Long-Term Faithful Recapitulation of Transglutaminase1–Deficient Lamellar Ichthyosis in a Skin-HumanizedMouse Model, and Insights from Proteomic StudiesJournal of Investigative Dermatology (2012) 132, 1918–1921; doi:10.1038/jid.2012.65; published online 22 March 2012

TO THE EDITORTransglutaminase 1 (TG1)–deficient la-mellar ichthyosis (LI) is associated withincreased mortality in the neonatalperiod and has a marked impact onquality of life. No efficient treatment isavailable; current therapy only relievessome symptoms (Oji and Traupe,2006). For the development of newtherapeutic approaches and to furtherinvestigate molecular mechanisms un-derlying the pathophysiology of LI, astable, long-lived preclinical model isneeded, which fully recapitulates thehuman skin phenotype.

In vivo studies in human skin arelimited by ethical and practical consid-erations. The Tgm1�/� mouse does notrecapitulate the human skin phenotype.Moreover, the mice die within the firsthours of life (Matsuki et al., 1998)because of impaired barrier function,underscoring the importance of TG1 forbarrier formation (Candi et al., 2005).Only transplanted Tgm1–/– mouse skinresembled the skin seen in severeichthyosis (Kuramoto et al., 2002). Cho-ate and colleagues showed the feasibilityof ex vivo and in vivo gene transfer for LI(Choate et al., 1996; Choate and Kha-vari, 1997), although their model systemonly allowed short-term human skinregeneration. Moreover, a model of ratskin with TG1 deficiency was described(O’Shaughnessy et al., 2010).

Using optimized tissue engineeringand surgical conditions enabling stablehuman skin engraftment in athymic nudemice (Garcıa et al., 2011), we have beenable to develop a robust skin-human-ized mouse model for TG1-deficient

LI, involving persistent engraftment ofbioengineered human skin, suitablefor long-term, preclinical studies andfor the investigation of molecular me-chanisms.

We analyzed two LI patients andidentified the compound heterozygousTGM1 mutations c.377G4A (p.Arg126His) and c.876þ 2T4C (p.Glu253Valfs*2; patient 1) and c.428G4A(p.Arg143His) and c.877-2A4G(p.Phe293Serfs*38 or p.Phe293Valfs*2;patient 2), which were previously de-scribed (Farasat et al., 2009). Punchbiopsies obtained from these patientswere used to isolate keratinocytes andfibroblasts after obtaining written, in-formed consent of the probands andinstitutional approval in accordancewith the Declaration of Helsinki Princi-ples. At 4–6 weeks after grafting ofbioengineered skin equivalents, regen-erated human skin grafts became visibleand persisted in the recipient animalslonger than 20 weeks, indicating stableengraftment of epidermal stem cells(data not shown).

Macroscopically, epidermal hyper-plasia and an increased scaly hyperker-atosis completely matched the humanskin phenotype. Using a human-speci-fic antibody against involucrin, wecould confirm the human origin of thegrafts and clearly delineate the borderbetween mouse and human skin. Lightmicroscopy displays a very thick andpacked stratum corneum (SC) in the LIgrafts (Figure 1a).

TG1 activity and protein were absentin LI skin/grafts. Ultrastructural analysisrevealed cholesterol clefts in the SC of

the LI skin/grafts, which are importantdiagnostic markers typical for TG1deficiency (Pigg et al., 1998). In contrast,skin/grafts derived from healthy indivi-duals were completely normal (Figure 1b).

The presence and spatial distributionof TG1 substrates, differentiation mar-kers, and dermoepidermal junction(DEJ) constituents were assessed byimmunostaining. The comparison be-tween normal graft/skin and LI graft/skin showed that normal grafts mirrorednormal skin and, similarly, alterationsof epidermal differentiation in LI graftsmatched those found in LI skin. Thediffuse and upwardly shifted distribu-tion of loricrin and involucrin indicatesinsufficient cross-linking of these pro-teins into the cornified envelope (CE).Filaggrin was expressed in the periph-ery of cells from the upper spinous layerin all samples. The early differentiationmarker keratin 10, localized in viablesuprabasal cell layers, remained un-changed (Figure 1c). Integrin-a6 dec-orates the DEJ in a continuous, linearmanner. Collagen VII was present alongthe DEJ, indicating its correct formationin the grafts (Figure 1d).

To further explore changes in differ-entiation and to relate them to theprevious results, we present, to ourknowledge previously unreported, dataon the analysis of the proteome of LIepidermis. Comparison between normaland LI skin/grafts provides valuable in-sights into molecular mechanisms in-volved in LI pathophysiology. Analysis ofthe skin-humanized mouse model yieldsvery similar data to those observed in LIskin. Altogether, 147 proteins wereidentified (Supplementary Table S1 on-line). We focused on TG1 substrates andsome conspicuous proteins, which may

LETTERS TO THE EDITOR

1918 Journal of Investigative Dermatology (2012), Volume 132 & 2012 The Society for Investigative Dermatology

Abbreviations: CE, cornified envelope; DEJ, dermoepidermal junction; KPRP, keratinocyte proline-richprotein; LI, lamellar ichthyosis; NMF, natural moisturizing factor; TEWL, transepidermal water loss; TG1,transglutaminase 1

give insights into molecular differentia-

tion mechanisms (Table 1).The numbers of unique peptides from

identified proteins in the grafts werehighly consistent with those found inhuman skin biopsies. No TG1 peptidescould be detected in LI skin. The fourTG1 peptides detected in one of three LIgrafts (Table 1) were likely derivedfrom surrounding mouse tissue, as thesequences are identical in mouse andhuman.

Notably, the number of unique filag-grin peptides decreased markedly inboth LI skin and grafts. Profilaggrin isproteolytically processed during kerati-nocyte differentiation and subsequentlydegraded into hydrophilic amino acids,

their metabolites, and ions that contri-bute to moisture retention in the SC.Expression of filaggrin and its hydrolysisinto these natural moisturizing factors(NMFs) are influenced by the SC micro-environment, including local pH, exter-nal humidity, and transepidermal waterloss (TEWL) (O’Regan et al., 2008). Thereduction of filaggrin detected by pro-teome analysis in LI skin/grafts may beattributed to an increased hydrolysis intoNMFs to compensate for TEWL.

In contrast, the number of uniqueloricrin peptides increased in LI incomparison with normal skin/grafts.Loricrin becomes extensively cross-linked to numerous CE componentsby different TGs (Candi et al., 2005).

Insufficient intermolecular oligomeriza-tion by TG1 could result in an enhancedaccessibility of trypsin cleavage sitesduring sample processing, resulting inan increase of unique peptides. Simi-larly, a marked increase of uniquepeptides of keratinocyte proline-richprotein (KPRP) was found in LI skin/grafts. KPRP expression was markedlyincreased in psoriatic lesions, suggestingthat it could be extensively cross-linkedby TG1 like small proline-rich proteinsor some late CEs (Lee et al., 2005).

Interestingly, we did not observechanges in the number of uniquepeptides of involucrin, one of the firstproteins to be cross-linked to initiate CEassembly by forming a monomolecular

Mouse Human

lor

int-α6

col VII col VII col VII col VII

int-α6 int-α6 int-α6

lor lor lor

inv inv inv inv

flg flg flg flg

K10 K10 K10 K10

SCSC

LI graft

LI skin

CCCC

CC

LI graft

Normal graft

Normal skin

Ultr

astr

uctu

reT

G1-

prot

ein

TG

1-ac

tivity

Normal graft

LI skin LI graftNormal skin Normal graft

LI skin LI graftNormal skin Normal graft

i i ii iii iv

v vi vii viii

i ii iii iv

v vi vii viii

ix x xi xii

i ii iii iv

v vi vii viii

ix x xi xii

xiii xiv xv xvi

iv v

ii iii

Figure 1. Characterization of the skin-humanized mouse model for transglutaminase 1–deficient lamellar ichthyosis. (a) Keratinocytes were seeded onto a

fibroblast-populated fibrin-based matrix. Skin equivalents were grafted orthotopically onto the back of athymic nude mice. (i) Normal human regenerated skin

and (ii) regenerated lamellar ichthyosis (LI) skin (patient 1) 12 weeks after grafting; (iii) peroxidase staining of involucrin (human specific) confirms the

human skin phenotype; (iv) semithin sections, methylene blue staining, and morphology of normal and (v) LI grafts showing a very thick and packed stratum

corneum (SC). Bar: iv, v¼ 50mm. (b i–iv) Transglutaminase 1 (TG1) activity, (v–viii) TG1 protein, and (ix–xii) ultrastructure. Normal skin/grafts show the typical

pericellular distribution of TG1 activity/protein in the stratum granulosum. Ultrastructurally, no cholesterol clefts are visible. In contrast, LI skin and LI grafts

lacked TG1 activity/protein but displayed cholesterol clefts as typical ultrastructural markers. (c i–iv) Characterization of human and regenerated skin by

immunostaining. TG1 substrates such as loricrin (lor), (v–viii) involucrin (inv), and (ix–xii) filaggrin (flg) in LI samples show a more diffuse and slightly shifted

staining pattern when compared with the normal samples. (xiii–xvi) Keratin 10 (K10) was expressed in suprabasal layers in all four samples. (d i–iv) The

distributions of integrin-a6 (int-a6) and (v–viii) collagen VII (col VII) in LI skin/grafts in comparison with normal skin/grafts are visualized by immunostaining.

These components of the dermoepidermal junction show a comparable staining in all samples, indicating a correct formation of the junction zone. Slides were

counterstained with 40,6-diamidino-2-phenyl indole. CC, cholesterol clefts.

www.jidonline.org 1919

K Aufenvenne et al.Skin-Humanized Mouse Model of TG1 Deficiency

layer adjacent to the cell membrane(Candi et al., 2005). We speculate thatthis location does not alter its accessi-bility to trypsin.

Desmoplakin, a structural cytoskele-ton constituent, which participates inkeratinocyte adhesion, and other des-mosomal/corneodesmosomal proteinssuch as desmoglein 1, desmocollin 1,junction plakoglobin, and corneodes-mosin display a marked increase inunique peptide numbers in LI skin/grafts. From a clinical perspective, it islikely that increased expression ofdesmosomal adhesion molecules re-sults in increased ‘‘stickiness’’ of cor-neocytes and thus explains why in LIthe scales are large and plate-like, asnormal, invisible desquamation did notoccur. The rigid adhesion of cells isthought to be a compensatory effect toprevent TEWL (Steinert et al., 1998).

We conclude that the LI skin-huma-nized mouse model faithfully recapitu-lates the human disease phenotypeand concomitant molecular changesand can be used as an excellent toolfor testing new therapeutic approachesfor this up-to-now untreatable geno-dermatosis.

The study was approved by theInstitutional Review Board of the

University Hospital of Munster. Allpatients enrolled gave their written,informed consent. All animal studieshave been approved by Centro deInvestigaciones Energeticas Medioam-bientales y Tecnologicas’s InstitutionalReview Board and all experimentalprocedures were conducted accordingto European and Spanish laws andregulations.

CONFLICT OF INTERESTThe authors state no conflict of interest.

ACKNOWLEDGMENTSThis work was supported by the Bundesminister-ium fur Bildung und Forschung as part of theNetwork for Rare Diseases NIRK (grant numbers:01GM0901 and 01GM0902); the Foundation forIchthyosis and Related Skin Types (F.I.R.S.T.); theNational Institutes of Health (NIH) (grant number:P42 ES004699); and the Selbsthilfe Ichthyosee.V. FL was supported in part by the Institudothe Salud Carlos III (ISCIII) (grant number:PI081054), MDR was supported by the Ministeriode Ciencia y Innovacion (MICINN) (grant number:SAF2010-16976). HCH was further supportedby the Deutsche Forschungsgemeinschaft (DFG)(grant number: HE3119/5-1) and Koln Fortune(grant number: 79/2011). We are grateful to all thepatients and other probands who participated inthe study. The excellent technical assistance ofBlanca Duarte, Androniki Kolovou, Marc Nate-bus, Anette Peffekoven, and the Proteomics CoreFacility (Brett Phinney, Rich Eigenheer), Universityof California, Davis, is gratefully acknowledged.Special thanks to Mrs Brigitte Willis.

Karin Aufenvenne1, Robert H. Rice2,Ingrid Hausser3, Vinzenz Oji1,Hans Christian Hennies4,5,Marcela Del Rio6,7, Heiko Traupe1

and Fernando Larcher6,7

1Department of Dermatology, UniversityHospital Munster, Munster, Germany;2Department of Environmental Toxicology,University of California, Davis, Davis,California, USA; 3Department of Dermatology,University Hospital Heidelberg, Heidelberg,Germany; 4Cologne Center for Genomics,Division of Dermatogenetics, University ofCologne, Cologne, Germany; 5Cologne Clusterof Excellence on Cellular Stress Responses inAging-Associated Diseases, University ofCologne, Cologne, Germany; 6EpithelialBiomedicine Division, CIEMAT andCIBERER-U714, Madrid, Spain and7Bioengineering Department, Carlos IIIUniversity (UC3M), Madrid, SpainE-mail: karin.aufenvenne@ukmuenster.de

SUPPLEMENTARY MATERIAL

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

REFERENCES

Candi E, Schmidt R, Melino G (2005) Thecornified envelope: a model of cell death inthe skin. Nat Rev Mol Cell Biol 6:328–40

Choate KA, Khavari PA (1997) Direct cutaneousgene delivery in a human genetic skindisease. Hum Gene Ther 8:1659–65

Choate KA, Medalie DA, Morgan JR et al. (1996)Corrective gene transfer in the human skindisorder lamellar ichthyosis. Nat Med 2:1263–7

Farasat S, Wei MH, Herman M et al. (2009) Noveltransglutaminase-1 mutations and genotype-phenotype investigations of 104 patients withautosomal recessive congenital ichthyosis inthe USA. J Med Genet 46:103–11

Garcıa M, Larcher F, Hickerson RP et al. (2011)Development of skin-humanized models ofPachyonychia Congenita. J Invest Dermatol131:1053–60

Kuramoto N, Takizawa T, Takizawa T et al. (2002)Development of ichthyosiform skin compen-sates for defective permeability barrier func-tion in mice lacking transglutaminase 1.J Clin Invest 109:243–50

Lee WH, Jang S, Lee JS et al. (2005) Molecularcloning and expression of human keratino-cyte proline-rich protein (hKPRP), an epider-mal marker isolated from calcium-induceddifferentiating keratinocytes. J Invest Derma-tol 125:995–1000

Matsuki M, Yamashita F, Ishida-Yamamoto A et al.(1998) Defective stratum corneum andearly neonatal death in mice lacking thegene for transglutaminase 1 (keratinocytetransglutaminase). Proc Natl Acad Sci 95:1044–9

Oji V, Traupe H (2006) Ichthyoses: differentialdiagnosis and molecular genetics. Eur JDermatol 16:349–59

Table 1. Numbers of unique peptides of transglutaminase 1, cornifiedenvelope components, and desmosome/corneodesmosome componentsidentified by MS/MS-analysis in two or three independent samples derivedfrom normal human skin, normal grafts, human LI skin, and LI grafts

Number of unique peptides

Protein MW (kDa) Normal skin Normal graft LI skin LI graft

Transglutaminase 1 90 17, 10 10, 11 0, 0 0, 41, 0

Filaggrin 435 74, 110 76, 84 10, 13 13, 12, 16

Keratinocyte proline-rich protein 64 35, 39 35, 34 70, 59 51, 55, 53

Involucrin 68 16, 3 3, 11 15, 5 14, 5, 10

Loricrin 26 2, 2 0, 2 29, 24 16, 18, 22

Late cornified envelope protein 1C 12 2, 2 0, 0 5, 4 9, 8, 8

Small proline-rich protein 1B 10 0, 0 0, 0 2, 2 2, 2, 3

Desmoplakin 332 77, 28 40, 36 146, 128 135, 136, 137

Junctional plakoglobin 82 30, 14 13, 19 51, 38 55, 56, 53

Desmoglein 1 114 25, 19 14, 24 53, 43 45, 39, 44

Desmocollin 1 100 7, 0 0, 10 16, 17 12, 8, 10

Corneodesmosin 52 0, 0 0, 0 3, 4 2, 4, 3

1Four unique TG1-peptides (GSGVNAAGDGTIR; GTNPSAWVGSVEILLSYLR; YDTPFIFAEVNSDK;NPLPVTLTNVVFR) detected in only one sample are identical to mouse TG1-peptides.

1920 Journal of Investigative Dermatology (2012), Volume 132

K Aufenvenne et al.Skin-Humanized Mouse Model of TG1 Deficiency

O’Regan GM, Sandilands A, McLean WH et al.(2008) Filaggrin in atopic dermatitis. J AllergyClin Immunol 124:R2–6

O’Shaughnessy RF, Choudhary I, Harper JI(2010) Interleukin-1 alpha blockade pre-vents hyperkeratosis in an in vitro model of

lamellar ichthyosis. Hum Mol Genet 19:2594–605

Pigg M, Gedde-Dahl Jr T, Cox D et al. (1998)Strong founder effect for a transglutami-nase 1 gene mutation in lamellar ichthyosisand congenital ichthyosiform erythroder-

ma from Norway. Eur J Hum Genet 6:589–96

Steinert PM, Candi E, Kartasova T et al. (1998) Smallproline-rich proteins are cross-bridging proteinsin the cornified cell envelopes of stratifiedsquamous epithelia. J Struct Biol 122:76–85

Homozygous Dominant Missense Mutation in Keratin 17Leads to Alopecia in Addition to Severe PachyonychiaCongenitaJournal of Investigative Dermatology (2012) 132, 1921–1924; doi:10.1038/jid.2011.484; published online 16 February 2012

TO THE EDITORHomozygosity for dominant mutationsin keratin genes is rare and hasonly been reported for epidermolysisbullosa simplex (EBS; Stephens et al.,1995; Hu et al., 1997; Oldak et al.,2011). The majority of pathogenicvariants reported in 23 keratin genesare heterozygous missense or smallin-frame insertion/deletion mutationsinherited in an autosomal domi-nant manner (http://www.interfil.org;Szeverenyi et al., 2008). A smallnumber of recessive cases have beenreported, mostly due to nonsense muta-tions (Yiasemides et al., 2008). Here,we report homozygosity for dominantmissense mutations in keratin 17 thatmodify the pachyonychia congenita(PC) phenotype. PC is an autosomaldominant skin disorder caused byheterozygous mutations in any one ofthe genes encoding keratins K6a, K6b,K16, or K17 (McLean et al., 2011). Themain characteristics are palmoplantarkeratoderma, plantar pain, and naildystrophy. Additionally, oral leukokera-tosis, follicular keratoses, and epidermalcysts often occur. PC due to mutations inK17 (termed PC-17) is more frequentlyassociated with neonatal teeth and wide-spread pilosebaceous cysts in adults.

Family 1, of Hispanic ancestry,showed autosomal dominant inheri-tance (Figure 1a). A consanguineousmarriage between affected individualsresulted in three affected (one homo-

zygous (proband) and two hetero-zygous) and one unaffected offspring.Both parents had some characteristicsof PC. The father had thickened nails,mild plantar hyperkeratosis and steato-cysts; the mother had steatocysts butreported neither nail changes norkeratoderma. The 10-year-old probandhad features typical of PC but wasmuch more severely affected than otherfamily members and had additionalfeatures (Figure 1b–g). Unusually, allnails were thickened at birth. He alsohad several neonatal teeth. At 4 monthshe developed leukokeratosis, and by7 months had blistering on his handsand feet, and thickening of palmo-plantar skin that is now localized topressure points (Figure 1b, c, e, and f).He continues to have painful blistersand has follicular keratosis (Figure 1g).He has no steatocysts to date, althoughthese generally develop at puberty. Themost unusual feature, not previouslyassociated with PC, was hair loss, firstnoted at 7 months and has continuedduring his lifetime (Figure 1d). Onexamination he had a circumscribedarea in the occipital region where thehairs were much shorter and thinner.Eyebrows were normal. His affectedbrother (25 years old) and sister(13 years old) had thickened toenails,slight thickening of fingernails withsplinter hemorrhages and widespreadsteatocysts. Neither sibling had kerato-derma, blistering, neonatal teeth nor

alopecia (Figure 1h–j). The brother’s1-year-old daughter had nail involve-ment and a neonatal tooth.

Genomic DNA was obtained withinformed consent and appropriateethical approval that complies withthe Declaration of Helsinki Principles(Western IRB study no. 20040468).DNA extraction and mutation detectionwere performed according to the pub-lished protocols that avoid pseudogenecontamination (Wilson et al., 2011).Two primer sets were used for eachmutation hotspot exon to ensure againstpolymorphism. By DNA sequencing ofKRT17, a homozygous dominant mis-sense mutation was identified, desig-nated p.Asn92Ser (protein), c.275A4G(DNA), in the proband of Family 1. Themildly affected parents and siblingswere heterozygous (Figure 1k–n). Muta-tion p.Asn92Ser is the most commonlyreported mutation in KRT17, occurringin 36% of PC-17 families (Wilson et al.,2011).

Family 2 was of Middle Easternancestry. The proband, a 32-year-oldmale, had thickened finger and toenailsby 2 years of age (Figure 2a and b).Painful blisters, calluses, fissures, andulcerations were present on hands andfeet (Figure 2c and e), and he hadfollicular hyperkeratosis (Figure 2fand g) but no oral leukokeratosis. Hedeveloped generalized alopecia at theage of 3 years and now has almosttotal alopecia (Figure 2d, at age 8).At the age of 7, the patient receivedoral etretinate (Tigason; 1 mg kg�1 perday) for 3 months, which reduced

See related commentary on pg 1757

Abbreviations: EBS, epidermolysis bullosa simplex; PC, pachyonychia congenita

www.jidonline.org 1921

NJ Wilson et al.Homozygous Missense Mutation in Keratin 17