Journal of Dermatological Science 59 (2010) 7–15

Re-epithelialization from human skin explant cultures is promoted byligand-activated HER3 receptor

Sofi Forsberg, Ola Rollman *

Department of Medical Sciences, Dermatology and Venereology, University Hospital, Uppsala University, SE-751 85 Uppsala, Sweden

A R T I C L E I N F O

Article history:

Received 25 August 2009

Received in revised form 20 March 2010

Accepted 26 March 2010

Keywords:

Tyrosine kinase receptors

ErbB3 protein

Neuregulin-1

Epidermis

Fluorescence microscopy

A B S T R A C T

Background: Ligand-stimulated epidermal growth factor receptor (EGFR/HER1) plays a fundamental role

in skin biology as potent transducer of mitotic and anti-apoptotic stimuli in keratinocytes. In human

epidermis, at least two additional EGFR family members – HER2 and HER3 – are expressed but their

biological functions in normal and diseased human skin remain obscure.

Objective: Here, we studied the expression and biological impact of HER3 in regenerating human

epidermis formed from skin explants adhered to acellular dermis.

Methods: Neoepidermal HER3 expression was examined by quantitative real-time reverse transcriptase

polymerase chain reaction, immunohistochemistry and Western blot analysis. The dynamic effect of

HER3 receptor stimulation by recombinant heregulin (HRG)-b1 was assessed by fluorescence imaging of

re-epithelialization.

Results: In the neoepidermis, HER3 mRNA and protein were detected with activated receptors being

immunolocalized at basal and low suprabasal levels. Exogenous HRG-b1 at 10–20 ng/ml increased the

outgrowth rate corresponding to approximately 30% the response of exogenous EGF. The growth-promoting

effect of HRG-b1 was associated with enhanced HER3 phosphorylation, keratinocyte proliferation and

thickening of viable neoepidermis whereas blockade of ligand-binding to HER3 delayed the outgrowth

process and inhibited both constitutive and ligand-induced HER3 phosphorylation. HER2 antagonism using

an anti-dimerization antibody, pertuzumab, impeded the re-epithelialization rate. In addition, a selective

HER2 kinase inhibitor, CP654577, downregulated phospho-HER3 expression suggesting that transactiva-

tion of kinase-deficient HER3 was accomplished through dimerization with HER2.

Conclusion: The study emphasizes the central role of EGFR in epidermal renewal and demonstrates that

HRG-activated HER3 contributes to the outgrowth process of epidermis in vitro.

� 2010 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights

reserved.

Contents lists available at ScienceDirect

Journal of Dermatological Science

journa l homepage: www.e lsev ier .com/ jds

1. Introduction

Renewal of epithelial tissues is dependent on growth factor-mediated cell signalling from transmembrane type I receptortyrosine kinases. In human skin, the epidermal growth factorreceptor EGFR (synonymous to HER1 or ErbB1) represents the mostthoroughly studied member within the EGFR superfamily(reviewed in Jorissen et al. [1]). Since its discovery, the pivotalrole of EGFR in support of epidermal cell proliferation, motility andsurvival has been firmly established [2]. Its contribution toepidermal cell growth is particularly noticeable in vitro becauseat least 90% of autocrine keratinocyte proliferation results from

Abbreviations: BrdU, 5-bromodeoxyuridine; DED, de-epidermized dermis; EGFR,

epidermal growth factor receptor; HER, human epidermal growth factor receptor;

HRG, heregulin; PBS, phosphate buffered saline; RT-qPCR, reverse transcription

quantitative real-time PCR.

* Corresponding author. Tel.: +46 18 6115086; fax: +46 18 6112680.

E-mail address: ola.rollman@medsci.uu.se (O. Rollman).

0923-1811/$36.00 � 2010 Japanese Society for Investigative Dermatology. Published b

doi:10.1016/j.jdermsci.2010.03.017

ligand-activated signalling from EGFR [3]. By contrast, theimportance of other EGFR family receptors, i.e. HER2 (ErbB2,neu), HER3 (ErbB3) and HER4 (ErbB4) in epithelial developmenthas gained far less attention in dermatological research. Althoughexpressed by differentiated epidermal keratinocytes, neither HER2nor HER3 has yet been tied to defined biological functions inhuman skin [4–7]. Moreover, HER4 is generally claimed to beexpressed at low or undetectable levels in normal adult epidermis[6–9].

In the present in vitro study, we analyzed the expression anddynamic effects through ligand-stimulated HER3 in culturedhuman epidermis using heregulin (HRG)-b1 as agonist. Thisligand belongs to a class of EGF-related glycoproteins [10] with keybiological functions during development of the heart, breast andnervous system [11]. Earlier attempts to evaluate the biologicalrole of HRGs (neuregulins) in skin keratinocytes are sparse andassociated with variable outcomes related to cell origin andspecific isoform selected. Among the HRGs, the b1-variant hasstrong HER3–4 binding affinity and high overall potency relative to

y Elsevier Ireland Ltd. All rights reserved.

S. Forsberg, O. Rollman / Journal of Dermatological Science 59 (2010) 7–158

the a-isoforms [12]. In mouse keratinocytes, HRG-b1 represents amitogenic factor [13], whereas in monolayer-cultured humankeratinocytes it yields no proliferative activity [14]. Furthermore,AU-565 mammary cancer cells reacts in a dose-dependent way toHRG stimulation since low concentrations support proliferativeresponses and high levels of ligand causes growth inhibition andinduction of differentiation markers [15]. The relative expressionof EGFR family members is another decisive factor for HRG-b1bioactivity. Breast cancer cells expressing high HER3:HER4 ratioseem disposed to HER3 activation, increased proliferation andreduced apoptosis [16]. In contrast, cells with high HER4expression respond to HRG-b1 treatment by HER4 activation,reduced proliferation and increased cell differentiation [17,18].

The current study shows that newly formed human epidermisexpresses HER1–4 mRNA, and that neoepidermal HER3 protein isconstitutively phosphorylated under basal culture conditions, i.e.

in presence of serum-containing medium. HRG-b1-stimulation ofneoepidermal cultures reinforces HER3 activation whereas expo-sure to antibodies interfering with ligand-binding to HER3 reducesreceptor phosphorylation and neoepidermal outgrowth. Theresults from experiments using anti-HER2-directed strategiessuggest that kinase-dead HER3 is transactivated by HER2 kinase.The study demonstrates that ligand-induced HER3–HER2 hetero-dimers are involved in the process of skin re-epithelialization.

2. Materials and methods

2.1. Skin explant culture, BrdU labelling and imaging of re-

epithelialization

Skin punch biopsies (2 mm diameter) were obtained fromsurplus normal tissue in conjunction with mammoplastic surgeryin women aged 40–63 years. Each biopsy was adhered to the centreof a 10-mm disc of de-epidermized dermis (DED) [19,20] aspreviously described [21]. In each well of a 6-well culture plate, 2explanted DEDs (4 in anti-HER3 culture series) were placed on amodified CellStrainer1 (70 mm pore size; Falcon, BD Biosciences,Bedford, MA) and grown in basal medium (classic keratinocytemedium used in one case) for 9 days at the air–liquid interface asdescribed [22,23].

‘Basal medium’ comprised Dulbecco’s Modified Eagle’s Medium(DMEM):Ham’s F12 medium (3:1), 100 mg/ml streptomycin,100 U/ml penicillin (PEST), nonessential amino acids (all Gibco,Paisley, Scotland, UK), 10% fetal bovine serum (FBS, PAALaboratories, Pasching, Austria) and 0.18 mM adenine (Sigma–Aldrich, St. Louis, MO). ‘Classic keratinocyte medium’ was basalmedium supplemented with 5 mg/ml insulin (Lilly, Solna, Sweden),10�10 M cholera enterotoxin and 0.5 mg/ml hydrocortisone (bothSigma–Aldrich). Medium renewal was after 72 h and then at 48-hintervals throughout the culture period. In experiments withexogenous ligand, EGF (from mouse submaxillary gland, Sigma–Aldrich) and recombinant HRG-b1 (Sigma–Aldrich) at 5, 10 or20 ng/ml, or vehicle (PBS) was added to the growth medium ateach medium renewal, or at 10 ng/ml during the final 1.5–2 h ofculture. Anti-HER3 antibody (mouse monoclonal, Ab-5, cloneH3.105.5, without BSA and azide, Thermo Scientific, Fremont, CA)was used at 10 mg/ml, cetuximab (Merck, Darmstadt, Germany)and pertuzumab (Genentech, San Francisco, CA) were used at100 nM (�15 mg/ml) during the entire culture period. Proliferativecells were labelled by incubation with 30 mM bromodeoxyuridine(BrdU, Roche Diagnostics, Mannheim, Germany) 4 h prior toharvest [23].

Prior to RNA and protein sampling, the cultures were washedthrice with ice-cold PBS. The explant was removed and theremaining neoepidermis scraped off from the DED using a dermalcurette. Samples grown under identical conditions were pooled.

For protein analysis, the detached neoepidermis was pelletted andfrozen at �70 8C until lysation.

The area of re-epithelialization was traced before each mediumrenewal (days 3, 5, 7 and 9) by fluorescence imaging of re-epithelialization (FIRE) technique [21] using fluorescein diacetate(Sigma–Aldrich) as viability indicator [24]. Neoepidermal out-growth (total fluorescent area minus explant area) was quantifiedusing a DP-soft image analysis program (Olympus, Hamburg,Germany).

Each series included explants and DEDs, respectively, from acommon donor. The study was approved by the local ethicscommittee at Uppsala University and performed according to theDeclaration of Helsinki Principles. Written informed consent wasobtained from all donors.

2.2. Immunohistochemistry and histometry

Tissue samples were fixed in 4% neutral-buffered formaldehydeprior to dehydration and paraffin embedment. Deparaffinizedcross-sections (5 mm) were immunohistochemically stained usingthe following primary antibodies: mouse anti-BrdU (1:200, cloneBu20a, Dako, Glostrup, Denmark), mouse anti-involucrin (1:500,clone Sy5, Nordic Biosite, Taby, Sweden), rabbit anti-loricrin(1:500, Abcam, Cambridge, UK), mouse anti-filaggrin (1:200, clone15C10, Novocastra, Newcastle upon Tyne, UK), rabbit anti-pHER3(Tyr 1289, 1:250, clone 21D3, Cell Signaling Technology) andmouse anti-HER3 (1:100, clone DAK-H3-IC, Dako). Enzymaticepitope retrieval (for involucrin and loricrin) was in 0.05% proteasetype VIII (Sigma–Aldrich) solution for 10 min at 37 8C. Heat-induced epitope retrieval (for BrdU and filaggrin) was performedusing a pressure cooker with slides in 10 mM citrate buffer pH 6.0or Reveal (Biocare Medical, HistoLab). For pHER3 and HER3staining, slides were run in 1 mM EDTA pH 8.0. Endogenousperoxidase was blocked in 0.6% H2O2 in methanol for 15 min or inPeroxidazed 1 (Biocare Medical) for 5 min. Nonspecific bindingwas blocked by pre-incubation with 10% normal serum (horse/goat, Vector Laboratories, Burlingame, CA) or Background sniper(Biocare Medical). Primary antibodies were incubated overnight at4 8C or 1–2 h at room temperature. Biotinylated anti-mouse oranti-rabbit secondary antibodies (1:200, 30 min, Vector Laborato-ries) were detected with ABC technique using diaminobenzidine(DAB, Vector laboratories) as chromogenic substrate. For pHER3and HER3 detection, the MACH3 system (rabbit and mouse,respectively, Biocare Medical) was used after primary antibodies,and Vulcan Fast Red Chromogen kit (Biocare Medical) was used forvisualization. Slides were counterstained with hematoxylin andmounted in glycerol jelly. The staining procedures includednegative controls omitting the primary antibody. Proliferative celldensity was defined as number of BrdU-positive cells per mmlength of the basal membrane zone (BMZ).

Histometric analysis was by Zeiss Axiovision LE version 4.4image analysis software (Carl Zeiss AB, Stockholm, Sweden) andincluded ‘neoepidermal thickness’, i.e. total area of viableneoepidermis/horizontal length of neoepidermis and ‘papilloma-tosis index’, i.e. length of the BMZ/horizontal length of theneoepidermis [21].

2.3. RNA preparation and cDNA synthesis

Total RNA was prepared from neoepidermis using TRIzol Reagentaccording to the manufacturers’ instructions (Invitrogen, Carlsbad,CA) including 45 mg GlycoBlue (Ambion Ltd, Huntingdon, UK).Homogenization was by frequent pipetting until the tissue wasdissolved. The RNA was resuspended in DEPC water (Invitrogen),quantified using a ND1000 spectrophotometer (NanoDrop Technol-ogies, Inc., Wilmington, DE) and stored at �70 8C.

S. Forsberg, O. Rollman / Journal of Dermatological Science 59 (2010) 7–15 9

First-strand complementary DNA (cDNA) was generated byreverse transcription using 3 mg total RNA mixed with an oligo-d(T)16 primer (75 pmol, Applied Biosystems, Foster City, CA)adjusting the volume to 20.6 ml with DEPC water. After the firstannealing step (70 8C for 10 min in a PTC-200 apparatus, Bio-RadLaboratories, Hercules, CA), a mixture rendering final concentra-tions of 500 mM of each dNTP, 12.5 ng/ml random hexamers (bothGE Healthcare, Uppsala, Sweden), 1� first-strand buffer, 10 mMDTT and 7 U/ml M-MLV reverse transcriptase (all Invitrogen) wasadded to yield a total reaction volume of 40 ml. After a secondannealing step at 25 8C for 10 min, an elongation step at 37 8C for60 min was performed before enzyme inactivation at 70 8C for15 min. The obtained products were diluted in 560 ml DEPC waterto 5 ng/ml final concentration, aliquoted and stored at �70 8C.

2.4. Reverse transcription quantitative real-time PCR (RT-qPCR)

The number of transcripts was determined by RT-qPCR in a 25-ml reaction volume including 1� iQTM SYBR Green Supermix (Bio-Rad) and 0.2 mM primers using cDNA as template (�25 ng totalRNA). Samples were analyzed in triplicate in an MyiQ (Bio-Rad)using the following settings: 3 min at 95 8C, 40 cycles of 15 secondsat 95 8C plus 1 min at 60 8C and final melt curve analysis.Simultaneous amplification of known amounts of PCR productgenerated a standard curve for comparison. To correct for variableefficiencies in cDNA synthesis, all mRNA values were related to theamount of the housekeeping gene cyclophilin A. Primers wereobtained from Applied Biosystems, and the sequences (given in 50–30-orientation, F = forward, R = reverse) used were as follows:cyclophilin A (F: TTC ATC CTA AAG CAT ACG GGT CCT G, R: GCT TGCCAT CCA ACC ACT CAG TC), EGFR (F: AGC AGT GAC TTT CTC AGCAAC, R: TTC TGG CAG TTC TCC TCT CC), HER2 (F: CTG GTG GAT GCTGAG GAG TAT CTG, R: GCG GTG CCT GTG GTG GAC), HER3 (F: GGTGCT GGG CTT GCT TTT, R: CGT GGC TGG AGT TGG TGT TA), HER4 (F:TGT GAG AAG ATG GAA GAT GGC, R: GTT GTG GTA AAG TGG AATGGC) and HRG1-b (detecting all b-isoforms) adapted from Memonet al. [25] (F: TAG GAA ATG ACA GTG CCT C, R: CGT AGT TTT GGCAGC GA).

2.5. Protein lysate preparation and cell culture

Frozen neoepidermal pellets were lysed on ice for 15–30 min incold LCW-lysis buffer pH 7.5 (20 mM Tris pH 7.5 (Bio-Rad), 0.5%(w/v) Triton X-100, 0.5% (w/v) deoxycholate, 10 mM EDTA, 30 mMNaPyro-P (all Sigma–Aldrich), 150 mM NaCl (Scharlau, PortAdelaide, Australia)). To the lysis buffer, phosphatase and proteaseinhibitors (0.5 mM Na3VO4 (Sigma–Aldrich) and Complete Mini(Roche)) were added immediately prior to use. Lysates werecleared by centrifugation at 12,000 � g for 10–30 min at 4 8C andsupernatants were stored at�70 8C. The protein concentration wasdetermined with Bio-Rad protein assay (Bio-Rad) using bovineserum albumin as standard.

Normal human epidermal keratinocytes were cultured inEpiLife medium complemented with human keratinocyte growthsupplement (HKGS) and gentamycin/amphotericin (all fromCascade Biologics Inc., Portland, OR). Starvation of keratinocyteswas performed by excluding all supplements from the EpiLifemedium. T47D and A431 cell lines were cultured in DMEM,supplemented with 10% FBS and 1% antibiotics (PEST). T47D breastcarcinoma cells express HER1–4 [26] whereas A431 cells (derivedfrom epidermoid squamous cell carcinoma) express all HERsexcept HER4 [7]. Cells were incubated at 37 8C in 5% CO2 andsubcultivated with trypsin-EDTA (Cascade Biologics). Ten ng/ml ofeither HRG-b1 (for 5 or 10 min) or EGF (for 5 min) was added priorto harvest. Anti-HER3 antibody was used at 10 mg/ml and anti-HRG-b1 (rabbit polyclonal, Ab-2, without BSA/azide, Thermo

Scientific) at 50 mg/ml. Stock solutions (10 mM) of ZD1839/gefitinib (AstraZeneca, Macclesfield, UK) and CP654577 (Pfizer,Groton, CT) were prepared in DMSO and used at 1 mM concentra-tion. At the time of harvest, cells were washed thrice with ice-coldPBS and lysed as described above.

2.6. Immunoprecipitation

HER4 was immunoprecipitated by incubating lysates overnightat 4 8C with a rabbit monoclonal anti-HER4 antibody (1:50, clone111B2, Cell Signaling Technology, Danvers, MA) in a 200-mlvolume (corresponding to 62 mg proteins for T47D cells and�300 mg for neoepidermal and A431 lysates) followed by a 2-hincubation with protein-A sepharose beads slurry (GE Healthcare,Uppsala, Sweden). The beads were washed five times in LCW-lysisbuffer and proteins solubilized in 2� SDS sample buffer prior toimmunoblotting.

2.7. SDS-PAGE and Western blot analysis

Proteins were separated by SDS polyacrylamide gel electro-phoresis (7% gel) and transferred to a nitrocellulose membrane bysemidry transfer (Trans-blot1 SD, Bio-Rad). All components(papers, gel and membrane) were first soaked in semidry transfersolution (50 mM Tris, 40 mM glycine, 4% (w/v) SDS and 20% (v/v)methanol). Blocking was with 5% Membrane Blocking Agent (GEHealthcare). Membranes were incubated overnight at 4 8C withrabbit anti-pHER3 (Tyr 1289, 1:1000, clone 21D3), rabbit anti-HER3 (1:1000, clone 1B2), rabbit anti-HER4 (1:1000, clone 111B2),or rabbit anti-b-actin (1:1000, clone 13E5) antibodies (all from CellSignaling Technology). Primary antibodies were detected using ahorseradish peroxidase-conjugated secondary anti-rabbit IgG-antibody (Jackson Immunoresearch, Suffolk, UK) at 1:50,000dilution for 1 h. Visualization was by enhanced chemilumines-cence, ECL plus on ECL Hyperfilm (GE Healthcare). All washingsteps and dilutions were in Tris-buffered saline (TBS) with 1% (v/v)Tween-20 (Merck) and 1% (w/v) bovine serum albumin (IgG- andprotease-free, Jackson). The size of protein bands was related to aRainbow marker (GE Healthcare). For reprobing, blots werestripped with 0.4 M NaOH for 10 min and reblocked before addingthe next primary antibody.

2.8. Statistical analysis

Differences in radial growth rates between treated and controlsamples were estimated by a mixed effects model [27] aspreviously described [23]. Statistical analysis was with SAS version9.1 (SAS Institute Inc., Cary). Absolute differences of means in BrdUdensity and histometric data between ligand- and vehicle-exposedgroups were calculated for within each experimental series andwere subjected to paired t-test using GraphPad Prism version 4.00(GraphPad Software, Inc., San Diego, CA). Probability valuesP < 0.05 were considered statistically significant.

3. Results

3.1. HRG-b1 stimulates neoepidermal outgrowth

The dynamic effects of HRG-b1 were traced by fluorescenceimaging of skin explants grown in basal medium over a 9-dayculture period. Exogenous HRG-b1 at 10 or 20 ng/ml concentra-tions enhanced the outgrowth rates significantly although lessmarkedly than by corresponding concentrations of EGF (Table 1).Overall, HRG-b1 at 10 and 20 ng/ml levels had a growthstimulatory effect �30% that of exogenous EGF. Dose-relatedresponses were observed for HRG-b1 at 5 ng/ml vs. 10 ng/ml

Table 1Dynamic effects of exogenous HRG-b1 and EGF on re-epithelialization.

Ligand Concentration (ng/ml) Outgrowth rate (mm/day)a Pb

HRG-b1 5 422�5 0.150

10 446�8 <0.001

20 439�6 0.002

EGF 5 500�5 <0.001

10 520�6 <0.001

20 528�6 <0.001

Vehicle 408�7

a Estimated radial growth rates (mm/day) of neoepidermis cultured on 10-mm DEDs in basal medium supplemented with HRG-b1 or EGF at three different concentrations.

Mean values� SEM from four independent experiments are given. SEM values are adjusted to study. n = 16 in each group, except for EGF at 20 ng/ml (n = 15).b Statistical significance of difference in growth rates between treatment and vehicle (PBS) group. Insert: visualization of neoepidermal outgrowth (green area) generated

from a central skin explant glued onto acellular dermis (day 7). Living cells are traced by fluorescein diacetate probing and fluorescence microscopy.

Fig. 1. Effect of HRG-b1 on neoepidermal cell proliferation and histometric indices.

Skin explants were fixed onto DEDs and grown in basal culture medium for 9 days in

presence of HRG-b1, EGF (both 10 ng/ml) or vehicle (PBS). Prior to harvest, cultures

were incubated with BrdU for 4 h before formalin fixation and paraffin embedment.

Tissue cross-sections were stained for anti-BrdU and hematoxylin for evaluation of

proliferation activity and histometric analysis. Data are presented as average

differences of means between ligand- and vehicle-exposed samples from four

separate experimental series (n = 4/series) normalizing the control to zero. Mean

absolute values of controls are given. (a) Proliferative cell density (BrdU-positive

cells in neoepidermis/mm basal membrane). (b) Neoepidermal thickness (mm, area

of viable neoepidermis/horizontal length of neoepidermis). (c) Papillomatosis index

(length of basal membrane/horizontal length of neoepidermis). Error bars indicate

SEM. *P < 0.05, **P < 0.001, paired t-test of differences between ligand and control

regarding series.

S. Forsberg, O. Rollman / Journal of Dermatological Science 59 (2010) 7–1510

(P = 0.02), and for EGF at 5 ng/ml vs. 10 ng/ml or 20 ng/ml (P = 0.04and P = 0.006, respectively). In a separate experiment using 50 ng/ml of HRG-b1, no additional increase in growth rate was observed(data not shown).

3.2. HRG-b1 increases keratinocyte proliferation and

neoepidermal thickness

The proliferative and morphological effects of HRG-b1 werestudied in organ cultures treated with HRG-b1 vs. vehicle. At20 ng/ml, the frequency of BrdU-positive cells in ligand-stimulatedneoepidermis increased by 43% as compared to control samples(Fig. 1a). The mitogenic effects of lower doses of HRG-b1 weremoderate and did not reach statistical significance. For compari-son, exogenous EGF at 20 ng/ml increased the number of BrdU-labelled cells on average 85%. Histometric analysis showed thatHRG-b1 increased the mean neoepidermal thickness in samplesexposed to all test concentrations (Fig. 1b). At 20 ng/ml dose level,the thickness increased by on average 23% with no concomitantchange in the papillomatosis index as compared to controlcultures.

3.3. HRG-b1 does not affect immunostaining patterns of late

differentiation markers

Immunohistochemically, the expression of terminal differenti-ation markers involucrin, loricrin and filaggrin were not notablyaltered in neoepidermis grown in presence of exogenous HRG-b1,EGF or vehicle (Fig. 2). Occasionally, the staining intensity ofloricrin and involucrin was slightly weaker in HRG-b1-exposedsamples, but the overall visual impression was that all markerswere similarly expressed independent of treatment.

3.4. Neoepidermal cultures express HER3 transcript and protein

The mRNA expression of neoepidermal EGFR family membersand HRG-b isoforms were investigated by RT-qPCR assay.Transcripts of EGFR, HER2 and HER3 were readily identified inthe neoepidermis whereas HER4 mRNA was detected only atexceedingly low levels (Table 2). Furthermore, no HER4 product ofthe expected transcript size was seen by subsequent agarose gelanalysis (data not shown). Transcripts of the HER3 ligand HRG-b,including mRNA of b1–3-isoforms, were also found in culturedneoepidermis.

Having disclosed HER3 and ligand mRNA expression by ourmodel, we next analyzed the state of HER3 phosphorylation.Western blot analysis of neoepidermis grown under basalconditions displayed phospho-HER3 (Tyr 1289), which was

enhanced in HRG-b1—but not in EGF-exposed tissues (Fig. 3).HER4 protein was not detected in the neoepidermis by Westernblotting of total lysates or after HER4 immunoprecipitation.

In the neoepidermis, HER3 immunostaining showed a chicken-wire pattern of cell membranes in the upper stratum spinosum andstratum granulosum (Fig. 4b). This labelling pattern was similar tothat of HER2 [22] with no accentuation of the HER3 stainingintensity at the advancing edge like that observed in skin wounds[28]. Staining with an antibody specific for phosphorylated HER3,however, displayed a mixed cytoplasmic and predominantlymembranous pattern of keratinocytes residing in the basal andlow-to-intermediate suprabasal compartment (Fig. 4d) similar tothat of neoepidermal EGFR [22]. The staining intensity of

Fig. 2. Expression of differentiation markers in cultured neoepidermis. Immunolocalization pattern of the differentiation markers involucrin (a–c), loricrin (d–f) and filaggrin

(g–i) in neoepidermis cultured for 9 days on 10-mm DEDs in basal medium supplemented with either HRG-b1, EGF (both at 10 ng/ml), or vehicle (PBS). Bar = 50 mm.

Table 2Expression of EGFR superfamily receptors and HRG1-b mRNA in cultured

neoepidermis.

Transcript Mean mRNA expression

(ratio to housekeeping gene �103)a

EGFR 544

HER2 530

HER3 1365

HER4 0.0045

HRG1-bb 16

a Mean mRNA expression in neoepidermis grown in basal culture medium (see

Section 2). Values are expressed relative to the housekeeping gene cyclophilin, n = 4.b Detects b1–b3-isoforms of HRG1.

S. Forsberg, O. Rollman / Journal of Dermatological Science 59 (2010) 7–15 11

neoepidermal HER3 and pHER3 expression gradually diminishedin horizontal direction towards its advancing edge. Analogous tothe proliferation marker Ki-67 [21], the HER3 receptors were moreabundantly expressed in the proximal and mid-radial sections ofneoepidermis. Based on visual assessment, there was no obviousdifference in the staining intensity or distribution of HER3 andpHER3 with reference to treatment, either vehicle or HRG-b1. Ascompared to native human skin (Fig. 4a and c), organ-cultured

Fig. 3. Expression of phosphorylated HER3 and HER4 in cultured neoepidermis. Weste

grown in absence or presence of exogenous HRG-b1 or EGF at 10 ng/ml. T47D cells (expr

reference cells. The panels are representative of several independent experiments. (a) Ne

(lanes 3–4) or classic (lanes 7–8) keratinocyte medium supplemented � HRG-b1 (final 2 h

harvest. Lysate amounts of 20 or 10 mg proteins, for neoepidermis and cells, respectively, w

immunoblotting against pHER3 (Tyr 1289). For internal control, blots were reprobed again

antibody against HER4 in a 200-ml volume (corresponding to 62 mg proteins for T47D cells an

SDS-PAGE and blotted against HER4. HRG-b1 was added to cultures for 2 h prior to harve

epidermis displayed more intense HER3 and – in particular pHER3– immunolabelling.

3.5. HER3-directed antibodies inhibit constitutive receptor

phosphorylation in keratinocytes and neoepidermis

To explore the impact of HER3 antagonism at ligand level weanalyzed pHER3 status by Western blotting. As shown in Fig. 5a (lane1), pHER3 was expressed by keratinocytes in monolayer culture andnot further upregulated by the addition of HRG-b1 (Fig. 5a, lane 4).This may be explained by activation of HER3 through endogenousproduction of HRGs under basal culture conditions. The inability of aneutralizing anti-HRG-b1 antibody to abate phosphorylation ofHER3 suggests that other HRGs except for the b1 isoform might beinvolved in basal HER3 activation (Fig. 5a, lanes 3 and 6).Nonetheless, phosphorylation of HER3 in keratinocytes wasinhibited by applying an antibody that prevents ligand-binding toHER3 both in presence or absence of exogenous ligand (Fig. 5a, lanes2 and 5). A similar result was observed in the explant model;constitutively phosphorylated HER3 was slightly increased afterHRG-b1 exposure (Fig. 3 and 6, lane 4) and markedly blocked inpresence of the anti-HER3 antibody (Fig. 6, lane 6).

rn blot analysis of total lysates and HER4 immunoprecipitates from neoepidermis

essing high levels of HER3) and A431 cells (expressing high levels of EGFR) served as

oepidermis from normal human skin was grown on 10-mm DEDs for 9 days in basal

) or EGF (final 1.5 h). T47D and A431 cells were treated � HRG-b1 or EGF 5 min before

ere separated on a 7% SDS-PAGE gel and the effect on HER3 activation was analyzed by

st HER3. (b) Neoepidermis and cells were subjected to immunoprecipitation using an

d�300 mg for neoepidermal and A431 lysates). Immunoprecipitates were separated by

st.

Fig. 4. Expression and localization of HER3 in normal and cultured skin. Expression of HER3 (a and b) and pHER3 (c and d) in normal (a and c) and cultured (b and d) human

skin. Biopsies were obtained from normal skin of a healthy individual. Cultured skin (neoepidermis) was generated from normal skin explants adhered to DEDs and grown for

9 days in basal medium. The tissue was formalin fixed, paraffin embedded, sectioned (5 mm) and immunohistochemically stained for HER3 or pHER3 (Tyr 1289) as described

in Section 2. Control experiments omitting the primary antibody showed no background staining (results not shown). Bar = 50 mm.

S. Forsberg, O. Rollman / Journal of Dermatological Science 59 (2010) 7–1512

3.6. HER3 activation in cultured keratinocytes is blocked by

an HER2 kinase inhibitor

Since HER3 virtually lacks intrinsic catalytic activity it isdependent on a kinase-active heterodimerization partner to becomephosphorylated and recruit downstream signalling proteins. Ingeneral, EGFR or – in particular – HER2 is likely to fulfil this role. To

Fig. 5. Effect of anti-HER3 and anti-HRG-b1-targeting antibodies and small-

molecule kinase inhibitors on HER3 phosphorylation in cultured keratinocytes. The

effect of antagonizing HER strategies on HER3 activation in normal human

keratinocytes was investigated by Western blotting against pHER3 (Tyr 1289).

HRG-b1-treated T47D cells served as positive control for HER3 phosphorylation. In

both cell types, HRG-b1 (10 ng/ml) was added 10 min prior to cell lysis. The lower

part of each blot was incubated with b-actin antibody to confirm equal loading of

the gels. The panels are representative of several independent experiments. (a)

Keratinocytes were exposed to antibodies against HER3 (10 mg/ml) or HRG-b1

(50 mg/ml) in Epilife medium deprived of growth supplements over the final 27 h of

culture. Lysates corresponding to 10 mg of total proteins were used. (b) HER3

phosphorylation in cells treated with different combinations of exogenous HRG-b1

and two HER tyrosine kinase inhibitors. The keratinocytes were starved for 25 h

before addition of ZD1839 (inhibits EGFR kinase) or CP654577 (inhibits HER2

kinase) at 1 mM during the final 2 h of culture. Amounts of 20 or 10 mg total

proteins, for keratinocytes and T47D cells, respectively, were used for analysis.

elucidate which HER3 dimerization partner(s) may be operative inhuman keratinocytes, cells in culture were treated with small-molecule tyrosine kinase inhibitors preferentially directed at EGFR(ZD1839) or HER2 (CP654577). As shown in Fig. 5b (lanes 3 and 6),exposures to CP654577 – as opposed to ZD1839 – blocked thephosphorylation of HER3 in epidermal keratinocytes suggesting thatHER2 kinase is the preferential activator of HER3 in epidermalkeratinocytes.

3.7. Re-epithelialization is inhibited by an HER2 anti-dimerization

antibody

The effect of HER2 kinase blockade on neoepidermal growthwas recently assessed by comparing CP654577 with non-HER2-specific kinase inhibitors, e.g. ZD1839 [22]. In that study, CP654577at 1 mM drug level inhibited re-epithelialization though lessefficiently than EGFR kinase inhibitors. Here we compared thegrowth-inhibitory potential of the anti-HER3 antibody with that ofthe anti-HER2 dimerization antibody, pertuzumab, and the EGFR-blocking antibody, cetuximab. As illustrated in Fig. 7, the averagegrowth rate of neoepidermis exposed to anti-HER3 antibody (mean304 � 6.8 mm/day) was reduced by 10% compared to the control

Fig. 6. Effect of ligand-stimulation and anti-HER3 targeting antibodies on HER3

phosphorylation in cultured neoepidermis. Neoepidermis was grown from normal

human skin on 10-mm DEDs in basal medium for 9 days. Anti-HER3 antibodies

were present throughout the entire culture period. HRG-b1 (10 ng/ml) was added

during 2 h immediately prior to harvest. T47D cells treated with HRG-b1 (10 ng/ml,

10 min) served as positive control for HER3 phosphorylation. Protein amounts

loaded on gel were 10 mg for T47D cells and 20 mg for neoepidermis. Parallel

blotting (lower part of the membrane) against b-actin confirmed equal loading

between paired control and treated samples.

Fig. 7. Effect of HER blockade on re-epithelialization. Mean radial outgrowths (mm)

of neoepidermis from normal human skin explants cultured in basal medium on 10-

mm DEDs for 9 days in absence or presence of three different HER-directed

antibodies during the entire culture period. Error bars indicate SEM. P-values

denote statistical significance (mixed effects model, see Section 2) of mean

differences in radial growth rates (mm/day) between antibody-exposed and

matched control group. (a) Anti-HER3 antibody was used at 10 mg/ml. Values

represent means from one experimental series. n = 12 in each group. (b and c)

Pertuzumab and cetuximab were used at 100 nM concentration. Values represent

means from four experimental series. n = 16 in each group.

S. Forsberg, O. Rollman / Journal of Dermatological Science 59 (2010) 7–15 13

group (339 � 6.8 mm/day, P = 0.002). This growth reduction coincid-ed with blockage of neoepidermal HER3 phosphorylation by anti-HER3 exposure (Fig. 6). By comparison, the mean outgrowth rates ofcultures treated with pertuzumab and cetuximab were reduced by18% and 46%, respectively.

4. Discussion

Regulation of cell activities through EGFR family membersassumes stringent control and interaction of individual HERs, theirligands and downstream signalling components. Here, we focusedspecifically on the expression and biological activity of ligand-stimulated HER3 during re-epidermalization. As experimental set-up, we grew human skin explants on a substrate of cell-free dermisand attached a gentle imaging procedure. This skin model excludessystemic factors and concentrates on epidermal keratinocytessince cell-specific markers of other resident skin cells such asLangerhans cells, fibroblasts and melanocytes vanish duringcultivation [21]. An inherent advantage of this model is that in

vivo-like neoepidermis is generated at the air–liquid borderallowing quantitative outgrowth measurements prior to tissueanalysis. Using this technique, we recently demonstrated thatEGFR was immunolocalized at basal and adjacent suprabasal levelsof cultured neoepidermis [22] similar to that in hyperproliferative

human skin [6,7,29]. Targeted inhibition of neoepidermal EGFRkinase efficiently reduced the outgrowth rate from skin explants[22] illustrating the importance of EGFR signalling in regeneratingepidermis. By contrast, HER2 was expressed mainly at higherlayers of the neoepidermis, and an HER2-selective kinase inhibitorhad more limited effects on the outgrowth course. These resultsagree with the concepts that EGFR mainly associates withkeratinocyte proliferation, whereas HER2 is more closely linkedwith cell differentiation [4–7]. Here, we extended our investiga-tions on neoepidermal HERs by analyzing HER3 expression and thepotency of HRG-b1 (HER3 and HER4 ligand) relative to EGF (EGFRligand) in human skin explants.

Exogenous HRG-b1 increased the rate of re-epithelializationalthough not as clearly as observed with EGF. The response to HRG-b1 may have resulted from combined cellular mechanisms sinceHRGs may affect proliferation, migration and apoptosis ofepithelial cells. For example, HRG-a is known to stimulateHER3-mediated motility of human adult keratinocytes in culture[28]. Moreover, HRG-b1 facilitates locomotion of breast carcinomacells via b1-integrin upregulation [30] and reorganization ofmotile actin cytoskeleton structures [31]. However, it is not knownwhether HRG-b1 directs migration of skin keratinocytes throughsuch mechanisms or via altered expression of cytokeratinsassociated with regenerating epidermis. With regard to terminaldifferentiation, no HRG-b1-related effects on involucrin, loricrinand filaggrin expression were observed in our study as evaluatedby immunohistochemistry (Fig. 2) and microarray gene analysis(preliminary data not shown) of neoepidermis. This contrasts withthe report by De Potter [14] who found significant upregulation ofkeratin 1/10, involucrin and loricrin mRNA in HRG-b1-exposedhuman keratinocyte cultures.

The receptor targeted by HRG-b1 in our model was probablyHER3 exclusively since (i) EGFR has virtually no affinity for thisligand [32], (ii) HER2 lacks ligand-binding sites and (iii) HER4protein expression was not evident in cultured neoepidermis asjudged by Western blot analysis. Although transcripts of all EGFRfamily members were detected in neoepidermis, the amount ofHER4 mRNA was exceedingly low. This coincides with previousobservations that HER4-as opposed to HER3-is either absent oronly sparsely detected in normal skin when analyzed by Northernblot or immunohistochemical techniques [6–9]. Although theprecise role of HER3 in normal skin remains unraveled, thisreceptor has attracted major interest in cancer research sinceamplification or overexpression of HER3 is coupled to evolution ofseveral malignancies, e.g. cancer of the breast [33], stomach [34],lung [35] and uterine cervix [36]. In skin oncology, HER3 has beenassociated with skin carcinoma [37,38] and melanoma [39].Furthermore, overexpression of HER3 was recently proposed asa marker of poor prognosis in malignant melanoma [40].

HER3 differs from other EGFR family members due to its lack ofintrinsic kinase activity and inability to form active homodimersafter ligand-binding [41]. Thus a prerequisite for ligand-stimulatedHER3 to start signal transduction is that the cytoplasmatic tailbecomes phosphorylated from an adjacent kinase. To appraisewhether EGFR or HER2 kinase cross-activated HER3 in neoepi-dermis, selective tyrosine kinase inhibitors and monoclonalantibodies directed at either receptor were used. We found thatthe HER2 kinase inhibitor CP654577, but not the EGFR kinaseinhibitor ZD1839, had a suppressive effect on HER3 phosphoryla-tion in epidermal keratinocytes. Although not a proof ofneoepidermal HER2–HER3 dimerization, this suggests that HER2tyrosine kinase rather than EGFR kinase may serve as a majortransactivator of ligand-stimulated HER3 in neoepidermis. Furthersupport for this explanation was gained from experiments usingHER-targeting monoclonal antibodies. Cetuximab is known tointeract with ligand-binding to domain III of the extracellular EGFR

S. Forsberg, O. Rollman / Journal of Dermatological Science 59 (2010) 7–1514

domain thereby competitively reducing downstream signallingfrom this receptor [42]. Therefore, the growth-inhibitory effect ofcetuximab on epithelialization was probably a consequence ofselective EGFR antagonism without direct involvement of HER2 orHER3. By contrast, pertuzumab binds at the dimerization arm ofHER2 thus preventing heterodimerization of HER2–HER3 andsignalling from this receptor combination [43,44]. The findingsstrengthen the view that ligand-bound HER3 preferably associateswith HER2 [45], and that HER2–HER3 heterodimers constitute anactive signalling complex in epidermal cells analogous to that ine.g. fetal lung [46] and bronchial epithelial cells [47]. Whichpostreceptor proteins were recruited by HRG-b1-stimulation ofneoepidermis remains to be elucidated. HER3 has a large numberof docking sites for the p85 unit of phosphatidylinositol 3-kinase[48,49] and is associated with activation of the Akt kinase pathwaythat drives pro-growth and anti-apoptotic responses [50].

The neoepidermal expression of HER3 immunostaining corre-sponded to the reported pattern of HER2 localization inintermediate and high suprabasal strata [6,22]. Provided thatligand-stimulated HER3 was activated by HER2 kinase such spatialcolocalization of the two receptors would be anticipated. Prolifer-ative responses coupled to HER3 phosphorylation in the upperneoepidermis might be partly explained by activated heterodimerspromoting effects such as HRG-b1-induced VEGF secretion asdescribed in breast and lung cancer cells [51]. Since receptors forVEGF may transfer proliferative signals in basal and suprabasalcells in the epidermis such indirect effects of HRG-b1 cannot beexcluded.

Phosphorylated HER3 was immunolocalized at basal and lowsuprabasal levels, i.e. almost the reverse pattern to that of HER3.This agrees with the pro-growth responses associated with HER3activation and high expression of HRG transcript by undifferenti-ated keratinocytes [14]. Nevertheless, it raises the questionwhether the anti-HER3 antibody binds with equal affinity tophosphorylated vs. unphosphorylated species in neoepidermaltissue. We are not aware of any previous report that would help tointerpret the immunohistochemical staining pattern of thisantibody in skin. Since the epitope for the anti-HER3 antibody isclose to the phosphorylation site one may speculate thatconformational changes induced by ligand-stimulation affect thebinding properties of the antibody.

Altogether, our data suggest that cell signalling from ligand-activated HER3 stimulates regeneration of epidermis in vitro

though EGFR mediates more overt effects on outgrowth dynamics.Further studies are required to elucidate the full repertoire of HER3ligands, dimeric partners and transduction molecules that areoperative in native human skin. In particular, the issue of receptorcompartmentalization and cellular mechanisms that connectphosphorylated HER3 with keratinocyte-specific biologicalresponses need explanation to appreciate the biological role ofHER3 in normal and hyperproliferative epidermis.

Conflict of interest

The authors state no conflict of interest.

Acknowledgements

The Department of Plastic Surgery, Uppsala University Hospital,is acknowledged for supplying us with skin samples. Thanks aredue to Pfizer, AstraZeneca and Genentech Inc. for generous supplyof tyrosine kinase inhibitors and antibodies, respectively. Weacknowledge Lisa Wernroth at Uppsala Clinical Research Centre(UCR) for statistical advice, Hans Torma for technical advice on RT-qPCR and Arne Ostman for fruitful discussions. This study was

supported by the Finsen Welander Foundation and the SwedishPsoriasis Association.

References

[1] Jorissen RN, Walker F, Pouliot N, Garrett TP, Ward CW, Burgess AW. Epidermalgrowth factor receptor: mechanisms of activation and signalling. Exp Cell Res2003;284:31–53.

[2] Cohen S, Ushiro H, Stoscheck C, Chinkers M. A native 170,000 epidermalgrowth factor receptor-kinase complex from shed plasma membrane vesicles.J Biol Chem 1982;257:1523–31.

[3] Piepkorn M, Lo C, Plowman G. Amphiregulin-dependent proliferation ofcultured human keratinocytes: autocrine growth, the effects of exogenousrecombinant cytokine, and apparent requirement for heparin-like glycosami-noglycans. J Cell Physiol 1994;159:114–20.

[4] Prigent SA, Lemoine NR, Hughes CM, Plowman GD, Selden C, Gullick WJ.Expression of the c-erbB-3 protein in normal human adult and fetal tissues.Oncogene 1992;7:1273–8.

[5] Press MF, Cordon-Cardo C, Slamon DJ. Expression of the HER-2/neu proto-oncogene in normal human adult and fetal tissues. Oncogene 1990;5:953–62.

[6] Piepkorn M, Predd H, Underwood R, Cook P. Proliferation–differentiationrelationships in the expression of heparin-binding epidermal growth factor-related factors and erbB receptors by normal and psoriatic human keratino-cytes. Arch Dermatol Res 2003;295:93–101.

[7] Stoll SW, Kansra S, Peshick S, Fry DW, Leopold WR, Wiesen JF, et al. Differentialutilization and localization of ErbB receptor tyrosine kinases in skin comparedto normal and malignant keratinocytes. Neoplasia 2001;3:339–50.

[8] Plowman GD, Culouscou JM, Whitney GS, Green JM, Carlton GW, Foy L, et al.Ligand-specific activation of HER4/p180erbB4, a fourth member of the epi-dermal growth factor receptor family. Proc Natl Acad Sci USA 1993;90:1746–50.

[9] Srinivasan R, Poulsom R, Hurst HC, Gullick WJ. Expression of the c-erbB-4/HER4 protein and mRNA in normal human fetal and adult tissues and in asurvey of nine solid tumour types. J Pathol 1998;185:236–45.

[10] Falls DL. Neuregulins: functions, forms, and signaling strategies. Exp Cell Res2003;284:14–30.

[11] Britsch S. The neuregulin-I/ErbB signaling system in development and disease.Adv Anat Embryol Cell Biol 2007;190:1–65.

[12] Lu HS, Chang D, Philo JS, Zhang K, Narhi LO, Liu N, et al. Studies on the structureand function of glycosylated and nonglycosylated neu differentiation factors.Similarities and differences of the alpha and beta isoforms. J Biol Chem1995;270:4784–91.

[13] Marikovsky M, Lavi S, Pinkas-Kramarski R, Karunagaran D, Liu N, Wen D, et al.ErbB-3 mediates differential mitogenic effects of NDF/heregulin isoforms onmouse keratinocytes. Oncogene 1995;10:1403–11.

[14] De Potter IY, Poumay Y, Squillace KA, Pittelkow MR. Human EGF receptor(HER) family and heregulin members are differentially expressed in epider-mal keratinocytes and modulate differentiation. Exp Cell Res 2001;271:315–28.

[15] Bacus SS, Huberman E, Chin D, Kiguchi K, Simpson S, Lippman M, et al. A ligandfor the erbB-2 oncogene product (gp30) induces differentiation of humanbreast cancer cells. Cell Growth Differ 1992;3:401–11.

[16] Karamouzis MV, Badra FA, Papavassiliou AG. Breast cancer: the upgraded roleof HER-3 and HER-4. Int J Biochem Cell Biol 2007;39:851–6.

[17] Muraoka-Cook RS, Caskey LS, Sandahl MA, Hunter DM, Husted C, Strunk KE,et al. Heregulin-dependent delay in mitotic progression requires HER4 andBRCA1. Mol Cell Biol 2006;26:6412–24.

[18] Sartor CI, Zhou H, Kozlowska E, Guttridge K, Kawata E, Caskey L, et al. Her4mediates ligand-dependent antiproliferative and differentiation responses inhuman breast cancer cells. Mol Cell Biol 2001;21:4265–75.

[19] Regnier M, Asselineau D, Lenoir MC. Human epidermis reconstructed ondermal substrates in vitro: an alternative to animals in skin pharmacology.Skin Pharmacol 1990;3:70–85.

[20] Rikimaru K, Moles JP, Watt FM. Correlation between hyperproliferation andsuprabasal integrin expression in human epidermis reconstituted in culture.Exp Dermatol 1997;6:214–21.

[21] Lu H, Rollman O. Fluorescence imaging of re-epithelialization from skinexplant cultures on acellular dermis. Wound Repair Regen 2004;12:575–86.

[22] Forsberg S, Ostman A, Rollman O. Regeneration of human epidermis onacellular dermis is impeded by small-molecule inhibitors of EGF receptortyrosine kinase. Arch Dermatol Res 2008;300:505–16.

[23] Forsberg S, Saarialho-Kere U, Rollman O. Comparison of growth-inhibitoryagents by fluorescence imaging of human skin re-epithelialization in vitro.Acta Derm Venereol 2006;86:292–9.

[24] Johnson I. Fluorescent probes for living cells. Histochem J 1998;30:123–40.[25] Memon AA, Sorensen BS, Melgard P, Fokdal L, Thykjaer T, Nexo E. Expression of

HER3, HER4 and their ligand heregulin-4 is associated with better survival inbladder cancer patients. Br J Cancer 2004;91:2034–41.

[26] Beerli RR, Graus-Porta D, Woods-Cook K, Chen X, Yarden Y, Hynes NE. Neudifferentiation factor activation of ErbB-3 and ErbB-4 is cell specific anddisplays a differential requirement for ErbB-2. Mol Cell Biol 1995;15:6496–505.

[27] Pinheiro JC, Bates DM. Mixed-effects models in S and S-PLUS. New York:Springer; 2000.

S. Forsberg, O. Rollman / Journal of Dermatological Science 59 (2010) 7–15 15

[28] Schelfhout VR, Coene ED, Delaey B, Waeytens AA, De Rycke L, Deleu M, et al.The role of heregulin-alpha as a motility factor and amphiregulin as a growthfactor in wound healing. J Pathol 2002;198:523–33.

[29] Nanney LB, McKanna JA, Stoscheck CM, Carpenter G, King LE. Visualization ofepidermal growth factor receptors in human epidermis. J Invest Dermatol1984;82:165–9.

[30] Adelsman MA, McCarthy JB, Shimizu Y. Stimulation of beta1-integrin functionby epidermal growth factor and heregulin-beta has distinct requirements forerbB2 but a similar dependence on phosphoinositide 3-OH kinase. Mol BiolCell 1999;10:2861–78.

[31] Adam L, Vadlamudi R, Kondapaka SB, Chernoff J, Mendelsohn J, Kumar R.Heregulin regulates cytoskeletal reorganization and cell migration throughthe p21-activated kinase-1 via phosphatidylinositol-3 kinase. J Biol Chem1998;273:28238–46.

[32] Jones JT, Akita RW, Sliwkowski MX. Binding specificities and affinities of egfdomains for ErbB receptors. FEBS Lett 1999;447:227–31.

[33] Naidu R, Yadav M, Nair S, Kutty MK. Expression of c-erbB3 protein in primarybreast carcinomas. Br J Cancer 1998;78:1385–90.

[34] Hayashi M, Inokuchi M, Takagi Y, Yamada H, Kojima K, Kumagai J, et al. Highexpression of HER3 is associated with a decreased survival in gastric cancer.Clin Cancer Res 2008;14:7843–9.

[35] Yi ES, Harclerode D, Gondo M, Stephenson M, Brown RW, Younes M, et al. Highc-erbB-3 protein expression is associated with shorter survival in advancednon-small cell lung carcinomas. Mod Pathol 1997;10:142–8.

[36] Fuchs I, Vorsteher N, Buhler H, Evers K, Sehouli J, Schaller G, et al. Theprognostic significance of human epidermal growth factor receptor correla-tions in squamous cell cervical carcinoma. Anticancer Res 2007;27:959–63.

[37] Krahn G, Leiter U, Kaskel P, Udart M, Utikal J, Bezold G, et al. Coexpressionpatterns of EGFR, HER2, HER3 and HER4 in non-melanoma skin cancer. Eur JCancer 2001;37:251–9.

[38] Wimmer E, Kraehn-Senftleben G, Issing WJ. HER3 expression in cutaneoustumors. Anticancer Res 2008;28:973–9.

[39] Stove C, Stove V, Derycke L, Van Marck V, Mareel M, Bracke M. The heregulin/human epidermal growth factor receptor as a new growth factor system inmelanoma with multiple ways of deregulation. J Invest Dermatol 2003;121:802–12.

[40] Reschke M, Mihic-Probst D, van der Horst EH, Knyazev P, Wild PJ, Hutterer M,et al. HER3 is a determinant for poor prognosis in melanoma. Clin Cancer Res2008;14:5188–97.

[41] Berger MB, Mendrola JM, Lemmon MA. ErbB3/HER3 does not homodimerizeupon neuregulin binding at the cell surface. FEBS Lett 2004;569:332–6.

[42] Li S, Schmitz KR, Jeffrey PD, Wiltzius JJ, Kussie P, Ferguson KM. Structural basisfor inhibition of the epidermal growth factor receptor by cetuximab. CancerCell 2005;7:301–11.

[43] Agus DB, Akita RW, Fox WD, Lewis GD, Higgins B, Pisacane PI, et al. Targetingligand-activated ErbB2 signaling inhibits breast and prostate tumor growth.Cancer Cell 2002;2:127–37.

[44] Franklin MC, Carey KD, Vajdos FF, Leahy DJ, de Vos AM, Sliwkowski MX.Insights into ErbB signaling from the structure of the ErbB2-pertuzumabcomplex. Cancer Cell 2004;5:317–28.

[45] Graus-Porta D, Beerli RR, Daly JM, Hynes NE. ErbB-2, the preferred hetero-dimerization partner of all ErbB receptors, is a mediator of lateral signaling.Embo J 1997;16:1647–55.

[46] Patel NV, Acarregui MJ, Snyder JM, Klein JM, Sliwkowski MX, Kern JA. Neur-egulin-1 and human epidermal growth factor receptors 2 and 3 play a role inhuman lung development in vitro. Am J Respir Cell Mol Biol 2000;22:432–40.

[47] Polosa R, Prosperini G, Tomaselli V, Howarth PH, Holgate ST, Davies DE.Expression of c-erbB receptors and ligands in human nasal epithelium. JAllergy Clin Immunol 2000;106:1124–31.

[48] Fedi P, Pierce JH, di Fiore PP, Kraus MH. Efficient coupling with phosphatidy-linositol 3-kinase, but not phospholipase C gamma or GTPase-activatingprotein, distinguishes ErbB-3 signaling from that of other ErbB/EGFR familymembers. Mol Cell Biol 1994;14:492–500.

[49] Prigent SA, Gullick WJ. Identification of c-erbB-3 binding sites for phospha-tidylinositol 30-kinase and SHC using an EGF receptor/c-erbB-3 chimera. EmboJ 1994;13:2831–41.

[50] Carraway 3rd KL, Soltoff SP, Diamonti AJ, Cantley LC. Heregulin stimulatesmitogenesis and phosphatidylinositol 3-kinase in mouse fibroblasts trans-fected with erbB2/neu and erbB3. J Biol Chem 1995;270:7111–6.

[51] Yen L, You XL, Al Moustafa AE, Batist G, Hynes NE, Mader S, et al. Heregulinselectively upregulates vascular endothelial growth factor secretion in cancercells and stimulates angiogenesis. Oncogene 2000;19:3460–9.