K(AW

IBG

IKmg

CJUN

R9

H©P

iller Cell Immunoglobulin-Like ReceptorKIR) Genes in the Basque Population:ssociation Study of KIR Gene Contentsith Type 1 Diabetes Mellitus

zortze Santin, Guiomar Pérez de Nanclares,egoña Calvo, Ayman Gaafar, Luis Castaño,

EPV-N Group, and Jose Ramon Bilbao

dpKeuvptotcHS2

K

ABSTRACT: Killer cell immunoglobulin-like receptors(KIR) form a group of regulatory molecules that specifi-cally recognize HLA class I molecules, modulating cyto-lytic activity of natural killer cells. The number of KIRgenes can vary between individuals, significant allelicvariations have been described, and KIR genes are orga-nized in a complex and heterogeneous family. In thepresent study we have performed KIR genotype analysis inthe Basque general population. Additionally, we havetested the possible association between KIR gene content(in combination with its HLA ligand) and type 1 diabetesmellitus (T1DM). KIR genotyping was performed using acommercial sequence-specific primer amplification geno-typing kit and amino acid position 80 of HLA-C wasgenotyped by specific amplification and direct sequenc-ing. Haplotypes and genotypes were deduced based on

previous studies, and frequencies were compared between t

NK natural killer

STTV

s(swsbiaers

46006376; Fax: 34 946006076; E-mail: rbilbao@hcru.osakidetza.net.Received December 24, 2005.

uman Immunology 67, 118–124 (2006)American Society for Histocompatibility and Immunogenetics, 2006

ublished by Elsevier Inc.

isease and control groups. All KIR genes tested wereresent in Basques and several of these genes (KIR2DS5,IR3DS1, and KIR2DL2) displayed significant differ-nces from corresponding genes in other Caucasoid pop-lations. In general, Basques present an increase in acti-ating KIR gene frequency and, consequently, theroportion of B haplotypes is higher. Three novel haplo-ypes were identified in the Basque population. Overall,ur results confirm the particular genetic characteristics ofhe Basque population. No association between KIR geneontent and susceptibility to T1DM was observed.uman Immunology 67, 118–124 (2006). © American

ociety for Histocompatibility and Immunogenetics,006. Published by Elsevier Inc.

EYWORDS: Killer cell immunoglobulin-like recep-

ors; Basque population; type 1 diabetes; HLA-C

ABBREVIATIONSCTLA-4 cytotoxic T lymphocyte antigen 4HLA human leukocyte antigenINS insulin geneKIR killer cell immunoglobulin-like receptor

SP sequence-specific primer amplification1DM type 1 diabetes mellitusLR toll-like receptorNTR variable number of tandem repeats

NTRODUCTIONiller cell immunoglobulin-like receptors (KIRs) areembers of the immunoglobulin superfamily and form a

roup of regulatory molecules expressed on the surface of

From the Endocrinology and Diabetes Research Group, Hospital deruces, Barakaldo-Bizkaia, Spain (I.S., G.P.dN., B.C., A.G., L.C.,.R.B.); and Departments of Pediatrics (L.C.) and Nursing (J.R.B.),niversity of the Basque Country, Bilbao, Spain.GEPV-N Group: Basque-avarre Pediatrics and Endocrinology Group.

Address reprint requests to: Jose Ramon Bilbao, Endocrinology and Diabetesesearch Group, Hospital de Cruces, Barakaldo E48903 Bizkaia, Spain; Tel: 34

ome subsets of lymphoid cells, including natural killerNK) cells and certain T cells. These surface receptors arepecific for allelic forms of HLA class I molecules andere first identified by their ability to confer some

pecificity to NK cytolytic action [1, 2]. Interactionetween NK receptors and HLA class I molecules resultsn the generation of signals that modulate the cytolyticctivity of NK cells, extending their function beyond thelimination of infected or transformed cells toward theegulation of cytokine production, cytotoxicity, and

timulation of Th1 responses, so that NK cells may act as

0198-8859/06/$–see front matterdoi:10.1016/j.humimm.2006.02.036

adlsptihtlfspawadrdwcgasgigbogCPwsn

dab6tpHisgtibbfacs

voKwc[osdKTwd

MTbpgvfpj

uGj(Dtbi2(ctotsp3tuet

ilweg(

119KJR genotyping in Basque General population and T1DM patients

link between innate and adaptive immunity [3]. Toate, 16 functional KIR genes and two pseudogenesocated in the leukocyte receptor complex on chromo-ome 19q13.4 have been identified. KIR moleculesresent a very high degree of structural homology be-ween them and can be divided on functional groundsnto inhibitory and activating receptors. Products of in-ibitory KIR genes are characterized by long cytoplasmicails (“L” KIR genes) and transmit inhibitory signalseading to the general shutdown of NK cell effectorunctions [4]. In contrast, activating KIR molecules havehort cytoplasmic tails (“S” KIR genes) and their signalromotes NK cell activity. On the other hand, KIR genesre organized in a highly polymorphic, multigene familyith considerable allelic polymorphism and their associ-

tions define two main groups of haplotypes, A and B,epending on their gene content [5–7]. A haplotypes areelatively homogenous and are characterized by a pre-ominance of genes encoding inhibitory receptors,hereas B haplotypes are more diverse and generally

ontain several activating genes [8]. A subset of KIRenes common to both groups (KIR3DL2, KIR3DL3,nd KIR2DL4) is known as “framework” genes. Thetudy of KIR genes has revealed a high variability of KIRene contents within individuals and evidence of signif-cant allelic polymorphisms in several of the known KIRenes. Population-based genotyping studies have alsoeen performed and different gene frequencies have beenbserved [9–14]. In the present study we performed KIRenotype analysis in the Basque general population, aaucasian group of unknown origin living in the westernyrenees (northern Spain and southwestern France),hich maintains distinctive genetic characteristics re-

ulting from the relatively low admixture with othereighboring populations [15–17].

Type 1 diabetes mellitus (T1DM) is an autoimmuneisease in which pancreatic beta cells are destroyed by anberrant immune response. The major T1DM suscepti-ility locus is located in the HLA region on chromosomep21 and accounts for around 50% of the familial clus-ering of T1DM in Caucasians [18]. Several additionalutative susceptibility loci have been found outside theLA region; candidate regions include a VNTR in the

nsulin gene (11p15.5) and CTLA4/CD28 on chromo-ome 2q33 [19] in addition to innate immune responseenes, including MIC, TLR, and KIR genes [20–22]. Inhe latter, a genetic imbalance between activating andnhibitory KIR genes could be involved in the suscepti-ility to different autoimmune diseases. Costimulationy activating KIR genes, in combination with an insuf-icient regulation by inhibitory KIR, could facilitate thectivation of autoreactive T cells that are targeting betaells and contribute to the loss of tolerance for islet

elf-antigens [22]. Several studies have associated acti- m

ating KIR genes with rheumatoid arthritis [23, 24], andther groups have reported that activating KIR2DS1 andIR2DS2 genes are associated with psoriatic arthritishen the HLA class I ligands for inhibiting KIR mole-

ules (KIR2DL1, KIR2DL2, and KIR2DL3) are absent25]. There is some evidence in favor of the implicationf KIR genes in the development of T1DM [22] althoughome evidence of lack of association with autoimmuneiseases has also been published [26]. To test whetherIR gene content contributes to the development of1DM, we also analyzed KIR genotype in combinationith its HLA-C80 ligand in Basque patients with theisease.

ATERIAL AND METHODShe study population consisted of 71 unrelated healthylood donors of Basque origin, genotyped to define KIRrofiles in the Basque population, together with 76eographically and ethnically matched unrelated indi-iduals with type I diabetes mellitus, genotyped to testor genetic association with the disease. All studies wereerformed after informed consent obtained from the sub-ects or their parents.

Genomic DNA was extracted from peripheral bloodsing NucleoSpin DNA extraction kit (Macherey-NagelmbH, Düren, Germany). KIR genotyping of all sub-

ects was performed using a commercial SSP typing kitPel-Freez KIR genotyping SSP kit; Dynal Biotech, Browneer, WI; USA) according to the manufacturer’s instruc-

ions. Amino acid position 80 of HLA-C was genotypedy specific amplification and rapid fluorescent sequenc-ng. In brief, 100 ng of genomic DNA was amplified in5-�l PCRs using 10 pmol each amplification primer5=-tattgggaccgggagacacag-3= and 5=-ggaggaggcgc-cgtc-3=) and annealing at 64°C. Each sequencing reac-ion contained 2.5 �l of unpurified PCR product, 5.5 �lf ddH2O, 2 �l of 5X sequencing buffer (Applied Biosys-ems, Foster City, CA, USA), 1 �l of BigDye version 3.1equencing reaction mix (Applied Biosystems), and 100mol of sequencing primer (5=-tgggaccgggagacacagaa-=). Thermocycling was performed as recommended byhe manufacturer, excess dye terminators were removedsing gel filtration in a 96-well format (Autoseq96; Am-rshan Biosciences, Little Chalfont, England), and reac-ions were run on an ABI PRISM 3100-Avant machine.

KIR haplotypes and genotypes were deduced accord-ng to Hsu et al. [27]. KIR gene frequencies were calcu-ated for the control and T1DM groups and differencesere tested in 2 � 2 contingency tables using Fisher’s

xact test. Frequencies for activating and inhibitory KIRenes in the presence of their putative class I ligandHLA-C80) were calculated and analyzed in the same

anner. To control for the confounding effect of HLA-C

ptswC

RBWaar

mdopgi(Bam[Kpttt[tdatedfCa[

mh1Bts

T

KKKKKKKKKKK

a

b

Fh

120 I. Santin et al.

osition 80 variants, stratification analyses calculatinghe Mantel–Haenszel weighted odds ratio (OR) and as-ociated p values were performed. All statistical analysesere performed using Epi Info v.6.0 (Centers for Diseaseontrol, Atlanta, GA, USA).

ESULTS AND DISCUSSIONasque General Populationithin human populations, diversity for KIR proteins is

chieved via a combination of variable gene content andllelic polymorphism within each locus [28, 29] as aesult of different combinations of duplication, point

ABLE 1 Observed KIR gene frequencies (%) inBasque population and other Caucasians(frequencies of framework genes andpseudogenes not shown)

KIR Gene

Populations

p valueBasque

(n � 71)Caucasiansa

(n � 698)

IR2DL1 99 92 NSb

IR2DL2 62 46 0.01IR2DL3 96 91 NSIR2DL5 67 — —IR3DL1 90 92 NSIR2DS1 54 46 NSIR2DS2 63 52 NSIR2DS3 25 30 NSIR2DS4 85 90 NSIR2DS5 43 29 0.01IR3DS1 63 43 0.0009

According to Niokou et al. [9].

NS, nonsignificant.

IGURE 1 KIR haplotypes defined in Basque individuals

aplotypes (BH1-BH3) were identified in Basques.

utation, and recombination events that have occurreduring evolution [30, 31]. KIR genotyping in a samplef Basques from the general population revealed theresence of all KIR genes tested (Table 1). Frameworkenes KIR2DL4, KIR3DL2, and KIR3DL3 were presentn all of the samples studied and KIR pseudogenesKIR2DP1 or KIR3DP1) were present in 98.6% of theasque individuals. The most frequent locus encoding anctivating KIR protein was KIR2DS4 (85%) in agree-ent with that previously described in other populations

9–14]. Two other activating genes, KIR2DS3 andIR2DS5, were the least frequent loci in the Basqueopulation (25 and 43%, respectively) but, interestingly,he frequency of KIR2DS5 was significantly higher thanhat observed in other populations [9–11, 13], in con-rast to the idea that this is a relatively uncommon gene29]. Two other genes, activating KIR3DS1 and inhibi-ory KIR2DL2, were also increased among Basques. Theifferences in the frequencies of KIR2DL2, KIR2DS5,nd KIR3DS1 compared to cumulative frequencies for aotal of 698 Caucasoid individuals calculated by Niokout al. [9] were statistically significant (p � 0.05). Theistribution of KIR gene frequencies among Caucasiansrom France, which is geographically close to the Basqueountry, was similar to that among other Caucasians,nd the same divergences from the Basques are present14].

Following the haplotype subtraction strategy and no-enclature defined by Hsu et al. [27], 16 different

aplotypes were defined in our Basque sample set (Figure), 3 of which have not been previously described (BH1,H2, and BH3). Of these, 13 haplotypes were present in

he general population, where the most commonly ob-erved haplotype was haplotype 1, occurring 39 times

the general population and T1DM patients. Three novel

from

(shvBrhiEWtKBhiptft

fhBlcn

dooesf(tpwh

Fgc

121KJR genotyping in Basque General population and T1DM patients

31%), whereas haplotype 2 was found only on 9 occa-ions (7%). In consequence, the overall frequency of Aaplotypes (haplotypes 1 and 2: those lacking all acti-ating KIR genes, except for KIR2DS4) was 38%, whilehaplotypes (containing up to six activating KIR genes)

eached a frequency of 62%, a proportion significantlyigher than that described in other populations, includ-ng Caucasians (English, Greek, and Palestinian) andastern Asians (Chinese, Japanese, and Thai) [8–10, 13].ith the exception of KIR2DS3, the remaining genes

hat are characteristic of B haplotypes (KIR2DL5,IR2DL2, KIR2DS1, and KIR2DS5) are increased inasques and contribute to the overall increase in Baplotype frequency. In general, B haplotypes carry anncreased number of activating genes, and it has beenroposed that populations that have been more exposedo certain pathogens may suffer from a selective pressureavoring more efficiently responding B haplotypes over

IGURE 2 KIR genotypes found in our population, previoenotypes (BG1-BG13) are composed of previously undescribeombinations for 19 genotypes (UKN1-UKN19).

hose of the A group, as is the case in a study in Hindus p

rom northern India, where a similar frequency of Baplotypes (63%) has been found [12]. Additionally,asques have been relatively isolated during human evo-

ution, and population-related phenomena could alsoontribute to allele frequency discrepancies compared toeighboring groups.

From the studies performed so far, more than 100ifferent KIR genotypes have been identified on the basisf gene content alone [32]. In our test population, basedn the genotypes described by Hsu et al. [27] and Jiangt al. [13], genotypes were deduced for 89% of theamples. A total of 27 different genotypes were identi-ied, 12 of which have not been previously describedFigure 2). The most frequent genotype in our popula-ion was C genotype (14%), in contrast to other Eurasianopulations such as Indians, Vietnamese, or Caucasians,here AF is the most frequent genotype. On the otherand, the average number of KIR loci per genotype in our

described by Hsu et al. [27] or Jiang et al. [13]. Thirteenlotype combinations. It was not possible to deduce haplotype

uslyd hap

opulation is 13; the variation in the number of loci

raKbaanctrafp

A

Togcingtp(oidKpacIiacIKgpa0bgHac

T

I

A

P

N

T

KKKK

KKKK

T

122 I. Santin et al.

esults predominantly from the presence or absence ofctivating KIR genes because the presence of inhibitoryIR genes among genotypes is relatively constant. It haseen suggested that this variation in the number ofctivating KIR genes could provide the basis for targetednd diversified NK cell responses to infection [9]. Ge-otypes AF and AG are A genotypes because they wereomposed of two A haplotypes, (1-2) and (1-1), respec-ively, and account for 16% of Basque genotypes; theemaining 84% correspond to “B genotypes” which carryt least one “B haplotype” in accordance with the higherrequency of B haplotypes observed in the Basqueopulation.

ABLE 2 Distribution of KIR gene frequencies (%)in Basque T1DM patients and generalpopulation

KIR GeneT1DM

(n � 76)

Generalpopulation(n � 71)

nhibitoryKIR2DL1 97 98KIR2DL2 52 62KIR2DL3 93 95KIR2DL5 49 66KIR3DL1 89 90

ctivatingKIR2DS1 48 54KIR2DS2 52 63KIR2DS3 24 25KIR2DS4 80 85KIR2DS5 35 43KIR3DS1 53 63

seudogenesKIR2DP1, KIR3DP1 99 99

o significant differences were observed.

ABLE 3 Distribution of KIR2DL1, KIR2DL2/3, KIRwith their putative HLA-C ligand (%) in T

HLA-C-Lys80 (�)

CTRL T1DM

IR2DL1 (�) 74 34IR2DL1 (�) 2 2IR2DS1 (�) 44 15IR2DS1 (�) 31 21

HLA-C-Asn80 (�)

CTRL T1DM

IR2DL2/2DL3 (�) 31 81IR2DL2/2DL3 (�) 0 0IR2DS2 (�) 20 42IR2DS2 (�) 11 39

he stratified analysis showed that differences are not significantly different (p�0.

ssociation Study with T1DM

o determine the influence of KIR genes in the devel-pment of type 1 diabetes mellitus, we analyzed the KIRene content in T1DM patients of Basque origin andompared it to the general population. In contrast to thedea that the number of activating KIR genes per phe-otype may influence disease susceptibility through aene dosage effect [22], no significant differences be-ween activating or inhibitory gene frequencies in T1DMatients and controls were observed in our populationTable 2). As shown in Figures 1 and 2, several previ-usly undescribed genotypes and a novel haplotype weredentified in the disease group, but, again, differencesid not reach statistical significance. Since the effects ofIR genes on disease susceptibility might depend on theresence of their putative HLA ligands in an individual,ctivating and inhibitory KIR genes were also analyzed inombination with aminoacid position 80 of HLA-C.ndividuals positive for inhibitory KIR2DL1 or activat-ng KIR2DS1 together with group 2 (Lys80) HLA-Clleles were less frequent among patients (34 and 15%)ompared to control subjects (74 and 44%), respectively.n contrast, the presence of inhibitory KIR2DL2/IR2DL3 or activating KIR2DS2 in the presence ofroup 1 (Asn80) HLA-C alleles was more frequent amongatients (81 and 42%) compared to healthy controls (31nd 20%). Differences were statistically significant (p �.05) and suggestive of an association with the disease,ut stratified analysis showed that discrepancies amongroups were biased due to the differential distribution ofLA-C80 alleles (Table 3). Our results suggest that the

ssociation with the disease is determined by the HLAomponent but, due to the strong linkage disequilibrium

1, and KIR2DS2 gene frequencies in combinationpatients and healthy controls

HLA-C-Lys80 (�)

L T1DM p OR (95% CI)

65 0.76 1.90 (0.05–74.28)0

31 0.44 1.48 (0.62–3.59)34

LA-C-Asn80 (�)

L T1DM p OR (95% CI)

19 0.44 undetermined0

15 0.76 1.27 (0.48–3.32)5

2DS1DM

CTR

240

1113

H

CTR

690

4820

05).

wc

CTowgopshWlTebpdgmKi

ATRSfBM9

R

1

1

1

1

1

1

1

1

1

1

123KJR genotyping in Basque General population and T1DM patients

ithin the MHC, it is not clear whether HLA-C directlyontributes to the genetic risk.

ONCLUSIONhe singularity of the Basque population compared tother human groups of Caucasian origin is also evidenthen KIR gene frequencies are considered. In general,enetic isolates, as the Basques, offer several advantagesver outbred populations for the identification of com-lex disease genes: it is generally easier to obtain atandardized phenotype definition, together with a moreomogeneous genetic background closer to Hardy–

einberg equilibrium genotype frequencies. Neverthe-ess, in contrast to other reports, no association with1DM has been observed, but this does not completelyxclude KIR genes as putative determinants of diseaseecause additional levels of variation, namely allelicolymorphisms, have not been investigated. Finally, un-er the hypothesis of the implication of KIR genes in theenetic risk of developing autoimmune responses, itight be interesting to perform association studies ofIR genes with other autoimmune disorders in this

solated population.

CKNOWLEDGMENTS

his work was funded by Grants RGDM-G03/212 andCMN-C03/08 from the Instituto de Salud Carlos III of thepanish Ministry of Health. I.S. is supported by a predoctoralellowship from the University of the Basque Country. G.P.N.,.C., and J.R.B. are FIS researchers supported by the Spanishinistry of Health Fellowships, Nos. 03/0064, 03/0062, and

9/3076, respectively.

EFERENCES

1. Harel-Bellan A, Quillet A, Marchiol C, DeMars R,Tursz T, Fradelizi D: Natural killer susceptibility ofhuman cells may be regulated by genes in the HLAregion on chromosome 6. Proc Natl Acad Sci USA83:5688, 1986.

2. Moretta A, Tambussi G, Bottino C, Tripodi G, Merli A,Ciccone E, Pantaleo G, Moretta L: A novel surface antigenexpresed by a subset of human CD3-CD16� natural killercells. Role in cell activation and regulation of cytolyticfunction. J Exp Med 171:695, 1990.

3. Rajagopalan S, Long EO: Understanding how combina-tions of HLA and KIR genes influence disease. J Exp Med201:1025, 2005.

4. Lanier, LL: NK cell receptors. Annu Rev Immunol 16:359, 1998.

5. Martin AM, Freitas EM, Witt CS, Christiansen FT: Thegenomic organization and evolution of natural killer immu-noglobulin-like receptor (KIR) gene cluster. Immunogenetics

51:268, 2000.

6. Steffens U, Vyas Y, Dupont B, Selvakumar A: Nucleotideand amino acid sequence alignment for human killer cellinhibitory receptors (KIR). Tissue Antigens 51:398,1998.

7. Shilling HG, Guethlein LA, Cheng NW, Gardiner CM,Rodriguez R, Tyan D, Parham P: Allelic polymorphismsynergizes with variable gene content to individualizehuman KIR genotype. J Immunol 168:2307, 2002.

8. Yawata M, Yawata N, McQueen KL, Cheng NG, Gueth-lein LA, Rajalingam R, Shilling HG, Parham P: Predom-inance of group A KIR haplotypes in Japanese associatedwith diverse NK cell repertoires of KIR expression. Im-munogenetics 54:543, 2002.

9. Niokou D, Spyropoulou-Vlachou M, Darlamitsou A,Stavropoulos-Giokas C: Distribution of killer cell immu-noglobulin-like receptors in the Greek population. HumImmunol 64:1167, 2003.

0. Norman PJ, Stephens HA, Verity DH, Chandanay-ingyong D, Vaughan RW: Distribution of natural killercell immunoglobulin-like receptor sequences in three eth-nic groups. Immunogenetics 52:195, 2001.

1. Toneva M. Lepage V, Lafay G, Dulphy N, Busson M,Lester S, Vu-Trien A, Michaylova A, Naumova E, Mc-Cluskey J, Charron D: Genomic diversity of natural killercell receptor genes in three populations. Tissue antigens57:358, 2001

2. Rajalingam R, Krausa P, Shilling HG, Stein JB, Balamu-rugan A, McGinnis MD, Cheng NW, Mehra NK, ParhamP: Distinctive KIR and HLA diversity in a panel of northIndian Hindus. Immunogenetics 53:1009, 2002.

3. Jiang K, Zhu FM, Lv QF, Yan LX: Distribution of killercell immunoglobulin-like receptor genes in the ChineseHan population. Tissue antigens 65:556, 2005.

4. Denis L, Sivula J, Gourraud PA, Kerdudou N, Chout R,Ricard C, Moisan JP, Gagne K, Partanen J, Bignon JD:Genetic diversity of KIR natural killer cell markers inpopulations from France, Guadeloupe, Finland, Senegaland Reunion. Tissue Antigens 66:267, 2005.

5. Comas D, Mateu E, Calafell F, Perez-Lezaun A, Bosch E,Martinez-Arias R, Bertranpetit J: HLA class I and class IIDNA typing and the origin of basques. Tissue Antigens51:30, 1998.

6. Aguirre A, Vicario A, Mazon LI, Estomba A, Martinez dePancorbo M, Arrieta Pico V, Perez Elortondo F, LostaoCM: Are the Basques a Single and a Unique Population?Am J Hum Genet 49:450, 1991.

7. Bertranpetit J, Cavalli-Sforza LL: A genetic reconstructionof the history of the population of the Iberian Peninsula.Ann Hum Genet 55:51, 1991.

8. Steck AK, Bugawan TL, Valdes AM, Emery LM, Blair A,Norris JM, Redondo MJ, Babu SR, Erlich HA, EisenbarthGS, Rewers MJ: Association of non-HLA genes with typeI diabetes autoimmunity. Diabetes 54:2482, 2005.

9. Pugliese A: Genetics of type 1 Diabetes. Endocrinol

Metab Clin N Am 33:1, 2004.

2

2

2

2

2

2

2

2

2

2

3

3

3

124 I. Santin et al.

0. Novota P, Kolostova K, Pinterova D, Novak J, Weber P,Treslova L, Kovar J, Andel M, Cerna M: Association ofMHC class I chain related gene-A microsatellite polymor-phism with the susceptibility to T1DM and LADA inCzech adult patients. Int J Immunogenet 32:273, 2005.

1. Pirie FJ, Pegoraro R, Motala AA, Rauff S, Rom L, Gov-ender T, Esterhuizen TM: Toll-like receptor 3 gene poly-morphisms in South African Blacks with type 1 diabetes.Tissue antigens 66:125, 2005.

2. van der Slik AR, Koeleman BP, Verduijn W, BruiningGJ, Roep BO, Giphart MJ: KIR in type 1 diabetes:Disparate distribution of activating and inhibitory naturalkiller cell receptors in patients versus HLA-matched con-trol subjects. Diabetes 52:2639, 2003.

3. Yen JH, Moore BE, Nakajima T, Scholl D, Schaid DJ,Weyand CM, Goronzy JJ: Major histocompatibility com-plex class l-recognizing receptors are disease risk gene inrheumatoid arthritis. J Exp Med. 193:1159, 2001.

4. Namekawa T, Snyder MR, Yen JH, Goehring BE, Leib-son PJ, Weyand CM, Goronzy JJ: Killer cell activatingreceptors function as costimutatory molecules onCD4�CD28null T cells expanded in rheumatoid arthri-tis. J Immunol. 165:1138, 2000.

5. Martin MP, Nelson G, Lee JH, Pellet F, Gao X, Wade J,Wilson MJ, Trowsdale J, Gladman D, Carrington M:Cutting edge: susceptibility to psoriatic arthritis: influ-ence of activating killer Ig-like receptor genes in theabsence of specific HLA-C alleles. J Immunol 169:2818,

2002.

6. Moodie SJ, Norman PJ, King AL, Fraser JS, Curtis D,Ellis HJ, Vaughan RW, Ciclitira PJ: Analysis of candidategenes on chromosome 19 in coeliac disease: an associationstudy of the KIR and LILR gene clusters. Eur J Immu-nogenet 29:287, 2002.

7. Hsu KC, Liu XR, Selvakumar A, Mickelson E, O’ReillyRJ, Dupont B: Killer Ig-like receptor haplotype analysisby gene content: evidence for genomic diversity with aminimum of six basic framework haplotypes, each withmultiple subsets. J Immunol 169:5118, 2002.

8. Selvakumar A, Steffens U, Dupont B: Polymorphism anddomain variability of human killer cell inhibitory recep-tors. Immunol Rev 155:183, 1997.

9. Uhrberg M, Valiante NM, Shum BP, Shilling HG, Lie-nert-Weidenbach K, Corliss B, Tyan D, Lanier LL, Par-ham P: Human diversity in killer cell inhibitory receptorgenes. Immunity 7:753, 1997.

0. Kakhoo S, Rajalingam R, Shum BP, Weidenbach K,Flodin L, Muir DG, Canavez F, Cooper SL, Valiante NM,Lanier LL, Parham P: Rapid evolution of NK cell receptorsystems demonstrated by comparison of chimpazees andhumans. Immunity 12:687, 2000.

1. Shilling H, Lienert-Weidenbach K, Valiante N, UhrbergM, Parham P: Evidence for recombination as a mechanismfor KIR diversification. Immunogenetics 48:413, 1998.

2. Carrington, M, Norman, P: The KIR Gene Cluster. Be-thesda (MD): National Library of Medicine (US), National

Center for Biotechnology Information, 2003.