Association of NKG2D gene variants with susceptibility and severity of rheumatoid arthritis

Summary

NKG2D (KLRK1) is a C‐type lectin receptor present on natural killer (NK) cells, γδ, CD8⁺ and CD4⁺ T cells. Upon ligand binding, NKG2D mediates activatory and co‐stimulatory signals to NK cells and activated CD4⁺ T cells, respectively. Polymorphisms in NKG2D predispose to infectious diseases, cancer, transplantation and autoimmune disorders. We studied the influence of this NK receptor polymorphism on predisposition to and modification of the disease phenotype in patients with rheumatoid arthritis (RA). Eight different single nucleotide polymorphisms (SNP) in the NKG2 gene were genotyped in 236 patients with RA and 187 controls using Taqman 5' nuclease assays. NKG2D genotype/allele frequency did not differ between patients and controls. Subgroup analysis showed that the frequency of A allele of NKG2D9 and T allele of NKG2D10 was significantly higher in patients with deformities (a marker of severe disease) [11 versus 5%, Pc = 0·03, odds ratio (OR) = 2·44, 95% confidence interval (CI) = 1·09‐5·98 and 10 versus 4%, Pc = 0·04, OR = 2·45, 95% CI = 1·05‐6·39, respectively], while the frequency of alleles G of NKG2D9 and A of NKG2D10 was greater in patients without deformities (Pc = 0·03, OR = 0·41, 95% CI = 0·17‐0·91 and Pc = 0·04, OR = 0·41, 95% CI = 0·16‐0·96). Similar trends of association were observed with deforming phenotype of RA in female patients and deforming young onset RA subgroups. Haplotype analysis revealed that the frequency of haplotype G‐C‐A‐G‐A‐T‐C‐C was higher in patients than in controls (12 versus 8%, P = 0·04, OR = 1·61, 95% CI = 1·01‐2·55), suggesting that it may predispose to RA. Our study suggests that the NKG2D gene polymorphisms may modify the risk of development and severity of RA.

Introduction

Rheumatoid arthritis (RA), a systemic autoimmune disease, is characterized by irreversible joint damage and deformities. It affects 0·5–1·0% of adults in developed countries 1. Synovial inflammation and hyperplasia, autoantibody production, cartilage and bone destruction over time characterize RA. The accompanying disabilities in RA impact significantly upon quality of life and socio‐economic status of the patient, the family and society at large 2. The most significant risk factors for RA include female gender, advanced age and a family history of such disorders. The presence of familial tendency suggests a genetic predisposition to the development of RA. Genetic predisposition is thought to account for approximately one‐half to two‐thirds of the risk of developing RA 3. The precise mechanism through which genetic factors act is still unclear. It is likely to be triggered by complex and stochastic interactions between genetic and environmental factors 2.

Human and mouse natural killer (NK) cells express both activating and inhibitory cell surface receptors. Fine‐tuning of immune surveillance (activated or quiescent state of NK cells) involves the integration of both activating and inhibitory receptor signalling 4. Among many NK cell receptors, the most studied receptors are the killer‐cell immunoglobulin‐like receptors (KIR). Several polymorphisms in the KIR genes have been identified, and the association of KIR genes with RA have been reported in Caucasian 5, Polish 6, Taiwanese 7, Mexican 8 and Iranian 9 populations. In a case–control study involving 100 North Indian subjects with RA the presence of KIRDS1 and KIRDS2 increased the risk of RA, whereas KIRDL1, KIRDL2 and KIRDL3 were protective. The presence of KIRDS1 and KIRDS3 increased the risk of extra‐articular complications and deforming disease in patients with RA 10. KIR gene polymorphisms are reported to influence disease severity and treatment response to methotrexate 11, 12.

In humans, the NK gene complex on chromosome 12p12–p13 harbours 19 genes that encode for C‐type lectins NKG2 and CD94. NKG2D lies amid this cluster of genes, referred to as ‘NK complex’ (NKC). Several genes expressed by NK cells, including KLRD1 (CD94) on the centromeric side and KLRC4 (NKG2F), KLRC3 (NKG2E), KLRC2 (NKG2C) and KLRC1 (NKG2A) on the telomeric side, are located within this complex. NKG2D (KLRK1) encoded by killer cell lectin‐like receptor subfamily K is an activatory C‐type lectin receptor on NK cells, gamma delta T (γδ T) cells, CD8⁺ and CD4⁺ T cells. The NKG2D receptor binds to a variety of ligands, including major histocompatibility complex (MHC) class‐I polypeptide sequence A (MICA), MHC class‐I polypeptide sequence B (MICB), UL16 binding protein (ULBP)−1, ULBP‐2, ULBP‐3, ULBP‐4 and retinoic acid early transcript 1 G (RAET1G). Ligand binding of NKG2D mediates activatory signals to NK cells and co‐stimulatory signals to activated CD4⁺ T cells. NKG2D expressed on the surface of most cytotoxic CD8⁺ T cells acts as a co‐stimulatory receptor for naive CD8⁺ T cells, enabling their transformation to activated CD8⁺ T cells. Regulation of ligand expression is important for immune homeostasis, as inappropriate expression in normal tissues favours autoimmune processes. Conversely, failure to up‐regulate ligand expression in pathological conditions favours the development of cancers or dissemination of intracellular infections 13.

Most of the published literature related to NKG2D in autoimmune diseases concerns MIC molecules, MICA and MICB. Genetic analyses have revealed a predisposition to develop RA in Tunisian patients bearing the A/A genotype of the single nucleotide polymorphism (SNP) MICA‐250. The same study also revealed increased risk of seropositive RA in those individuals carrying the G allele and G/G genotype of MICA‐250 SNP and the (val) allele and (val/val) genotype of MICAmet129val SNP 14. An autoimmune response is triggered if MIC molecules are induced non‐specifically by an inflammatory process or expressed inappropriately on the cells in predisposed individuals 15. Dysregulated receptor/ligand expression was first reported in RA for NKG2D/MICA 16. High levels of interleukin (IL)−15 and tumour necrosis factor (TNF)‐α found in the serum and inflamed joints of RA patients induce expansion of autoreactive CD4⁺CD28⁻NKG2D⁺ cells in peripheral blood and synovial tissue 16. MICA and MICB up‐regulated in RA synoviocytes are capable of activating these autoreactive autologous T cells in an NKG2D‐dependent manner 17. While an abundance of soluble MICA (sMICA) has been found in the serum and synovial fluid of RA patients, it is unable to down‐regulate the activating NKG2D receptors on NK cells due to massive NKG2D up‐regulation driven by elevated levels of IL‐15 and TNF‐α. These reports suggest that dysregulation of NKG2D and abnormal expression of MIC in the local tissue environment can cause autoreactive T cell stimulation, contributing to the pathogenesis of RA 15. There is a paucity of data on the role of NKG2D polymorphism on disease susceptibility, severity and outcomes in RA, with the exception of one study from Korea 18. Hence, this study was carried out to identify the contribution of polymorphisms in NKG2D gene towards risk of the development of RA and their influence on disease characteristics and response to methotrexate‐based synthetic disease‐modifying anti‐rheumatic drug (DMARD) therapy in South Indian Tamil patients.

Methods

Study design

This was a case–control immunogenetic study conducted at the Jawaharlal Institute of Postgraduate Medical Education and Research (JIPMER), Puducherry, a major tertiary care referral center in South India.

Study subjects

Ethnic Tamil South Indian patients fulfilling the 2010 modified American College of Rheumatology Criteria (ACR) 19 for rheumatoid arthritis were recruited for the study. A total of 236 patients (215 females and 21 males) were enrolled. Patients with inactive disease or other co‐morbid conditions were excluded. The mean age of patients at enrolment was 40·9 ± 11·1 years (range = 18–72 years) and age of onset was 37·3 ± 11·6 years (range = 17–72 years). Their clinical characteristics are presented in Table 1. All patients received methotrexate (MTX) as initial therapy up to a maximum of 25 mg per week. In case of inadequate response, sulphasalazine (SSZ) up to 2 g per day and hydroxychloroquine (HCQ) up to 400 mg per day, with or without low‐dose prednisolone (< 7·5 mg per day) was added. Treatment response was assessed using European League Against Rheumatism (EULAR) response criteria 20 at the end of 6 months or after 8 continuous weeks of a stable combination of DMARD therapy, whichever was later. Clinical characteristics assessed were age of onset (young onset RA, or YORA – onset before 55 years); presence or absence of deformities; and presence or absence of extra‐articular features related to RA (rheumatoid nodules, sicca symptoms, lung involvement, neuropathy and vasculitis).

Table 1.

Characteristics of the patients with rheumatoid arthritis

Characteristics	n = 236 (%)
Clinical features
Female RA	215 (91)
YORA	206 (87)
Erosive deforming disease	142 (60)
Extra articular manifestations	53 (22)
Autoantibody status
IgM‐, RF‐positive	177 (75)
ACPA‐positive	137 (58)
Treatment response
Good responders	54 (23)
Moderate responders	158 (67)
Non‐responders	24 (10)

Ig = immunoglobulin; RA = rheumatoid arthritis; YORA = young onset RA; RF = rheumatoid factor; ACPA = anti‐citrullinated peptide antibody.

A total of 187 (157 females and 30 males) unrelated age‐, sex‐ and ethnicity‐matched healthy individuals without family history of autoimmune disorders were included as healthy controls (HC). Written informed consent was obtained from all study participants before enrolment. The study was reviewed and approved by the Institutional Ethics Committee.

Autoantibody measurement

Immunoglobulin (Ig)M rheumatoid factor (RF) in the serum of patients was measured by Nephelometry (Seimens Health Care, Marburg, Germany N Prospec^®; Dade Behring, Germany) using kits supplied by Siemens Healthcare Diagnostics Inc. (Marburg, Germany). RF titres above 10 IU/ml were considered positive. ACPA status was determined by sandwich enzyme‐linked immunosorbent assay (ELISA) method using a second‐generation commercial ELISA kit (Biosystems, Barcelona, Spain). Patients with values of ACPA above 15 IU/ml were labelled as ACPA‐positive.

DNA extraction

Genomic DNA from patients and controls was extracted from peripheral blood leucocytes using a standard protocol. The quality and quantity of DNA were analysed by spectrophotometry (Nanodrop, ThermoScientific, Fremont, CA, USA).

NKG2D and NKG2A genotyping

We evaluated seven single nucleotide polymorphisms (SNPs) in NKG2D gene (rs1049174: NKC‐3, rs2255336: NKC‐4, rs2617160: NKC‐7, rs2246809: NKC‐9, rs2617169: NKC‐10 and rs2617170: NKC‐11, rs2617171: NKC‐12) and one SNP in NKG2A gene (rs1983526: NKC‐17) using a predesigned TaqMan 5' nuclease assay (Applied Biosystems, Foster City, CA, USA) with allele‐specific fluorogenic oligonucleotide probes allowing discrimination of each pair of alleles under investigation. Briefly, 50 ng of DNA was amplified in a 10 µl reaction containing 5 µl of TaqMan genotyping master mix, 0·25 µl of ×40 allelic discrimination mix and 3·75 µl of DNase and RNase free water using the following polymerase chain reaction (PCR) cycling conditions: preread at 60°C for 1 min, hold at 50°C for 2 min, 95°C for 10 min followed by 40 cycles of 95°C for 15 s, 60°C for 1 min and postread at 60°C for 1 min for all the above polymorphisms of NK receptors except NKG2D12. PCR conditions for NKG2D12 was preread at 60°C for 1 min, hold at 50°C for 2 min, 95°C for 10 min followed by 40 cycles of 95°C for 15 s, 60°C for 1 min 30 s and postread at 60°C for 1 min 21. PCR plates were run and read in the ABI StepOne Plus™ Sequence Detection System (Applied Biosystems). Results were analysed using allelic discrimination software.

Statistical analysis

Sample size was calculated using the CaTS power calculator for genetic studies 22, taking into account the described minor allele frequency with significance level of 0·05 and power of 0·8. Allele and genotype frequencies of the studied SNPs were compared between patients and controls using the χ² test with Yates' correction. P‐values less than 0·05 were considered significant. Both odds ratio (OR) and 95% confidence interval (95% CI) were calculated to assess the relative risk conferred by a specific allele or genotype. Gender, age of onset, disease activity, erosive/deforming disease, extra‐articular manifestations, autoantibody (RF, ACPA) positivity and response to DMARD therapy were the characteristics included in the analysis. Deviation from Hardy–Weinberg equilibrium was analysed using χ² testing. The statistical software GraphPad InStat was used for carrying out the statistical analysis. Haplotype and linkage disequilibrium (LD) analysis between the three SNPs was performed with the Haploview program version 4.2 (www.broadinstitute.org/haploview). Lewontin's D' measure was used to estimate the intermarker coefficient of LD. Two markers with a D' value greater than 0·7 were defined to be part of the same haplotype block.

Results

Distribution of NK receptor genotypes/alleles

Eight different SNPs clustered in the NK complex (rs1049174, rs2255336, rs2617160, rs2246809, rs2617169, rs2617170, rs2617171 and rs1983526) were genotyped in 236 patients and 187 controls. The genotype/allele frequency did not differ between patients and controls for all the eight different polymorphisms analysed in this study. However, the CC genotype of NKC‐12 (rs2617171) (22 versus 15%, Pc = 0·07, OR = 1·64, 95% CI = 0·97–2·84) and the CC genotype of NKC‐17 (rs1983526) (38 versus 29%, Pc = 0·06, OR = 1·52, 95% CI = 0·97–2·34) were higher in patients compared to controls, albeit not to a level of statistical significance, suggesting a possible role in susceptibility to RA (Table 2).

Table 2.

Distribution of natural killer (NK) receptor alleles and genotypes in study subjects

	RA, n = 236, %	HC, n = 187, % P		Pc	OR (95% CI)
NK3 (rs1049174)
CC	86 (36)	75 (40)	n.s.
CG	97 (41)	83 (44)	n.s.
GG	53 (23)	29 (16)	n.s.
NK4 (rs2255336)
GG	197 (84)	157 (84)	n.s.
GA	36 (15)	30 (16)	n.s.
AA	3 (1)	0 (0)	n.s.
NK7 (rs2617160)
AA	49 (21)	26 (14)	n.s.
AT	105 (44)	86 (46)	n.s.
TT	82 (35)	75 (40)	n.s.
NK9 (rs2246809)
AA	2 (1)	0 (0)	n.s.
AG	35 (15)	30 (16)	n.s.
GG	199 (84)	157 (84)	n.s.
NK10 (rs2617169)
AA	202 (85·5)	158 (84·5)	n.s.
TA	33 (14)	28 (15)	n.s.
TT	1 (0·5)	1 (0·5)	n.s.
NK11 (rs2617170)
CC	85 (36)	75 (40)	n.s.
CT	100 (42)	84 (45)	n.s.
TT	51 (22)	28 (15)	n.s.
NK12 (rs2617171)
CC	53 (22)	28 (15)	0·05	0·07	1·64 (0·97–2·84)
CG	101 (43)	86 (46)	n.s.
GG	82 (35)	73 (39)	n.s.
NK17 (rs1983526)
CC	90 (38)	54 (29)	0·05	0·06	1·52 (0·99–2·34)
CG	102 (43)	98 (52)	0·06	0·08	0·69 (0·46–1·04)
GG	44 (19)	35 (19)	n.s.

Frequencies were determined by χ² test. RA = rheumatoid arthritis; HC = healthy controls; Pc = Yates' corrected P; OR = odds ratio; 95% CI = 95% confidence interval; n.s. = not significant. Pc values < 0·05 were considered significant. Values of Pc and OR mentioned only in those cases where Pc was statistically significant or approaching statistical significance.

Association of NK receptor polymorphism with clinical phenotypes

Analysis of the association of the NK receptor polymorphisms with clinical characteristics (namely age at onset, presence or absence of deformities, extra‐articular manifestations and autoantibody status) was performed. We observed that the frequency of GG genotype and G allele of NK9 (rs2246809) (91 versus 80%, Pc = 0·04, OR = 0·41, 95% CI = 0·16–0·96 and 95 versus 89%, Pc = 0·03, OR = 0·41, 95% CI = 0·17–0·91, respectively) and AA genotype and A allele of NK10 (rs2617169) (92 versus 81%, Pc = 0·04, OR = 0·40, 95% CI = 0·15–0·97 and 96 versus 90% Pc = 0·04, OR = 0·41, 95% CI = 0·16–0·95, respectively) was significantly higher in individuals without deformities compared to those with deformities. We observed that the frequencies of A allele of NK9 (rs2246809) and the T allele of NK10 (rs2617169), respectively, were significantly higher in patients with deformities (11 versus 5%, Pc = 0·03, OR = 2·44, 95% CI = 1·09–5·98 and 10 versus 4%, Pc = 0·04, OR= 2·45, 95% CI = 1·05–6·39). A similar trend of association was observed in female patients and those with YORA having deformities (Table 3).

Table 3.

Association of NK receptor polymorphism with deformities

	Total RA		Pc	OR (95% CI)	Female RA		Pc	OR (95% CI)	YORA		Pc	OR (95% CI)
Deformities	+	–			+	–			+	–
(%)	(%)	(%)			(%)	(%)			(%)	(%)
NK9
A	11	5	0·03	2·44 (1·09–5·98)	11	5	0·05	2·26 (1·01–5·57)	10	4	0·04	2·50 (1·07–6·51)
G	89	95	0·03	0·41 (0·17–0·91)	89	95	0·05	0·44 (0·18– 0·99)	90	96	0·04	0·40 (0·15–0·94)
AA	1	0	n.s.		1	0	n.s.		1	0	n.s.
AG	19	9	n.s.		20	11	n.s.		18	9	n.s.	2·21 (0·89– 5·99)
GG	80	91	0·04	0·41 (0·16–0·96)	79	89	0·07	0·45 (0·17–1·05)	81	91	0·05	0·41 (0·15–0·99)
NK10
A	90	96	0·04	0·41 (0·16–0·95)	90	95	0·06	0·44 (0·17–1·03)	90	96	0·04	0·38 (0·14–0·93)
T	10	4	0·04	2·45 (1·05–6·39)	10	5	0·04	2·28 (0·97–5·95)	10	4	0·04	2·64 (1·08–7·38)
AA	81	92	0·04	0·40 (0·15–0·97)	80	90	0·07	0·43 (0·16–1·05)	82	92	0·04	0·37 (0·13–0·95)
TA	18	8	0·06	2·39 (0·98–6·42)	19	10	0·09	2·22 (0·91–5·99)	18	8	0·06	2·56 (0·997–7·37)
TT	1	0	n.s.		1	0	n.s.		1	0	n.s.

RA = rheumatoid arthritis; YORA = young onset RA; Pc = Yates' corrected P‐value; OR = odds ratio; 95% CI = 95% confidence interval. Pc values < 0·05 were considered significant.

Association of NK receptor polymorphism with treatment response

Polymorphisms in the NK receptor complex were not found to be associated with response to treatment with methotrexate‐based synthetic DMARD therapy.

Association of NK receptor haplotypes with RA susceptibility

Linkage analysis was performed for the eight bi‐allelic SNPs of various genes in the NK cell receptor complex, including NKG2D3, NKG2D4, NKG2D7, NKG2D9, NKG2D10, NKG2D11, NKG2D12 and NKG2A17 on chromosome 12. Haplotypes were constructed for 236 patients and 187 controls. We observed that the frequency of haplotype G‐C‐A‐G‐A‐T‐C‐C was higher in patients than in controls (12 versus 8%, P = 0·04, OR = 1·61, 95% CI = 1·01–2·55) (Fig. 1, Table 4), suggesting that the presence of this haplotype predisposed to the development of RA.

Figure 1.

Linkage disequilibrium of natural killer (NK) cell receptor polymorphisms. Haplotypes of various single nucleotide polymorphisms (SNPs) in NK cell receptors were constructed using the Haploview software version 4.2. Colour scheme of the linkage disequilibrium map is based on the standard D'/LOD option in the software. Dark squares indicate high r ² and bright squares indicate low r ² values. Values in squares are D' between single markers. [Colour figure can be viewed at wileyonlinelibrary.com]

Table 4.

Frequency of natural killer (NK) receptor haplotypes

1‐2‐3‐4‐5‐6‐7‐8	RA, n = 472 (%)	HC, n = 374 (%)	P‐value	OR (95% CI)
G‐C‐A‐G‐A‐T‐C‐C	58 (12)	30 (8)	0·04	1·61(1·01‐2·55)
G‐C‐T‐G‐A‐T‐G‐C	32 (6·8)	16 (4·3)	n.s.
G‐C‐T‐G‐A‐T‐G‐G	42 (8·9)	46 (12·3)	n.s.
C‐C‐A‐G‐A‐C‐C‐C	78 (16·5)	56 (15)	n.s.
C‐C‐T‐G‐A‐C‐G‐C	44 (9·3)	39 (10·4)	n.s.
C‐C‐T‐G‐A‐C‐G‐G	113 (24)	105 (28)	n.s.

1‐rs1049174, 2‐rs2255336, 3‐rs2617160, 4‐rs2246809, 5‐rs2617169, 6‐rs2617170, 7‐rs2617171, 8‐rs1933526. RA = rheumatoid arthritis; HC = healthy control; OR = odds ratio; 95% CI = 95% confidence interval. P‐values < 0·05 are considered significant.

Discussion

NKG2D dysregulation and abnormal MIC expression can cause autoreactive T cell stimulation and promote pathogenic process in RA 15. It is likely that polymorphisms in NK cell receptors such as NKG2D may play a major role in modulating NK cell activation. To address this issue, eight different SNPs in the NK gene complex (rs1049174, rs2255336, rs2617160, rs2246809, rs2617169, rs2617170, rs2617171 and rs1983526) were genotyped in RA patients and healthy controls. We observed that the genotype/allele frequency did not differ between patients and controls for all the eight polymorphisms analysed. However, the G‐C‐A‐G‐A‐T‐C‐C haplotype was more frequent in patients with RA than in healthy controls, suggesting that NKD2D may play a role in predisposing to development of RA.

In our study, we observed that the frequency of allele A (rs2246809) and T (rs2617169) was significantly higher in patients with deformities, while that of the other alleles G (rs2246809) and A (rs2617169) was higher in patients without deformities. A similar trend of association was observed with the deforming disease in female patients and in those with young age of onset. However, NKG2D polymorphisms did not appear to influence the phenotype related to RF and ACPA autoantibodies.

Park et al., from Korea, reported that the NKG2D (rs225536) thr/thr genotype conferred risk of developing RA. They also reported that the increased risk was noted in individuals with NKG2C (rs1141715) ser/ser and NKG2D (rs225536) thr/thr genotypes and that the presence of these two polymorphisms together increase RA risk by 12‐fold 18. No association was noted with other variants of NKG2D (rs1049174 and rs1049172). Contrary to their report that NK cell receptor polymorphism did not affect the antibody production and joint erosions, we observed that NKG2D polymorphisms were associated with the development of deformities in South Indian Tamil patients. The differences in our findings could be explained by the larger number of SNPs studied by us in a different population. Therefore, it is quite possible that different polymorphisms of NKG2D influence different disease characteristics in RA.

Our findings compare favourably with studies in other autoimmune disorders, such as diabetes mellitus, paediatric inflammatory bowel disease (IBD) and systemic lupus erythematosus (SLE). In these diseases, NKG2D has been reported to play a major role in disease severity and progression by altering the distribution and signal strength of receptors 23, 24, 25, 26. Non‐depleting anti‐NKG2D treatment during the prediabetic stage of non‐obese diabetic (NOD) mice prevented the development of diabetes by impairing the expansion and function of autoreactive CD8⁺ T cells, suggesting that NKG2D is essential for disease progression 23. In paediatric IBD, prophylactic anti‐NKG2D therapy attenuated the development of colitis. It also reduced severity when administered to animals with mild but not severe colitis 24, 25. In the case of systemic autoimmune diseases such as SLE, the NKG2D rs2255336 (NKC‐4) ala/ala genotype was found to have a high prevalence in German and Spanish patients 26, while the Thr allele (rs2255336) was reported to be protective in SLE patients from Poland 27. Kabalak et al. hypothesized that this SNP does not affect the cell surface expression of NKG2D but changes the binding affinity to its adapter DAP10, thus altering the signal strength. In the presence of the ala/ala genotype of NKC‐4, activation of CD4⁺CD28^–NKG2D⁺ cells is reduced, leading to severe disease in SLE 26. Similar mechanisms may be operative in RA, responsible for early onset of the disease and a severe phenotype with deformities reported by us.

Hayashi et al. reported the presence of two different haplotype blocks in the NK gene complex. Within these two blocks, allele combinations had either weak cytotoxic activity or high NK activity, which predisposed these patients to develop malignancy differentially 21. In our study, we identified the G‐C‐A‐G‐A‐T‐C‐C haplotype as a possible risk haplotype for RA. The alleles of this haplotype were found to be associated with high NK cell activity (HNK) in the study referred to above. Assessment of the degree of cytotoxic activity exhibited by this haplotype in patients with rheumatoid arthritis could form the basis of future investigation.

To conclude, our study shows that G‐C‐A‐G‐A‐T‐C‐C haplotype of the NKG2D gene may predispose to RA. In addition, NK cell receptor polymorphisms may also play a role in influencing disease expression, with certain variants causing severe deforming disease. Given the reported abundance of TNF and IL‐15 in RA synovium and their ability to activate autoreactive autologous T cells in an NKG2D‐dependent manner, it is plausible that polymorphisms in MICA and NKG2D influence the threshold of effector cell activation in autoimmune disorders such as RA, leading to a poor outcome of the disease 16. Further functional studies are required to understand the underlying cellular interactions/mechanisms influenced by NK receptor polymorphisms in modifying the phenotype of RA.

Key messages

• Individual NKG2D receptor polymorphisms are not risk factors for RA.

• The G‐C‐A‐G‐A‐T‐C‐C haplotype of NKG2D increases the risk for RA.

• Polymorphisms in NKC‐9 and NKC‐10 are risk modifiers, and may be responsible for a severe disease phenotype of RA.

Author contributions

V. S. N., R. T., R. K. and D. C. contributed to the conception and design of the work. V. S. N., D. P. M. and V. K. J. contributed to acquisition of clinical data. Laboratory data were generated by C. M. M.; C. M. M., R. T., R. K., D. C., D. P. M., V. K. J. and V. S. N. contributed to analysis and interpretation of the data, C. M. M., D. P. M. and V. K. J. drafted the paper and V. S. N., R. T., R. K. and D. C. critically revised the paper for important intellectual content. C. M. M., R. T., R. K., D. C., D. P. M., V. K. J. and V. S. N. gave approval of the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Disclosures

The authors declare they have no disclosures.