mitotic activity and lesion depth of border-line melanocytic lesions. Br J Cancer 105:1023–9

Jiang L, Lv X, Li J et al. (2011) The status ofmicroRNA-21 expression and its clinical sig-nificance in human cutaneous malignantmelanoma. Acta Histochem 114:582–8

Krengel S, Hauschild A, Schafer T (2006) Mela-noma risk in congenital melanocytic naevi: asystematic review. Br J Dermatol 155:1–8

Mitchell PS, Parkin RK, Kroh EM et al. (2008)Circulating microRNAs as stable blood-basedmarkers for cancer detection. Proc Natl AcadSci USA 105:10513–8

Robert C, Thomas L, Bondarenko I et al. (2011)Ipilimumab plus dacarbazine for previouslyuntreated metastatic melanoma. N Engl J Med364:2517–26

Satzger I, Mattern A, Kuettler U et al. (2012)MicroRNA-21 is upregulated in malignant

melanoma and influences apoptosis of mela-nocytic cells. Exp Dermatol 21:509–14

Volinia S, Calin GA, Liu CG et al. (2006) AmicroRNA expression signature of humansolid tumors defines cancer gene targets. ProcNatl Acad Sci USA 103:2257–61

Zhang L, Huang J, Yang N et al. (2006) MicroRNAsexhibit high frequency genomic alterations inhuman cancer. Proc Natl Acad Sci USA103:9136–41

Inhibition of the Stem Cell Marker Nestin Reduces TumorGrowth and Invasion of Malignant MelanomaJournal of Investigative Dermatology (2013) 133, 1384–1387; doi:10.1038/jid.2012.508; published online 7 February 2013

TO THE EDITORNestin is a class VI intermediate fila-ment protein that was first described asa neural stem cell marker (Lendahlet al., 1990). It is expressed throughoutthe dermis in the early embryo and inhair follicles after birth (Sellheyerand Krahl, 2010). Nestin-positive hairfollicle cells located above the folliclebulge region can differentiate intovarious cell types during woundhealing (Amoh et al., 2009).

Nestin expression has also beenreported in various neoplasms. Nestinknockdown inhibits migration, invasion,and metastasis of pancreatic cancercells (Matsuda et al., 2011), suggestingthat the protein could be a noveltherapeutic target for tumors. Inmelanoma, nestin overexpression hasbeen observed in advanced stages ofthe disease (Brychtova et al., 2007), inthe invading front (Piras et al., 2010)and at sites of melanoma metastases(Klein et al., 2007). The number ofcirculating nestin-positive melanomacells in the peripheral blood of patientscorrelates with a poor prognosis (Fusiet al., 2011). Therefore, in this study, weused a gene silencing strategy toinvestigate the potential effectiveness ofa nestin-targeting therapy in malignantmelanomas.

Nestin was found to be stronglyexpressed in A375 and MeWo mela-noma cells, but weakly expressed inG361 cells (Supplementary Figure S1a,b, c online and Supplementary Materialonline). We first suppressed nestinexpression in A375 cells using shorthairpin RNA (shRNA) (Matsuda et al.,2011). Nestin expression was lower incells transfected with nestin shRNA (Sh)than in cells transfected with ascrambled sequence shRNA (Sc) andnontreated (wild) cells (SupplementaryFigure S1d and e online), confirming theefficiency of the knockdown. Controland nestin shRNA–transfected cellsexhibited similar morphology (Supple-mentary Figure S2a online), whereasF-actin polymerization was increasedin the cytoplasm of Sh cells comparedwith Sc and wild cells (SupplementaryFigure S2c online, arrows), which wassimilar to that observed in pancreaticcancer cells (Matsuda et al., 2011).Super-high-resolution images clearlyrevealed the formation of F-actin fibersin Sh cells (Supplementary Figure S2donline, arrows). Nestin and F-actincolocalized to the periphery of Sc cells(Supplementary Figure S2d online,arrowheads), but not in Sh cells. Thecolocalization of nestin and F-actinobserved in this study suggests that

nestin regulates F-actin organization inmelanoma cells in a manner similar tosynemin, an intermediate filament pro-tein, which was reported to modulateactin dynamics (Pan et al., 2008).

The growth of Sh cells was lower thanthat of Sc cells, as indicated by manualcounting of cell number (Figure 1a). AWST-8 assay with Sh cells confirmedthese results (Figure 1b). Nestin hasbeen reported to regulate cell prolifera-tion through the mitogen-activated pro-tein kinase (MAPK) signaling pathway(Johannessen et al., 2009; Xue andYuan, 2010); therefore, we analyzedalterations of MAPK using Phospho-MAPK Arrays. Phosphorylated AKTwas the only protein affected by nestinsilencing; no inhibitory effects wereobserved for ERK, p38, JNK, or CREB.Decreased phosphorylation of AKT inSh cells was confirmed by westernblotting (Figure 1c). This finding suggeststhat AKT activation is important fornestin function.

The effect of nestin silencing onmelanoma cell migration and invasionwas next analyzed using a modifiedBoyden chamber assay. The migrationand invasion of Sh cells into thematrigel layer was less than that of Sccells (Figure 1d and e). Next, weanalyzed the effect of nestin silencingusing siRNA targeting nestin (siA;Supplementary Figure S1f online).Knockdown of nestin with siRNA inA375 cells also suppressed cell growth

Abbreviations: F-actin, filamentous-actin; MAPK, mitogen-activated protein kinase; Sc, scrambledsequence shRNA; Sh, nestin shRNA; shRNA, short hairpin RNA; siRNA, small interfering RNA

1384 Journal of Investigative Dermatology (2013), Volume 133

M Akiyama et al.Nestin in Malignant Melanoma

(Figure 1f), migration (Figure 1g), andinvasion (Figure 1h).

We performed a sphere-formationassay to determine in vitro tumorigeni-city of cells after treatment withnestin shRNA (Santini et al., 2012).The number of spheres in Sh cells waslower than that in Sc and wild cells(Figure 2a (arrows) and b). The spheresin A375 cells expressed nestin protein(Figure 2c, right panel; green fluore-scence) and higher levels of nestinmRNA than non-sphere cells(Figure 2d).

In a xenotransplantation experimentusing nude mice, subcutaneoustumors originating from Sh cells weresmaller and weighed significantly

less (Figure 2e) compared with thoseoriginating from Sc cells (Po0.05).Moreover, after tail vein injection ofthese cells into nude mice, liver meta-stases were observed in Sc mice, but notin Sh mice (Figure 2f, upper panels;arrows). Immunohistochemical stainingof liver sections using an anti-HLAantibody showed that the liver areaoccupied by the tumor was less in theSh group than in the Sc group (Figure 2f,lower panels, arrows; and Figure 2g).

We also examined the effects ofreduced nestin expression on cell pro-liferation, migration, and invasion usingnestin-targeting siRNAs (siA and siB) inmelanoma cells that express differentlevels of nestin (Supplementary Figure

S3a online). These siRNAs reduced theexpression of nestin in MeWo cells,which strongly express nestin mRNA,but did not reduce the expression inG361 cells (Supplementary Figure S3aonline). The effect of nestin siRNAon cell growth was more apparent inMeWo cells compared with the othercell types (Supplementary Figure S3bonline). In addition, cell movement,migration, and invasion were significantlyinhibited in MeWo cells (Supple-mentary Figure S3c and d online), butnot in G361 cells (Supplementary FigureS3c and e online), after treatment withnestin siRNA. However, these findingsdo not eliminate the possibility thatdecreased proliferative abilities have an

× 105

8 1.5

p-AKT

Sc Sh

AKT

GAPDH

1.0

0.5

0

Abs

orba

nce

at 4

50 n

m

**

6

4

2

0

Wild

Wild

Wild

Wild siN siA W

ild siN siA Wild siN siA

50

2.0 *

1.0

0

2.0 30

20

10

0

**

*1.0

0

Abs

orba

nce

at 4

50 n

m

Mig

rate

d ce

llspe

r fie

ld

Inva

ded

cells

per

field

30

20

10

0

Inva

ded

cells

per

fiel

d

0Mig

rate

d ce

lls p

er fi

eld

Sc

Sc

Sh

Wild

Sc

Sh

Sh Wild Sc Sh

**

**

Sc ShW

ild Sc Sh

Cel

l num

ber

per

wel

l

Figure 1. Cell growth and motility in nestin shRNA or nestin siRNA–transfected A375 human melanoma cells. (a) Manual cell count growth assay and

(b) WST-8 cell growth assay of wild-type (Wild), scrambled sequence RNA–transfected (Sc), and nestin shRNA–transfected (Sh) A375 cells. (c) Western blot

of phosphorylated AKT, total AKT, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). (d) Boyden chamber assay for cell migration, and (e) cell invasion

through matrigel. Bar¼ 100mm. (f) WST-8 cell growth assay of wild, negative control siRNA–transfected (siN), and nestin siRNA–transfected (siA) A375 cells.

(g) Cell migration and (h) invasion in nestin siRNA–transfected A375 cells. *Po0.05 and **Po0.01 vs. wild, Sc, or siN cells.

www.jidonline.org 1385

M Akiyama et al.Nestin in Malignant Melanoma

effect on metastasis and sphere forma-tion of melanoma cells.

We next tested the effects of nestinsiRNA on the cell behavior of non-tumorigenic cells. Human epidermalmelanocytes (HEMn-LP) expressed lowlevels of nestin (Supplementary FigureS3a online). Moreover, nestin siRNAtransfection of these cells did notdecrease nestin mRNA levels and hadno effect on cell growth and motility(Supplementary Figure S4 online).

Tumor stem cells that possess self-renewal ability and multilineage poten-tial have been considered to haveimportant roles in tumor development(Reya et al., 2001). Melanoma stemcells seem to possess a strong ability to

efflux Hoechst33342 dye, which is acharacteristic of stem cells, and thesecells express nestin (Grichnik et al.,2006). Melanoma spheroid cells havebeen reported to possess a high numberof tumor stem–like cells (Fang et al.,2005). We found that sphere-formingmelanoma cells express high levels ofnestin, and that knockdown of nestinsuppresses the sphere-forming ability,suggesting a relationship betweennestin expression and melanomastem–like cell functions.

In conclusion, we have foundthat reducing the expression of nestinin melanoma cells decreases the cellgrowth, migration, invasion, and sphere-forming ability in vitro and tumor

growth and metastasis in vivo, in part,through alteration of the expressionpattern of F-actin and regulation of theMAPK pathway. Therefore, nestin maybe a therapeutic candidate for treatingmalignant melanoma.

CONFLICT OF INTERESTThe authors state no conflict of interest.

ACKNOWLEDGMENTSWe thank Masahito Hagio, Tetsushi Yamamoto,Taeko Suzuki, Yoko Kawamoto, Kiyoko Kawahara,and Yuji Yanagisawa (Departments of Pathologyand Integrative Oncological Pathology) for techni-cal assistance. This work was supported by Leave aNest, a Grant-in-Aid for Scientific Research (MA),and a Grant-in-Aid for Young Scientific Research(A, No. 22689038 to YM).

Michiko Akiyama1,2, Yoko Matsuda1,Toshiyuki Ishiwata1, Zenya Naito1 andSeiji Kawana2

1Departments of Pathology and IntegrativeOncological Pathology, Nippon MedicalSchool, Tokyo, Japan and 2Department ofDermatology, Graduate School of Medicine,Nippon Medical School, Tokyo, JapanE-mail: ishiwata@nms.ac.jp

SUPPLEMENTARY MATERIAL

Supplementary material is linked to the onlineversion of the paper at http://www.nature.com/jid

REFERENCES

Amoh Y, Kanoh M, Niiyama S et al. (2009)Human and mouse hair follicles containboth multipotent and monopotent stem cells.Cell Cycle 8:176–7

Brychtova S, Fiuraskova M, Hlobilkova A et al.(2007) Nestin expression in cutaneousmelanomas and melanocytic nevi. J CutanPathol 34:370–5

Fang D, Nguyen TK, Leishear K et al. (2005) Atumorigenic subpopulation with stem cellproperties in melanomas. Cancer Res 65:9328–37

Fusi A, Reichelt U, Busse A et al. (2011) Expressionof the stem cell markers nestin and CD133 oncirculating melanoma cells. J Invest Dermatol131:487–94

Grichnik JM, Burch JA, Schulteis RD et al. (2006)Melanoma, a tumor based on a mutant stemcell? J Invest Dermatol 126:142–53

Johannessen TC, Wang J, Skaftnesmo KO et al.(2009) Highly infiltrative brain tumours showreduced chemosensitivity associated with astem cell-like phenotype. Neuropathol ApplNeurobiol 35:380–93

Klein WM, Wu BP, Zhao S et al. (2007) Increasedexpression of stem cell markers in malignantmelanoma. Mod Pathol 20:102–7

Lendahl U, Zimmerman LB, McKay RD (1990)CNS stem cells express a new class of

Sc Sh

100

*

*

50

0

3.0*

2.0

1.0

0

Non-s

pher

e

Spher

eNes

tin/1

8S r

RN

A

Tum

or w

eigh

t (g)

1*

0

Wild Sc Sh

Sc Sh

(%)

2

1

0Met

asta

tic a

rea

Wild Sc Sh

Wild Sc Sh

Num

ber

ofsp

here

s pe

r w

ell

Figure 2. Sphere formation and in vivo analyses in nestin shRNA–transfected A375 cells. (a) Phase-

contrast images of spheres. Bar¼ 100mm. (b) Number of spheres per well. *Po0.05 vs. scrambled

sequence RNA–transfected cells (Sc). (c) Phase-contrast (left, bar¼ 10mm) and fluorescence images of

nestin (right, bar¼1mm) in A375 spheres. (d) The quantitative reverse transcription-PCR analysis of nestin

mRNA in sphere-forming cells and cultured cells in plates (non-sphere). *Po0.05 vs. non-sphere cells. (e)

Weight of subcutaneous tumors. *Po0.05 vs. Sc cells. (f) Liver metastases in mice injected with A375 cells

through the tail vein. Macroscopic images (upper panels) and immunohistochemistry of HLA in the liver

(lower panel). Upper panels, bar¼ 1 mm; lower panels, bar¼ 100mm. (g) The percentage of the metastatic

tumor area in the liver. *Po0.05 vs. wild cells. Sh, nestin shRNA–transfected cells.

1386 Journal of Investigative Dermatology (2013), Volume 133

M Akiyama et al.Nestin in Malignant Melanoma

intermediate filament protein. Cell 60:585–95

Matsuda Y, Naito Z, Kawahara K et al. (2011)Nestin is a novel target for suppressing pan-creatic cancer cell migration, invasion andmetastasis. Cancer Biol Ther 11:512–23

Pan Y, Jing R, Pitre A et al. (2008) Intermediatefilament protein synemin contributes to themigratory properties of astrocytoma cells byinfluencing the dynamics of the actin cyto-skeleton. FASEB J 22:3196–206

Piras F, Perra MT, Murtas D et al. (2010) The stemcell marker nestin predicts poor prognosis inhuman melanoma. Oncol Rep 23:17–24

Reya T, Morrison SJ, Clarke MF et al. (2001) Stemcells, cancer, and cancer stem cells. Nature414:105–11

Santini R, Vinci MC, Pandolfi S et al. (2012)HEDGEHOG-GLI signaling drives self-renewal and tumorigenicity of humanmelanoma-initiating cells. Stem Cells30:1808–18

Sellheyer K, Krahl D (2010) Spatiotemporal expres-sion pattern of neuroepithelial stem cell mar-ker nestin suggests a role in dermalhomeostasis, neovasculogenesis, and tumorstroma development: a study on embryonicand adult human skin. J Am Acad Dermatol63:93–113

Xue XJ, Yuan XB (2010) Nestin is essential formitogen-stimulated proliferation of neuralprogenitor cells. Mol Cell Neurosci 45:26–36

Coculture Model of Sensory Neurites and Keratinocytes toInvestigate Functional Interaction: Chemical Stimulationand Atomic Force Microscope–Transmitted MechanicalStimulation Combined with Live-Cell ImagingJournal of Investigative Dermatology (2013) 133, 1387–1390; doi:10.1038/jid.2012.471; published online 13 December 2012

TO THE EDITORThere is growing evidence for func-tional interactions between nociceptivenerve endings and non-neuronal cellsmodulating sensory function in healthand disease. In the skin, non-neuronalcells like keratinocytes actively partici-pate in nociceptor sensitization and thusin encoding of noxious stimuli (Lumpkinand Caterina, 2007; Gold and Gebhart,2010). In vivo approaches for live-cellimaging of the interplay of sensoryterminals and surrounding cells arehampered by the small diameter of theendings and the difficult experimentalaccessibility. Moreover, selectivelystimulating one type of cell whilerecording responses from both ischallenging. Our goal was therefore toestablish an in vitro model that wouldallow studying potential functionalinteractions between nociceptiveterminals and keratinocytes.

To this end, we developed a cocul-ture model of sensory endings andkeratinocytes in a compartmented Cam-penot chamber (Figure 1a; Campenotet al., 2009; Roggenkamp et al., 2012).Compartmented chambers for coculture(Chateau et al., 2007; Roggenkamp

et al., 2012) have the advantage of aspatial segregation of neurites (Ns) fromtheir somata (Figure 1a) compared witha coculture in one compartment(Ulmann et al., 2009; Pereira et al.,2010a,b). Moreover, in the Campenotchamber, fluid isolation between thecompartments allows different culturemedia and factors for neuronalsomata in one compartment andfor Ns and keratinocytes in the other,mimicking the in vivo structural andenvironmental conditions. In ourexperiments, isolated somata ofporcine dorsal root ganglion neurons(Obreja et al., 2008) were grown inthe central compartment; Ns outgrowninto the lateral compartments served asa model for sensory endings. TheseNs were cocultured with primaryisolated porcine keratinocytes (seeSupplementary information).

Using this approach, we obtainedareas of confluent keratinocytes in spa-tial contact with sensory Ns (Figure 1band c). By using atomic force micro-scopy (AFM), we visualized nanoscaledsurface topographies of adjacent Ns andkeratinocytes (Figure 1d). In the past,AFM had been established as a tool for

the characterization of cell–cell contactscomprising also the desmosomal junc-tions between the keratinocytes (Funget al., 2010). In comparison, our imagessuggest cell–cell contacts not onlybetween keratinocytes but also possiblybetween Ns and keratinocytes. For func-tional investigations, we implementedlive-cell imaging combined with chemi-cal or mechanical stimulation, allowingselective activation of one cell typewhile recording responses simulta-neously from both types. Responses tothese stimuli were visualized using thenon-ratiometric calcium dye Fluo8 acet-oxymethyl ester. To activate only Ns, amembrane-depolarizing concentrationof KCl was added to the somata contain-ing central compartment. The induceddepolarization is transmitted along theNs into the lateral compartment,indicated by a calcium response(Supplementary Figure S1a online). Forspecific activation of nociceptive Ns,capsaicin was added to the lateral com-partment (Supplementary Figure S1bonline; Caterina et al., 2007).

In coculture of Ns and keratinocytesin the lateral compartment, we observedan interaction between both cell types.In the experiment shown in Figure 2a–d,KCl was first applied to the centralcompartment. After an immediateAbbreviations: AFM, atomic force microscopy; N, neurite

www.jidonline.org 1387

A Klusch et al.Coculture Model of Sensory Neurites and Keratinocytes