Sa.63. Utilizing Long-range Genomic Identity to Assess Type 1A Diabetes Risk

Abstracts DR, B and A alleles, we identified 82 different haplotype groups (N = 2627 haplotypes within the 82 groups). Amongst the 82 haplotype groups, the number of HLA-matched (DR-BA) chromosomes varied from 510 for the DR3-B8-A1 (8.1) haplotype and 236 for the DR4-B15-A2 haplotype, to 10 chromosomes for 8 different haplotype groups. With high density SNP analysis across the MHC region (N = 2,918 SNPs), we determined the percentage and extent to which each of these haplotypes (within a haplotype group) were identical for essentially all SNPs. Of the 82 HLA defined haplotype groups, all but 1 group contained haplotypes that were SNP conserved from HLA-DR to HLA-B. The distance of identity varied from 0.98 Mb to 4.9 Mb (entire set of SNPs analyzed). When the number of chromosomes conserved from HLA-DR to HLA-B within each group were summed across all 82 groups, 42% of all 4,386 founder chromosomes were in haplotype groups and demonstrated the dramatic SNP conservation (Table). Haplotypes conserved over several megabases are both common and diverse with significant implications for transplantation, autoimmunity and localizing disease-relevant polymorphisms. Table: Percentage of all founder chromosomes SNP-identical for MHC regions DR to B: 42% (1824/4386); DR to A: 31% (1342/4386); DR to F: 30% (1294/ 4386); DR to 29.3 Mb: 25% (1109/4386); complete (34.2 Mb to 29.3 Mb): 8% (347/4386). This research utilizes resources provided by the Type 1 Diabetes Genetics Consortium. doi:10.1016/j.clim.2008.03.283

Sa.63. Utilizing Long-range Genomic Identity to Assess Type 1A Diabetes Risk Kelly Johnson, Erin Baschal, Jean Jasinski, George Eisenbarth. University of Colorado Denver, Aurora, CO Advances in technology now permit high-density SNP analysis across large genomic regions, allowing for definition of complex topography extending across millions of base pairs. The Type 1 Diabetes Genetics Consortium (T1DGC) analyzed 1,240 participating families (6,297 individuals) for 2,918 SNPs using Illumina exon-centric and mapping MHC panels. We utilized four “sentinel” 8.1 SNPs and HLA typing to define conserved 8.1 haplotypes. Using such a strict definition, 98% of identified chromosomes were ≥ 99% identical for the complete MHC region (HLA-DR to HLA-A). The conserved 8.1 haplotype was present in 40% of DR3 case chromosomes and 51% of DR3 control chromosomes (p = 0.009, OR = 0.65), indicating that the 8.1 haplotype is lower risk than other DR3 haplotypes despite identical DR and DQ alleles. DR3/3 individuals with an 8.1 haplotype (120 cases and 115 controls) were also at lower risk for type 1A diabetes than DR3/3 non-8.1 individuals [80 cases and 31 controls (p = 2.8E-4, OR = 0.40)]. The “Basque” conserved haplotype (DR3-B18A30) has been reported to be higher risk than other DR3 haplotypes, but after removing individuals with an 8.1 haplotype, the Basque haplotype was not associated with diabetes (p = 1, OR = 1.1) compared to the remaining DR3/3 individuals. Given the frequency (12% of all founder chromosomes) of the 8.1 haplotype, its extreme conservation and the lower risk of conserved 8.1 haplotypes than

S101 other DR3 haplotypes for type 1A diabetes suggests that the 8.1 haplotype should be studied for influence on other autoimmune disorders. This research utilizes resources provided by the Type 1 Diabetes Genetics Consortium. doi:10.1016/j.clim.2008.03.284

Sa.64. Analysis of T Cell Receptor Hybridomas with the Conserved anti-insulin B:9-23 BDC12-4.1 T Cell Receptor (TCR) Alpha Chain Li Zhang, Jean Jasinski, Masakazu Kobayashi, Bennett Davenport, Kelly Johnson, George Eisenbarth. University of Colorado, Aurora, CO There is evidence that insulin peptide B:9-23 is a primary autoantigen necessary for diabetes. The majority of the CD4 T cells cloned from islets of NOD mice recognized insulin and utilized conserved Vα and Jα chains with no conservation of beta chains to target the insulin B:9-23 peptide. We hypothesized TCR beta chains will be also important for targeting insulin B:9-23 peptide. We produced T cell hybridomas from mice restricted to having the complete 12-4.1 TCR α and β chain or restricted to utilizing the single 12-4.1 alpha chain and “random” endogenous β chains. With the complete TCR, 29% of the hybridomas responded to multiple insulin B: 9-23 peptides, in particular alanine replacing cysteine at position B19 (insulin B19A) and the native B: 9-23 peptide (insulin 1 and insulin 2). In vivo, immunized young NOD mice with B:19A peptide induced higher insulin autoantibodies than B:9-23 peptide. In contrast, no hybridoma derived from the 12-4.1 TCRα+Cα-/mice, utilizing multiple different Vβ chains, responded to insulin B:9-23 peptides (0/350 vs. 55/189, P b 0.0001) despite identical response to stimulation by anti-CD3 antibody. In summary: 1. Replacing native cysteine with alanine enhances anti-B:9-23 T cell response. 2. Though the TCR β chain does not show evidence of conservation as does the alpha chain, TCR β chain sequences crucially limit recognition of the insulin B:9-23 peptide with the canonical 12-4.1 anti-B:9-23 alpha chain. It is suggested the conserved Vα, Jα antiB:9-23 TCR play a limited set of Vβ chains drives a high protect of anti-insulin autoimmunity. doi:10.1016/j.clim.2008.03.285

Sa.65. Biosimulations Predict Optimal Oral Insulin/ Anti-CD3 and Oral Insulin/Exendin-4 Combination Treatment Regimens for the Reversal of Diabetes in the Non-Obese Diabetic (NOD) Mouse Yanan Zheng, 1 Damien Bresson, 2 Matthew Fradkin, 2 Jason Chan, 1 Matthias von Herrath, 2 Chan Whiting.3 1 Entelos, Inc., Foster City, CA; 2La Jolla Institute for Allergy and Immunology, La Jolla, CA; 3Ingenuity Systems, Inc., Redwood City, CA To date, no single pharmacologic agent has been identified that can reverse hyperglycemia and maintain normoglycemia in recent-onset human type 1 diabetes (T1D). However, combinations of existing agents may prove more effective.