Targeting CD22 reprograms B-cells and reverses autoimmune diabetes

Abstract

Objectives: To investigate a B-cell-depleting strategy to reverse diabetes in naïve NOD mice.

Research design and methods: We targeted the CD22 receptor on B-cells of naïve NOD mice to deplete and reprogram B-cells to effectively reverse autoimmune diabetes.

Results: Anti-CD22/cal monoclonal antibody (mAb) therapy resulted in early and prolonged B-cell depletion and delayed disease in pre-diabetic mice. Importantly, when new-onset hyperglycemic mice were treated with the anti-CD22/cal mAb, 100% of B-cell-depleted mice became normoglycemic by 2 days, and 70% of them maintained a state of long-term normoglycemia. Early therapy after onset of hyperglycemia and complete B-cell depletion are essential for optimal efficacy. Treated mice showed an increase in percentage of regulatory T-cells in islets and pancreatic lymph nodes and a diminished immune response to islet peptides in vitro. Transcriptome analysis of reemerging B-cells showed significant changes of a set of proinflammatory genes. Functionally, reemerging B-cells failed to present autoantigen and prevented diabetes when cotransferred with autoreactive CD4(+) T-cells into NOD.SCID hosts.

Conclusions: Targeting CD22 depletes and reprograms B-cells and reverses autoimmune diabetes, thereby providing a blueprint for development of novel therapies to cure autoimmune diabetes.

FIG. 1.

Depletion studies. Splenocytes were extracted from normoglycemic 10-week-old NOD mice (n = 5) and were analyzed by flow cytometry for CD19 and CD22 expression on B220⁺ cells and CD138⁺ cells (plasma cells). CD19 and CD22 were similarly expressed on B220⁺ cells (A and B), and CD22 was expressed on CD138⁺ cells (C). We then examined by flow cytometry the infiltrating cells in the pancreata of 4-, 8-, and 12-week-old and hyperglycemic NOD mice (>14 weeks old) (n = 5 mice/group). Most of the infiltrate is constituted by CD45⁺CD19⁺ cells (B-cells) (D). B-cell pancreatic infiltration in NOD mice peaked around 8–10 weeks (P < 0.05; D), whereas CD45⁺CD3⁺ cells (T-cells) remained stable over time (E). The percentage of CD45⁺CD19⁺ cells (B-cells) was significantly higher than CD45⁺CD3⁺ cells (T-cells) in the pancreata of 8-week-old NOD mice (B-cells, 65.1 ± 5.0 vs. T-cells, 30.2 ± 3.2%, P = 0.004) (E). Two injections (160 μg i.p. 5 days apart, day 0 and day 5) of anti-CD22/cal mAb elicits a quick and profound depletion of B-cells in the peripheral blood of 10-week-old NOD mice (n = 6 mice/group) by 1 week that lasts for 6–7 weeks (F and H). Control NOD mice did not appear to be depleted (F and G), whereas the group treated with unconjugated anti-CD22 mAb shows a transient and partial B-cell depletion (F and I). At 8–10 weeks after depletion, B-cells recovered almost completely (F and H).

FIG. 2.

Diabetes prevention studies. We observed a significant delay in diabetes onset in anti-CD22/cal mAb–treated female 10-week-old NOD mice (n = 20) compared with controls (n = 30, P < 0.01) (A). The calicheamicin alone–treated group developed diabetes similarly to untreated controls (n = 10, P < 0.01 vs. anti-CD22/cal mAb–treated NOD mice) (A). Unconjugated anti-CD22 treatment slightly delayed diabetes onset (n = 10, P = 0.06 vs. untreated controls) (A). At 35 weeks of age, an increase in the percentage of CD4⁺CD25⁺FoxP3⁺ cells was evident in the pancreatic lymph nodes of anti-CD22/cal mAb–treated NOD mice (n = 4) compared with 10-week-old untreated control NOD (n = 4, P = 0.02) and compared with hyperglycemic >14-week-old NOD mice (n = 4, P = 0.009) (B). CD4⁺ cells extracted from splenocytes of anti-CD22/cal mAb–treated NOD mice at 35 weeks of age produced less IFN-γ when challenged with the BDC2.5 peptide compared with CD4⁺ cells extracted from splenocytes of untreated age-matched control hyperglycemic NOD mice (P = 0.001) and 10-week-old NOD mice (P = 0.04) (n = 4 mice/group) (C). Isolated autoreactive BDC2.5 TCR Tg⁺ CD4⁺ cells were transferred into NOD.SCID mice previously reconstituted with NOD splenocytes and then treated with anti-CD22/cal mAb or left untreated. Fewer autoreactive BDC2.5 TCR Tg⁺ CD4⁺ cells were recovered (D) in the anti-CD22/cal mAb–treated NOD.SCID hosts (E, top quadrant) compared with the untreated controls (E, bottom quadrant). Insulitis score analysis revealed better-preserved islets in the anti-CD22/cal mAb–treated NOD mice at 15 and 35 weeks of age (F). (Please http://dx.doi.org/10.2337/db08-0420 for a high-quality digital representation of this figure.)

FIG. 3.

Histology of prevention studies. At baseline, NOD mice showed mild perinsulitis (A1) with many B220⁺ cells (A2) and some CD3⁺ cells (A3) but still with well-preserved insulin and glucagon staining (A5 and A6). FoxP3⁺ cells are merely present at baseline (A4). Interestingly, at 15 weeks of age, treated NOD mice showed reduced infiltrate (B1) with no B220⁺ cells (B2) and fewer CD3⁺ cells (B3), whereas in the control, B220⁺ and CD3⁺ cells are abundantly represented with increased infiltrate (C1–C3). At 35 weeks of age, the treated group showed cleaner pancreata compared with the untreated control hyperglycemic NOD mice (D1 and E1). B220⁺ and CD3⁺ cells did not infiltrate the islets in the treated group (D2 and D3), whereas in the controls, islets were extensively infiltrated by B220⁺ and CD3⁺ cells (E2 and E3). Islet morphology is well-preserved in the treated group at 15 and 35 weeks of age (B5, B6, D5, and D6) but not in the control group (C5, C6, E5, and E6). FoxP3 staining of islet infiltrate revealed a persistent reduced expression of FoxP3 in untreated compared with treated NOD mice at 15 and 35 weeks of age, particularly when compared with the massive presence of T-cells in the control (B4, D4, C4, and E4). (Please see http://dx.doi.org/10.2337/db08-0420 for a high-quality digital representation of this image.)

FIG. 4.

Hyperglycemia reversal studies. A rapid reversal of hyperglycemia was observed in all treated hyperglycemic NOD mice (A). Six of 10 remained normoglycemic in the long term. None of the untreated newly hyperglycemic control NOD mice reverted from hyperglycemia (B). After 5 days from hyperglycemia onset (n = 6), anti-CD22/cal mAb was not able to restore normoglycemia in the long term (B). Either calicheamicin alone (GG5/cal) (n = 5) or CD22 unconjugated treatment failed to restore normoglycemia in the long term (B). Proinflammatory cytokines (IL-17, TNF-α, and slightly IFN-γ) are reduced 10 days after treatment compared with untreated controls (C). CD4⁺CD25⁺FoxP3⁺ cell percentage increases in anti-CD22/cal mAb–treated long-term tolerant compared with untreated control NOD mice in the pancreatic lymph nodes (anti-CD22/cal mAb–treated vs. normoglycemic 10-week-old mice, P = 0.007, and vs. hyperglycemic, P = 0.03; D) and in the spleen as well (anti-CD22/cal mAb–treated vs. normoglycemic 10-week-old mice, P = 0.02, and vs. hyperglycemic, P = 0.01; E). Insulitis score confirmed that anti-CD22/cal mAb–treated NOD mice showed better preserved and less infiltrated islets compared with untreated control NOD mice both at baseline and 10 days after hyperglycemia onset (F).

FIG. 5.

Histology of hyperglycemia reversal studies. Untreated, hyperglycemic mice at baseline show islets heavily infiltrated by lymphocytes (A1) predominantly composed of B220⁺ and CD3⁺ cells (A2 and A3) with few FoxP3⁺ Tregs (A4). Few insulin-positive cells and more glucagon-positive cells can be detected (A5 and A6). Ten days after treatment with anti-CD22/cal mAb, islets appeared scarcely infiltrated compared with untreated controls (B1 and C1), with few B220⁺ and CD3⁺ cells (B2, B3, C2, and C3) but with an increase in FoxP3⁺ cells (B4 and C4). In treated animals but not in the untreated controls, islets show abundant stainable insulin (B5 and C5) and glucagon (B6 and C6). Two histological patterns are seen in the treated group 100 days after treatment: Many of the islets show essentially no lymphoid infiltrate at all (D1–D3), and few cells stain for insulin whereas more stain for glucagon (D5 and D6). A smaller subset of islets show an abundant B220⁺/CD3⁺ infiltrate (E1, E2, and E3). However, the infiltrate remains largely confined to the periphery of the islets, with a greater percentage of FoxP3⁺ Tregs (E4). Glucagon is easily detected (E6), but insulin staining is low (E5). (Please see http://dx.doi.org/10.2337/db08-0420 for a high-quality digital representation of this image.)

FIG. 6.

Transcriptome analysis of reemerging B-cells. We extracted B-cells (using CD19 magnetic beads) from 10-week-old NOD mice, from hyperglycemic NOD mice, and from the reemerging B-cell pool from age-matched B-cell–depleted NOD mice in which the B-cell repertoire is recovered. A gene array analysis was performed to evaluate gene expression of >40,000 genes. Genes that are differentially expressed in naïve B-cells extracted from normoglycemic 10-week-old or hyperglycemic NOD mice and reemerging B-cells are shown in the heat map (A–C). Blue represents lesser expression and red higher expression. Two hundred genes are downregulated in the reemerging B-cells compared with naïve B-cells from 10-week-old NOD mice (A). Thirty-eight genes are downregulated in the reemerging B-cells compared with naïve B-cells from hyperglycemic NOD mice (B). Twenty-one genes are downregulated similarly in the reemerging B-cells compared with naïve B-cells from 10-week-old and hyperglycemic NOD mice (C).

FIG. 7.

Functional studies of reemerging B-cells. FACS analysis of CD80, CD86, CD40, Class II, and IgM did not reveal any differences between reemerging and naïve B-cells extracted from splenocytes (the latter from either normo- or hyperglycemic NOD mice) (representative of five mice; A). Interestingly, we observed by FACS analysis a higher percentage of anergic B-cells (B220⁺CD93⁺CD23⁺IgM^lo cells) in the reemerging B-cell pool compared with naïve B-cells from hyperglycemic age-matched untreated NOD mice (representative of five mice [B, with anergic B-cells circled]). We customized an in vitro assay in which B-cells are used as APCs and autoreactive BDC2.5 TCR Tg⁺ CD4⁺ cells are used as responders in the presence of the BDC2.5 peptide. When reemerging B-cells were APCs, a lower IFN-γ production by BDC2.5 TCR Tg⁺ CD4⁺ cells was evident compared with when naïve B-cells were used (C). Supernatants were collected from the experiment described above, and cytokine profile was assessed with a Luminex assay. Interestingly, when reemerging B-cells, but not naïve B-cells, were used as APCs, BDC2.5 autoreactive CD4⁺ cells downregulated the production of proinflammatory cytokines (IL-2, IL-17, TNF-α, and IFN-γ) (F–I). We then coadoptively transferred CD19⁺ cells (obtained from reemerging or from naïve B-cell pool) into NOD.SCID recipients with diabetogenic CD4⁺ cells obtained from hyperglycemic NOD mice. When reemerging B-cells, but not naive B-cells, were transferred, they completely abrogated the onset of diabetes mediated by the transfer of diabetogenic CD4⁺ cells (n = 5 mice/group) (D). We also analyzed the percentage of CD4⁺CD25⁺FoxP3⁺ cells in the spleen of NOD.SCID recipients of diabetogenic CD4⁺ T-cells and reemerging B-cells or controls (B-cells from hyperglycemic animals or no cells) at day 30 after adoptive transfer. E: There is no difference between the groups (n = 5 mice/group).