Zeeuwen 2003-the human cystatin m

ORIGINAL ARTICLE

The Human Cystatin M/E Gene (CST6): ExclusionCandidate Gene For Harlequin Ichthyosis

Patrick L. J. M. Zeeuwen, BeverlyA. Dale,n Gys J. de Jongh, Ivonne M. J. J. vanVlijmen-Willems,Philip Fleckman,n Janet R. Kimball,n Karen Stephens,n and Joost SchalkwijkDepartment of Dermatology, University Medical Center Nijmegen, Nijmegen, the Netherlands; nDepartments of Oral Biology, Biochemistry, LaboratoryMedicine, and Medicine/Dermatology & Medical Genetics, University of Washington, Seattle,Washington, USA

Cystatin M/E is a recently discovered cysteine protei-nase inhibitor whose expression is largely con¢ned tocutaneous epithelia. In human skin it is expressed insweat glands, hair follicles, and stratum granulosum ofthe epidermis where it presumably acts as a substratefor transglutaminase. Very recently we reported that anull mutation in the mouse cystatin M/E gene (Cst6)causes the murine ichq phenotype, which is character-ized by abnormalities in corni¢cation and desquama-tion, demonstrating an essential role for cystatin M/Ein the ¢nal stages of epidermal di¡erentiation.We hereobtained the complete sequence of the human cystatinM/E gene (CST6), which provides a tool to investigateCST6 as a candidate gene in skin diseases characterizedby abnormal corni¢cation. The involvement of CST6 inharlequin ichthyosis in humans was evaluated by se-

quencing the entire coding region and intron^exonboundaries for mutations in 11 sporadic harlequinichthyosis patients. No CST6 mutations were detectedin this group, which comprised type 1 and type 2 harle-quin ichthyosis patients. Disturbed transcription/trans-lation due to mutations in regulatory and noncodingregions of cystatin M/E was unlikely because cystatinM/E protein expression was observed in all patientsexamined, as assessed by immunohistochemistry.Although our results indicate that CST6 is not a majorgene contributing to type 1 and 2 harlequin ichthyosis,these data may facilitate further analysis of the role ofcystatin M/E in normal human skin and other geneticdisorders of corni¢cation. Key words: cysteine proteinases/hair follicle/mouse mutation/skin/stratum corneum. J InvestDermatol 121:65 ^68, 2003

We have previously reported that a new memberof the human cystatin superfamily, named cy-statin M/E (Ni et al, 1997; Sotiropoulou et al,1997), has an unusually tissue-speci¢c expres-sion pattern, which is largely limited to cuta-

neous epithelia (Zeeuwen et al, 2001, 2002a). Cystatin M/E is a 14kDa secreted protein that shares 30^35% homology with the hu-man family 2 cystatins. It has a similar overall structure such as asignal peptide and two intrachain disul¢de bonds, but possessesthe unusual characteristic of being a glycoprotein. The reportedtissue-speci¢c expression pattern is in contrast with other cystatinfamily members, which are mainly ubiquitously expressed. CST6has been assigned to chromosome 11q13 (Stenman et al, 1997),whereas all other family type 2 cystatin genes are clustered in anarrow region on chromosome 20p11.2 (Schnittger et al, 1993),which further suggests that cystatin M/E is an atypical proteinand only distantly related to the other known family members.We have shown that cystatin M/E could act as a substrate for epi-dermal transglutaminases suggesting a role in the formation ofthe stratum corneum (Zeeuwen et al, 2001), which integrity ismaintained by continuous renewal and shedding of old corneo-

cytes at the skin surface. This process of desquamation is the ¢nalevent in terminal di¡erentiation of the epidermis and is likely tobe regulated by the concerted action of proteolytic enzymes andtheir inhibitors (Egelrud and Lundstrom, 1991; Kalinin et al,2001). Recently we demonstrated that homozygosity for a nullmutation in the cystatin M/E gene (Cst6) of ichq mice is responsi-ble for juvenile lethality and defects in epidermal corni¢cationand desquamation (Zeeuwen et al, 2002b). Sundberg et al (1997)previously described the mouse ichq mutant, which presents asan ichthyosiform dermatitis. The ichq mouse phenotype includeshyperkeratosis, abnormally large mitochondria, absence of lamel-lar granules, and a characteristic keratin expression pattern in theinterfollicular epidermis; some of these morphologic and bio-chemical features are similar to human type 2 harlequin ichthyo-sis (HI), suggesting that the ichq mouse might be a good modelfor the human disorder. HI (MIM 242500) is a severe congenitalskin disorder usually leading to a stillborn fetus or early neonataldeath. Its clinical features at birth include ectropion, eclabium, eardysmorphology, and a thickened ¢ssured epidermis (Williamsand Elias, 1987). Histopathologically, HI is characterized by exces-sive epidermal and follicular hyperkeratosis. Biochemical and ul-trastructural abnormalities have suggested genetic heterogeneityand division into three subtypes (Dale et al, 1990; Dale and Kam,1993; Akiyama et al, 1998).These three subtypes have been de¢nedbased on expression of epidermal structural proteins; in types 1and 2 pro¢laggrin is expressed but not processed to ¢laggrin,whereas type 3 lacks pro¢laggrin; types 2 and 3 both have kera-tins 6 and 16 in addition to the normal keratins 5/14 and 1/10 seen

Address correspondence and reprint requests to: Patrick L. J. M. Zeeu-wen, Department of Dermatology, University Medical Center Nijmegen,Nijmegen Center for Molecular Life Sciences, PO Box 9101, 6500 HBNijmegen, the Netherlands. Email: [email protected]

Manuscript received November 4, 2002; revised December 16, 2002;accepted for publication February 20, 2003

0022-202X/03/$15.00 . Copyright r 2003 by The Society for Investigative Dermatology, Inc.

65

in all three subtypes. As HI is an extremely rare congenital andusually lethal disease in humans, its etiology and pathogenesishave been di⁄cult to investigate. Prior to the identi¢cation ofCst6 as the causative gene in ichq mice, several candidate genes inthe ichq region of mouse chromosome 19 (Capn1, Plcb3, Rela, andIkka/Chuk) and in the homologous region on human chromo-some 11q13 (CAPN1 and RELA) were tested and excluded as cau-sative for the HI phenotype (Dunnwald et al, 2003); however, theidenti¢cation of Cst6 as causative in ichq mice provides an excel-lent tool to investigate CST6 as a candidate gene in humanHI. Furthermore, this knowledge could be valuable to study thepathophysiology and biochemical mechanisms of skin diseasescharacterized by abnormal corni¢cation. In this study, we reportCST6 genomic organization and the mutational analysis ofthe entire coding region of the gene in 11 individuals a¡ectedwith HI.

MATERIALS AND METHODS

Cloning and sequencing An incomplete sequence of CST6 wasobtained by aligning (ClustalW, http://searchlauncher.bcm.tmc.edu/)genomic and mRNA sequences found in the public database (GenBankaccession nos. U62800, AJ328230, NT_033241, AK092391). In order toclose the remaining sequence gap in intron 1, we designed twooligonucleotide primers based on the known cDNA sequence of cystatinM/E, which could amplify a genomic fragment that includes intron 1;forward primer, 50 -ATGGCGCGTTCGAACCTCC-30; and reverseprimer, 50 -GTACTTGATGCCGGCCACC-30. Polymerase chain reactions(PCR) were carried out using the GC-RICH PCR System (RocheDiagnostics, Mannheim, Germany) according to the protocol providedby the manufacturer. As a template for the PCRwe used genomic DNAisolated from the whole blood of healthy volunteers. The E 800 bpampli¢cation product was cloned into plasmid Topo-TA (Invitrogen,Carlsbad, California) and the insert was sequenced, using the vectorprimers, at the DNA Sequencing Facility, Department of HumanGenetics, UMCN, Nijmegen, the Netherlands. The GC-rich part ofintron 1 was sequenced by BaseClear Labservices, Leiden, theNetherlands. The sequencing primers (¢rst sense then anti-sense) thatare used to complete the CST6 sequence were, 50 -ATGGGGTGGATCGGGGAGGA-30; and 50 -GGGTCTTGACCCTTGATCCTCACT-30.

Patients and mutation analysis Eleven sporadic cases of HI, whichcomprised type 1 and type 2, were investigated. A group of eight caseswere classi¢ed as two type 1 HI cases and six type 2 HI cases (Dale et al,1990). Four of these cases were previously described (Dale et al, 1990; Daleand Kam, 1993). The remaining three cases could not be classi¢ed byphenotype. Two of them were previously published as a case report(Herterich et al, 1993; Stewart et al, 2001), and one case was not publishedbefore. Exons of the CST6 gene were PCR ampli¢ed using intron-speci¢cprimers and 100 ng genomic DNA (Table I). PCRwere carried out using aDNA thermal cycler (TB1, Biometra, G˛ttingen, Germany) in 25 mLmixtures. The following bu¡er conditions were used: 10 mM Tris^HCl,pH 9.0, 1 to 1.5 mM magnesium chloride, 50 mM potassium chloride,0.1% Triton X-100, all four deoxyribonucleoside triphosphates (each at200 mM), 1 unit ofTaq DNA Polymerase (Promega, Madison,Wisconsin),and 20 pmol of each primer. After an initial incubation of 6 min at 941Campli¢cation was conducted for 35 cycles as follows: 1 min at 941C, 1 minat annealing temperature and 2 min at 721C. An additional 10 min at 721Cwas used for the last cycle. For precise PCR conditions see Table I.Ampli¢cation products were puri¢ed, and sequencing was performed onboth strands with the same ampli¢cation primers at the DNA SequencingFacility, Department of Human Genetics, UMCN, Nijmegen, theNetherlands.

Immunohistochemistry Human autopsy or biopsy material from HIpatients was processed for immunohistochemistry as previously described(Zeeuwen et al, 2002a). Immunohistochemical staining was performed aspreviously described using a⁄nity-puri¢ed polyclonal rabbit antibodiesdirected against human cystatin M/E (Zeeuwen et al, 2001). Antigenretrieval procedures were not necessary.

RESULTS

Genomic organization of CST6 At the time we started thisinvestigation the CST6 gene was partially represented in thepublic databases by a number of genomic and mRNA entries,but hitherto no complete genomic sequence was available.Using a system that facilitates PCR on GC-rich templates wecompleted the genomic sequence of cystatin M/E (deposited asCST6 in the GenBank database, accession no. AY145051). Atpresent, the full-length gene can also be extracted from asequence derived from human genomic chromosome 11 DNA,which has recently been deposited in the public database(GenBank accession no. AP001201); however, there are stillminor divergences between the sequenced part of intron 1 ofthis entry compared with our sequence data. The CST6 gene isorganized in three exons and two introns, and from thetranslation start site to the translation termination site it spans1354 bp of genomic DNA. Exon^intron boundaries wereestablished using the cystatin M/E cDNA sequence as describedpreviously (Sotiropoulou et al, 1997), and by the identi¢cation ofconserved splice site consensus sequences. Analysis of thegenomic sequence revealed a polyadenylation signal in the 30 -noncoding region of the gene; however, no identi¢able CAAT-box or TATA-box were found in the 50 -£anking sequence. TheATG start codon of CST6 lies in a Kozak consensus sequence(Kozak, 1987). Database analysis (http://fruit£y.org/cgi-bin/seq_tools/promoter.pl) predicted three transcription start sitesat, respectively, ^75 bp, ^42 bp, and ^34 bp upstream tothe translation initiation codon. This is in accordance with theaverage length of the expressed sequence tags found in thepublic database, in which the start positions vary betweenapproximately ^60 bp and ^20 bp.

Mutation analysis of CST6 in HI patients The major e¡ectdue to the absence of cystatin M/E in ichq mice appears to be onthe corni¢cation or desquamation process, which most likelyinduces juvenile lethality of these mice (Zeeuwen et al, 2002b).As this matches the most striking clinical feature of HI inhumans, we investigated CST6 as a disease causing candidategene in human HI. The entire protein-coding region plusintron^exon boundaries of CST6 were screened for mutations in11 HI patients using a set of intron-speci¢c primers (Table I). Nocausative mutations for HI have been detected in the CST6 gene.

Cystatin M/E expression in HI patients The expression ofcystatin M/E was examined in skin biopsies of HI type 1 andtype 2 patients using a⁄nity-puri¢ed polyclonal antibodiesagainst recombinant cystatin M/E. We found cystatin M/Eexpression in di¡erentiated keratinocytes of the infundibularepithelium of hair follicles (Fig 1a), and in the stratumgranulosum of the interfollicular epidermis (Fig 1b). In contrastto the expression pattern as previously observed in normal adultskin (Fig 1d) (Zeeuwen et al, 2001, 2002a) we found no cystatinM/E immunolocalization in the secretory coil epithelium of

Table I. Primers and PCR conditions used for analysis of the CST6 gene

Exon Forward Reverse Product Size (bp) AnnealingTemp (1C) MgCl2 (mM)

1 GCTCGGCACTCACGGCTCTG GGAGGCAGAATGCGACCAGGC 102 59 1ATGGCGCGTTCGAACCTCC TCCCAGCACCCCGCCCGCC 263 55 1

2 GGGCAACAGGAGAATATATC CAAACCAGTTCAGGATGGAC 285 54 1.53 TCAGTGATTGTCCCTCTCTTG TGACAGATACGGCCCACGG 277 53 1

66 ZEEUWEN ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure1. Immunohistology of HI skin. Formalin-¢xed sections of biopsies were incubated with puri¢ed rabbit anti-cystatin M/E polyclonal antibo-dies. (A) Hair follicle in epidermal skin of HI type 1. Strong staining of di¡erentiated keratinocytes of the infundibular epithelium. (B) Skin of HI type 2.Staining of cystatin M/E was observed in the stratum granulosum just beneath the thick stratum corneum. SC, stratum corneum; SG, stratum granulosum;SB, stratum basale. (C) Sweat glands in the skin of HI type 2. The secretory coils (arrows) as well as the ductal part of the glands are both completelynegative. (D) The secretory coil epithelium of the eccrine sweat glands in normal adult skin is strongly positive (arrows), whereas the ductal part shows amore di¡use staining for cystatin M/E (arrowheads). (E) Skin of age-matched control (neonatal skin, 3 mo of age) shows a normal cystatin M/E stainingpattern of the stratum granulosum. (F) Sweat glands in neonatal skin (1 mo of age). The secretory coil epithelium of the eccrine sweat glands shows a verydi¡use to absent staining for cystatin M/E. Scale bars: (A) 50 mm; (B^F) 25 mm.

EXCLUSION OF CST6 AS CANDIDATE GENE FOR HI 67VOL. 121, NO. 1 JULY 2003

eccrine sweat glands (Fig 1c). Samples of neonatal skin (1 and 3mo of age) that were used as age-matched controls also showedan almost complete absence of cystatin M/E expression in thesecretory coil epithelium of eccrine sweat glands (Fig 1f),whereas the staining pattern in the stratum granulosum of theseneonatal skin was similar to that of adult epidermis (Fig 1e).

DISCUSSION

In this paper, we report the genomic organization of the humanCST6 gene and mutational analysis in HI patients. CST6 containsthree exons within E 1.4 kb of genomic sequence. Characteriza-tion of the genomic sequence of CST6 shows some di¡erenceswith the other cystatin 2 family members.We found that the sizeof both introns was remarkably smaller compared with other cy-statin family 2 members (e.g., cystatin S, SA, SN, C, D). Thesecystatins harbor introns of, respectively, 1.1 to 2.2 kb, whereasthe introns of the CST6 gene are only 542 bp and 365 bp long.Besides these family 2 cystatins contain an identi¢able TATA-boxin their 50 -£anking sequence, whereas CST6 lacks this conservedsequence.As we have recently reported that a null mutation in the mouse

cystatin M/E gene causes the HI phenotype, we evaluated the in-volvement of CST6 in human HI patients. In this study, we couldnot detect any disease causing mutations in CST6 coding regionsor splice sites among the 11 HI patients we have analyzed.Withthis work and that of Dunnwald et al (2003), three genes in the11q13 region have been eliminated as causative for human type 1and 2 HI. A normal cystatin M/E expression pattern in the skinof HI patients was observed, except for the secretory coils ofsweat glands, where cystatin M/E appeared to be absent. The pre-sence of immunoreactive cystatin M/E protein suggests that de-fects in transcription/translation due to mutations in noncodingsequences, are unlikely to be the cause of disease in this group ofpatients. The cystatin M/E de¢ciency in the sweat glands of theseneonatal skin samples was surprising, as in normal adult humanskin the sweat glands are strongly positive. To investigate this weused age-matched controls for the expression of the protein. In-terestingly, we con¢rmed the absence of strong cystatin M/E ex-pression in sweat glands, in samples of neonatal skin (1 and 3 moof age). This ¢nding con¢rms a previous observation that di¡er-ences exist in the maturation and regulation of di¡erentiation be-tween epidermis and sweat ducts (Dale et al, 1990).Taken together,these data indicate that CST6 is unlikely to be a major gene re-sponsible for type 1 and 2 HI. The genomic organization of theCST6 gene and the designed primer sets described in this studywill facilitate further investigation of other HI patients, in parti-cular type 3 patients, which were not included in this study. AsHI is a very heterogeneous disease, it could still be possible thatthe underlying gene defect in some sporadic HI cases could bedue to mutations in CST6. These data also may facilitate studiesregarding the role of cystatin M/E in normal skin or as a candi-date gene for other genetic disorders of corni¢cation and desqua-mation.We speculate that cystatin M/E plays an important part inkeratinization, and is a crucial protein in pathways leading toterminal di¡erentiation of epidermal keratinocytes. The analysisof the function of cystatin M/E and the identi¢cation of its puta-tive target proteinase is clearly a direction for future research.

We thank Dr H.A. van den Bergen (Laboratory of Pathology, Isala Klinieken,Zwolle, the Netherlands), Prof. Peter Steijlen (Department of Dermatology, Univer-sity Medical Center, Nijmegen, the Netherlands), Dr Jill Clayton-Smith (Depart-ment of Clinical Genetics, St Mary’s Hospital, Manchester, UK), Dr Nicole JoyeŁand Dr Mary Gonzales (Laboratoire D’embryologie Pathologique et de CytogeŁ neŁ -

tique, Ho“ spital Saint-Antoine, Paris, France), Dr Keith Meredith (Neonatology,Memorial Hospital, Colorado Springs, Colorado), Dr Lakshmi Mehta (Genetics,North Shore Hospital, Manhasset, NewYork), Dr Michael He¡ernan (Dermatology,Washington University, St Louis, Missouri), Dr Linda Seely and Dr Sheldon Kor-ones (Reproductive Genetics, University ofTennessee, Memphis,Tennessee), DrVin-nie Biggs (Pediatrics, Northern Navajo Medical Center, Shiprock, New Mexico), DrDavid Witt (Dermatology, Kaiser-Permanente, San Jose, California), Dr DenizeMain (Genetics, California Paci¢c Medical Center, San Francisco, California) andDr Ephraim Kam (Seattle, Washington), and Dr M. Levy (Dermatology, TexasChildrens Hospital, Houston, Texas) for providing patient samples. This work was¢nancially supported by grant 902.11.092 from the Netherlands Organization forScienti¢c Research (NWO), by US PHS NIH grant P01 AR21557 and by theOdland Endowed Research Fund.

REFERENCES

Akiyama M, Dale BA, Smith LT, Shimizu H, Holbrook KA: Regional di¡erence inexpression of characteristic abnormality of harlequin ichthyosis in a¡ected fe-tuses. Prenat Diagn 18:425^436, 1998

Dale BA, Kam E: Harlequin ichthyosis. Variability in expression and hypothesis fordisease mechanism. Arch Dermatol 129:1471^1477, 1993

Dale BA, Holbrook KA, Fleckman P, Kimball JR, Brumbaugh S, Sybert VP: Het-erogeneity in harlequin ichthyosis, an inborn error of epidermal keratinization:Variable morphology and structural protein expression and a defect in lamellargranules. J Invest Dermatol 94:6^18, 1990

Dunnwald M, Zuberi AR, Stephens K, Le R, Sundberg JP, Fleckman P, Dale BA:The ichq mutant mouse, a model for the human skin disorder harlequinichthyosis: Mapping, keratinocyte culture, and consideration of candidategenes involved in epidermal growth regulation. Exp Dermatol 2003: in press,

Egelrud T, Lundstrom A: A chymotrypsin-like proteinase that may be involvedin desquamation in plantar stratum corneum. Arch Dermatol Res 283:108^112,1991

Herterich R, Hofweber K, Hammerschmid K, Steijlen PM, Happle R: Harlequin-Ichthyosis. Erfolgreiche Therapie mit oralen Retinoiden. Padiatr Praxis Z Kin-der-Jugendmedizin 46:93^99, 1993

Kalinin A, Marekov LN, Steinert PM: Assembly of the epidermal corni¢ed cellenvelope. J Cell Sci 114:3069^3070, 2001

Kozak M: At least six nucleotides preceding the AUG initiator codon enhancetranslation in mammalian cells. J Mol Biol 196:947^950, 1987

Ni J, Abrahamson M, Zhang M, et al: Cystatin E is a novel human cysteine protei-nase inhibitor with structural resemblance to family 2 cystatins. J Biol Chem272:10853^10858, 1997

Schnittger S, Rao VV, Abrahamson M, Hansmann I: Cystatin C (CST3), the candi-date gene for hereditary cystatin C amyloid angiopathy (HCCAA), and othermembers of the cystatin gene family are clustered on chromosome 20p11.2.Genomics 16:50^55, 1993

Sotiropoulou G, Anisowicz A, Sager R: Identi¢cation, cloning, and characterizationof cystatin M, a novel cysteine proteinase inhibitor, down-regulated in breastcancer. J Biol Chem 272:903^910, 1997

Stenman G, Astrom AK, Roijer E, Sotiropoulou G, Zhang M, Sager R: Assignmentof a novel cysteine proteinase inhibitor (CST6) to 11q13 by £uorescence in situhybridization. Cytogenet Cell Genet 76:45^46, 1997

Stewart H, Smith PT, Gaunt L, et al: De novo deletion of chromosome 18q in a babywith harlequin ichthyosis. AmJ Med Genet 102:342^345, 2001

Sundberg JP, Boggess D, Hogan ME, et al: Harlequin ichthyosis (ichq): A juvenilelethal mouse mutation with ichthyosiform dermatitis. Am J Pathol 151:293^310,1997

Williams ML, Elias PM: Genetically transmitted, generalized disorders of corni¢ca-tion. The ichthyoses. Dermatol Clin 5:155^178, 1987

Zeeuwen PL, Vlijmen-Willems IM, Jansen BJ, et al: Cystatin M/E expression isrestricted to di¡erentiated epidermal keratinocytes and sweat glands: Anew skin-speci¢c proteinase inhibitor that is a target for cross-linking bytransglutaminase. J Invest Dermatol 116:693^701, 2001

Zeeuwen PL, Vlijmen-Willems IM, Egami H, Schalkwijk J: Cystatin M/Eexpression in in£ammatory and neoplastic skin disorders. Br J Dermatol147:87^94, 2002a

Zeeuwen PL, Vlijmen-Willems IM, Hendriks W, Merkx GF, Schalkwijk J: A nullmutation in the cystatin M/E gene of ichq mice causes juvenile lethality anddefects in epidermal corni¢cation. Hum Mol Genet 11:2867^2875, 2002b

68 ZEEUWEN ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY