2000 Oxford University Press Human Molecular Genetics, 2000, Vol. 9, No. 10 15331542

A candidate gene for psoriasis near HLA-C, HCR (Pg8),is highly polymorphic with a disease-associatedsusceptibility alleleKati Asumalahti, Tarja Laitinen, Raija Itkonen-Vatjus1, Marja-Liisa Lokki2, Sari Suomela3,Erna Snellman4, Ulpu Saarialho-Kere3 and Juha Kere5,+

Department of Medical Genetics, Haartman Institute, University of Helsinki, 00014 Helsinki, Finland, 1Department ofDermatology, Oulu University Central Hospital, 90401 Oulu, Finland, 2Finnish Red Cross Blood Transfusion Service,Department of Tissue Typing, 00310 Helsinki, Finland, 3Department of Dermatology, Helsinki University CentralHospital, 00250 Helsinki, Finland, 4Department of Dermatology, Central Hospital of Pijt-Hme, 15850 Lahti, Finlandand 5Finnish Genome Center, University of Helsinki, 00014 Helsinki, Finland

Received 15 February 2000; Revised and Accepted 11 April 2000 DDBJ/EMBL/GenBank accession no. AF216493

A susceptibility gene for psoriasis, a chronic skindisorder, resides in chromosome 6p near the HLA-Clocus. Sequencing of the region has allowed theidentification of a new gene, HCR. We found that HCRis highly polymorphic with at least 12 coding vari-ants. An association study of the new HCR polymor-phisms and the previously suggested susceptibilityalleles HLA-Cw*0602 and corneodesmosin allele 5(CD*5) with psoriasis revealed a specific HCR variantassociated with psoriasis susceptibility. However,the HLA-Cw*0602 allele was rarer in controls andassociated with a stronger relative risk. Associationanalysis did not support CD*5 as a psoriasis suscep-tibility allele in our sample of patients (n = 100) andpopulation-matched controls (n = 93) from anisolated population. We found HCR to be overex-pressed in keratinocytes of psoriatic lesionscompared with paired samples of healthy skin. Ourresults suggest a potential role for HCR in the patho-genesis of psoriasis.

INTRODUCTIONPsoriasis is a chronic skin disorder characterized by keratino-cyte hyperproliferation, angiogenesis and inflammatory leuko-cyte infiltrates (1). It shows familial clustering, and in recentyears, several candidate loci for psoriasis have been proposed(PSORS15) (25). The oldest and still strongest evidenceexists for a susceptibility locus in the HLA region (PSORS1,OMIM#177900); in the earlier studies for the HLA-A andHLA-B loci (6,7) and later for the HLA-C locus (8,9). By DNA-based genotyping, the HLA-Cw*0602 allele has been found inmany populations to be the best marker for the major risk alleleof psoriasis. The proportion of patients who are HLA-Cw*0602positive (5480%), is four to five times higher than in popula-tion-based controls (1020%) (10,11). Also linkage results

from genome-wide scans place a major psoriasis locus onchromosome 6p21.3 (12,13). Recently, the susceptibilityregion for psoriasis has been tentatively narrowed to aninterval of a few hundred kilobases either centromeric or telo-meric to HLA-C (14,15).

A typical feature of psoriasis is lymphocyte, macrophageand neutrophil infiltration in the skin lesions. The HLA class Iantigens (HLA-A, -B, -C) have an important role in immuneresponse as antigen presenting molecules to CD8+ Tlymphocytes (16). Current knowledge of the pathogenesis ofpsoriasis suggests, however, that CD4+ T lymphocytes mayhave a more important role in the disease process than CD8+cells (17). The class I antigens can also inhibit NK cell-medi-ated cytotoxicity through killer cell-inhibitory receptors(KIRs), but the significance of NK cells in psoriasis hasremained contradictory (1,18). Therefore, the psoriasis suscep-tibility gene may not be HLA-C, but instead another gene inlinkage disequilibrium (LD) with the HLA-C locus. Indeed, thecorneodesmosin gene allele 5 (CD*5) has been found to belinked with psoriasis by the transmission disequilibrium testand CD has been suggested to be a prime functional candidatefor the psoriasis susceptibility gene (1921).

The HLA region has been completely sequenced, and most ifnot all of its genes have been predicted by sequence analysis(22). In the immediate neighborhood of HLA-C, there are onlyfour known functional genes: HLA-B, OTF3, TCF19 and CD(Fig. 1). We determined the structure of a new gene predictedfrom the genomic sequence. This gene was first called Pg8 (forputative gene 8, a suggested tricohyalin homolog) (23) andmore recently HCR (for -helix coiled coil rod homolog) (15)(GenBank accession no. AB029331). Due to its genomic posi-tion between HLA-C and CD and from preliminary data on itsexpression in keratinocytes, HCR must be considered as aninteresting candidate for a psoriasis susceptibility gene.

We screened the HCR gene for sequence variations andfound that it is highly polymorphic with at least 12 coding vari-ants. The HCR polymorphisms were analyzed together withHLA-Cw*0602 and CD*5 for association with psoriasis in

+To whom correspondence should be addressed. Tel: +358 9 191 26538; Fax: +358 9 191 26789; Email: juha.kere@helsinki.fi

1534 Human Molecular Genetics, 2000, Vol. 9, No. 10

patients and population-matched controls from a Finnishsubisolate. We found a specific HCR variant associated withpsoriasis susceptibility, whereas association analysis did notsupport CD*5 as a psoriasis susceptibility gene. We foundHCR to be overexpressed in keratinocytes at psoriatic lesions.Our results suggest a potential role for HCR in the pathogen-esis of psoriasis.

RESULTS

Gene structure and expression patternWe predicted the exonintron structure of the HCR gene fromthe sequence of the genomic clone Y24c027 (GenBank acces-sion no. AC004195) by using two different gene prediction

programs, FGENES and GENSCAN. To verify the codingsequence, we designed primers specific for the predictedexons; these were then used in RTPCR with keratinocyteRNA, and the amplified products sequenced. Compared withthe verified cDNA sequence, FGENES had predicted the intra-genic exon structure exactly right, whereas GENSCAN hadmerged exons 12 and 13 into one exon. The gene consists of 16exons, varying in size from 46 to 304 bp (Fig. 2). The codingsequence of HCR was confirmed to be 2271 bp, encoding aprotein of 757 amino acids (GenBank accession no.AF216493). HCR was shown to be expressed by RTPCR atvariable levels in all human tissues tested, most abundantly inheart, liver, skeletal muscle, kidney and pancreas, and moreweakly in lung and placenta (Fig. 3).

Detection of HCR polymorphismsTo detect sequence variations, we screened all exons in fivepsoriasis patients and one control by parallel SSCP analysisand direct sequencing of PCR products. Two patients werechosen among the HLA-Cw*0602-positive and three amongthe HLA-Cw*0602-negative to enrich for different variants.Altogether, 18 polymorphisms were found in the coding regiondistributed in eight exons (Table 1). All sequence variants weresingle nucleotide polymorphisms (SNPs). Eleven SNPs werepredicted to cause an amino acid change, nine of them non-conservative, and one SNP introduced a stop codon. Exons 2and 8 were the most polymorphic, with five and four SNPs,

Figure 1. A map of the HLA region between HLA-B and CD. Filled bars indi-cate genes and open bars indicate pseudogenes.

Table 1. The SNPs of the HCR gene coding sequence with the corresponding amino acid changes

The sequence of the clone Y24c027 (GenBank accession no. AC004195) has been used as the genomic reference.aNon-conservative amino acid change.bIn the Finnish samples in position +2119 only base A was detected.

Exon Base number (coding sequence) Base number (AC004195) Polymorphism Amino acid changeExon 2 +249 +10696 GA ArgGln (R76Q)a

+251 +10698 CT TrpArg (W77R)a+269 +10716 CT TrpArg (W83R)a+421 +10868 CT no change

+436 +10883 GC ArgSer (R138S)aExon 4 +715 +14628 CG no change

+769 +14682 AC GluAsp (E249D)Exon 8 +1193 +16946 TC TrpArg (W391R)a

+1194 +16947 GA TrpSTOP codon

+1193 and +1194 +16946 and +16947 TC and GA TrpGln (W391Q)a+1219 +16972 CT no change

+1229 +16982 TC no change

Exon 12 +1667 +20442 TG CysGly (C549G)Exon 13 +1824 +20695 GA ArgGln (R601Q)aExon 14 +1855 +21999 GA no change

+1861 +22005 GT GlnHis (Q613H)aExon 15 +2119b +22390 GA no change

+2122 +22393 AT no change

Exon 16 +2271 +22788 GC CysSer (C750S)a

Human Molecular Genetics, 2000, Vol. 9, No. 10 1535

respectively. We did not screen introns systematically, butfound seven non-coding SNPs in the small introns 3, 7, 10 and12 that were included in amplicons containing flanking exons.The Mendelian segregation of each SNP was confirmed in 22families.

Patients and controls

Altogether, 142 psoriasis patients and 210 family members(total of 100 families) were recruited from the central-eastern

part of Finland (the Kainuu province) during January 1996. Inaddition, we recruited 93 population-based controls. Samplecollection was approved by the Ethical Review Boards of theKainuu Central Hospital and the Department of MedicalGenetics, University of Helsinki, and all samples were usedwith informed consent.

The clinical characteristics of the study subjects are shown inTable 2. The subgroup with familial psoriasis consists ofpatients (n = 71) who have at least one affected first-degree

Figure 2. The coding sequence and structure of the HCR gene. Exon boundaries are indicated with black arrows below the sequence. The 16 exons are illustratedwith black rectangles with the exon number above them. Exons with polymorphism are marked with asterisks.

1536 Human Molecular Genetics, 2000, Vol. 9, No. 10

family member. Fifty-three of familial psoriasis patients haddeveloped psoriasis before the age of 40 (type I psoriasis).

Association analyses

All unrelated psoriasis patients (n = 100, one from each family)and 93 population-based controls were included in an associa-tion study. HLA-Cw*0602 was genotyped using the PCRSSPmethod, and CD*5 using PCR and restriction digestion. HCRSNPs were screened using SSCP, direct sequencing or restric-tion digestion of PCR products.

Thirty-seven of the patients (37%) and eight of the controls(9%) were HLA-Cw*0602-positive (P = 0.0000031). TwoHCR SNPs in exon 2 associated significantly with psoriasis(Table 3). Both SNPs changed an amino acid from tryptophanto arginine, and in all instances they occurred together; we callthis the ArgArg allele (HCR*ArgArg). Forty-two percent ofthe patients (42/100) compared with 19% of the controls (18/93)had HCR*ArgArg, showing significant excess of HCR*ArgArgamong patients (P = 0.00068, Table 4). Both HLA-Cw*0602and HCR*ArgArg associations were even stronger amongtype I psoriasis patients (52 versus 9%, P = 0.00000015 and 61versus 19%, P = 0.0000099, respectively).

Figure 3. The expression of HCR mRNA in different tissues studied by RTPCR. The upper panel shows the expression of HCR (amplicon size 929 bp)and the lower panel the expression of the GAPDH control housekeeping gene(amplicon size 226 bp). The tissues are (from the left) heart, placenta, lung,liver, skeletal muscle, kidney, pancreas (MTC panel I, Clontech) and keratino-cyte (separate experiment).

Table 2. Clinical characteristics of patients diagnosed with familial or sporadic psoriasis

One patient from each family (n = 100) was used for the association study.aP < 0.02.bAn individual can have more than one option.cP = 0.02.dP = 0.0004.

Clinical characteristics All (n = 142) Familial (n = 71) Sporadic (n = 71)Gender

Male 77 (54%) 40 (56%) 37 (52%)Age at the time of the study

median (years) 47 47 47range (years) 583 683 580Age at the time of clinical diagnosis

median (years) 30 30 30range (years) 380 380 375Age at onset 40 years 36 (25%) 11 (16%) 25 (35%)aunknown 12 (9%) 8 (11%) 4 (6%)PASI-score at the time of study

median 3.35 3.2 3.5

range 018.6 015.0 018.6

Type of skin lesionsb

plaque 121 (85%) 58 (82%) 63 (89%)guttate 14 (10%) 8 (11%) 6 (9%)inversa 9 (6%) 3 (4%) 6 (9%)unknown 2 (1%) 1 (1%) 1 (1%)Nail lesions 63 (44%) 27 (38%) 36 (51%)Anamnestic information ofb

systemic treatment (metotrexate/acitretin) 29 (20%) 9 (13%) 20 (28%)cUV-therapy 79 (56%) 29 (41%) 50 (70%)dhospitalization 46 (32%) 22 (31%) 24 (34%)

Human Molecular Genetics, 2000, Vol. 9, No. 10 1537

All the HLA-Cw*0602-positive patients (37/37) were alsoHCR*ArgArg positive. In the control group, HLA-Cw*0602was also always found in association with HCR*ArgArg (8/8).Among the HLA-Cw*0602-negative patients, 8% (5/63) hadHCR*ArgArg, which did not differ significantly from theproportion of HLA-Cw*0602-negative HCR*ArgArg-positivecontrols (10/85, 12%). These results suggested that HLA-Cw*0602 and HCR*ArgArg are in strong linkage disequilib-rium among both patients and controls.

We next considered two CD SNPs that together form theallele CD*5 (+619 T, +1243 C) and that have been reported toassociate with psoriasis (1921). In our material, CD*5 wasthe major allele in both the patient and control groups. Eighty-six percent (86/100) of patients and 85% (79/93) of controlshad CD*5 and thus we could not detect any disease association(Table 4). We did not observe any significant deviations fromthe HardyWeinberg equilibrium for HCR or CD gene.

Haplotype analysis

We genotyped all family members in 30 psoriasis families tostudy an eight-marker haplotype including HLA-Cw*0602,five HCR SNPs representing different exons (+269, +715,+1667, +2122, +2271) and two CD SNPs (+619, +1243). Two

additional CD SNPs at +1215 and +1236 were genotyped todifferentiate between the CD2 and CD3 haplotypes (21) in allHLA-Cw*0602-positive chromosomes. Sixty-nine eight-marker haplotypes could be determined unambiguously and 15of the chromosomes were HLA-Cw*0602-positive. Altogether17 different haplotypes were detected.

Among the HLA-Cw*0602-positive chromosomes, a singlefive-marker haplotype was seen for HCR (Table 5). All 13HLA-Cw*0602-positive patient chromosomes had the sameHCR susceptibility haplotype, but for the CD gene, the haplo-type was broken into three different haplotypes. None of theHLA-Cw*0602-negative haplotypes associated with psoriasisby chi-squared analysis.

Tissue in situ hybridizationWe examined HCR expression in paired biopsies frompsoriasis lesions and normal-looking skin from six psoriasispatients and from three unaffected individuals by in situhybridization (Fig. 4). HCR expression appeared strong inkeratinocytes of psoriatic lesions. The samples from healthy-appearing skin of psoriasis patients and from unaffected indi-viduals remained negative even after 40 days of exposure.

Table 3. Allele frequencies of the HCR gene in psoriasis patients compared with population-based controls

aPatients n = 50, controls n = 50.

Base Polymorphism Patients (n = 100) Controls (n = 93) P2+249 GA 9 (9%) 9 (10%)+251 CT 42 (42%) 18 (19%) 0.00068+269 CT 42 (42%) 18 (19%) 0.00068+421 CT 18 (18%) 31 (33%) 0.014+436 GC 81 (81%) 85 (91%)

+715 CG 43 (43%) 40 (43%)+769 AC 57 (57%) 53 (57%)

+1193a TC 9 (18%) 11 (22%)+1194a GA 1 (2%) 1 (2%)+1219a CT 3 (6%) 5 (10%)+1229a TC 40 (80%) 34 (68%)

+1667 TG 96 (96%) 90 (97%)

+1824 GA 13 (13%) 14 (15%)

+1855 GA 49 (49%) 44 (47%)+1861 GT 4 (4%) 3 (3%)

+2122 AT 22 (22%) 29 (31%)

+2271 GC 88 (88%) 87 (94%)

1538 Human Molecular Genetics, 2000, Vol. 9, No. 10

These results suggested upregulation of HCR in psoriatickeratinocytes.

DISCUSSIONPSORS1 at 6p21.3 has been the most consistently observedsusceptibility locus for psoriasis in both linkage and associa-tion studies. Our data suggest that HCR (Pg8) has many char-acteristics of a psoriasis susceptibility gene located in the HLAregion and, specifically, it emerges as a better candidate thanCD. This conclusion is based on several lines of evidence.

The region of interest for a psoriasis susceptibility gene hasbeen narrowed down to a few hundred kilobase interval aroundthe HLA-C locus (14,15). The role of HLA-C as a causativegene in the pathogenesis of psoriasis is unclear. Of the non-HLA genes in the region, CD has been suggested to be involvedin skin desquamation, making it a tempting candidate for apsoriasis susceptibility gene. CD was reported to carry anHLA-C-independent effect and CD*5 was suggested as a riskallele (1921). HCR localizes in the most promising intervalfor a psoriasis susceptibility gene, between HLA-C and CD,but we are not aware of any studies assessing its role inpsoriasis directly. The OTF3 and TCF19 genes, located veryclose to HCR, are most likely to have housekeeping functions

and no polymorphism has been described in these genes so far(24).

We verified the structure of HCR and found that its aminoacid sequence shows little homology with known proteins; thestrongest homologies are with various myosins. The corre-sponding porcine gene in the MHC region has been cloned(GenBank accession no. AJ251914). The secondary structureof the HCR protein is predicted to consist mainly of -helicalcoils and it may be a nuclear protein with a leucine zippermotif. The predictions do not allow specific suggestionsbeyond speculations about the physiological role of HCR.

We found that HCR is highly polymorphic, including 11predicted amino acid substitutions and one truncated variant.Of these variants one, named HCR*ArgArg, is associatedwith psoriasis [relative risk (RR) 2.1, 95% confidence interval(CI) 1.72.6, P = 0.00068]. It was present in all psoriasis-asso-ciated HLA-Cw*0602 chromosomes, but its frequency washigher than that of HLA-Cw*0602 among control chromo-somes; thus HLA-Cw*0602 associated with psoriasis evenmore strongly (RR 4.3, 95% CI 3.15.0, P = 0.0000031).HCR*ArgArg was also detected in HLA-Cw*0602-negativechromosomes, suggesting that it might be an older variant thanHLA-Cw*0602. The proposed different age of the two variantsmight explain the difference in the relative risk and strength of

Figure 4. Expression of HCR mRNA in human skin. (A) Dark-field image of a biopsy from psoriatic skin hybridized with an antisense cRNA probe for HCRdisplaying abundant signal in keratinocytes. The patient had been treated with only topical corticosteroids. (B) A parallel section from the same tissue block hybrid-ized with a sense cRNA probe for HCR is devoid of signal. (C) A bright-field image of HCR mRNA-positive keratinocytes shown in (A). (D) Dark-field image ofa parallel in situ hybridization experiment, performed on a paired biopsy of normal looking skin from the same individual as in (A), is negative for HCR mRNA inthe epidermis. All samples were counterstained with hematoxylin and eosin.

Human Molecular Genetics, 2000, Vol. 9, No. 10 1539

association. However, we cannot exclude the possibility thatHCR*ArgArg together with HLA-Cw*0602 forms the riskallele. Further association studies of both genes in larger sets ofpsoriasis patients and controls from different populations areobviously needed to verify and refine these suggestions.

We found the previously suggested psoriasis susceptibilityallele CD*5 at a high frequency (8586%) among bothpsoriasis patients and controls. This observation arguesstrongly against CD*5 as a major psoriasis susceptibilityallele, because the prevalence of psoriasis is 12% in the popu-lation studied; a strong effect of such a common variant shouldcause a much higher disease incidence. Furthermore, ourhaplotype analysis showed that a single susceptibility haplo-type was intact across the HLA-C and HCR genes, but split intothree haplotypes in CD. When intragenic SNP haplotypes forboth HCR and CD were determined among HLA-Cw*0602-positive chromosomes, a single five-marker haplotype wasdetected for HCR, whereas three different four-marker haplo-types were present for CD, including non-CD*5 haplotypes.Thus, we conclude that HCR is a better candidate than CD fora psoriasis susceptibility gene by genetic association analysis.

HCR was expressed at high levels in the keratinocytes inpsoriatic skin lesions whereas in paired samples from normalappearing skin it was barely detectable. At least one gene,amphiregulin, has been found to be overexpressed in psoriaticskin lesions, and transgenic mice with overexpression ofamphiregulin in the skin develop lesions similar to psoriasis(25). It will be informative to study similarities and differencesin the expression of both genes in the psoriatic process.

In conclusion, we have studied HCR, a new highly polymor-phic gene near HLA-C, which by position, genetic associationresults and expression pattern in normal and psoriatic skinmust be considered a candidate gene for psoriasis suscepti-bility. Its primary sequence suggests -helical rod structure butgives little additional insight to its physiological role. Furtherassociation and functional studies with HCR are highlywarranted to find out its role in the pathogenesis of psoriasis.

MATERIAL AND METHODS

Gene structureExonintron structure of the HCR (Pg8) gene was predictedfrom a genomic clone (GenBank accession no. AC004195)using two different gene prediction programs GENSCAN(http://ccr-081.mit.edu/GENSCAN.html ) and FGENES (http://dot.imgen.bcm.tmc.edu:9331/gene-finder/gf.html ). To amplifythe HCR cDNA, RNA was extracted (RNeasy Mini Kit,Qiagen, Valencia, CA) from a primary keratinocyte cell line(54A III/IV) and RTPCR [M-MLV Reverse Transcriptase,Promega (Madison, WI) and Random Hexamer primers,Research Genetics (Huntsville, AL)] was carried out using themanufacturers protocol. The coding regions of the HCR genewere identified from the cDNA by direct sequencing using theprimer pairs: CCA CCT GGC TCT CAG ACA TT and GTGCAG CCT CTG AAC CTC TT (exons 1 and 2); AGC AGGCTG AGG TGA TCGT and CCC TTC AGC TGC TTA ACAGAG (exons 26); ATT CCC TGG AGC CTG AGT TT andCCT CCT GCT GGA TGA GGC (exons 611); GGT CAC

Table 4. Frequency of the HLA-Cw*0602, HCR*ArgArg and CD*5 alleles among all psoriasis and typeI psoriasis patients compared with controls

aP = 0.0000031, RR 4.3 (95% CI, 3.15.0).bP = 0.00000015.cP = 0.00068, RR 2.2 (95% CI, 1.72.6).dP = 0.0000099.

Locus All psoriasis patients (n = 100) Type I patients (n = 31) Controls (n = 93)HLA-Cw*0602 37 (37%)a 16 (52%)b 8 (9%)HCR*ArgArg 42 (42%)c 19 (61%)d 18 (19%)CD*5 86 (86%) 25 (81%) 79 (85%)

Table 5. Extended haplotypes in HLA-Cw*0602-positive chromosomes

A single HCR haplotype was seen in all HLA-Cw*0602-positive chromosomes (dark gray). For CD, the haplotype was broken intothree different haplotypes in the risk chromosomes. For HCR, base T at +269 represents the ArgArg variant. For CD, bases T at+619 and C at +1243 together represent CD*5 (light gray).

Haplotype Number of chromosomes

HLA-C HCR CD Psoriasis (n = 13) Control (n = 2)+269 +715 +1667 +2122 +2271 +619 +1215 +1236 +1243

0602 T C T A G T A T C 5 1

0602 T C T A G T G T C 6 0

0602 T C T A G C G T C 2 0

0602 T C T A G T A T T 0 1

1540 Human Molecular Genetics, 2000, Vol. 9, No. 10

AGA TGT GAG CCT TG and GCT GGA GCA TCT GTCAAG GT (exons 1116). The primers were designed using thePrimer3 program (26). A rough starting point for transcriptionwas detected by PCR amplification using alternative forwardprimers designed to stepwise upstream and to find the sitewhere no PCR product was obtained from cDNA. GenomicDNA was used as a positive control and all amplified productswere sequenced. The end of transcription was detected bycomparing several ESTs with the sequence obtained fromkeratinocyte cDNA and genomic DNA. The open readingframe was detected with the program DNAWorks.

PCR assays were carried out in a 50 l volume containing5 l cDNA, 1 PCR buffer (10 mM TrisHCl, 50 mM KCl,0.1% Triton X-100), 440 M dNTPs, 2.2 mM MgCl2, 1 Mprimer mix and 2 U of DNA polymerase (DyNAzyme EXT,Finnzymes, Espoo, Finland). The samples were denatured for3 min at 94C, followed by 32 cycles each of 60 s at 94C and180 s at 68C. Purification of the PCR products was performedusing a gel extraction kit (Qiagen). Sequencing was performedby dye-terminator chemistry in both directions using an ABI373A sequencer.

Patient and control samplesWe informed people in Kainuu about the study through thelocal psoriasis patient organization, health care centers and thedermatological clinic of Kainuu Central Hospital. The Kainuupopulation represents a late settled region of Finland (27). Atthe end of the 16th century, Kainuu was inhabited by only afew hundred unrelated founders. The population remainedsmall until rapid expansion occurred at the end of the 19thcentury, leading to the present population of ~100 000. Basedon our genealogical study, 65% of the patients grandparentsoriginated in Kainuu proper, and most of the remaining grand-parents originated in the neighboring, often southern munici-palities. Our sample set represents 510% of all psoriasispatients in the region.

Based on self-reported psoriasis, probands were selected foran interview and a clinical examination was performed by asenior dermatologist (R. Itkonen-Vatjus). The diagnosis wasconfirmed for 142 patients accepted into the study. To allowhaplotyping, 210 unaffected family members were alsorecruited from 100 families. Both the probands and theirfamily members donated blood samples and filled out a healthquestionnaire. All participants gave written, informed consentfor access to their medical records to verify disease history,and the Ministry of Social and Health Affairs granted permis-sion to access the patients medical records. As population-based controls, we used DNA samples from healthy individ-uals from the same recruitment area. DNA was extracted fromvenous blood using a standard non-enzymic method. The studyprotocol was approved by the Ethical Review Board of theKainuu Central Hospital and the Department of MedicalGenetics, University of Helsinki.

Screening for polymorphismsFive psoriasis patients and one control sample were used forscreening sequence variations with 12 different primer pairscovering all exons: CCC TCC CAC TTT CAA GCTC andTTC CAG TGA GGA AGG GTC AC (exon 1); CAC CTGCAC TAA CCT GTC TTTG and TTT CTA CCC CTG CAT

TCA CC (exon 2); CTT CTT TCC GCA GCT GTC CT andTCC CTA AGT CTG CAC ACA GAT (exons 3 and 4); TACAGA GGG GCT GCT TTC CT and GCT GAG GGT GAGGGG TCT (exon 5); AAA GAT GCC ACC TCC TTC CT andGAG GGA ATA CCG GGA GAA AA (exon 6); CTG CCCAGC TCT CTC TCCT and CTC CAT CCC TGA TAC CTGCT (exons 7 and 8); GGA TCA GTG ACT TGT GCC CT andGTG GCT CGC AGT TGT CCT AC (exon 9); TTT CTC CCTGCT TTT TCC CT and CTC ATC CTC TCC ACC CTC TG(exons 10 and 11); TCC TTT TAG GGG AGG CAG AG andAAG GCC CTA TCC ACC CTG (exons 12 and 13); TGCCTT GGC CTC TCT GTA GT and TCT GCC CTC CTG TCTCCT AC (exon 14); GCT CTA TCC GGG CTA GGT TT andCCC TTG TCC CTT TGT GCTT (exon 15); TGG TGC TCATCT GCT GTC TT and CTT TCC CTC CAA CTG TCA GC(exon 16).

PCR assays were carried out in 50 l vol containing 50 ng ofgenomic DNA, 1 PCR buffer (10 mM TrisHCl, 50 mM KCl,0.1% Triton X-100), 100 M dNTPs, 1 mM MgCl2, 0.12 Mprimer mix, 1% DMSO and 2 U of DNA polymerase(DyNAzyme II, Finnzymes). The samples were denatured for5 min at 96C, followed by 3538 cycles each of 30 s at 96C,120 s at 5764C and 120 s at 72C. Elongation was performedfor 5 min at 72C. Purified (PCR purification kit, Qiagen) PCRproducts were sequenced using an ABI 373A sequencer anddye-terminator chemistry.

Association analyses

DNA samples were genotyped for HLA-Cw*0602 using theClass I SSP ARMSPCR method (28) and for the HCRsequence variations using SSCP, sequencing and restrictiondigestion. The primers and PCR protocols used were asdescribed above. The six fully sequenced samples were used asreferences, but we also sequenced all variant band patternsobserved in SSCP to confirm the polymorphisms. SSCP gelwas prepared in a 50 ml volume containing 1012.5 ml of 2MDE gel solution (FMC BioProducts, Rockland, ME), 6 ml of5 TBE, 100 l of 10% APS and 40 l of TEMED, and thesamples were electrophoresed for 1720 h at 24 W. In exon 8,the SNPs were identified using direct sequencing in 50 patientsand 50 controls. In addition, seven polymorphisms were veri-fied using altered restriction site recognizing enzymes. Diges-tions were performed in 20 l reactions containing 10 l ofPCR product and 2.5 U of either PstI (+249), AvaII (+269),Tsp509I (+421), BsmFI (+715), MslI (+1667), SfaNI (+2122)or MwoI (+2271) and the appropriate buffer provided by themanufacturer (New England Biolabs, Beverly, MA).

The SNPs of CD*5 were amplified with primer pairs S5/S6(+619) and S15/S16 (+1243) and analyzed with restriction siterecognizing enzymes (MnlI for +619 and HphI for +1243) asdescribed previously (29). The SNPs at +1215 and +1236 weredetected by sequencing the primer pair S15/S16 product.

Statistical significance of the differences in the allelefrequencies between patients and controls was calculated usingthe chi-square test. Relative risk was calculated using theformula [a/(a + c)]/[b/(b + d)], where a is the number ofpatients with the risk allele; c the number of patients withoutthe risk allele; b and d are the equivalent values in the controls,respectively.

Human Molecular Genetics, 2000, Vol. 9, No. 10 1541

Haplotype classificationThe haplotypes were constructed manually. For each family,the chromosome was classified as trait-associated if it occurredin any affected family member and as a control if it occurredonly in unaffected family members. Each chromosome wascounted only once per family.

Expression analysesThe tissue expression of the HCR gene was studied by PCRamplification (Human Multiple Tissue cDNA panel I, Clon-tech, Palo Alto, CA) using the primer pair GGT CAC AGATGT GAG CCT TG and GCT GGA GCA TCT GTC AAG GT(exons 1115). GAPDH was used as a housekeeping controlgene and was amplified using the primer pair GAA GGT GAAGGT CGG AGT CA and CTT CTA CCA CTA CCC TAA AG(PE Biosystems, Foster City, CA). The PCR amplifications forboth genes were performed in 50 l reactions using 35 cycles.

Skin specimens for in situ hybridization were obtained fromthe Department of Dermatology, University of Helsinki. Thefollowing subgroups of histological sections were examined:three samples of normal skin from different parts of the body,paired biopsies from psoriasis plaques and normal looking skinfrom six psoriasis patients. For in situ hybridization, a probewas generated by PCR amplification of the keratinocyte cellline 54A III/IV by RTPCR with random hexamer primers asdescribed above as well as primers ATT TAG GTG ACA CTATAC att ccc tgg agc ctg agt tt and TAA TAC GAC TCA CTATAc ctc ctg ctg gat gag gc, introducing T7 and Sp6 RNApolymerase promoter sequences at opposite ends of the 628 bpgene-specific product. In vitro transcribed antisense and senseRNA probes were labeled with [35S]UTP. In situ hybridizationwas performed at 50C overnight on formalin-fixed paraffin-embedded specimens using 4 104 c.p.m./ml of labeled probewith subsequent washing under stringent conditions, includingtreatment of RNase A (30,31). Following autoradiography for2040 days, the photographic emulsion was developed and theslides were stained with hematoxylin and eosin for micro-scopy.

ACKNOWLEDGEMENTSThe authors wish to thank Ms Liisa Rajasalo and Ms PivikkiPajunen and their staff in the Kainuu Central Hospital for theirinvaluable help in the recruitment of the patients. We also wishto thank Mr Vesa Ollikainen for his valuable adviceconcerning the statistical analysis. This study was supportedby the Academy of Finland, the Sigrid Juselius Foundation andHelsinki University Research Funds.

REFERENCES1. Barker, J.N.W.N. (1991) The pathophysiology of psoriasis. Lancet, 338,

227230.2. Tomfohrde, J., Silverman, A., Barnes, R., Fernandez-Vina, M.A., Young,

M., Lory, D., Morris, L., Wuepper, K.D., Statsny, P., Menter, A. et al.(1994) Gene for familial psoriasis susceptibility mapped to the distal endof human chromosome 17q. Science, 264, 11411144.

3. Matthews, D., Fry, L., Powles, A., Weber, J., McCarthy, M., Fisher, E.,Davies, K. and Williamson, R. (1996) Evidence that a locus for familialpsoriasis maps to chromosome 4q. Nature Genet., 14, 231233.

4. Capon, F., Novelli, G., Semprini, S., Clementi, M., Nudo, M., Vultaggio,P., Mazzani, C., Gobello, T., Botta, A., Fabrizi, G. et al. (1999) Searching

for psoriasis susceptibility genes in Italy: genome scan and evidence for anew locus on chromosome 1. J. Invest. Dermatol., 112, 3235.

5. Enlund, F., Samuelsson, L., Enerbck, C., Inerot, A., Wahlstrm, J., Yhr,M., Torinsson, ., Riley, J., Swanbeck, G. and Martinsson, T. (1999) Pso-riasis susceptibility locus in chromosome region 3q21 identified inpatients from southwest Sweden. Eur. J. Hum. Genet., 7, 783790.

6. Russel, T.J., Schultes, L.M. and Kuban, D.J. (1972) Histocompatibility(HL-A) antigens associated with psoriasis. N. Engl. J. Med., 287, 738740.

7. White, S.H., Newcomer, V.D., Mickey, M.R. and Terasaki, P.I. (1972)Disturbance of HL-A antigen frequency in psoriasis. N. Engl. J. Med.,287, 740743.

8. Tiilikainen, A., Lassus, A., Karvonen, J., Vartiainen, P. and Julin, M.(1980) Psoriasis and HLA-Cw6. Br. J. Dermatol., 102, 179184.

9. Asahina, A., Akazaki, S., Nakagawa, H., Kuwata, S., Tokunaga, K., Ishi-bashi, Y. and Juji, T. (1991) Specific nucleotide sequence of HLA-C isstrongly associated with psoriasis vulgaris. J. Invest. Dermatol., 97, 254258.

10. Mallon, E., Bunce, M., Wojnarowska, F. and Welsh, K. (1997) HLA-CW*0602 is a susceptibility factor in type I psoriasis, and evidence Ala-73 is increased in male type I psoriatics. J. Invest. Dermatol., 109, 183186.

11. Enerbck, C., Martinsson, T., Inerot, A., Wahlstrm, J., Enlund, F., Yhr,M. and Swanbeck, G. (1997) Evidence that HLA-Cw6 determines earlyonset of psoriasis, obtained using sequence-specific primers (PCR-SSP).Acta Derm. Venereol., 77, 273276.

12. Trembath, R.C., Clough, R.L., Rosbotham, J.L., Jones, A.B., Camp, R.D.,Frodsham, A., Browne, J., Barber, R., Terwilliger, J., Lathrop, G.M. et al.(1997) Identification of a major susceptibility locus on chromosome 6pand evidence for further disease loci revealed by a two stage genome-widesearch in psoriasis. Hum. Mol. Genet., 6, 813820.

13. Nair, R.P., Henseler, T., Jenisch, S., Stuart, P., Bichakjian, C.K., Lenk,W., Westphal, E., Guo, S.W., Christophers, E., Voorhees, J.J. et al. (1997)Evidence for two psoriasis susceptibility loci (HLA and 17q) and twonovel candidate regions (16q and 20p) by genome-wide scan. Hum. Mol.Genet., 6, 13491356.

14. Balendran, N., Clough, R.L., Arguello, J.R., Barber, R., Veal, C., Jones,A.B., Rosbotham, J.L., Little, A.M., Madrigal, A., Barker, J.N.W.N. et al.(1999) Characterization of the major susceptibility region for psoriasis atchromosome 6p21.3. J. Invest. Dermatol., 113, 322328.

15. Oka, A., Tamiya, G., Tomizawa, M., Ota, M., Katsuyama, Y., Makino, S.,Shiina, T., Yoshitome, M., Iizuka, M., Sasao, Y. et al. (1999) Associationanalysis using refined microsatellite markers localizes a susceptibilitylocus for psoriasis vulgaris within a 111 kb segment telomeric to the HLA-C gene. Hum. Mol. Genet., 8, 21652170.

16. Schwartz, B.D. (1990) The human major histocompatibility human leuko-cyte antigen (HLA) complex. In Stites, D.P. and Terr, A.I. (eds), Basicand Clinical Immunology. 7th edn. Appleton & Lange, San Mateo, CA, pp4560.

17. Ortonne, J.P. (1999) Recent developments in the understanding of thepathogenesis of psoriasis. Br. J. Dermatol., 140, 17.

18. Nickoloff, B.J., Wrone-Smith, T., Bonish, B. and Porcelli, S. (1999)Response of murine and normal human skin to injection of allogeneicblood-derived psoriatic immunocytes: detection of T cells expressingreceptors typically present on natural killer cells, including CD94, CD158,and CD161. Arch. Dermatol., 135, 546552.

19. Allen, M.H., Veal, C., Faassen, A., Powis, S.H., Vaughan, R.W., Trem-bath, R.C. and Barker, J.N.W.N. (1999) A non-HLA gene within the MHCin psoriasis. Lancet, 353, 1589.

20. Tazi-Ahnini, R., Camp, N.J., Cork, M.J., Mee, J.B., Keohane, S.G., Duff,G.W. and Giovine, F.S. (1999) Novel genetic association between the cor-neodesmosin (MHC S) gene and susceptibility to psoriasis. Hum. Mol.Genet., 8, 11351140.

21. Jenish, S., Koch, S., Henseler, T., Nair, R.P., Elder, J.T., Watts, C.E.,Westphal, E., Voorhees, J.J., Christophers, E. and Krnke, M. (1999) Cor-neodesmosin gene polymorphism demonstrates strong linkage disequilib-rium with HLA and association with psoriasis vulgaris. Tissue Antigens,54, 439449.

22. The MHC Sequencing Consortium (1999) Complete sequence and genemap of a human major histocompatibility complex. Nature, 401, 921923.

23. Guillaudeux, T., Janer, M., Wong, G.K., Spies, T. and Geraghty, D.E.(1998) The complete genomic sequence of 424 015 bp at the centromericend of the HLA class I region: gene content and polymorphism. Proc. NatlAcad. Sci. USA, 95, 94949499.

1542 Human Molecular Genetics, 2000, Vol. 9, No. 10

24. Nair, R.P., Chia, N.V.C., Jenisch, S., Stuart, P., Epperson, J., Henseler, T.,Westphal, E., Arndt, J.L., Christophers, E., Voorhees, J.J. et al. (1999)Investigation of TCF19 as a candidate gene for psoriasis susceptibility. Br.J. Dermatol., 141, 981.

25. Cook, P.W., Piepkorn, M., Clegg, C.H., Plowman, G.D., DeMay, J.M.,Brown, J.R. and Pittelkow, M.R. (1997) Transgenic expression of thehuman amphiregulin gene induces a psoriasis-like phenotype. J. Clin.Invest., 100, 22862294.

26. Rosen, S. and Skaletsky, H.J. (1996, 1997) Primer3. Code available athttp://www-genome.wi.mit.edu/genome_software/other/primer3.html

27. de la Chapelle, A. (1993) Disease gene mapping in isolated human popu-lations: the example of Finland. J. Med. Genet., 30, 857865.

28. Tonks, S., Marsh, S.G.E., Bunce, M. and Bodmer, J.G. (1999) Moleculartyping for HLA class I using ARMS-PCR: Further developments follow-ing the 12th International Histocompatibility Workshop. Tissue Antigens,53, 175183.

29. Ishihara, M., Yamagata, N., Ohno, S., Naruse, T., Ando, A., Kawata, H.,Ozawa, A., Ohkido, M., Mizuki, H., Shiina, T. et al. (1996) Genetic poly-morphisms in the keratin-like S gene within the human major histocom-patibility complex and association analysis on the susceptibility topsoriasis vulgaris. Tissue Antigens, 48, 182186.

30. Prosser, I.W., Stenmark, K.R., Suthar, M., Crouch, E.C., Mecham, R.P.and Parks, W.C. (1989) Regional heterogeneity of elastin and collagengene expression in intralobar arteries in response to hypoxic pulmonaryhypertension as demonstrated by in situ hybridization. Am. J. Pathol., 135,10731088.

31. Saarialho-Kere, U.K., Kovacs, S.O., Pentland, A.P., Olerud, J.E., Welgus,H.G. and Parks, W.C. (1993) Cell-matrix interactions modulate interstitialcollagenase expression by human keratinocytes actively involved inwound healing. J. Clin. Invest., 92, 28582866.