Human Immunology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

www.ashi-hla.org

journal homepage: www.elsevier .com/locate /humimm

The Raikas – A unique combination of high prevalence of type 1 diabetessusceptibility genes and near zero incidence of the disease

http://dx.doi.org/10.1016/j.humimm.2014.09.0090198-8859/� 2014 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.

⇑ Corresponding authors at: Department of Transplant Immunology and Immu-nogenetics, All India Institute of Medical Sciences, New Delhi 110029, India.

E-mail addresses: umakanga@hotmail.com (U. Kanga), narin98@hotmail.com(N.K. Mehra).

1 Contributed equally to this work.

Please cite this article in press as: Bhat DK et al. The Raikas – A unique combination of high prevalence of type 1 diabetes susceptibility genes and neincidence of the disease. Hum Immunol (2014), http://dx.doi.org/10.1016/j.humimm.2014.09.009

Deepali K. Bhat a,1, Uma Kanga a,⇑,1, Neeraj Kumar a, R.P. Agrawal b, Manish Mourya a, Mani Kalaivani c,Tanvir Kaur d, Narinder K. Mehra a,⇑a Department of Transplant Immunology and Immunogenetics, All India Institute of Medical Sciences, New Delhi, Indiab Department of Medicine, SP Medical College, Bikaner, Rajasthan, Indiac Department of Biostatistics, All India Institute of Medical Sciences, New Delhi, Indiad Indian Council of Medical Research, Ansari Nagar, New Delhi, India

a r t i c l e i n f o

Article history:Received 27 February 2014Accepted 25 September 2014Available online xxxx

Keywords:Raika communityCamel milkType1 diabetesHLAHaplotypes

a b s t r a c t

The Raikas, a camel rearing tribal group living in the Thar desert of Rajasthan has been reported with avery low incidence of diabetes. We analysed the frequency distribution of HLA alleles in this communityand compared the same with the non-Raika group living in the same geographic location and also that ofthe healthy North Indian (NI) population. The data revealed an exceptionally high phenotype frequencyof HLA-DRB1*03 in this community (53%) as compared to the non-Raika group (27.73%, p = 7.9E-05) andthe NI population (14.6%, p = 7.65E06). Further analysis revealed the occurrence of four major DRB1*03haplotypes in the Raikas: (i) A*26-B*08-DRB1*03 (AH8.2, 11.76%); (ii) A*24-B*08-DRB1*03 (AH8.3,8.82%); (iii) A*02-B*08-DRB1*03 (3.78%); (iv) A*01-B*08-DRB1*03 (AH8.1v, 0.84%); all of which occurredwith a several fold higher frequency in the Raikas than the other two groups. These haplotypes have beenreported to be positively associated with T1D in the NI population. The apparent lack of T1D and/or otherautoimmune diseases in the Raikas despite the higher occurrence of known disease associated HLAalleles/haplotypes is intriguing and highlights the quintessential role of the environmental factors, foodhabits and level of physical activity in the manifestation of T1D. Possible influence of other protectionconferring genes located on, as yet undefined chromosomal locations cannot be ruled out.� 2014 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights

reserved.

1. Introduction

Genes and environment have been intricately associated witheach other both in the occurrence and manifestation of a diseaseespecially if it is multifactorial and/or polygenic. Type 1 diabetes(T1D) is one such disease in which >40 genetic loci have beenimplicated [1], of which HLA complex plays the most importantrole with �50% genetic risk conferred by genes within its variousloci. The HLA Class II haplotypes of the major histocompatibilitycomplex (MHC) region particularly DR3-DQ2 and DR4-DQ8 havebeen reported to associate with T1D in most Caucasian populationsas >80% of the patients carry either of the two haplotypes or both[2,3]. In the Indian population, we have earlier shown a positive

association of multiple HLA-DR3-DQ2 haplotypes with T1D [4,5]as well as celiac disease [6] with almost no contribution of theDR4-DQ8 haplotypes. In addition to class II alleles, several HLAclass I alleles e.g., HLA-A*01, A*02, A*11, A*24, C*05, B*08, B*18,B*39 and B*50, have been reported to associate with T1D suscepti-bility either independent of HLA class II or due to their strong link-age disequilibrium with HLA-DRB1-DQA1-DQB1 haplotypes [7–10]. Association of HLA class I alleles, namely A*24, A*30 andB*18 with disease progression has also been reported [11,12].

The extended haplotype; A1-B8-DR3 often referred to as theancestral haplotypes AH 8.1 has been reported to be associatedwith several autoimmune diseases including T1D in the Caucasians[13–15]. However, in the Indian population, this haplotype occurswith a highly reduced frequency and has been replaced by a vari-ant form (AH8.1v) which is not disease associated. On the contrary,several other DR3 associated haplotypes occur in India of which atleast 3 confer high risk for T1D development. These include: (i)HLA-A*26-B*08-DRB1*03 (AH8.2) (ii) A*24-B*08-DRB1*03 (AH8.3)and (iii) A*02-B*08-DRB1*03 [4,5,16,17]. In Caucasians,

ar zero

Fig. 1. Geographic location of the state of Rajasthan and the Raika inhabiting region from where the samples were collected through onsite medical camps. Star (*) highlightsthe location of the city of Bikaner.

2 D.K. Bhat et al. / Human Immunology xxx (2014) xxx–xxx

DRB1*15:01-DQB1*06:02 haplotype has been reported to conferprotection towards T1D and has been shown to reduce the diseaserisk [18]. Therefore, it could easily be speculated that the T1D sta-tus of a population is a reflection of the distribution of T1D predis-posing and protection conferring HLA genes in that population. Forexample, the high incidence of T1D reported in Sardinia (Italy)with 38.9/100,000 cases per year is in consensus with the highestfrequency of predisposing HLA-DRB1*03 allele in that population[www.allelefrequencies.net]. Incidence of T1D varies greatlyamong various populations of the world being the lowest in China(0.1/100,000 cases per year) and highest in the Scandinavian coun-tries, particularly Finland where it stands 40/100,000 cases peryear [19,20]. In the latter populations, >85% of T1D patients haveeither HLA-DR3 or DR4 or both [3,21]. Although DR3 frequency islow in Finland, the hygiene hypothesis, i.e., decline of infectionsas a result of better hygiene is one of the major explanations forthe increased frequency of T1D in such countries [22,23]. Further,exposure to diabetes causing viruses, environmental and chemicalfactors and variations in vitamin D intake [24,25] have been sug-gested as additional factors to explain the increased occurrenceof T1D in industrialized countries.

We report here a closely bred community inhabiting the northwestern region of India in the state of Rajasthan (Fig. 1). It is a tri-bal, nomadic group living in the Thar desert of Rajasthan withcamel rearing as their major occupation. Members of this commu-nity can easily be marked by their lean healthy built apparentlyattributable to their physically active lifestyle and a staple dietcomprising of Bajra (Pennisetum glaucum) and camel milk. Epide-miological studies conducted on the Raikas have consistently high-lighted the very low to zero incidence of diabetes in this population[26]. Additionally, epidemiological surveys carried out by theEndocrinologist involved in present study (RPA) and other physi-cians of the local hospitals/clinics in the area indicate occasionalT1D cases observed in the Raika population during the past onedecade or more. Since HLA genes contribute significantly towardssusceptibility to T1D, we tested the notion that low or zero occur-

Please cite this article in press as: Bhat DK et al. The Raikas – A unique combinatincidence of the disease. Hum Immunol (2014), http://dx.doi.org/10.1016/j.hu

rence of T1D in the Raikas could be reflective of a similar low fre-quency of T1D susceptibility HLA genes in this population.Accordingly, the HLA profile of this community was studied andcompared with that of the Non-Raikas who also inhabit the samearea but do not belong to the Raika community. The two groups:Raika and non-Raika differ from each other in their life style andeating habits. For example, Raikas rear camels whereas very fewnon-Raikas do so. Additionally, Raika community is known to reg-ularly consume camel milk as a part of their staple diet whereasthe non-Raikas are not frequent camel milk drinkers, as observedduring filling the patient information sheet and the clinical pro-forma at the time of subject recruitment. Further, we comparedthe HLA profile of these two groups with a cohort of the NorthIndian population mainly belonging to Delhi and its neighbouringregions.

2. Material and methods

2.1. Sample collection

With an aim to obtain sporadic samples, onsite medical campswere organised in an area �50 km around the city of Bikaner, at sixdifferent villages namely Gadwala, Baderan, Tanwarwala, Sansarde-san, Ranasar and Rebaribass inhabiting the Raikas and distributedover the Thar desert. A total of 270 unrelated individuals (143 Rai-ka and 127 non-Raika) were recruited for the study after takingtheir informed consent and as per the approval of the institutionalethics committee. The DNA was extracted by the salting out proce-dure and only those samples with OD260/OD280 ratio between 1.7and 2.0 were used for the study. A total of 253 samples (134 Raikasand 119 non-Raikas) were finally utilized for the study.

2.2. Study subjects

The demographic data of 270 individuals belonging to the twocommunities included in the study are summarized in Table 1. A

ion of high prevalence of type 1 diabetes susceptibility genes and near zeromimm.2014.09.009

Table 1Demographic details of the two study groups. The mean BMI indicates the fitnessstatus of the individuals of both groups (BMI values ranging between 20 and 25 isconsidered healthy).

Feature Raika(N = 143)

Non Raika(N = 127)

p value

Sex distribution 0.02Male 82 55Female 61 72

Age rangeMean (±SD)

14–85 years44.1 (±17.3)

14–71 years42.3 (±14.8) 0.35

History of diabetes (No.)T1DT2D

2 (1.4)11

15 (11.8)114

<0.001

Camel milk drinkingN (%) 118 (82.5) 33 (25.9) <0.001

Fasting blood glucose (mg/dl)Mean (±SD) 93.02 (±22.96) 109.4 (±44.5) 0.002

HbA1c (%)Mean (±SD) 5.21 (±0.63) 5.62 (±1.27) 0.001

BMIMean (± SD) 20.95 (±4.57) 21.87 (±4.91) 0.11

Abbreviations: T1D, type 1 diabetes; T2D, type 2 diabetes.p value is mentioned in bold for features where difference is statistically significant.

D.K. Bhat et al. / Human Immunology xxx (2014) xxx–xxx 3

one to one questionnaire including questions on demographic andclinical details was asked from the study population. The partici-pants informed us of the community they belonged to (Raika ornon-Raika) and provided information on their physical activity,dietary habits, camel milk drinking profile and disease status inthe family including the children. The study questionnaire specifi-cally included questions regarding several autoimmune diseaseslike rheumatoid arthritis (history of joint pain/swelling), coeliacdisease (wheat sensitivity), vitiligo (skin pigmentation) and physi-cal examination of thyroid was recorded. Community heads wereinvolved whenever language difficulties were encountered. Outof the 143 Raika individuals, 82 (57.3%) were males and 61(42.7%) females with a male/female ratio of 3:2 while the non-Rai-ka study group comprised of more females than males (43%) with amale/female ratio of 2:3. Among Raikas, most women joined menfor camel rearing while the non-Raika women stayed home forhousehold activity; hence the difference in female representationin both communities. Both the study groups matched for the agewith a median value of 44 years for the Raikas and 42 years forthe Non-Raikas. The age range varied from 14 to 85 years for Raikasand 14 to 71 years for the non-Raikas. Children were reluctant topermit sample collection, therefore our study participants are allabove 14 years of age. Further, both groups had a similar BMI,HbA1C levels, but the mean fasting blood glucose level was slightlyhigher in the non-Raikas. ADA criteria was used for diabetes diag-nosis [27]. T1D diagnosis was ascertained by body habitus, ketosisprone status and age of the patient. Only two individuals with his-tory of diabetes were observed in the entire Raika community –one of these was diagnosed to have type1 diabetes and the othertype 2 diabetes. On the other hand, the Non-Raika group had onecase of T1D and fourteen cases of T2D.

We defined camel milk consumers if they gave a history ofdrinking at least 400–500 ml of milk daily for 5 days in a week.

Table 2Camel milk drinking profile of Raika and Non-Raika groups. Nearly three quarters ofRaikas consume camel milk in large quantities (71.2% drinking >400 ml).

Quantity of camel milk Raikas N = 118N (%)

Non Raikas N = 33N (%)

<200 ml 6 (5.1) 2 (6.1)200–400 ml 28 (23.7) 10 (30.3)400–500 ml 14 (11.9) 10 (30.3)>500 ml 70 (59.3) 11 (33.3)

Please cite this article in press as: Bhat DK et al. The Raikas – A unique combinatincidence of the disease. Hum Immunol (2014), http://dx.doi.org/10.1016/j.hum

Majority of the Raikas (82.5%) gave a history of camel milk con-sumption as compared to only 25.9% of non-Raikas (p < 0.001). Ofthe 33 non-Raika camel milk drinkers, only 11 (33%) fulfilled thecriteria of camel milk consumers. Details on the quantity of camelmilk consumption in the two study groups are given in Table 2.

2.3. HLA genotyping

HLA class I and II genes were investigated by the PolymeraseChain Reaction (PCR) (Biorad C1000, USA) of the genomic DNA fol-lowed by hybridization with sequence specific oligonucleotide(SSO) probes immobilized on nylon strips. The biotin labelledamplicons were chemically denatured and hybridized to an arrayof immobilized SSO probes. The amplicon–probe complex wasvisualized using a colorimetric reaction. A positive amplicon–probereaction resulted in the formation of a blue colour complex and thesamples hit pattern were recorded and interpreted as per the PMPsoftware for genotype analysis [Reverse Line Strip Assay kits (Invit-rogen, USA)].

The HLA typing by PCR-SSP was performed using sequence spe-cific primers to obtain high resolution information on select HLAclass II alleles viz. DRB1*03,*04,*13 and *15 (BAG Healthcare, Ger-many). The amplicons were visualized on 2% agarose gel electro-phoresis followed by the visualization with ethidium bromidestaining. Interpretation of gel electrophoresis was done as perthe hit table provided by the manufacturer.

2.4. Statistical analysis

Statistical analysis was performed using Statistical Package forSocial Sciences (SPSS) version 16.0 for windows. Statistical com-parisons between groups were performed using chi-square testwith yates correction. Haplotypes were constructed on the basisof previous data on the North Indian healthy population [28] anddiabetes cohort [17]. Data were presented as number (percentage)or mean (±SD) as appropriate. The demographic and clinical char-acteristics between Raikas and non-Raikas were compared usingchi-square test/Fisher’s exact test. Quantitative parameters werecompared between the groups using student’s ‘t’ test for indepen-dent samples. The comparison of camel milk consumptionbetween the groups was done using chi-square test for trend.Results were considered significant when p values were lower than0.05.

3. Results

3.1. HLA-DR/DQ alleles in Raikas and non-Raikas and their comparisonwith the healthy North Indian population

Considering the very low incidence of T1D in the Raika and thenon-Raika communities, we first studied the distribution of HLAClass II genes in the two population groups. Among the Raikagroup, HLA-DRB1*03 was the most frequent phenotype (53%) fol-lowed by DRB1*13 (33.58%) and DRB1*15 (27.61%) (Table 3). Thisis in contrast to the situation in non-Raikas in which the threealleles occurred with a frequency of 27.7%, 35.3% and 29.4%,respectively. High resolution analysis revealed that DRB1*03 wasrepresented by *03:01 in each individual while DRB1*15 was pre-dominantly *15:01 (80%) followed by *15:02 (16%). DRB1*04 highresolution analysis revealed only two alleles, *04:03 as the pre-dominant allele with 04:01 being rare. The predominantDRB1*13 allele was 13:01 followed by 13:02 and 13:04. Majorityof the DRB1*13 were in linkage with DQB1*06:03. All DR3+ve indi-viduals in both Raikas and non-Raikas carried DQB1*02:01 allele

ion of high prevalence of type 1 diabetes susceptibility genes and near zeroimm.2014.09.009

Table 3Percent phenotype frequency of HLA class II alleles (DRB1 and DQB1) in the Raika, Non-Raika and the North Indian populations.

HLA class II alleles (DRB1 and DQB1) Raika (R) (N = 134) Non-Raika (NR) (N = 119) North Indians (NI) (N = 335) R vs NR R vs NI NR vs NI(%PF) (%PF) (%PF) p value p value p value

DRB1*01 18.65 5.88 5.1 4.215E-03 7.655E-06 NSDRB1*03 53.00 27.73 14.6 7.920E-05 7.52E-11 2.264E-03DRB1*04 10.44 20.16 16.4 0.04 NS NSDRB1*07 7.46 10.08 26.3 NS 1.084E-05 4.143E-04DRB1*08 0.74 3.36 2.1 NS NS NSDRB1*09 1.50 0.84 0.9 NS NS NSDRB1*10 9.70 14.28 11.0 NS NS NSDRB1*11 4.47 15.96 17.3 4.435E-03 4.493E-04 NSDRB1*12 0.74 4.20 6.9 NS 0.01 NSDRB1*13 33.58 35.29 22.1 NS 0.01 6.634E-03DRB1*14 12.68 12.60 6.9 NS NS NSDRB1*15 27.61 29.41 25.5 NS 4.098E-03 0.01DRB1*16 1.50 0.0 0.7 NS NS NSDQB1*02 60.44 33.61 38.9 3.493E-05 5.155E-04 NSDQB1*03 18.65 39.49 53.5 4.20E-04 3.851E-09 0.03DQB1*04 0.74 5.0 0.7 0.09 NS NSDQB1*05 39.55 30.25 32.6 NS NS NSDQB1*06 48.50 58.0 51.4 NS NS NS

Abbreviations: R, Raika; NR, Non-Raika; NI, North Indian; (%PF), percentage phenotype frequency.Bold indicates the relevant HLA allele, its frequency and p value.

4 D.K. Bhat et al. / Human Immunology xxx (2014) xxx–xxx

and it occurred with a frequency of 60.4% in the Raikas as com-pared to 33.6% in the non-Raikas.

We next compared the distribution of the HLA class II allelesand their associated DQ alleles among the Raikas and non-Raikaswith the healthy North Indian (NI) population (Table 3). The datarevealed that the T1D predisposing HLA DRB1*03:01 occurred witha significantly high frequency in the Raikas as compared to boththe non-Raika (p = 7.9E-05) and the healthy North Indian popula-tion (p = 7.52E-11). On the other hand, the distribution ofDRB1*04 did not show significant differences between the Raikasand the North Indians. Interestingly, the frequencies of DRB1*15in both Raikas (27.6%) and non-Raikas (29.4%) were significantlyreduced as compared to the North Indian population (42.4%,p = 4.09E-03 and p = 0.01, respectively). However, the frequencyof HLA-DQB1*03 allele was significantly reduced in the Raikas(18.65%), as compared to that in the non-Raikas (39.49%,p = 4.2E-04) and North Indians (53.5%, p = 3.85E-09). Additionally

Table 4Percent phenotype frequency of HLA class I alleles (A and B) in the Raika and the Non-Ra

HLA allele Raika (R) (N = 119) Non-Raika (NR) (N = 103) North(%PF) (%PF) (%PF)

A*01 15.12 20.38 22A*02 36.13 29.12 28.9A*03 5.04 15.53 16.1A*11 21.84 36.90 27.8A*24 34.45 23.30 28.8A*26 26.05 12.62 9.6A*31 16 14.56 6.2A*32 6.72 6.8 6.4A*33 4.20 8.73 16.7B*07 5.88 11.65 10.8B*08 49.60 18.44 15.7B*15 0.84 9.70 8.5B*18 3.36 1.94 7.5B*35 13.44 35.92 22.7B*37 8.40 8.73 3.5B⁄39 8.40 1.94 1.6B*40 14.28 17.47 19.7B*50 5.04 0.97 5.5B*51 42.0 36.90 16.3B*52 7.56 7.76 17.0B*58 6.72 11.60 9.8

Abbreviations: R, Raika; NR, Non-Raika; NI, North Indian; (%PF), percentage phenotype frBold indicates the relevant HLA allele, its frequency and p value.

Please cite this article in press as: Bhat DK et al. The Raikas – A unique combinatincidence of the disease. Hum Immunol (2014), http://dx.doi.org/10.1016/j.hu

the frequency of DQB1*06 was increased in non-Raikas (58%) incomparison to North Indians (51.4%). No significant differenceswere observed for the frequency distribution of HLA-DQB1*04and DQB1*05 alleles among the three groups under study.

3.2. Distribution of HLA-A and B alleles

The HLA-A and B allelic distribution was studied in the 119 Rai-kas and 103 non-Raikas (Table 4). Distribution analysis revealedthat four alleles HLA-A*02 (36.13%), HLA-A*24 (34.45%), HLA-A*26 (26.05%) and HLA-A*11 (21.84%) were predominant. Of theseHLA-A*02 (29.12%) and A*11 (36.9%) were also frequently observedin non-Raikas, although HLA-A*26, was significantly reduced ascompared to the Raikas (12.6% vs 26.05%, p = 7.56E-07) and HLA-A*03 was significantly increased (p = 2.62E-03). HLA-B*08 overdominated the HLA-B allelic distribution of Raikas (49.6%), ascompared to its frequency of only 18.44% (p = 2.54E-06) in the

ika populations.

Indians (NI) (N = 706) R vs NR R vs NI NR vs NIp value

NS NS NSNS NS NS0.01 2.629E-03 NS0.02 NS NSNS NS NS0.01 7.569E-07 NSNS 4.445E-04 4.586E-03NS NS NSNS 6.595E-04 0.05NS NS NS2.541E-06 7.500E-11 NS6.402E-03 5.7090E-03 NSNS NS 0.051.701E-04 0.03 5.015E-03NS 0.02 0.02NS 4.675E-05 NSNS NS NSNS NS NSNS 2.685E-10 1.212E-06NS 0.01 0.02NS NS NS

equency.

ion of high prevalence of type 1 diabetes susceptibility genes and near zeromimm.2014.09.009

Table 5Distribution of known T1D susceptibility conferring HLA-A-B-DRB1*03 haplotypes in the three populations under study. Indian T1D predisposing haplotypes AH8.2 and AH8.3were observed in much higher frequencies as compared with the healthy North Indian population (7.4E-11, 7.5E-11, respectively). Note that AH8.2 occurred around 10 timeshigher in the Raikas.

Ancestral haplotype 3 locus haplotype Raika (R) (2n = 238) Non-Raika (NR) (2n = 206) North Indians (NI) (2n = 1412) p value p value p value

No. +ve % HF No. +ve %HF No. +ve %HF R vs NR R vs NI NR vs NI

AH 8.1v A*01-B*08-DRB1*03 2 0.84% 1 0.48% 5 0.35% NS NS NSAH 8.2 A*26-B*08-DRB1*03 28 11.76% 7 3.39% 21 1.48% 2.0E-03 7.4E-11 NSAH 8.3 A*24-B*08-DRB1*03 21 8.82% 4 1.94% 15 1.06% 3.3E-03 7.5E-11 NS

A*02-B*08-DRB1*03 9 3.78% 4 1.94% 1 0.07% NS 2.6E-10 1.2E-04Others Ax-B*08-DRB1*03 4 1.68% 3 0.5% 9 0.63% NS NS NSOthers Ax-Bx-DRB1*03 7 2.94% 9 4.36% 63 4.5% NS NS NS

Abbreviations: R, Raika; NR, Non-Raika; NI, North Indian; (%HF), percentage haplotype frequency.Bold indicates most relevant HLA haplotypes, the frequency and p value.

D.K. Bhat et al. / Human Immunology xxx (2014) xxx–xxx 5

non-Raikas. The other predominant HLA-B allele in the Raikas wasHLA-B*51 (42%). Both of these alleles comprised of more than halfof the HLA-B gene pool of the Raikas. However in non-Raikas, HLA-B*51 (36.9%), B*35 (35.92%), B*08 (18.44%) and B*40 (17.47) werethe predominant alleles.

3.3. Three locus haplotype analysis

Three locus HLA-A-B-DR3 haplotypes were defined in 119 Rai-kas and 103 non-Raikas resulting in 238 and 206 haplotypes inrespective groups. HLA-B*08 and DRB1*03 were observed to bethe major alleles in the Raikas and non-Raikas, hence the threelocus HLA-A-B*08-DRB1*03 haplotypes were analysed in both thegroups and compared with those in the North Indian population(Table 5). An appreciable heterogeneity was observed in the distri-bution of 3 locus DR3+ve haplotypes. Among Raikas, four majorhaplotypes; A*26-B*08-DRB1*03 (AH8.2, 11.76%), A*24-B*08-DRB1*03 (AH8.3, 8.82%), A*02-B*08-DRB1*03 (3.78%) and A*01-B*08-DRB1*03 (AH8.1, 0.84%) were observed in the decreasingorder. These represented about one quarter of the total haplotypesin Raikas and �8% in non-Raikas. Of these haplotypes, AH8.2 wasthe most common in both groups, although its frequency was 3times higher in the Raikas as compared to non-Raikas (11.76% vs3.39%, p = 2.0E-03) and North Indians (7.4E-11). The second mostcommon haplotype in the Raikas was A*24-B*08-DRB1*03 with ahaplotype frequency of 8.82% and was significantly increased ascompared to non-Raikas (1.94%, p = 3.3E-03) and North Indians(1.06%, p = 7.5E-11). HLA-A*02-B*08-DRB1*03 was comparable inRaikas and non-Raikas (3.78% and 1.94%, respectively), however

Fig. 2. Frequency distribution of HLA-DRB1*03 allele in various populations of the worldallele in Italy Sardinia is comparable with the high incidence of T1D (38.9) in this popu

Please cite this article in press as: Bhat DK et al. The Raikas – A unique combinatincidence of the disease. Hum Immunol (2014), http://dx.doi.org/10.1016/j.hum

these were significantly higher than those in the North Indians(0.07%, p = 2.6E-10 and 1.2E-04, respectively).

4. Discussion

The magnitude of the association of HLA-DRB1*03 with T1D canbe easily exemplified by the disease hot spot Sardinia (Italy) withan annual incidence of 38.9 per 100,000 and the highest DRB1*03allelic frequency of 55.7% in the world [17]. Fig. 2 shows the directproportionality between the DRB1*03 allelic frequencies and T1Dincidence [[19,20,29,30], www.allelefrequencies.net], indicating astrong association of this allele with the disease. We present herea unique population called the Raikas inhabiting the Thar desertof Rajasthan who despite a very high phenotype frequency ofDRB1*03 (53%) and DQB1*02 (60.4%), presents a near zero inci-dence of the T1D as evident in the previous survey and currentstudies conducted by the diabetologist (RPA) involved in the pres-ent study [26,31]. Hence the notion that low or zero occurrence ofT1D in the Raikas could be reflective of a similar low frequency ofT1D susceptibility HLA genes in this population does not hold true.The protective effect of other HLA alleles in lowering T1D incidencein this population cannot be ruled out, for example the relativelylow frequency of HLA-DRB1*04 which is strongly associated withT1D in Caucasians, absence of predisposing haplotypeDRB1*04:05-DQB1*03:02 and majority of DR4 allele beingDRB1*04:03 which is associated with diabetes protection. Furtherwe observed low frequency of HLA-DQ3 (DQ9) that is reported tobe associated with T1D in the Japanese and Chinese [32–34]. Alsothe increased occurrence of T1D negatively associated allele,

along with the known incidence of T1D in each of them. Note the frequency of thislation.

ion of high prevalence of type 1 diabetes susceptibility genes and near zeroimm.2014.09.009

6 D.K. Bhat et al. / Human Immunology xxx (2014) xxx–xxx

DRB1*13 (with predominant presence of haplotype DRB1*13:01-DQB1*06:03) in the Raikas could be an additional HLA related fac-tor. The negative association of DRB1*13:01-DQA1*01:03-DQB1*06:03 with T1D has recently been reported in populationfrom Turkey [35].

Diabetes related childhood mortality could be an additional fac-tor resulting in the very low prevalence of the disease in this pop-ulation. However, during field surveys/camps no case of childhooddiabetes was noticed or reported by the families. Higher intake ofVitamin D and exposure to infections early in life, both having pre-ventive and protective effect on autoimmunity could be additionalpossible reasons for the low incidence of diabetes in this tribalgroup. Although Vitamin D levels were not measured during pres-ent or earlier surveys, higher levels are expected as the commu-nity’s lifestyle of going to fields for camel rearing, exposes themto enough sunlight. Preventive effect of vitamin D on disease isbased on experimental model showing tolerogenic dendritic cellsinduced by vitamin D receptor ligands enhance regulatory T cellsthat inhibit autoimmune diseases [25].

This study was initiated primarily to define the population dis-tribution of HLA alleles in the two communities. High frequency ofT1D predisposing HLA-DRB1*03 phenotype (53%) and the nearabsence of disease in the Raikas intrigued us to study DR3associated haplotypes in this population. The analysis revealedan unexpectedly high frequency of the known T1D predisposingHLA-B*08-DRB1*03 haplotypes in the Raikas when compared tothe healthy North Indian population. We reported that T1D inthe North Indian populations is associated with multiple DR3+haplotypes that fall in three groups: B8-DR3, B50-DR3 and B58-DR3 [4,17]. Of these, the former is uniquely heterogeneous sharingmultiple extended haplotypes designated as AH8.2 (HLA-A*26-B*08-DRB1*03); AH8.3 (A*24-B*08-DRB1*03) and others[4,5,16,17]. This is in contrast to the Caucasian populations inwhich one single B8-DR3 haplotype (AH8.1) predominates and isdisease associated. In the Indian population, however, AH8.1 isunrelated with the disease and differs from the ancestral AH8.1at several loci between HLA-B and DR. It has therefore beenreferred as a variant AH8.1v haplotype. The two main DR3 haplo-types associated with T1D in North Indian population are AH8.2and AH 8.3 with a relative risk of 4.2 and 4.5, respectively [17].The overall distribution of DR3 haplotypes in the Raikas and nonRaikas was quite similar to that observed in North Indians withB8-DR3 being the predominant haplotype. Of the other two, B50-DR3 haplotype was observed only in the Raikas while B58-DR3was observed in both groups, albeit more frequent in non-Raikas.The B8-DR3 group, showed a similar level of heterogeneity in thispopulation as observed in the North Indian group. It is intriguinghowever, that despite this; the Raikas remain characterised bythe lack of T1D.

Being a tribal group with the ritual of marrying within theircommunity might explain the overrepresentation of HLA-DRB1*03 allele in the Raikas. However, despite the overrepresenta-tion of the T1D predisposing HLA-DRB1*03 allele and its associatedhaplotypes, near zero incidence of T1D was observed in this popu-lation. This could be attributed either to the multifactorial natureof T1D, influence of other T1D negatively associated HLA alleles/haplotypes or an overriding influence of environmental triggers.Several studies have highlighted the requirement of an externaltrigger ranging from dietary factors, physical built to viral infec-tions in the manifestation of the disease [36–38]. The indispens-able role of environmental triggers in T1D manifestation issupported by studies where the percent incidence of disease getsincreased when the population has migrated from a low incidencegeographic region to a high incidence region [39]. Hence involve-ment of environmental pressures in clinical T1D disease progres-sion highlights the requirement of an intricate nexus between

Please cite this article in press as: Bhat DK et al. The Raikas – A unique combinatincidence of the disease. Hum Immunol (2014), http://dx.doi.org/10.1016/j.hu

the predisposing genes and the environment. Several hypothesishave been put forward to explain the increase in diabetes inci-dence and these include, exposure to diabetes causing viruses,changes in exposure to some environmental or chemical factors,variations in vitamin D intake which has protective effect andthe role of gut microbiota that controls autoimmunity [24].

Life style and food habits comprise an integral part of studying atribal ethnic population like the Raikas. Belonging to the lowincome group, Raikas mainly survive on Bajra grown in their fieldsand camel milk. As shown in Tables 1 and 2, >80% of the Raikas arecamel milk drinkers with �24% drinking camel milk daily. Accord-ing to previous reports, the near zero diabetes status in Raikas isattributable to their regular camel milk drinking habit [26] and across over study comparing cow milk and camel milk in T2D sub-jects from Raika and non Raikas showed hypoglycaemic effect ofcamel milk by reducing insulin resistance [40]. The chemical anal-ysis of the camel milk has also confirmed some insulin like proper-ties accounting for its anti-diabetic role [41]. A recent report onstreptozotocin induced T1D mice model has revealed that camelwhey protein treatment improved healing and closure of diabeticwounds by enhancing expression of chemokines like MIP-1a,MIP-2 and CX3CL1 [42].

The anti-diabetic effect of camel milk has been proven in animalstudy suggesting that the effect was dependent on the quantity ofmilk and remained effective even after stopping its consumption[43]. The beauty about camel milk is that the lack of coagulum for-mation allows it to pass rapidly through the stomach together withthe specific insulin like protein/insulin and it remains available forabsorption at the intestinal level. Even if some insulin getsdestroyed, enough amount still remains available for absorptionin intestine. Beneficial effect of camel milk on beta cell function,insulin dose, lipid profile, plasma insulin, plasma c-peptide anddiabetes quality of life score has recently been reported [31]. Addi-tionally, studies using camel milk as an adjuvant therapy in treat-ment of T1D subjects revealed improved metabolic control [44].Two randomized control trials on T1D patients from westernRajasthan (geographic region where study population of Raikaand non-Raika were enrolled) showed improved glycemic controland reduction in doses of insulin [45,46]. A strong correlation hasalso been observed between T1D progression and viral infectionsparticularly of the enteroviral type [47]. A close temporal correla-tion has been reported between the appearance of diabetes-associ-ated autoantibodies and enteroviral (EV) infection [48,49]. Theabsence of the EV viral infection which can act as a trigger forthe manifestation of disease is an area of active future research.In this regard, a recent study has confirmed the anti-infectiousaction of camel milk in mice infected with Salmonella enterica [50].

It is reasonable to argue that in addition to living in a protectiveenvironment, the Raikas carry a protective gene for T1D that couldplay a dominant role over the susceptibility HLA allele/haplotypes.Despite a strong genetic component, progression of T1D is depen-dent on very strong requirement of environmental triggers. There-fore T1D susceptible HLA genes along with environmental triggersmay actually result in progression of the disease to clinical T1D.This highlights the possibility that if this population is exposedto susceptible environment, their progression to T1D owing tothe repertoire of susceptible risk haplotypes cannot be ruled out.Our results highlight the need to further define the gene-environ-ment interaction in the context of a multifactorial disease like T1D.

Acknowledgments

The study was supported by a research grant from the IndianCouncil of Medical Research (ICMR), New Delhi, India (No. 57/4/Diab/06-NCD-II). The authors sincerely acknowledge the staff of

ion of high prevalence of type 1 diabetes susceptibility genes and near zeromimm.2014.09.009

D.K. Bhat et al. / Human Immunology xxx (2014) xxx–xxx 7

the Diabetes Care and Research Centre at Sardar Patel, Medical Col-lege, Bikaner, for their support in organizing the medical camps.

References

[1] Barrett JC, Clayton DG, Concannon P, Akolkar B, Cooper JD, Erlich HA, et al.Genome-wide association study and meta-analysis find that over 40 loci affectrisk of type 1 diabetes. Nat Genet 2009;41:703–7.

[2] Mosaad YM, Auf FA, Metwally SS, Elsharkawy AA, El-Hawary AK, Hassan RH,et al. HLA-DQB1* alleles and genetic susceptibility to type 1 diabetes mellitus.World J Diabetes 2012;3:149–55.

[3] Noble JA, Valdes AM. Genetics of the HLA region in the prediction of type 1diabetes. Curr Diab Rep 2011;11:533–42.

[4] Kanga U, Vaidyanathan B, Jaini R, Menon PS, Mehra NK. HLA haplotypesassociated with type 1 diabetes mellitus in North Indian children. HumImmunol 2004;65:47–53.

[5] Mehra NK, Kumar N, Kaur G, Kanga U, Tandon N. Biomarkers of susceptibilityto type 1 diabetes with special reference to the Indian population. Indian J MedRes 2007;125:321–44.

[6] Kaur G, Sarkar N, Bhatnagar S, Kumar S, Rapthap CC, Bhan MK, et al. Pediatricceliac disease in India is associated with multiple DR3-DQ2 haplotypes. HumImmunol 2002;63:677–82.

[7] Nejentsev S, Howson JM, Walker NM, Szeszko J, Field SF, Stevens HE, et al.Localization of type 1 diabetes susceptibility to the MHC class I genes HLA-Band HLA-A. Nature 2007;450:887–92.

[8] Howson JM, Walker NM, Clayton D, Todd JA, Type 1 Diabetes GeneticsConsortium. Confirmation of HLA class II independent type 1 diabetesassociations in the major histocompatibility complex including HLA-B andHLA-A. Diabetes Obes Metab 2009;11:31–45.

[9] Ziegler AG, Nepom GT. Prediction and pathogenesis in type1diabetes.Immunity 2010;32:468–78.

[10] Noble JA, Valdes AM, Varney MD, Carlson JA, Moonsamy P, Fear AL, et al. HLAclass I and genetic susceptibility to type 1 diabetes: results from the Type 1Diabetes Genetics Consortium. Diabetes 2010;59:2972–9.

[11] Tait BD, Colman PG, Morahan G, Marchinovska L, Dore E, Gellert S, et al. HLAgenes associated with autoimmunity and progression to disease in type 1diabetes. Tissue Antigens 2003;61:146–53.

[12] Mbunwe E, Van der Auwera BJ, Vermeulen I, Demeester S, Van Dalem A, BaltiEV, et al. HLA-A*24 is an independent predictor of 5-year progression todiabetes in autoantibody-positive first-degree relatives of type 1 diabeticpatients. Diabetes 2013;62:1345–50.

[13] Price P, Witt C, Allcock R, Sayer D, Gariepp M, Kok CC, et al. The genetic basisfor the association of 8.1 ancestral haplotypes (A1, B8, DR3) with multipleimmunopathological diseases. Immunol Rev 1999;167:257–74.

[14] Witt CS, Price P, Kaur G, Cheong K, Kanga U, Sayer D, et al. The common HLA-B8-DR3 haplotype in northern India is different from that found in Europe.Tissue Antigens 2002;60:474–80.

[15] Candore G, Campagna AM, Cuppari I, Di Carlo D, Mineo C, Caruso C. Geneticcontrol of immune response in carriers of the 8.1 ancestral haplotype:correlation with levels of IgG subclasses: its relevance in the pathogenesis ofautoimmune diseases. Ann N Y Acad Sci 2007;1110:151–8.

[16] Kumar N, Kaur G, Mehra NK. Genetic determinants of type 1 diabetes: immuneresponse genes. Biomarkers Med 2009;3:153–73.

[17] Kumar N, Kaur G, Tandon N, Kanga U, Mehra NK. Genomic evaluation of HLA-DR3(+) haplotypes associated with type 1 diabetes. Ann N Y Acad Sci2013;1283:91–6.

[18] Ettinger RA, Papadopoulos GK, Moustakas AK, Nepom GT, Kwok WW. Allelicvariation in key peptide-binding pockets discriminates between closelyrelated diabetes-protective and diabetes-susceptible HLA-DQB1*06 alleles. JImmunol 2006;1(176):1988–98.

[19] Green A. Patterson CC; EURODIAB TIGER study group-Europe and diabetes.Trends in the incidence of childhood-onset diabetes in Europe 1989–1998.Diabetologia 2001;44(Suppl. 3):B3–8.

[20] DIAMOND Project Group. Incidence and trends of childhood type 1 diabetesworldwide 1990–1999. Diabet Med 2006;23:857–66.

[21] Awdeh ZL, Yunis EJ, Audeh MJ, Fici D, Pugliese A, Larsen CE, et al. A geneticexplanation for the rising incidence of type 1 diabetes, a polygenic disease. JAutoimmun 2006;27:174–81.

[22] Okada H, Kuhn C, Feillet H, Bach JF. The ‘hygiene hypothesis’ for autoimmuneand allergic diseases: an update. Clin Exp Immunol 2010;160:1–9.

[23] Bach JF. The effect of infections on susceptibility to autoimmune and allergicdiseases. N Engl J Med 2002;347:911–20.

[24] Bach JF, Chatenoud L. The hygiene hypothesis: an explanation for the increasedfrequency of insulin dependent diabetes. Cold Spring Harb Perspect Med2012;4:a007799.

Please cite this article in press as: Bhat DK et al. The Raikas – A unique combinatincidence of the disease. Hum Immunol (2014), http://dx.doi.org/10.1016/j.hum

[25] Adorini L. Tolerogenic dendritic cells induced by vitamin D receptor ligandsenhance regulatory T cells inhibiting autoimmune diabetes. Ann N Y Acad Sci2003;987:258–61.

[26] Agrawal RP, Budania S, Sharma P, Gupta R, Kochar DK, Panwar RB, et al. Zeroprevalence of diabetes in camel milk consuming Raica community of north-west Rajasthan, India. Diabetes Res Clin Pract 2007;76:290–6.

[27] Genuth S, Alberti KG, Bennett P, Buse J, Defronzo R, Kahn R, et al. Expertcommittee on the diagnosis and classification of diabetes mellitus. Follow-upreport on the diagnosis of diabetes mellitus. Diabetes Care 2003;26:3160–7.

[28] Mehra NK, Taneja V, Kailash S, Raizada N, Vaidya MC. Distribution of HLAantigens in a sample of the North Indian Hindu population. Tissue Antigens1986;27:64–74.

[29] Karvonen M, Tuomilehto J, Libman I, LaPorte R. A review of the recentepidemiological data on incidence of type 1 (insulin-dependent) diabetesmellitus World Health Organization DIAMOND project group. Diabetologia1993;36:883–92.

[30] Hua F, Shui-xian S, Zhao-wen C, Jia-jun W, Ting-ting Y, LaPorte RE, et al. Chinahas the lowest confirmed incidence of childhood diabetes in the world.Diabetes Care 1994;17:1206–8.

[31] Agarwal RP, Tantia P, Jain S, Agarwal R, Agarwal V. Camel milk: a possible boonfor type 1 diabetic population. Cell Mol Biol 2013;59:99–107.

[32] Sugihara S, Sakamaki T, Konda S, Murata A, Wataki K, Kobayashi Y, et al.Association of HLA-DR, DQ genotype with different beta-cell functions atIDDM diagnosis in Japanese children. Diabetes 1997;46(11):1893–7.

[33] Kikuoka N, Sugihara S, Yanagawa T, Ikezaki A, Kim HS, Matsuoka H, et al.Cytotoxic T lymphocyte antigen 4 gene polymorphism confers susceptibility totype 1 diabetes in Japanese children: analysis of association with HLAgenotypes and autoantibodies. Clin Endocrinol 2001;55(5):597–603.

[34] Penny MA, Jenkins D, Mijovic CH. DRw9-DQw9 haplotypes and insulin-dependent diabetes mellitus in Chinese subjects. Tissue Antigens1991;38(4):191.

[35] Saruhan-Direskeneli G, Uyar FA, Bas F, Gunoz H, Bundak R, Saka N, et al. HLA-DR and -DQ associations with insulin dependent diabetes mellitus in apopulation of Turkey. Hum Immunol 2000;61:296–302.

[36] Akerblom HK, Vaarala O, Hyöty H, Ilonen J, Knip M. Environmental factors inthe etiology of type 1 diabetes. Am J Med Genet 2002;115:18–29.

[37] Peng H, Hagopian W. Environmental factors in the development of type 1diabetes. Rev Endocr Metab Disord 2006;7:149–62.

[38] Hyöty H, Taylor KW. The role of viruses in human diabetes. Diabetologia2002;45:1353–61.

[39] Knip M, Veijola R, Virtanen SM, Hyöty H, Vaarala O, Akerblom HK.Environmental triggers and determinants of b-cell autoimmunity and type 1diabetes. Diabetes 2005(suppl. 2):S125–136.

[40] Agrawal RP, Sharma P, Gafoorunissa SJ, Ibrahim SA, Shah B, Shukla DK, et al.Effect of camel milk on glucose metabolism in adults with normal glucosetolerance and type 2 diabetes in Raica community: a crossover study. ActaBiomed 2011;82:181–6.

[41] Malik A, Al-Senaidy A, Skrzypczak-Jankun E, Jankun J. A study of the anti-diabetic agents of camel milk. Int J Mol Med 2012;30:585–92.

[42] Badr G, Badr BM, Mahmoud MH, Mohany M, Rabah DM, Garraud O. Treatmentof diabetic mice with undenatured whey protein accelerates the woundhealing process by enhancing the expression of MIP-1a, MIP-2, KC, CX3CL1and TGF-b in wounded tissue. BMC Immunol 2012;13:32–40.

[43] Sboui A, Khorchani T, Djegham M, Agrebi A, Elhatmi H, Belhadj O. Anti-diabeticeffect of camel milk in alloxan-induced diabetic dogs: a dose–responseexperiment. J Anim Physiol Anim Nutr 2010;94:540–6.

[44] Mohamad RH, Zekry ZK, Al-Mehdar HA, Salama O, El-Shaieb SE, El-Basmy AA,et al. Camel milk as an adjuvant therapy for the treatment of type 1 diabetes:verification of a traditional ethnomedical practice. J Med Food 2009;12:461–5.

[45] Agrawal RP, Beniwal R, Kochar DK, Tuteja FC, Ghorui SK, Sahani MS, et al.Camel milk as an adjunct to insulin therapy improves long-term glycemiccontrol and reduction in doses of insulin in patients with type-1 diabetes A 1year randomized controlled trial. Diabetes Res Clin Pract 2005;68:176–7.

[46] Agrawal RP, Jain S, Shah S, Chopra A, Agarwal V. Effect of camel milk onglycemic control and insulin requirement in patients with type 1 diabetes: 2-years randomized controlled trial. Eur J Clin Nutr 2011;65:1048–52.

[47] Yeung WC, Rawlinson WD, Craig ME. Enterovirus infection and type 1 diabetesmellitus: systematic review and meta-analysis of observational molecularstudies. Br Med J 2011;342:d35.

[48] Lönnrot M, Korpela K, Knip M, Ilonen J, Simell O, Korhonen S, et al. Enterovirusinfection as a risk factor for ß-cell autoimmunity in a prospectively observedbirth cohort: The Finnish Diabetes Prediction and Prevention (DIPP) Study.Diabetes 2000;49:1314–8.

[49] Oikarinen S, Martiskainen M, Tauriainen S, Huhtala H, Ilonen J, Veijola R, et al.Enterovirus RNA in blood is linked to the development of type 1 diabetes.Diabetes 2011;60:276–9.

[50] Cardoso RR, Ponte M, Leite V. Protective action of camel milk in miceinoculated with Salmonella enterica. Isr Med Assoc J 2013;15:5–8.

ion of high prevalence of type 1 diabetes susceptibility genes and near zeroimm.2014.09.009