B lymphocytes play a key role in type 1 diabetes (T1D) development by serving as a subset of APCs preferentially supporting the expansion of autoreactive pathogenic T cells. As a result of their pathogenic importance, B lymphocyte–targeted therapies have received considerable interest as potential T1D interventions. Unfortunately, the B lymphocyte–directed T1D interventions tested to date failed to halt β cell demise. IgG autoantibodies marking humans at future risk for T1D indicate that B lymphocytes producing them have undergone the affinity-maturation processes of class switch recombination and, possibly, somatic hypermutation. This study found that CRISPR/Cas9-mediated ablation of the activation-induced cytidine deaminase gene required for class switch recombination/somatic hypermutation induction inhibits T1D development in the NOD mouse model. The activation-induced cytidine deaminase protein induces genome-wide DNA breaks that, if not repaired through RAD51-mediated homologous recombination, result in B lymphocyte death. Treatment with the RAD51 inhibitor 4,4′-diisothiocyanatostilbene-2, 2′-disulfonic acid also strongly inhibited T1D development in NOD mice. The genetic and small molecule–targeting approaches expanded CD73+ B lymphocytes that exert regulatory activity suppressing diabetogenic T cell responses. Hence, an initial CRISPR/Cas9-mediated genetic modification approach has identified the AID/RAD51 axis as a target for a potentially clinically translatable pharmacological approach that can block T1D development by converting B lymphocytes to a disease-inhibitory CD73+ regulatory state.

This article is featured in In This Issue, p.4181

Although the autoimmune destruction of insulin-producing pancreatic β cells underlying the development of type 1 diabetes (T1D) is ultimately mediated by the combined activity of CD4+ and CD8+ T cells, it is clear in the NOD mouse model, and likely in humans, that B lymphocytes play an additional key pathogenic role (19). Studies in NOD mice indicate that B lymphocytes contribute to T1D by being the subset of APCs that most efficiently supports the expansion of pathogenic T cell responses (1012). This is due to the presence of B lymphocytes expressing plasma membrane–bound Ig molecules capable of efficiently capturing and internalizing β cell autoantigens for subsequent processing and presentation to diabetogenic T cells (10, 12). Similar populations of pathogenic B lymphocytes also likely contribute to T1D development in humans, given the presence of circulating β cell Ag-specific autoantibodies that are critical biomarkers for identifying individuals at high risk for future disease (13).

Most autoantibodies in humans with, or at risk for, T1D are of an IgG isotype, indicating that the B lymphocytes producing them have undergone affinity maturation (13). Affinity maturation is the process occurring within germinal centers (GCs) by which B lymphocytes undergo Ig diversification and clonal selection. Ig diversification occurs through somatic hypermutation (SHM) and class switch recombination (CSR), whereas clonal selection results from competitive interaction with follicular helper T (Tfh) cells (14). Selective pressures within GCs result in the preferential expansion of B lymphocytes with greater affinity for their cognate Ag. In autoimmune diseases, such as T1D, aberrant selection processes lead to expansion of self-reactive B lymphocytes, which may become autoantibody-secreting cells or retain their surface Ig to serve as potentially more effective APCs (15). Although previous findings suggest that affinity maturation is important to T1D pathogenesis (16), the significance of CSR/SHM processes to disease progression has yet to be elucidated. Furthermore, it remains unclear whether B lymphocytes must undergo CSR/SHM to become effective autoreactive APCs supporting T1D pathogenesis.

Because of their role in supporting pathogenic T cell responses, there has been considerable interest in determining whether B lymphocyte–targeted approaches could provide an effective T1D intervention. A clinical trial found that transient treatment with the B lymphocyte–depleting CD20-specific rituximab Ab allowed for early (1 y), but not long-term (2 y), preservation of C-peptide production in recent-onset T1D patients (17). The lack of long-term protection may be attributable, at least in part, to the rebound of B lymphocytes following transient rituximab treatment. However, in NOD mice, pancreatic islet–infiltrating B lymphocytes lose cell surface expression of CD20 and, thus, are rendered resistant to depletion by a rituximab-like murine anti-CD20 Ab (18). These results indicate a need to identify alternative strategies that may provide a more effective B lymphocyte–directed T1D-intervention approach.

In the current study, we evaluated the contribution of CSR/SHM to T1D development and whether specifically targeting B lymphocytes undergoing these processes could provide an effective therapeutic intervention. As a first step, we used CRISPR-Cas9 technology to directly ablate, in NOD mice, the activation-induced cytidine deaminase gene (Aicda) necessary for initiating CSR/SHM processes (1921). Aicda ablation significantly inhibited T1D development.

The Aicda-encoded activation-induced cytidine deaminase protein (AID) initiates SHM and CSR by inducing point mutations and dsDNA breaks (DSBs) in Ig gene sequences. Additionally, AID generates DSBs elsewhere throughout the genome (22, 23) that are normally repaired by RAD51 complex–mediated homologous recombination (HR). In the absence of HR, B lymphocytes in which SHM/CSR has been initiated undergo cell death (23). The small molecule 4,4′-diisothiocyanatostilbene-2, 2′-disulfonic acid (DIDS) inhibits RAD51-mediated HR (23). DIDS has been used as a model agent in previous studies, indicating that RAD51 blockade could be considered an intervention for eliminating AID+ B cell lymphomas (23). In this study, we used DIDS treatment to determine whether targeting the AID/RAD51 axis could provide a B lymphocyte–directed intervention for T1D. Interestingly, genetic and DIDS-mediated disruption of the AID/RAD51 axis increased the numbers of CD73+ B lymphocytes that exerted regulatory processes actively suppressing T1D. Together, these studies indicate that therapies capable of expanding regulatory B lymphocytes, potentially including those targeting affinity-maturation processes, may ultimately represent clinically translatable T1D-intervention strategies.

NOD and C57BL/6J (B6) mice are maintained at The Jackson Laboratory under specific pathogen–free conditions. Lymphocyte-deficient NOD.Cg-PrkdcscidEmv30b/Dvs (NOD-scid) mice were described previously (24). B6.Cg-Aicda<tm1Hon>/HonRbrc mice were kindly provided to K.D.M. by Dr. T. Honjo (Graduate School of Medicine, Kyoto University). CRISPR-Cas9 technology was used to directly generate NOD.Aicda−/− mice by cytoplasmic microinjection of NOD/ShiLtDvs zygotes with 100 ng/μl Cas9 mRNA and 50 ng/μl the following single-guide RNAs (sgRNAs), with the upper case letters being the complement to the targeted genomic sequence: 5′-gaaattaatacgactcactataggAGTCACGCTGGAGACCGATAgttttagagctagaaatagc-3′ or 5′-gaaattaatacgactcactataggACTTCTTTTGCTTCATCAGAgttttagagctagaaatagc-3′ targeting exon 1 or 2, respectively, of Aicda (Supplemental Fig. 1A). Exon 1 and 2 sgRNAs were microinjected into 47 and 39 zygotes, respectively. These microinjected zygotes were then transplanted into three and two recipient females, respectively. Tail DNA from surviving progeny was sequenced and identified 100 and 14.3% targeting efficiency for exon 1 (14/14) and exon 2 (2/14). Mosaic founder mice identified as carrying a mutation in the targeted region of Aicda were backcrossed to NOD/ShiLtDvs mice. The resulting N1 progeny were screened for germline transmitted mutations by PCR amplification of exon 1 with the primers 5′-TCACACAACAGCACTGAAGC-3′ and 5′-ACCCAAAAGACCTGAGCAGA-3′ or exon 2 with the primers: 5′-CGCTCAGCTACCTTGCCTAT-3′ and 5′-CGAAGTCCAGTGAGCAGGA-3′. PCR products were purified and analyzed by sequencing on an ABI 3730 DNA analyzer (Applied Biosystems) using the forward or reverse primer. Mutant sequences were separated from wild-type (WT) using the Poly Peak Parser package (25) for R. A 2-bp deletion along with a 309-bp insertion was selected for a line targeting exon 1 (referred to as Line 1 in the text; formal name: NOD/ShiLtDvs-Aicda<em1Cml>/Dvs), and a 13-bp deletion was selected for a line targeting exon 2 (referred to as Line 26 in the text; formal name: NOD/ShiLtDvs-Aicda<em2Cml>/Dvs). N1 mutants were intercrossed to fix the mutations to homozygosity, after which lines were maintained by brother–sister matings. Mutant mice were genotyped by amplification-length polymorphisms using the same primers used for sequencing (Supplemental Fig. 1B). Mice were matched by age and sex for experimentation, but no specific randomization method was performed to form experimental groups. The Jackson Laboratory’s Institutional Animal Care and Use Committee approved all protocols involving mice.

B lymphocytes were purified from 8-wk-old female NOD or B6 mice. Total RNA was extracted using an RNeasy Micro Kit (QIAGEN). Primers for Aicda RT-PCR (26) are 5′-CAGGGACGGCATGAGACCT-3′ and 5′-TCAGCCTTGCGGTCTTCACA-3′, and primers for Gapdh are 5′-GAGAAACCTGCCAAGTATGATGAC-3′ and 5′-TGATGGTATTCAAGAGAGTAGGGAG-3′ (27). RNA was used to synthesize cDNA with a MessageSensor RT Kit (Thermo Fisher) and Random Decamers (Invitrogen). Power SYBR Green (Applied Biosystems) was used to determine the expression of Aicda and Gapdh. Quantitative PCR (qPCR) was run using an Applied Biosystems ViiA7, and data were analyzed using A&B RUO software.

Development of T1D was assessed by weekly monitoring of glycosuria with Ames Diastix (Bayer), with disease onset defined by two consecutive readings ≥ 0.25% (corresponds to blood glucose ≥ 300 mg/dl).

Quantitative mean insulitis scores were determined using the previously described calculation method (28). Briefly, Bouin’s fixed pancreata were sectioned at three nonoverlapping levels. Slides were stained with aldehyde fuchsin and H&E. Islets were scored by a blinded observer as follows: 0, no visible lesions; 1, peri-insular noninvasive leukocytic aggregates; 2, <25% islet destruction; 3, 25–75% islet destruction; and 4, 75–100% islet destruction. The final score was determined by dividing the cumulative score for each pancreas by the number of total islets (≥20 per mouse) examined. If no β cell–containing islets were found across three sections, the analyzed mouse received an insulitis score of 4. The final insulitis score incorporates the proportion of islets in analyzed mice that had undergone each level of destruction.

Infiltrating islet-associated leukocytes were isolated for flow cytometry, as previously described (18).

Splenic B lymphocytes were purified using anti-CD43 MicroBeads (Miltenyi Biotec), according to the manufacturer’s protocol. Purified B lymphocytes were cultured at a concentration of 1 × 106 cells per milliliter in RPMI 1640 (Life Technologies) supplemented with 2 mM l-glutamine (Life Technologies), 71.5 μM 2-ME (BP176-100; Fisher Scientific), 100 U/ml penicillin (Sigma-Aldrich), 100 μg/ml streptomycin (Sigma-Aldrich), and 10% (v/v) heat-inactivated FBS (Atlanta Biologicals) under stimulation with recombinant murine IL-4 (25 ng/ml; PeproTech) and anti-CD40 (1 μg/ml; BD Biosciences) or LPS (25 μg/ml; Sigma-Aldrich) at 37°C (5% CO2) for 48 h. At 48 h, cultures were restimulated as described above. At 72 h, cultures were diluted with media back to 1 × 106 cells per milliliter. Class switching to IgG1 isotype was quantified by flow cytometric analysis after 96 total hours in culture.

Leukocytes among the indicated samples were phenotyped by flow cytometry using LSRII SORP (BD Biosciences) or Attune Cytometer (Thermo Fisher Scientific) instrumentation, and data were analyzed using FlowJo software (TreeStar). For Aicda−/− experiments, single-cell suspensions of splenocytes were lysed with Gey’s buffer to remove RBCs, as previously described (29). For all experiments, propidium iodide or DAPI was used for live/dead discrimination. All analyses were done on gated singlets (live) cells; gating strategies are shown in Supplemental Fig. 2A, 2B. For DIDS-based experiments, prior to singlet/live gating, TER-119+ cells were excluded (Supplemental Fig. 2C) via gating strategies in lieu of lysis procedures. The following fluorochrome-conjugated Abs (clones) were used for these studies: CD95 (J02), GL-7 (GL7), CD43 (S7), CD69 (H1.2F3), CD21 (7G6), CD45.1 (A20), CD62L (MEL-14), CD8a (53–6.7), CD19 (1D3), B220 (RA3-6B2), CD4 (GK1.5), CD23 (B3B4), TER-119 (TER119), and streptavidin (all from BD Biosciences); CXCR5 (L138-D7), CD21 (7G6), IgD (11-26c.2a), CD80 (16-10A1), GL-7 (GL7), CD73 (TY/11.8), CD39 (Duha59), CD45.1 (A20), PD-1 (RMP1-30), PD-L2 (TY25), CD4 (GK1.5), CD23 (B3B4), B220 (RA3-6B2), and TCRβ (H57-597) (all from BioLegend); CD19 (1D3), IgM (II/41), CD45.1 (A20), and CD44 (IM7) (all from eBioscience); CD8α (53-6.7), CD19 (1D3), and CD16/CD32 (2.4G2) (all from Tonbo Biosciences); and IgG1 (1070-09) (SouthernBiotech). Splenic B lymphocyte subsets were characterized by flow cytometry, as previously described (30).

Quantification of serum Ig isotypes was performed as described (31). Coating Abs were as follows: IgG1, IgG2b, IgG3, and IgM (SouthernBiotech). Detection Ab was goat anti-mouse κ coupled to alkaline phosphatase (SouthernBiotech). Ab standards included IgG1 and IgG2b (SouthernBiotech) and IgG3 and IgM (Sigma-Aldrich). Plates were analyzed using an Infinite M200 Pro running Magellan 7.0 software (Tecan).

Splenic B lymphocytes and T cells were purified by negative depletion of CD11b+, CD11c+, TER-119+, and CD3ε+ populations and, in some experiments, CD73+ (for B lymphocyte purification) or CD11b+, CD11c+, TER-119+, and B220+ cells (for T cell purification), using biotin-conjugated Abs and streptavidin MicroBeads (Miltenyi Biotec). NOD-scid recipients were injected i.v. with the indicated B lymphocytes and/or T cells and were monitored for T1D development. The following specific biotinylated Abs were used: CD11b (M1/70), CD11c (HL3), CD3ε (145-2C11), TER-119 (TER119), and B220 (RA3-6B2) (all from BD Biosciences). CD73 (TY/11.8) was obtained from BioLegend.

B lymphocytes from 6–8-wk-old NOD.Aicda−/− or NOD mice were purified from pooled spleens and pancreatic lymph nodes (PLNs) by negative depletion of CD11b+, CD11c+, CD3ε+, and TER-119+ cells using Biotin Binder Dynabeads (Invitrogen). CD73+ and CD73 B lymphocytes from the purified pools were then sorted directly into Heat-Inactivated HyClone FBS (GE Healthcare Life Sciences) using a FACSAria II SORP sorter (BD Biosciences) equipped with an 85-μm nozzle. CD4+ CD73 T cells were purified from spleens of 6–8-wk-old NOD mice by negative depletion of CD11b+, CD11c+, TER-119+, B220+, CD8+, and CD73+ cells using streptavidin MicroBeads (Miltenyi Biotec) and the above-described Abs. Purified CD4+ CD73 T cells were labeled with 5 μM Cell Proliferation Dye eFluor 670 (eBioscience). Sorted B lymphocytes and labeled T cells were washed with serum-free X-VIVO 20 medium (Lonza) three times to remove residual serum and were cocultured under stimulation with soluble anti-CD40 (1 μg/ml), plate-bound anti-CD3ε (5 μg/ml), and soluble anti-CD28 (2 μg/ml; all from BD Biosciences) in the presence of 10 μM AMP and 0 or 100 μM α,β-methyleneadenosine 5′-diphosphate (APCP; all from Sigma-Aldrich) in serum-free X-VIVO 20 medium supplemented with 10 mM HEPES (Lonza), 1 mM sodium pyruvate (Life Technologies), 71.5 μM 2-ME (Fisher Scientific), 100 U/ml penicillin, 100 μg/ml streptomycin (both from Sigma-Aldrich), and nonessential amino acids (Lonza) at 37°C for 4 d. T cell proliferation was assessed by flow cytometry, and data were analyzed using FlowJo software. Percentage suppression was calculated relative to the mean proliferation index of T cells in the presence of CD73 B lymphocytes with 0 μM AMP and 0 μM APCP.

CD73+ or CD73 B lymphocytes were sorted as described above and cultured at 5.0 × 105 cells per milliliter in X-VIVO 20 medium with 0 or 10 μg/ml LPS (Sigma-Aldrich) for 3 d. IL-10 concentration in culture supernatants was determined using a Mouse IL-10 ELISA MAX Kit (BioLegend).

Beginning at 6, 8, or 10 wk of age, female NOD mice were injected i.p. weekly with 0, 10, or 50 mg/kg DIDS (CAS 67483-13-0; Santa Cruz Biotechnology) for the indicated periods of time. Vehicle was 0.1 M potassium bicarbonate.

Serum insulin autoantibodies (IAAs) were detected as described as part of the International Workshop on Lessons From Animal Models for Human Type 1 Diabetes (32).

Library preparation and data analysis for Ig repertoire sequencing were performed as previously described (33), with minor alterations. Briefly, female NOD/ShiLtDvs mice received weekly DIDS (50 mg/kg) or vehicle treatment from 8 to 16 wk of age, after which PLN B lymphocytes (CD45.1+ CD19+ B220+) were sort purified using a FACSAria II. Similarly, purified PLN-resident B lymphocytes from 8-wk-old unmanipulated NOD and NOD.Aicda−/− mice served as controls. Purified cells were washed with PBS, resuspended in Buffer RLT, and snap-frozen. Total RNA was purified using an RNeasy Micro Kit (QIAGEN). RNA integrity number determined, using the Bioanalyzer RNA 6000 Pico Kit (Agilent), for all samples was 10. Total RNA was reverse transcribed using SMARTScribe Reverse Transcriptase (Takara) with IgH isotype-specific primers (1 μM each) IgHG1/2: 5′-CAGGGATCCAKAGTTC-3′, IgHG3: 5′-CAGGGCTCCATAGTTC-3′, IgHM: 5′-GATGACTTCAGTGTTGT-3′, and IgHD: 5′-AGTGGCTGACTTCCAA-3′; and 1 μM template-switch adapter (5′-AAGCAGUGGTAUCAACGCAGAGUNNNNUNNNNUNNNNUCTTrGrGrGrG-3′, U, deoxyuridine) to introduce a unique molecular identifier (UMI) sequence to the 3′-end of each cDNA molecule. Reverse-transcription reactions were set-up as follows: 5 μl of RNA and 2 μl of IgH isotype-specific primers (10 μM each) were heated to 72°C for 3 min and then incubated at 42°C for 3 min to anneal primers. Then the following components were added to each reaction for a final concentration of 1× first-strand buffer, 2 mM DTT, 1 μM 5′-template switch adapter, and 1 mM each dNTP (Invitrogen), 40 U RNaseOUT (Invitrogen), and 200 U SMARTScribe Reverse Transcriptase. cDNA was then treated with 5 U uracil-DNA glycosylase (New England Biolabs) to degrade residual template switch adapter and then purified using a NucleoSpin Gel and PCR Purification Kit (Takara). First PCR was performed using cDNA equivalent to 2.5 ng of RNA and Platinum SuperFi DNA Polymerase (Invitrogen) with the following primers: M1SS: 5′-AAGCAGTGGTATCAACGCA-3′, mIGG12_r2: 5′-ATTGGGCAGCCCTGATTAGTGGATAGACHGATG-3′, mIGG3_r2: 5′-ATTGGGCAGCCCTGATTAAGGGATAGACAGATG-3′, mIGM_r2: 5′-ATTGGGCAGCCCTGATTGGGGGAAGACATTTGG-3′, and mIGD_r2: 5′-ATTGGGCAGCCCTGATTCTCTGAGAGGAGGAAC-3′. First PCR products were purified using a NucleoSpin Gel and PCR Purification Kit and then amplified, using Platinum SuperFi DNA Polymerase to introduce sample-specific dual-end barcodes, with the primers M1S: 5′-(N)4–6(XXXXXX)CAGTGGTATCAACGCAGAG-3′ and Z: 5′-(N)4–6(XXXXXX)ATTGGGCAGCCCTGATT-3′, where XXXXXX represents a sample-specific 6-nt barcode. Second PCR products were purified using a NucleoSpin Gel and PCR Purification Kit. For each sample, 300 ng of purified second PCR product was used for individual sequencing library preparation using a NEBNext Ultra II DNA Library Prep Kit with NEBNext Singleplex Oligos for Illumina (New England Biolabs). Concentrations for each library were determined using a Qubit dsDNA BR Assay Kit (Life Technologies); an equal quantity of each library was pooled and run on a 1.5% agarose gel, and a 600–800-bp band was excised and extracted using a NucleoSpin Gel and PCR Purification Kit. Pooled, gel-extracted libraries were then quality controlled using a Bioanalyzer High Sensitivity Kit (Agilent) and quantified by qPCR using a Library Quantification Kit/Illumina GA/ABI Prism with the Illumina Internal Control (both from Kapa Biosystems).

Sequencing libraries were spiked with a 30% PhiX Control v3 and loaded at 10 pM concentration for asymmetric 400 + 225–bp paired-end sequencing performed using a MiSeq sequencer running a MiSeq Reagent Kit v3 (all from Illumina). Sequencing data were processed using MIGEC (34), as follows. First, raw reads were demultiplexed based on each sample’s two 6-nt barcodes and then split into molecular identifier groups (MIGs) consisting of reads sharing the same UMI. MIGs consisting of fewer than five reads were discarded. Then, using only the first reads, consensus sequences were determined independently for 5′- and 3′-end reads within each MIG. 5′- and 3′-end read consensus sequences were then merged using the MiTools Merge Utility (https://github.com/milaboratory/mitools), with the minimum sequence similarity set to 70% for reads with overlapping parts, to generate full-length IgH sequences for each MIG. Using MiXCR (35), full-length Ig sequences were then mapped to the mouse reference germline sequences, and the resulting alignments were used to assemble clonotypes. When assembling clonotypes, a specific mutation probability of 10−4 was used for frequency-based correction of PCR or sequencing errors. Assembled clonotypes were then exported for diversity analyses using VDJtools (36). Clonotypic diversity estimates were calculated based on CDR3 nucleotide sequence and V- and J-segment usage.

All graphs and statistics were generated using Prism 6 (GraphPad) or Excel (Microsoft). Specific statistical analyses are listed within the respective figure legends. The Mann–Whitney, Wilcoxon, and t test analyses were all two-tailed.

B lymphocytes in NOD mice can undergo CSR/SHM, enhancing their capacity to process and present autoantigenic epitopes to diabetogenic T cells (16). However, it is unknown whether CSR and SHM are naturally important contributors to the diabetogenic activity of B lymphocytes in NOD mice, and by extension, in disease-susceptible humans. An initial finding supporting this possibility was that basal expression levels of Aicda necessary to induce CSR/SHM was higher in NOD B lymphocytes than in those from nonautoimmune-prone B6 mice (Fig. 1A). These data support the presence of a significantly higher baseline level of autoantigen-activated Aicda-expressing B lymphocytes in NOD mice than in B6 controls and are consistent with previous observations that GCs are expanded in autoimmune-prone strains (37).

FIGURE 1.

NOD B lymphocytes spontaneously express high levels of Aicda, and ablation of this gene inhibits T1D development. (A) B lymphocytes purified from individual spleens of 8-wk-old NOD (n = 3) and B6 (n = 3) female mice were tested for Aicda expression via qPCR. Data are representative of three independent experiments. (B) Representative flow cytometry plots showing class switching to IgG1 of purified B lymphocytes from NOD and NOD.Aicda−/− Lines 1 and 26 mice stimulated in culture with anti-CD40 (1 μg/ml) and murine IL-4 (25 ng/ml) for 96 h; quantitative data pooled from three experiments (NOD = 7, Line 1 = 5, Line 26 = 5) are summarized in (C). (D) Female T1D incidence in NOD controls compared with NOD.Aicda−/− Line 1 (p < 0.0001) and NOD.Aicda−/− Line 26 (p = 0.001) mice. (E) Insulitis scores on a scale of 0 (no visible lesion) to 4 (75–100% islet destruction) for female NOD.Aicda−/− Lines 1 and 26 mice remaining free of overt T1D at the end of the disease incidence study. (F) Insulitis for female NOD and NOD.Aicda−/− Line 1 mice at 7, 11, and 18 wk of age. (G) Representative islet for each insulitis scoring level. Arrows point to lymphocytic infiltration of islets. Scale bar, 200 μm. All bar graphs and scatter plots show mean ± SEM. The Student t test was used to determine the p value shown for gene expression, and the Mann–Whitney test was used for class switching. The p values for T1D incidence were determined using Mantel–Cox log-rank tests.

FIGURE 1.

NOD B lymphocytes spontaneously express high levels of Aicda, and ablation of this gene inhibits T1D development. (A) B lymphocytes purified from individual spleens of 8-wk-old NOD (n = 3) and B6 (n = 3) female mice were tested for Aicda expression via qPCR. Data are representative of three independent experiments. (B) Representative flow cytometry plots showing class switching to IgG1 of purified B lymphocytes from NOD and NOD.Aicda−/− Lines 1 and 26 mice stimulated in culture with anti-CD40 (1 μg/ml) and murine IL-4 (25 ng/ml) for 96 h; quantitative data pooled from three experiments (NOD = 7, Line 1 = 5, Line 26 = 5) are summarized in (C). (D) Female T1D incidence in NOD controls compared with NOD.Aicda−/− Line 1 (p < 0.0001) and NOD.Aicda−/− Line 26 (p = 0.001) mice. (E) Insulitis scores on a scale of 0 (no visible lesion) to 4 (75–100% islet destruction) for female NOD.Aicda−/− Lines 1 and 26 mice remaining free of overt T1D at the end of the disease incidence study. (F) Insulitis for female NOD and NOD.Aicda−/− Line 1 mice at 7, 11, and 18 wk of age. (G) Representative islet for each insulitis scoring level. Arrows point to lymphocytic infiltration of islets. Scale bar, 200 μm. All bar graphs and scatter plots show mean ± SEM. The Student t test was used to determine the p value shown for gene expression, and the Mann–Whitney test was used for class switching. The p values for T1D incidence were determined using Mantel–Cox log-rank tests.

Close modal

We next tested whether Aicda expression by NOD B lymphocytes was critical to their diabetogenic activity. CRISPR-Cas9 technology was used to directly target exon 1 or 2 (containing a potential alternative start site) of Aicda in NOD zygotes (Supplemental Fig. 1A–D). This allowed for the subsequent generation of pure NOD background stocks carrying the exon 1 (Line 1)–targeted or exon 2 (Line 26)–targeted Aicda allele in a homozygous state (Supplemental Fig. 1A–D). We then assessed the ability of purified B lymphocytes from the Line 1 or 26 stocks to undergo CSR to the IgG1 isotype following anti-CD40 and IL-4 stimulation in vitro. Similar to the case reported for a B6 background stock (22), B lymphocytes from Lines 1 and 26 NOD.Aicda−/− mice displayed significantly decreased CSR (Fig. 1B, 1C). Analyses were also carried out to confirm that the reduced ability of NOD.Aicda−/− B lymphocytes to undergo CSR was not the result of off-target mutations caused by CRISPR-Cas9. To do this, each NOD.Aicda−/− line was crossed to the established B6.Cg-Aicda<tm1Hon>/HonRbrc stock, and purified B lymphocytes from the resulting F1 hybrids were tested for CSR capacity. The observation of similarly reduced CSR in F1 offspring for each line demonstrates that this phenotype is the result of our novel direct-in-NOD disruption of Aicda (Supplemental Fig. 1E, 1F).

We then compared the female rate of T1D development for each NOD.Aicda−/− line with that of WT NOD controls. Both NOD.Aicda−/− lines exhibited similar significantly reduced rates of T1D (Fig. 1D). Heterozygous NOD.Aicda+/− mice developed T1D at an intermediate rate (data not shown). Interestingly, Lines 1 and 26 NOD.Aicda−/− mice remaining free of overt T1D at the end of the incidence study were still characterized by significant levels of insulitis (Fig. 1E). Next, we compared the kinetics of insulitis progression in female NOD and NOD.Aicda−/− mice at 7, 11, and 18 wk of age. Insulitis progression was not delayed in Line 1 NOD.Aicda−/− mice (Fig. 1F, 1G), indicating that this could not explain their resistance to overt T1D. Together, these data indicate that, despite the continued presence of islet-infiltrating leukocytes in NOD.Aicda−/− mice, the diabetogenic nature of these cells has been altered by genetic ablation of AID.

To initially investigate the basis for the T1D resistance of NOD.Aicda−/− mice, we compared them with WT controls for the presence of various immature and mature B lymphocyte populations within spleens and PLNs. Gating strategies for subsequently described flow cytometric studies are depicted in Supplemental Fig. 2. Line 1 mice were used for all subsequent studies because of better breeding proclivity and ease of genotyping compared with the equally T1D-resistant Line 26 mice.

Selection processes to prune autoreactive B lymphocytes operate at the immature transitional-1 (T1) stage in the spleen (38). The frequency of splenic T1 B lymphocytes is increased in NOD.Aicda−/− mice, whereas subsequent developmental stage transitional-2 (T2) cells are decreased (Fig. 2A). A shift in the ratio of T1/T2 cells might reflect increased tolerogenic culling of autoreactive B lymphocytes in NOD.Aicda−/− mice (39). Alternatively, this shift may be the result of the decreased susceptibility of Aicda−/− B lymphocytes to apoptosis at the T1 stage (40). Despite the differing distribution of T1 and T2 subsets, total splenic and PLN B lymphocyte yields were higher in NOD.Aicda−/− mice than in WT controls (Fig. 2B). Consistent with reports of AID ablation in other strains (41), Fas+ GL7+ GC B lymphocytes were expanded in the spleens and PLNs of NOD.Aicda−/− mice (Fig. 2C, 2D). This correlates with the large GC sizes observed in spleens of Aicda−/− mice (data not shown). Also, the proportions of splenic CD80+ and PD-L2+ memory-like B lymphocytes (42) were increased in NOD.Aicda−/− mice (Fig. 2E). Furthermore, the IgM-dominated serum isotype profile of NOD.Aicda−/− mice parallels that of other Aicda−/− strains (Fig. 2F), as well as those seen in AID-deficient patients (43). Finally, the percentages of B lymphocytes among islet-infiltrating leukocytes (CD45.1+) were similar in NOD and NOD.Aicda−/− mice (Fig. 2G). The proportions of islet-infiltrating B lymphocytes in NOD.Aicda−/− mice expressing the CD69 activation and CD80 or PD-L2 memory-like markers (42) were similar and greater, respectively, than those in NOD controls (Fig. 2H). Together, these data indicate that genetically ablating AID in NOD mice increases the numbers of peripheral B lymphocytes displaying a more predominant GC and memory-like phenotype.

FIGURE 2.

NOD.Aicda−/− B lymphocytes are numerically increased and display a more predominant GC phenotype than do those from NOD controls. (A) Flow cytometric analyses comparing splenic T1 (B220+ CD19+ CD21 CD23), T2 (B220+ CD19+ CD21hi CD23+), marginal zone (MZ; B220+ CD19+ CD21hi CD23), and follicular (FO; B220+ CD19+ CD21+ CD23+) B lymphocyte subsets in 8-wk-old female NOD and NOD.Aicda−/− mice (n = 8 per group; one representative of three experiments). (B) Enumeration of total B lymphocytes (CD45.1+ B220+ CD19+) in spleen and PLNs of 7–10-wk-old NOD.Aicda−/− mice (n = 21) compared with NOD controls (n = 19) pooled from four individual experiments. (C) Representative flow cytometry contour plots showing single-cell live B lymphocyte–gated events for analysis of splenic and PLN GC (Fas+ GL7+) B lymphocytes in 7-wk-old female NOD and NOD.Aicda−/− mice; quantitative data (n = 7 per group) from one experiment are summarized in (D). (E) Percentage of CD80+ or PD-L2+ cells among splenic or PLN B lymphocytes in 10-wk-old female NOD (n = 8) and NOD.Aicda−/− (n = 10) mice from one experiment. (F) Quantification, by ELISA, of serum Ig isotypes in 6–7-wk-old male NOD mice (n = 7) and NOD.Aicda−/− mice (n = 10) pooled from two experiments. (G) Analysis of islet-infiltrating B220+ cells among CD45.1+ leukocytes in 10–12-wk-old female NOD mice (n = 18) and NOD.Aicda−/− mice (n = 16) pooled from five experiments. (H) Proportions of islet-infiltrating B lymphocytes with a CD69+, CD80+, or PD-L2+ phenotype in 10–12-wk-old female NOD mice (n = 18 for CD69 and CD80, n = 13 for PD-L2) or NOD.Aicda−/− mice (n = 16 for CD69 and CD80, n = 14 for PD-L2) pooled from five experiments. All bar graphs show mean ± SEM. All p values were calculated using Mann–Whitney analyses.

FIGURE 2.

NOD.Aicda−/− B lymphocytes are numerically increased and display a more predominant GC phenotype than do those from NOD controls. (A) Flow cytometric analyses comparing splenic T1 (B220+ CD19+ CD21 CD23), T2 (B220+ CD19+ CD21hi CD23+), marginal zone (MZ; B220+ CD19+ CD21hi CD23), and follicular (FO; B220+ CD19+ CD21+ CD23+) B lymphocyte subsets in 8-wk-old female NOD and NOD.Aicda−/− mice (n = 8 per group; one representative of three experiments). (B) Enumeration of total B lymphocytes (CD45.1+ B220+ CD19+) in spleen and PLNs of 7–10-wk-old NOD.Aicda−/− mice (n = 21) compared with NOD controls (n = 19) pooled from four individual experiments. (C) Representative flow cytometry contour plots showing single-cell live B lymphocyte–gated events for analysis of splenic and PLN GC (Fas+ GL7+) B lymphocytes in 7-wk-old female NOD and NOD.Aicda−/− mice; quantitative data (n = 7 per group) from one experiment are summarized in (D). (E) Percentage of CD80+ or PD-L2+ cells among splenic or PLN B lymphocytes in 10-wk-old female NOD (n = 8) and NOD.Aicda−/− (n = 10) mice from one experiment. (F) Quantification, by ELISA, of serum Ig isotypes in 6–7-wk-old male NOD mice (n = 7) and NOD.Aicda−/− mice (n = 10) pooled from two experiments. (G) Analysis of islet-infiltrating B220+ cells among CD45.1+ leukocytes in 10–12-wk-old female NOD mice (n = 18) and NOD.Aicda−/− mice (n = 16) pooled from five experiments. (H) Proportions of islet-infiltrating B lymphocytes with a CD69+, CD80+, or PD-L2+ phenotype in 10–12-wk-old female NOD mice (n = 18 for CD69 and CD80, n = 13 for PD-L2) or NOD.Aicda−/− mice (n = 16 for CD69 and CD80, n = 14 for PD-L2) pooled from five experiments. All bar graphs show mean ± SEM. All p values were calculated using Mann–Whitney analyses.

Close modal

T cells are the ultimate mediators of β cell destruction in T1D. Therefore, we tested whether T cells in T1D-resistant NOD.Aicda−/− mice were quantitatively and/or qualitatively distinct from those in NOD controls. Surprisingly, yields of splenic CD4+ and CD8+ T cells were increased in NOD.Aicda−/− mice (Fig. 3A). Corresponding with increased GC B lymphocytes, fully activated CXCR5+ PD-1+ Tfh cells were also expanded in spleens and PLNs of NOD.Aicda−/− mice (Fig. 3B, 3C). Although numerically expanded, we sought to determine whether the intrinsic diabetogenic activity of T cells in NOD.Aicda−/− mice was decreased. Purified NOD or NOD.Aicda−/− T cells did not differ in their ability to transfer T1D to lymphocyte-deficient NOD-scid recipients (Fig. 3D). Together, these data indicate that genetic ablation of AID in NOD mice leads to an expansion of total peripheral and Tfh subset T cells, and the observed disease protection is not due to a decrease in their diabetogenic potential.

FIGURE 3.

Numbers of total peripheral and GC phenotype CD4+ T cells are increased in NOD.Aicda−/− mice. (A) Yields of splenic or PLN-resident CD4 (CD45.1+ TCRβ+ CD4+) and CD8 (CD45.1+ TCRβ+ CD8+) T cells in 7–10-wk-old female NOD (n = 19) and NOD.Aicda−/− (n = 21) mice pooled from four experiments. (B) Representative flow cytometric contour plots showing single-cell live CD4 T cell–gated events for analysis of splenic and PLN full Tfh cells (CXCR5+ PD-1+) in 7-wk-old female mice, with quantitative data (n = 7 per group) summarized in (C), one representative of three experiments. (D) T1D development in NOD-scid recipients injected with 2.5 × 106 purified T cells from 7–8-wk-old female NOD.Aicda−/− or NOD mice. All bar graphs show mean ± SEM. All p values in bar graphs were calculated using Mann–Whitney analysis. T1D incidence p value was calculated using Mantel–Cox analysis.

FIGURE 3.

Numbers of total peripheral and GC phenotype CD4+ T cells are increased in NOD.Aicda−/− mice. (A) Yields of splenic or PLN-resident CD4 (CD45.1+ TCRβ+ CD4+) and CD8 (CD45.1+ TCRβ+ CD8+) T cells in 7–10-wk-old female NOD (n = 19) and NOD.Aicda−/− (n = 21) mice pooled from four experiments. (B) Representative flow cytometric contour plots showing single-cell live CD4 T cell–gated events for analysis of splenic and PLN full Tfh cells (CXCR5+ PD-1+) in 7-wk-old female mice, with quantitative data (n = 7 per group) summarized in (C), one representative of three experiments. (D) T1D development in NOD-scid recipients injected with 2.5 × 106 purified T cells from 7–8-wk-old female NOD.Aicda−/− or NOD mice. All bar graphs show mean ± SEM. All p values in bar graphs were calculated using Mann–Whitney analysis. T1D incidence p value was calculated using Mantel–Cox analysis.

Close modal

As described above, when transferred in the absence of B lymphocytes, purified T cells from NOD.Aicda−/− donors efficiently induced T1D development in NOD-scid recipients. This indicated that T1D resistance in NOD.Aicda−/− mice does not result from a T cell–intrinsic effect. Therefore, we assessed how AID-deficient B lymphocytes might influence such pathogenic effectors. To initially test this, equal numbers of purified NOD splenic T cells and purified NOD or NOD.Aicda−/− splenic B lymphocytes were transferred into NOD-scid recipients. NOD T cells induced T1D far less efficiently in NOD-scid recipients when cotransferred with Aicda−/− B lymphocytes than when cotransferred with WT B lymphocytes (Fig. 4A). These results indicated that Aicda−/− B lymphocytes may actively suppress diabetogenic T cell responses.

FIGURE 4.

NOD T cells adoptively transfer T1D less efficiently in the presence of Aicda-deficient B lymphocytes characterized by an expansion of the CD73+ subset. (A) T1D development in NOD-scid recipients injected with 2.5 × 106 purified T cells from 7–8-wk-old female NOD donors admixed with 2.5 × 106 B lymphocytes from 7–8-wk-old NOD mice (n = 10) or NOD.Aicda−/− mice (n = 9). (B) Representative contour plots for B220 versus CD73 staining pattern of splenic B220+ CD19+ B lymphocytes from female NOD and NOD.Aicda−/− mice. (C) Quantification of the percentage of CD73+ cells among spleen, PLN, and islet B lymphocytes in 7–12-wk-old NOD mice (spleen/PLN: n = 15, islets: n = 18) and NOD.Aicda−/− mice (spleen/PLN: n = 17, islets: n = 16). Spleen and PLN data are pooled from three experiments; islet data are pooled from five experiments. Quantification of the percentage of CD39+ CD73+, CD39+ CD73, and CD39 CD73+ cells among splenic (D), PLN (E), or islet (F) B220+ CD19+ B lymphocytes in 7–12-wk-old female NOD mice (spleen/PLN: n = 15, islets: n = 7) and NOD.Aicda−/− mice (spleen/PLN: n = 17, islets: n = 8) pooled from two individual experiments. T1D incidence p value was calculated using Mantel–Cox analysis. All bar graph p values were calculated using Mann–Whitney analysis, and all bar graphs show mean ± SEM.

FIGURE 4.

NOD T cells adoptively transfer T1D less efficiently in the presence of Aicda-deficient B lymphocytes characterized by an expansion of the CD73+ subset. (A) T1D development in NOD-scid recipients injected with 2.5 × 106 purified T cells from 7–8-wk-old female NOD donors admixed with 2.5 × 106 B lymphocytes from 7–8-wk-old NOD mice (n = 10) or NOD.Aicda−/− mice (n = 9). (B) Representative contour plots for B220 versus CD73 staining pattern of splenic B220+ CD19+ B lymphocytes from female NOD and NOD.Aicda−/− mice. (C) Quantification of the percentage of CD73+ cells among spleen, PLN, and islet B lymphocytes in 7–12-wk-old NOD mice (spleen/PLN: n = 15, islets: n = 18) and NOD.Aicda−/− mice (spleen/PLN: n = 17, islets: n = 16). Spleen and PLN data are pooled from three experiments; islet data are pooled from five experiments. Quantification of the percentage of CD39+ CD73+, CD39+ CD73, and CD39 CD73+ cells among splenic (D), PLN (E), or islet (F) B220+ CD19+ B lymphocytes in 7–12-wk-old female NOD mice (spleen/PLN: n = 15, islets: n = 7) and NOD.Aicda−/− mice (spleen/PLN: n = 17, islets: n = 8) pooled from two individual experiments. T1D incidence p value was calculated using Mantel–Cox analysis. All bar graph p values were calculated using Mann–Whitney analysis, and all bar graphs show mean ± SEM.

Close modal

CD39 and CD73 are ectoenzymes that catalyze the conversion of extracellular ATP to AMP and then to adenosine, respectively. There is evidence that alterations to the CD39/CD73 purinergic metabolic pathway play a major role in several autoimmune diseases (44). Adenosine signaling through the adenosine A2a receptor (A2aR) can inhibit TCR signaling and prevent cellular activation (45). In a murine colitis model, CD73 expression reportedly marks a subset of B lymphocytes with the capacity to suppress T cell responses through adenosine production (46). Thus, we compared the extent to which B lymphocytes from Aicda-intact and -deficient NOD mice express CD39 and/or CD73. Levels of CD73+ B lymphocytes were greater within spleens, PLNs, and islets of NOD.Aicda−/− mice compared with NOD controls (Fig. 4B, 4C). We observed higher proportions of CD73+ B lymphocytes that either did or did not coexpress CD39 within spleens, PLNs, and islets of NOD.Aicda−/− mice compared with NOD controls but found no differences in the CD39+ CD73 fraction (Fig. 4D–F). Together, these data indicate that the ability of NOD T cells to induce T1D development is reduced in the presence of NOD.Aicda−/− B lymphocytes, which contain an expanded CD73+ subset that was previously reported to exert regulatory activity (46).

As noted above, CD73 expression can mark a population of suppressive B lymphocytes. Aicda−/− B lymphocytes engrafted into NOD-scid recipients retain increased CD73 expression (Fig. 5A). Thus, we assessed whether the expanded CD73+ B lymphocytes in NOD.Aicda−/− mice exert T1D suppressive effects in adoptive-transfer experiments. Purified NOD T cells transferred T1D to NOD-scid recipients with significantly greater efficiency when admixed with Aicda−/− B lymphocytes that had been depleted of the CD73+ subset (Fig. 5B). We then carried out in vitro studies to determine whether Aicda-deficient NOD B lymphocytes exert suppressive effects on such pathogenic effectors through regulatory activity mediated by the expanded CD73+ subset. Sorted CD73+ and CD73 B lymphocytes from NOD or NOD.Aicda−/− mice were cultured with anti-CD40, anti-CD3/CD28–stimulated CD73 CD4+ WT T cells, and AMP, with or without the CD73 inhibitor APCP. On a per-cell basis, CD73+ B lymphocytes from AID-intact and -deficient NOD mice suppressed T cell proliferation equally and to a significantly greater extent than did the CD73 subset in the presence of AMP; addition of APCP diminished this effect (Fig. 5C, Supplemental Fig. 3A).

FIGURE 5.

Expanded AID-deficient CD73+ B lymphocytes exert regulatory-like activity suppressing diabetogenic T cell responses. (A) Splenic engraftment levels of CD73+ B lymphocytes from NOD (n = 9) or NOD.Aicda−/− (n = 8) donors admixed with NOD T cells (2.5 × 106 of each cell type) 4 wk posttransfer into NOD-scid recipients. Results from one experiment. (B) T1D development in NOD-scid recipients injected with 2.5 × 106 purified T cells from 7–8-wk-old female NOD mice admixed with 2.5 × 106 total or CD73-depleted B lymphocytes purified from 7–8-wk-old female NOD.Aicda−/− donors. (C) A total of 1.0 × 105 CD73-depleted purified CD4+ T cells from 7–10-wk-old male NOD mice were labeled with Cell Proliferation Dye eFluor 670 and cocultured for 4 d with 1.0 × 105 CD73+ or CD73 B lymphocytes from pooled spleens of 7–10-wk-old male NOD (n = 6 biological replicates) or NOD.Aicda−/− (n = 10 biological replicates) mice under stimulation conditions consisting of soluble anti-CD40 (1 μg/ml), plate-bound anti-CD3ε (5 μg/ml), and soluble anti-CD28 (2 μg/ml), with 0 (baseline) or 10 μM AMP in the presence or absence of 100 μM APCP. Quantification of percentage suppression from baseline. (D) A total of 5 × 104 sort-purified NOD (n = 5 biological replicates) or NOD.Aicda−/− (n = 10 biological replicates) CD73+ or CD73 B lymphocytes were cultured for 3 d with 0 or 10 μg/ml LPS, and culture supernatant IL-10 levels were measured by ELISA. CD73-mediated in vitro suppression and IL-10 production data shown are combined from three and two individual experiments, respectively. All bar graphs show mean ± SEM. T1D incidence study p value was calculated using Mantel–Cox analysis. The Wilcoxon test was performed for the suppression assay. Mann–Whitney analysis was performed for IL-10 production.

FIGURE 5.

Expanded AID-deficient CD73+ B lymphocytes exert regulatory-like activity suppressing diabetogenic T cell responses. (A) Splenic engraftment levels of CD73+ B lymphocytes from NOD (n = 9) or NOD.Aicda−/− (n = 8) donors admixed with NOD T cells (2.5 × 106 of each cell type) 4 wk posttransfer into NOD-scid recipients. Results from one experiment. (B) T1D development in NOD-scid recipients injected with 2.5 × 106 purified T cells from 7–8-wk-old female NOD mice admixed with 2.5 × 106 total or CD73-depleted B lymphocytes purified from 7–8-wk-old female NOD.Aicda−/− donors. (C) A total of 1.0 × 105 CD73-depleted purified CD4+ T cells from 7–10-wk-old male NOD mice were labeled with Cell Proliferation Dye eFluor 670 and cocultured for 4 d with 1.0 × 105 CD73+ or CD73 B lymphocytes from pooled spleens of 7–10-wk-old male NOD (n = 6 biological replicates) or NOD.Aicda−/− (n = 10 biological replicates) mice under stimulation conditions consisting of soluble anti-CD40 (1 μg/ml), plate-bound anti-CD3ε (5 μg/ml), and soluble anti-CD28 (2 μg/ml), with 0 (baseline) or 10 μM AMP in the presence or absence of 100 μM APCP. Quantification of percentage suppression from baseline. (D) A total of 5 × 104 sort-purified NOD (n = 5 biological replicates) or NOD.Aicda−/− (n = 10 biological replicates) CD73+ or CD73 B lymphocytes were cultured for 3 d with 0 or 10 μg/ml LPS, and culture supernatant IL-10 levels were measured by ELISA. CD73-mediated in vitro suppression and IL-10 production data shown are combined from three and two individual experiments, respectively. All bar graphs show mean ± SEM. T1D incidence study p value was calculated using Mantel–Cox analysis. The Wilcoxon test was performed for the suppression assay. Mann–Whitney analysis was performed for IL-10 production.

Close modal

It has been reported previously that B lymphocytes with an ability to suppress T1D development in NOD mice primarily do so through IL-10 secretion (47). Similarly, the expanded CD73+ B lymphocyte population in NOD.Aicda−/− mice, which exhibit the capacity to suppress diabetogenic T cell responses, secrete significantly greater levels of IL-10 upon LPS stimulation than do their CD73 counterparts (Fig. 5D). Together, these results indicate that diminished T1D development in NOD.Aicda−/− mice is due, at least in part, to an expansion of CD73+ B lymphocytes with the capacity to suppress pathogenic T cell responses. CD73+ B lymphocytes suppress T cell responses through the generation of adenosine by this ectoenzyme and potentially by their ability to secrete IL-10 upon stimulation. These collective results also indicate that genetic ablation of AID inhibits T1D development in NOD mice by quantitatively increasing, rather than qualitatively changing, immunoregulatory CD73+ B lymphocytes.

The results of our genetic studies show that AID-dependent processes play an important role in B lymphocyte contributions to T1D development in NOD mice. By inhibiting RAD51-mediated HR repair of AID-initiated off-target DSBs, treatment with the small molecule DIDS induces the death of B lymphocytes in which SHM and CSR processes have been initiated (23). Therefore, we hypothesized that targeting the AID/RAD51 pathway by DIDS treatment might provide an effective B lymphocyte–directed T1D intervention. To initially test this possibility, we stimulated purified NOD B lymphocytes with anti-CD40 and IL-4 in the presence of vehicle or 150 μM DIDS in vitro. Similar to the case for those from B6 control mice, after 4 d of stimulation, fewer NOD B lymphocytes were recovered from DIDS-containing cultures (Fig. 6A). In addition, fewer anti-CD40/IL-4–stimulated B lymphocytes were recovered from cultures containing (E)-5-acetamido-2-(4-(3-isopropylthioureido)-2-sulfonatostyryl)benzenesulfonate, another small molecule RAD51 inhibitor that is 1500-fold more potent than DIDS (Fig. 6A, Supplemental Fig. 4).

FIGURE 6.

DIDS diminishes in vitro expansion and Ig usage diversity of NOD B lymphocytes. (A) Cellular yields of purified B lymphocytes from B6 or NOD (n = 3 biological replicates per group) mice cultured for 96 h (1 × 106 cells per milliliter) with anti-CD40 (1 μg/ml) and murine IL-4 (25 ng/ml) in the presence of vehicle, 150 μM DIDS, or 100 nM (E)-5-acetamido-2-(4-(3-isopropylthioureido)-2-sulfonatostyryl)benzenesulfonate. Data are representative of one of three experiments. (B) A total of 136,000–400,000 purified PLN B lymphocytes from the indicated NOD experimental groups was sequenced for IgH gene usage diversity (n = 3 mice per group) in one experiment. Sequences with early stop codons or frame-shift mutations were filtered to display only “functional” clones. Data are presented in a rarefaction plot showing clonal diversity as a function of unique cDNA molecules sequenced. Solid and dashed lines are interpolated and extrapolated regions, respectively, with points marking the exact sample size and observed diversity. The shaded area represents the 95% confidence interval. (C) Chaos diversity estimate index. Observed diversity (D) and Efron–Thisted estimate (E) after 500 iterations of downsampling to 500 reads. Scatter plots show mean ± SEM. The p values were calculated using the Student t test.

FIGURE 6.

DIDS diminishes in vitro expansion and Ig usage diversity of NOD B lymphocytes. (A) Cellular yields of purified B lymphocytes from B6 or NOD (n = 3 biological replicates per group) mice cultured for 96 h (1 × 106 cells per milliliter) with anti-CD40 (1 μg/ml) and murine IL-4 (25 ng/ml) in the presence of vehicle, 150 μM DIDS, or 100 nM (E)-5-acetamido-2-(4-(3-isopropylthioureido)-2-sulfonatostyryl)benzenesulfonate. Data are representative of one of three experiments. (B) A total of 136,000–400,000 purified PLN B lymphocytes from the indicated NOD experimental groups was sequenced for IgH gene usage diversity (n = 3 mice per group) in one experiment. Sequences with early stop codons or frame-shift mutations were filtered to display only “functional” clones. Data are presented in a rarefaction plot showing clonal diversity as a function of unique cDNA molecules sequenced. Solid and dashed lines are interpolated and extrapolated regions, respectively, with points marking the exact sample size and observed diversity. The shaded area represents the 95% confidence interval. (C) Chaos diversity estimate index. Observed diversity (D) and Efron–Thisted estimate (E) after 500 iterations of downsampling to 500 reads. Scatter plots show mean ± SEM. The p values were calculated using the Student t test.

Close modal

We next determined whether DIDS treatment affected the Ig repertoire of NOD mice in a similar manner to genetic ablation of AID. We reasoned that the inability of NOD.Aicda−/− B lymphocytes to initiate CSR/SHM processes would result in a decreased Ig gene coding repertoire diversity relative to WT controls. By extension, we also hypothesized that DIDS-mediated disruption of the AID/RAD51 axis would recapitulate such decreases in Ig repertoire diversity. Therefore, using established protocols (33), we used high-throughput sequencing to characterize full-length IgH mRNAs from purified PLN-resident B lymphocytes of individual 8-wk-old NOD and NOD.Aicda−/− mice, as well as from NOD mice treated with vehicle or 50 mg/kg DIDS from 8 to 16 wk of age. To minimize overestimations of diversity, the protocol uses template switch reverse transcription to incorporate a 12-bp UMI at the 3′-end of each cDNA molecule. This allows for correction of sequencing and PCR errors by grouping multiple reads originating from single cDNA molecules to form consensus reads. Rarefaction and extrapolated Chao diversity estimate (48) analyses revealed decreased IgH diversity among B lymphocytes from NOD.Aicda−/− and DIDS-treated NOD mice compared with untreated and vehicle-treated NOD controls, respectively (Fig. 6B, 6C). To ensure that sampling bias resulting from variances in sequencing depth was not responsible for the observed differences in diversity, each sample was down-sampled to 500 consensus reads. After 500 iterations of resampling, Aicda−/− and DIDS-treated B lymphocytes had decreased diversity, as measured by the observed diversity and lower bound total diversity (Efron–Thisted) estimate (49), compared with their respective controls (Fig. 6D, 6E). Together, the in vitro and in vivo data indicate that NOD B lymphocytes are as susceptible to DIDS as previously reported for those from the B6 strain (23), and treatment with this small molecule decreases IgH repertoire diversity in a manner similar to Aicda ablation.

We next tested whether in vivo treatment with DIDS exerted T1D protective effects in NOD mice. Starting at 6, 8, or 10 wk of age, female NOD mice received weekly i.p. injections of vehicle or DIDS at 10 mg/kg (low-dose) or 50 mg/kg (high-dose; 6 and 8 wk start only). Regardless of start time or dose, DIDS treatment exerted strong T1D protective effects to 24 wk of age, at which time the disease incidence in controls reached 90% (Fig. 7A–C).

FIGURE 7.

DIDS treatment inhibits T1D development. Starting at 6 (A), 8 (B), or 10 (C) wk of age, female NOD mice were treated with vehicle or DIDS (10 or 50 mg/kg) on a weekly basis and monitored for T1D development. The p values were calculated using Mantel–Cox analysis. (D) Serum was harvested from the cohort of mice in (A) at the initiation of treatment and retrospectively typed for IAAs. Graph represents the percentage of IAA+ or IAA mice in each treatment group that did or did not progress to T1D (IAA+ DIDS: n = 9; IAA+ Vehicle: n = 3; IAA DIDS: n = 11; IAA Vehicle: n = 7). The p values were calculated using the Fisher exact test. (E) Insulitis scores for NOD female mice treated weekly starting at 8 wk of age with vehicle (n = 10) or DIDS at a dose of 10 mg/kg (n = 8) or 50 mg/kg (n = 14). Data show mean ± SEM. The p values were calculated using Mann–Whitney analysis.

FIGURE 7.

DIDS treatment inhibits T1D development. Starting at 6 (A), 8 (B), or 10 (C) wk of age, female NOD mice were treated with vehicle or DIDS (10 or 50 mg/kg) on a weekly basis and monitored for T1D development. The p values were calculated using Mantel–Cox analysis. (D) Serum was harvested from the cohort of mice in (A) at the initiation of treatment and retrospectively typed for IAAs. Graph represents the percentage of IAA+ or IAA mice in each treatment group that did or did not progress to T1D (IAA+ DIDS: n = 9; IAA+ Vehicle: n = 3; IAA DIDS: n = 11; IAA Vehicle: n = 7). The p values were calculated using the Fisher exact test. (E) Insulitis scores for NOD female mice treated weekly starting at 8 wk of age with vehicle (n = 10) or DIDS at a dose of 10 mg/kg (n = 8) or 50 mg/kg (n = 14). Data show mean ± SEM. The p values were calculated using Mann–Whitney analysis.

Close modal

The presence of IAAs is an important criterion that is used to identify humans at high future risk for T1D for inclusion in possible disease-intervention trials (13). Thus, just prior to treatment initiation, serum was collected from mice depicted in Fig. 7A and retrospectively typed for IAAs. Importantly, DIDS treatment prevented progression to overt T1D when initiated in already IAA+ NOD mice (IAA+: vehicle 2/3, DIDS 0/9; IAA: vehicle 6/7, DIDS 0/11) (Fig. 7D). Additionally, NOD mice treated with DIDS from 8 wk of age had decreased levels of insulitis compared with controls (Fig. 7E). These collective data indicate that DIDS treatment inhibits progression to overt T1D, even when initiated in NOD mice that had already developed high ongoing levels of β cell autoimmunity marked by the presence of IAAs.

To initially investigate how the reagent may elicit T1D protection, NOD mice treated with DIDS from 8 to 16 wk of age were evaluated for phenotypic changes in B lymphocyte populations. As observed following genetic ablation of AID, NOD mice treated with 50 mg/kg DIDS from 8 to 16 wk of age were characterized by increased numbers of total splenic and PLN B lymphocytes (Fig. 8A). Additionally, similar to the case elicited by genetic ablation of AID, DIDS treatment led to the expansion of splenic and PLN GC B lymphocyte compartments (Fig. 8B). B lymphocytes were also proportionally increased among islet-infiltrating lymphocytes (Fig. 8C). Further phenocopying NOD.Aicda−/− mice, the proportions of splenic, PLN, and islet-resident CD73+ B lymphocytes with potential immunosuppressive capacity were increased by DIDS treatment (Fig. 8D). Furthermore, the DIDS-elicited increase in islet CD73+ B lymphocytes occurred in the CD73+ CD39+ and CD73+ CD39 subpopulations (Fig. 8E). Finally, islet-infiltrating B lymphocytes had similar levels of CD69 but increased CD80 and PD-L2 expression (Fig. 8F). This was also similar to the expansion of B lymphocytes with a memory-like phenotype within the islets of NOD.Aicda−/− mice. Despite this observed expansion of islet-infiltrating memory-like B lymphocytes in DIDS-treated mice, flow cytometric analyses of bone marrow–resident B lymphocytes revealed that, relative to vehicle-treated controls, DIDS treatment led to a proportional decrease in recirculating mature (Hardy fraction F) B lymphocytes (Fig. 8G, Supplemental Fig. 2F).

FIGURE 8.

DIDS treatment alters B lymphocyte profiles in a manner similar to that elicited by Aicda ablation. NOD female mice were treated with vehicle or 50 mg/kg DIDS from 8 to 16 wk of age. (A) Yield of total splenic and PLN CD19+ B220+ B lymphocytes (n = 7 vehicle, n = 8 DIDS; combined from two experiments). (B) Percentage of splenic and PLN GC (Fas+ GL7+) B lymphocytes (n = 7 vehicle, n = 8 DIDS; combined from two experiments). (C) Percentage of B220+ leukocytes among islet CD45.1+ cells (combined from two experiments). (D) Percentage of CD73+ B cells among B220+ CD19+ B lymphocytes (n = 7 Vehicle Spleen/PLN, n = 8 DIDS Spleen/PLN, n = 13 Vehicle Islet, n = 14 DIDS Islet; combined from two experiments). (E) Percentage of CD73+ CD39+, CD73 CD39+, and CD73+ CD39 cells among islet B220+ lymphocytes (n = 13 Vehicle, n = 14 DIDS; combined from two experiments). (F) Percentage of CD69+, CD80+, and PD-L2+ cells among islet B220+ cells (n = 13 Vehicle, n = 14 DIDS; combined from two experiments). (G) Proportions of Hardy fraction D (B220+ CD43 IgM IgD), fraction E (B220+ CD43 IgM+ IgD), and fraction F (B220+ CD43 IgM+ IgD+) B lymphocyte subsets in bone marrow (one experiment). Mann–Whitney analysis was performed for all bar and scatter plot graphs; data are mean ± SEM.

FIGURE 8.

DIDS treatment alters B lymphocyte profiles in a manner similar to that elicited by Aicda ablation. NOD female mice were treated with vehicle or 50 mg/kg DIDS from 8 to 16 wk of age. (A) Yield of total splenic and PLN CD19+ B220+ B lymphocytes (n = 7 vehicle, n = 8 DIDS; combined from two experiments). (B) Percentage of splenic and PLN GC (Fas+ GL7+) B lymphocytes (n = 7 vehicle, n = 8 DIDS; combined from two experiments). (C) Percentage of B220+ leukocytes among islet CD45.1+ cells (combined from two experiments). (D) Percentage of CD73+ B cells among B220+ CD19+ B lymphocytes (n = 7 Vehicle Spleen/PLN, n = 8 DIDS Spleen/PLN, n = 13 Vehicle Islet, n = 14 DIDS Islet; combined from two experiments). (E) Percentage of CD73+ CD39+, CD73 CD39+, and CD73+ CD39 cells among islet B220+ lymphocytes (n = 13 Vehicle, n = 14 DIDS; combined from two experiments). (F) Percentage of CD69+, CD80+, and PD-L2+ cells among islet B220+ cells (n = 13 Vehicle, n = 14 DIDS; combined from two experiments). (G) Proportions of Hardy fraction D (B220+ CD43 IgM IgD), fraction E (B220+ CD43 IgM+ IgD), and fraction F (B220+ CD43 IgM+ IgD+) B lymphocyte subsets in bone marrow (one experiment). Mann–Whitney analysis was performed for all bar and scatter plot graphs; data are mean ± SEM.

Close modal

DIDS treatment and genetic ablation of Aicda elicit similar phenotypic alterations in NOD B lymphocytes. Although not elicited by Aicda ablation, we considered it important to determine whether DIDS treatment directly affects the diabetogenic capacity of NOD T cells. Total CD4+ and CD8+ T cells were numerically increased in the PLNs, but not spleens, of DIDS-treated mice (Fig. 9A). As observed upon genetic ablation of AID, the proportions of Tfh cells were increased in the spleens and PLNs of DIDS-treated NOD mice (Fig. 9B). The proportion of total T cells was decreased among islet-infiltrating leukocytes (Fig. 9C) after DIDS treatment. Additionally, there was a decrease in the proportion of the CD8+ subset among islet-infiltrating T cells (Fig. 9D). Islet CD4+ T cells displayed increased CD69 expression after DIDS treatment (Fig. 9E). DIDS treatment did not significantly alter CD69 expression of islet CD8+ T cells, but it marginally increased IL-7Rα levels (Fig. 9E). Finally, compared with vehicle-treated controls, in DIDS-treated mice, islet-infiltrating CD4+ T cells had a decreased naive and increased central memory phenotype, whereas CD8+ T cells had a decreased effector and an increased central memory phenotype (Fig. 9F). Together, these data indicate that DIDS treatment reduces effector CD8+ T cells and drives the accumulation of central memory CD4+ and CD8+ T cells within the islets of NOD mice.

FIGURE 9.

DIDS’ effects on B lymphocytes indirectly suppresses diabetogenic T cell responses. NOD female mice were treated weekly with 50 mg/kg DIDS or vehicle from 8 to 16 wk of age. (A) Percentage of CD4+ and CD8+ T cells among live TER-119 splenic or PLN-resident leukocytes (non-RBCs) (n = 7 Vehicle, n = 8 DIDS; combined from two experiments). (B) Quantification of percentage of full Tfh cells among CD4+ TCRβ+ cells in the spleen and PLN (n = 7 Vehicle, n = 8 DIDS; combined from two experiments). (C) Percentage of TCRβ+ leukocytes among CD45.1+ cells within the islets. (D) Percentage of CD4+ or CD8+ cells among islet T cells (n = 13 Vehicle, n = 14 DIDS; combined from two experiments). (E) Percentage of CD69+ and IL-7Rα+ cells among islet CD4+ and CD8+ T cells (n = 13 Vehicle, n = 14 DIDS; combined from two experiments). (F) Percentage of naive (CD44 CD62L+), effector (CD44+ CD62L), and central memory (CD44+ CD62L+) CD4+ and CD8+ T cells in the islets of DIDS- or vehicle-treated mice (n = 13 vehicle, n = 14 DIDS; combined from two experiments). (G) Female NOD mice were injected with vehicle or 50 mg/kg DIDS from 8 to 16 wk of age. Splenic T cells were then purified from each treatment group and transferred (3 × 106) into NOD-scid recipients (n = 16 per group) that were subsequently monitored for T1D development. (H) A total of 2.5 × 106 purified total T lymphocytes from 6-wk-old female NOD mice was transferred into NOD-scid mice that subsequently began weekly treatment with 0 or 50 mg/kg DIDS (n = 10 per group) and were monitored to T1D. Incidence study p values were calculated using Mantel–Cox analysis. Mann–Whitney analysis was performed for all bar graphs; data show mean ± SEM.

FIGURE 9.

DIDS’ effects on B lymphocytes indirectly suppresses diabetogenic T cell responses. NOD female mice were treated weekly with 50 mg/kg DIDS or vehicle from 8 to 16 wk of age. (A) Percentage of CD4+ and CD8+ T cells among live TER-119 splenic or PLN-resident leukocytes (non-RBCs) (n = 7 Vehicle, n = 8 DIDS; combined from two experiments). (B) Quantification of percentage of full Tfh cells among CD4+ TCRβ+ cells in the spleen and PLN (n = 7 Vehicle, n = 8 DIDS; combined from two experiments). (C) Percentage of TCRβ+ leukocytes among CD45.1+ cells within the islets. (D) Percentage of CD4+ or CD8+ cells among islet T cells (n = 13 Vehicle, n = 14 DIDS; combined from two experiments). (E) Percentage of CD69+ and IL-7Rα+ cells among islet CD4+ and CD8+ T cells (n = 13 Vehicle, n = 14 DIDS; combined from two experiments). (F) Percentage of naive (CD44 CD62L+), effector (CD44+ CD62L), and central memory (CD44+ CD62L+) CD4+ and CD8+ T cells in the islets of DIDS- or vehicle-treated mice (n = 13 vehicle, n = 14 DIDS; combined from two experiments). (G) Female NOD mice were injected with vehicle or 50 mg/kg DIDS from 8 to 16 wk of age. Splenic T cells were then purified from each treatment group and transferred (3 × 106) into NOD-scid recipients (n = 16 per group) that were subsequently monitored for T1D development. (H) A total of 2.5 × 106 purified total T lymphocytes from 6-wk-old female NOD mice was transferred into NOD-scid mice that subsequently began weekly treatment with 0 or 50 mg/kg DIDS (n = 10 per group) and were monitored to T1D. Incidence study p values were calculated using Mantel–Cox analysis. Mann–Whitney analysis was performed for all bar graphs; data show mean ± SEM.

Close modal

We next tested whether DIDS treatment had lasting effects on T cell diabetogenic activity. Female NOD mice, treated once weekly with vehicle or 50 mg/kg DIDS from 8 to 16 wk of age, then became donors of purified splenic T cells that were transferred into NOD-scid recipients that were subsequently monitored for T1D development. T cells from DIDS-treated mice transferred T1D to NOD-scid recipients with significantly less efficiency compared with those from control donors (Fig. 9G). We envisioned two possible (nonexclusive) explanations for this result: B lymphocytes in the DIDS-treated donors expand autoreactive T cell populations less efficiently than in vehicle-treated mice and DIDS can act directly to suppress diabetogenic T cell activity.

To distinguish between the above possibilities, we tested the direct effect of DIDS on the diabetogenic function of NOD T cells in the absence of B lymphocytes. Purified splenic T cells from 5–6-wk-old female NOD mice were transferred into NOD-scid recipients. Starting 3 d posttransfer, the NOD-scid recipients began weekly 50 mg/kg DIDS or vehicle treatment and were monitored for T1D development. After engraftment with NOD T cells, NOD-scid recipients treated with DIDS or vehicle developed T1D at an equivalent rate (Fig. 9H). These results indicate that DIDS treatment does not directly affect the diabetogenic activity of NOD T cells. Thus, the diminished ability of T cells from DIDS-treated NOD mice to adoptively transfer T1D cannot be attributed to a direct reduction in their pathogenicity; rather, it is likely due to DIDS limiting the ability of B lymphocytes to support the expansion of pathogenic effectors.

We tested whether coinfusion of B lymphocytes from control or DIDS-treated NOD mice differentially affected the ability of diabetogenic T cells to transfer disease to NOD-scid recipients. NOD female mice were injected once weekly, from 8 to 16 wk of age, with vehicle or 50 mg/kg DIDS. Purified splenic T cells from the vehicle-treated mice were then cotransferred into NOD-scid recipients with equal numbers of purified total B lymphocytes from vehicle or DIDS-treated donors. As previously noted, similar to the case of NOD.Aicda−/− mice, DIDS treatment expands CD73+ B lymphocyte populations. Therefore, we also transferred T cells from vehicle-treated controls with CD73-depleted B lymphocytes from DIDS-treated mice. T1D development was significantly decreased in NOD-scid recipients of pathogenic T cells coinfused with total B lymphocytes from DIDS-treated donors compared with vehicle-treated donors (Fig. 10A). However, depletion of the CD73+ subset significantly enhanced the ability of B lymphocytes from DIDS-treated donors to support T1D development in NOD-scid recipients coinfused with pathogenic T cells (Fig. 10A). Hence, similar to the case with genetic ablation of AID, DIDS treatment of NOD mice induces a quantitative increase in CD73+ B lymphocytes with a capacity to actively suppress T1D development (Fig. 10A).

FIGURE 10.

CD73+ B lymphocytes in DIDS-treated mice are T1D suppressive. NOD female mice were treated weekly with 50 mg/kg DIDS or vehicle from 8 to 16 wk of age. (A) NOD-scid recipients were infused with 2 × 106 purified splenic T cells from vehicle-treated donors admixed with an equal number of purified total splenic B lymphocytes from vehicle (Total VB), CD73-depleted B lymphocytes from DIDS-treated mice (CD73 DB), or total B lymphocytes from DIDS-treated mice (Total DB). The NOD-scid recipients were then monitored for T1D development over 8 wk. (B) A total of 1.0 × 105 NOD T cells was labeled with Cell Proliferation Dye eFluor 670 and cocultured for 4 d with 1.0 × 105 CD73+ or CD73 B lymphocytes from pooled spleens of vehicle-treated (n = 6 biological replicates) or DIDS-treated (n = 6 biological replicates) mice under stimulation conditions consisting of soluble anti-CD40 (1 μg/ml), plate-bound anti-CD3ε (5 μg/ml), and soluble anti-CD28 (2 μg/ml) with 0 (baseline) or 10 μM AMP in the presence or absence of 100 μM APCP. Quantification of percentage of suppression from baseline. (C) A total of 1 × 105 sort-purified CD73+ or CD73 B lymphocytes from vehicle- or DIDS-treated NOD mice (n = 6 biological replicates per group) were cultured for 3 d with 10 μg/ml LPS, and culture supernatant IL-10 levels were measured by ELISA. CD73-mediated in vitro suppression and IL-10 production data shown are each combined from two individual experiments. All bar graphs show mean ± SEM. Incidence study p value was calculated using Mantel–Cox analysis. The Wilcoxon test was performed for suppression assay. Mann–Whitney analysis was performed for IL-10 production.

FIGURE 10.

CD73+ B lymphocytes in DIDS-treated mice are T1D suppressive. NOD female mice were treated weekly with 50 mg/kg DIDS or vehicle from 8 to 16 wk of age. (A) NOD-scid recipients were infused with 2 × 106 purified splenic T cells from vehicle-treated donors admixed with an equal number of purified total splenic B lymphocytes from vehicle (Total VB), CD73-depleted B lymphocytes from DIDS-treated mice (CD73 DB), or total B lymphocytes from DIDS-treated mice (Total DB). The NOD-scid recipients were then monitored for T1D development over 8 wk. (B) A total of 1.0 × 105 NOD T cells was labeled with Cell Proliferation Dye eFluor 670 and cocultured for 4 d with 1.0 × 105 CD73+ or CD73 B lymphocytes from pooled spleens of vehicle-treated (n = 6 biological replicates) or DIDS-treated (n = 6 biological replicates) mice under stimulation conditions consisting of soluble anti-CD40 (1 μg/ml), plate-bound anti-CD3ε (5 μg/ml), and soluble anti-CD28 (2 μg/ml) with 0 (baseline) or 10 μM AMP in the presence or absence of 100 μM APCP. Quantification of percentage of suppression from baseline. (C) A total of 1 × 105 sort-purified CD73+ or CD73 B lymphocytes from vehicle- or DIDS-treated NOD mice (n = 6 biological replicates per group) were cultured for 3 d with 10 μg/ml LPS, and culture supernatant IL-10 levels were measured by ELISA. CD73-mediated in vitro suppression and IL-10 production data shown are each combined from two individual experiments. All bar graphs show mean ± SEM. Incidence study p value was calculated using Mantel–Cox analysis. The Wilcoxon test was performed for suppression assay. Mann–Whitney analysis was performed for IL-10 production.

Close modal

Next, we tested whether the CD73+ population expanded by DIDS treatment can directly suppress T cell responses. On a per-cell basis, CD73+ B lymphocytes from vehicle- and DIDS-treated mice equally suppressed T cell proliferation in response to anti-CD3ε and anti-CD28 stimulation to a significantly greater extent than did the CD73 subset in the presence of AMP (Fig. 10B). This suppressive effect was diminished upon addition of APCP (Fig. 10B, Supplemental Fig. 3B). Finally, we examined IL-10 production by sorted CD73+ and CD73 B lymphocytes from vehicle- and DIDS-treated mice. Surprisingly, CD73+ B lymphocytes from DIDS-treated mice produced less IL-10 upon LPS stimulation than did those from vehicle-treated controls (Fig. 10C). These results indicate that pharmacological targeting of RAD51 activity inhibits T1D development in NOD mice in a manner similar to that elicited by Aicda ablation, including a quantitative increase in CD73+ regulatory B lymphocytes (Bregs).

We have demonstrated that disruption of the AID/RAD51 axis, through genetic or pharmacological means, strongly inhibits T1D development in NOD mice, and this protection is due, at least in part, to the expansion of specific CD73+ B lymphocyte populations with capacities to regulate pathogenic T cell responses. Additionally, we provide the first evidence, to our knowledge, that, although CSR and SHM processes are important B lymphocyte–intrinsic processes in T1D pathogenesis, their absence does not diminish autoreactive T cell development. Although purified T cells from NOD.Aicda−/− mice are effective at adoptively transferring T1D to NOD-scid recipients, inhibiting CSR and SHM processes leads to an expansion of Bregs controlling these effectors. These findings reveal an unexpected link between blocking AID/RAD51-dependent affinity-maturation processes and Breg development. Thus, disruption of the AID/RAD51 axis may represent a previously unrealized means of in vivo Breg expansion.

Ablation of AID and DIDS targeting of RAD51 results in accumulation of CD73+ B lymphocytes capable of suppressing T cells, at least in part through the production of adenosine (Figs. 5C, 10B). Previous studies demonstrated that adenosine signaling via A2aR inhibits CD8+ T cell downregulation of IL-7Rα, preventing the differentiation of memory T cells to effectors (50). Therefore, the shift in islet-infiltrating CD8+ T cell populations in DIDS-treated NOD mice from an effector to a central memory phenotype (Fig. 9F), paired with increased IL-7Rα expression (Fig. 9E), further supports the conclusion that, in this system, diabetogenic T cells are suppressed through an adenosine-mediated mechanism. Therefore, although not precluding the possibility of other nonoverlapping adenosine-independent suppression pathways, the population of CD73+ B lymphocytes expanded in AID-deficient or DIDS-treated NOD mice appears to largely inhibit diabetogenic T cell responses through a mechanism that is dependent on activity of this ectoenzyme.

It should also be noted that, although CD73-depleted B lymphocytes from DIDS-treated mice could support the ability of coinfused pathogenic T cells to transfer T1D to NOD-scid recipients, they did so less efficiently than total B lymphocytes from vehicle-treated control donors. This could indicate that, in addition to expanding CD73+ Bregs, DIDS treatment diminishes the ability of other B lymphocyte populations in NOD mice to support diabetogenic T cell activity. This could include DIDS treatment supporting the expansion of CD73 Bregs capable of suppressing T cells through adenosine-independent mechanisms. Should this be the case, the mechanism by which such CD73 Bregs mediate suppression is likely to be IL-10 independent, because this population produces little of the cytokine in response to stimulation.

It is unknown why CD73+ B lymphocytes from DIDS-treated mice produce less LPS-induced IL-10 than do those from controls. IL-10 secretion has been recognized as a major means of Breg-mediated immune suppression (51). However, there have also been reports of Bregs capable of suppressing systemic lupus erythematosus and experimental autoimmune encephalomyelitis development through IL-10–independent mechanisms (52). IL-10 production by B lymphocytes requires strong stimulation, whereas adenosine generation by CD73+ B lymphocytes is constitutive, providing that its substrate is present (46). The ability of B lymphocytes from DIDS-treated NOD mice to inhibit diabetogenic T cell responses (Fig. 10A, 10B), despite decreased IL-10 production (Fig. 10C), supports the conclusion that such disease-protective effects are largely the result of CD73-mediated adenosine production. This observation is consistent with a report suggesting that Breg-mediated adenosine production may play a more important role in immunosuppression than IL-10 secretion (46). Additionally, it should be noted that IL-10−/− B lymphocytes have reduced CD73 expression, suggesting that this cytokine likely directly regulates CD73 expression by B lymphocytes (46). Therefore, DIDS treatment of mice made genetically deficient in IL-10 expression or treated with an Ab blocking this cytokine would not discriminate between the regulatory contributions of CD73-mediated adenosine production and direct action by IL-10. The future creation of B lymphocyte–specific CD73−/− NOD mice will help to uncouple the individual regulatory contributions provided by adenosine generation and IL-10 secretion.

This study has focused on the role of the AID/RAD51 axis in GC affinity-maturation processes contributing to T1D. However, the AID pathway has been shown to play an important role in B lymphocytes outside of the splenic or lymph node GC environment. For example, AID has been implicated in the central tolerance of B lymphocytes in C57BL/6 and BALB/c background strains (53). Another study demonstrated a role for AID in central and peripheral tolerance of human B lymphocytes (54). However, discerning the contribution of AID to these processes in NOD mice is complicated by strain-specific defects in B lymphocyte central tolerance (55). Therefore, the effect of targeting AID/RAD51 on B lymphocyte central tolerance in NOD is unclear and warrants further investigation. Additionally, manipulation of AID affects gut-associated B lymphocytes and alters intestinal microflora (56). Gut microflora alterations can drastically impact T1D development (57). Thus, examining the impact of disrupting the AID/RAD51 axis on GALT homeostasis is warranted. Additionally, disruption of the AID/RAD51 axis could conceivably result in a decreased ability to clear infection. Due to institutional policy, we cannot infect mice with microbes at The Jackson Laboratory. Therefore, the possible effects of targeting the AID/RAD51 axis on the ability to clear pathogens will need to be the subject of other investigators’ research. Additionally, agents that might specifically increase CD73+ Bregs without impacting the AID/RAD51 axis should be investigated.

A previous rituximab-mediated pan–B lymphocyte–targeting clinical trial did not permanently halt β cell demise (17). Studies in NOD mice indicate that this may be due, at least in part, to downregulation of cell surface CD20 expression by islet-infiltrating B lymphocytes (18). This provides a likely explanation for why anti-CD20 immunotherapy only protects NOD mice from T1D if initiated prior to IAA development. Other studies have provided evidence that Breg expansion has a potential to confer strong T1D inhibitory effects, and pan-B lymphocyte–depletion regimens could deplete these protective populations (58). Our results provide the first indication, to our knowledge, that future clinically applicable pharmaceutical agents targeting the AID/RAD51 axis could convert some B lymphocytes to a T1D-protective CD73+ regulatory phenotype. Furthermore, because this immunomodulatory therapy retains efficacy, even when initiated at a late prodromal autoantibody-positive stage of T1D development, it also represents a significant improvement upon previous B lymphocyte–targeted disease-intervention strategies.

In summary, these studies show that the genetic or pharmacologic blockade of B lymphocyte affinity-maturation processes in NOD mice drives diversion to a CD73+ regulatory phenotype capable of inhibiting autoimmune T1D development. Therefore, pharmacological targeting of RAD51 to block diabetogenic B lymphocyte activity, either directly or by converting them to an immunoregulatory state, might ultimately represent a clinically translatable disease-intervention approach. Because RAD51 is a multiprotein complex, this area of the AID pathway has other potential targets that might be exploited. The identification and use of pharmacological agents that potentially directly block AID activity as a novel T1D-intervention approach should also be explored. These studies also indicate that further research into the role of the purinergic immunoregulatory pathway in T1D pathogenesis is warranted. Furthermore, this pathway could also represent a potentially clinically relevant immunomodulatory target for the treatment of various other autoimmune diseases, either through increasing adenosine production or through administration of A2aR agonists. Together, these initial studies reveal that therapeutic targeting of the AID/RAD51 axis in B lymphocytes is a previously unrealized area of research for T1D therapy development.

We thank staff within The Jackson Laboratory’s Genome Technologies group, Genetic Engineering Technologies group, Flow Cytometry service, and Research Animal Facility for technical support. We also thank Susanne Sattler (Imperial College, London, U.K.) for critical review of the manuscript.

This work was supported by National Cancer Institute Grant P30CA034196. M.A.A. is supported by National Institutes of Health Grant P01-AI42288. D.V.S. is supported by National Institutes of Health Grants DK-46266, DK-95735, and OD-020351. C.M.L. is supported by National Institutes of Health Grant DK101735. K.D.M. is supported by National Institutes of Health Grant CA138646 and The Jackson Laboratory Principal Investigator Grant TJL DIF FY13 KDM. J.J. Racine is supported by National Institutes of Health Fellowship 1F32DK111078.

The online version of this article contains supplemental material.

Abbreviations used in this article:

Aicda

activation-induced cytidine deaminase gene

AID

activation-induced cytidine deaminase protein

APCP

α,β-methyleneadenosine 5′-diphosphate

A2aR

adenosine A2a receptor

B6

C57BL/6J

Breg

regulatory B lymphocyte

CSR

class switch recombination

DIDS

4,4′-diisothiocyanatostilbene-2-2′disulfonic acid

DSB

dsDNA break

GC

germinal center

HR

homologous recombination

IAA

insulin autoantibody

NOD-scid

NOD.Cg-PrkdcscidEmv30b/Dvs

PLN

pancreatic lymph node

qPCR

quantitative PCR

sgRNA

single-guide RNA

SHM

somatic hypermutation

Tfh

follicular helper T

T1

transitional-1

T1D

type 1 diabetes

T2

transitional-2

UMI

unique molecular identifier

WT

wild-type.

1
Serreze
D. V.
,
Chapman
H. D.
,
Varnum
D. S.
,
Hanson
M. S.
,
Reifsnyder
P. C.
,
Richard
S. D.
,
Fleming
S. A.
,
Leiter
E. H.
,
Shultz
L. D.
.
1996
.
B lymphocytes are essential for the initiation of T cell-mediated autoimmune diabetes: analysis of a new “speed congenic” stock of NOD.Ig mu null mice.
J. Exp. Med.
184
:
2049
2053
.
2
Akashi
T.
,
Nagafuchi
S.
,
Anzai
K.
,
Kondo
S.
,
Kitamura
D.
,
Wakana
S.
,
Ono
J.
,
Kikuchi
M.
,
Niho
Y.
,
Watanabe
T.
.
1997
.
Direct evidence for the contribution of B cells to the progression of insulitis and the development of diabetes in non-obese diabetic mice.
Int. Immunol.
9
:
1159
1164
.
3
Wong
F. S.
,
Visintin
I.
,
Wen
L.
,
Granata
J.
,
Flavell
R.
,
Janeway
C. A.
.
1998
.
The role of lymphocyte subsets in accelerated diabetes in nonobese diabetic-rat insulin promoter-B7-1 (NOD-RIP-B7-1) mice.
J. Exp. Med.
187
:
1985
1993
.
4
Noorchashm
H.
,
Noorchashm
N.
,
Kern
J.
,
Rostami
S. Y.
,
Barker
C. F.
,
Naji
A.
.
1997
.
B-cells are required for the initiation of insulitis and sialitis in nonobese diabetic mice.
Diabetes
46
:
941
946
.
5
Bouaziz
J. D.
,
Yanaba
K.
,
Venturi
G. M.
,
Wang
Y.
,
Tisch
R. M.
,
Poe
J. C.
,
Tedder
T. F.
.
2007
.
Therapeutic B cell depletion impairs adaptive and autoreactive CD4+ T cell activation in mice.
Proc. Natl. Acad. Sci. USA
104
:
20878
20883
.
6
Fiorina
P.
,
Vergani
A.
,
Dada
S.
,
Jurewicz
M.
,
Wong
M.
,
Law
K.
,
Wu
E.
,
Tian
Z.
,
Abdi
R.
,
Guleria
I.
, et al
.
2008
.
Targeting CD22 reprograms B-cells and reverses autoimmune diabetes.
Diabetes
57
:
3013
3024
.
7
Hu
C. Y.
,
Rodriguez-Pinto
D.
,
Du
W.
,
Ahuja
A.
,
Henegariu
O.
,
Wong
F. S.
,
Shlomchik
M. J.
,
Wen
L.
.
2007
.
Treatment with CD20-specific antibody prevents and reverses autoimmune diabetes in mice.
J. Clin. Invest.
117
:
3857
3867
.
8
Xiu
Y.
,
Wong
C. P.
,
Bouaziz
J. D.
,
Hamaguchi
Y.
,
Wang
Y.
,
Pop
S. M.
,
Tisch
R. M.
,
Tedder
T. F.
.
2008
.
B lymphocyte depletion by CD20 monoclonal antibody prevents diabetes in nonobese diabetic mice despite isotype-specific differences in Fc gamma R effector functions.
J. Immunol.
180
:
2863
2875
.
9
Pearson
J. A.
,
Wong
F. S.
,
Wen
L.
.
2016
.
The importance of the Non Obese Diabetic (NOD) mouse model in autoimmune diabetes.
J. Autoimmun.
66
:
76
88
.
10
Hulbert
C.
,
Riseili
B.
,
Rojas
M.
,
Thomas
J. W.
.
2001
.
B cell specificity contributes to the outcome of diabetes in nonobese diabetic mice.
J. Immunol.
167
:
5535
5538
.
11
Mariño
E.
,
Tan
B.
,
Binge
L.
,
Mackay
C. R.
,
Grey
S. T.
.
2012
.
B-cell cross-presentation of autologous antigen precipitates diabetes. [Published erratum appears in 2013 Diabetes 62: 662.]
Diabetes
61
:
2893
2905
.
12
Silveira
P. A.
,
Johnson
E.
,
Chapman
H. D.
,
Bui
T.
,
Tisch
R. M.
,
Serreze
D. V.
.
2002
.
The preferential ability of B lymphocytes to act as diabetogenic APC in NOD mice depends on expression of self-antigen-specific immunoglobulin receptors.
Eur. J. Immunol.
32
:
3657
3666
.
13
Bonifacio
E.
,
Ziegler
A. G.
.
2010
.
Advances in the prediction and natural history of type 1 diabetes.
Endocrinol. Metab. Clin. North Am.
39
:
513
525
.
14
Victora
G. D.
,
Nussenzweig
M. C.
.
2012
.
Germinal centers.
Annu. Rev. Immunol.
30
:
429
457
.
15
Murphy
K.
,
Travers
P.
,
Walport
M.
,
Janeway
C.
.
2008
.
Janeway’s Immunobiology.
Garland Science
,
New York
.
16
Wan
X.
,
Thomas
J. W.
,
Unanue
E. R.
.
2016
.
Class-switched anti-insulin antibodies originate from unconventional antigen presentation in multiple lymphoid sites.
J. Exp. Med.
213
:
967
978
.
17
Pescovitz
M. D.
,
Greenbaum
C. J.
,
Krause-Steinrauf
H.
,
Becker
D. J.
,
Gitelman
S. E.
,
Goland
R.
,
Gottlieb
P. A.
,
Marks
J. B.
,
McGee
P. F.
,
Moran
A. M.
, et al
Type 1 Diabetes TrialNet Anti-CD20 Study Group
.
2009
.
Rituximab, B-lymphocyte depletion, and preservation of beta-cell function.
N. Engl. J. Med.
361
:
2143
2152
.
18
Serreze
D. V.
,
Chapman
H. D.
,
Niens
M.
,
Dunn
R.
,
Kehry
M. R.
,
Driver
J. P.
,
Haller
M.
,
Wasserfall
C.
,
Atkinson
M. A.
.
2011
.
Loss of intra-islet CD20 expression may complicate efficacy of B-cell-directed type 1 diabetes therapies.
Diabetes
60
:
2914
2921
.
19
Chaudhuri
J.
,
Tian
M.
,
Khuong
C.
,
Chua
K.
,
Pinaud
E.
,
Alt
F. W.
.
2003
.
Transcription-targeted DNA deamination by the AID antibody diversification enzyme.
Nature
422
:
726
730
.
20
McBride
K. M.
,
Gazumyan
A.
,
Woo
E. M.
,
Barreto
V. M.
,
Robbiani
D. F.
,
Chait
B. T.
,
Nussenzweig
M. C.
.
2006
.
Regulation of hypermutation by activation-induced cytidine deaminase phosphorylation.
Proc. Natl. Acad. Sci. USA
103
:
8798
8803
.
21
Muramatsu
M.
,
Kinoshita
K.
,
Fagarasan
S.
,
Yamada
S.
,
Shinkai
Y.
,
Honjo
T.
.
2000
.
Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme.
Cell
102
:
553
563
.
22
Hasham
M. G.
,
Donghia
N. M.
,
Coffey
E.
,
Maynard
J.
,
Snow
K. J.
,
Ames
J.
,
Wilpan
R. Y.
,
He
Y.
,
King
B. L.
,
Mills
K. D.
.
2010
.
Widespread genomic breaks generated by activation-induced cytidine deaminase are prevented by homologous recombination.
Nat. Immunol.
11
:
820
826
.
23
Lamont
K. R.
,
Hasham
M. G.
,
Donghia
N. M.
,
Branca
J.
,
Chavaree
M.
,
Chase
B.
,
Breggia
A.
,
Hedlund
J.
,
Emery
I.
,
Cavallo
F.
, et al
.
2013
.
Attenuating homologous recombination stimulates an AID-induced antileukemic effect.
J. Exp. Med.
210
:
1021
1033
.
24
Serreze
D. V.
,
Leiter
E. H.
,
Hanson
M. S.
,
Christianson
S. W.
,
Shultz
L. D.
,
Hesselton
R. M.
,
Greiner
D. L.
.
1995
.
Emv30null NOD-scid mice. An improved host for adoptive transfer of autoimmune diabetes and growth of human lymphohematopoietic cells.
Diabetes
44
:
1392
1398
.
25
Hill
J. T.
,
Demarest
B. L.
,
Bisgrove
B. W.
,
Su
Y. C.
,
Smith
M.
,
Yost
H. J.
.
2014
.
Poly peak parser: method and software for identification of unknown indels using sanger sequencing of polymerase chain reaction products.
Dev. Dyn.
243
:
1632
1636
.
26
Morgan
H. D.
,
Dean
W.
,
Coker
H. A.
,
Reik
W.
,
Petersen-Mahrt
S. K.
.
2004
.
Activation-induced cytidine deaminase deaminates 5-methylcytosine in DNA and is expressed in pluripotent tissues: implications for epigenetic reprogramming.
J. Biol. Chem.
279
:
52353
52360
.
27
Cartwright
R.
,
Dunn
A. M.
,
Simpson
P. J.
,
Tambini
C. E.
,
Thacker
J.
.
1998
.
Isolation of novel human and mouse genes of the recA/RAD51 recombination-repair gene family.
Nucleic Acids Res.
26
:
1653
1659
.
28
Johnson
E. A.
,
Silveira
P.
,
Chapman
H. D.
,
Leiter
E. H.
,
Serreze
D. V.
.
2001
.
Inhibition of autoimmune diabetes in nonobese diabetic mice by transgenic restoration of H2-E MHC class II expression: additive, but unequal, involvement of multiple APC subtypes.
J. Immunol.
167
:
2404
2410
.
29
Julius
M. H.
,
Herzenberg
L. A.
.
1974
.
Isolation of antigen-binding cells from unprimed mice: demonstration of antibody-forming cell precursor activity and correlation between precursor and secreted antibody avidities.
J. Exp. Med.
140
:
904
920
.
30
Silveira
P. A.
,
Chapman
H. D.
,
Stolp
J.
,
Johnson
E.
,
Cox
S. L.
,
Hunter
K.
,
Wicker
L. S.
,
Serreze
D. V.
.
2006
.
Genes within the Idd5 and Idd9/11 diabetes susceptibility loci affect the pathogenic activity of B cells in nonobese diabetic mice.
J. Immunol.
177
:
7033
7041
.
31
Christianson
G. J.
,
Brooks
W.
,
Vekasi
S.
,
Manolfi
E. A.
,
Niles
J.
,
Roopenian
S. L.
,
Roths
J. B.
,
Rothlein
R.
,
Roopenian
D. C.
.
1997
.
Beta 2-microglobulin-deficient mice are protected from hypergammaglobulinemia and have defective antibody responses because of increased IgG catabolism.
J. Immunol.
159
:
4781
4792
.
32
Bonifacio
E.
,
Atkinson
M.
,
Eisenbarth
G.
,
Serreze
D.
,
Kay
T. W.
,
Lee-Chan
E.
,
Singh
B.
.
2001
.
International workshop on lessons from animal models for human type 1 diabetes: identification of insulin but not glutamic acid decarboxylase or IA-2 as specific autoantigens of humoral autoimmunity in nonobese diabetic mice.
Diabetes
50
:
2451
2458
.
33
Turchaninova
M. A.
,
Davydov
A.
,
Britanova
O. V.
,
Shugay
M.
,
Bikos
V.
,
Egorov
E. S.
,
Kirgizova
V. I.
,
Merzlyak
E. M.
,
Staroverov
D. B.
,
Bolotin
D. A.
, et al
.
2016
.
High-quality full-length immunoglobulin profiling with unique molecular barcoding.
Nat. Protoc.
11
:
1599
1616
.
34
Shugay
M.
,
Britanova
O. V.
,
Merzlyak
E. M.
,
Turchaninova
M. A.
,
Mamedov
I. Z.
,
Tuganbaev
T. R.
,
Bolotin
D. A.
,
Staroverov
D. B.
,
Putintseva
E. V.
,
Plevova
K.
, et al
.
2014
.
Towards error-free profiling of immune repertoires.
Nat. Methods
11
:
653
655
.
35
Bolotin
D. A.
,
Poslavsky
S.
,
Mitrophanov
I.
,
Shugay
M.
,
Mamedov
I. Z.
,
Putintseva
E. V.
,
Chudakov
D. M.
.
2015
.
MiXCR: software for comprehensive adaptive immunity profiling.
Nat. Methods
12
:
380
381
.
36
Shugay
M.
,
Bagaev
D. V.
,
Turchaninova
M. A.
,
Bolotin
D. A.
,
Britanova
O. V.
,
Putintseva
E. V.
,
Pogorelyy
M. V.
,
Nazarov
V. I.
,
Zvyagin
I. V.
,
Kirgizova
V. I.
, et al
.
2015
.
VDJtools: unifying post-analysis of T cell receptor repertoires.
PLoS Comput. Biol.
11
:
e1004503
.
37
Luzina
I. G.
,
Atamas
S. P.
,
Storrer
C. E.
,
daSilva
L. C.
,
Kelsoe
G.
,
Papadimitriou
J. C.
,
Handwerger
B. S.
.
2001
.
Spontaneous formation of germinal centers in autoimmune mice.
J. Leukoc. Biol.
70
:
578
584
.
38
Petro
J. B.
,
Gerstein
R. M.
,
Lowe
J.
,
Carter
R. S.
,
Shinners
N.
,
Khan
W. N.
.
2002
.
Transitional type 1 and 2 B lymphocyte subsets are differentially responsive to antigen receptor signaling.
J. Biol. Chem.
277
:
48009
48019
.
39
Su
T. T.
,
Rawlings
D. J.
.
2002
.
Transitional B lymphocyte subsets operate as distinct checkpoints in murine splenic B cell development.
J. Immunol.
168
:
2101
2110
.
40
Kuraoka
M.
,
Holl
T. M.
,
Liao
D.
,
Womble
M.
,
Cain
D. W.
,
Reynolds
A. E.
,
Kelsoe
G.
.
2011
.
Activation-induced cytidine deaminase mediates central tolerance in B cells.
Proc. Natl. Acad. Sci. USA
108
:
11560
11565
.
41
Dahlberg
C. I.
,
He
M.
,
Visnes
T.
,
Torres
M. L.
,
Cortizas
E. M.
,
Verdun
R. E.
,
Westerberg
L. S.
,
Severinson
E.
,
Ström
L.
.
2014
.
A novel mouse model for the hyper-IgM syndrome: a spontaneous activation-induced cytidine deaminase mutation leading to complete loss of Ig class switching and reduced somatic hypermutation.
J. Immunol.
193
:
4732
4738
.
42
Tomayko
M. M.
,
Steinel
N. C.
,
Anderson
S. M.
,
Shlomchik
M. J.
.
2010
.
Cutting edge: hierarchy of maturity of murine memory B cell subsets.
J. Immunol.
185
:
7146
7150
.
43
Revy
P.
,
Muto
T.
,
Levy
Y.
,
Geissmann
F.
,
Plebani
A.
,
Sanal
O.
,
Catalan
N.
,
Forveille
M.
,
Dufourcq-Labelouse
R.
,
Gennery
A.
, et al
.
2000
.
Activation-induced cytidine deaminase (AID) deficiency causes the autosomal recessive form of the Hyper-IgM syndrome (HIGM2).
Cell
102
:
565
575
.
44
Antonioli
L.
,
Pacher
P.
,
Vizi
E. S.
,
Haskó
G.
.
2013
.
CD39 and CD73 in immunity and inflammation.
Trends Mol. Med.
19
:
355
367
.
45
Cekic
C.
,
Linden
J.
.
2016
.
Purinergic regulation of the immune system.
Nat. Rev. Immunol.
16
:
177
192
.
46
Kaku
H.
,
Cheng
K. F.
,
Al-Abed
Y.
,
Rothstein
T. L.
.
2014
.
A novel mechanism of B cell-mediated immune suppression through CD73 expression and adenosine production.
J. Immunol.
193
:
5904
5913
.
47
Kleffel
S.
,
Vergani
A.
,
Tezza
S.
,
Ben Nasr
M.
,
Niewczas
M. A.
,
Wong
S.
,
Bassi
R.
,
D’Addio
F.
,
Schatton
T.
,
Abdi
R.
, et al
.
2015
.
Interleukin-10+ regulatory B cells arise within antigen-experienced CD40+ B cells to maintain tolerance to islet autoantigens.
Diabetes
64
:
158
171
.
48
Colwell
R. K.
,
Chao
A.
,
Gotelli
N. J.
,
Lin
S.-Y.
,
Mao
C. X.
,
Chazdon
R. L.
,
Longino
J. T.
.
2012
.
Models and estimators linking individual-based and sample-based rarefaction, extrapolation and comparison of assemblages.
J. Plant Ecol.
5
:
3
21
.
49
Britanova
O. V.
,
Putintseva
E. V.
,
Shugay
M.
,
Merzlyak
E. M.
,
Turchaninova
M. A.
,
Staroverov
D. B.
,
Bolotin
D. A.
,
Lukyanov
S.
,
Bogdanova
E. A.
,
Mamedov
I. Z.
, et al
.
2014
.
Age-related decrease in TCR repertoire diversity measured with deep and normalized sequence profiling.
J. Immunol.
192
:
2689
2698
.
50
Bono
M. R.
,
Fernández
D.
,
Flores-Santibáñez
F.
,
Rosemblatt
M.
,
Sauma
D.
.
2015
.
CD73 and CD39 ectonucleotidases in T cell differentiation: beyond immunosuppression.
FEBS Lett.
589
:
3454
3460
.
51
Mauri
C.
,
Bosma
A.
.
2012
.
Immune regulatory function of B cells.
Annu. Rev. Immunol.
30
:
221
241
.
52
Ray
A.
,
Basu
S.
,
Williams
C. B.
,
Salzman
N. H.
,
Dittel
B. N.
.
2012
.
A novel IL-10-independent regulatory role for B cells in suppressing autoimmunity by maintenance of regulatory T cells via GITR ligand.
J. Immunol.
188
:
3188
3198
.
53
Berland
R.
,
Fernandez
L.
,
Kari
E.
,
Han
J. H.
,
Lomakin
I.
,
Akira
S.
,
Wortis
H. H.
,
Kearney
J. F.
,
Ucci
A. A.
,
Imanishi-Kari
T.
.
2006
.
Toll-like receptor 7-dependent loss of B cell tolerance in pathogenic autoantibody knockin mice.
Immunity
25
:
429
440
.
54
Meyers
G.
,
Ng
Y. S.
,
Bannock
J. M.
,
Lavoie
A.
,
Walter
J. E.
,
Notarangelo
L. D.
,
Kilic
S. S.
,
Aksu
G.
,
Debré
M.
,
Rieux-Laucat
F.
, et al
.
2011
.
Activation-induced cytidine deaminase (AID) is required for B-cell tolerance in humans.
Proc. Natl. Acad. Sci. USA
108
:
11554
11559
.
55
Henry-Bonami
R. A.
,
Williams
J. M.
,
Rachakonda
A. B.
,
Karamali
M.
,
Kendall
P. L.
,
Thomas
J. W.
.
2013
.
B lymphocyte “original sin” in the bone marrow enhances islet autoreactivity in type 1 diabetes-prone nonobese diabetic mice.
J. Immunol.
190
:
5992
6003
.
56
Wei
M.
,
Shinkura
R.
,
Doi
Y.
,
Maruya
M.
,
Fagarasan
S.
,
Honjo
T.
.
2011
.
Mice carrying a knock-in mutation of Aicda resulting in a defect in somatic hypermutation have impaired gut homeostasis and compromised mucosal defense.
Nat. Immunol.
12
:
264
270
.
57
Kriegel
M. A.
,
Sefik
E.
,
Hill
J. A.
,
Wu
H. J.
,
Benoist
C.
,
Mathis
D.
.
2011
.
Naturally transmitted segmented filamentous bacteria segregate with diabetes protection in nonobese diabetic mice.
Proc. Natl. Acad. Sci. USA
108
:
11548
11553
.
58
Matsushita
T.
,
Yanaba
K.
,
Bouaziz
J. D.
,
Fujimoto
M.
,
Tedder
T. F.
.
2008
.
Regulatory B cells inhibit EAE initiation in mice while other B cells promote disease progression.
J. Clin. Invest.
118
:
3420
3430
.

K.D.M. is a founder and shareholder in Cyteir Therapeutics, Inc. K.D.M., M.G.H., and C.M.L. hold United States Patent No. US20130184342 A1: “Methods and compositions for treatment of cancer and autoimmune disease.” The other authors have no financial conflicts of interest.

Supplementary data