Muhr P, Renne J, Schaefer V et al. (2010) Primaryhuman keratinocytes efficiently induceIL-1-dependent IL-17 in CCR6þ T cells. ExpDermatol 19:1105–7

Nicol C, Cesur O, Forrest S et al. (2013) AnRNA aptamer provides a novel approachfor the induction of apoptosis by targeting

the HPV16 E7 oncoprotein. PLoS One 8:e64781

Reyes-Reyes EM, Teng Y, Bates PJ (2010) A newparadigm for aptamer therapeutic AS1411action: uptake by macropinocytosis andits stimulation by a nucleolin-dependentmechanism. Cancer Res 70:8617–29

Wittmann M, Doble R, Bachmann M et al.(2012) IL-27 regulates IL-18 binding proteinin skin resident cells. PLoS One 7:e38751

Zhou J, Bobbin ML, Burnett JC et al. (2012) Currentprogress of RNA aptamer-based therapeutics.Front Genet 3:234

Epidermal CCL27 Expression Is Regulated during SkinDevelopment and Keratinocyte DifferentiationJournal of Investigative Dermatology (2014) 134, 855–858; doi:10.1038/jid.2013.394; published online 17 October 2013

TO THE EDITORThe adult human skin is a highly activeimmunological organ containing con-siderable numbers of T cells (Clarket al., 2006), which are fundamentalfor the maintenance of barrier integrityand immune surveillance (Kupperand Fuhlbrigge, 2004). The greatmajority of resident T cells exhibit amemory/effector phenotype and expressthe skin-homing marker cutaneouslymphocyte� associated antigen (CLA;Clark et al., 2006) and severalchemokine receptors (CCR4, CCR6,CCR8, and CCR10; Campbell et al.,1999; Homey et al., 2002; Schaerliet al., 2004; McCully et al., 2012).The CCR10 ligand CCL27/CTACK(cutaneous T-cell-attracting chemokine)is specifically expressed by epidermalkeratinocytes (KCs), and its over-production is associated with anenhanced influx of CCR10þCD4þ

T cells into the skin in patients withinflammatory skin diseases such asatopic or allergic contact dermatitisand psoriasis (Homey et al., 2002).This suggests a distinct role for CCR10in controlling effector T–cell migrationunder specific inflammatory conditions(Reiss et al., 2001; Homey et al., 2002;Sigmundsdottir et al., 2007) as opposedto the recently proposed homeostaticimmune surveillance T-cell trafficking(McCully et al., 2012). Recently,Di Nuzzo et al. (2009) reported that,

in contrast to adult skin, fetal humanskin harbors mainly naive CLA�CD4þ

T cells and less CLAþ T cells.Confirming and extending this work,we have shown that only low numbersof CD3þ T cells (up to 20 times loweras compared with adult skin) werepresent in the skin during midgestationand only rare T cells during the firsttrimester, which were exclusivelylocated in the dermal compartment(Schuster et al., 2012). After birth,memory T cells can be readilyidentified in the dermis and in theepidermis and their number increaseswith age (Akgun et al., 2012). As it isunclear which factors/chemokinescontrol T-cell immigration into the skinduring ontogeny, we investigated in thisstudy whether the scarcity of T cells inprenatal human skin correlates withCCL27 levels produced by epidermalKCs. To this purpose, we analyzed itsexpression and regulation in prenataland adult human KCs in monolayerand in organotypic skin cultures in vitro.

In line with a previous publication(Homey et al., 2002), we detectedstrong CCL27 expression in adultepidermis preferentially in basal KCsand the stratum corneum (Figure 1c,inset). The specificity of the epidermalstaining was confirmed by blockingexperiments using recombinant CCL27for pre-absorption of the antibody(Supplementary Figure S1 online).

CCL27 staining of dermal cells, how-ever, was not inhibited providing furtherevidence for an indirect CCL27 reactiv-ity of dermal cells via the binding ofCCL27 to its receptor CCR10 as sug-gested previously (Homey et al., 2002).In contrast, epidermal CCL27 wasvirtually absent in embryonic skin(12 weeks estimated gestational age(EGA) (Figure 1a) and was marginallypositive in fetal skin (24 weeks EGA),showing strongest expression in theuppermost layer (Figure 1b), indicatinga differentiation-dependent regulation ofCCL27 in KCs. To compare the CCL27secretory potential of fetal and adultKCs, we cultured primary KCs for24 hours in the absence or presence ofthe toll-like receptor-3 ligand poly(I:C),a potent inducer of a variety of chemo-kines including CCL27 (Lebre et al.,2007). We found that CCL27 mRNAwas comparably expressed in both,unstimulated fetal and adult KCs(Figure 1d, Ct-value¼29), whereasrelease of CCL27 protein was detectableonly in adult KCs (Figure 1e). Poly (I:C)induced a strong increase in CCL27expression and secretion in adult KCsand a moderate, although significant,increase in fetal KCs (Figure 1d and e).When analyzing lysates of fetal andadult KCs, we found a significant intra-cellular expression level of CCL27 notonly in KCs from adult but also fromthose of fetal donors (Figure 2c). Thestriking differences in the CCL27 stain-ing pattern between prenatal and adultskin together with the fact that theprenatal epidermis is not fully differen-tiated yet (Dale et al., 1985) promptedAccepted article preview online 13 September 2013; published online 17 October 2013

Abbreviations: CLA, cutaneous lymphocyte–associated antigen; EGA, estimated gestational age; KC,keratinocyte

M Mildner et al.CCL27 Secretion Profile during Skin Development

www.jidonline.org 855

us to further investigate the effect ofKC differentiation on CCL27 produc-tion and secretion in vitro. As shownin Figure 2, CCL27 mRNA (Figure 2a)and release of CCL27 protein (Figure 2b)were strongly increased in differentiatedadult KCs. In contrast, induction ofdifferentiation in fetal KCs failed toenhance CCL27 production and release(Figure 2a–c). Analysis of the differentia-tion-associated proteins loricrin and ker-atin 10 (Supplementary Figure S2a and bonline) revealed that in KCs of both agegroups, markers of late KC differentia-tion were readily induced, althoughexpression levels were lower in fetalcells. Although our results suggest thatdifferentiation of KCs may influenceCCL27 release, our assumption needsto be supported by detailed biochemicalanalyses. In contrast to fetal KCs inmonolayer cultures in vitro (Figure 2a–c), CCL27 expression was distinctlypresent in the upper layers of fetalepidermis (Figure 1b), suggesting thatin vivo additional factors might be pre-sent and necessary for CCL27 upregula-tion. These findings extend our previous

work, which showed that CCL27 is notreleased by cultured fetal skin single-cell suspensions either (Schuster et al.,2012). Furthermore, our observationthat differentiation strongly increasesCCL27 release by adult KCs in vitro(Figure 2), whereas suprabasal KCs inadult skin (Figure 1c) show a weakerstaining intensity compared with basalKCs and the stratum corneum, mightindicate that also in vivo suprabasal,i.e., differentiated KCs strongly releaseCCL27 resulting in a weaker cell–associated immune reactivity.

To further elaborate the differentia-tion-dependent expression of CCL27 inadult KCs, we used an organotypic skinmodel, recently established in ourlaboratory (Mildner et al., 2010). Thisskin model resembles normal humanskin and can be investigated atdifferent time points of epidermaldifferentiation. After 1 day, the epider-mis is already multilayered but no stra-tum corneum is present and after 5 daysthe model shows all hallmarks of a fullydifferentiated epidermis including astratum corneum (Figure 2d and e).

High levels of CCL27 expression werefound in all epidermal layers already atthe early time point (Figure 2d and f).However, as this model reflects epider-mal regeneration (such as in woundhealing) rather than epidermal develop-ment during ontogeny, our data suggestthat an enhanced CCL27 release mayhave a role in regenerating epidermis.Indeed, the CCL27/CCR10 axis has beenimplicated to be important for epider-mal wound healing (Inokuma et al.,2006). The late differentiated organo-typic skin showed a strong CCL27expression (Figure 2e and f) and astaining pattern comparable to that ofadult human skin (Figure 1c).

In summary, we show that adult butnot prenatal human KCs produce andrelease large amounts of the T-cell che-moattractant CCL27. Moreover, wedemonstrate that epidermal CCL27secretion is strongly dependent on KCdifferentiation. Our data further implythat the triggers responsible for CCL27expression differ between epidermalregeneration of postnatal skin and epi-dermal development of prenatal skin.

12-Week EGA

E

D

30

100

Poly (I:C)Poly (I:C)

200500

1,500

2,500

20

FetalAdult

FetalAdult

***

***

***

******

***

**

CC

L27

(rel

. exp

ress

ion)

CC

L27

(pg

ml–1

)

10

00–– – –+ + ++

24-Week EGA Adult skin

Figure 1. Regulation of CCL27 expression in KCs during skin development. Immunofluorescence of cryostat sections from embryonic (a), fetal (b),

and adult (c) skin samples (n¼5) revealed no-to-modest CCL27 positivity in embryonic skin and a strong staining intensity of the outermost epidermal

layer in fetal skin, whereas in adult epidermis both the basal and the cornified layers displayed strong staining (D, dermis; E, epidermis; red: CCL27

staining, blue: nuclei staining; scale bars¼ 40mm). (d) Relative CCL27 mRNA expression of fetal and adult KCs in vitro is shown (n¼ 5). Quantitative reverse

transcription-PCR revealed similar mRNA levels in untreated fetal and adult KCs. Poly (I:C) induced a significantly higher CCL27 mRNA expression in adult

KCs as compared with fetal KCs. (e) CCL27 ELISA of supernatants from fetal and adult KCs is shown (n¼ 5). In contrast to fetal KCs, a low baseline CCL27 secretion

was observed in adult KCs. Poly (I:C) induced a strong upregulation of CCL27 secretion in adult but not in fetal KCs. Each value represents one independent

donor performed in triplicate. Bars represent the mean±SEM of all donors. **Po0.005, ***Po0.0005. KC, keratinocyte.

M Mildner et al.CCL27 Secretion Profile during Skin Development

856 Journal of Investigative Dermatology (2014), Volume 134

Thus, changes in the KC differentiationprocess as well as the microenvironmentafter birth seem to favor a gradient forthe influx of T cells suggesting a role forthe CCL27/CCR10 chemokine system inthe control of immigrating cutaneousT cells in the developing human skin.

CONFLICT OF INTERESTThe authors state no conflict of interest.

ACKNOWLEDGMENTSWe thank C. Fiala and W. Eppel for the organiza-tion of prenatal human skin samples and S. Karnerfor technical support. This project was funded, inpart, by on Austrian Science Foundation grant(p19474-B13 to A.E-B).

Michael Mildner1,3, Marion Prior2,3,Maria Gschwandtner1,Christopher Schuster2, Erwin Tschachler1

and Adelheid Elbe-Burger2

1Research Division of Biology and Pathobiologyof the Skin, Department of Dermatology,Medical University of Vienna, Vienna, Austriaand 2Laboratory of Cellular and MolecularImmunobiology of the Skin, Division ofImmunology, Allergy and Infectious Diseases,Department of Dermatology, MedicalUniversity of Vienna, Vienna, AustriaE-mail: adelheid.elbe-buerger@meduniwien.ac.at3These authors contributed equally to this work.

SUPPLEMENTARY MATERIAL

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

REFERENCES

Akgun J, Prior M, Elbe-Burger A (2012) Most T cellsin human neonatal skin are not naive. J ClinExp Dermatol Res S2:004

Campbell JJ, Haraldsen G, Pan J et al. (1999)The chemokine receptor CCR4 in vascular

recognition by cutaneous but not intestinalmemory T cells. Nature 400:776–80

Clark RA, Chong B, Mirchandani N et al. (2006)The vast majority of CLAþ T cells are residentin normal skin. J Immunol 176:4431–9

Dale BA, Holbrook KA, Kimball JR et al. (1985)Expression of epidermal keratins and filaggrinduring human fetal skin development. J CellBiol 101:1257–69

Di Nuzzo S, Pavanello P, Masotti A et al. (2009)Densities, distribution and phenotypic expres-sion of T cells in human fetal skin. ArchDermatol Res 301:753–5

Homey B, Alenius H, Muller A et al. (2002) CCL27-CCR10 interactions regulate T cell-mediatedskin inflammation. Nat Med 8:157–65

Inokuma D, Abe R, Fujita Y et al. (2006) CTACK/CCL27 accelerates skin regeneration via accu-mulation of bone marrow-derived keratino-cytes. Stem Cells 24:2810–6

Kupper TS, Fuhlbrigge RC (2004) Immune surveil-lance in the skin: mechanisms and clinicalconsequences. Nat Rev Immunol 4:211–22

600

400

CC

L27

(pg

ml–1

)

CC

L27

(pg

ml–1

)

CC

L27

(rel

. exp

ress

ion)

200

0

600

400

200

0

15

Fetal

**

***

***

*

***

**

**

Adult

FetalAdult

FetalAdult

10

5

0

Prol

Prol

Diff Diff Prol

Prol

Diff Diff Prol

Prol

Diff Diff

Lysate

Late differentiation

2,500

2,000

CC

L27

(pg

ml–1

)

1,500

500

Early Late0

1,000

Supernatant

Early differentiation

CC

L27

H&

E

Differentiation

Figure 2. CCL27 expression and secretion are regulated during KC differentiation. (a) CCL27 mRNA expression in proliferating (Prol) and differentiated

(Diff) fetal and adult KCs in vitro is shown (n¼5–6). Whereas KC differentiation led to an upregulation of CCL27 mRNA expression in adult KCs, no regulation

was observed in fetal KCs. Analysis of secreted (b) and intracellular (c) CCL27 by ELISA showed an enhanced secretion in adult but not in fetal KCs (n¼3).

(d and e) Organotypic skin either at an early- or a late-differentiation state (n¼ 3) is shown. H&E staining displayed a multilayered but not cornified epithelium

in the early-differentiated sample (d, upper panel) and a fully differentiated epidermis also showing a cornified layer in the late-differentiated sample

(e, upper panel). Immunofluorescence of cryostat sections from organotypic skin revealed a strong staining already in all layers of the early-differentiated sample

(d, lower panel) and a staining pattern similar to adult human skin in the late-differentiated sample (e, lower panel; red: CCL27 staining, blue: nuclei staining,

scale bars¼ 40mm). (f) CCL27 ELISA of secreted CCL27 from organotypic skin cultures is shown (n¼ 3 for each group). Each value represents one independent

donor, each done in triplicate. Bars represent the mean±SEM of all donors. *Po0.05, **Po0.005, ***Po0.0005. H&E, hematoxylin and eosin; KC, keratinocyte.

M Mildner et al.CCL27 Secretion Profile during Skin Development

www.jidonline.org 857

Lebre MC, van der Aar AM, van Baarsen L et al.(2007) Human keratinocytes express func-tional Toll-like receptor 3, 4, 5, and 9. J InvestDermatol 127:331–41

McCully ML, Ladell K, Hakobyan S et al. (2012)Epidermis instructs skin homing receptorexpression in human T cells. Blood 120:4591–8

Mildner M, Jin J, Eckhart L et al. (2010) Knockdownof filaggrin impairs diffusion barrier function

and increases UV sensitivity in a human skinmodel. J Invest Dermatol 130:2286–94

Reiss Y, Proudfoot AE, Power CA et al. (2001) CCchemokine receptor (CCR)4 and the CCR10ligand cutaneous T cell-attracting chemokine(CTACK) in lymphocyte trafficking to inflamedskin. J Exp Med 194:1541–7

Schaerli P, Ebert L, Willimann K et al. (2004) Askin-selective homing mechanism for human

immune surveillance T cells. J Exp Med199:1265–75

Schuster C, Vaculik C, Prior M et al. (2012) Pheno-typic characterization of leukocytes in prenatalhuman dermis. J Invest Dermatol 132:2581–92

Sigmundsdottir H, Pan J, Debes GF et al. (2007)DCs metabolize sunlight-induced vitamin D3to ‘program’ T cell attraction to the epidermalchemokine CCL27. Nat Immunol 8:285–93

Identifying a Hyperkeratosis Signature in AutosomalRecessive Congenital Ichthyosis: Mdm2 InhibitionPrevents Hyperkeratosis in a Rat ARCI ModelJournal of Investigative Dermatology (2014) 134, 858–861; doi:10.1038/jid.2013.374; published online 17 October 2013

TO THE EDITORAutosomal recessive congenital ichthyo-sis (ARCI) is a Mendelian disorder ofkeratinization. LI patients are born aserythrodermic ‘‘collodion’’ babies butsoon become covered with scales. Themost commonly mutated genes in ARCIare Transglutaminase 1, Ichthyin, andthe Arachidonate Lipoxygenases 12band 3 (Russell et al., 1995; Jobardet al., 2002; Fischer, 2009). Hyper-keratosis (scaling) is a cardinal featureof disorders of keratinization, which arecharacterized by abnormal barrierfunction. For example, Tgm1-null micedie perinatally because of transepi-dermal water loss after birth, butwhen this skin is grafted onto nudemice, they form scales similar to ARCIpatients (Kuramoto et al., 2002). Theepidermis is hyperproliferative, pro-duces increased non-polar lipid, andthe cornified layer is thickened. Thisscaling is the principle phenotype inARCI patients, and the time taken toremove scale significantly affects theirquality of life. Oral retinoids, thecurrent treatment for ARCI, efficientlyreduces scale formation but are alsorarely associated with side effects,some serious (Rood et al., 2007).

Therefore, there is a need to developnew treatments for ARCI.

Current research on alternate ARCItreatments is focused on gene therapy(Choate et al., 1996), or enzyme replace-ment therapy, adding recombinant Tgm1to the skin (Aufenvenne et al., 2012). Agreater understanding of the commonmolecular mechanisms that lead to hyper-keratosis would facilitate the discovery ofmore directed therapies to reduce scalingin ARCI. Therefore, we investigated com-mon gene expression changes in twodifferent rat epidermal keratinocyte ARCImodels, to identify the common net-works of differentially expressed genesthat are required for hyperkeratosis and,within these molecular networks, to iden-tify new ‘‘drug-able’’ targets for the reduc-tion of hyperkeratosis.

Two separate Arachidonate 12-lipoxy-genase (Alox12b) short hairpin RNA(shRNA)–knockdown (kd) rat epider-mal keratinocyte lines (SupplementaryFigure S1a and S1b online) were com-pared with an existing Tgm1 shRNA–kdline (O’Shaughnessy et al., 2010). BothTgm1- and Alox12b-kd organotypiccultures were hyperkeratotic (Figure 1a).Comparison of Tgm1-kd with Alox12b-kdcultures revealed a common increase in

keratin 1 expression, but keratin 10 reduc-tion was observed only in Tgm1-kdcultures (Figure 1a and SupplementaryFigure S1c online), suggesting both com-mon and distinct programs of altered geneexpression in the two models. Prin-cipal component analysis confirmed that,although phenotypically similar, overallgene expression differences were diver-gent (Figure 1b and Supplementary FigureS2 online). Despite these differences, weidentified 155 genes that were commonlyupregulated and 78 genes that were com-monly downregulated in both Tgm1- andAlox12b-kd keratinocyte cultures (Supple-mentary Table S1 online). We identifiedupregulated genes related to the positiveregulation of apoptosis and downregulatedgenes related to protein phosphorylation(Supplementary Figure S3 online).

The only significant network of inter-related genes was related to the genesinvolved in the positive regulation ofapoptosis, and within this network weidentified the p53 ubiquitin ligaseMdm2 as an attractive candidate geneto target in order to reduce hyperker-atosis, as mice overexpressing Mdm2display a hyperkeratotic epidermis in ap53-independent manner (Alkhalafet al., 1999). Mdm2 upregulation inboth Tgm1- and Alox12b-kd cultureswas confirmed by quantitative reversetranscriptase in real-time PCR (Figure 1cand d). In organotypic culture, cytoplasmicAccepted article preview online 4 September 2013; published online 17 October 2013

Abbreviations: ARCI, autosomal recessive congenital ichthyosis; kd, knockdown; pAkt, phosphorylated Akt

G Youssef et al.Mdm2 Upregulation in Autosomal Recessive Congenital Ichthyosis

858 Journal of Investigative Dermatology (2014), Volume 134