MHC class I chain-related gene A-A5·1 allele is associated with ulcerative colitis in Chinese population


Prof Bing Xia, Department of Internal Medicine & Geriatrics, Wuhan University, Zhongnan Hospital, Donghu Road 169, Wuhan 430071, Hubei Province, P.R. of China.


The human MHC class I chain-related gene A (MICA) plays a role in regulating protective responses by intestinal epithelial Vδ1 γ δ T cells and the polymorphism of MICA were reported to be related to several autoimmune diseases. The present study aimed to investigate the association of the microsatellite polymorphisms of TM region of MICA gene with the susceptibility to ulcerative colitis (UC) in Chinese population. The microsatellite polymorphisms of the MICA were genotyped in unrelated 86 Chinese patients with UC and 172 ethnically matched healthy controls by a semiautomatic fluorenscently labelled PCR method. All the subjects were the Chinese with Han nationality. The frequency of MICA-A5·1 homozygous genotype and A5·1 allele were significantly increased in UC patients compared with healthy controls (22·1%versus 7%, P = 0·0009, Pc = 0·0126, OR = 3·781, 95%CI: 1·738–8·225 and 30·2%versus 17·4%, P = 0·0014, Pc = 0·007, OR = 2·051, 95%CI: 1·336–3·148, respectively). Adjusted the effects of gender and age at onset, MICA-A5·1 homozygous genotype and A5·1 allele were also increased in the UC patients. Moreover MICA-A5·1 allele was significantly increased in frequency in the female UC patients (38·2%versus 21·0%, P = 0·0095, Pc = 0·0475, OR = 2·326, 95%CI: 1·234–4·382). Logistic regression analysis also revealed that gender was independently associated with UC patients carried MICA-A5·1 allele (P = 0·046, OR (male) = 0·511, 95% CI: 0·264–0·987). Although the UC patients with extensive colitis (32·5%versus 17·4% in the healthy controls, P = 0·005, Pc = 0·025) and the UC patients with extraintestinal manifestations (36%versus 17·4% in the healthy controls, P = 0·0039, Pc = 0·0195) were more likely to carry the MICA-A5·1 allele, EIMs was associated with extent of disease (P < 0·0001, OR (with EIMs) = 3·511, 95% CI 1·747–7·056) and MICA-A5·1 allele was not associated with UC patients with extensive colitis or with EIMs in the logistic regression analysis. Therefore, the MICA-A5·1 homozygous genotype and A5·1 allele were closely associated with UC and the MICA-A5·1 allele was positively associated with the female UC patients in Chinese population.


The aetiology of inflammatory bowel disease (IBD), including ulcerative colitis (UC) and Crohn's disease (CD) is not totally understood. The epidemiological studies in western countries have shown strong evidence of genetic susceptibility for IBD, such as the concordance rates in twin pairs and multipley affected families [1,2]. Recently the genome wide scans and numerous replication studies found that the susceptibility loci of IBD were on chromosomes 16q (IBD1), 12q (IBD2), 6p (IBD3) and 14q (IBD4) [3–8] and suggested that IBD involves polygenic effects. HLA genes were located in  the  IBD3  region  and  were  confirmed  to  be  associated  with IBD in Japanese studies and part studies in western Caucasians. However, the association of HLA and IBD in Chinese population is not clear.

HLA class I gene encodes the class I peptides and plays a role in immune reactions. The disease predisposition and clinical phenotype of UC were found to be related to HLA class I alleles in several studies. HLA-B*52 is well known to be strongly associated with UC in Japanese patients [9,10]. HLA-B27 and Bw35 were associated with the risk of UC in Caucasians [11] and HLA-B*A33 and Cw6 were associated with UC in Asian-Indian patients [12]. In Spain, HLA-B7 was found to be overrepresented in distal UC patients compared with those with extensive UC [13].

Recently, a new family of nonclassical MHC genes, the human MHC class I chain-related genes (MIC), has recently been identified, which spanned 11 kb of a DNA stretch located 41·2 kb centromeric to the HLA-B gene [14,15]. To date at least seven MIC loci have been identified. MICA and MICB are functionally expressed genes and MICC, MICD, MICE, MICF and MICG are pseudogenes. MICA is composed of six exons that encode the leader peptide including three extracelluar domains (α1, α2 and α3), a transmembrane (TM) segment and a carboxy-terminal cytoplasmic tail. Its domain structure is similar to MHC class I antigens without β2-microglobulin [14]. MICA was expressed on gastrointestinal epithelium, where MICA molecules play a role in regulating protective responses by intestinal epithelial Vδ1 γ δ T cells [16,17]. Like MHC class I genes, the MICA locus was highly polymorphic. The microsatellite repeats (GCT/AGC)n within TM region encode for a variable number of alanines (A). The different alleles are named in A4, A5, A6 and A9 with the number of triplets that they contain. A5·1, an allelic form with an extra insertion of G/C (GGCT/AGCC), causes to a truncated protein [18,19].

Many studies have shown that MICA microsatellite alleles were associated with several autoimmune diseases, including insulin-dependent diabetes mellitus (IDDM) [20], Addison's disease [21], coeliac disease [22], rheumatoid arthritis [23], Behcet's disease [18,24] and IBD [13,25–28]. Among these studies some authors reported the relation between the genetic predisposition for UC and MICA microsatellite alleles. Sugimura et al. [25] showed that the MICA-A6 allele in the Japanese was closely related to the disease susceptibility in UC and A6 homozygous patients with UC had significantly earlier onset of the disease than patients without the A6 allele. Seki et al. [26] also found that a significant increase in MICA-A6 among Japanese UC patients could be attributed to linkage disequilibrium with HLA-B52, although they were unable to determine which genomic region would be predominantly responsible for UC on the short arm of chromosome 6. However, these results could not be repeated in studies of German and British patients, which showed that MICA alleles were not associated with UC or CD [27,28]. Later Fdez-Morera et al. [13] studied north Spain Caucasoids and found that the MICA-A5·1 allele was protective against extensive forms of UC and MICA-A5 may predispose to a worse progression of the disease.

Over the last 20 years the prevalence of UC has increased in China and we found that there are ethnic differences for distribution of HLA-DRB1 genes (our unpublished observations) and several other gene polymorphisms [29–31]. Since MICA is important in immune regulation in autoimmune diseases, we are interested to study microsatellite polymorphisms of MICA gene in relation to susceptibility to UC in Chinese patients.

Meterials and methods


A total of 86 UC patients were studied. The demographic characters and clinical features of the patients are shown in Table 1. The diagnosis of UC was based on conventional clinical, endoscopic and pathohistological criteria as described by Lennard-Jones [32]. The extent of the disease was grouped as the extensive colitis (inflammation beyond the splenic flexure) and the distal colitis (inflammation below left sided colon) by endoscopy examination. The extraintestinal manifestations (EIMs) of the patients were reviewed by records of the case-history and interviews with the patients. A group of 172 unrelated ethnically matched blood donors were used as healthy controls. All subjects were unrelated Chinese with Han nationality (Table 1). The study was approved by ethics committee of Wuhan University Medical School and all the subjects gave informed consent.

Table 1.  Demographic characteristics and clinical features of the patients with UC.
 UC (n = 86)HC (n = 172)
Gender (male/female)48/38110/62
Age at onset (year)
 Mean ± SD (range)38·6 ± 13·4 (16–79) 42·3 ± 13·5 (21–71)
Course of disease (year)
 Mean ± SD (range) 3·7 ± 1·3 (1–18) 
Family history 0 
Smoke habit (%)
 Yes20 (23·3) 
 No66 (76·7) 
Extent of disease (%)
 Extensive colitis40 (46·5) 
 Distal colitis46 (53·5) 
Colectomy (%)  1 (1·2) 
Extraintestinal manifestations (%)
 Arthritis17 (19·8) 
 Erythema nodosum  4 (4·7) 
 Oral aphtous ulcer 4 (4·7) 
 Oculopathy 2 (2·3) 
 Clubbing fingers 1 (1·2) 

DNA isolation

Genomic DNA was isolated from EDTA anticoagulated peripheral venous blood by a conventional proteinase K digestion/phenol-chloroform extraction method.

Typing of microsatellite polymorphisms in the TM region of the MICA gene

The microsatellite polymorphisms in TM region of MICA were genotyped by a semiautomatic fluorescent-labelled PCR method as previously described [18]. PCR primers were: 5′-CCTTTTTTTCAGGGAAAGTGC-3′ and 5′-CCT TACCATCTCCAGAAACTGC-3′ which was labelled at the 5′ end with HEX Amidite (Biosearch Technologies, Novato, CA, USA). The PCR reaction was carried out in a 15 µl volume containing 100 ng of genomic DNA, 2·5 mM MgCl2, 10×GeneAmp PCR Buffer II 3 µl, 200 µM of each dNTP, 6 pmol of each primer, 0·5 units of AmpliTaq Gold DNA Polymerase (Perkin Elmer, Foster City, USA). PCR conditions were as follows: an initial denaturation at 95 °C for 12min, followed by 10 cycles of 94 °C for 15 s, 55 °C for 15 s and 72 °C for 30 s and another 30 cycles of 89 °C for 15 s, 55 °C for 15 s and 72 °C for 30 s, then a final extension of 10 min at 72 °C in a GeneAmpR PCR System 9700 (PE Applied Biosystems). The number of microsatellite polymorphisms in the TM region of the MICA gene was determined in ABI PRISM 3700 Genetic Analyser (Perkin Elmer Applied Biosystems, Foster City, USA) using Genescan 3·7 and Genotyper 3·7 software and a size standard marker of GENESCAN 400HD (ROXTM) standard (Applied Biosystems). Five distinct alleles consisting of CGT repetitions were designated as A4 (179 bp), A5 (182 bp), A5·1 (183 bp), A6 (185 bp), and A9 (194 bp).

Statistical analysis

All data were analysed using SPSS12 statistical software (SPSS, Inc., Chicago, IL, USA). The genotype and allelic frequencies were calculated by direct counting and statistical comparisons were performed by the χ2 test with Yates’ correction or Fisher's exact test. Odds ratios (OR) and 95% confidence intervals (CI) were calculated for the disease in carriers of the specific alleles. Bonferroni multiple correction was used for the corrected P (Pc) by multiplying the P-value by the number of the statistical tests in order to avoid the α1 error. The Pc < 0·05 was considered to be statistically significant. Logistic regression was used to analyse the influence of the polymorphism of MICA adjusted gender and age at onset. And this analysis was also performed to use the extent of disease and EIMs as dependent variable, respectively, and independent variables including the MICA alleles.


Distribution of MICA allele and genotype frequencies in patients with UC and healthy controls

The frequencies of MICA allele and genotype between UC patients and the healthy controls are summarized in Table 2. The MICA-A5·1 allele was significantly increased in frequency in UC patients compared to healthy controls (30·2%versus 17·4%, P = 0·0014, Pc = 0·007, OR = 2·051, 95%CI: 1·336–3·148). Logistic regression analysis to examine the effects of alleles as well as risk factors, gender and age at onset in the UC patients revealed that the MICA-A5·1 allele (P = 0·003, OR = 1·924, 95%CI: 1·243–2·978), age at onset (P = 0·006, OR = 0·980, 95%CI: 0·966–0·994) and gender (P = 0·033, OR (male) = 0·671, 95%CI: 0·465–0·968), in descending order of statistical significance, affected the prevalence of UC. There were 14 MICA genotypes found in the UC patients and healthy controls. Among them, the frequency of MICA-A5·1/A5·1 homozygous genotype was increased in the UC patients (22·1%versus 7% in healthy controls, P = 0·0009, Pc = 0·0126, OR = 30781, 95%CI: 1·738–8·225). Adjusted for the impact of gender and age at onset, the MICA-A5·1/A5·1 homozygous genotype was also significantly differently distributed between the UC patients and healthy controls (P = 0·0004, OR = 1·729, 95%CI: 1·191–2·509). The other alleles or genotypes of MICA did not shown significant association with UC in Chinese patients when given Bonferroni correction.

Table 2.  MICA genotype and allele frequencies in patients with ulcerative colitis (UC) and healthy controls (HC).
GenotypeGenotype frequenciesAlleleAllele frequencies
(n = 86)
n (%)
(n = 172)
n (%)
(2n = 192)
n (%)
HC (2n = 344)
n (%)
  • *

    P = 0·0009, Pc = 0·0126, OR = 3·781, 95%CI: 1·738–8·225;

  • P = 0·0014, Pc = 0·007, OR = 2·051, 95%CI: 1·336–3·148.

A4/A4 2 (2·3) 1 (0·6)A4 9 (5·2) 42 (12·2)
A4/A5 2 (2·3)15 (8·7)   
A4/A5·1 2 (2·3)14 (8·1)   
A4/A6 0 (0) 4 (2·3)   
A4/A9 1 (1·2) 7 (4·1)   
A5/A526 (30·2)57 (33·1)A566 (38·4)155 (45·1)
A5/A5·1 0 (0) 0 (0)   
A5/A6 5 (5·8)12 (7·0)   
A5/A9 7 (8·1)14 (8·1)   
A5·1/A5·119 (22·1)*12 (7·0)A5·152 (30·2) 60 (17·4)
A5·1/A6 3 (3·5) 9 (5·2)   
A5·1/A9 9 (10·5)13 (7·6)   
A6/A6 5 (5·8) 3 (1·7)A620 (11·6) 38 (11·0)
A6/A9 2 (2·3) 7 (4·1)   
A9/A9 3 (3·5) 4 (2·3)A925 (14·5) 49 (14·2)

The increased frequency of the MICA-A5·1 allele in female patients with UC

The results above suggest gender could be an important risk factor for UC patients. The distribution of MICA allele frequencies in female and male patients is shown in Table 3. The frequency of A5·1 was significantly increased in the female patients with UC compared with healthy female controls (38·2%versus 21·0%, P = 0·0095, Pc = 0·0475, OR = 2·326, 95%CI: 1·234–4·382). Although the frequencies of the A5 were decreased in the female patients with UC compared with the female controls, the difference was not statistically significant after P-value was corrected. At the same time, no statistical differences were found both between the male patients and the corresponding healthy controls and between the male and female patients.

Table 3.  Distribution of MICA microsatellite polymorphisms in female and male patients with ulcerative colitis (UC) and healthy controls (HC).
(2n = 76)
n (%)
(2n = 124)
n (%)
(2n = 96)
n (%)
(2n = 220)
n (%)
  • *

    UC female patients versus healthy female controls: P = 0·0095, Pc = 0·0475, OR = 2·326, 95%CI: 1·234–4·382

A4 3 (3·9)13 (10·5)6 (6·3)29 (13·2)
A524 (31·6)58 (46·8)42 (43·8)97 (44·1)
A5·129 (38·2)*26 (21·0)23 (24·0)34 (15·5)
A6 7 (9·2)9 (7·3)13 (13·5)29 (13·2)
A913 (17·1)18 (14·5)12 (12·5)31 (14·1)

Association of the MICA alleles with subgroups of UC

When patients were stratified according to their clinical manifestations, disease extent and EIMs (Table 4), the frequency of the A5·1 allele was significantly increased both in the patients with the extensive colitis (32·5%versus 17·4% in healthy controls, P = 0·005, Pc = 0·025, OR = 2·279, 95%CI: 1·322–3·929) and in patients with EIMs (36%versus 17·4% in healthy controls, P = 0·0039, Pc = 0·0195, OR = 2·663, 95%CI: 1·402–5·056) and the frequency of the A4 was also further decreased in patients with the extensive colitis than in the healthy controls (2·5%versus 12·2%, P = 0·0075, Pc = 0·0375, OR = 0·1844, 95%CI: 0·04366–0·7786). When comparison between patients with extensive colitis and distal colitis and comparison between patients with EIMs and without EIMs were performed, the differences of the distributions of the MICA alleles did not reach statistical significance.

Table 4.  Association of MICA alleles with subgroups of ulcerative colitis (UC) according to the disease extent, extraintestinal manifestations (EIMs)
Distal colitis
(2n = 92)
Extensive colitis
(2n = 80)
(2n = 50)
(2n = 122)
(2n = 344)
  • *

    versus healthy controls: P = 0·0075, Pc = 0·0375, OR = 0·1844, 95%CI: 0·04366–0·7786;

  • versus healthy controls: P = 0·005, Pc = 0·025, OR = 2·279, 95%CI: 1·322–3·929;

  • versus healthy controls: P = 0·0039, Pc = 0·0195, OR = 2·663, 95%CI: 1·402–5·056

A47 (7·6) 2 (2·5)* 1 (2·0)8 (6·6) 42 (12·2)
A535 (38·0)31 (38·8)15 (30·0)51 (41·8)155 (45·1)
A5·126 (28·3)26 (32·5)18 (36·0)34 (27·9) 60 (17·4)
A613 (14·1) 7 (8·8) 5 (10·0)15 (12·3) 38 (11·0)
A911 (12·0)14 (17·5)11 (22·0)14 (23·0) 49 (14·2)

When logistic regression analysis was performed considering MICA-A5·1 allele as a dependent variable, gender, age at onset, extent of disease and EIMs as independent variables, the only associated independent variable was gender (P = 0·046, OR(male) = 0·511, 95% CI: 0·264–0·987). When extent of disease was considered as a dependent variable and independent variables including gender, age at onset, EIMs and MICA-A5·1 allele, only EIMs was associated with extent of disease (P < 0·0001, OR(with EIMs) = 3·511, 95% CI 1·747–7·056).

The frequencies of MICA alleles in different ethnic population

Compared with the previous published data [25,27] from the healthy controls in different ethnic populations, some differences were found among the Chinese, Japanese and Germans (Table 5).

Table 5.  The frequencies of MICA alleles in Chinese, Japanese and German populations.
(n = 172)
(n = 132)
(n = 154)
  • *

    MICA-A5 allele in the three population: chi-squared (3 

  • ×

    2 table): 76·057, P < 0·0001;

  • MICA-A5·1 allele in the three population: chi-squared (3 

  • ×

    2 table): 49·825, P < 0·0001;

  • MICA-A6 allele in the three population: chi-squared (3 

  • ×

    2 table): 23·440, P < 0·0001.

A4 42 (12·2)38 (14·4) 36 (11·7)
A5155 (45·1)*80 (30·3) 42 (13·6)
A5·1 60 (17·4)39 (14·8)114 (37·0)
A6 38 (11·0)67 (25·4) 69 (22·4)
A9 49 (14·2)40 (15·2) 47 (15·3)


In the present study, we have investigated a trinucleotide repeat polymorphisms in TM region of MICA in the Chinese patients with UC. Our results were not consistent with the previous studies in Japanese and Europeans. We did not show that the A6 allele was associated with UC, but found that MICA-A5·1/A5·1 homozygous genotype and the A5·1 allele were associated with susceptibility to UC. Moreover, logistic regression analysis confirmed that the MICA-A5·1 allele was more likely to be related to female patients with UC, as it is the first report and was not found in other nationalities. When the different risk factors such as gender, age at onset, disease extent and EIMs were compared multivariately in UC patients, the logistic regression analysis presented gender (male) as the independent risks factor that negatively effects the association of MICA-A5·1 allele with UC patients. That means that female UC patients have a higher frequency of MICA-A5·1 allele than male ones, which confirmed the result above. Patients with extensive colitis are more likely to have EIMs, which can explain why a higher frequency of MICA-A 5·1 allele existed both in the UC patients with extensive colitis and in those with EIMs than healthy controls. But MICA-A5·1 allele was not associated with UC patients with extensive colitis or with EIMs in the logistic regression analysis.

We selected previous published articles from different ethnic populations according to the same methods for MICA genotyping and statistical methods. Comparing the frequencies of the MICA allele in Chinese healthy controls with the Japanese and German, we found a significantly difference in the distribution of the MICA alleles in the different populations. This finding may explain ethnic differences in MICA gene distribution.

We found the MICA-A5·1/A5·1 homozygous genotype and the A5·1 allele were associated with susceptibility to UC in the Chinese population. And MICA-A5·1 allele carries a nucleotide insertion resulting in a premature stop codon that may encode a secreted or a truncated protein. As for the MICA protein, it and the closely related MICB were recognized by intestinal epithelial T cells expressing diverse Vδ1 γ δ TCRs [17] and Suemizu et al. [33] found that the cytoplasmic tail-deleted protein as the naturally occurring A5·1 allele are aberrantly transported to the apical surface of epithelial cells of the intestinal tissue, while native expression of MICA (the full-length MICA protein) is sorted to the basolateral surface,  the  site  of  putative  interaction  with  intraepithelial  T and NK lymphocytes. Groh et al. [34] reported that circulating tumour-derived soluble MICA caused the down-regulation of NKG2D and in turn severe impairment of the responsiveness of tumour-antigen-specific effector T cells. So we have a hypothesis that the secreted or a truncated protein encoded by MICA-A5·1 allele may influence immunological function located in intestinal tract or some other systems and this possible mechanism may play an important role in the pathogenesis of UC, especially perform special function in the onset of the female UC patient. Then functional consequences of the transmembrane polymorphism of MICA and other genes linkage disequilibrium with MICA may be studied further in Chinese population.


This study was supported by a grant from the State Natural Science Foundation of China (30370638) to Prof Bing Xia.