The Relationship Between Human Leukocyte Antigen Class II Genes and Type 1 Diabetes, Autoimmune Thyroid Diseases, and Autoimmune Polyendocrine Syndrome Type 3

Common genetic risk factors have been associated with type 1 diabetes (T1D) and autoimmune thyroid diseases (AITD). Graves’ disease (GD) and Hashimoto’s thyroiditis (HT) are typical AITD. T1D and AITD are major components of autoimmune polyendocrine syndrome (APS)-2 and/or APS-3. The human leukocyte antigen (HLA) has been extensively studied in these diseases [1]. However, population studies have shown that HLA associations may vary depending on the ethnic origin [2]. In Caucasian populations, the highest-risk HLA haplotype for T1D is DRB1*03:01-DQA1*05-DQB1*02 (DR3) and/or DRB1*04-DQA1*03:01-DQB1*03:02 (DR4) [2, 3], and the corresponding haplotype for AITD is DR3 [4, 5]. DRB1*15-DQB1*06:02 and DRB1*07:01-DQA1*02:01 (DR7) haplotypes confer strong protection against both T1D [2, 3] and AITD [6, 7]. However, in the Japanese population, the DR3 haplotype is absent, and the DR4 and DR7 haplotypes are rare [8–10], which may be more helpful for examining the susceptibility and resistance to T1D and AITD of HLA DR-DQ haplotypes, with the exception of DR3, DR4, and DR7.

In this chapter, we will review HLA class II genes that confer susceptibility and resistance to T1D and AITD, and discuss the relationship between HLA class II genes and T1D, AITD, and APS-3. Furthermore, we focus on amino acids at position 74 of the HLA-DRβ1 chain, position 52 of the HLA-DQα1 chain, and position 57 of the HLA-DQβ1 chain as key factors involved in susceptibility and resistance to T1D and AITD, and we discuss key amino acids and their involvement in susceptibility and resistance to T1D and AITD.

Nomenclature
In 1980, Neufeld and Blizzard suggested a classification of APS based on clinical criteria alone, and described four main types [30]. Of the four types, APS-2 and APS-3 are mainly associated with AITD and/or T1D. APS-2 is characterized by Addison's disease (AD) associated with AITD and/or T1D. APS-2 is quite rare with an incidence of 1.4-4.5 cases for every 100,000 individuals [31,32]. While all patients with APS-2 have AD [30,[32][33][34][35], AITD and T1D are reported to occur in 69-82% and 30-52% of patients with APS-2, respectively [30,34,35]. APS-3 has been defined as an association between a clinical entity of AITD and an additional autoimmune disease such as T1D (Type 3A), chronic atrophic gastritis, pernicious anemia (Type 3B), vitiligo, alopecia, myasthenia gravis (Type 3C). AD and/or chronic hypoparathyroidism were categorically excluded from APS-3 [30]. Although AITD consists of HT, idiopathic myxedema, asymptomatic thyroiditis, GD, endocrine ophthalmopathy, and pretibial myxedema, GD or HT comprise the majority of AITD. Thus, in discussing the relationship between HLA and T1D and/or AITD, it is necessary to focus on APS type 3A (APS-3A) rather than APS-2 or APS type 3B/3C, and GD or HT may be considered as AITD.
In Caucasian populations, including those in Northern Europe, the incidence rates of T1D are high, in excess of 30 cases/100,000 individuals per year. In contrast, the Japanese population has one of the lowest incidence rate of T1D in the world, at 1.6 cases/100,000 individuals per year, suggesting that the Japanese population may either lack an important susceptibility gene or have a unique T1D protective gene [36,37]. However, AITD is the most frequent autoimmune disease in the population, present in approximately 7-8% of the general population [38,39]. When thyroid disease is caused by environmental factors such as levels of iodine, incidence rates have been found to vary between locations and over time [40][41][42][43]. Studies regarding the incidence rates of AITD have come from a limited range of geographical areas. Therefore, it is difficult to comment on the absence or presence of variances in incidence rates of AITD between different geographical locations. Coexistence of T1D and AITD is common, with 15 to 30% of T1D subjects having AITD [44][45][46], whereas the prevalence of glutamic acid decarboxylase antibodies (GADAb) in AITD patients is around 5% [47,48]. There is a need to distinguish T1D with AITD (T1D+AITD, APS-3A) from T1D without AITD (T1D-AITD). Conversely, we may not need to distinguish AITD with T1D from AITD without T1D (AITD-T1D). Abbreviations are listed in Table 1.

T1D
HLA class II genes are closely related to the onset of T1D in all ethnic groups. Recently, Thomson et al. investigated whether HLA DR-DQ haplotypes and genotypes show the same relative predispositional effects across populations and ethnic groups using data from 38 studies worldwide [49]. They introduced a new static, the patient/control (P/C) ratio of haplotype or genotype frequencies within a study that allows comparison of absolute penetrance values within and across studies. Mean P/C ratios are listed in Table 2. When the mean P/C ratio is >1.10, we consider that the haplotype confers susceptibility to T1D, whereas when the mean P/C ratio is <0.90, we consider that the haplotype confers protection against T1D. When the mean P/C ratio is 0.90-1.10, we consider the haplotype as neutral to T1D.

AITD
The results of HLA association studies in AITD have been less consistent than in T1D. Moreover, data on HLA haplotypes in HT have been less definitive than on those in GD. A general methodological problem has been disease definition [50]; though the diagnosis of GD may be relatively straightforward, the definition of HT has been more controversial. Three varieties of thyroid autoantibodies are commonly used and widely available in clinical diagnostic laboratories: anti-thyroglobulin antibodies (TgAb), anti-thyroid peroxidase autoantibodies (TPOAb), and antibodies to thyrotropin receptor (TRAb). TgAb and TPOAb are found in almost 100% of patients with HT, whereas these antibodies are also detectable in 50% to 90% of patients with GD and are common in the general population. The low levels of TPOAb and TgAb found in many individuals are of uncertain significance in the presence of normal thyroid function [51]. Table 3 shows previous reports on the relationship between HLA class II and AITD. The most probable HLA-DR and -DQ haplotypes were deduced from linkage disequilibria [8][9][10]. Alleles in parentheses following the reference number indicate that the reference reported susceptibility or resistance of the allele, but not the haplotype, to the disease. There is no parenthesis following the reference number if the references reported susceptibility or resistance with 4-digit DRB1-DQB1, DQA1-DQB1, or DRB1-DQA1-DQB1 haplotypes. In cases with more than 2 haplotypes sharing the same allele, the allele is listed redundantly in each haplotype. However, considering the ethnicities that the references examined, the allele is removed from the corresponding haplotypes; for example, the DRB1*08:02-DQA1*03:01-DQB1*03:02 haplotype is rare in Caucasian populations [49] and thus in the reference examining Caucasian populations, the DQA1*03:01 allele is listed only in the DRB1*04:01-DQA1*03:01-DQB1*03:02 haplotype, and not in the DRB1*08:02-DQA1*03:01-DQB1*03:02 haplotype.   [54]. Further investigations in other ethnic groups may be necessary to confirm whether their conclusions are widely applicable.

T1D-AITD and T1D+AITD
Few previous reports have been published on the relationship between HLA class II and T1D-AITD. In contrast, there are a number of reports concerning the relationship between HLA class II and T1D+AITD, which includes T1D+GD and T1D+HT. The results are shown in Table  4. As in Table 3, alleles in parentheses following the reference number indicate that the reference reported susceptibility or resistance of the allele, but not the haplotype, to the disease. There is no parenthesis following the reference number if the references reported the susceptibility or resistance of 4-digit DRB1-DQB1, DQA1-DQB1, or DRB1-DQA1-DQB1 haplotypes to the disease. In cases with more than 2 haplotypes sharing the same allele, the allele is listed redundantly in each haplotype. However, with consideration of the ethnicities that the references examined, the allele may be removed from the corresponding haplotypes.  T1D+GD   Table 4. Effects of HLA DR-DQ genes on T1D-AITD and T1D+AITD

Relationship between GD and amino acid
Badenhoop et al. demonstrated that Arg at position 52 of the DQα1 chain plays an important role in susceptibility to GD [27]. It was recently shown that Arg at position 74 of the DRβ1 chain is important for the development of GD in a significant number of patients [16,17]. Further analysis has shown that the presence of Gln at position 74 of the DRβ1 chain was protective for GD [16]. Table 5 shows the susceptibility and resistance of HLA DR-DQ genes to GD, and amino acids at position 74 of the DRβ1 chain and position 52 of the DQα1 chain. When more than 2 references reported susceptibility, we considered that the haplotype confers susceptibility to GD (abbreviated as "S"). When more than 2 references reported protection against the disease, we considered that the haplotype confers protection against GD (abbreviated as "P"). When only one reference reported susceptibility, we considered that the haplotype either confers susceptibility or is neutral to GD (abbreviated as "S/N"). When only one reference reported a protective effect, we considered that the haplotype either confers protection against or is neutral to GD (abbreviated as "P/N"). Badenhoop et al. showed that susceptibility to GD is conferred by the DQA1*05:01 allele as well as Arg at position 52 of the DQα1 chain [27]. DRβ-Arg-74 and DRβ-Gln-74 are always present on DR3 and DR7, respectively [16]. These amino acids are indicated in bold. The amino acids at position 52 of the DQα1 chain that are encoded by the haplotypes listed in Table 5 are Arg, Gln, and Ser. The effect on GD of the haplotypes which encode Arg or Ser at position 52 of the DQα1 chain varies from susceptible to protective. Amino acids at position 74 of the DRβ1 chain that are encoded by the haplotypes listed in Table 5 are Ala, Arg, Gln, and Leu. The effect on GD of these haplotypes also varies from susceptible to protective. However, haplotypes that encode

Relationship between T1D±AITD and amino acid
It is well known that DQα-Arg-52 confer susceptibility to T1D [15]. Todd et al. demonstrated that DQβ-Asp-57 is neutral or negatively associated with T1D, and that Ala, Val, or Ser at position 57 of the DQβ1 chain is positively associated with T1D [13]. Table 6 Table 6, areas of the effect on T1D are shaded in the haplotypes that conflict with the theory that DQα-Arg-52 or "non-Asp" at position 57 of the DQβ1 chain confers susceptibility to T1D, and that DQβ-Asp-57 confers protection against T1D. Table 6 also shows the effects of HLA DR-DQ genes on AITD, T1D-AITD, and T1D+AITD. When more than 2 references reported susceptibility to the disease, we considered that the haplotype confers susceptibility (abbreviated as "S"), regardless of a single report demonstrating that the haplotype confers protection against the disease. When more than 2 references reported protection against the disease, we considered that the haplotype confers protection (abbreviated as "P"), regardless of one report demonstrating to the disease. When only one reference reported susceptibility, we considered that the haplotype confers susceptibility or is neutral (abbreviated as "S/N"). When only one reference reported protection against the disease, we considered that the haplotype confers protection or is neutral (abbreviated as "P/ N"). Recently, Menconi et al. demonstrated that amino acids at position 74 of the DRβ1 chain play an important role in susceptibility and resistance to APS-3A, i.e., T1D+AITD as well as GD [18]. DRβ-Tyr-26, DRβ-Leu-67, DRβ-Lys-71, and DRβ-Arg-74 are positively associated with APS-3A, while DRβ-Ala-71 and DRβ-Gln-74 are negatively associated with APS-3A. These amino acids are indicated in bold in Table 6.  While these haplotypes encode Val at position 57 of the DQβ1 chain, they confer protection or are neutral to T1D. Although they confer protection against AITD and T1D+AITD, they tend to confer susceptibility to T1D-AITD (S/N in Table 6). Since 15 to 30% of subjects with T1D have AITD [44][45][46], the effect of AITD on T1D may result in resistance of subjects with these haplotypes to T1D.  These haplotypes are the major haplotypes which confer susceptibility to T1D in East Asians, especially in the Japanese population where the DR3 haplotype is absent and the DR4 haplotype is rare [8][9][10]. While these haplotypes encode Asp at position 57 of the DQβ1 chain, they confer susceptibility to T1D+AITD. DRβ-Leu-67 in the DRB1*04:05-DQA1*03:03-DQB1*04:01 haplotype and DRβ-Tyr-26 in the DRB1*09:01-DQA1*03:02-DQB1*03:03 haplotype might play an important role in susceptibility to T1D+AITD. Since 15 to 30% of subjects with T1D have AITD [44][45][46], the effect of these amino acids on T1D may be susceptibility, and that on T1D-AITD might be susceptibility or neutrality, which is weaker than that on T1D or T1D+AITD. There are few reports concerning the effect of HLA DR-DQ genes on T1D-AITD in the Caucasian [57], Japanese [59,61,62], and Taiwan Chinese [60] populations ( Table 4). The DRB1*07:01-DQA1*02:01-DQB1*02:02 and DRB1*16:02-DQA1*01:02-DQB1*05:02 haplotypes are rare in the Japanese population [8][9][10]49]. Therefore, it is difficult to explain the protective or neutral effect of these haplotypes with "non-Asp" at position 57 of the DQβ1 chain on T1D by examining the effect of these haplotypes on T1D-AITD. However, Menconi et al. demonstrated that DRβ-Gln-74 is negatively associated with T1D+AITD, although they failed to demonstrate that the DR7 allele, which encodes Gln at position 74 of the DRβ1 chain, confers protection against T1D+AITD [18] ( Table 6). The DR3 and DR4 haplotypes encode Ala at position 57 of the DQβ1 chain, which confers strong susceptibility to T1D [2]. Since the DR7 haplotype also encodes Ala at position 57 of the DQβ1 chain, the effect of this haplotype might potentially result in susceptibility to T1D-AITD. Since 15 to 30% of subjects with T1D have AITD [44][45][46], DRβ-Gln-74 might play a role in protection against T1D.

Conclusion
T1D and AITD share common genetic risk factors. The prevalence of given HLA haplotypes varies among populations, but given the same DR and DQ haplotypes, the influence of HLA on T1D and/or AITD is similar on populations throughout the world. By clarifying the region of the diseases on which certain reports were focused, we can explain to some extent and speculate on the relationship between HLA haplotypes, specific amino acids, and T1D and/or AITD.