Subcongenic Analyses Reveal Complex Interactions between Distal Chromosome 4 Genes Controlling Diabetogenic B Cells and CD4 T Cells in Nonobese Diabetic Mice This information is current as of October 2, 2021. Jessica Stolp, Yi-Guang Chen, Selwyn L. Cox, Vivien Henck, Wenyu Zhang, Shirng-Wern Tsaih, Harold Chapman, Timothy Stearns, David V. Serreze and Pablo A. Silveira

J Immunol published online 25 June 2012 Downloaded from http://www.jimmunol.org/content/early/2012/06/25/jimmun ol.1200120

Supplementary http://www.jimmunol.org/content/suppl/2012/06/25/jimmunol.120012 http://www.jimmunol.org/ Material 0.DC1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on October 2, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2012 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published June 25, 2012, doi:10.4049/jimmunol.1200120 The Journal of Immunology

Subcongenic Analyses Reveal Complex Interactions between Distal Chromosome 4 Genes Controlling Diabetogenic B Cells and CD4 T Cells in Nonobese Diabetic Mice

Jessica Stolp,*,† Yi-Guang Chen,‡,x Selwyn L. Cox,*,† Vivien Henck,* Wenyu Zhang,* Shirng-Wern Tsaih,x Harold Chapman,{ Timothy Stearns,{ David V. Serreze,{ and Pablo A. Silveira*,†

Autoimmune type 1 diabetes (T1D) in humans and NOD mice results from interactions between multiple susceptibility genes (termed Idd) located within and outside the MHC. Despite sharing ∼88% of their genome with NOD mice, including the H2g7 MHC haplotype and other important Idd genes, the closely related nonobese resistant (NOR) strain fails to develop T1D because of resistance alleles in residual genomic regions derived from C57BLKS mice mapping to chromosomes (Chr.) 1, 2, and 4. We Downloaded from previously produced a NOD background strain with a greatly decreased incidence of T1D as the result of a NOR-derived 44.31- Mb congenic region on distal Chr. 4 containing disease-resistance alleles that decrease the pathogenic activity of autoreactive B and CD4 T cells. In this study, a series of subcongenic strains for the NOR-derived Chr. 4 region was used to significantly refine genetic loci regulating diabetogenic B and CD4 T cell activity. Analyses of these subcongenic strains revealed the presence of at least two NOR-origin T1D resistance genes within this region. A 6.22-Mb region between rs13477999 and D4Mit32, not previously known to contain a locus affecting T1D susceptibility and now designated Idd25, was found to contain the main NOR gene(s) http://www.jimmunol.org/ dampening diabetogenic B cell activity, with Ephb2 and/or Padi2 being strong candidates as the causal variants. Penetrance of this Idd25 effect was influenced by genes in surrounding regions controlling B cell responsiveness and anergy induction. Conversely, the gene(s) controlling pathogenic CD4 T cell activity was mapped to a more proximal 24.26-Mb region between the rs3674285 and D4Mit203 markers. The Journal of Immunology, 2012, 189: 000–000.

utoreactive T cells are responsible for destroying pan- chronic treatment with anti-IgM Abs were strongly protected from creatic b cells during the development of type 1 diabetes T1D (3, 4). More recent studies showed that temporal B cell de- (T1D) in humans and the NOD mouse model (1, 2). A pletion in NOD mice at different stages of disease, using Abs or A by guest on October 2, 2021 pathogenic role for B cells in this disease first became evident chimeric proteins targeting CD20, CD22, or the B lymphocyte when NOD mice made deficient in this population through in- stimulator (Blys/BAFF), could be effective at delaying, inhibit- troduction of an inactivated Igm H chain gene (NOD.Igmnull)or ing, or reversing the onset of T1D (5, 6). These findings were mirrored in patients with recent-onset T1D treated with the B cell- depleting CD20-specific rituximab mAb, who, on average, exhibi- *Immunology Program, Garvan Institute of Medical Research, Darlinghurst, New ted decreased loss of endogenous insulin production and reduced South Wales 2010, Australia; †St. Vincent’s Clinical School, University of New requirement for exogenous insulin over a 1-y follow-up period (7). South Wales, Sydney, New South Wales 2052, Australia; ‡Department of Pediatrics, Max McGee National Research Center for Juvenile Diabetes, Medical College of This confirmed the role of B cells in the pathogenesis of human Wisconsin, Milwaukee, WI 53226; xHuman and Molecular Genetics Center, Medical T1D, as well as showed that targeting this population could be a { College of Wisconsin, Milwaukee, WI 53226; and The Jackson Laboratory, Bar potentially effective therapeutic strategy. Harbor, ME 04609 Multiple lines of evidence indicate that the primary pathogenic Received for publication January 11, 2012. Accepted for publication May 22, 2012. function of B cells in T1D consists of acting as APCs for CD4 This work was supported by grants from the National Health and Medical Research T cells. First, NOD.Igmnull mice were shown to produce poor CD4 Council of Australia (402727 and 427620), the National Institutes of Health (DK46266 and DK077443), the American Diabetes Association, and the Juvenile T cell responses to T1D-associated autoantigens (8–10). Second, Diabetes Research Foundation. J.S. and S.L.C. were recipients of Postgraduate Re- autoreactive CD4 T cell responses and the development of insulitis search Awards from the University of New South Wales. and T1D could only be restored by reconstituting NOD.Igmnull The microarray datasets presented in this article have been submitted to the National mice with NOD B cells and not autoantibodies (9). Finally, strong Center for Biotechnology Information/Gene Expression Omnibus (http://www.ncbi. nlm.nih.gov/geo/) under accession number GSE37294. T1D protection was observed in NOD bone marrow (BM) chimera Address correspondence and reprint requests to Dr. Pablo A. Silveira, Garvan Insti- mice whose B cells were made specifically deficient in MHC class tute of Medical Research, 384 Victoria Street, Darlinghurst, New South Wales 2010, II molecules (11). The pathogenic APC function of B cells was Australia. E-mail address: [email protected] shown to be particularly dependent on their capacity to specifi- The online version of this article contains supplemental material. cally capture Ags via surface Ig (12, 13). Thus, like T cells, B cells Abbreviations used in this article: B6, C57BL/6; B10, C57BL/10; BKS, C57BLKS; contributing to T1D must arise from defects in self-tolerance BM, bone marrow; Chr., chromosome; FL, full length; FO, follicular; HEL, hen egg lysozyme; IgHEL, hen egg lysozyme-specific Ig; MZ, marginal zone; NOR, non- mechanisms, leading to the generation and activation of autore- obese resistant; NR4, NOD.NOR-Chr4; q-rtPCR, quantitative real-time PCR; SBM, active clones. Indeed, defects in various tolerance-induction mecha- syngeneic bone marrow; sHEL, soluble hen egg lysozyme; SNP, single nucleotide nisms, including receptor editing (14), anergy (15–17), and peripheral polymorphism; T1D, type 1 diabetes; tg, transgenic; TJL, The Jackson Laboratory. deletion of self-reactive clonotypes (15, 18), were shown to con- Copyright Ó 2012 by The American Association of Immunologists, Inc. 0022-1767/12/$16.00 tribute to the generation of autoreactive B cells in NOD mice.

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1200120 2 GENETIC CONTROL OF DIABETOGENIC CD4 T AND B CELLS

Similar defects in B cell tolerance were recently reported in following stimulation through the BCR and CD40, irrespective of humans developing T1D (14, 19, 20). Given the important con- whether they developed in the presence or absence of their cog- tribution of B cells to T1D, it was likely that their pathogenic nate soluble neo–self-Ag (soluble HEL [sHEL]), similar B cells capacity would be controlled by genes conferring susceptibility to from the NR4 stock anergized properly in response to self-Ag this disease. exposure. A subsequent study found that the NR4 congenic re- Predisposition to T1D in both NOD mice and humans is de- gion also had the capacity to intrinsically dampen the pathogenic termined by multiple susceptibility genes located within and activity of CD4 T cells contributing to T1D (32). It remains to be outside the MHC (21, 22). In NOD mice, these genes have been elucidated whether the CD4 T cell effect is controlled by the same mapped to .40 loci (termed Idd) on 16 chromosomes (Chr.) (21). or a distinct NOR gene(s) as that controlling B cells. Although the MHC class II genes within the NOD H2g7 haplotype The NOR-derived congenic interval in NR4 mice is large (44.31 (Idd1) are the most influential determinants of T1D in this strain, Mb, containing 889 known or predicted genes), making identifi- contributions by non-MHC susceptibility genes are also necessary cation of genes regulating the pathogenic activity of B and CD4 for disease development (21). This is highlighted by the closely T cells within this region difficult. In this study, we describe the related nonobese resistant (NOR) mouse strain, which, despite generation of a series of novel subcongenic strains for the NOR- sharing ∼88% of its genome with NOD mice, including the H2g7 derived Chr. 4 region. B and CD4 T cells from subcongenic strains MHC haplotype, exhibits strong protection from T1D (23). This were used in various experimental strategies to significantly refine protection is encoded within C57BLKS (BKS)-derived genetic genetic regions within distal Chr. 4 that modulate development of regions (itself a combination of C57BL/6 [B6] and DBA/2 T1D by controlling the pathogenic function of these lymphocyte genomes) in NOR mice on Chr. 1, 2, and 4 (23–26). The NOR- populations. derived T1D resistance locus on Chr. 2, termed Idd13, had not Downloaded from been observed in outcrosses of NOD mice with other strains. Materials and Methods Conversely, the NOR-derived T1D resistance genes on Chr. 1 and Mice 4 were mapped to similar regions as were the previously identified Idd5.2/5.3 and Idd9.1/9.2/11 loci, respectively, which were lo- All mice were housed at the Garvan Institute or The Jackson Laboratory (TJL) animal facilities under standard specific pathogen-free conditions calized using NOD outcrosses with C57BL/10 [B10] or B6 with unrestricted access to food and acidified water. NOD/Lt mice were disease-resistant strains (27–29). However, given their diverse purchased from Australian BioResources (Moss Vale, NSW, Australia) or http://www.jimmunol.org/ origin from the BKS strain, it is also possible that the NOR- produced at TJL. Derivation of the NR4 strain (at N5) using a speed- derived T1D resistance genes on Chr. 1 or 4 are not completely congenic strategy was described previously (26). The proximal boundary of the congenic region in these mice lies between rs3674285 (104.99 Mb) overlapping with those of B6 or B10 origin. and D4Mit31 (106.78 Mb). Screening of additional genetic markers iden- Our previous studies showed that reconstitution of lethally ir- tified a more distal boundary for the congenic region than originally de- radiated NOD.Igmnull mice with syngeneic BM (SBM) and puri- scribed, which lies between D4Mit129 (148.30 Mb) and D4Mit127 (149.30 fied NOD B cells prevented rejection of grafted B cells by host Mb). Seven new subcongenic strains (NR4S1–S4 and NR4S6–7 at N7 and CTL and restored their susceptibility to T1D (9, 30). In contrast, NR4S5 at N8) containing smaller NOR-derived congenic regions on Chr. 4 were developed by outcrossing the NR4 strain with NOD mice and then those reconstituted with SBM only remained resistant to disease. backcrossing heterozygous progeny with NOD mice for two or three ad- by guest on October 2, 2021 null Interestingly, reconstitution of NOD.Igm recipients with equiva- ditional generations. Progeny of the first backcross were screened for lent levels of B cells from NOR mice, instead of NOD mice, recombinations within the original congenic interval by determining the resulted in a significantly reduced incidence of disease (30). To origin of microsatellite markers at the proximal (D4Mit31) and distal (D4Mit129) boundaries. Those that were NOD homozygous at one of these identify the NOR-derived Idd resistance loci responsible for null markers were screened with an additional 22 microsatellite (primer dampening the diabetogenic activity of B cells, NOD.Igm mice sequences obtained from the Mouse Genome Informatics Web site [http:// were reconstituted with B cells from NOD stocks containing www.informatics.jax.org/] or Ref. 27) or single nucleotide polymorphism NOR-derived congenic regions on Chr. 1, 2, or 4 (30). Only those (SNP) markers (carried out by KBioscience, Hoddesdon, U.K.) to refine B cells containing a congenic region on Chr. 4 (NOD.NOR-Chr4 the location of recombinations. The NR4S5 strain, containing two recom- binations within the original NR4 congenic interval, was generated by an [NR4]), spanning from 104.99 to 149.30 Mb (26), recapitulated additional backcross to NOD of a subcongenic mouse that was NOD/NOR the T1D resistance conferred by NOR B cells (30). Thus, genes heterozygous between microsatellites D4Mit203 and D4Mit129. Mice with within the Idd9.1/9.2/11 regions, or alternatively a previously un- informative recombinations underwent a final backcross to NOD to pro- recognized locus, are important for inhibiting diabetogenic B cell duce male and female offspring with the same recombination, which were intercrossed to fix the congenic regions to homozygosity. Genera- activity in NOR mice. tion of NOD.Igmnull (3), NOD.CD4null, NOD.AI4ab-tg (33), and NOD. NOD and NR4 B cells were subsequently compared for a num- B6-Ptprcb (NOD.CD45.2) (34) mice, as well as NOD and NR4 mice ber of characteristics to establish potential mechanisms by which expressing IgHEL and sHEL transgenes (12, 15, 30), were described polymorphic genes within the distal region of Chr. 4 controlled previously. Only female mice were used in experiments, which were their diabetogenic activity. Although a previous study reported that approved by the Garvan Institute/St. Vincent’s Hospital or TJL Animal Ethics Committees. genes within distal Chr. 4 contributed to exaggerated marginal zone (MZ) B cell development in NOD mice (31), splenic B cell subsets Assessment of T1D development were comparable in NOD and NR4 mice (30). In addition, dele- T1D development was assessed by measuring glycosuric values weekly tion of splenic transitional-1 B cells stimulated through the BCR with Ames Diastix (Bayer Corporation Diagnostics Division, Elkhart, IN). was equivalent between strains. However, mature B cells from Values $500 mg/dl were considered indicative of overt diabetes. lymph nodes (mainly of the follicular [FO] subset) or spleen (FO and MZ subsets) of NR4 congenic mice proliferated more vigor- Mixed BM and B cell chimeras ously in response to BCR stimulation than did those from NOD CTLs in otherwise unmanipulated NOD.Igmnull mice destroy passively mice; this was accentuated by CD40 costimulation (30). The NR4 transferred B cells (9). Therefore, B cell reconstitution was achieved by null congenic region was also shown to contribute to the maintenance lethally irradiating NOD.Igm mice between 4 and 5 wk of age with two 550-rad split doses from an x-ray source, followed by i.v. injection with of anergy in self-reactive B cells (30). Thus, in contrast to hen egg 3 3 106 anti-CD4 (GK1.5) and anti-CD8 (53-6.7, eBioscience, San Diego, lysozyme (HEL)-specific Ig (IgHEL) transgenic (tg) B cells on the CA)-coated SBM cells admixed with 7 3 106 purified splenic B cells from NOD background, which showed equivalent levels of proliferation the indicated strains. Splenic B cells were purified using the previously The Journal of Immunology 3 described (9) MACS negative-depletion system (Miltenyi Biotec, Bergisch vided by the manufacturer. Levels of gene transcripts were determined Gladbach, Germany), which routinely achieved .93% purity, as deter- with a PRISM7900 HT machine (Applied Biosystems, Foster City, CA) mined by flow cytometry. BM/B cell recipients were monitored weekly using triplicate q-rtPCR reactions containing 50–100 ng cDNA and 0.9 between 8 and 23 wk postreconstitution for T1D development. Upon mM of each primer in LightCycler-RNA SYBR Green I Master Mix disease onset or at the end of the incidence study (23 wk postreconsti- (Roche, Mannheim, Germany). Primers for each gene were designed such tution), spleens from recipient mice were assessed by flow cytometry to that one would bind across an exon–exon boundary (sequences in Sup- determine the percentage reconstitution of B cells and CD4 T cells using plemental Table I). The capacity of primers to quantify target genes by q- previously published methods and reagents (9). Flow cytometry was also rtPCR was initially established using triplicate sets of 4-fold serial dilu- used to establish reconstitution levels of lymphocytes expressing CD45.2 tions (1/1, 1/4, 1/16, 1/64, and 1/256) of NR4 or NOD spleen cDNA. or CD45.1 allotypic markers in NOD.Igmnull recipients using the 104 and Amplification efficiency of all primers ranged between 95.7 and 99.6% A20 Ab clones (BD Biosciences), respectively. (data not shown). Expression of target genes in B cell samples were nor- malized to the Hprt housekeeping gene. Mixed BM and CD4 T cell chimeras NOD.CD4null mice were lethally irradiated (two 650-rad split doses from Results 137 6 a[ Cs] source) and reconstituted i.v. with 5 3 10 anti-CD8 (53-6.7; Determining the mode of inheritance of the NOR-derived Chr. eBioscience)-coated SBM cells admixed with 5 3 106 purified CD4 T cells from the indicated strains. Splenic CD4 T cells were purified using the 4 gene(s) that dampens the diabetogenic activity of B cells previously described MACS negative-depletion strategy (32), which rou- In a previous study, we showed that, in contrast to the low incidence . tinely achieved 92% purity, as determined by flow cytometry. BM/CD4 of T1D exhibited by NR4 mice (25% by 30 wk), mice that were T cell recipients were monitored weekly between 8 and 20 wk post- reconstitution for T1D development. NOD/NOR heterozygous at this congenic region developed a higher incidence of disease that was not significantly different from Adoptive transfer of AI4 T cells NOD homozygous littermates (71 versus 92% by 30 wk, respec- Downloaded from At 6–8 wk of age, the indicated recipient mice were sublethally irradiated tively; NS, log-rank test) (26). This result indicated that a NOR- (600 rad from a [137Cs] source) and injected i.v. with 5 3 106 NOD. derived allele(s) in the original full-length (FL) Chr. 4 congenic null 6 Rag1 .AI4ab-tg splenocytes (∼1 3 10 AI4 T cells). T1D development interval must be homozygous to confer significant resistance to was monitored daily for up to 2 wk posttransfer. T1D. To establish whether the gene(s) decreasing the diabetogenic B cell-proliferation assays capacity of B cells in the NR4 congenic region was consistent with this mode of inheritance, we compared the T1D incidence of http://www.jimmunol.org/ B cells were purified from pooled spleens of three mice of the indicated null strains (6–8 wk of age) using the EasySep Mouse B cell enrichment kit, lethally irradiated NOD.Igm mice reconstituted with SBM plus followed by separation using an EasySep magnet, according to the man- MACS-purified B cells from NOD, NR4, or (NOD3NR4)F1 ufacturer’s instructions (StemCell Technologies, Vancouver, BC, Canada). donors (Fig. 1). The T1D incidence of NOD.Igmnull recipients . This method routinely achieved 98% purity of B cells, as determined by reconstituted with NR4 B cells differed significantly from those flow cytometry. Triplicate aliquots of 1 3 105 B cells were cultured in complete RPMI 1640 medium with 10 mg/ml LPS (Sigma-Aldrich, St. reconstituted with NOD B cells after 23 wk (31.3% [n = 16] , Louis, MO) or 10 mg/ml AffiniPure goat anti-mouse IgM F(ab9)2 frag- versus 69.6% [n = 23]; p 0.05, log-rank test). However, T1D ments (Jackson ImmunoResearch Laboratories, West Grove, PA) and 5 mg/ development did not differ in a statistically significant manner in ml CD40-specific mAb HM40-3 (BD Biosciences, San Jose, CA), alone or recipients reconstituted with (NOD3NR4)F1 or NOD B cells. in combination. Control wells contained no stimulatory agents. Prolifera- tion of cells was assessed by addition of 1 mCi/well [3H]thymidine This indicated that a NOR-derived allele(s) on distal Chr. 4 must by guest on October 2, 2021 (Amersham Pharmacia Biotech, Uppsala, Sweden) in the final 24 h of a 72- be homozygous to significantly suppress diabetogenic B cell de- h incubation period using previously described methods (12). velopment. Differences in T1D susceptibility between cohorts was unlikely due to expansion of contaminating CD4 or CD8 T cells Microarrays and quantitative real-time PCR from the purified donor B cell preparations. This was made evi- null For microarrays, B cells from three independent lots of seven or eight dent in two lethally irradiated NOD.Igm mice (expressing pooled spleens from NOD or NR4 mice, respectively, were purified using CD45.1) reconstituted with SBM plus B cells purified from NOD. a previously described MACS negative-depletion strategy (9). B cells from 7 CD45.2 congenic mice, whose spleens, 6 wk postengraftment, each lot were resuspended at 1 3 10 cells/ml for 2 h in complete RPMI + 1640 medium, alone or with 10 mg/ml AffiniPure goat anti-mouse IgM F contained 1.2 and 6.0% CD45.2 CD4 T cells, 2.4 and 6.4% (ab9)2 fragments (Jackson ImmunoResearch Laboratories). Total RNA was then purified from B cells using TRIzol reagent (Invitrogen, Carlsbad, CA) and subjected to one-cycle linear amplification, biotin labeling, and frag- mentation, according to the protocol provided, by the GeneChip Expres- sion 39-Amplification One-cycle kit (Affymetrix, Santa Clara, CA). Ten micrograms of biotin-labeled and fragmented cRNA from two BCR- stimulated and three unstimulated NOD B cell samples plus three BCR- stimulated and three unstimulated NR4 B cell samples were hybridized to Mouse Genome 430 2.0 Chips (Affymetrix) for 16 h at 45˚C and scanned with a GeneChip Console Scanner (Affymetrix). All microarray data were submitted to the Gene Expression Omnibus public database (http://www. ncbi.nlm.nih.gov/geo/) under accession number GSE37294. Analysis of expression data was performed using GeneSpring 12.0 software (Agilent Technologies, Santa Clara, CA). All samples were normalized by Robust Multiarray Analysis on a per-gene basis to the median of the total samples. Differential expression of genes was determined using one- or two-way ANOVA tests, with Benjamini and Hochberg false detection rate correction (q-value) to account for multiple testing. Gene symbols used in the figures and tables are those prescribed by the Mouse Genomic Nomenclature FIGURE 1. NOR-derived allele(s) on distal Chr. 4 must be homozygous Committee. For quantitative real-time PCR (q-rtPCR), B cells were puri- to significantly suppress diabetogenic B cell development. T1D incidence fied from three pools of three spleens each from the indicated strains of Igmnull mice (independent of microarray samples), using the aforementioned was monitored weekly for up to 23 wk in cohorts of female NOD. EasySep Mouse B cell enrichment kit. RNA was then immediately mice that were lethally irradiated at 4–5 wk of age and reconstituted with 6 6 extracted from each sample using TRIzol reagent (without culture) and 3 3 10 SBM admixed with 7 3 10 splenic B cells from NOD (n = 23), transcribed into cDNA using M-MuL V reverse transcriptase (New Eng- NR4 (n = 16), and (NOD3NR4)F1 (n = 15) mouse strains. The p values land BioLabs, Ipswich, MA), via the first-strand synthesis protocol pro- for comparison of incidence curves by log-rank analysis are shown. 4 GENETIC CONTROL OF DIABETOGENIC CD4 T AND B CELLS

CD45.2+ CD8+ T cells, and 99.3 and 99.2% CD45.2+ B cells, NR4 strains but not each other (Fig. 2C). In contrast, the NR4S5 respectively (data not shown). mouse strain developed T1D at a similar incidence and rate (86%, n = 27) as did NOD mice (Fig. 2B), indicating that the central Generation of novel subcongenic lines of the NR4 mouse strain section of the NR4 congenic region between D4Mit72 and To facilitate identification of the NOR-derived distal Chr. 4-region rs13477999 (Fig. 2A), which encompasses the Idd11 locus (27), gene(s) controlling T1D resistance, as well as the decreased path- does not contain a NOR origin gene conferring resistance to T1D ogenic activity of B cells in the original FL NR4 congenic strain, development. Our observation that none of the subcongenic strains subcongenic lines were generated as outlined in Materials and showed the same level of T1D protection as did the original FL Methods. This resulted in the generation of five novel subcongenic NR4 congenic strain provides evidence of more than one distal NR4 mouse lines (referred to as NR4S1–S5) containing smaller Chr. 4-region gene contributing to disease resistance in NOR homozygous NOR-derived regions spanning different sections of mice. This was also confirmed by the fact that there was no the original FL 44.31-Mb congenic interval. The genetic bound- overlap in the NOR-derived regions of two subcongenic strains aries of the NOR-derived regions in the NR4S1–S5 subcongenic showing T1D resistance: NR4S4 and NR4S2 (Fig. 2A, 2B). strains, relative to the original NR4 strain and previously identified Therefore, these two subcongenic strains define at least two loci Idd9 and Idd11 genes (27, 28), are illustrated in Fig. 2A. containing NOR T1D-resistance genes on Chr. 4, which span the intervals between the markers rs3674285 (104.99 Mb) and T1D susceptibility in NR4 subcongenic mouse lines D4Mit203 (129.25 Mb) and rs13477999 (135.18 Mb) and To refine the NOR-derived region(s) on Chr. 4 responsible for D4Mit127 (149.30 Mb). These regions partially overlap with the conferring protection from T1D, the natural incidence of disease in T1D-protective Idd9.1 and Idd9.2 loci, respectively, which were female mice of the NR4S1–S5 subcongenic lines was compared found to be present in the B10 strain (28). However, given that the Downloaded from with that of the NOD and the original FL NR4 congenic strain T1D-protective genes in the distal region of NOR Chr. 4 originally through 40 wk of age (Fig. 2B). More than 90% of female NOD derive from the BKS strain, the possibility cannot be excluded that mice developed T1D by 40 wk of age compared with only 23% of their identity differs from those responsible for the Idd9.1 and/or the NR4 stock (p , 0.0001, log-rank analysis, Fig. 2C), consistent Idd9.2 effects. with the results of a previous study on the same strains performed in an independent animal facility (26). The NR4S1 (65%, n = 33), Diabetogenic capacity of B cells in NR4 subcongenic mouse http://www.jimmunol.org/ NR4S2 (63%, n = 29), NR4S3 (73%, n = 36), and NR4S4 (59%, n lines = 31) mouse strains all showed an intermediate incidence of T1D We next refined the location of the distal Chr. 4-region gene(s) (Fig. 2B) that differed significantly from both the NOD and FL suppressing the diabetogenic capacity of NOR B cells. T1D de- by guest on October 2, 2021

FIGURE 2. Generation of NR4 subcongenic strains reveals the presence of more that one T1D-resistance gene in the distal Chr. 4 region of NOR mice. (A) Schematic illustration showing boundaries of NOR-derived congenic regions on Chr. 4 in the NR4 and NR4S1–S5 subcongenic mouse lines. Ori- entation of Chr. 4 is marked by a circle symbolizing the centromere. Rectangular boxes within the broken lines represent NOR-derived congenic intervals. Boundaries of congenic intervals were determined by the polymorphic microsatellite and SNP markers shown. Markers are ordered according to their physical position in megabases according to Ensembl (http://www.ensembl.org/), with the exception of WEHI markers, whose positions were obtained from Ref. 27. Colors within the congenic regions mark NOR genome of NOD (black), B6 (light green), or DBA/2 (dark green) origin, as previously defined by Reifsnyder et al. (26). Regions of unknown origin between markers are represented by hatched lines. For comparison, the most current positions of the Idd9.1 (116.08–134.49 Mb), Idd9.2 (144.97–149.10 Mb), and Idd11 (129.63–129.64 Mb) loci are shown (27, 28). (B) Cumulative incidence of T1D in female cohorts of NOD (n = 33), NR4 (n = 29), NR4S1 (n = 33), NR4S2 (n = 29), NR4S3 (n = 36), NR4S4 (n = 31), and NR4S5 (n = 27) mice through to 40 wk of age. (C) The p values from comparisons of T1D incidence curves using log-rank analysis. The Journal of Immunology 5 velopment was assessed in NOD.Igmnull mice that were lethally and nondiabetic mice of the various recipient groups (Table I). The irradiated and reconstituted with SBM plus splenic B cells from strain of B cell donor contributed to variance in B cell (p = 0.03), the T1D-resistant subcongenic strains NR4S1–S4 (Fig. 3A). Dis- but not CD4 T cell (NS), reconstitution levels in the different ease incidence in these groups up to 23 wk was compared with recipient groups. However, this effect was not strong given that control cohorts of NOD.Igmnull recipients reconstituted with SBM Bonferroni posttests did not identify a significant difference in the and splenic B cells from NOD or FL NR4 strains. Mirroring our percentages of B cells between recipients of any two specific previous study (30), NOD.Igmnull mice reconstituted with NOD B cell donors in either the diabetic or nondiabetic groups. Added B cells developed a high incidence of T1D (68%, n = 33), which to this, it is unlikely that differences in T1D susceptibility between was reduced by approximately half (36%, n = 30) in recipients of recipient groups were a consequence of variance in B cell re- NR4 B cells (Fig. 3A) despite similar levels of B cell reconsti- constitution levels, because in all transfers, there was no signifi- tution between groups (Table I). Compared with those receiving cant difference in B cell proportions in spleens of diabetic versus NOD B cells, there was a significant reduction in the incidence of nondiabetic mice (Table I; NS, two-way ANOVA, Bonferroni T1D shown by NOD.Igmnull mice reconstituted with either NR4S3 posttests). Furthermore, recipient groups with a higher incidence (20%, n = 10) or NR4S4 (35%, n = 17) B cells (Fig. 3A) In ad- of T1D (i.e., NOD, NR4S1, and NR4S2 donors) did not neces- dition, these two groups of recipients developed a delayed onset of sarily have higher mean levels of B cell reconstitution than did T1D (NR4S3: 21.2 6 0.9 wk, NR4S4: 21.5 6 0.5 wk) compared those with a lower incidence (i.e., NR4, NR4S3, and NR4S4 with the NR4 B cell-reconstituted groups (18.5 6 0.6 wk), sug- donors) (NS, Student t test). gestive of an enhanced level of protection (Fig. 3B). The differ- The results of these incidence studies indicate that the main ence in mean onset time of T1D was significant between NOD and NOR-derived gene(s) responsible for dampening the diabetogenic NR4S4 B cell-reconstituted recipients (p = 0.006, Student t test). activity of B cells lies within the 6.22-Mb overlap of the NR4S3 and Downloaded from Given that only two NOD.Igmnull mice containing NR4S3 B cells NR4S4 congenic regions (i.e., between rs13477999 and D4Mit32, developed T1D in this study, the difference in mean time of dis- Fig. 2A). This region of Chr. 4 has not been previously identified ease onset compared with NR4 B cell recipients was not statisti- to contain an Idd locus. Thus, we now designate this newly cally significant; however, the trend for delay was still strong identified locus on distal Chr. 4 differentially controlling the de- (p = 0.08). In contrast, NOD.Igmnull mice reconstituted with velopment of diabetogenic B cells in NOD and NOR mice as B cells from NR4S1 (50%, n = 16) and NR4S2 (54%, n = 13) Idd25. However, it is interesting that the protective B cell phe- http://www.jimmunol.org/ subcongenic strains showed no significant decrease in T1D inci- notype mediated by the Idd25 resistance variant seems to be ab- dence compared with recipients of NOD B cells (Fig. 3A). rogated by a combined interaction between at least two other NOR Two-way ANOVA tests were performed to examine differences genes within the intervals bound by D4Wehi6 to D4Mit69 and in B cell or CD4 T cell reconstitution levels in spleens of diabetic D4Mit179 to D4Mit127, which are both present in the NR4S1 strain but do not coexist in the NR4S3 and NR4S4 strains (Fig. 2A). The original FL NR4 congenic region also dampens diabe- togenic B cell activity, despite containing all of the above-mentioned subregions. Thus, it is likely that another NOR origin gene(s) in by guest on October 2, 2021 the most proximal region, bound by the markers rs3674285 and D4Mit203, also contributes to the dampened diabetogenic activity of B cells. However, this NOR gene(s) does not mediate a reduc- tion in pathogenic B cell activity when inherited alone, given that B cells from the NR4S2 congenic stock are as diabetogenic as those from standard NOD mice (Fig. 3A). Hyperresponsiveness of B cells in NR4 subcongenic mouse lines B cells from FL NR4 mice undergo significantly higher levels of proliferation upon BCR or BCR plus CD40 stimulation than do those from NOD mice (30). This difference in responsiveness to stimulation was evident in B cells from spleen (FO and MZ subsets) or lymph nodes (FO subset only). To identify the genetic region(s) within the original FL NR4 congenic region that is re- sponsible for this B cell hyperproliferative phenotype, we sub- jected purified splenic B cells from the NR4S1–5 congenic strains to in vitro stimulation with anti–IgM-F(ab9)2 fragments or anti- CD40 Abs alone or in combination, or LPS. Consistent with our previous study, NR4 B cells proliferated significantly more than FIGURE 3. B cells from NR4S3 and NR4S4 subcongenic mouse strains those from NOD mice when stimulated with anti–IgM-F(ab9)2 or exhibit decreased diabetogenic capacity. (A) T1D incidence was monitored anti–IgM-F(ab9)2 plus anti-CD40 (1.3 6 0.07- and 1.5 6 0.1-fold null for up to 23 wk in cohorts of female NOD.Igm mice that were lethally higher, respectively, Fig. 4A, 4B). This difference was probably 3 6 irradiated (1100 rad) at 4–5 wk of age and reconstituted with 3 10 SBM the direct consequence of variability in signaling downstream of admixed with 7 3 106 purified splenic B cells from NOD (n = 33), NR4 the BCR, given the observed equivalent responses between B cells (n = 30), NR4S1 (n = 16), NR4S2 (n = 13), NR4S3 (n = 10), or NR4S4 (n = 17) strains. *p , 0.05, **p , 0.01 compared with NOD B cell stimulated through CD40 alone or through the TLR4 receptor with reconstituted group (log-rank analysis). (B) Graph comparing mean time of LPS (Fig. 4C, 4D). Of the subcongenic B cells, only those isolated disease onset in diabetic NOD.Igmnull mice reconstituted with NOD (n = from the NR4S1 stock exhibited the hyperproliferative phenotype 23), NR4 (n = 11), NR4S3 (n = 2), or NR4S4 (n = 6) B cells. The p values of the FL NR4 strain in response to stimulation with BCR, with or between groups are shown (Student t test). without CD40 costimulation (1.5 6 0.1- and 1.4 6 0.2-fold higher 6 GENETIC CONTROL OF DIABETOGENIC CD4 T AND B CELLS

Table I. Levels of B220+ and CD4+ lymphocytes in SBM plus B cell-reconstituted NOD.Igmnull recipients

B Cell Donor Used Diabetic (%; Mean 6 SEM) Nondiabetic (%; Mean 6 SEM) To Reconstitute NOD.Igmnull Micea n B220+b CD4+c n B220+b CD4+c NOD 22 41.8 6 6.9 44.9 6 6.0 11 37.8 6 3.4 40.8 6 2.4 NR4 11 43.2 6 8.0 29.6 6 4.9 19 47.8 6 3.7 28.9 6 2.8 NR4S1 8 40.6 6 3.6 39.4 6 3.0 8 37.3 6 3.9 31.8 6 3.2 NR4S2 7 26.0 6 4.3 38.5 6 3.7 6 23.3 6 2.6 44.2 6 3.0 NR4S3 2 31.3 6 7.4 34.9 6 7.2 8 22.6 6 1.3 38.9 6 3.4 NR4S4 6 28.0 6 4.3 43.4 6 3.9 11 31.2 6 2.5 36.0 6 2.8 aNOD.Igmnull mice were lethally irradiated at 4–5 wk of age and reconstituted with 3 3 106 SBM and 7 3 106 splenic B cells from the indicated strain. Proportions of B220+ B cells and CD4+ T cells in spleens of each recipient mouse were determined by flow cytometry after T1D onset or at the end of the study (23 wk postreconstitution). bTwo-way ANOVA test of B220+ reconstitution dataset did not find significant interaction between diabetes status and B cell donor. Furthermore, diabetes status did not contribute significantly to variation in B220+ reconstitution. In contrast, the strain of B cell donor was found to significantly contribute to variation in B220+ reconstitution (p = 0.03). However, Bonferroni posttests did not identify a significant difference between any two specific B cell donors in diabetic or nondiabetic groups. cTwo-way ANOVA test of the CD4+ T cell reconstitution dataset did not find significant interaction between disease status and B cell donor. In addition, neither diabetes status nor B cell donor contributed significantly to variation in CD4 T cell reconstitution. Downloaded from proliferation than NOD, respectively, Fig. 4A, 4B). NR4S4 and and D4Mit179 and D4Mit127 that are not coexistent in the NR4S4 NR4S3 B cells did not exhibit a hyperproliferative response to and NR4S3 lines. Interestingly, these were also the two regions that BCR stimulation. This suggested that the hyperproliferation phe- combined to lessen the ability of the NOR gene(s) within the notype of NR4S1 B cells is the result of an interaction between rs13477999–D4Mit32 region to suppress diabetogenic B cell ac- NOR genes located in two regions bound by D4Wehi6 and D4Mit69 tivity (Figs. 2A, 3A). These results dissociate the induction of http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 4. Hyperresponsive and anergy phenotypes are not necessary for decreased diabetogenic activity of NR4 B cells. (A–D) To test hyper- responsiveness, B cells were purified from pooled spleens of three 6–8-wk-old female NOD, NR4, NR4S1, NR4S2, NR4S3, NR4S4, and NR4S5 mice. 5 Triplicate aliquots of 1 3 10 B cells were stimulated in culture with 10 mg/ml anti–IgM-F(ab9)2 (A), 10 mg/ml anti–IgM-F(ab9)2 and 5 mg/ml anti-CD40 (B), 5 mg/ml anti-CD40 (C), or 10 mg/ml LPS (D). Proliferation over the final 24 h of a 72-h incubation period was measured by [3H]thymidine incor- poration and is shown as the mean (6 SEM) fold change compared with NOD B cells in two to four independent experiments for each congenic strain. *p , 0.05, **p , 0.01, ***p , 0.001, compared with NOD B cells (one-way ANOVA, Tukey posttest). (E–H) To test anergy induction, B cells were purified from pooled spleens of three 6–8-wk-old female IgHEL single-tg or IgHEL/sHEL double-tg mice with NOD, NR4, (NOD3NR4)F1, (NR4S23NR4)F1, (NR4S33NR4)F1, (NR4S43NR4)F1, or (NR4S53NR4)F1 genetic backgrounds. Triplicate aliquots of 1 3 105 B cells were stimulated in culture with 10 mg/ml anti–IgM-F(ab9)2 (E), 10 mg/ml anti–IgM-F(ab9)2 and 5mg/ml anti-CD40 (F), 5 mg/ml anti-CD40 (G), or 10 mg/ml LPS (H). Proliferation over the final 24 h of a 72-h incubation period was measured by [3H]thymidine incorporation. Bars represent the mean (6 SEM) fold change in proliferation of B cells from IgHEL/sHEL double-tg mice compared with single IgHEL counterparts in two to four independent experiments for each subcongenic background, whereas the experiments using NOD, NR4, and (NOD x NR4)F1 background controls were performed five, four, and eight times, respectively. *p , 0.05 compared with NOD and (NOD3NR4)F1 background group (one-way ANOVA, Tukey posttest). The Journal of Immunology 7 a stronger BCR-signaling response from the ability of a NOR- B cells. Moreover, the clear difference in responsiveness of self- derived Chr. 4 gene(s) to dampen diabetogenic B cell activity. reactive B cells in mice homozygous for the NR4S3 subcongenic region compared with those carrying the FL NR4 congenic seg- B cell anergy in NR4 subcongenic mouse lines ment indicates that a NOR-derived region between D4Mit179 and Our previous study demonstrated that homozygous expression of D4Mit127 (Fig. 2A) must also contribute to the correction of B cell the FL NOR-derived Chr. 4 congenic region on the NOD back- anergy induction. The gene within this region can only mediate its ground resulted in the correction of faulty B cell anergy induction effect on B cell anergy when inherited in combination with a ho- in the IgHEL/sHEL tg model (30). Based on the observation that mozygous NOR-derived gene(s) in the NR4S2 congenic interval. the NOR-derived FL NR4 congenic interval must be inherited This conclusion is supported by the finding that strains homozygous in a homozygous manner to significantly suppress diabetogenic for NOR genes in the distal region alone (i.e., NR4S1 and NR4S4 B cell development (Fig. 1), an array of F1 hybrids was made to hybrid strains) displayed defects in B cell anergy induction that fix the various NR4 subcongenic intervals to homozygosity to were indistinguishable from that in standard NOD mice (Fig. 4E, assess their effects on B cell anergy induction. The strategy con- 4F). B cells showing partially corrected tolerance induction in the sisted of crossing the previously developed FL NR4 congenic NR4S2 line were as diabetogenic as those from standard NOD strain expressing both IgHEL and sHEL transgenes (30) with the mice, whereas NR4S3 and NR4S4 B cells with dampened patho- NR4S2, NR4S3, NR4S4, or NR4S5 subcongenic mouse lines. The genic activity exhibited anergy defects equivalent to those from resulting hybrid progeny were homozygous at the NOR-derived NOD mice (Figs. 3A, 4E, 4F). Thus, it appears that homozygous region shared between the NR4 and subcongenic strain but het- expression of NOR genes in both the rs3674285 to D4Mit203 and erozygous throughout the remainder of the FL congenic region, D4Mit179 to D4Mit127 intervals, which combinatorially restore while also expressing a combination of hemizygous IgHEL and anergy-induction processes, does not independently suppress dia- Downloaded from sHEL transgenes. Anergy was assessed by measuring the prolif- betogenic B cell activity but may augment this process. erative response of B cells from IgHEL/sHEL double-tg mice versus IgHEL single-tg mice from each hybrid cross following Differentially expressed genes from microarray analysis of NOD versus NR4 B cells in vitro stimulation with LPS or anti–IgM-F(ab9)2 fragments and anti-CD40 Abs, alone or in combination (Fig. 4E–H). The occu- Microarrays were conducted on B cells purified from spleens of pancy of HEL on the majority of BCRs of B cells of double-tg, but NOD and NR4 mice to highlight differentially expressed genes http://www.jimmunol.org/ not single-tg, mice precluded us from using specific Ag stimula- within the distal Chr. 4 locus. B cells were either cultured in media tion (17). B cells from NOD, (NOD3NR4)F1, and NR4 back- alone (unstimulated) or with BCR cross-linking anti–IgM-F(ab9)2 ground mice expressing the same transgenes were used as fragments (stimulated) for 2 h before RNA was extracted for controls. Consistent with our previous study (30), HEL-specific transcript analysis. Several significant expression differences B cells in NOD mice were not effectively anergized when ma- were noted between NOD and NR4 B cell samples using a one- turing in the presence of sHEL, and thus proliferated at similar way ANOVA test, and, as expected, most of these genes (24/32 levels as did naive B cells in culture, irrespective of the stimuli mapped probes) were situated within the distal Chr. 4 region (Fig. 4E–H). Conversely, a significant reduction in proliferation encompassed by the FL NR4 congenic interval (Fig. 5A, Sup- by guest on October 2, 2021 was observed in NR4 B cells from IgHEL/sHEL double-tg mice plemental Fig. 1A, 1B, Supplemental Table II). These 24 dif- compared with IgHEL single-tg mice in response to anti–IgM-F ferentially expressed genes represented 1.8% of the total probe (ab9)2 with or without anti-CD40 (Fig. 4E, 4F). Such a reduction sets (n = 1299) that detect transcripts within the Chr. 4 congenic was not evident when B cells were stimulated with LPS or anti- region. Interestingly, none of the genes that were differentially CD40 alone (Fig. 4G, 4H). B cells from (NOD3NR4)F1 mice expressed between NOD and NR4 B cells changed significantly expressing IgHEL and sHEL transgenes showed a similar defect in upon 2 h of BCR cross-linking (Supplemental Fig. 1A, 1B). B cell anergy as did NOD background mice (Fig. 4E, 4F), high- Nevertheless, polymorphic genes within the distal Chr. 4 region lighting the requirement for homozygous expression of NOR had the capacity to differentially alter stimulation-responsive alleles in the FL NR4 congenic region to correct this immuno- genes located at other sites, given that more than approxi- regulatory process. Of the hybrid subcongenic progeny, only mately half of the genes that changed significantly after BCR B cells homozygous for the NR4S2 region exhibited a significant cross-linking ($2-fold change and q , 0.1, one-way ANOVA) decrease in proliferation following stimulation with anti–IgM-F were not shared by the NOD and NR4 strains (Supplemental Fig. (ab9)2 alone or in combination with anti-CD40 after having de- 1A, Supplemental Table III). veloped in the presence, as opposed to the absence, of their cog- Of the 24 genes within distal Chr. 4 that were deemed to be nate Ag (Fig. 4E, 4F). This decrease (∼30%) was not as distinct as differentially expressed between NOD and NR4 B cells (Fig. 5A), that observed for IgHEL/sHEL double-tg B cells from the FL NR4 11 were chosen for further examination by q-rtPCR because of strain (∼50%). This indicated that the NR4S2 subcongenic region their potential involvement in tolerance mechanisms (e.g., cell contributes to a partial decrease in the responsiveness of self- death), BCR signaling, and lymphocyte function or as a conse- reactive B cells but is not the sole region affecting this pheno- quence of large differences in expression levels between mouse type in the original FL NR4 congenic mouse strain. In this regard, strains. Differential expression of these genes was analyzed in an it is worth highlighting the result of tg B cells derived from hybrid independent set of cDNA derived from purified NOD and NR4 mice homozygous for the NR4S3 region. Despite also being ho- splenic B cells that were not subjected to culture. Consistent with mozygous throughout the NR4S2 region, NR4S3 B cells showed the results of the microarrays, the expression of the selected genes equivalent levels of proliferation in response to anti–IgM-F(ab9)2 Lrp8, Ccdc28b, Ccdc21, Ephb2, and Txlna were higher in B cells and anti–IgM-F(ab9)2 plus anti-CD40 Abs, whether they matured from NR4 mice than from NOD mice, whereas Zfp69, SIII, Tceb3, in the presence or absence of their cognate self-Ag, similar to Padi2, 2610305D13Rik, and mTOR were more highly expressed NOD background mice (Fig. 4E, 4F). This result implies that the by B cells from NOD mice compared with NR4 mice (Fig. 5B). NOR region differentiating the NR4S2 and NR4S3 strains (be- The only gene that was not consistent with the results of the tween D4Mit203 and D4Mit32, Fig. 2A) contains a gene(s) that microarrays was Ptprf, which did not show a significant difference contributes to the aberrant induction of anergy in self-reactive in expression between NOD and NR4 B cells by q-rtPCR. 8 GENETIC CONTROL OF DIABETOGENIC CD4 T AND B CELLS

B cells from the subcongenic NR4S3 and NR4S4 strains showed a decrease in diabetogenic activity (Fig. 3A). Of the genes con- firmed to be differentially expressed between NOD and NR4 B cells by microarray, only Tceb3, Ephb2, and Padi2 are located within the overlap of the NR4S3 and NR4S4 congenic regions (between rs13477999 and D4Mit32, Fig. 2A). Therefore, we assessed the differential expression of these genes by q-rtPCR in NR4S3 and NR4S4 B cells compared with those from NOD and NR4 mice (Fig. 5C). Consistent with their expression in NR4 B cells, we observed a significant decrease in the expression of Padi2 and a significant increase in the expression of Ephb2 in NR4S3 and NR4S4 B cells compared with those from NOD mice. Although a significant decrease in Tceb3 expression was also observed in NR4S3 B cells relative to NOD B cells, this difference was not evident in the comparison with NR4S4 B cells. Confir- mation of differential Padi2 and Ephb expression within the NR4S3 and NR4S4 mouse strains implicates these genes as po- tential Idd25 candidates for the control of diabetogenic B cell activity in NOD mice. Downloaded from Subcongenic analysis of the NR4 region at the level of CD4 T cells We previously showed that the FL NR4 congenic interval also conferred T1D resistance at the level of CD4 T cells (32). The complex interaction of the NR4-region genes described above raised the possibility that some of the B cell phenotypes are http://www.jimmunol.org/ influenced to some extent by the quality of CD4 T cell help. Conversely, B cell phenotypes controlled by polymorphic NR4- region genes could also differentially effect activation of diabe- togenic CD4 T cells. In an independent effort carried out at TJL, two additional NR4 subcongenic lines (NR4S6 and NR4S7) were generated to further define the critical interval that regulates the diabetogenic activity of CD4 T cells (Fig. 6A). T1D-incidence studies performed at TJL demonstrated that both subcongenic by guest on October 2, 2021 lines were less susceptible to disease than were standard NOD mice (Fig. 6B). However, neither subregion alone was sufficient to confer the same level of T1D protection provided by the FL NR4 congenic interval. These results are consistent with the observa- FIGURE 5. Differentially expressed genes in NOD versus NR4 B cells tions at the Garvan Institute and confirm that at least two genes A within the distal Chr. 4 locus. ( ) Differentially expressed genes between NOD contribute to T1D resistance in NR4 mice. , and NR4 splenic B cells (q 0.05 false detection rate, one-way ANOVA), We next asked whether any of the NOR-derived Chr. 4 sub- which map within the region of distal Chr. 4 and distinguish the two strains, regions could suppress the diabetogenic activity of CD4 T cells. We were determined using Affymetrix Mouse 430 2.0 microarray chips after a 2-h b culture in media alone (n = 3 for NOD, n =3forNR4)orwith10mg/ml of anti- previously showed that the ability of adoptively transferred cell- autoreactive AI4-abTCR-tg CD8+ effectors to induce rapid T1D IgM-F(ab9)2 fragments (n =2forNOD,n = 3 for NR4). None of the Chr. 4 gene probes showed significant changes in expression with stimulation (see Sup- onset in sublethally irradiated NOD recipients was enhanced by plemental Fig. 1A, 1B). Therefore, each line represents the mean fold change in help provided by pathogenic CD4 T cells (32). Therefore, the expression between combined NOD (n =5)versusNR4(n = 6) B cell groups susceptibility of NR4S6 and NR4S7 mice to AI4-induced T1D for each differentially expressed gene probe plotted against its physical position development was examined to initially evaluate the pathogenic (in megabases) according to the Ensembl database. The location of the con- helper activity of their CD4 T cells. As shown in Fig. 6C, the genic region in NR4 mice is demarcated by the thick gray line. All differentially ability of AI4 T cells to induce the rapid onset of T1D was sig- expressed genes reside in genomic areas that are B6 derived in NR4 mice, with nificantly reduced in NR4S6 mice, but not in NR4S7 mice. These the exception of Padi2 (underlined), which is in a DBA/2-derived region. (B) results suggested that CD4 T cells from the NR4S6 stock were less Differentially expressed genes in microarray analyses were confirmed by performing q-rtPCR of selected genes using cDNA from three independent lots diabetogenic than were those from the standard NOD mice. This of purified NOD and NR4 splenic B cells that were not subjected to culture. q- was indeed the case, as demonstrated in a combined BM/CD4 rtPCR was conducted on triplicate aliquots of cDNA and normalized to ex- T cell-reconstitution system. NR4S6 CD4 T cells had a reduced null pression of the housekeeping gene Hprt. Data are presented as mean fold ability to promote T1D development in NOD.CD4 recipients change differences in expression in three NR4 samples relative to the mean of compared with those from standard NOD mice (Fig. 6D). In the NOD B cell samples. *p , 0.05, **p , 0.01, ***p , 0.001 (Student t test). contrast, NR4S7 and NOD CD4 T cells had comparable T1D- (C) q-rtPCR was performed for the Tceb3, Ephb2,andPadi2 genes using cDNA promoting capacity in NOD.CD4null recipients. The NR4S6 con- obtained from three independent lots of splenic B cells from 6–8-wk-old female genic region partially overlaps that in NR4S5 mice (Figs. 2A, 6A). NOD, NR4, NR4S3, and NR4S4 mice. Gene expression was normalized to Because the congenic interval in the NR4S5 stock did not confer Hprt. Data for each strain are presented as the mean fold change difference in T1D resistance (Fig. 2B), it appears that the NOR-derived region expression relative to the mean of the NOD B cell samples. *p , 0.05, **p , 0.01, ***p , 0.001, compared with NOD (one-way ANOVA, Tukey posttest). regulating diabetogenic CD4 T cells can be defined in a 24.26-Mb region between the markers rs3674285 and D4Mit203. This region The Journal of Immunology 9

FIGURE 6. Genes within the NR4S6 subcongenic region regulate T1D development by modulating the pathogenic activity of CD4 T cells. (A) Polymorphic microsatellite and SNP markers were genotyped to define the boundaries of the NR4S6 and NR4S7 subcongenic lines developed at TJL. Colors within the congenic regions mark NOR genome of NOD (black), B6 (light gray), or DBA/2 (dark gray) origin. For comparison, the most current positions of the Idd9.1 (116.08–134.49 Mb), Idd9.2 (144.97–149.10 Mb), and Idd11 (129.63–129.64 Mb) loci are shown (27, 28). (B) Cumulative incidence of T1D in female cohorts of NOD (n = 60), NR4 (n = 11), NR4S6 (n = 34), and NR4S7 (n = 29) mice was monitored through to 30 wk of age at TJL. **p , 0.01, ***p , 0.001, compared with NOD group (log-rank analysis). (C) Five million NOD.Rag1null.AI4-tg splenocytes were transferred into the indicated sublethally (600 rad) irradiated female recipients (n = 10/group). T1D de- Downloaded from velopment was monitored daily for up to 2 wk post- transfer. **p , 0.01, compared with NOD recipients (log-rank analysis). (D) T1D development in NOD. CD4null females lethally irradiated (1300 rad) at 4–6 wk of age and reconstituted with equal numbers (5 3 106) of SBM and purified CD4 T cells from NR4S6

(n = 13), NR4S7 (n = 11), or NOD mice (n = 16) was http://www.jimmunol.org/ followed for 20 wk. *p , 0.05, compared with NOD CD4 T cell recipients (log-rank analysis).

partially overlaps with the previously defined Idd9.1 locus (Fig. (144.97–149.10 Mb) susceptibility loci, respectively, mapped us- 6A) (28). In the B cell studies, the same region was shown to ing B10 congenic regions on the NOD background (28). Although promote anergy induction. Collectively, these results suggest that the NOR and B10 strains may share T1D-resistance alleles at by guest on October 2, 2021 genes within this region may regulate T1D through modulating the these loci, the distinct origin of the NOR genome, which is pri- interplay between B and CD4 T cells. marily derived from NOD, B6, and DBA/2 strains, appears to also endow it with other novel genes regulating disease (23–26). Discussion Of the seven subcongenic strains described in this study, only the Congenic NOD mouse strains have facilitated the discovery of NR4S5 strain, with a NOR-derived congenic interval spanning various T1D-susceptibility genes in this model (e.g., b2 micro- from 128.60 to 135.19 Mb (Fig. 2A), did not contribute any globulin, Il2, and Ctla4) (24, 35, 36), which have provided im- protection from the development of T1D (Fig. 2B). This result was portant insights into the molecular and cellular mechanisms surprising, given that refinement of the locus housing the Idd11 influencing the human disease (21, 22). We previously described gene using NOD mice with B6-derived congenic regions showed the generation of the NR4 strain (26), which contained a 44.31-Mb that significant resistance to T1D was conferred within a 6.9-kb NOR-derived congenic region on distal Chr. 4 that decreased interval between 129.63 and 129.64 Mb that lies within a pre- susceptibility to T1D in NOD background mice by dampening the dicted gene of unknown function named Ak005651 (27). Given pathogenic activity of B and CD4 T cells (30, 32). To significantly that the NOR strain was found to have an identical sequence to B6 refine the location of genes contributing to these phenotypes, we mice over this 6.9-kb interval (27), we would have predicted the generated seven novel subcongenic NOD mouse lines (NR4S1– NR4S5 strain to show significant protection from disease. This S7) with truncated NOR-derived regions spanning different parts result may be suggestive of genetic drift in the B6-derived region of the original FL NR4 congenic interval (Figs. 2A, 6A). T1D- surrounding the Idd11 locus in NOR mice, which may possess incidence studies revealed that six of the seven subcongenic polymorphisms that decrease the penetrance of the protective strains displayed some protection from disease, although none Ak005651 allele. exhibited the same level of resistance seen in the original NR4 The other incentive for generating NR4 subcongenic mouse lines congenic mouse strain (Figs. 2B, 6B). This implied that the NOR- was to map the NOR gene(s) within the distal end of Chr. 4 that derived region contained more than one gene controlling T1D reduced the diabetogenic activity of B cells, as well as to determine resistance, which was confirmed by the fact that the NR4S2 and whether B cell hyperresponsiveness and the restoration of B cell NR4S4 lines containing nonoverlapping subcongenic regions anergy remained linked to this gene. Our experiments with the afforded similar levels of disease protection. The result of these subcongenic strains demonstrated that the original NR4 congenic studies mapped the NOR genes conferring T1D resistance within interval could be divided into four subregions (termed R1–R4) that the Chr. 4 locus to two regions: rs3674285 (104.99 Mb) to control different aspects of B cell function (Fig. 7). B cells from D4Mit203 (129.25 Mb) and rs13477999 (135.18 Mb) to the NR4S3 and NR4S4 subcongenic strains had a reduced ca- D4Mit127 (149.30 Mb) (Fig. 7). These partly overlap with the pacity to induce T1D in NOD.Igmnull mice compared with those previously described Idd9.1 (128.37–131.18 Mb) and Idd9.2 from standard NOD donors (Fig. 3A). This led us to hypothesize 10 GENETIC CONTROL OF DIABETOGENIC CD4 T AND B CELLS

trolling these associated phenotypes may have the capacity to modulate the diabetogenic activity of B cells. In this regard, it is worth noting that B cells from the NR4S1 strain, whose congenic interval also encompasses the R3 region (Fig. 2A), showed a similar diabetogenic capacity as those from standard NOD mice (Fig. 3A). The potential reason for this is that, in addition to R3, the congenic interval in NR4S1 mice contains two additional regions located between 129.64 and 135.18 Mb (R2) and between 141.40 and 149.30 Mb (R4) (Figs. 2A, 7). Our experiments showed that the combination of NOR genes within the R2 and R4 regions was necessary for B cells to display hyperresponsiveness to BCR stimulation (Fig. 4A, 4B). NR4S3 and NR4S4 B cells were not hyperresponsive because they only contained the R2 or R4 region, respectively (Figs. 2A, 7). From these experiments, we predict that the hyperresponsive phenotype encoded by NOR genes within R2 and R4 leads to increased diabetogenic activity of B cells, overriding the protective B cell phenotype encoded by R3. Despite containing the R2, R3, and R4 regions (and the hy- perresponsive phenotype), B cells from the original NR4 congenic mouse were less diabetogenic than were those from NR4S1 donors Downloaded from (Fig. 3A). This difference indicated that an additional region on NOR Chr. 4, between 104.99 and 129.25 Mb (R1) (Fig. 7), was capable of decreasing the diabetogenic activity of B cells. None- theless, it is important to note that genes within the R1 region are incapable of independently decreasing diabetogenic B cell activ- FIGURE 7. Model delineating the distinct segments controlling patho- http://www.jimmunol.org/ genic B and CD4 T cell function in the distal Chr. 4 locus of NOR mice. ity, as demonstrated by the high incidence of disease in NOD. mnull Results of our experiments in the NR4S1–S7 subcongenic strains identified Ig mice reconstituted with NR4S2 B cells (Fig. 3A). NOR four regions (R1–R4) on distal Chr. 4 that contain genes controlling genes within R1, as well as the R4 region, were shown to con- pathogenic B and CD4 T cell function. Markers demarcating the bound- tribute to the decreased responsiveness (i.e., anergy) of self- aries of each region are shown. T1D is independently controlled by NOR- reactive B cells to BCR stimulation (Fig. 4E, 4F). Such a find- derived genes in the R1 region and the R3–R4 region. The hyperresponsive ing may explain the capacity of an R1-region gene(s) to reduce the phenotype of NR4 B cells to BCR stimulation is the result of an interaction diabetogenic capacity of B cells in an augmentative, but not in- of genes within R2 and R4. The maintenance of anergy in NR4 B cells is dependent, fashion. controlled by genes within R1 and R4, with counter-balancing negative In addition to their effects on B cells, NOR genes in the R1 region regulation by R2. R3 contains the gene(s) responsible for dampening the by guest on October 2, 2021 were shown to contribute to T1D resistance by dampening the diabetogenic activity of NR4 B cells. The penetrance of this gene is inhibited by genes within R2 and R4, potentially by making B cells hy- pathogenic activity of CD4 T cells. This raised the possibility that perresponsive to BCR stimulation. Genes within R1 and R4 could also modulation of CD4 T cells by R1 genes may indirectly regulate contribute to the dampening of B cell diabetogenic activity in the presence B cell diabetogenic activity and anergy. We do not believe this to be of the R3 region, potentially through their effects on B cell anergy. Finally, the case, given that our previous experiments showed that the B cell the R1 region contains a gene(s) responsible for dampening the diabeto- anergy defect in NOD mice is an intrinsic trait controlled by genes genic function of NR4 CD4 T cells. For comparison, the most current within Chr. 1 and 4 loci (17) and, as previously mentioned, the R1 positions of the Idd9.1, Idd9.2, and Idd11 resistance loci (in megabases) region on its own had little effect on the diabetogenic activity of from B10 or B6 mice are shown (27, 28). B cells (Fig. 3A). Alternatively, although the congenic region in NR4S4 mice was shown to have a strong dampening effect on the that the principal gene(s) controlling the diabetogenic capacity of diabetogenic activity of B cells (Fig. 3A), a virtually identical B cells is encompassed within the overlap of the NR4S3 and congenic interval in the NR4S7 strain (Fig. 6A) had no effect on NR4S4 congenic regions. This subregion, termed R3, consists of CD4 T cell pathogenesis (Fig. 6C, 6D). From these studies, we a 6.22-Mb interval located between 135.18 and 141.40 Mb (Fig. conclude that the NOR distal Chr. 4 region contains distinct sets of 7). This region has not previously been associated with the de- resistance alleles that mediate independent effects on B and CD4 velopment of T1D and, thus, is likely to represent a distinct sus- T cells contributing to T1D. Inheritance of both sets of NOR genes ceptibility gene from those mapped to Idd9.1, Idd9.2, and Idd11 is likely to significantly affect the interplay between these cell loci, which we now designate as Idd25. Interestingly, the NOD. types, leading to greater protection from T1D. Igmnull recipients reconstituted with NR4S3 and NR4S4 B cells Through our analyses of subcongenic mouse strains, we man- also exhibited a delayed onset of disease compared with those aged to reduce the region containing a gene(s) controlling the reconstituted with NR4 B cells (Fig. 3B). Hence, the gene(s) diabetogenic activity of B cells from 44.31 Mb (containing 889 dampening diabetogenic B cell activity in the NR4S3 and NR4S4 genes) to a 6.22-Mb region (containing 152 genes) encompassing strains may have been separated from NOR genes with opposing the newly identified Idd25 locus contributing to this phenotype. effects that exist within the FL congenic region. The smaller size of this region makes it much more amenable to Importantly, neither NR4S3 nor NR4S4 B cells exhibited the sequencing, which has the capacity to reveal all SNPs between hyperresponsiveness (Fig. 4A, 4B) or the restoration of anergy NOD and NOR origin genes that may be responsible for the Idd25 phenotype (Fig. 4E, 4F) that characterize B cells from the NR4 effect regulating B cell diabetogenic function. In the meantime, strain (30). This confirmed that these phenotypes are not directly our microarray (Fig. 5A) and q-rtPCR studies (Fig. 5B, 5C) associated with the NOR gene(s) dampening diabetogenic B cell identified two promising Idd25 candidate genes that were differ- activity. Nevertheless, our experiments suggest that genes con- entially expressed between NOD and NR4 B cells in this 6.22-Mb The Journal of Immunology 11 region: Ephb2 and Padi2. EphB2 encodes a protein tyrosine kinase 7. Pescovitz, M. D., C. J. Greenbaum, H. Krause-Steinrauf, D. J. Becker, S. E. Gitelman, R. Goland, P. A. Gottlieb, J. B. Marks, P. F. McGee, receptor with the capacity to bind various Ephrin B ligands on A. M. Moran, et al; Type 1 Diabetes TrialNet Anti-CD20 Study Group. 2009. neighboring cells (37). Although the function of Ephb2 has been Rituximab, B-lymphocyte depletion, and preservation of beta-cell function. N. studied most extensively in the nervous system (38), a role for this Engl. J. Med. 361: 2143–2152. 8. Falcone, M., J. Lee, G. Patstone, B. Yeung, and N. Sarvetnick. 1998. receptor was more recently revealed in T cells, where it was B lymphocytes are crucial antigen-presenting cells in the pathogenic autoim- shown to potentially act as a costimulatory molecule that alters mune response to GAD65 antigen in nonobese diabetic mice. J. Immunol. 161: TCR-mediated signaling (39). As a result, mice containing a de- 1163–1168. 9. Serreze, D. V., S. A. Fleming, H. D. Chapman, S. D. Richard, E. H. Leiter, and letion of EphB2 exhibit defects in T cell development (40) and R. M. Tisch. 1998. B lymphocytes are critical antigen-presenting cells for the peripheral T cell responses (41, 42). As we also found in this initiation of T cell-mediated autoimmune diabetes in nonobese diabetic mice. J. Immunol. 161: 3912–3918. study, expression of EphB2 was reported for B cells, although its 10. Wheat, W., R. Kupfer, D. G. Gutches, G. R. Rayat, J. Beilke, R. I. Scheinman, function in this population remains to be elucidated (37). Padi2 and D. R. Wegmann. 2004. Increased NF-kappa B activity in B cells and bone encodes the Ca2+-dependent enzyme peptidylarginine deaminase- marrow-derived dendritic cells from NOD mice. Eur. J. Immunol. 34: 1395– 1404. 2 that catalyzes the conversion of arginine residues to citrulline 11. Noorchashm, H., Y. K. Lieu, N. Noorchashm, S. Y. Rostami, S. A. Greeley, (43). This posttranslational conversion results in an ablation of the A. Schlachterman, H. K. Song, L. E. Noto, A. M. Jevnikar, C. F. Barker, and positive charge in proteins, potentially altering protein structure A. Naji. 1999. I-Ag7-mediated antigen presentation by B lymphocytes is critical in overcoming a checkpoint in T cell tolerance to islet beta cells of nonobese and function, as well as their recognition and presentation by the diabetic mice. J. Immunol. 163: 743–750. immune system (44–46). In humans, PADI2 and its closely related 12. Silveira, P. A., E. Johnson, H. D. Chapman, T. Bui, R. M. Tisch, and D. V. Serreze. 2002. The preferential ability of B lymphocytes to act as diabe- family member, PADI4, were identified as susceptibility genes for togenic APC in NOD mice depends on expression of self-antigen-specific im- multiple sclerosis and rheumatoid arthritis, respectively (47, 48). munoglobulin receptors. Eur. J. Immunol. 32: 3657–3666. In both diseases, excess citrullination of proteins caused by 13. Hulbert, C., B. Riseili, M. Rojas, and J. W. Thomas. 2001. B cell specificity Downloaded from contributes to the outcome of diabetes in nonobese diabetic mice. J. Immunol. overexpression of PADI2 or PADI4 leads to a breakdown in B and 167: 5535–5538. T cell tolerance (43, 47, 49, 50). Excessive expression of Padi2 by 14. Panigrahi, A. K., N. G. Goodman, R. A. Eisenberg, M. R. Rickels, A. Naji, and NOD B cells could play a similar role in the breakdown of tol- E. T. Luning Prak. 2008. RS rearrangement frequency as a marker of receptor editing in lupus and type 1 diabetes. J. Exp. Med. 205: 2985–2994. erance to pancreatic b cell Ags. 15. Silveira, P. A., J. Dombrowsky, E. Johnson, H. D. Chapman, D. Nemazee, and In conclusion, the refinement of the large congenic region in D. V. Serreze. 2004. B cell selection defects underlie the development of dia-

betogenic APCs in nonobese diabetic mice. J. Immunol. 172: 5086–5094. http://www.jimmunol.org/ NR4 mice into smaller subcongenic regions revealed a complexity 16. Acevedo-Sua´rez, C. A., C. Hulbert, E. J. Woodward, and J. W. Thomas. 2005. of interactions between multiple genes on Chr. 4 that regulate Uncoupling of anergy from developmental arrest in anti-insulin B cells supports susceptibility to T1D by controlling the pathogenic activity of B the development of autoimmune diabetes. J. Immunol. 174: 827–833. 17. Cox, S. L., J. Stolp, N. L. Hallahan, J. Counotte, W. Zhang, D. V. Serreze, and CD4 T cells. Identifying these gene(s) can provide significant A. Basten, and P. A. Silveira. 2010. Enhanced responsiveness to T-cell help insights into the mechanisms underlying the development of T1D causes loss of B-lymphocyte tolerance to a b-cell neo-self-antigen in type 1 and have the potential to highlight new molecular pathways that can diabetes prone NOD mice. Eur. J. Immunol. 40: 3413–3425. 18. Quinn, W. J., III, N. Noorchashm, J. E. Crowley, A. J. Reed, H. Noorchashm, be targeted as a disease-intervention approach with greater spec- A. Naji, and M. P. Cancro. 2006. Cutting edge: impaired transitional B cell ificity than current pan-B or T cell-depleting drugs (6, 51). The production and selection in the nonobese diabetic mouse. J. Immunol. 176: distinct set of polymorphisms in the NOR background also makes 7159–7164. 19. Menard, L., D. Saadoun, I. Isnardi, Y. S. Ng, G. Meyers, C. Massad, C. Price, by guest on October 2, 2021 our subcongenic mice useful for mapping allelic variants affecting C. Abraham, R. Motaghedi, J. H. Buckner, et al. 2011. The PTPN22 allele several other immune defects in NOD mice that have been linked encoding an R620W variant interferes with the removal of developing autore- active B cells in humans. J. Clin. Invest. 121: 3635–3644. to the distal Chr. 4 region (21). 20. Habib, T., A. Funk, M. Rieck, A. Brahmandam, X. Dai, A. K. Panigrahi, E. T. Luning Prak, A. Meyer-Bahlburg, S. Sanda, C. Greenbaum, et al. 2012. Altered B cell homeostasis is associated with type I diabetes and carriers of the PTPN22 Acknowledgments allelic variant. J. Immunol. 188: 487–496. We thank Christopher Brownlee for technical assistance with flow cytom- 21. Driver, J. P., D. V. Serreze, and Y. G. Chen. 2011. Mouse models for the study of etry and Jenny Kingham, Eric Schmeid, John Fisher, and other staff of the autoimmune type 1 diabetes: a NOD to similarities and differences to human Garvan Institute Biological Testing Facility and Australian BioResouces for disease. Semin. Immunopathol. 33: 67–87. 22. Todd, J. A. 2010. Etiology of type 1 diabetes. Immunity 32: 457–467. help with animal husbandry. We also thank Tom Brodnicki for providing 23. Serreze, D. V., M. Prochazka, P. C. Reifsnyder, M. M. Bridgett, and E. H. Leiter. microsatellite primers. 1994. Use of recombinant congenic and congenic strains of NOD mice to identify a new insulin-dependent diabetes resistance gene. J. Exp. Med. 180: 1553–1558. Disclosures 24. Hamilton-Williams, E. E., D. V. Serreze, B. Charlton, E. A. Johnson, The authors have no financial conflicts of interest. M. P. Marron, A. Mullbacher, and R. M. Slattery. 2001. Transgenic rescue implicates beta2-microglobulin as a diabetes susceptibility gene in nonobese diabetic (NOD) mice. Proc. Natl. Acad. Sci. USA 98: 11533–11538. 25. Serreze, D. V., M. Bridgett, H. D. Chapman, E. Chen, S. D. Richard, and References E. H. Leiter. 1998. Subcongenic analysis of the Idd13 locus in NOD/Lt mice: 1. Anderson, M. S., and J. A. Bluestone. 2005. The NOD mouse: a model of im- evidence for several susceptibility genes including a possible diabetogenic role mune dysregulation. Annu. Rev. Immunol. 23: 447–485. for beta 2-microglobulin. J. Immunol. 160: 1472–1478. 2. Kelemen, K., E. Liu, and H. W. Davidson. 2005. The role of T cells in beta cell 26. Reifsnyder, P. C., R. Li, P. A. Silveira, G. Churchill, D. V. Serreze, and damage in NOD mice and humans. In Type I Diabetes: Molecular, Cellular and E. H. Leiter. 2005. Conditioning the genome identifies additional diabetes re- Clinical Immunology, Online Edition Version 2.5. G. S. Eisenbarth, ed. Springer. sistance loci in Type I diabetes resistant NOR/Lt mice. Genes Immun. 6: 528– 3. Serreze, D. V., H. D. Chapman, D. S. Varnum, M. S. Hanson, P. C. Reifsnyder, 538. S. D. Richard, S. A. Fleming, E. H. Leiter, and L. D. Shultz. 1996. 27. Tan, I. K., L. Mackin, N. Wang, A. T. Papenfuss, C. M. Elso, M. P. Ashton, B lymphocytes are essential for the initiation of T cell-mediated autoimmune F. Quirk, B. Phipson, M. Bahlo, T. P. Speed, et al. 2010. A recombination hotspot diabetes: analysis of a new “speed congenic” stock of NOD.Ig mu null mice. J. leads to sequence variability within a novel gene (AK005651) and contributes to Exp. Med. 184: 2049–2053. type 1 diabetes susceptibility. Genome Res. 20: 1629–1638. 4. Noorchashm, H., N. Noorchashm, J. Kern, S. Y. Rostami, C. F. Barker, and 28. Yamanouchi, J., M. C. Puertas, J. Verdaguer, P. A. Lyons, D. B. Rainbow, A. Naji. 1997. B-cells are required for the initiation of insulitis and sialitis in G. Chamberlain, K. M. Hunter, L. B. Peterson, L. S. Wicker, and P. Santamaria. nonobese diabetic mice. Diabetes 46: 941–946. 2010. Idd9.1 locus controls the suppressive activity of FoxP3+CD4+CD25+ 5. Serreze, D. V., H. D. Chapman, M. Niens, R. Dunn, M. R. Kehry, J. P. Driver, regulatory T-cells. Diabetes 59: 272–281. M. Haller, C. Wasserfall, and M. A. Atkinson. 2011. Loss of intra-islet CD20 29. Hunter, K., D. Rainbow, V. Plagnol, J. A. Todd, L. B. Peterson, and L. S. Wicker. expression may complicate efficacy of B-cell-directed type 1 diabetes therapies. 2007. Interactions between Idd5.1/Ctla4 and other type 1 diabetes genes. J. Diabetes 60: 2914–2921. Immunol. 179: 8341–8349. 6. Marin˜o, E., P. A. Silveira, J. Stolp, and S. T. Grey. 2011. B cell-directed therapies 30. Silveira, P. A., H. D. Chapman, J. Stolp, E. Johnson, S. L. Cox, K. Hunter, in type 1 diabetes. Trends Immunol. 32: 287–294. L. S. Wicker, and D. V. Serreze. 2006. Genes within the Idd5 and Idd9/11 di- 12 GENETIC CONTROL OF DIABETOGENIC CD4 T AND B CELLS

abetes susceptibility loci affect the pathogenic activity of B cells in nonobese 41. Yu, G., J. Mao, Y. Wu, H. Luo, and J. Wu. 2006. Ephrin-B1 is critical in T-cell diabetic mice. J. Immunol. 177: 7033–7041. development. J. Biol. Chem. 281: 10222–10229. 31. Rolf, J., V. Motta, N. Duarte, M. Lundholm, E. Berntman, M. L. Bergman, 42. Alfaro, D., J. J. Garcı´a-Ceca, T. Cejalvo, E. Jime´nez, E. J. Jenkinson, L. Sorokin, S. L. Cardell, and D. Holmberg. 2005. The enlarged population of G. Anderson, J. J. Mun˜oz, and A. Zapata. 2007. EphrinB1-EphB signaling marginal zone/CD1d(high) B lymphocytes in nonobese diabetic mice maps to regulates thymocyte-epithelium interactions involved in functional T cell de- diabetes susceptibility region Idd11. J. Immunol. 174: 4821–4827. velopment. Eur. J. Immunol. 37: 2596–2605. 32. Chen, Y. G., F. Scheuplein, M. A. Osborne, S. W. Tsaih, H. D. Chapman, and 43. Vossenaar, E. R., A. J. Zendman, W. J. van Venrooij, and G. J. Pruijn. 2003. D. V. Serreze. 2008. Idd9/11 genetic locus regulates diabetogenic activity of PAD, a growing family of citrullinating enzymes: genes, features and involve- CD4 T-cells in nonobese diabetic (NOD) mice. Diabetes 57: 3273–3280. ment in disease. Bioessays 25: 1106–1118. 33. Graser, R. T., T. P. DiLorenzo, F. Wang, G. J. Christianson, H. D. Chapman, 44. Tarcsa, E., L. N. Marekov, G. Mei, G. Melino, S. C. Lee, and P. M. Steinert. D. C. Roopenian, S. G. Nathenson, and D. V. Serreze. 2000. Identification of 1996. Protein unfolding by peptidylarginine deiminase. Substrate specificity and a CD8 T cell that can independently mediate autoimmune diabetes development structural relationships of the natural substrates trichohyalin and filaggrin. J. in the complete absence of CD4 T cell helper functions. J. Immunol. 164: 3913– Biol. Chem. 271: 30709–30716. 3918. 45. Arita, K., H. Hashimoto, T. Shimizu, K. Nakashima, M. Yamada, and M. Sato. 34. Hamilton-Williams, E. E., X. Martinez, J. Clark, S. Howlett, K. M. Hunter, 2004. Structural basis for Ca(2+)-induced activation of human PAD4. Nat. D. B. Rainbow, L. Wen, M. J. Shlomchik, J. D. Katz, G. F. Beilhack, et al. 2009. Struct. Mol. Biol. 11: 777–783. Expression of diabetes-associated genes by dendritic cells and CD4 T cells 46. Arita, K., T. Shimizu, H. Hashimoto, Y. Hidaka, M. Yamada, and M. Sato. 2006. drives the loss of tolerance in nonobese diabetic mice. J. Immunol. 183: 1533– Structural basis for histone N-terminal recognition by human peptidylarginine 1541. deiminase 4. Proc. Natl. Acad. Sci. USA 103: 5291–5296. 35. Araki, M., D. Chung, S. Liu, D. B. Rainbow, G. Chamberlain, V. Garner, 47. Suzuki, A., R. Yamada, X. Chang, S. Tokuhiro, T. Sawada, M. Suzuki, K. M. Hunter, L. Vijayakrishnan, L. B. Peterson, M. Oukka, et al. 2009. Genetic M. Nagasaki, M. Nakayama-Hamada, R. Kawaida, M. Ono, et al. 2003. Func- evidence that the differential expression of the ligand-independent isoform of tional haplotypes of PADI4, encoding citrullinating enzyme peptidylarginine CTLA-4 is the molecular basis of the Idd5.1 type 1 diabetes region in nonobese deiminase 4, are associated with rheumatoid arthritis. Nat. Genet. 34: 395–402. diabetic mice. J. Immunol. 183: 5146–5157. 48. Kenealy, S. J., M. C. Babron, Y. Bradford, N. Schnetz-Boutaud, J. L. Haines, 36. Yamanouchi, J., D. Rainbow, P. Serra, S. Howlett, K. Hunter, V. E. Garner, J. B. Rimmler, S. Schmidt, M. A. Pericak-Vance, L. F. Barcellos, R. R. Lincoln, A. Gonzalez-Munoz, J. Clark, R. Veijola, R. Cubbon, et al. 2007. Interleukin-2 et al; American-French Multiple Sclerosis Genetics Group. 2004. A second- gene variation impairs regulatory T cell function and causes autoimmunity. Nat. generation genomic screen for multiple sclerosis. Am. J. Hum. Genet. 75: Downloaded from Genet. 39: 329–337. 1070–1078. 37. Pasquale, E. B. 2008. Eph-ephrin bidirectional signaling in physiology and 49. Kim, J. K., F. G. Mastronardi, D. D. Wood, D. M. Lubman, R. Zand, and disease. Cell 133: 38–52. M. A. Moscarello. 2003. Multiple sclerosis: an important role for post- 38. Palmer, A., and R. Klein. 2003. Multiple roles of ephrins in morphogenesis, translational modifications of myelin basic protein in pathogenesis. Mol. Cell. neuronal networking, and brain function. Genes Dev. 17: 1429–1450. Proteomics 2: 453–462. 39. Wu, J., and H. Luo. 2005. Recent advances on T-cell regulation by receptor 50. Lundberg, K., S. Nijenhuis, E. R. Vossenaar, K. Palmblad, W. J. van Venrooij, tyrosine kinases. Curr. Opin. Hematol. 12: 292–297. L. Klareskog, A. J. Zendman, and H. E. Harris. 2005. Citrullinated proteins have

40. Stimamiglio, M. A., E. Jime´nez, S. D. Silva-Barbosa, D. Alfaro, J. J. Garcı´a- increased immunogenicity and arthritogenicity and their presence in arthritic http://www.jimmunol.org/ Ceca, J. J. Mun˜oz, T. Cejalvo, W. Savino, and A. Zapata. 2010. EphB2-mediated joints correlates with disease severity. Arthritis Res. Ther. 7: R458–R467. interactions are essential for proper migration of T cell progenitors during fetal 51. Kaufman, A., and K. C. Herold. 2009. Anti-CD3 mAbs for treatment of type 1 thymus colonization. J. Leukoc. Biol. 88: 483–494. diabetes. Diabetes Metab. Res. Rev. 25: 302–306. by guest on October 2, 2021 SUPPLEMENTARY MATERIAL

Supplementary Table I. Primers used for quantitative real-time PCR.

Gene Forward Primer (5’-3’) Reverse Primer (5’-3’) Apitd1 TCTCTGCCAGGAAGTCACG GAAAAGTCACCTCGGAGATCG Ccdc21 GACAAGCAAAGGAAGCATATCTC AGTGTTCTCTCGCTGCAACTC Ccdc28b CAATTCCATACAGAAGCTACACTTG CACAGTTTGAAGAGGAGAGTAGCAG Ephb2 CTGTCCTCTGGCATCAACCT AAAGCTGGTGTAGTCCGGTATC Lrp8 CTCCAGTGAAGAGTGTCCTGC CACAGTCCTTCTCTCCGTCG Mtor ACAGATCCTGACCCTGATATGC ATTCAGAGCCACAAACAGAGC Padi2 AGCGACAAGGTCACTGTCAATACT ATGCCTTCTTAGGATTGTTCTTCTC Ptprf AGCTGTGCCCTTCAAGATCC CGCTACTGCCACGATTATC Tceb3 GGTTCCAGTAGAGCGAAATAGTG CCAGAGGGTTTCCAGCTTTC Zfp69 CCAGAGGATCCCATTTTAGATGTG CCTCGTGATGCCATGATCG Hprt CAATGCAAACTTTGCTTTCCC AAATCCAACAAAGTCTGGCC

Supplementary Table II. Microarray probes detecting differentially expressed transcripts between NOD vs. NR4 B cells. NOD vs NR4 q-value Probe Gene Chr: Position (bases) fold-changea (FDR)b 1449038_at Hsd11b1 1:195047829-195068553 -2.41 8.73E-03 1442347_at Lrp8 4:107547998-107549437 -2.13 3.68E-02 1420044_at Osbpl9 4:108733579-108734079 2.49 2.24E-02 1420843_at Ptprf 4:117880771-117964010 -1.58 4.08E-02 1458274_at Zfp69 4:120602743-120603230 7.92 2.24E-02 1454913_at 9930104L06Rik 4:124621223-124621905 -2.60 3.58E-02 1459840_s_at Ccdc28b 4:129296518-129296771 -7.36 1.91E-03 1454861_at Txlna 4:129303323-129305534 -2.18 1.19E-02 1418630_at Khdrbs1 4:129391062-129419549 -2.82 2.03E-02 1435982_at Stx12 4:132409417-132410076 -2.14 4.46E-02 1442318_at Nudc 4:133087684-133088210 7.33 2.03E-03 1432391_at Ccdc21 4:133720120-133743000 -15.16 1.57E-06 1434117_at Tceb3 4:135559282-135560774 1.72 2.95E-03 1454022_at Ephb2 4:136325304-136391889 -2.53 3.93E-02 1418252_at Padi2 4:140462258-140508501 3.51 2.03E-03 1448390_a_at Dhrs3 4:144482695-144517545 9.41 2.24E-02 1436574_at 1700029I01Rik 4:145420139-146555866 -2.83 4.27E-03 1442183_at Gm13238 4:145430789-146566430 -5.47 1.25E-03 1451477_at Gm13139 4:145371445-146555000 3.91 2.94E-03 1446175_at - 4:145425823-146561350 -9.38 4.41E-03 1428604_at 2610305D13Rik 4:146986045-147076644 38.06 5.08E-06 1424213_at Ubiad1 4:147808603-147818888 1.45 4.15E-02 1451478_at Mtor 4:147869291-147874571 2.41 1.26E-02 1453067_at Apitd1 4:148502307-148511728 4.22 6.18E-03 1423994_at Kif1b 4:148550427-148681740 2.27 4.79E-02 1424588_at Srgap3 6:112667964-112670977 -13.16 1.55E-04 1445537_at 2310039F13Rik 6:112708194-112708767 -20.41 5.48E-04 1460442_at Rps19 7:25669654-25674825 3.83 1.55E-04 1435703_at Ubash3b 9:40819175-40820310 -1.54 3.58E-02 1448862_at Icam2 11:106238969-106244292 14.94 1.02E-02 1417826_at Akr1e1 13:4591735-4608407 -3.99 2.24E-02 1435088_at Nsd1 13:55417742-55419687 -3.69 1.25E-03 1452452_at - Unmapped -5.97 1.66E-03 1452426_x_at - Unmapped -29.55 6.98E-06 a Mean fold-change (FC) in expression between five NOD and six NR4 B cell samples (combination of BCR-stimulated and unstimulated samples). Positive fold changes shown for probes where NOD>NR4, while negative fold changes denote probes where NR4>NOD. b A FDR cutoff of q<0.05 was used to define differentially expressed transcripts between NOD and NR4 samples.

2 Supplementary Table III. Microarray probes detecting differentially expressed transcripts in BCR-stimulated vs. unstimulated B cells from NOD or NR4 strains only, or in both strains.

NOD FC NR4 FC Probe Gene Chr: Position (b) (s vs. us)a q-valuec (s vs. us)b q-valuec 1444878_at - 1:80662178-80663151 2.52 8.89E-02 1452666_a_at Tmcc2 1:134252891-134287857 -2.01 5.24E-02 1449363_at Atf3 1:192994177-193007212 2.51 2.90E-02 1449270_at Plxdc2 2:16277917-16673950 2.24 7.08E-02 1450749_a_at Nr4a2 2:56959927-56967666 2.08 5.09E-02 1450750_a_at Nr4a2 2:56959927-56967666 2.49 8.08E-02 1447863_s_at Nr4a2 2:56960532-56960760 2.47 5.76E-02 1432678_at Itgav 2:83607291-83609658 2.44 7.20E-02 1415874_at Spry1 3:37538858-37543520 2.82 5.24E-02 1416022_at Fabp5 3:57331738-10016612 2.06 4.11E-02 1436836_x_at Cnn3 3:121160605-121161125 -2.16 3.34E-02 1455570_x_at Cnn3 3:121160772-121161113 -2.21 4.11E-02 1417539_at Slc35a1 4:34610505-34634676 -2.07 5.24E-02 1429262_at Rassf6 5:91032109-91060635 2.16 6.07E-02 1446055_at - 5:124915788-124916508 2.11 2.02E-02 1454551_at 9530034D02Rik 6:31280626-31281731 2.40 2.02E-02 1437671_x_at Prss23 7:96657609-96657905 3.35 5.95E-02 1431057_a_at Prss23 7:96657618-96666049 2.19 4.51E-02 1453547_at 1810046K07Rik 9:51097798-51137022 2.85 7.97E-02 1444402_at Zc3h12c 9:51920089-51920815 2.30 2.70E-02 1457824_at Plscr1 9:92147080-92147413 2.48 2.22E-02 1440432_at - 9:123621990-123622647 2.73 7.08E-02 1438588_at Plagl1 10:12835007-12835966 2.09 5.24E-02 1430581_at Bclaf1 10:20048836-20053697 2.49 4.92E-02 1437992_x_at Gja1 10:56110005-56110221 -2.59 2.22E-02 1422708_at Pik3cg 12:32857997-32893335 -2.01 2.02E-02 1428851_at 1300014I06Rik 13:34719704-34758192 -2.31 4.11E-02 1421477_at Cplx2 13:54479945-54481604 -2.09 4.11E-02 1436387_at Homer1 13:94129653-94131011 2.02 4.08E-02 1440573_at Erbb2ip 13:104688064-104688764 2.67 7.11E-02 1426156_at Lats2 14:58341026-58353137 2.31 3.67E-02 1443086_at Alcam 16:52261222-52261845 -2.20 2.64E-02 1440298_at Treml2 17:48451176-48451863 -2.05 3.34E-02 1444299_at A430093F15Rik 19:10830372-10860533 2.07 8.73E-02 1418248_at Gla X:131122894-131135505 2.13 4.95E-02 1450330_at Il10 1:132916423-132921547 3.19 2.82E-02 1428789_at Ralgps2 1:158734296-158735895 -2.43 4.60E-02 1431704_a_at Ralgps2 1:158743211-158869698 -2.12 4.16E-03 1457773_at Slamf6 1:173882570-173883300 -2.74 2.90E-02 1454877_at Sertad4 1:194670624-194677337 -3.12 3.97E-02 1426875_s_at Srxn1 2:151935895-151937106 2.06 7.28E-02 1419942_at Npn3 2:151936327-151936928 2.42 4.21E-02 1436480_at Dpp7 2:25207809-25208394 2.12 4.21E-02 1416021_a_at Gm3601 3:10012604-10016612 2.24 4.69E-02 1425454_a_at Il12a 3:68495448-68502013 -2.56 2.90E-02 1449176_a_at Dck 5:89194268-89212270 -2.06 2.02E-02 1448364_at Ccng2 5:93696282-93705253 -2.19 3.68E-02 1458100_at - 6:122538959-122539639 -2.10 4.67E-02 1419769_at Cd22 7:31650847-31665288 -2.51 7.86E-02 1419768_at Cd22 7:31650847-31665288 -2.18 8.19E-02 1443969_at Irs2 8:10986976-11004939 2.04 2.90E-02

3 1434955_at March1 8:68993022-68994332 -2.48 4.21E-02 1422027_a_at Ets1 9:32503610-32562627 -2.02 2.90E-02 1440157_at Scml4 10:42582798-42583132 -2.56 1.94E-02 1429413_at Cpm 10:117066775-117122359 -4.08 7.04E-02 1457339_at Ccm2 11:6446890-6496747 -2.02 1.94E-02 1418421_at Bcl6b 11:70037628-70042674 2.70 1.94E-02 1423339_s_at Ccdc16 11:82577846-82579757 -2.11 3.68E-02 1444376_at Sesn1 10:41622229-41623046 -2.04 2.82E-02 1441482_at - 11:110261063-110261748 -3.79 3.68E-02 1420342_at Gdap10 12:33506736-33511770 2.20 2.02E-02 1450350_a_at Jdp2 12:86940967-86980827 2.10 2.90E-02 1437221_at Rrm2b 13:3352742-3353251 -2.00 9.01E-03 1437222_x_at Rrm2b 13:3352742-3353251 -2.26 4.16E-03 1458206_at LOC218805 14:21185185-21185948 -2.00 4.16E-03 1448860_at Rem2 14:55094936-55099273 2.51 2.90E-02 1417750_a_at Slc25a37 14:69859904-69903160 -2.04 4.31E-02 1442445_at 2610027H17Rik 14:79123667-79124134 2.71 2.90E-02 1453287_at Ankrd33b 15:31253710-31297323 2.04 2.90E-02 1438030_at Rasgrp3 17:75926335-75928391 -2.25 2.63E-02 1430143_at 4930426D05Rik 18:21810065-21814608 -2.59 2.90E-02 1421912_at Slc23a1 18:35773951-35786881 -2.24 2.90E-02 1450188_s_at Lipg 18:75099035-75120914 2.70 2.90E-02 1428114_at Slc14a1 18:78296832-78298443 -4.47 2.90E-02 1418353_at Cd5 19:10792632-10813464 -2.09 6.13E-02 1431106_a_at 9530053H05Rik X:4789659-4791091 -2.12 1.94E-02 1442263_at Rgs13 1:145985796-145986427 2.32 3.68E-02 2.27 5.80E-02 1417601_at Rgs1 1:146091799-146096198 3.31 2.02E-02 2.79 1.94E-02 1442116_at Gm1012 2:118102844-118103494 3.36 5.76E-02 5.80 4.21E-02 1453590_at Arl8 2:14976908-14998230 2.78 3.34E-02 3.42 2.08E-02 1451680_at Srxn1 2:151931260-151936197 2.58 7.97E-02 2.16 5.16E-02 1417679_at Gfi1 5:108145672-108153363 2.25 2.02E-02 2.21 2.63E-02 1428207_at Bcl7a 5:123794462-123824091 -2.55 2.02E-02 -3.20 1.90E-02 1422134_at Fosb 7:19888045-19895100 2.55 5.76E-02 2.17 5.80E-02 1428834_at Dusp4 8:35882847-35885006 2.98 8.31E-02 3.10 1.14E-02 1453220_at Fam55b 9:48126088-48148943 -4.66 4.22E-02 -5.64 5.80E-03 1453360_a_at Tex9 9:72308907-72339763 -2.14 8.73E-02 -2.02 5.86E-02 1429963_at Mapk6 9:75256133-75257797 2.06 2.02E-02 2.04 3.08E-02 1419004_s_at Bcl2a1 9:88618122-114239689 2.14 2.85E-02 2.23 4.80E-02 1427683_at Egr2 10:67000625-67004935 2.57 2.02E-02 2.39 7.15E-03 1427682_a_at Egr2 10:67000625-67004935 2.09 4.95E-02 2.22 4.16E-03 1453009_at Cpm 10:117122298-117124401 -4.17 9.25E-02 -7.05 5.36E-02 1420710_at Rel 11:23641516-23670970 2.2 3.34E-02 2.07 2.63E-02 1421578_at Ccl4 11:83476088-83478182 3.51 4.95E-02 3.15 4.16E-03 1419561_at Ccl3 11:83461344-83462880 4.11 4.95E-02 2.66 7.42E-02 1422631_at Ahr 12:36182645-36219576 2.01 2.70E-02 2.05 1.94E-02 1450695_at Ahr 12:36182645-36219576 2.18 4.95E-02 2.19 7.20E-02 1440162_x_at A630043P06 13: 12569175-12612715 -3.11 9.04E-02 -3.08 3.44E-02 1437800_at Edaradd 13:12569079-12569626 -4.52 4.95E-02 -6.63 3.71E-02 1421174_at Irf4 13:30841126-30858850 2.64 3.34E-02 2.26 6.75E-03 1438658_a_at S1p3 13:51517303-51517499 2.3 4.95E-02 2.24 1.03E-02 1438511_a_at 1190002H23Rik 14:79688553-79688817 2.66 4.95E-02 3.77 9.11E-02 1424942_a_at Myc 15:61818976-61821802 3.09 4.46E-02 2.46 4.60E-02 1418936_at Maff 15:79178119-79189504 3.59 4.95E-02 3.87 2.90E-02 1431394_a_at Lrrk2 15:91610808-91646111 -2.43 3.34E-02 -4.43 1.14E-02 1416505_at Nr4a1 15:101097281-101105230 2.72 4.95E-02 2.67 5.55E-02

4 1443553_at - 17:52002007-52002542 -2.27 8.73E-02 -2.37 8.19E-02 1452408_at LOC547214 17:7515626-13535099 2.47 7.08E-02 2.31 4.60E-02 1428837_at Klhl14 18:21708879-21810583 -2.83 2.02E-02 -4.53 2.98E-02 1421262_at Lipg 18:75099035-75120914 2.1 8.31E-02 3.56 2.90E-02 1447084_at Nfatc1 18:80844173-80844704 2.16 4.95E-02 2.09 4.08E-02 1425761_a_at Nfatc1 18:80845197-80909810 2.17 4.11E-02 2.01 2.82E-02 1441887_x_at Sh3bgrl X:106395423-106395710 2.82 2.02E-02 3.17 3.04E-02 1448595_a_at Bex1 X:132748510-132749998 7.48 8.89E-02 3.17 9.91E-02 a Mean fold-change (FC) in expression between two BCR-stimulated (s) and three unstimulated (us) NOD B cell samples. Positive fold-changes are shown for probes where transcripts were greater in stimulated vs. non- stimulated B cells; while negative fold-changes denote the reverse. b Mean fold-change in expression between three BCR-stimulated and three unstimulated NR4 B cell samples. c A fold change ≥ 2 and FDR cutoff of q<0.1 was used to identify differentially expressed transcripts between BCR stimulated and unstimulated groups for each strain. Grey areas denote no significant changes (q>0.1) of gene expression in NOD or NR4 B cell groups.

5

SUPPLEMENTARY FIGURE 1. Genes from distal Chr. 4 conditions the genome such that B cells from NOD and NR4 mice exhibit a different genetic response to stimulation. (A) The Venn diagram on the right shows the number of probes that detected differentially expressed transcripts by one-way ANOVA (FDR, q<0.05) between NOD (n=5) vs. NR4 (n=6) B cell samples (dark blue); or ≥2-fold (FDR, q<0.1) between BCR-stimulated (stim.) vs. unstimulated (unstim.) NOD (red; n=2 and 3, respectively) or NR4 (green; n=3 and 3, respectively) B cell samples. Probes deemed significant in multiple tests are shown in overlapping segments. Lists of significant probes for each category are shown in Tables II-V. No probes exhibited significance via two-way ANOVA (FDR, q<0.1) comparing the interaction between strain and stimulation. The approximate chromosomal locations of differentially expressed probes, colored according to Venn diagram, are illustrated in the schematic diagram on the left. (B) Heat-map depicting the normalized signal of individual NOD and NR4 samples for differentially expressed genes on Chr. 4.

6