Association of Killer Cell Immunoglobulin-Like Receptors and Their HLA-Ligands with Type 1 Diabetes Among South Indian Population

ABSTRACT Background Type 1 diabetes (T1D) is a multifactorial autoimmune disease, involving strong genetic components with familial predisposition. Killer cell immunoglobulin-like receptors (KIRs) found on the surface of NK cells have ligands of human leukocyte antigen (HLA) class I that are associated with T1D. The present study evaluates the influence of KIR genes and their HLA-ligands in the aetiology of T1D among the South Indian population. Methods A total of 125 T1D patients, along with their parents (n = 126) and siblings (n = 52) were recruited. PCR-based genotyping was performed for KIR genes and HLA class I ligands. The gene frequencies were compared between patients and siblings/parents. Linkage-disequilibrium (LD) analysis was performed to assess the genetic association between KIR gene pairs. Results The results show significant differences in HLA-ligands of KIR genes between patients and parents. The HLA-C1C1 homozygosity was found to be a predisposing risk factor, and HLA-C1C2 heterozygosity was protective towards T1D along with either the activating KIR2DS2 or inhibitory KIRs 2DL1, 2DL2, 2DL3. However, the frequency of inhibitory KIR3DL1 significantly increased in the presence of HLA-B Bw4 Ile80 in parents when compared to patients showing a protective effect on T1D. Two pairs of KIR genes, 2DS4-3DL1 and 2DS1-2DL5, showed strong LD in patients, siblings and parents. Conclusion The KIR-HLA ligand combinations have a significant effect on T1D aetiology among the South Indian population. This study defines a pattern for family-based association studies with genotypic information about KIR genes and their HLA-ligands, providing the first evidence towards T1D among the South Indian population.


Introduction
Type 1 diabetes (T1D) is an organ-specific, multifactorial autoimmune disease characterized by the complete destruction of pancreatic β-cells leading to absolute insulin dependency. Genetic and environmental factors play crucial roles in the predisposition to T1D (Bluestone et al. 2010). The incidence of T1D is mostly observed in children and young adults, although it occurs at any age (Lammi et al. 2008). The Diabetes Atlas 2021 released by the International Diabetes Federation illustrates that India has the World's highest number of children and young adults with T1D (Brussels 2021). The underlying mechanisms that lead to the development of T1D still remain unexplained. Natural Killer (NK) cells play a vital role in modulating the innate immune response with their ability to produce cytokines and chemokines (Trinchieri 1989). Killer cell immunoglobulin-like receptors (KIRs) present on the NK cell surface, are the key receptors having a major role in regulating the function of NK cells (Lanier 1998). KIRs are also expressed on the subsets of T lymphocytes, especially on CD8+ cells (Parham 2005).
The KIR region is about 150 kb and lies on chromosome 19q13.4 within the lymphocyte receptor complex (Uhrberg et al. 1997;Wilson et al. 2000). Inhibitory KIRs with long cytoplasmic tails and activating KIRs with short cytoplasmic tails trigger inhibitory and activating signals, respectively, upon binding to the relevant ligands (Lanier 1998). The genes that encode KIRs are highly polymorphic, and there are 16 KIR genes, of which seven genes encode inhibitory KIRs, six genes encode activating KIRs, two are pseudogenes and KIR2DL4 has both inhibitory and activating functions (Parham 2005;Wilson et al. 2000). The KIR genes KIR2DL4, 3DL2, 3DL3 and 3DP1 are the framework genes and are universally present in all individuals (Wilson et al. 2000). Inter-individual variability in the KIR gene family depends upon the variation in gene number, particularly in a set of activating KIR genes and allelic polymorphism (McQueen and Parham 2002). NK cells have various functions based on different immune perspectives. The activity of NK cells is controlled by the balance between inhibitory and activating KIRs upon binding to specific human leukocyte antigen (HLA) class I ligands on target cells (Lanier 1998;Zhi et al. 2011). The presence of a varied number of activating and inhibitory KIR genes has been reported to be strongly associated with various inflammatory and autoimmune diseases (Enciso-Vargas et al. 2021;Kusnierczyk 2013). The present study could be evident on the involvement of KIRs and their HLA-ligands towards T1D susceptibility among the South Indian population.
Inhibitory and activating KIRs have signalling motifs, which are triggered upon binding to their corresponding HLA-ligands (motifs) such as HLA-A (A3/A11), HLA-B (Bw4 and Bw6) and HLA-C (C1 and C2) molecules (Vilches and Parham 2002). However, for several KIRs, the HLA-ligands are yet to be identified. The combination of KIRs and HLA-ligands act as important immunogenetic factors in human diseases. Increased activation of KIRs and decreased inhibitory KIRs with respective HLA-ligand interactions has been proposed to result in autoimmune diseases (Martin et al. 2002;Nelson et al. 2004). The presence or absence of the KIRs on NK cells, and their HLA-C ligands on target cells create varied immune responses that could be associated with T1D. Shastry et al. (2008) have reported that the absence of 2DL2 and HLA-C1 together with the absence of 2DS1 and 2DS2 is protective, whereas the presence of 2DL2 and HLA-C1 with the absence of 2DS1 and 2DS2 act as predisposing factors in T1D.
There are numerous genetic studies on the polygenic KIRs and multi-allelic HLA class I ligands in association with different autoimmune diseases like rheumatoid arthritis (McGeough et al. 2012), Graves' disease (Dastmalchi et al. 2014), systemic lupus erythematosus (Tozkir et al. 2016), multiple sclerosis (Jelcic et al. 2012), ulcerative colitis (Jones et al. 2006) including T1D (Mehers et al. 2011;Middleton et al. 2006;Sanjeevi et al. 2016;van der Slik et al. 2003;Zhi et al. 2011). In spite of several population studies, there is only one study that reports the association of KIR/HLA-ligands and T1D among the Indian population (Sanjeevi et al. parents. The allele and genotype frequencies were determined by direct counting method and compared between patient groups and siblings/parents by means of chisquare test, odds ratio (OR) and 95% confidence interval (CI) using Epi info v7 with Yate's correction. Haplotype frequencies and linkage disequilibrium (LD) analysis were computationally inferred in patient groups, siblings and parents using Haploview v4.2 (Barrett et al. 2005). LD analysis was performed to assess the genetic association between KIR gene pairs. A p-value of <0.05 was considered to be statistically significant.

Distribution of KIR genes and HLA-ligands
Genotyping of KIR genes in patients, siblings and parents revealed the presence of all the tested KIR genes. The distribution of KIR and HLA-ligand genes in patients, siblings and parents is tabulated in Table 1. The four framework genes and inhibitory KIR2DL1 were present in all the individuals. The KIR and HLA genotype frequencies of the studied population adhered to HWE (p > .05) among patients, siblings and parents except for HLA-A Bw4 in siblings (p = .031). The individuals carried six to eight inhibitory KIR genes and one to six activating KIR genes. Increased frequency was observed in siblings with six inhibitory KIR genes (OR = 0.16; p c = 0.0201) when compared to group C; however, decreased frequency was observed in siblings with seven inhibitory KIR genes when compared to patients (OR = 2.93; p c = 0.0071), especially, group A (OR = 3.96; p c = 0.0077) and group C (OR = 3.49; p c = 0.0191).
The KIR ligand, HLA-C2, was observed more frequently in parents when compared to patients (OR = 0.53; p c = 0.0195) particularly, group C (OR = 0.36; p c = 0.0224). Parents had an increased frequency of HLA-B Bw4 (OR = 0.5; p c = 0.0431), especially HLA-B Bw4 Ile80 (OR = 0.49; p c = 0.0403) compared to group B. However, HLA-A Bw4 frequency was higher in patients (OR = 1.87; p c = 0.0195) especially, group B (OR = 2.34; p c = 0.0111) when compared to parents. HLA-C1C1 homozygosity was higher in patients (OR = 1.87; p c = 0.0195) specifically, group C (OR = 2.8; p c = 0.0224) when compared to parents. Similarly, increased frequencies of the HLA-B Bw6 homozygosity (absence of HLA-B Bw4) were observed in patients, especially among group B (OR = 2.06; p c = 0.0331) when compared to parents. Meanwhile, HLA-C1C2 heterozygosity was higher in parents when compared to patient groups (OR = 0.41; p c = 0.0014).

KIR genotypes, haplotypes and LD analysis
A total of 38 different GIDs with distinct KIR gene content were observed in this study population (Table 2). Patients exhibited 28 GIDs, siblings with 16 GIDs and parents with 30 GIDs. Eleven GIDs were detected only in one individual, and the GID 5 was most frequent among the study population. The frequencies of GIDs 1, 5 and 6 were higher in siblings when compared to patients. The haplotypes carried by the individuals were determined, in which Bx haplotypes were more common than AA haplotypes among the study groups. The centromeric and telomeric haplotype frequencies of patients compared with siblings and parents are tabulated in Supplementary Table S1.  Table S2b). Logs of likelihood odds ratios (LODs) of KIR genes in patients, siblings and parents are given in Supplementary Table S2c. Two pairs of KIR genes 2DS4-3DL1 and 2DS1-2DL5 showed strong LD in patients, siblings and parents.   (12)

Stratification analysis of KIRs and HLA-ligands
The frequencies of KIR along with their HLA-ligand genes in patients, siblings and parents were analysed for their susceptibility and protection towards T1D. The KIR/HLA ligand combinations with most inhibition confer protection and most activation confer susceptibility towards T1D. Based on these findings, a hypothetical model was proposed to show the effect of KIR/HLA ligand combinations in the development of T1D (Figure 2). The HLA-C1C1 homozygosity showed susceptibility and the HLA-C1C2 heterozygosity showed protection with either the inhibitory KIRs or the activating KIRs (Table 3)   The KIR genes with and without their HLA ligands having significant frequencies in patients, siblings and parents are presented in Table 4. The frequency of activating KIR3DS1 in the absence of its ligand HLA-B Bw4 was higher in patients (OR = 2.86; p c = 0.0383), particularly, group B (OR = 3.21; p c = 0.036) than in siblings. In the absence of HLA-A Bw4, the frequency of KIR3DL1 was higher in siblings (OR = 0.39; p c = 0.0264) as well as in parents (OR = 0.37; p c = 0.005) when compared to group B. An increased frequency of KIR3DL1 was observed in the presence of HLA-B Bw4 among parents when compared to patients (OR = 0.53; p c = 0.0194), especially, group B (OR = 0.44; p c = 0.0145). More precisely, an increased frequency of 3DL1 with HLA-B Bw4 Ile80 was also observed in parents when compared to patients (OR = 0.49; p c = 0.0103), especially, groups A and B (Figure 2c).
KIRs with their HLA-C ligand combinations in patients, siblings and parents are shown in Supplementary Table S3. The presence of KIR2DL1 and 2DS1, in the absence of HLA-C2 ligand, showed a positive association with T1D when compared between patients and parents (OR = 1.88; p c = 0.0418). While analysing the combination of three inhibitory (KIR 2DL1, 2DL2, 2DL3) and two activating (KIR 2DS1, 2DS2) KIR genes along with their HLA-C ligands (HLA-C1, C2), the frequency of the presence of three inhibitory KIRs and one activating KIR (2DS2) with HLA-C1 and C2 was observed to be higher in siblings compared to patients (OR = 0.19: p c = 0.0319) (Figure 2f). In addition, the frequency of the combination of two inhibitory KIRs (2DL1, 2DL3) with HLA-C2, in the absence of KIR2DL2, two activating KIRs and HLA-C1 was higher in siblings than in patients (OR = 0.21; p c = 0.0255) (Figure 2e). However, the frequency of three inhibitory KIRs and two activating KIRs with HLA-C1 in the absence of HLA-C2 was higher in patients (OR = 2.48; p c = 0.027), specifically group C (OR = 3.64; p c = 0.0212) when compared to parents (Figure 2a).    (52) Patients (

Discussion
The NK cells have been implicated with the dual function of either protecting or enhancing autoimmunity by KIRs, influencing the pathogenesis of autoimmune diseases (Lanier 1998). KIR-HLA interactions are crucial in regulating the functions of NK cells that determine the fate of the target cell (Parham 2005). Studies have emphasized the effects of KIR and HLA-ligand interactions in various autoimmune diseases (Martin et al. 2002;Mehers et al. 2011;van der Slik et al. 2003). In this line, this is the first familial study to evaluate the diversity of KIR genes and HLA-ligands as well as their interactions towards T1D susceptibility among the South Indian population.
Results of this study show significant differences in the frequencies of HLA-ligands, but not in KIR genes among patients, siblings and parents. Similar results for KIR genes with no significant differences have been documented among Finnish  and Chinese Han (Zhi et al. 2011) populations. The presence of HLA-C1C2 heterozygosity shows protection towards T1D along with either the inhibitory KIRs 2DL1, 2DL2, 2DL3 or activating KIR2DS2. Likewise, Mehers et al. (2011) have reported a protective effect of HLA-C1C2 with 2DL1 or 2DL3 on T1D among the British population. In the same way, reports on HLA-C1C2 heterozygosity have been shown to be associated with T1D (Zhi et al. 2011) and multiple sclerosis (Garcia-Leon et al. 2011). Further, HLA-C1C2 heterozygosity with 2DL2 also shows a protective association in an autoimmune disease, Vogt-Koyanagi-Harada (Levinson et al. 2016). In contrast, the HLA-C1C1 homozygosity could be considered as a predisposing risk factor for the development of T1D along with either the activating KIRs 2DS1, 2DS2 or inhibitory KIRs 2DL1, 2DL2, 2DL3. This is in agreement with earlier studies reporting the susceptibility to T1D in HLA-C1C1 carriers with 2DL1 or 2DS1 (Mehers et al. 2011) and with KIR2DL3 (Osman et al. 2016), in T1D. In addition, Nelson et al. (2004) have reported that HLA-C1C1 with 2DS1 and/or 2DS2 favours susceptibility to psoriatic arthritis. With these observations, our results evidently conclude that the diversity of activating and inhibitory KIRs and their effect with respective HLA-C ligands contribute to either susceptibility or protection for the aetiology of T1D among the South Indian population.
Data of this study show a protective effect of HLA-B Bw4 with T1D. Similar reports have been evidenced in multiple sclerosis (Garcia-Leon et al. 2011;Lorentzen et al. 2009) and T1D (Osman et al. 2016). Increased frequency of HLA-B Bw4 with inhibitory KIR3DL1 has been suggested to have an important role in NK cell development (Anfossi et al. 2006). This is in line with the present study that shows a protective effect on T1D and has also been reported in multiple sclerosis (Garcia-Leon et al. 2011). In addition, HLA-B Bw6 homozygosity could be associated with susceptibility to T1D in our study, which has been previously reported among multiple sclerosis patients (Garcia-Leon et al. 2011).
Even though the data of the present study did not signify the association of individual KIR genes with the pathogenesis of T1D, the relative proportion of activating and inhibitory KIRs that determine the susceptibility or protection towards T1D has been investigated. Nevertheless, our results show a significantly increased number of inhibitory KIRs, which suggest a protective association with T1D in the South Indian population. A similar report towards the presence of more inhibitory KIRs among the Greek population further strengthens our findings (Varla-leftherioti et al. 2003). Moreover, in this study, a higher frequency of Bx haplogroup is observed among the studied South Indian population, which is in line with the previous report among the North Indian population (Rajalingam et al. 2002). The analysis of the family-based data in the present study provides the first evidence for an association between T1D and the KIR genes with strong LD. Several KIR gene pairs in the study population emphasize their strong linkage with each other. Augusto et al. (2012) have reported highly positive LD values for the KIR gene pairs with KIR2DL5 and other B haplotype activating genes (2DS3, 2DS5) in the Southern Brazilian population.
In this study, KIR2DS1 shows a distinct susceptibility to T1D in the absence of HLA-C2. Similar observation has been reported in psoriatic arthritis (Martin et al. 2002). Furthermore, the presence of activating KIR2DS2 combined with HLA-C1 in the absence of HLA-C2 observed in our study indicates an increased susceptibility to T1D among the studied population. This notion is supported by the observations made among the Dutch population with T1D (van der Slik et al. 2003). In addition, a protective association of inhibitory KIRs 2DL1, 2DL3 with HLA-C2 in the absence of KIRs 2DL2, 2DS1, 2DS2 and HLA-C1 has been reported by Shastry et al. (2008) in T1D patients, which is in line with the present study.
In conclusion, the correlation of the KIR-HLA ligand combinations among patients, siblings and parents has been established. Significant variations in KIR-HLA interactions are mostly found between patients and parents than with siblings among the South Indian population. KIR-HLA combinations show significant differences among T1D patients compared with siblings and parents, while no significance is observed in KIR gene frequencies. Thus, the combination of KIR with the respective HLA ligand is considered to be decisive in regulating the NK cell function and has a significant effect on T1D aetiology. Our findings indicate the complexity of KIR-HLA combination variability that determines the susceptibility or protection against T1D among the studied population. Finally, this study defines a pattern for family-based association studies with genotypic information of KIR genes and their HLA-ligands, providing first evidence towards the development of T1D among the South Indian population. Furthermore, a thorough understanding of the association between KIR-HLA genes and T1D will require a large family set, with siblings to uncover potentially higher statistical significance.