RTKs activatedand expressed above the threshold (Supplementary TablesS1 and S2).RTK activation patterns were variable, even among cell

lines with BRAF and NRAS activating mutations. Suchmutations occur in 44% and 12%, respectively, of themelanoma cultures analyzed in this report. Examples ofarray variability are found in pervanadate-treated NBMEL,which have higher levels of phosphorylated macrophagestimulating 1 protein receptor (MST1R) than melanomacell lines YUHEIK and YULAC (Fig. 1B). Similarly,YUHEIK cells showed more activated EGFR than eithernewborn melanocytes or YULAC cells, whereas YULACcells had lower activation of KIT and higher activation ofIGF1R (Fig. 1B). Despite these differences, the insulinreceptor family (IR, IGF1-R), EGFR family (especiallyEGFR and HER3), MET family (MET and RON), andTAM family (TYRO3, AXL, and MERTK) were com-monly activated in melanoma cultures (Tables 1 and 2). Aranking of receptor activation across all melanoma cellstrains was determined by summing the phosphorylationintensities for each individual RTK (SupplementaryTable S3). Interestingly, the 15 RTKs with the highestoverall activation included receptors that had not beenpreviously reported to be active in melanoma, includingTYRO3, AXL, MERTK, and EPHB2 (SupplementaryTable S3). RTKs such as EGFR, FGFR3, IR, IGF1R,and TYRO3 were frequently activated with and withoutpervanadate treatment whereas HER3, MET, and RONwere more commonly detected after pervanadate treatment,perhaps reflecting faster turnover of phosphotyrosine byendogenous phosphatases (Tables 1 and 2). The frequencyof detection of phosphorylated EGFR and TYRO3decreased with pervanadate treatment, possibly indicatinga biological change such as heterologous desensitizationthrough activation of other RTKs.In both pervanadate-treated and untreated cultures, it was

common to find multiple phosphorylated RTKs (Tables 1and 2). Nonsupervised clustering was used to evaluatepossible co-associations (Supplementary Tables S4 andS5). Without pervanadate, there was some co-associationof IGF1R and TYRO3 phosphorylation, with a subgroupshowing common activation with IR, FGFR3, and EPHB2(Supplementary Table S4). With pervanadate treatment,activation of multiple RTKs across cell lines was morecommon (Supplementary Table S5). Recently, a relatedset of primary array data analyzing 15 separate melanomalines was reported. This dataset identified active RTKs such






as HER3, IGF1R, and TYRO3 in non-pervanadate–treatedsamples, but these results were not validated (23).

Validation of array dataWe focused on validating TYRO3, MERTK, AXL,

HER3, MET, KIT, and IGF1R because they were eitheramong the most common or most novel active RTKsidentified and because they represent attractive cancertherapeutic targets (Fig. 1C and D). With the immuno-blotting reagents available, we were able to determine theabundance of TYRO3 and MERTK, but could not assesstheir phosphorylation status for comparison with the phos-pho-array data. Although high levels of phosphorylation inpervanadate-treated cells permitted analysis of total celllysates, immunoprecipitated phosphorylated RTKs couldalso be detected in untreated cells. The pattern of receptoractivation in the melanoma cell lines analyzed in Figure 1did not vary substantially with pervanadate treatment(Fig. 1C and D). We considered the immunoblots moreaccurate in the instances where the arrays did not correlatewith the immunoblots. Hence, the RTK arrays provided auseful, but imprecise, indication of which RTKs are activein a given cell line.

RTK activation in melanomaTo identify receptors that might be activated through

autocrine circuits or overexpression, we compared thephosphorylation levels determined by phospho-array ana-lysis to the transcriptional levels of the RTKs and theirrespective ligands. In untreated cell lines, a few RTKphosphorylations correlated with ligand mRNA expressionincluding FGFR3::FGF13, AXL::GAS6, and HER4 withassorted ligands (Supplementary Table S6). Similarly, inpervanadate-treated cells, moderate correlations were foundbetween RTK phosphorylation and ligand expression forAXL::GAS6, EPHA2::EFNA3, and FGFR3::FGF13. Inboth datasets, receptor phosphorylation was moderatelyassociated with RTK mRNA expression for AXL, whereasERBB4 and EGFR also showed correlations in untreatedcells and KIT mRNA expression was correlated with KITactivation in pervanadate-treated samples (SupplementaryTable S6).The small number of cases for which receptor phosphor-

ylations were identified in association with high ligandexpression represent candidates for autocrine loops. Here,the limited sensitivity of the RTK arrays coupled with thepossible chronic downregulation of ligand-activated RTKs

A B

with PVno PV NBMEL YUHEIK YULAC

MST1R

EGFR

KIT

IGF1R

YUDOSO with PV

MST1R

FGFR2a

RET

TYRO3

YU

CA

SY

UD

OS

OY

UH

EF

YU

KS

IY

ULA

CY

UP

LAY

UR

IFY

US

IK

D

YU

CA

SY

UD

OS

OY

UH

EF

YU

KS

IY

ULA

CY

UP

LAY

UR

IFY

US

IKIg

G

pAXLAXL

pHER3HER3

pIGF1RIGF1R

pKITKIT

pMETMET

TYRO3MERTKGAPDH

IP

IP

IP

IP

TCL

TCL

TCL

TCL

TCL

TCL

C

TCL

without PV with PV

pAXLAXL

pHER3HER3

pIGF1RIGF1R

pKITKIT

pMETMET

TYRO3MERTKGAPDH

Figure 1. Analysis of RTKphosphorylation. A, comparison ofYUDOSO cell lysates incubated inthe absence (left) or presence(right) of pervanadate (PV) prior tolysis. Each pair of vertical spotsrepresents one receptor. Pairedspots in the 4 corners are positivePTyr controls, and appear darkerat left because of the longer filmexposure time required for thelower signals without pervanadatetreatment. B, comparison of RTKactivation profiles amongpervanadate-treated NBMEL,YUHEIK, and YULAC cells. Spotsrepresenting specific receptorsare boxed in each panel. C and D,lysates from nontreated (C) andpervanadate-treated (D)melanoma cell lines were analyzedby immunoblotting for total andTyr-phosphorylated receptors.Total cell lysates (TCL) wereanalyzed by immunoblotting withanti-RTK and anti–phospho-RTKantibodies. RTK IP were analyzedwith anti–phospho-RTK, withcontrol immunoglobulin (IgG) IP asnegative control. All immunoblotshave been cropped.

Tworkoski et al.





means that there may be many false-negatives. Additionally,elevated expression can move the designated RTK abovephospho-array threshold for detection, which may skew theresults of our screen. The co-associations of RTK phos-phorylation with high receptor mRNA levels may, however,identify cases for which RTK overexpression contributes toreceptor activation.The strongest correlations were found in comparison

of receptor and ligand mRNA expression levels. For asubset of RTK/ligand pairs, including FGFR3 withseveral of its ligands, PDGFRb::PDGFb, AXL::GAS6,and ERBB4::neuregulin 3 (NRG3) NRG3, there werestrong associations of mRNA levels (SupplementaryTable S6). These RTK/ligand pairs represent candidatesfor autocrine activation, although for some pairs subcellularcompartmentalization or the need for proteolytic activationof prohormones may limit RTK activity.

RTK knockdown influences melanoma cellproliferation, migration, and signalingThe phospho-array survey identified IGF1R as active in

most of the melanoma cell lines (Tables 1 and 2). Becausecells were maintained in medium supplemented withserum, which is rich in ligand IGF1, we explored whetherIGF1R phosphorylation was affected by serum concentra-tion. Three melanoma cell lines were serum-starved for a

period of 24 to 48 hours and immunoblotting was used tomeasure the presence of total and phosphorylated IGF1R(Fig. 2A). The persistence of IGF1R phosphorylation indi-cates that serum IGF1 is not the primary factor influencingIGF1R activation.The high prevalence of IGF1R activation in our panel of

melanoma cell lines suggested that it might be functionallyimportant. Four melanoma cell lines characterized by phos-phorylated IGF1R were infected with empty vector(pLKO), scrambled shRNA (SC), or shRNA against IGF1R(Tables 1 and 2; Fig. 2B). IGF1R knockdown correspondedwith decreased cell proliferation and a reduction in AKTsignaling, in accordance with earlier studies (ref. 6; Fig. 2Band C).AXL and HER3 were selected for further validation of the

data generated by the phospho-array screen. We focused onthese 2 receptors because they are relatively unexamined inthe context of melanoma pathogenesis and because TAMfamily members are poorly characterized in cancers ingeneral. Lentiviral vectors encoding shRNA directed againstAXL or HER3 were used to knock down the respectivereceptors singly and in combination (Fig. 3A). Thegrowth of YURIF cells, which do not express AXL orHER3 (Tables 1 and 2; Fig. 1C and D), was not affectedby lentiviral-mediated knockdown of either RTK (Fig. 3B).In contrast, knockdown of AXL or HER3 strongly inhibited

Figure 2. IGF1R activation inmelanoma cell signaling andproliferation. A, melanoma celllines were maintained in 0.1%FBS or 5% FBS for 24 or 48 hoursprior to lysis. Lysates wereanalyzed for phospho-IGF1R,IGF1R, and the GAPDH loadingcontrol. Three biological replicatesare shown. Image J analysis wasused to quantify phospho-IGF1Rrelative to GAPDH for eachsample. Average ratios � 1 SD forthe samples shown, from left toright, are as follows: 0.78 � 0.29,1.1 � 0.08, 0.79 � 0.29, 0.81 �0.16, 0.76 � 0.07, 0.76 � 0.01,0.77 � 0.09, 0.69 � 0.24, 0.81 �0.26, 0.81 � 0.21, 0.50 � 0.01,0.47 � 0.26. B, expression ofphospho-IGF1R, IGF1R,phospho-AKT, and AKT inmelanoma cells infected withlentiviruses encoding backboneonly (pLKO), scrambled shRNA(SC), or shRNA for IGF1R. Allimmunoblots have been croppedto improve readability. C, growthof lentivirus-infected cellsassessed over a 72-hour period byusing the CellTiter-GloLuminescent Cell Viability Assay. 0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

YUHEF YUGEN8 YULAC YUTICA

T72

/T0 Mock

pLKO

SC

IGF1R

A B

C

5FB

S 2

4 h

0.1F

BS

24

h5F

BS

48

h0.

1FB

S 4

8 h

5FB

S 2

4 h

0.1F

BS

24

h5F

BS

48

h0.

1FB

S 4

8 h

5FB

S 2

4 h

0.1F

BS

24

h5F

S 4

8 h

0.1F

BS

48

h

moc

kpL

KO

SC

IGF

1R

moc

kpL

KO

SC

IGF

1R

moc

kpL

KO

SC

IGF

1R

moc

kpL

KO

SC

IGF

1R

pIGF1RIGF1R

pAktAkt

GAPDH

Cell proliferation

YUKSI YULAC YUHEF YUKSI

YULAC YUTICA

pIGF1RIGF1R

GAPDHpIGF1R

IGF1RGAPDHpIGF1R

IGF1RGAPDH

YUTICA

pIGF1RIGF1R

pAktAkt

GAPDH

pIGF1RIGF1R

pAktAkt

GAPDH

pIGF1RIGF1R

pAktAkt

GAPDH






proliferation of YUDOSO cells (Fig. 3B), which expressactive AXL and HER3 (Fig. 1C and D). A milder reductionin proliferation was obtained by targeting active HER3 inYUROL cells (Tables 1 and 2; Fig. 3B). YUKSI cells, whichexpress AXL but have low AXL activity (Tables 1 and 2;Figs. 1C and D, and 3A), displayed only slight reductions inproliferation andmigration upon AXL knockdown (Fig. 3D

and E). These alterations in proliferation and migrationcorrelated with a slight reduction in phospho-STAT3(Fig. 3A). Knockdown of active AXL in YUSIV, YUSTE,and YUTICA melanoma cells evoked a more pronounceddecrease in cell proliferation, cell migration, and STAT3phosphorylation (Tables 1 and 2; Fig. 3A, D, and E).Sporadic reductions in phospho-AKT were also observed

0

20

40

60

80

100

120

YUKSI YUSIV YUSTE YUTICA

% M

igra

tio

n

Mock

SC

AXL

0.00

1.00

2.00

3.00

4.00

5.00

YUKSI YUSIV YUSTE YUTICA

T72

/T0

Mock

pLKO

SC

AXL

0.00

1.00

2.00

3.00

4.00

5.00

YUROL

T72

/T0

Mock

pLKO

SC

HER3

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

YUDOSO YURIF

T72

/T0

Mock

pLKO

SC

AXL

HER3

AXL+HER3

A B

C

D E

moc

kpL

KO

SC

AX

LH

ER

3A

XL+

HE

R3

AxlHER3

GAPDH

moc

kpL

KO

SC

HE

R3

moc

kpL

KO

SC

AX

L

moc

kpL

KO

SC

AX

L

moc

kpL

KO

SC

AX

L

moc

kpL

KO

SC

AX

L

HER3GAPDH

pAxlAxl

pSTAT3STAT3

GAPDH

Cell proliferation

Cell proliferation

Cell proliferation Cell migration

YUDOSO YUROL

YUSIV YUKSI

YUSTE YUTICA

pAxlAxl

pSTAT3STAT3

GAPDH

pAxlAxl

pSTAT3STAT3

GAPDH

pAxlAxl

pSTAT3STAT3

GAPDH

Figure 3. Impact of AXL and HER3 knockdown on cell proliferation, migration, and signaling. A, expression of phospho-AXL (pAXL), AXL, HER3,phospho-STAT3 (pSTAT3), and STAT3 with GAPDH loading control in melanoma cells infected with empty backbone (pLKO), scrambled shRNA (SC),or shRNA for AXL and/or HER3. Immunoblots have been cropped. B–D, growth of lentivirus-infected cells assessed over a 72-hour period by the CellTiter-GloLuminescent Cell Viability Assay. Melanoma cells are grouped on the basis of the viruses used; cell lines in the same panel were generally analyzed indifferent experiments. E, migration of melanoma cells infected with control (SC) or AXL shRNA lentivirus relative to mock-infected cells. AXL knockdownwas achieved by AXL 1 shRNA sequence.

Tworkoski et al.





on AXL knockdown (data not shown). Further validation ofthese effects was obtained by targeting AXL by using asecond shRNA in 4 melanoma cell lines (SupplementaryFig. S1). These results support the biological relevance ofAXL and HER3 in melanoma cells and substantiate theimportance of targeting activated receptors.

Melanoma cell lines predict presence of RTKs intumorsTo verify that expression of the RTKs identified in our

screen is not a cell culture artifact, snap-frozen tissuesderived from the same original tumors were analyzed viaqRT-PCR. AXL mRNA was identified by transcriptionalprofiling in YUDOSO, YUSIV, and YUSTE, but not inYUCOT melanoma cells (Table 3). AXL mRNA andprotein were variably detected in YULAC cells (Fig. 1Cand D; Table 3). qRT-PCR analysis of frozen tumor tissuesrevealed that YUDOSO, YUSIV, YUSTE, and YULACtumors all express AXL, but the YUCOT tumor did not(Fig. 4A). Similarly, HER3 mRNA was detected by tran-scriptional profiling of the YUCOT, YULAC, YUDOSO,YUSIV, and YUSTE melanoma cell strains as well as byqRT-PCR analysis of tissue from the corresponding tumors(Table 3; Fig. 4B).As expected, there was no direct correlation between the

activity of an RTK in a cell line (Tables 1 and 2) and thepresence of mRNA in the corresponding tumor (Fig. 4Aand B). Nonetheless, activated AXL and HER3 were onlydetected in melanoma cell lines for which the tumorsexpressed the corresponding receptors. The differences inrank expression of the RTKs in tumors versus cell lines mayreflect tumor to cell strain differences, tumor heterogeneity,or differential infiltration by nontumor cells in the tumormicroenvironment.

Discussion

We report here the first broad survey of functionallyactivated RTKs in a panel of early-passage melanoma cellstrains. This survey represents one facet of an integrated

analysis of a reference set of melanoma cell lines and corre-sponding tumor tissue that will eventually encompass tran-scriptionprofiling, epigenetic analysis, copynumber analysis,transcriptome and exome resequencing, and drug responseprofiling.We found that the insulin receptor family (IGF1R,IR), MET family (MET, MST1R), EGFR family (EGFR,HER3,HER4), and TAM family (TYRO3, AXL,MERTK)were commonly activated in melanoma (Tables 1 and 2).IGF1R was frequently and strongly activated in our

panel of cell strains, and this activation did not seem to

Table 3. Receptor expression and activation in cell lines

Activationof AXL on array

Transcriptionalexpression of AXL

Activation ofHER3 on array

Transcriptionalexpressionof ERBB3

Cell line �PV þPV �PV þPV

NBMEL 0 0 842 0 1 943YUCOT 0 0 466 0 5 15,164YUDOSO 0 3 2,500 2 5 21,188YULAC 0 0 425 0 5 12,623YUSIV 0 1 8,327 0 0 9,717YUSTE 0 1 8,351 0 5 10,251

NOTE: Whole genome RNA expression values determined for AXL (NM_021913) and ERBB3 (HER3; NM_001982) in melanoma celllines. Receptor activation values are from RTK capture array analyses of pervanadate-treated (þPV) and nontreated (�PV) cells.

A

B

01,0002,0003,0004,0005,0006,0007,0008,0009,000

mR

NA

rel

ativ

e to

n

egat

ive

con

tro

lm

RN

A r

elat

ive

to

neg

ativ

e co

ntr

ol

AXL expression

0

100

200

300

400

500

600ERBB3 expression

RIF-c

ell

NBMEL-

cell

COT-tu

mor

DOSO-tum

or

LAC-tu

mor

SIV-tu

mor

STE-tum

or

MB23

1-ce

ll

NBMEL-

cell

COT-tu

mor

DOSO-tum

or

LAC-tu

mor

SIV-tu

mor

STE-tum

or

Figure 4. Analysis of RTK mRNA in tumor samples. A and B, relativeexpression of AXL and ERBB3 mRNA from tumors. AXL and HER3expression were measured by qRT-PCR and normalized to GAPDH, andthe resulting expression is shown relative to negative controls for AXL(YURIF) and HER3 (MB231).






be serum-dependent (Tables 1 and 2; Fig. 2A). A previousreport showed that knockdown of IGF1R can greatly reducemelanoma cell growth and that targeting IGF1R reducedsignaling through both the AKT and the mitogen-activatedprotein kinase pathways (6). Our data also support thenotion that knockdown of IGF1R may reduce cell growthin an AKT-dependent manner (Fig. 2). Activation ofIGF1R promotes liver metastases in uveal melanoma andactivated IGF1R, along with activated HER3, is known tocontribute to Herceptin resistance in breast cancer (24–27).Signaling through IGF1R also compensates for therapeuticinhibition of mTOR in breast and prostate cancer cells (28).It is possible that activated IGF1R might serve an analogousrole in melanoma and it will therefore be importantto determine the mechanism of IGF1R activation inmelanoma.A recent study showed that HER4 may be somatically

mutated in up to 19% of melanomas, and our survey revealsfrequent activations of the EGFR family members inmelanoma (29). While this work was in progress, elevatedHER3 expression and activation was reported in melanomacells relative to normal melanocytes, with knockdown ofHER3 reducing melanoma cell proliferation, migration,and invasion (30–31). Reduced HER3 expression was alsocorrelated with reduced phospho-AKT and increased p27expression (31). The findings presented here are consistentwith these reports, because our melanoma cell strainsfrequently show higher HER3 activation and mRNA thanmelanocytes (Tables 1 and 2; Supplementary Tables S1 andS2) and because HER3 knockdown reduces melanoma cellproliferation (Fig. 3). It has been proposed that activation ofHER3 by NRG1 contributes to melanoma developmentand progression (30), but we did not identify a correlationbetween HER3 activation and the expression of ligandsNRG1 or NRG2 (Supplementary Table S6). One explana-tion may be that NRG1 is important in early stages ofmelanoma but its expression does not persist in later stages;this is consistent with the reported higher frequency ofactivated HER3 in primary compared with metastaticmelanomas (30). The phospho-RTK array analysis alsoreveals coactivation of MET with EGFR family memberssuch as HER3 (Tables 1 and 2). Because coordinateactivation of MET with ERBBs often modulates resistanceto EGFR inhibitors in NSCLC, combined inhibition ofMET and ERBBs may be beneficial for a subset of mela-nomas as well (32).The TAM receptor family is well-studied in inflamma-

tion but is not yet completely characterized in carcinogen-esis. Previous work shows that AXL mRNA is elevated inmelanomas relative to normal melanocytes, in agreementwith our transcriptional data (Supplementary Tables S1 andS2; ref. 33). Although TYRO3 and MERTK were reportedto be preferentially expressed in melanomas, we found thatTYRO3 is highly expressed in both melanocytes and mel-anomas and thatMERTK expression is, on average, lower inmelanoma cells compared with normal melanocytes (Sup-plementary Tables S1 and S2; refs. 33–35). Interestingly,the phospho-arrays revealed that at least one member of the

TAM family was activated in the majority of the melanomacell lines, even without pervanadate treatment (Tables 1 and2). Further, activation of AXL and MERTKmay have somedegree of mutual exclusivity, because AXL and MERTKactivation rarely occurred concurrently in the melanoma celllines.Suppression of TYRO3 reduces melanoma cell prolifera-

tion, sensitizes cells to chemotherapeutics, decreases colonyformation, and reduces tumor formation and growth (34).Further, AXL is required for GAS6-dependent uveal mel-anoma cell survival (36). In our study, knockdown ofactivated AXL noticeably affects cell proliferation in aprimary cutaneous melanoma line (YUDOSO), 3 meta-static cutaneous lines (YUGEN8, YUSIV, YUSTE), and 2cutaneous lines that had metastasized to the lung (YUHEF,YUTICA; Fig. 3; Supplementary Fig. S1). Targeting AXLin a subset of these cell lines also reduces cell migration(Fig. 3). It appears that this decrease in migration andproliferation correlates with a reduction in STAT3 signaling(Fig. 3; Supplementary Fig. S1). Thus, AXL signaling maycontribute to the constitutive activation of STAT3 that isfrequently observed in human melanoma cell lines andtumors (37).Some of the RTKs identified by our screen (IGF1R,

MET, and KIT) have been previously described in mela-nomas. Others, including HER3, HER4, members of theTAM family, EPHB2, and MST1R, have either not beenreported as active, or are poorly characterized in this disease.These findings have therapeutic implications because someof the receptors identified in the melanoma cell lines,specifically AXL and HER3, were found in the correspond-ing tumor tissue (Table 3; Fig. 4). Targeting AXL in a cellline that either does not express phosphorylated AXL(YUKSI) or in a cell line that does not express the receptorat all (YURIF) had little to no effect on cell behavior (Fig. 3).Interestingly, targeting the activated receptors appears to beeffective regardless of NRAS and BRAF mutation status.Both IGF1R and AXL knockdown alters cell behavior inBRAF/NRAS wild-type (WT) cells (YUHEF, YUSIV),NRAS mutant cells (YUDOSO, YUTICA), and BRAFmutant (YUGEN8, YUKSI, YULAC, YUSTE) melanomacell lines. Although HER3 knockdown was not examined ina BRAF mutant line, it is effective in NRAS (YUDOSO)mutant and WT (YUROL) cell lines (Fig. 3). These find-ings are important because expression and activation ofRTKs was largely independent of mutation status, andbecause activation of nonmutated RTKs is a commonmechanism of resistance in tumor cells attacked withRTK or pathway-directed therapeutics (38). In fact, in anonbiased survey of protein kinase cDNAs for the ability topromote melanoma cell resistance to the activated BRAFinhibitor PLX4720, AXL was identified as the RTK bestable to rescue cell proliferation (39). Another study identi-fied upregulation and activation of PDGFRb receptor as aroute to resistance in BRAFmutated melanoma treated withPLX4032 (40). Perhaps coexpression of ligand–receptorconjugate pairs, as we found for several RTKs includingPDGFRb, foreshadows routes to resistance.

Tworkoski et al.





These findings document variability in RTK expressionand activation among melanomas. Identification of cancerdrivers by gene mutation analysis is attractive because thetechnology is accurate and sensitive and because muta-tions in protein kinases and other signaling proteins canhave strong predictive value. Nonetheless, carcinogenicalterations in RTKs and signaling pathways are alsoinduced by changes in expression, subcellular localization,and proteolysis of ligands. Because RTKs are mainlyregulated through protein–protein interactions and pro-tein modifications, functional analyses that measure pro-tein changes continue to provide important insights intocancer at the discovery and treatment levels. For instance,one of the early indications of anaplastic lymphoma kinase(ALK) activation in a subset of lung cancers was derivedfrom a mass spectrometry survey of lung cancer tumorsand cell lines (41). The fact that HER3, AXL, and IGF1Rknockdown was effective regardless of NRAS and BRAFstatus suggests that inhibition of RTKs may be valuable incombination with RAF or MEK inhibitors (Fig. 3). TheRTKs identified in this survey may provide new thera-

peutic opportunities for driver inhibition while improvingour ability to prevent therapeutic resistance.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were declared.

Acknowledgments

Fibroblasts, melanocytes, and melanoma cell lines, as well as the NimbleGen geneexpression data were provided by the Specimen Resource Core of the Yale SPORE inSkin Cancer. We thank James McCusker for hierarchical clustering analysis.

Grant Support

This work was funded in part by the NSF Graduate Research Fellowship Program(K. Tworkoski), the Harry J. Lloyd Charitable Trust (D.F. Stern), and the YaleSPORE in Skin Cancer, funded by the National Cancer Institute Grant 1 P50CA121974 (R. Halaban, principal investigator).

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

Received November 15, 2010; revised March 17, 2011; accepted April 13, 2011;published OnlineFirst April 26, 2011.

References1. Califano J, Nance M. Malignant melanoma. Facial Plast Surg Clin

North Am 2009;17:337–48.2. Melanoma Skin Cancer Overview [Internet]. American Cancer

Society. 2010 [Cited 2010 April 1]. Available from: http://www.cancer.org/Cancer/SkinCancer-Melanoma/OverviewGuide/index.

3. Hersey P, Bastholt L, Chiarion-Sileni V, Cinat G, Dummer R, Egger-mont AM, et al. Small molecules and targeted therapies in distantmetastatic disease. Ann Oncol 2009;20:vi35–40.

4. Hodi FS, Friedlander P, Corless CL, Heinrich MC,Mac Rae S, Kruse A,et al. Major response to imatinib mesylate in KIT-mutated melanoma.J Clin Oncol 2008;26:2046–51.

5. Puri N, Ahmed S, Janamanchi V, Tretiakova M, Zumba O, Krausz T,et al. c-Met is a potentially new therapeutic target for treatment ofhuman melanoma. Clin Cancer Res 2007;13:2246–53.

6. Yeh AH, Bohula EA,Macaulay VM. Humanmelanoma cells expressingV600E B-RAF are susceptible to IGF1R targeting by small interferingRNAs. Oncogene 2006;25:6574–81.

7. Terheyden P, Houben R, Pajouh P, Thorns C, Zillikens D, Becker JC.Response to imatinib mesylate depends on the presence of theV559A-mutated KIT oncogene. J Invest Dermatol 2010;130:314–6.

8. Jiang X, Zhou J, Yuen NK, Corless CL, Heinrich MC, Fletcher JA, et al.Imatinib targeting of KIT-mutant oncoprotein in melanoma. Clin Can-cer Res 2008;14:7726–32.

9. Kenessey I, Keszthelyi M, Kramer Z, Berta J, Adam A, Dobos J, et al.Inhibition of c-Met with the specific small molecule tyrosine kinaseinhibitor SU11274 decreases growth and metastasis formation ofexperimental human melanoma. Curr Cancer Drug Targets2010;10:332–42.

10. Halaban R, ZhangW, Bacchiocchi A, Cheng E, Parisi F, Ariyan S, et al.PLX4032, a selective BRAF(V600E) kinase inhibitor, activates the ERKpathway and enhances cell migration and proliferation of BRAFmelanoma cells. Pigment Cell Melanoma Res 2010;23:190–200.

11. Hoek K, Rimm K, Williams KR, Zhao H, Ariyan S, Lin A, et al.Expression profiling reveals novel pathways in the transformationof melanocytes to melanomas. Cancer Res 2004;64:5270–82.

12. HalabanR, KrauthammerM, PelizzolaM,Cheng E, KovacsD, SznolM,et al. Integrative analysis of epigenetic modulation in melanoma cellresponse to decitabine: clinical implications. PLoSOne 2009;4:e4563.

13. Graeber TG, Eisenberg D. Bioinformatic identification of potentialautocrine signaling loops in cancers from gene expression profiles.Nat Genet 2001;29:295–300.

14. Koorstra JB, Karikari CA, Feldmann G, Bisht S, Rojas PL, OfferhausGJ, et al. The Axl receptor tyrosine kinase confers an adverse prog-nostic influence in pancreatic cancer and represents a new therapeu-tic target. Cancer Biol Ther 2009;8:618–26.

15. Engelman JA, Janne PA, Mermel C, Pearlberg J, Mukohara T, Fleet C,et al. Cantley, ErbB-3 mediates phosphoinositide 3-kinase activity ingefitinib-sensitive non-small cell lung cancer cell lines. Proc Natl AcadSci U S A 2005;102:3788–93.

16. Riedemann J, Sohail M, Macaulay VM. Dual silencing of the EGF andtype 1 IGF receptors suggests dominance of IGF signaling in humanbreast cancer cells. Biochem Biophys Res Commun 2007;355:700–6.

17. Lutterbach B, Zeng Q, Davis LJ, Hatch H, Hang G, Kohl NE, et al. Lungcancer cell lines harboring MET gene amplification are dependent onMet for growth and survival. Cancer Res 2007;67:2081–8.

18. Kauraniemi P, Kallioniemi A. Activation of multiple cancer-associatedgenes at the ERBB2 amplicon in breast cancer. Endocr Relat Cancer2006;13:39–49.

19. Lemmon MA, Schlessinger J. Cell signaling by receptor tyrosinekinases. Cell 2010;141:1117–34.

20. Menendez JA, Lupu R. Transphosphorylation of kinase-dead HER3and breast cancer progression: a new standpoint or an old conceptrevisited? Breast Cancer Res 2007;9:111.

21. Gjerdrum C, Tiron C, Hoiby T, Stefansson I, Haugen H, Sandal T, et al.Axl is an essential epithelial-to-mesenchymal transition-induced reg-ulator of breast cancermetastasis and patient survival. Proc Natl AcadSci U S A 2010;107:1124–9.

22. Narayan M, Wilken JA, Harris LN, Baron AT, Kimbler KD, Maihle NJ.Trastuzumab-induced HER reprogramming in "resistant" breast car-cinoma cells. Cancer Res 2009;69:2191–4.

23. Easty DJ, Gray SG, O’Byrne KJ, O’Donnell D. Bennett DC. Receptortyrosine kinases and their activation in melanoma. Pigment CellMelanoma Res 2011. [Epub ahead of print].

24. All-Ericsson C, Girnita L, Seregard S, Bartolazzi A, Jager MJ, LarssonO. Insulin-like growth factor-1 receptor in uveal melanoma: a predictorfor metastatic disease and a potential therapeutic target. InvestOphthalmol Vis Sci 2002;43:1–8.

25. Topcu-Yilmaz P, Kiratli H, Saglam A, Soylemezoglu F, Hascelik G.Correlation of clinicopathological parameters with HGF, c-Met, EGFR,and IGF-1R expression in uveal melanoma. Melanoma Res 2010;20:126–32.






26. Browne BC, Crown J, Venkatesan N, Duffy MJ, Clynes M, Slamon D,et al. Inhibition of IGF1R activity enhances response to trastuzumab inHER-2-positive breast cancer cells. Ann Oncol 2011;22:68–73.

27. Huang X, Gao L, Wang S, McManaman JL, Thor AD, Yang X, et al.Heterotrimerization of the growth factor receptors erbB2, erbB3, andinsulin-like growth factor-i receptor in breast cancer cells resistant toherceptin. Cancer Res 2010;70:1204–14.

28. O’Reilly KE, Rojo F, She QB, Solit D, Mills GB, Smith D, et al. mTORinhibition induces upstream receptor tyrosine kinase signaling andactivates Akt. Cancer Res 2006;66:1500–8.

29. Rudloff U, Samuels Y. A growing family: adding mutated Erbb4 as anovel cancer target. Cell Cycle 2010;9. [Epub ahead of print].

30. Buac K, XuM, Cronin J, Weeraratna AT, Hewitt SM, PavanWJ. NRG1/ERBB3 signaling in melanocyte development and melanoma: inhibi-tion of differentiation and promotion of proliferation. Pigment CellMelanoma Res 2009;22:773–84.

31. Reschke M, Mihic-Probst D, Van Der Horst EH, Knyazev P, Wild PJ,Hutterer M, et al. HER3 is a determinant for poor prognosis inmelanoma. Clin Cancer Res 2008;14:5188–97.

32. Agarwal S, Zerillo C, Kolmakova J, Christensen JG, Harris LN, RimmDL, et al. Association of constitutively activated hepatocyte growthfactor receptor (Met) with resistance to a dual EGFR/Her2 inhibitor innon-small-cell lung cancer cells. Br J Cancer 2009;100:941–9.

33. Quong RY, Bickford ST, Ing YL, Terman B, Herlyn M, Lassam NJ.Protein kinases in normal and transformed melanocytes. MelanomaRes 1994;4:313–9.

34. Zhu S, Wurdak H, Wang Y, Galkin A, Tao H, Li J, et al. A genomicscreen identifies TYRO3 as a MITF regulator in melanoma. Proc NatlAcad Sci U S A 2009;106:17025–30.

35. Gyorffy B, Lage H. A Web-based data warehouse on gene expres-sion in human malignant melanoma. J Invest Dermatol 2007;127:394–9.

36. vanGinkel PR, Gee RL, Shearer RL, Subramanian L,Walker TM, AlbertDM, et al. Expression of the receptor tyrosine kinase Axl promotesocular melanoma cell survival. Cancer Res 2004;64:128–34.

37. Niu G, Bowman T, Huang M, Shivers S, Reintgen D, Daud A, et al.Roles of activated Src and Stat3 signaling in melanoma tumor cellgrowth. Oncogene 2002;21:7001–10.

38. McDermott U, Pusapati RV, Christensen JG, Gray NS, Settleman J.Acquired resistance of non-small cell lung cancer cells to MET kinaseinhibition is mediated by a switch to epidermal growth factor receptordependency. Cancer Res 2010;70:1625–1634.

39. Johannessen CM, Boehm JS, Kim SY, Thomas SR, Wardwell L,Johnson LA, et al. COT drives resistance to RAF inhibitionthrough MAP kinase pathway reactivation. Nature 2010;468:968–72.

40. Nazarian R, Shi H,WangQ, Kong X, Koya RC, Lee H, et al. Melanomasacquire resistance to B-RAF(V600E) inhibition by RTK or N-RASupregulation. Nature 2010;468:973–7.

41. Rikova K, Guo A, Zeng Q, Possemato A, Yu J, Haack H, et al. Globalsurvey of phosphotyrosine signaling identifies oncogenic kinases inlung cancer. Cell 2007;131:1190–203.

Tworkoski et al.





2011;9:801-812. Published OnlineFirst April 26, 2011.Mol Cancer Res Kathryn Tworkoski, Garima Singhal, Sebastian Szpakowski, et al. Targets in MelanomaPhosphoproteomic Screen Identifies Potential Therapeutic

Updated version

10.1158/1541-7786.MCR-10-0512doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2011/06/17/1541-7786.MCR-10-0512.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/9/6/801.full#ref-list-1

This article cites 37 articles, 16 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/9/6/801.full#related-urls

This article has been cited by 9 HighWire-hosted articles. Access the articles at:

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

SubscriptionsReprints and

[email protected] at

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

Permissions

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

.http://mcr.aacrjournals.org/content/9/6/801To 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-10-0512

http://mcr.aacrjournals.org/content/suppl/2011/06/17/1541-7786.MCR-10-0512.DC1

http://mcr.aacrjournals.org/content/9/6/801.full#ref-list-1

http://mcr.aacrjournals.org/content/9/6/801.full#related-urls

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/9/6/801


phosphoproteomic screen identifies potential therapeutic … · signaling and regulation...

Documents