MICA, HLA-B Haplotypic Variation in Five Population Groups of Sub-Saharan African Ancestry

Genes and Immunity (2003) 4, 500–505 & 2003 Nature Publishing Group All rights reserved 1466-4879/03 $25.00 www.nature.com/gene MICA, HLA-B haplotypic variation in five population groups of sub-Saharan African ancestry

W Tian1, DA Boggs2, G Uko1, A Essiet3, M Inyama4, B Banjoko5, T Adewole6, W-Z Ding1, M Mohseni1, R Fritz7, D-F Chen1, LJ Palmer8 and PA Fraser1 1Center for Blood Research, Harvard Medical School, Boston, MA, USA; 2Department of Epidemiology, Harvard School of Public Health, Boston, MA, USA; 3Department of Surgery, University of Calabar Teaching Hospital, Calabar, Nigeria; 4Blue Cross Hospital, Lagos, Nigeria; 5Department of Pathology, University College Hospital, Ibadan, Nigeria; 6Nigerian Institute of Medical Research, Yaba, Nigeria; 7National Marrow Donor Program, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA; 8Channing Laboratory, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA

The human major histocompatibility complex (MHC) class I chain-related gene A (MICA), located 46 kb centromeric to HLA-B, encodes a stress-inducible protein, which is a ligand for the NKG2D receptor. In addition to its primary role in immune surveillance, data suggest that MICA is involved in the immune response to transplants and in susceptibility to some diseases. In this study, 152 subjects from the Yoruba (n¼74), Efik (n¼32), and Igbo (n¼46) tribes of southern Nigeria, 39 nationwide African-American stem cell donors, and 60 African-American individuals residing in the metropolitan Boston area were studied for MICA, HLA-B allelic variation, haplotypic diversity, and linkage disequilibrium (LD). MICA and HLA-B exhibited a high degree of genetic diversity among the populations studied. In particular, MICA allele and HLA-B–MICA haplotype frequencies and LD in the Efik and Igbo tribes were significantly different from the other study groups. HLA-B and MICA loci demonstrated significant global LD in all five populations (P-values o0.00001). LD also varied in a haplotype-specific manner. A novel MICA allele was detected in the Boston population. These findings are important from an anthropologic perspective, and will inform future HLA-linked disease association studies in related ethnic groups of African-derived ancestry. Genes and Immunity (2003) 4, 500–505. doi:10.1038/sj.gene.6364017

Keywords: MICA; HLA; SSOP; linkage disequilibrium; sub-Saharan Africans

Introduction NKG2D.10 Recent studies11,12 suggest that MICA might be an additional histocompatibility locus. The human major histocompatibility complex (MHC) Sub-Saharan African populations studied to date encompasses a 4 Mb genomic segment on chromosome exhibit a high degree of HLA gene diversity.13,14 6p21.3. From telomere to centromere, the MHC is Although several HLA gene frequency studies and two divided into class I, III, and II regions. On the 2 Mb class MICA reports include data on African-American sub- I segment, multiple MHC class I chain-related (MIC) loci jects,15–20 there is little information available regarding have been identified.1–3 The genes at these loci represent their ancestral population groups. Presently, the informa- a second lineage of mammalian MHC class I genes. tion on MICA allelic frequencies in sub-Saharan Africans MICA (MHC class I chain-related gene A; MIM 600169) is limited to a single cell line of South African origin maps about 46 kb centromeric to the HLA-B locus and (9021-RSH) in the International Histocompatibility Work- encodes a cell-surface glycoprotein that is expressed in shop collection.21 In the current study, we investigated keratinocytes and gastrointestinal epithelium,1,4 and in MICA and HLA-B allelic variation and haplotypic several other cell types.5 MICA functions as one of the diversity in three Nigerian tribal populations and two ligands for NKG2D/DAP10,6 an activating immunor- African-American populations. eceptor complex expressed on natural killer (NK) cells, gd T cells, and CD8 þ ab T cells.7 MICA triggers the cytolysis mediated by NKG2D-bearing cells7 and also costimulates CD8 þ ab T cells in pathogen-specific Results immune response.8 MICA and HLA-B allelic distributions MICA displays a high degree of allelic polymorphism In accordance with our previous report,19 MICA allele within the nonclassical HLA gene loci.9 Some of these assignment was based on the integrated information of polymorphisms modulate binding affinities with exons 2–5. A total of 9–15 MICA alleles were observed in the five study groups. MICA*00201, *004, and *00801 were commonly observed in all groups, with combined Correspondence: Dr W Tian, Center for Blood Research, Harvard Medical School, Boston, MA, USA. E-mail: [email protected] gene frequencies ranging from 71.6 to 80.4% (Table 1). Received 09 February 2003; revised 06 June 2003; accepted 09 June Pairwise comparisons of MICA allele frequencies re- 2003 vealed significant overall differences between the three MICA, HLA-B haplotypic variation W Tian et al 501 Nigerian tribes. MICA allele frequencies in the Efik and Accuracy of imputed HLA-B–MICA haplotypes Igbo tribes were also significantly different from each of To assess the reliability of the deduced HLA-B–MICA the African-American populations (data not shown). haplotypes reported here, 33 African-American families In all, 21–33 HLA-B alleles were observed in the five were typed for HLA-B, MICA at high resolution in populations, indicating a more heterogeneous allelic this study. A total of 57 parental samples were distribution at the HLA-B locus (Table 1). HLA-B*5301 determined for HLA-B–MICA haplotype by a maximum was the most common allele in all populations except the likelihood method and by family segregation. All Efik tribe, in which HLA-B*1503 predominated. Overall families were informative and produced unambiguous HLA-B allele frequencies in the Efik tribe were signifi- phase. Concordance between the haplotypes imputed cantly different from each of the other populations. In using the maximum likelihood method and by addition, HLA-B allele frequencies in the Yoruba and family segregation was found in 107/114 (93.9%) Igbo tribes were significantly different from the nation- haplotypes. Each of the remaining seven haplotypes wide African-American study population (data not involved a distinct HLA-B allele and had a frequency shown). less than 5% in the study populations (data not shown).

MICA and HLA-B Hardy–Weinberg equilibrium and HLA-B–MICA haplotype frequencies and linkage Ewens–Watterson homozygosity tests disequilibrium The MICA and HLA-B allelic distributions in each study In all, 12 haplotypes had a frequency Z5% in at least one population were consistent with Hardy–Weinberg pro- population (Table 2) and were chosen for linkage portions, with the exception of HLA-B in the Igbo tribe disequilibrium (LD) analysis. HLA-B*5301–MICA*00201 (P¼0.01). Observed values of the Ewens–Watterson was the most common haplotype in the Boston African- homozygosity statistic at the MICA locus were consistent American, Yoruba, and Igbo groups. In the nationwide with neutral expectations for all populations. The African-American and Efik populations, HLA-B*4201– observed Ewens–Watterson homozygosity statistics at MICA*004 and HLA-B*1503–MICA*00801 predomi- the HLA-B locus in the Efik and Igbo tribes were nated, respectively. Pairwise comparisons revealed that significantly higher than neutrality expectations overall HLA-B–MICA haplotype frequencies in the Efik (P¼0.04 and 0.03, respectively), whereas the Yoruba tribe and Igbo tribes were significantly different from each showed a marginal value (P¼0.06). other and from both African-American groups (data not

Table 1 MICA and HLA-B allele frequencies

Allele frequency

Nationwide African-Americans Boston African-Americans Yoruba Efik Igbo Allele a (N¼39) (N¼60) (N¼74) (N¼32) (N¼46)

MICA *00201 0.256 0.300 0.331 0.172 0.424 *004 0.295 0.233 0.270 0.219 0.130 *00801 0.231 0.183 0.142 0.328 0.250 *00802 0.051 0.075 0.068 0.031 0.011 *00902 — 0.033 0.007 0.078 — *041 0.013 0.017 0.020 0.031 0.098 Otherb 0.154 0.158 0.162 0.141 0.087

HLA-B *0702 0.077 0.042 0.068 0.063 0.033 *1503 0.038 0.050 0.047 0.141 0.098 *1510 0.077 0.075 0.074 0.031 0.011 *3501 0.051 0.075 0.088 0.063 0.065 *3910 — 0.017 0.027 — 0.054 *4201 0.179 0.067 0.041 0.031 0.044 *4403 0.051 0.083 0.047 0.031 0.054 *4501 0.026 0.025 0.034 0.094 0.022 *4901 0.026 0.017 0.027 0.109 0.011 *5101 — 0.050 0.027 0.031 0.022 *5301 0.205 0.150 0.155 0.078 0.239 *5801 0.013 0.050 0.088 0.016 0.065 *5802 0.026 0.025 0.034 0.078 0.044 Otherc 0.231 0.275 0.243 0.234 0.239 aAlleles with a frequency Z5% in at least one population. bOther MICA alleles with a frequency o5% include MICA*001, *00701, *00901, *010, *011, *01201, *015, *016, *017, *018, *019, *021, *030, *045, *046, MICA*CHAH. cOther HLA-B alleles with a frequency o5% include HLA-B*0706, *0801, *1302, *1401, *1402, *1403, *1501, *1516, *1517, *1525, *1801, *1803, *2703, *2705, *3502, *3701, *3801, *3906, *4001, *4102, *4202, *4402, *4410, *5001, *5201, *5501, *5601, *5701, *5702, *5703, *5704, *7801, *8101.

Genes and Immunity MICA, HLA-B haplotypic variation W Tian et al 502 Table 2 HLA-B–MICA haplotype frequencies and haplotype-specific LD estimates

Nationwide African-Americans Boston African-Americans Yoruba Efik Igbo (N¼39) (N¼60) (N¼74) (N¼32) (N¼46) Haplotype (HLA-B–MICA) Frequency D (s.e.) Frequency D (s.e.) Frequency D (s.e.) Frequency D (s.e.) Frequency D (s.e.)

0702–00801 0.077 0.054 (0.020) 0.042 0.035 (0.015) 0.046 0.042 (0.017) 0.063 — 0.033 — 1503–00801 0.039 0.034 (0.020) 0.050 0.040 (0.016) 0.047 0.061 (0.021) 0.141 0.126 (0.042) 0.098 0.049 (0.013) 1510–00802 0.051 0.043 (0.018) 0.075 0.080 (0.031) 0.068 0.064 (0.017) 0.031 0.029 (0.019) 0.011 — 3501–00201 0.051 — 0.075 0.047 (0.016) 0.054 — 0.031 — 0.065 — 4201–004 0.180 0.127 (0.025) 0.067 0.044 (0.015) 0.041 0.032 (0.014) 0.031 — 0.044 0.032 (0.014) 4403–004 0.051 0.047 (0.023) 0.083 0.061 (0.017) 0.047 — 0.031 — 0.022 — 4501–00902 — — — — — — 0.063 0.063 (0.028) — — 4901–004 0.026 — — — 0.027 — 0.109 0.108 (0.046) 0.011 — 5301–00201 0.167 0.087 (0.020) 0.117 0.068 (0.025) 0.128 0.082 (0.026) 0.047 — 0.152 — 5301–041 0.013 — 0.017 — 0.020 — 0.031 0.026 (0.017) 0.087 0.084 (0.034) 5801–00201 0.013 — 0.050 0.037 (0.016) 0.081 0.058 (0.016) 0.016 — 0.065 — 5802–00201 0.026 — 0.025 — 0.034 — 0.078 0.051 (0.018) 0.044 —

D with Po0.05 are shown.

shown). Significant global LD was observed in all five Table 3 (GCT)n polymorphism in exon 5 of 22 MICA alleles populations (data not shown). HLA-B*1503–MICA*00801 was observed to be in positive LD across the groups MICA allele (GCT)n in exon 5 (Table 2). *001 A4 A novel MICA allele *00201 A9 A previously unreported MICA allele was linked to an *004 A6 HLA-A*0201–B*1503–DRB1*1503 haplotype in one Afri- *00701 A4 *00801 A5.1 can-American family residing in Boston. When aligned *00802 A5.1 with MICA*019, the new MICA allele (tentatively named *00901 A6 MICA*CHAH) has an adenine substitution at position *00902 A6 751 in exon 4, resulting in an exon 4 sequence distinct *010 A5 from all known MICA alleles. It encodes glutamine *011 A6 (CAA) instead of arginine (CGA) at codon 251 of the a3 *01201 A4 *015 A9 domain. The Genbank accession numbers for this novel *016 A5 MICA allele are AF411923, AF411924, and AF411925. *017 A9 *018 A4 Trinucleotide repeat polymorphism of 22 types of *019 A5 MICA allele *021 A4 *030 A6 A total of 22 MICA alleles were detected in our sample *041 A9 panels, including MICA*021,*030 for which the (GCT) *045 A4 repeat number in exon 5 has not been previously *046 A9 reported (http://www.anthonynolan.com/HIG/seq/ MICA*CHAH A5 nuc/text/mica_nt.txt). The trinucleotide repeat showed a much restricted size polymorphism, in contrast to the high sequence diversification in exons 2–4 in the study We have therefore evaluated the HLA-B data in the populations (Table 3). present report relative to these previous findings. It was clear from previous studies23,34 that detection of HLA specificities in Nigerians was dependent on regional Discussion representation of the various linguistic and culturally This study was designed to investigate genetic diversity different groups. In this respect, our present report on in HLA-B and MICA in population groups where no HLA-B locus antigens confirms our previous serological previous MICA data existed. These data have demon- report23 in several ways, in that two major ethnic groups strated a high degree of genetic heterogeneity among from southern Nigeria included in that report were also sub-Saharan African-derived populations at these loci. well represented in this work. We have also detected The HLA complex contains three clusters of genes that HLA-B antigenic specificities reported for the first time are essential to immune response. Class I genes (HLA-A, by our ‘regional approach’ study in Nigerians but missed

-B, and -Cw) present endogenously derived peptides to previously. In particular, HLA-B8 and -B14 were not CD8 þ ab T cells. Their allelic repertoires have been detected by Okoye et al24, whereas HLA-B13 was not maintained to present an infinite number of antigenic reported in previous samples of West Africans.25 epitopes.22 Molecular typing of HLA-B class I or MHC By inclusion of the Efik group in the present study, we class I associated antigens (MICA) in Nigerian popula- have detected a sub-Saharan African population that tion groups has not been reported previously although appears to be genetically distinct from other Nigerian earlier studies performed by serology are known.23,24 tribal groups studied to date. This is best exemplified by

Genes and Immunity MICA, HLA-B haplotypic variation W Tian et al 503 the frequencies of some HLA-B–MICA haplotypes in the Pairwise comparisons revealed that overall HLA-B- Efik, some of which were not found in the Yoruba or MICA haplotype frequencies differed between each Igbo groups or in African-Americans. For instance, the African-American group and the Efik and Igbo tribes HLA-B*4501–MICA*00902 haplotype was observed only but not between the Yoruba tribe and the Boston-area in the Efik tribe. Furthermore, LD patterns within the African-Americans. This observation may be explained Efik population differed from the other study groups. by the fact that the Atlantic slave trade drew more Both the HLA-B*4901–MICA*004 and HLA-B*5802– heavily on the Yoruba, relative to the Efik and Igbo.31 MICA*00201 haplotypes were observed to be in LD Nonsynonymous substitutions are far more prevalent in the Efik group but not in any other population. In than synonymous ones in the MICA gene, suggesting contrast, the HLA-B*4201–MICA*004 haplotype was in some sort of selection. In our study, MICA genotype significant LD in all study groups reported here but not distributions fit a neutrality model in all five popula- in the Efik group. These findings may form the basis of tions. MICB and UL16-binding protein also function as the genetic reconstruction of the origins of the Efik, and NKG2D ligands.32 The presence of multiple NKG2D may help to substantiate earlier speculations that this ligands may reduce pathogen-driven selection on the group represents a distinct migration group,26 possibly MICA gene. In contrast, the excess of HLA-B homo- admixed with neighboring tribal groups in Cameroon.27 zygosity in the Nigerian tribal populations may have Our present study identified two haplotypes involving been shaped by directional selection to eliminate some the same HLA-B allele, namely HLA-B*5301–MICA*00201 endemic pathogens. The HLA-B allelic distribution in the and HLA-B*5301–MICA*041. HLA-B*5301–MICA*00201 Igbo tribe departed from Hardy–Weinberg equilibrium. is a common haplotype among African-Americans but, This observation may reflect the effects of inbreeding, among the Nigerians, it was in significant LD only in the outbreeding, relatively small sample size, or even Yoruba tribe. In contrast, the HLA-B*5301–MICA*041 misgenotyping. From the data available, we cannot haplotype was in LD exclusively in the Igbo and Efik assess breeding patterns in the Igbo. Resources are not tribes. This finding is of particular importance to the field available to permit additional sample collection from of transplantation medicine, as data continue to emerge Igbo tribal members. Misgenotyping seems unlikely about the role of MICA in transplant rejection. Subjects when we consider the consistency of results between with HLA-B*5301 may require MICA allelic assignment our previous family studies19 and this current analysis. to ensure optimal selection of allogeneic donors. Thus, the findings in the Igbo may require confirmation Although we hypothesize that the HLA-B*5301– in larger data sets in the future. MICA*00201 haplotype seen in African-Americans is A novel MICA allele was observed in an African- derived from their putative enslaved Yoruba ancestors, American family residing in Boston. It appears that this we cannot exclude the possible contribution of recent new allele is evolutionarily linked to MICA*019. This admixture with Caucasian populations to the high novel MICA variant was not found in any of the other frequency of this haplotype. populations involved in the present study, possibly HLA-B*15 encoding alleles displayed complex MICA suggesting a yet unknown ethnic origin. allelic association patterns in this study. HLA-B*1503 was In conclusion, our data demonstrate a high degree of in LD with MICA*00801 in all populations. HLA-B*1503 MICA gene diversity in sub-Saharan African-derived was also found to segregate with MICA*019 and populations. These data will provide a basis for future MICA*CHAH. HLA-B*1510 was in LD with MICA*00802 studies of the potential role of MICA in allogeneic stem except in the Igbo tribe, whereas HLA-B*1516 did not cell and solid organ transplantation as well as infectious demonstrate LD with MICA alleles in any of the five and autoimmune disease susceptibility in these popula- groups. MICA diversity on HLA-B*15 haplotypes could tions. MICA is an important candidate gene for a number have implications for risk of graft-vs-host disease of clinically significant diseases including diabetes,33 (GVHD) in African-American allogeneic stem cell trans- rheumatoid arthritis,34,35 and other autoimmune plant recipients. diseases.36–39 It is increasingly clear that knowledge of Our data demonstrate clear ethnic distinctions of the population-specific allelic and haplotypic structure is MICA allelic repertoire among sub-Saharan African- critical for genotype–phenotype association studies.40 derived populations. The similar findings in aggregate Our study will therefore inform such future studies in Caucasian samples20 and ethnic groups in Asian popula- African-American and sub-Saharan African populations. tions28 are consistent with a high degree of genomic diversity in the MICA gene. HLA-B and MICA loci demonstrated significant global Materials and methods LD in all five populations (P-values o0.00001). This finding is consistent with previous hypotheses that there Subjects has been very little recombination between the HLA-B The study included five distinct study populations. The and MICA loci during the successive duplications that first group comprised 60 unrelated healthy African- engender the present HLA class I region.29,30 It is worth American individuals residing in the metropolitan noting that LD also varied within this genomic region in Boston area, Massachusetts. The second group was a haplotype-specific manner. Strong LD was found composed of 39 African-American stem cell donors across populations in HLA-B*1503–MICA*00801 haplo- provided by the US National Marrow Donor Program type while confined to certain populations in several (NMDP). Of the 60 donor-recipient pairs provided to us haplotypes. Knowledge of haplotype-specific LD may in whom all recipients were African-American, 21 of the provide valuable information with regard to modern donors were not African-American, and therefore were populations’ ancestry, admixture, and the selective excluded from the analysis. The last three groups pressures maintaining such an LD. comprised 152 subjects recruited from three culturally

Genes and Immunity MICA, HLA-B haplotypic variation W Tian et al 504 and linguistically distinct tribes residing in southern the present study. PCR products of exon 4 derived from Nigeria—Yoruba (n¼74), Efik (n¼32), and Igbo (n¼46). both samples were therefore cloned using a TOPOs TA The Nigerian subjects were random volunteers from cloning kit according to the manufacturer’s protocol different locations within southern Nigeria. (Invitrogen, Carlsbad, CA, USA). Multiple inserts were Each study population consisted of healthy and sequenced in both directions following the protocol unrelated adults of self-reported ethnicity. Among the previously described.19 Boston subjects, 46.7% (28/60) were male and 53.3% (32/60) were female. Approximately two-thirds of the Statistical analysis 42 43 Nigerian study subjects were male and the remaining The programs Arlequin, CLUMP, MONTE-CAR- 44,45 46 one-third was female. In the Yoruba tribe, 66.2% (49/74) LO and GDA were used to analyze the data. P- were male and 33.8% (25/74) were female. In the Efik values were derived by empirical simulation where tribe, 62.5% (20/32) were male and 37.5% (12/32) were possible, and statistical significance was defined at the female. In the Igbo tribe, 67.4% (31/46) were male and 5% level. Comparisons of allele and haplotype frequen- 32.6% (15/46) were female. Among the NMDP donors, cies among populations and LD analyses were restricted Z 43.6% (17/39) were male and 56.4% (22/39) were female; to alleles and haplotypes, with a frequency 5% in at of the NMDP recipients, 56.4% (22/39) were male and least one study population. Exact, permutation-based 43 43.6% (17/39) were female. tests were used for all contingency table inference. All subjects voluntarily provided 10 ml of EDTA blood Consistency of genotype frequencies at each locus with samples from which DNA was extracted using Qiagen Hardy–Weinberg equilibrium was tested on a contin- kits (Valencia, CA, USA) according to the manufacturer’s gency table of observed vs predicted genotype frequen- specification. There were no missing genotypic data. All cies using a modified Markov-chain random walk 47 protocols were approved by the Institutional Review algorithm. The Ewens–Watterson’s homozygosity sta- 48,49 Boards of local hospitals. tistic was calculated with the MONTE CARLO program.44,45 LD was calculated using an iterative 50 Reference DNAs maximum-likelihood method and expressed as either 0 51 Genomic DNA from previously used heterozygous and D or D . The statistical significance of LD between homozygous cell lines19,20 were included in MICA MICA and HLA-B was tested using a likelihood-ratio genotyping. They comprised 9018 (MICA*001), 9085 test, whose empirical distribution was obtained by a 52 42 (MICA*001), 9068 (MICA*00201), 9027 (MICA*004), permutation procedure as implemented in Arlequin. 9040 (MICA*004), 9092 (MICA*00701), 9014 Maximum likelihood haplotype frequencies were (MICA*00801), 9048 (MICA*00801), 9087 (MICA*00801), imputed within each group of subjects using an 53 9016 (MICA*00901), 9032 (MICA*010), 9009 (MICA*010), expectation-maximization (EM) approach, as imple- 9031 (MICA*010), 9041 (MICA*016), 9052 (MICA*017), mented in the program Arlequin. The EM algorithm was 9008 (MICA*018), 9237 (MICA*01201/*019), and 9029 repeated from 20 different starting points. Standard (MICA*019). deviations were estimated using a parametric bootstrap procedure. HLA-B high-resolution typing A two-step strategy,19 low- or intermediate-resolution typing by PCR-SSOP followed by allelic typing by PCR- Acknowledgements SSP, was used for HLA-B allelic typing for all samples We thank Dr William H Hildebrand, Harriet Noreen, except for the 60 African-American donors, for which Dr Barbara Schmeckpeper, and Dr Peter Stastny for NMDP provided the HLA-B data. Confirmatory typing performing the HLA-B typing on the NMDP samples. was carried out for selected samples that deviated from expected HLA-B–MICA associations.19 References MICA high-resolution typing Exons 2–4 of the MICA gene were amplified separately 1 Bahram S, Bresnahan M, Geraghty DE, Spies T. A second as previously described.19 A panel of 64 allele- or group- lineage of mammalian major histocompatibility complex class specific probes was used to detect all known poly- I genes. Proc Natl Acad Sci USA 1994; 91: 6259–6263. morphic sites at exons 2–4 of the MICA gene and to 2 Leelayuwat C, Townend DC, Degli-Esposti MA, Abraham LJ, distinguish ambiguous allelic combinations encountered Dawkins RL. A new polymorphic and multicopy MHC gene family related to nonmammalian class I. Immunogenetics 1994; in our study. The (GCT)n microsatellite at exon 5 was 40: 339–351. 41 analyzed as described with previously reported mod- 3 The MHC sequencing consortium. Complete sequence and 19 ifications. gene map of a human major histocompatibility complex. Nature 1999; 401: 921–923. Cycle sequencing 4 Groh V, Bahram S, Bauer S, Herman A, Beauchamp M, Spies T. Low-frequency MICA alleles (MICA*017, *021, *030, Cell stress-regulated human major histocompatibility complex *041, *045, and *046) were confirmed by cycle sequen- class I gene expressed in gastrointestinal epithelium. Proc Natl cing19 following PCR-SSOP. Acad Sci USA 1996; 93: 12445–12450. 5 Zwirner NW, Fernandez-Vina MA, Stastny P. MICA, a new polymorphic HLA-related antigen, is expressed mainly by Identification of a novel MICA allele keratinocytes, endothelial cells, and monocytes. Immunoge- A new PCR-SSOP hybridization pattern, differing from netics 1998; 47: 139–148. that of MICA*019 in exon 4 by a single probe reaction, 6 Wu J, Song Y, Bakker AB et al. An activating immunoreceptor was observed in two members of an African-American complex formed by NKG2D and DAP10. Science 1999; 285: family previously studied,19 one of whom is included in 730–732.

Genes and Immunity MICA, HLA-B haplotypic variation W Tian et al 505 7 Bauer S, Groh V, Wu J et al. Activation of NK cells and T cells 30 Shiina T, Tamiya G, Oka A et al. Molecular dynamics of MHC by NKG2D, a receptor for stress-inducible MICA. Science 1999; genesis unraveled by sequence analysis of the 1,796,938-bp 285: 727–729. HLA class I region. Proc Natl Acad Sci USA 1999; 96: 8 Groh V, Rhinehart R, Randolph-Habecker J, Topp MS, Riddell 13282–13287. SR, Spies T. Costimulation of CD8ab T cells by NKG2D via 31 Curtin PD. The Atlantic Slave Trade. The University of engagement by MIC induced on virus-infected cells. Nat Wisconsin Press: Madison, WI, 1969. Immunol 2001; 2: 255–260. 32 Cosman D, Mullberg J, Sutherland CL et al. ULBPs, novel 9 Robinson J, Perez-Rodriguez M, Waller MJ et al. MICA MHC class I-related molecules, bind to CMV glycoprotein sequences 2000. Immunogenetics 2001; 53: 150–169. UL16 and stimulate NK cytotoxicity through the NKG2D 10 Steinle A, Li P, Morris DL et al. Interactions of human NKG2D receptor. Immunity 2001; 14: 123–133. with its ligands MICA, MICB, and homologs of the mouse 33 Gambelunghe G, Ghaderi M, Tortoioli C et al. Two distinct RAE-1 protein family. Immunogenetics 2001; 53: 279–287. MICA gene markers discriminate major autoimmune diabetes 11 Hankey KG, Drachenberg CB, Papadimitriou JC et al. MIC types. J Clin Endocrinol Metab 2001; 86: 3754–3760. expression in renal and pancreatic allografts. Transplantation 34 Martinez A, Fernandez-Arquero M, Balsa A et al. Primary 2002; 73: 304–306. association of a MICA allele with protection against rheuma- 12 Sumitran-Holgersson S, Wilczek HE, Holgersson J, Soder- toid arthritis. Arthritis Rheum 2001; 44: 1261–1265. strom K. Identification of the nonclassical HLA molecules, 35 Singal DP, Li J, Zhang G. Microsatellite polymorphism of the mica, as targets for humoral immunity associated with MICA gene and susceptibility to rheumatoid arthritis. Clin irreversible rejection of kidney allografts. Transplantation Exp Rheumatol 2001; 19: 451–452. 2002; 74: 268–277. 36 Glas J, Martin K, Brunnler G et al. MICA, MICB and C1_4_1 13 Hill AV, Allsopp CE, Kwiatkowski D et al. Extensive genetic polymorphism in Crohn’s disease and ulcerative colitis. Tissue diversity in the HLA class II region of Africans, with a focally Antigens 2001; 58: 243–249. predominant allele, DRB1*1304. Proc Natl Acad Sci USA 1992; 37 Orchard TR, Dhar A, Simmons JD et al. MHC class I chain-like 89: 2277–2281. gene A (MICA) and its associations with inflammatory bowel 14 Sanchez-Mazas A. African diversity from the HLA point of disease and peripheral arthropathy. Clin Exp Immunol 2001; view: influence of genetic drift, geography, linguistics, and 126: 437–440. natural selection. Hum Immunol 2001; 62: 937–948. 38 Bilbao JR, Martin-Pagola A, Vitoria JC et al. HLA-DRB1 and 15 Fernandez-Vina MA, Gao XJ, Moraes ME et al. Alleles at four MHC class 1 chain-related A haplotypes in Basque families HLA class II loci determined by oligonucleotide hybridization with celiac disease. Tissue Antigens 2002; 60: 71–76. and their associations in five ethnic groups. Immunogenetics 39 Salvarani C, Boiardi L, Mantovani V et al. Association of MICA 1991; 34: 299–312. alleles and HLA-B51 in Italian patients with Behcet’s disease. 16 Acton RT, Harman L, Go RC, Tseng ML, Bias W. Comparisons J Rheumatol 2001; 28: 1867–1870. of HLA phenotypes among African Americans from Alabama, 40 Terwilliger JD, Goring HH. Gene mapping in the 20th and 21st Maryland, and North Carolina. Transplant Proc 1993; 25: 2404– centuries: statistical methods, data analysis, and experimental 2407. design. Hum Biol 2000; 72: 63–132. 17 Just JJ, King MC, Thomson G, Klitz W. African-American HLA 41 Mizuki N, Ota M, Kimura M et al. Triplet repeat polymorph- class II allele and haplotype diversity. Tissue Antigens 1996; 48: ism in the transmembrane region of the MICA gene: a strong 636–644. association of six GCT repetitions with Behcet disease. Proc 18 Barger BO, Hudson SL, Shroyer TW et al. Influence of race on Natl Acad Sci USA 1997; 94: 1298–1303. renal allograft survival in the pre- and postcyclosporine era. 42 Schneider S, Roessli D, Excoffier L. Arlequin Version 2.000, Clin Transplant 1987: 217–233. Software for Population Genetics Data Analysis Genetics and 19 Tian W, Boggs DA, Ding WZ, Chen DF, Fraser PA. MICA genetic Biometry Laboratory, Department of Anthropology and polymorphism and linkage disequilibrium with HLA-B in 29 Ecology 2000. http://anthro.unige.ch/arlequin/. African-American families. Immunogenetics 2001; 53: 724–728. 43 Sham PC, Curtis D. Monte Carlo tests for associations between 20 Petersdorf EW, Shuler KB, Longton GM, Spies T, Hansen JA. disease and alleles at highly polymorphic loci. Ann Hum Genet Population study of allelic diversity in the human MHC class 1995; 59: 97–105. I-related MIC-A gene. Immunogenetics 1999; 49: 605–612. 44 Slatkin M. An exact test for neutrality based on the Ewens 21 Zhang Y, Lazaro AM, Lavingia B, Stastny P. Typing for all sampling distribution. Genet Res 1994; 64: 71–74. known MICA alleles by group-specific PCR and SSOP. Hum 45 Slatkin M. A correction to the exact test based on the Ewens Immunol 2001; 62: 620–631. sampling distribution. Genet Res 1996; 68: 259–260. 22 Parham P, Ohta T. Population biology of antigen presentation 46 Lewis PO, Zaykin D. GDA Version 1.1, Genetic Data by MHC class I molecules. Science 1996; 272: 67–74. Analysis Department of Ecology and Evolutionary Biology, 23 Uko GP, Asuquo MA, Dawkins RL et al. HLA class I University of Connecticut 2002. http://lewis.eeb.uconn.edu/ phenotype and gene frequencies in parts of southern Nigeria. lewishome. Dis Markers 1997; 13: 123–129. 47 Guo SW, Thompson EA. Performing the exact test of Hardy– 24 Okoye RC, Williams E, Alonso A et al. HLA polymorphisms in Weinberg proportion for multiple alleles. Biometrics 1992; 48: Nigerians. Tissue Antigens 1985; 25: 142–155. 361–372. 25 Dard P, Sanchez-Mazas A, Tiercy J-M et al. HLA-A, B, and DR 48 Ewens WJ. The sampling theory of selectively neutral alleles. differentiation among North and West African populations. Theor Popul Biol 1972; 3: 87–112. Histocompatibility testing 1991; 1: 632–636. 49 Watterson G. The homozygosity test of neutrality. Genetics 26 Greenberg JH. Languages in Africa. The Hague: Mouton, 1963. 1978; 88: 405–417. 27 Iwuagwu OC. Ethnic groups in Nigeria In: Agiri BA (ed). A 50 Weir BS. Genetic Data Analysis. Sinauer Associates, Inc.: Text on Nigerian Peoples and Culture Unpublished monograph, Sunderland, MA, 1990 pp 64–65. University of Lagos, Lagos 1996 pp 1–8. 51 Lewontin RC. On measures of gametic disequilibrium. 28 Komatsu-Wakui M, Tokunaga K, Ishikawa Y et al. MIC-A Genetics 1988; 120: 849–852. polymorphism in Japanese and a MIC-A-MIC-B null haplo- 52 Slatkin M, Excoffier L. Testing for linkage disequilibrium in type. Immunogenetics 1999; 49: 620–628. genotypic data using the expectation-maximization algorithm. 29 Shiina T, Tamiya G, Oka A et al. Nucleotide sequencing Heredity 1996; 76: 377–383. analysis of the 146-kilobase segment around the IkBL and 53 Excoffier L, Slatkin M. Maximum-likelihood estimation of MICA genes at the centromeric end of the HLA class I region. molecular haplotype frequencies in a diploid population. Mol Genomics 1998; 47: 372–382. Biol Evol 1995; 12: 921–927.

Genes and Immunity