MHC class I chain-like gene A (MICA) and its associations with inflammatory bowel disease and peripheral arthropathy

Abstract

MHC class I chain gene A (MICA) is a non-classical Class I gene which is expressed on the surface of epithelia without β2-microglobulin. The gene is found in the major histocompatibility complex (MHC) in tight linkage disequilibrium with human leucocyte antigen-B (HLA-B). Its precise function is unknown, but it interacts with γδ T cells of the intestinal immune system. This region of the MHC has been implicated in inflammatory bowel disease (IBD) pathogenesis by recent association mapping studies and this study was performed to examine the prevalence of MICA gene polymorphisms in IBD, in particular in type 2 peripheral arthropathy (PeA), which also has a strong HLA–B association. An assessment of the prevalence of MICA polymorphisms in IBD was made. Blood from 50 ulcerative colitis (UC) and 50 Crohn's disease controls was taken and MICA status determined using allele-specific PCR for 16 known alleles of MICA. A further 91 UC patients were recruited to confirm the results of this stage, and then the polymorphisms were studied in 52 type 1 and 45 type 2 PeA patients. The MICA status of these groups was compared with 118 blood and organ donor controls with appropriate correction for multiple comparisons. UC overall was associated with possession of MICA*007 in 32% compared to 11% of controls (P_c = 0·017). This association was confirmed in a second cohort of 91 patients (23% versus 11%, P = 0·02). These were independent of HLA class I status. Type 2 IBD PeA was associated with MICA*008 in 98% compared to 73% of controls (P = 0·0001). MICA*007 is associated with susceptibility to UC in our population and MICA*008 with type 2 IBD PeA. Further work is now required to assess the distribution and expression of MICA throughout the gut in health and disease.

Introduction

The pathogenesis of inflammatory bowel disease (IBD) is complex, and currently incompletely understood. However, many models of pathogenesis implicate the intestinal immune response. One key element in the regulation of this response involves antigen presentation by the human leukocyte antigen (HLA) system, and previous studies have demonstrated that HLA class II expression is up-regulated in inflamed ileum in Crohn's disease [1], and initiated in colonic epithelium in ulcerative colitis (UC) [2].

Genes for the HLA have also been a focus of genetic research in IBD. Genome-wide scanning has demonstrated linkage with the HLA region, particularly in UC [3,4], and association studies have demonstrated HLA associations. These include a widely replicated association between HLA-DRB1*0103 and extensive UC [5–7], and DRB1*03 with protection against fistulizing Crohn's disease [8]. Association studies have also implicated HLA in the pathogenesis of arthritis associated with IBD. The strongest associations are with HLA-DRB1*0103 in type 1 (pauci-articular) and with the class I antigen HLA-B*44 in type 2 (polyarticular), which is possessed by 62% of patients [9].

Association mapping studies have suggested recently that there are important susceptibility loci for UC near the TNFα gene, although this gene itself is probably not a susceptibility locus [10]. The MICA gene is one of a new group of genes that has been described recently which map to this area in the major histocompatibility complex (MHC) on chromosome 6. It is situated centromeric to HLA-B, between HLA-B and TNFα, and is in tight linkage disequilibrium with HLA-B. MICA (and the less well characterized MICB) [11–14], code for non-classical class I molecules, and like classical class I molecules it is expressed at the cell surface. However, unlike classical class I molecules it appears to be expressed without β2-microglobulin [11]. Its distribution appears to be restricted to epithelial surfaces, particularly the gastrointestinal tract [15,16].

MICA is known to be polymorphic, with at least 16 alleles having been described. The gene has five exons encoding a cytoplasmic tail, a transmembrane region and three extracellular domains (α1, α2 and α3) [11,13], and the polymorphisms are in the extracellular region where they may have a functional significance.

Thus MICA is a good functional and positional candidate for involvement in the genetic susceptibility to IBD, and UC in particular. In addition, the strong association found previously between type 2 PeA and class I HLA makes MICA an important possible candidate for involvement in type 2 PeA.

This study was undertaken to assess the prevalence of MICA allele polymorphisms in IBD compared to healthy controls, and to determine whether the presence of type 2 PeA was associated with specific MICA alleles.

Methods

EDTA blood was collected from 50 UC patients and 50 Crohn's disease patients without extra-intestinal manifestations attending the Oxford IBD Clinic in order to establish the prevalence of MICA alleles in the IBD population. A second cohort of 91 UC patients was recruited from Oxford and the neighbouring district in order to confirm the results from the first cohort. In addition DNA was obtained from patients identified previously as having type 1 or type 2 peripheral arthropathy. Fifty-two patients with type 1 and 45 with type 2 were studied. The ethnic composition of the patient and controls groups was similar, all patients being North European Caucasoid with less than 5% having Jewish ancestry.

DNA was separated from peripheral blood leucocytes by a modified salting-out technique, as described previously, and was dissolved in sterile water with concentrations of 10–100 ng/ml. MICA determination was performed using a PCR-based technique to detect the 16 polymorphisms described by Fodil et al. [13], according to the method described by Stephens et al. [17]. The PCR products were run for 20 min with a voltage of 200 mV on 1% agarose gels with 1% ethidium bromide and MICA alleles were assigned according to the primers that had amplified. In order to determine the role of linkage disequilibrium with HLA-B in the positive MICA associations HLA × B genotyping was performed according to the method described by Bunce et al. [18] on the second cohort of 91 patients and the PeA patients.

Data analysis

The MICA status of IBD patients was compared with healthy controls to test the hypothesis that MICA is a susceptibility gene to IBD. In patients with type 1 and type 2 IBD peripheral arthropathy MICA status was compared with healthy controls, IBD controls and with each other using contingency tables and Fisher's exact test to test the hypothesis that MICA is a susceptibility gene for IBD overall and to type 2 PeA. The Bonferroni correction was used to control for multiple comparisons using a factor of 16 (the number of alleles being tested). The MICA status was also analysed in relation to the HLA-B status of the patients as there is linkage disequilibrium between the two to ensure that the primary association was not with HLA-B.

Results

The results from the first cohort of 50UC patients and 50 Crohn's disease patients are shown in Table 1. There was a significant association between MICA*007 and UC, but not with Crohn's disease. As a consequence of this finding a further cohort of 91 UC patients underwent MICA genotyping. The results from this cohort are shown in Table 2. The association between MICA*007 and UC was maintained in this group. No class I HLA associations were present; in particular there was no association between IBD and HLA-B*27 (with which MICA*007 is in linkage disequilibrium). When the two datasets are combined MICA*007 is found in 37/141 patients (26%) (P = 0·001). The differences in MICA*007 between UC and controls are also significant when allele frequencies are analysed (38/282 versus 13/236, P = 0·003 for combined datasets).

Table 1.

MICA allele associations for inflammatory bowel disease compared to organ and blood donor controls

	Controls (n = 118)		UC (n = 50)		Crohn's disease (n = 50)
MICA allele	Number	%	Number	%	Number	%
MICA*002	41	35	18	36	13	26
MICA*004	22	19	7	14	10	20
MICA*006	5	4	2	4	1	2
MICA*007	13	11	16*	32	6	12
MICA*008	86	73	36	72	35	70
MICA*009	11	9	3	6	6	12
MICA*010	11	9	4	8	6	12
MICA*011	0	0	3	6	0	0
MICA*016	2	2	1	2	3	6

Alleles 1, 3, 5, 12, 14 and 15 were not seen in these patients.

^*P versus controls = 0·0011, P_c = 0·017.

Table 2.

Prevalence of common MICA genotypes in a second cohort of UC patients

	Controls (n = 118)		UC patients (n = 91)
MICA allele	Number	%	Number	%
MICA*002	41	35	20	22
MICA*004	22	19	15	16
MICA*006	5	6	4	4
MICA*007	13	11	21*	23
MICA*008	86	73	69	76
MICA*009	11	9	7	8
MICA*010	11	9	9	10
MICA*011	0	0	3	3
MICA*016	2	2	4	4

^*P = 0·02.

Fifty-two patients with type 1 and 45 with type 2 IBD PeA underwent MICA genotyping. The results are shown in Table 3. There was an association between MICA*007 and type 1 PeA, but in this group there is also an association with HLA-B27 [9], and linkage disequilibrium between MICA*007 and HLA-B*27 probably accounts for this finding. This is demonstrated in the healthy controls, where 46% of MICA*007-positive subjects were also HLA-B*27-positive. MICA*008 was the most common allele overall, but was increased significantly in type 2 PeA. This allele is linkage disequilibrium with HLA-B*44, but the strength of the association (98%) suggests that the MICA association itself may be of importance.

Table 3.

MICA allele associations in IBD type 1 and type 2 peripheral arthropathies

	Controls (n = 118)		IBD controls (n = 100)		Type 1 PeA (n = 52)		Type 2 PeA (n = 45)
MICA alleles	No.	%	No.	%	No.	%	No.	%
MICA*002	41	35	31	31	15	29	8	18
MICA*004	22	19	17	17	3*	6	9	20
MICA*006	5	4	3	3	3	6	0	0
MICA*007	13	11	22	22	15***	29	7	16
MICA*008	86	73	71	71	37	71	44**	98
MICA*009	11	9	9	9	2	4	3	7
MICA*010	11	9	10	10	5	10	3	7
MICA*011	0	0	3	3	0	0	0	0
MICA*016	2	2	4	4	3	6	1	2
HLA-B*27	6	5	9	9	14	27	3	7
HLA-B*44	43	36	28	28	5	10	27	60

^*P versus controls = 0·03, P_c = n.s.;

^**P versus controls = 0·0001, P_c = 0·0016;

^***P versus controls = 0·008, P_c = 0·013.

Discussion

This study demonstrates an association between MICA and susceptibility to ulcerative colitis but not Crohn's disease, a finding confirmed in a second separate dataset. This would be consistent with the findings of earlier studies, including the Oxford IBD genome screen, which demonstrated linkage with microsatellite markers around the MHC in UC but not Crohn's disease [3,5] although some recent studies have also found an association in Crohn's disease [19]. The MICA allele associated with UC, MICA*007, is in linkage disequilibrium with HLA-B*27. In controls 6/13 MICA*007-positive cases are HLA-B*27-positive. However, previous studies of HLA class I status have failed to demonstrate an association between susceptibility to UC overall and HLA-B27 [20,21] and this is confirmed in this study. However, we have demonstrated previously a significant association between type 2 IBD PeA and HLA-B*44 [9]. As with all association studies, it is not possible to determine whether this association is with HLA-B*44 itself or with a gene in tight linkage disequilibrium. This study demonstrates that type 2 IBD PeA is associated strongly with MICA*008, which is in linkage disequilibrium with HLA-B*44 (34/86 (40%) of MICA*008 patients are HLA-B*44). Given the strength of the MICA association it seems possible that this is the primary association, although there are many potential candidate genes in the MHC region. The role of MICA in the intestine is not yet clear, but recent studies have shown that MICA is unregulated in conditions of cellular stress, and that MICA binds to a receptor found on NK cells and intestinal lymphocytes − NKGPD2a [22] − causing cellular activation. Studies in human cell lines also suggest that MICA interacts with γδ lymphocytes which express α_Eβ₇ integrin on their surface [23] − a characteristic of a population of intestinal intraepithelial lymphocytes.

Thus MICA provides an exciting area of research in inflammatory disorders involving the gut. Further work is now required to confirm these associations and to attempt to narrow them in the light of new polymorphisms being described in the MICA gene. In addition, much needs to be done to elucidate the physiological function and expression of MICA and related molecules, such as MICB, in the healthy and diseased gut.