Roitberg-Tambur et al.

5. Norusis MJ. SPSS/PC+ Advanced Statistics 4.0. SPSSInc, Chicago, 1990.

6. Cassileth BR, Clark WH Jr, Lusk EJ,et al. How well dophysicians recognize melanoma andother problem lesions?JAM ACAD DERMATOL 1986;14:550-60.

7. Paine SL. Early detection ofskin cancer: knowledge, prac­tice and beliefs of Victorian general practitioners. MPHthesis, Monash University, 1991.

Journal of the American Academy of DermatologyDecember 1994

8. Koh HK, Caruso K, Gage I, et al. Evaluation of melano­ma/skin cancer screening in Massachusetts: preliminaryresults. Cancer 1990;65:375-9.

9. NIH Consensus Conference. Diagnosis and treatment ofearly melanoma. JAMA 1992;268:1314-8.

10. Burton RC, Coates MS, Hersey P, et aJ. Ananalysis of amelanoma epidemic.lnt J Cancer 1993;55:765-70.

Do specific pockets of HLA-C molecules predisposeJewish patients to psoriasis vulgaris?Anat Roitberg-Tambur, DMD,a Adam Friedmann, PhD,b, e Eli E. Tzfoni, MD,cShoshana Battat, MSc,d Rivka Ben Hammo/' Cilly Safirman, MSc,dKatsushi Tokunaga, MD,f Aldhiko Asahina, MD,g and Chaim Brautbar, PhDa, dJerusalem, Israel, and Tokyo, Japan

Background: Psoriasis vulgaris was reported to be associated with a specific alanine residueat position 73 of HLA-C alleles in Japanese patients.Objective: Our purposewasto determine the roleof HLA genesin susceptibility to psoriasisvulgaris in the Israeli Jewish population.Methods: Twenty-eightIsraeli patients wereanalyzed for their HLA classI and II specific­ities by means of serologic and molecular methods.Results:All patients possessed in their HLA-C antigens an alanine residue at position 73(p < 0.002). A significantly increasedfrequencyof HLA-Cw6 and ofCw7wasalsoobservedamongthe patients (p <0.02).Conclusion: Our studyclearlyshows that alanineinposition 73issignificantly associated withpsoriasis vulgaris in Jewish patients. Cw6 and Cw7 have a unique antigen-binding pocketcontaining both alanine at position 73 and a negatively charged aspartic acid at position 9.Theseresidues are most probably important in determiningtheconformation of the C pocketand in turn the nature of the peptide bound to it. We suggestthat this combination confersthe highest risk of the development of psoriasis vulgaris.(J AM ACAD DERMATOL 1994;31:964-8.)

Psoriasis vulgaris affects about 2% of the popula­tion in Western countries. Its pathogenesis remainunknown; however, autoimmune mechanisms havebeen implicated. Russell et al. l and White et a1.2

From The Lautenberg Center for General and Tumor Immunology,"Unit for the Development ofMolecular Biology," Department ofDermatology-Venereology," Tissue Typing Unit, Hadassah Univer­sityHospital and Hebrew University Hadassah Medical School," andthe DepartmentofGenetics, HebrewUniversity," Jerusalem; and theBlood Transfusion Service," and the Department ofDermatology,Faculty ofMedicine, The University ofTokyo,8 Tokyo.

Accepted for publication June 24, 1994.

Reprint requests: Prof. Chairn Brautbar, The Lautenberg Center forGeneral and Tumor Immunology, Hebrew University-HadassahMedical School, P.O. Box 1172, Jerusalem, 91010, Israel.

Copyright © 1994 by the American Academy ofDermatology, Inc.0190-9622/94 $3.00 +0 16/1/58664

964

were the first to report an association between HLAand susceptibility to psoriasis vulgaris. During thepast 20 years, several HLA antigens have been re­ported to have increased frequencies among patientswith psoriasis vulgaris in different racial or ethnicpopulations: HLA-AI, -A2, -Bl3, -BI7, -B3?, -B39,-Bw57, -Cw6, -Cv/l, -Cwll, and DR7.3-S Associa­tions with the complement components C4A4 andC4B4 were also described." The strongest correla­tion was found between psoriasis vulgaris andBLA-C antigens, specifically Cw6,S, 10,11 whereasother associations could be explained by linkage dis­equilibrium with Cw6, Cw7, or Cwll. Moreover,Ozawa et al.'? had shown a specific restriction frag­ment length polymorphism of the HLA-C regionthat correlates with susceptibility to psoriasis vul-

Journalof theAmerican Academy ofDermatologyVolume 31, Number 6

garis. Recently, Asahina et al. 13 investigated the as­sociation of specific nucleotide sequences of HLA-Calleles with psoriasis vulgaris in Japanese patients.Their results suggested that alanine at position 73(alanine 73) of HLA-C molecules can be a goodmarker for psoriasis vulgaris in the Japanese popu­lation. Alanine at position 73 was found in 81% ofpatients compared with 48% of healthy persons(p < 0.0001). It was therefore proposed that alanine73 may be important in determining susceptibility topsoriasis vulgaris.

The heterogeneous background of the Jewishpopulation, on the one hand, and its religion-basedseclusion on the other contribute to the formation ofa unique group within the Caucasian race, with arelatively typical HLA make-up. Analysis of HLAassociations with a certain disease in the Jewishpopulation may therefore provide better insight intothe role of these antigens in predisposition to thedisease. The present work was designed to studyHLA associations with psoriasis vulgaris in IsraeliJews at the serologic and molecular levels. Our spe­cific purpose was to examine the use of alanine 73 asa reliable marker for psoriasis vulgaris in Jews andto understand the role of this sequence in suscepti­bility to psoriasis vulgaris.

MATERIAL AND METHODS

Subjects

Twenty-eight unrelated Israeli Jewish patients withpsoriasisvulgaris (22 male, 6 female; 15 Ashkenazi, 13non-Ashkenazi) wereHLA typedwith serologic and mo­lecular methods. Two hundred fifty-eight healthy, unre­lated, ethnically matched persons were typed for HLAclass I and class II serologic specificities and oligotypedfor the determination of the HLA class II (DR,DQ) al­lelic variants.

HLA serologic typing

Peripheral blood lymphocytes were isolated by flota­tion onFicoll-Paque medium. HLA-A, -B,-C typing wasperformed bythe National Institutes of Health standardcomplement-dependent microlymphocytotoxicity tech­nique. HLA-DR and HLA-DQ typing was done onT-cell-depleted, B-cell-enriched lymphocytes by ex­tended incubation cytotoxicity testing.14 Alloantisera andmonoclonal antibodies used were those provided for theEleventh International Histocompatibility Workshop(IHW) set II, as well as of local origin.

DNA extraction and amplification

Genomic DNA was extracted by meansof a salting­out procedure. IS DNA samples were amplified by poly-

Roitberg-Tambur et al. 965

Table I. Phenotype frequencies of selected HLAclass I and class II antigens in Jewish patientswith psoriasis vulgaris (PV) and control subjects

PVpatients Controls

(%) (%)Phenotype (1/=28) (1/=258)

Al 43 33A2 36 31B13 0 8B17 7 12B37 7 19B39 11 3B46 0 0Cw4 14 31Cw6 25 18Cw7 39 27DR7 21 28

merasechainreaction (PCR) withHLA class II primer­pairs specific for HLA-DRBl,-DRB3, -DRB4,-DQAl,-DQBl, -DPAl, and -DPBlloci. Amplification was per­formed by 30 cycles of incubation at the correspondingdenaturation andannealing temperatures, followed by anelongation periodat 72° C. Afterelectrophoresis inaga­rosegels (2%),thePCR products were visualized with theaid of a UV transilluminator and blotted onto nylonmembranes (MSI, Westboro, Mass.).

Sequence-specific oligonucleotide probes (SSOPs)werelabeled withphosphorus 32.tetramethylammoniumchloride (TEMCl) solutions (3 mol/L, SigmaChemi­cals, St. Louis, Mo.) were used for hybridization andwashing procedures. All primers and probes forclass IIoligotypingwere provided bytheEleventh IHW andwereused according to the appropriate recommended proto­cols.l''

HLA-C amplification and oJigoprobehybridization

Specific amplification of HLA-Cfragments (residues45-88 of the first domain) was accomplished with theprimer pairs C[33P and C243PR. Two oligonucleotideprobes detecting a single nucleotide difference at position73 of HLA-C al domain were used: C208A-coding foralanineat position 73and C208B-coding for threonine atposition 73.Nucleotidesequenceofprimers and probes aswell as amplification and hybridization conditions followthe recommendations of Asahina et alP The HLA-Coligoprobing wasperformed forall 28 patients aswell asfor 91 of the control subjects.

Statistical analysis

The significance of association of HLA alleles withpsoriasis vulgaris was assessed by means ofa 2 X 2 table

Journal of the American Academy of Dermatology966 Roitberg-Tambur et al. December 1994

Table II. HLA class I typing of Israeli Jewish patients with psoriasis vulgaris

Patient No. HLA-A I HLA-B I HLA-C C208A*

1 1,24 35,38 4,9 +2 1,2 52,55 lO,-t +3 1,2 27,57 2,6 +4 1- 37,52 6- +, ,5 33,- 53,63 7,10 +6 2,68 53,- 4- +,7 1,24 27,39 7- +,8 23,30 49,52 6,7 +9 3- 35,58 6- +, ,

10 3,11.1 27,52 -,- +11 2,24 39,50 6,7 +12 2,26 37,38 6,- +13 1,2 8,44 6,7 +14 1,23 44,52 4,10 +15 2,28 14,44 8- +,16 1- 44,51 , +,17 26,31 38,39 6,7 +18 2,24 35,38 4,- +19 1,24 62,70 , +20 11.1,33 14,38 8- +,21 24,- 38,- , +22 1,2 8,38 7- +,23 2,33 14,41 7,8 +24 3,26 14,35 7,8 +25 3,26 8,35 7,9 +26 26,- 38,- , +27 1,68 35,44 6- +,28 1,11.1 8,63 7- +,

*Hybridization with probeC208Aindicating presenceof alanine at position 73 of HLA-C.tMinus sign indicates the possibility of homozygosity for a particular antigen or the presence of an unidentified antigen.

analysis and chi-square or Fisher's exact tests, as appro­priate.

RESULTS

Phenotype frequencies of selected HLA class Iand class II antigens in 28 patients with psoriasisvulgaris and 258 healthy unrelated control subjectsare summarized in Table I. The selection of the an­tigenspresented in this table was based on previouslypublished associations of HLA with psoriasis vul­garis in different populations. None of these associ­ations were confirmed in our study. Moreover, nosignificant associations were found between any sin­gle serologicallydetermined HLA class I or class IIantigens and psoriasis vulgaris in the present sampleof Jewish patients. However, when Cw6 and Cw7were grouped, a significant difference between pa­tients and controls was found (p < 0.02).

HLA-A, -B, -C typing and the hybridization re­sults of probe C208A in the patients with psoriasisvulgaris are shown in Table II. Positive hybridiza-

tion with the C208A probe indicates the presence ofthe amino acid alanine in position 73 of the HLA-Cfirst extracellular domain. One hundred percent ofthe patients (28 of 28) had this specific nucleotidesequence. In most cases the existence of alanine73 could be related to the presence of HLA-Cw4,Cw6, Cw7, C blank in linkage disequilibrium withB52(Cw12) or that ofCw8 in linkage disequilibriumwith B14. Four patients (Nos. 16, 19,21, and 26)gave positive hybridization results with C208A butwere not correlated with any of the aforementionedC locus specificities. In contrast, when a controlpopulation of 91 persons was analyzed for the pres­ence of alanine at position 73, only 70 of 91 controlsubjects carried this sequence in their HLA-C al­leles (p < 0.002).

HLA-DR and -DQ serologic typing as well asoligoprobe hybridization analysis of the DRBI,DRB3, DRB4, DRB5, DQA1, DQBl, DPA1, andDPBl alleles in the 28 Jewish patients with psoria­sis vulgaris were performed. The serologic and mo-

Journal of the American Academy of DermatologyVolume 31, Number 6 Roitberg-Tambur et al. 967

,.- - ..,~ - .. _ .... ,."

, . '

I,,I,,,

", ., ... , ... _...... .,. 'IIII

~",,

,

. .( .. .. J

,,,I,,

II.,,,

,,

"."," ...

Fig. 1. Schematic representation ofHLA-class I receptor-likepeptide binding groove.Darklinesrepresent a helices. Dashesindicate ,a-pleated sheets. The six pockets suited for bindingthe antigen side chains are termed A-E. Residues forming pocket C are indicated by theirposition number (9, 70, 73, 74, and 97).

I 10 20 30 40 so 60 70 80 90Consensus CSHSMRYFYT~V5RPGRCEPRFI~VCYVDDTQFVPFDSDAASPRCEPPAPWVEQECP£YWDPEVQlYKRQAQTDRVSLRNLRCYYNQSEA

CwlCw2C",JC",4Cw5C"'6roW?c",aC"'12CwlJC"'14

-----I--F-5 ------------S---------------------------------------------------------------------------------5----H---------------------------CR--------------------------H--K-- --------G-------c-----------H------------------D£------------Rl------------p---------------------­G-------5-5--W----------------------------------E-----------------------A---N--K---------D--------------------H-------------Q-----------------------------------------"--1------------------D--------------5------------------------------------------------A---N--K---------DG-------D--------------s------------------------------------------------A----------------D----------------------------------Q-------------------------------------------------------------------------------------------------------------------------------A-------------------------------------H---------------------------------------------------A-------------------------5-5-------------------------------------------------------------------------------

Fig. 2. Amino acid sequences of the al domain of HLA-C alleles. Nomenclature used af­ter Bodmer et al.22

CV'0101Cv'0201CV'OJOICv'0401CV'0501CV'0601CW'0702CV'0802CV"1201Cv'lJOlCv'1401

lecular results fully matched. No statistically signif­icant differences in antigen frequencies were foundbetween patients and control subjects.

DISCUSSION

Twenty-eight Israeli Jewish patients with psoria­sis vulgaris were typed for their HLA class I andclass II serologic specificities and for their class IIallelic variants at the molecular level. None of thepreviously reported HLA-A, -B, or -DR associationswith psoriasis vulgaris was confirmed in the presentstudy. When the HLA-C locus was analyzed forspecificnucleotide sequences, all patients were foundto have an alanine residue at position 73 of the firstextracellular domain ofHLA-C(p < 0.002). Twen-

ty-four of the 28 patients were typed as HLA-Cw4,-Cw6, -Cw7, or -Cw12 (CX52), alleles that areknown to share an alanine residue at position 73.Theremaining four patients were typed as HLA-C blankat the serologic level, but at the molecular level werefound to carry alanine at position 73 as well.

Psoriasis vulgaris is one of the rare examples inwhich predisposition to a disease is associated withHLA class I molecules . The mechanism underlyingthis association is intriguing given the fact that theHLA-C locus is less polymorphic than the HLA-Aor -B loci, and that a low level of expression ofHLA-C antigens is found on the cell surface, about10% ofHLA-A or HLA-B.l7

Crystallographic analysis of HLA class I mole-

968 Roitberg-Tamburet at.

cules1S-20 has revealed that these molecules have areceptor-like shape, with a peptide-binding groove.Six pockets, designated A-F, exist throughout thebinding groove (Fig. 1).21 Thesepockets are distinctwith regard to their amino acid composition andtheirelectron chargeproperties. Alterations ineitherof these parameters willcausedimensional and/orpolarity changes of the pocket, leading to differen­tialability ofthe HLA molecule to bindand presentimmunogenic peptides.

The significant association of alanineat position73 with psoriasis vulgaris in Jewish patients(p <0.002)and its criticallocation in the C pocketof HLA-C molecules drew our attention to otheramino acidsforming this pocket, that is, residues 9,70,74,and 97(Fig. I). Alignment ofknownHLA-Cfirst-domain amino acid sequences (Fig. 2) clearlyshow that residues 70 and 74 are identical in allHLA-C antigens (glutamine [Q] and aspartic acid[D], respectively). However, residue 9 lining thefloor of pocket C is rather polymorphic: phenylala­nine (F) in Cwl; tyrosine (Y) in Cw2, Cw5, andCw13; cystein (C) in Cw3; serine (5) in Cw4 andCw14; and asparticacid(D) in Cw6 and Cw7.Thefirstamino acid residues of Cw12(Cx52) have notbeen sequenced.

The presence of an acidic side chain (asparticacid) at the bottomof pocket C, togetherwith ala­nineat position 73, mayhavea cumulative effectonthe chargeand shapeof pocket C. This in turn mayfavor the lodging of a specific set ofdisease-trigger­ingantigenic peptides. Therefore it is not surprisingthat at the serologic level the onlysignificant asso­ciation found with psoriasis vulgaris in the Jewishpopulation is withCw6 and Cw7 (p < 0.02), whichhavebothalanine at position 73 andaspartic acidinposition 9 in pocket C.

In conclusion, the results are consistent with thecontention ofAsahina et al.13that alanine at position73of the firstdomain of HLA-C issignificantly as­sociated withpsoriasis vulgaris. Wealsosuggestthatpocket C in the HLA-C molecule maycontributetothedevelopment ofthedisease throughits affinity tospecific peptides.

We acknowledge the support of the Society of Re­search Associates of the Lautenberg Center and theConcern Foundation of Los Angeles.

REFERENCES

I. Russell TJ. Schultes LM, Kuban DJ. Histocompatibility(HLA) antigens associated with psoriasis. N Engl J Med1972;287:738-40.

Journal of the American Academy of DermatologyDecember 1994

2. White Sl-l, Newcomer VD, Mickey MR, et a1. Distributionof HLA antigen frequency in psoriasis. N Engl J Med1972;287:740-3.

3. Murray C, Mann DL, Gerber LN, et a1. Histocompatibil­ity alloantigens in psoriasis and psoriatic arthritis. J ClinInvest 1980;66:670-5.

4. Marcusson JA, Johannesson A, Moller E. HLA-A,B,CandDR antigens in psoriasis. Tissue Antigens 1981;17:525-9.

5. Green J, Montasser M. Investigations of the associations ofa number ofHLA antigens with psoriasis and psoriatic ar­thritis. Stat Med 1988;7:443-50.

6. Nakagawa H. Asahina A, Akazaki S, et a1. Association ofCw11 in Japanese patients with psoriasis vulgaris. TissueAntigens 1990;36:241-2.

7. Pitchappan RM, Roteeswaran A, Kakkaniah VN, et a1.HLA-Bw57 and DR7 association with psoriasis vulgaris inSouth India. Tissue Antigens \ 989;34: 133-7.

8. Tiilikainen A, Lassus A, Karvonen J, et a1. Psoriasis andHLA-Cw6. Br J DermatoI1980;102:179-84.

9. Nakagawa H, Akazaki S, Asahina A, et al. StudyofHLAclass I, class II and complement genes (C2, C4A, C4B andBF) in Japanese psoriatics and analysis of a newly-foundhigh-risk haplotype bypulsed field gel electrophoresis. ArchDermatol Res 1991;283:281-4.

10. Sakkas LI, Vaughan RW, Panayi GS, et a1. Application ofpolymerase chain reaction (PCR) to HLA-C locus typing.Eur J Immunogen 1991;18:185-7.

1J. Wuepper KD, Coulter SN, Haherman A. Psoriasis vul­garis: a geneticapproach. J Invest Dermato11990;95 :25-45.

12. Ozawa A, Ohkido M, Inoko H, et a1. Specific restrictionfragment length polymorphism on the HLA-C region andsusceptibility to psoriasis vulgaris. J Invest Dermatol 1988;90:402-5.

13. Asahina A, Akazaki S, Nakagawa H, et a1. Specific nucle­otide sequence of HLA-C is strongly associated with pso­riasis vulgaris. J Invest DermatoI1991;97:254-8.

14. Boyum A. Isolation of mononuclear cells and granulocytesfrom human blood. Scand J Clin Lab Invest SupplI967;21:77-90.

15. Miller SA, Dykes DD, Polesky HF. A simple salting outprocedure for extractingDNA from human nucleated cells.Nucleic Acids Res 1988;16:1215.

16. Kimura A, Sasazuki T. Eleventh International Histocom­patibility Workshop reference protocol for the HLA D N A­typing technique. In: Tsuji K, Aizawa M. Sasazuki T, eds.HLA 1991, Proceedings of the Eleventh International His­tocompatibility Workshop and Conference; vol l. Oxford:Oxford University Press, 1992:397-419.

17. Gussow D, Rein RS, Meijer I, et a1.Isolation expression andprimary structure of HLA-Cwl and HLA-Cw2 genes: ev­olutionary aspects. Immunogenetics 1987;25:3 \ 3-22.

18. Bjorkman PJ, Saper MA, Samraoui B, et a1. Structure ofthe human class I histocompatibility antigen, HLA-A2.Nature 1987;329:506-12.

19. Garrett TPJ, Saper MA, Bjorkman PJ, et a1. Specificitypockets for the side chains of peptide antigens in HLA­Aw68. Nature 1989;342:692-6.

20. Madden DR, Gorga Je, Strominger JL, et a1. The struc­ture of HLA-B27 reveals nonamer self-peptides bound inan extended conformation. Nature 1991;353:321-5.

21. Saper MA, Bjorkman PJ, Wiley DC. Refined structure ofthe human histocompatibility antigen HLA-A2 at 2.6Aresolution. J Mol Bioi 1991;219:277-319.

22. Bodmer JG, Marsh SGE, Albert ED, et a1.Nomenclaturefor factors of the HLA system, 1991. Tissue Antigens1992;39:161-73.