The co-inheritance of type 1 diabetes and multiple sclerosis in Sardinia cannot be explained by genotype variation in the HLA region alone

Transcription

1 Human Molecular Genetics, 2004, Vol. 13, No doi: /hmg/ddh319 Advance Access published on October 7, 2004 The co-inheritance of type 1 diabetes and multiple sclerosis in Sardinia cannot be explained by genotype variation in the HLA region alone Maria Giovanna Marrosu 1, Costantino Motzo 2, Raffaele Murru 1, Rosanna Lampis 2, Gianna Costa 1, Patrizia Zavattari 2, Daniela Contu 2, Elisabetta Fadda 1, Eleonora Cocco 1 and Francesco Cucca 2,3, * 1 Centro Sclerosi Multipla, Dipartimento di Neuroscienze and 2 Laboratorio di Immunogenetica, Dipartimento di Scienze Biomediche e Biotecnologie, University of Cagliari, Italy and 3 Centro di Genetica Clinica, Dipartimento di Scienze Biomediche, University of Sassari, Italy Received June 23, 2004; Revised August 30, 2004; Accepted September 26, 2004 Type 1 diabetes (T1D) and multiple sclerosis (MS) are two autoimmune diseases which exhibit a considerably higher incidence in Sardinia compared with the surrounding southern European populations. Surprisingly, a 5-fold increased prevalence of T1D has also been observed in Sardinian MS patients. Susceptibility to both disorders is associated with common variants of the HLA-DRB1 and -DQB1 loci. In this study, we determined the relative contribution of genotype variation of these loci to the co-occurrence of the two disorders in Sardinia. We genotyped 1052 T1D patients and 1049 MS patients (31 of whom also had T1D) together with 1917 ethnically matched controls. On the basis of the absolute risks for T1D of the HLA-DRB1-DQB1 genotypes, we established that these loci would only contribute to a 2-fold increase in T1D prevalence in MS patients. From this evidence, we conclude that shared disease associations due to the HLA-DRB1-DQB1 loci provide only a partial explanation for the observed increased prevalence of T1D in Sardinian MS patients. The data suggest that variation at other non-hla class II loci, and/or unknown environmental factors contribute significantly to the co-occurrence of these two traits. INTRODUCTION Type 1 diabetes (T1D) and multiple sclerosis (MS) are common inflammatory disorders which result from an autoimmune attack on the pancreatic b-cells and on the central nervous system, respectively. In a recent epidemiological survey of a cohort of Sardinian MS patients and their families, it has been found that the prevalence of T1D is increased in these patients and in their first degree relatives with respect to the general population (1). Notably, the prevalence of T1D was 5-fold higher in MS patients and 2-fold higher in their parents and siblings than in the general population (1). It has been commonly believed that T1D and MS do not occur together outside Sardinia. However, a recent report has indicated that adult T1D females from the United States have a 20-fold increase in the prevalence of MS in comparison with an ethnically matched female control population (2). Taken together, these findings provide evidence of a clustering of T1D and MS, at least in some population. There are also clear indications of T-cell cross-reactivity in T1D and MS. T-lymphocytes isolated from recent-onset T1D patients react against major central nervous system autoantigens characteristic of MS. Likewise, T-cells from MS patients react against the typical autoantigens of T1D (3). Furthermore, it is well established that T1D occurs more frequently in the patients and in the families of the patients with other autoimmune disorders, such as autoimmune thyroid disease, rheumatoid arthritis and coeliac disease (4,5), thus suggesting the existence of factors with a central role in general autoimmunity. Despite these observations, the reasons for the co-occurrence of T1D and MS or in general of autoimmune disorders or phenomena remain largely unknown. Numerous studies have provided a persuasive case that genetic variation within the HLA complex is involved in the predisposition to both T1D *To whom correspondence should be addressed at: Dipartimento di Scienze Biomediche e Biotecnologie, University of Cagliari, Via Jenner, Cagliari 09121, Italy. Tel: þ ; Fax: þ ; fcucca@mcweb.unica.it Human Molecular Genetics, Vol. 13, No. 23 # Oxford University Press 2004; all rights reserved

2 2920 Human Molecular Genetics, 2004, Vol. 13, No. 23 and MS. In these disorders, the HLA region shows a complex multi-locus and multi-allelic disease association with the main genetic effects encoded by variation at the HLA class II loci DQB1 and DRB1 (6 10). However, important differences in the HLA associations with T1D and MS were found, both in terms of content and relative contributions of the disease associated alleles and haplotypes (7 9). The HLA region encodes a major component of the disease risk in T1D, whereas it carries a much more modest albeit the only genetic effect so far consistently detected in MS (7,9). It has been proposed that the fact that T1D and MS are seen together in Sardinia but not in other population would be explained by the unique genetic make-up and distribution of HLA haplotypes in the Sardinian population (1,11). Notably, the HLA-DR3 (DRB DQB1 0201) haplotype is extremely frequent in this population and it is positively associated with both disorders (7 9). Moreover, another haplotype, HLA-DR2 (DRB DQB1 0602), is very rare in Sardinia (9) but very common in other ethnic groups where it represents the main HLA predisposing haplotype for MS (10,12) and a dominantly protective haplotype for T1D (8). Thus, in non-sardinian MS patients and families, the DRB DQB haplotype might have concealed the presence of T1D and MS shared autoimmunity genes (1,11). However, this contention and the putative masking effect of the DRB DQB haplotype are at odds with the findings of Dorman et al. (2) who reported that the two disorders also co-segregate in the USA population. Indeed, in this population, the DRB DQB haplotype is extremely frequent and represents the most predisposing HLA haplotype for MS (10). In general, it is unclear whether the co-inheritance of T1D and MS is merely accounted for by shared predisposing variation within the HLA region or whether it is explained, at least in part, by the presence of still unknown non-hla polymorphism(s) and/or environmental factors influencing autoimmune inflammation. In this study, we directly addressed this issue by establishing the relative contribution of the genes located in the HLA class II region to the observed increased prevalence of T1D in Sardinian MS patients. RESULTS First, we established the association of genotype variation at the DRB1 and DQB1 loci in a data set of 1052 T1D Sardinian patients and 1917 ethnically matched controls (see Materials and Methods). In T1D, there are important genotype effects (6,8,13) and therefore the association of this region with the disease is best accounted for by the complete mating type. We observed a continuum of association from positively to negatively associated genotypes with the overall profile of disease association skewed toward the negatively associated molecules (Table 1). Overall, there were nine genotypes showing evidence of positive association with the disease [i.e. genotypes in which the patient/control (P/C) ratio is significantly increased relative to a value of 1]. Of these nine genotypes, three provided the highest disease risk and were constituted by various combinations of DR3, DRB DQB and DR4, DRB1 04-DQB (in which DRB1 04 is equal to any DRB1 04 subtype different from DRB1 0403) (Table 1). Next, we evaluated the association of the various DRB1- DQB1 genotypes with MS in a sample set of 1049 MS patients (which included 31 patients with both MS and T1D, independent from the T1D patients previously considered). We found that the genotypes which dominated the association of the HLA region with T1D were only marginally predisposing to MS (Table 1). For instance the DR3/DR3 genotype shows a P/C ratio of 5.7 (95% CI: ) in T1D, whereas it shows a P/C ratio of 2.6 (95% CI: ) in MS (see Table 1 and Materials and Methods). An even more dramatic difference was seen for DR3/DR4-DQB1-0302, which shows a strong positive association with T1D (P/C ratio of 10.5, 95% CI: ), whereas it is only slightly more frequent in the MS patients than in the controls (P/C ratio of 1.7, 95% CI: ). Reciprocally, three genotypes show a positive association with MS but not with T1D. First, the DR3/DR4- DQB genotype is positively associated with MS (P/C ratio of 3.4, 95% CI: ) and neutrally associated with T1D (P/C ratio of 1.1, 95% CI: ). Second, the DR3/ DR2-DQB genotype is associated with high risk for MS (P/C ratio of 4.2, 95% CI: ) and it is negatively associated with T1D (P/C ratio of 0.3, 95% CI: ). Finally, also the DR3/DR13-DQB genotype is highly predisposing for MS (P/C ratio of 5.5, 95% CI: ) and negatively associated with T1D (P/C ratio of 0.3, 95% CI: ). The differences in the HLA class II disease associations between T1D and MS are also clearly illustrated by the different pair-wise odds ratio (POR) and related probabilities computed relative to the same reference genotype (Table 1 and Materials and Methods). When we considered the association of the various DRB1-DQB1 genotypes in the subgroup of 31 patients who developed both T1D and MS, we found an association with the disease very similar to that observed in the patients with only T1D (data not shown). Taken as a whole, the nine genotypes showing a positive association with T1D are seen in 80.8% of the T1D patients, in 83.9% of the patients with both T1D and MS but in just 27.8% of patients with MS only and in 20.9% of the general population. However, in striking contrast with the male excess observed in cases with only T1D (M : F ratio in this data set equal to 1.3, P ¼ ), in the sub-group of patients with both T1D and MS there is a significant female excess (F : M ratio equal to 3.6, P ¼ ). This ratio is even higher than that seen in the cases with MS alone (F : M ratio equal to 2.2, P ¼ ). These observations underline the presence of distinct sex-specific risk factors in T1D and MS. Overall, these data indicate an asymmetry in the association of the HLA class II genotypes with T1D and MS and suggest that the HLA class II loci are unlikely to contribute significantly to the observed co-segregation of T1D and MS. We directly addressed this issue by computing the expected number of MS patients with T1D, using the T1D absolute risk (AR) estimates of the general Sardinian population and the distribution of the various DRB1-DQB1 genotypes observed in the MS patients (Materials and Methods and Table 1). We found a striking difference between the number of expected and number of observed MS patients who also had T1D.

3 Table 1. Distribution of the DRB1-DQB1 genotypes in the Sardinian T1D and MS patients and controls Genotypes a T1D MS Controls T1D MS AR T1D Exp T1D in MS DRB1 DQB1 DRB1 DQB1 N % N % N % P/C ratio POR 95% CI P P/C ratio POR 95% CI P % N Other genotypes Total POR, pair-wise odds ratio; AR, absolute risk; P/C ratio, patient/control ratio. AR T1D refers to the genotype specific AR of developing T1D in the general Sardinian population. Exp T1D in MS refers to the expected number of MS patients with T1D in the genotyped samples. PORs are computed using DRB DQB1 0502/DRB DQB as the reference genotype (bold type). When the number of patients or controls with a given DRB1-DQB1 genotype was zero, the POR, P/C ratio and AR are computed by adding a value of 0.5 to both the patients and controls counts. a Only genotypes with a frequency 1% in the individual patients or control cohorts are shown in this table. Human Molecular Genetics, 2004, Vol. 13, No

4 2922 Human Molecular Genetics, 2004, Vol. 13, No. 23 The number of cases expected was 8.3, on the basis of the HLA class II risk alone, whereas the actual number observed was 31 among the 1049 independent MS cases. In order to avoid any ascertainment bias, we have extended these calculations and considered the global prevalence of T1D in the whole cohort of patients treated in the MS Center. Of the 1364 patients considered, 10.7 MS patients are expected to have T1D, whereas we observed 32 with both diseases (ratio of observed/expected, if only the HLA region was contributing to the correlated occurrence, is equal to 3.0, 95% CI: , P ¼ 0.001). DISCUSSION We have found that genotype variation at the HLA class II loci DRB1 and DQB1 can only partially account for the co-segregation of T1D and MS observed in the Sardinian population. These results indicate that shared non-drb1-dqb1 genetic and/or environmental factors influence the susceptibility to both MS and T1D. This is consistent with the observation that the two disorders show a co-occurrence in a population in which the HLA-DRB1-DQB1 associations with T1D and MS are largely distinct (2). The existence of loci which regulate inflammation and which are shared in different autoimmune diseases is consistent with preliminary work in animal models of T1D and other related diseases. The non-obese diabetic mouse is not only susceptible to T1D, which develops spontaneously, but also to experimental autoimmune encephalomyelitis (EAE), a model of MS which can be induced by immunizing the mouse with myelin antigens. A number of quantitative trait loci influencing susceptibility to these autoimmune disorders have been mapped (14 17) and some of them co-localize (4,18), raising the possibility that there are disease genes shared in T1D and EAE. For instance, the non-mhc autoimmune disease locus Idd3 not only controls the anti-islet inflammation observed in this model of T1D but also influences inflammation in the brain of the EAE model of MS (18). Consistent with our observations in the Sardinian patients, the mode of inheritance of the susceptibility to T1D and EAE observed in these mouse animal models suggests that the MHC genes are important, as are the non-mhc genes, but neither the MHC nor non-mhc genes alone are sufficient to cause the disease (19 21). From our data, we cannot rule out a role for additional non- DRB1-DQB1 loci located within the HLA region. The disease associations of these non-class II variants have not been fully elucidated in either T1D or MS and were not considered in this study. Nevertheless, a major contribution of other HLA genes seems unlikely since, within the HLA region, the DRB1-DQB1 class II loci account for the main genetic effects in both T1D and MS (6 10). Furthermore, we found that particular combinations of non-drb1-dqb1 HLA variants that further influence the HLA association with T1D in Sardinia (22) were not able to do so in MS (unpublished data). Finally, it is also possible that unknown environmental factors predisposing to general autoimmunity may contribute to the observed co-inheritance of T1D and MS. These permissive environmental factors could be more common in Sardinia when compared with other European regions. However, the fact that the two disorders also show a co-inheritance in other population (2) suggests that they are unlikely to be Sardiniaspecific. Furthermore, children with Sardinian parents, who were born and live in the Italian mainland in the region of Lazio, where T1D is about six times less frequent than in Sardinia, were found to have about the same incidence of T1D as the Sardinian children still living on the island (23). This observation is important because Sardinia represents the major exception to the general North South gradient of both T1D and MS incidence in Europe. It suggests that susceptibility alleles predisposing to autoimmunity might be highly prevalent in this population. The discovery of these variants will provide insights into the aetiology of autoimmunity in humans. MATERIALS AND METHODS Study participants We included in our analysis only people born and living in Sardinia and whose parents and grandparents were also born in Sardinia. All MS patients met the criteria of clinically or laboratory supported definite MS (24). All T1D patients met the criteria reported by the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus (25). The 1052 independent T1D patients considered in this study consisted of 505 cases from a family data set (in which the parents of the patients were also collected and HLA typed) and 547 additional sporadic patients. The average age at disease onset was 9.5 years (SD + 6.0) in the patients from the family data set and 12.1 years (SD + 7.8) in the sporadic patients, with an average at disease onset in the total T1D data set equal to 10.6 years (minimum age equal to 0.7, maximum age equal to 40 years). In order to ensure the independence of the data points, in 103 families with affected relatives only the proband was considered (78 affected siblings pairs, and 25 parent offspring pairs). From a cohort of 1364 patients that had the diagnosis of definite MS from the Multiple Sclerosis Clinic in Cagliari, between January 1, 1989 and December 31, 2002, it was possible to collect 1151 MS individuals. Again, in order to ensure the independence of the data points, only the probands were selected from 102 families with affected relatives (67 affected siblings pairs, including one with both T1D and MS, and 34 parent offspring pairs). Thus, we considered 1049 independent MS patients, who included 31 cases with both MS and T1D that were also independent from the individuals considered in the T1D sample set. This sample set of 1049 MS patients included 707 families in which both parents (N ¼ 596) or one parent (N ¼ 111) of the patients were collected and HLA typed and 342 patients in which the parental data were not available. The average age of onset for MS was 28.9 years, range 6 71 years, whereas the average age of onset for T1D in the patients with both MS and T1D was 17.2 years, range 2 37 years. Of the 31 patients with T1D and MS, 28 had T1D first (average interval between the T1D and MS onsets ¼ 12.4 years), two had MS first (average gap between the MS and T1D onsets ¼ 9.7 years) and one had the disease onset for both T1D and MS during the same year.

5 Human Molecular Genetics, 2004, Vol. 13, No HLA-DRB1 and -DQB1 typing The entire sample set was typed by PCR amplification of the polymorphic second exon of the HLA-DRB1-DQB1 gene and dot-blot analysis of amplified DNA with sequence-specific oligonucleotide probes as previously reported (8,9). The haplotype and genotype frequencies of 120 T1D patients and the haplotype frequencies of 868 MS patients have already been reported (9,13). The HLA-DRB1 and -DQB1 haplotypes and genotypes in T1D and MS families were established following the co-segregation of alleles within families. Only definite haplotypes from parental genotype data were considered. The DRB1-DQB1 haplotypes and genotypes in sporadic T1D and MS patients as well as in unrelated newborns and blood donors were assigned following the known patterns of LD in Sardinians (26). In case of rare associations, the haplotypes were accepted only when the haplotype present on the other chromosome was well defined. Ambiguous assignments were resolved by excluding those individuals. We did not observe a significant difference in the distribution of genotypes assigned with and without parental data from either control or patient groups (see Statistical analysis The association of the DRB1-DQB1 genotypes with T1D and MS was assessed using a case control design. The control data set included 625 newborns, 447 blood donors and 845 affected family based controls (AFBAC). The AFBAC samples were assembled from 376 simplex families with T1D and 469 simplex families with MS. This was done by selecting, family by family, the haplotypes not transmitted from the parents to the affected child at the HLA-DRB1-DQB1 loci (27). In this way, it is possible to assemble in each simplex family, a pseudo-individual carrying one genotype at HLA-DRB1- DQB1. Note that the individual blood donor, the newborn and the AFBAC sample sets were in Hardy Weinberg equilibrium and showed similar and not significantly different allele and genotype frequencies at the loci considered in this study (data not shown). The frequencies of the HLA class II genotypes observed in patients and in controls were compared by estimating the POR, P/C ratio and AR. When a disease is associated with more than one marker (allele, haplotype or genotype) at a given locus, the association of a test marker is influenced by the other associated markers. In this respect, the standard odds ratios, computed comparing one marker against all the others grouped, are not appropriate for an accurate estimate of the strength of the disease association. To alleviate this problem and to provide a more reliable computation of the relative risk of disease, in this study we have used PORs, in which the various genotypes are analyzed relative to one reference genotype. The resulting data points are arranged in a 2 2 contingency table and tested by Pearson s chi-squared test. In this study, as a reference marker we used the DRB DQB1 0502/DRB DQB genotype. This genotype was used because it is relatively common in the Sardinian population, it represents the homozygous state for the DRB DQB haplotype which is not associated with either T1D or MS and it shows an identical frequency in the T1D and MS sample sets. It is thus an appropriate reference for both disorders. The P/C ratio is defined as the frequency of a given genotype in the patients divided by the frequency in the controls. The genotype-specific ARs were calculated using the following formula: (a/b ) p; where a is the percentage of patients possessing a given HLA genotype, b the percentage of controls possessing that particular HLA genotype and p is the overall T1D prevalence. In this study, we considered a T1D prevalence of per hundred, which is the observed T1D prevalence in the age range 0 29 years in one of the Sardinian provinces (28). This prevalence estimate could be considered as representative of the whole island since no significant differences in disease incidence have been observed in different Sardinian subregions (29). Using this AR formula, we computed for each individual genotype the AR for T1D in the general Sardinian population, thus estimating the number of individuals per hundred who carry a given genotype and will develop the disease in the 0 29 years age range. To establish the expected number of MS patients with T1D, we considered genotype by genotype, the T1D AR estimates in the general Sardinian population and the observed distribution of genotypes in the MS patients. If for instance the T1D AR of a given genotype is equal to 2.6 per hundred and that genotype is present in 127 MS patients, the expected number of the MS patients carrying that genotype that will get also T1D will be equal to 3.3. The sum of the expected number of T1D patients for each individual genotype detected in the MS patients will correspond to the total expected number of MS patients with T1D in the MS cohort. This assuming that in the MS patients the non-hla background does not significantly increase the general population risk encoded by the HLA-genotypes (via gene gene interaction). Therefore, we tested the null hypothesis that the HLA class II genotypes were the only contributors to the observed increased prevalence of T1D in the MS cases. ACKNOWLEDGEMENTS We thank John Todd, Antonio Cao, Mauro Congia, Paolo Contu, Claudia Sardu for useful suggestions, Jamie Foster for editorial assistance and for help and advice, Efisio Angius, Paola Frongia, Margi Chessa and Rossella Ricciardi for help in collecting the Sardinian T1D families and for clinical information. We also thank the Juvenile Diabetes Research Foundation, the Italian Telethon, the Italian Multiple Sclerosis Society and the Regione Autonoma Sardegna Assessorato Sanita for financial support. REFERENCES 1. Marrosu, M.G., Cocco, E., Lai, M., Spinicci, G., Pischedda, M.P. and Contu, P. (2002) Patients with multiple sclerosis and risk of type 1 diabetes mellitus in Sardinia, Italy: a cohort study. Lancet, 359, Dorman, J.S., Steenkiste, A.R., Burke, J.P. and Songini, M. (2003) Type 1 diabetes and multiple sclerosis: together at last. Diabetes Care, 26, Winer, S., Astsaturov, I., Cheung, R., Gunaratnam, L., Kubiak, V., Cortez, M.A., Moscarello, M., O Connor, P.W., McKerlie, C., Becker, D.J. et al. (2001) Type I diabetes and multiple sclerosis patients target islet

6 2924 Human Molecular Genetics, 2004, Vol. 13, No. 23 plus central nervous system autoantigens; nonimmunized nonobese diabetic mice can develop autoimmune encephalitis. J. Immunol., 166, Vyse, T.J. and Todd, J.A. (1996) Genetic analysis of autoimmune disease. Cell, 85, Tait, K.F., Marshall, T., Berman, J., Carr-Smith, J., Rowe, B., Todd, J.A., Bain, S.C., Barnett, A.H. and Gough, S.C. (2004) Clustering of autoimmune disease in parents of siblings from the Type 1 diabetes Warren repository. Diabet. Med., 21, Noble, A.J., Valdes, A.M., Cook, M., Klitz, W., Thomson, G. and Erlich, H.A. (1996) The role of HLA class II genes in insulin-dependent diabetes mellitus: molecular analysis of 180 Caucasian, multiplex families. Am. J. Hum. Genet., 59, Zavattari, P., Lampis, R., Mulargia, A., Loddo, M., Angius, E., Todd, J.A. and Cucca, F. (2000) Confirmation of the DRB1-DQB1 loci as the major component of IDDM1 in the isolated founder population of Sardinia. Hum. Mol. Genet., 9, Cucca, F., Lampis, R., Congia, M., Angius, E., Nutland, S., Bain, S.C., Barnett, A.H. and Todd, J.A. (2001) A correlation between the relative predisposition of MHC class II alleles to type 1 diabetes and the structure of their proteins. Hum. Mol. Genet., 10, Marrosu, M.G., Murru, R., Murru, M.R., Costa, G., Zavattari, P., Whalen, M., Cocco, E., Mancosu, C., Schirru, L., Solla, E. et al. (2001) Dissection of the HLA association with multiple sclerosis in the founder isolated population of Sardinia. Hum. Mol. Genet., 10, Barcellos, L.F., Oksenberg, J.R., Begovich, A.B., Martin, E.R., Schmidt, S., Vittinghoff, E., Goodin, D.S., Pelletier, D., Lincoln, R.R., Bucher, P. et al. (2003) HLA-DR2 dose effect on susceptibility to multiple sclerosis and influence on disease course. Am. J. Hum. Genet., 72, Buzzetti, R., Pozzilli, P., Di Mario, U., Ballerini, C. and Massacesi, L. (2002) Multiple sclerosis and type I diabetes. Diabetologia, 45, Herrera, B.M. and Ebers, G.C. (2003) Progress in deciphering the genetics of multiple sclerosis. Curr. Opin. Neurol., 16, Cucca, F., Muntoni, F., Lampis, R., Frau, F., Argiolas, L., Silvetti, M., Angius, E., Cao, A., De Virgiliis, S. and Congia, M. (1993) Combinations of specific DRB1, DQA1, DQB1 haplotypes are associated with insulindependent diabetes mellitus in Sardinia. Hum. Immunol., 37, Encinas, J.A., Lees, M.B., Sobel, R.A., Symonowicz, C., Greer, J.M., Shovlin, C.L., Weiner, H.L., Seidman, C.E., Seidman, J.G. and Kuchroo, V.K. (1996) Genetic analysis of susceptibility to experimental autoimmune encephalomyelitis in a cross between SJL/J and B10.S mice. J. Immunol., 157, Sundvall, M., Jirholt, J., Yang, H.T., Jansson, L., Engstrom, A., Pettersson, U. and Holmdahl, R. (1995) Identification of murine loci associated with susceptibility to chronic experimental autoimmune encephalomyelitis. Nat. Genet., 10, Todd, J.A., Aitman, T.J., Cornall, R.J., Ghosh, S., Hall, J.R.S., Hearne, C.M., Knight, A.M., Love, J.M., McAleer, M.A., Prins, J.-B. et al. (1991) Genetic analysis of autoimmune type 1 diabetes mellitus in mice. Nature, 351, Ghosh, S., Palmer, S.M., Rodrigues, N.R., Cordell, H.J., Hearne, C.M., Cornall, R.J., Prins, J.B., McShane, P., Lathrop, G.M., Peterson, L.B. et al. (1993) Polygenic control of autoimmune diabetes in nonobese diabetic mice. Nat. Genet., 4, Encinas, J.A., Wicker, L.S., Peterson, L.B., Mukasa, A., Teuscher, C., Sobel, R., Weiner, H.L., Seidman, C.E., Seidman, J.G. and Kuchroo, V.K. (1999) QTL influencing autoimmune diabetes and encephalomyelitis map to a 0.15-cM region containing Il2. Nat. Genet., 21, Wicker, L.S., Todd, J.A. and Peterson, L.B. (1995) Genetic control of autoimmune diabetes in the NOD mouse. Ann. Rev. Immunol., 13, Fritz, R.B., Skeen, M.J., Chou, C.H., Garcia, M. and Egorov, I.K. (1985) Major histocompatibility complex-linked control of the murine immune response to myelin basic protein. J. Immunol., 134, Baker, D., Rosenwasser, O.A., O Neill, J.K. and Turk, J.L. (1995) Genetic analysis of experimental allergic encephalomyelitis in mice. J. Immunol., 155, Zavattari, P., Lampis, R., Motzo, C., Loddo, M., Mulargia, A., Whalen, M., Maioli, M., Angius, E., Todd, J.A. and Cucca, F. (2001) Conditional linkage disequilibrium analysis of a complex disease superlocus, IDDM1 in the HLA region, reveals the presence of independent modifying gene effects influencing the type 1 diabetes risk encoded by the major HLA-DQB1, -DRB1 disease loci. Hum. Mol. Genet., 10, Muntoni, S., Fonte, M., Marietti, G., Bizzarri, C., Crino, A., Ciampalini, P., Multari, G., Stuppa, M., Matteoli, M., Lucentini, L. et al. (1997) Incidence of insulin-dependent diabetes mellitus among Sardinian-heritage children born in Lazio region, Italy. Lancet, 349, Poser, C.M., Paty, D.W., Scheinberg, L., McDonald, W.I., Davis, F.A., Ebers, G.C., Johnson, K.P., Sibley, W.A., Silberberg, D.H. and Tourtellotte, W.W. (1983) New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann. Neurol., 13, Expert Committee on the Diagnosis and Classification of Diabetes Mellitus (1997) Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care, 20, Lampis, R., Morelli, L., Congia, M., Macis, M.D., Mulargia, A., Loddo, M., De Virgiliis, S., Marrosu, M.G., Todd, J.A. and Cucca, F. (2000) The inter-regional distribution of HLA class II haplotypes indicates the suitability of the Sardinian population for case control association studies in complex diseases. Hum. Mol. Genet., 9, Thomson, G. (1995) Mapping disease genes: family-based association studies. Am. J. Hum. Genet., 57, Frongia, O., Mastinu, F. and Sechi, G.M. (1997) Prevalence and 4-year incidence of insulin-dependent diabetes mellitus in the province of Oristano (Sardinia, Italy). Acta Diabetol., 34, Casu, A., Pascutto, C., Bernardinelli, L. and Songini, M. (2004) Type 1 diabetes among sardinian children is increasing: the Sardinian diabetes register for children aged 0 14 years ( ). Diabetes Care, 27,