Loss of intra-islet CD20 expression may complicate efficacy of B-cell-directed type 1 diabetes therapies

Abstract

Objective: Consistent with studies in NOD mice, early clinical trials addressing whether depletion of B cells by the Rituximab CD20-specific antibody provides an effective means for type 1 diabetes reversal have produced promising results. However, to improve therapeutic efficacy, additional B-cell-depleting agents, as well as attempts seeking diabetes prevention, are being considered.

Research design and methods: Autoantibodies, including those against insulin (IAAs), are used to identify at-risk subjects for inclusion in diabetes prevention trials. Therefore, we tested the ability of anti-CD20 to prevent diabetes in NOD mice when administered either before or after IAA onset.

Results: The murine CD20-specific 18B12 antibody that like Rituximab, depletes the follicular (FO) but not marginal zone subset of B cells, efficiently inhibited diabetes development in NOD mice in a likely regulatory T-cell-dependent manner only when treatment was initiated before IAA detection. One implication of these results is that the FO subset of B cells preferentially contributes to early diabetes initiation events. However, most important, the inefficient ability of anti-CD20 treatment to exert late-stage diabetes prevention was found to be attributable to downregulation of CD20 expression upon B cell entry into pancreatic islets.

Conclusions: These findings provide important guidance for designing strategies targeting B cells as a potential means of diabetes intervention.

FIG. 1.

Peripheral blood B cells in NOD mice are efficiently deleted by the IgG1 isotype of the 18B12 murine CD20-specific antibody. NOD female mice were initially injected intraperitoneally at 8 weeks of age with 10 mg/kg body wt of the IgG1 anti-CD20 isotype, with controls treated with a similar dose of the irrelevant 2B8 antibody (n = 3 per group). Proportions of B cells among peripheral blood leukocytes were then monitored over time. Depletion of peripheral blood B cells was safely maintained by repeated injection of the IgG1 18B12 antibody at 21-day intervals. PBL, peripheral blood.

FIG. 2.

Anti-CD20–mediated B-cell depletion strongly inhibits progression to overt diabetes development only when initiated in NOD mice that have not yet become IAA positive. A: Serum was collected from 10-week-old NOD female mice that then immediately began to receive treatments at 21-day intervals with the IgG1 CD20-specific or control antibody at a 10 mg/kg body wt dose. Serum samples were retrospectively assessed for the presence of IAAs and the mice monitored for diabetes development. Mice used for the analyses had a spread of eight separate birth dates and, thus, were entered into the treatment groups in a staggered fashion. Anti-CD20 treatment significantly suppressed diabetes development in the IAA-negative (P = 0.04) but not IAA-positive recipients (P = 0.14, Kaplan-Meier analyses). B: IAA-negative NOD female mice receiving a single anti-CD20 treatment at 10 weeks of age were not protected from subsequent diabetes development. Aby, antibody.

FIG. 3.

Selective depletion of the FO subset of mature B cells limits the initiation but does efficiently abrogate already established diabetogenic autoimmune responses in NOD mice. A: Splenic MZ B cells (CD21hi CD23⁻) undergo a greater age-dependent expansion in NOD than B6 female mice (n = 5 per strain at each time point; P = 0.02 by ANOVA). B: Splenic FO B cells (CD21int CD23hi) undergo a greater age-dependent expansion in B6 than NOD mice (P = 0.001 by ANOVA). C and D: Treatment with the IgG1 isotype of the 18B12 murine CD20-specific antibody deletes splenic FO but not MZ B cells in NOD mice. Treatment at 21-day intervals with the CD20-specific or control antibody was initiated at 5 weeks of age in a cohort of NOD female mice. At the indicated time points, a subset of mice in each group was then assessed for numbers of splenic MZ (C) and FO (D) B cells. E: Antibody responses to exogenous antigens are marginally suppressed in NOD mice selectively depleted of FO B cells. NOD mice were treated at 10 weeks of age with the control or IgG1 CD20-specific antibody (n = 5 per group) and then primed 9 days later with HEL. Serum was collected 11 days after antigen priming and assessed by enzyme-linked immunosorbent assay for relative levels of HEL-specific antibodies. Data are represented as mean OD₄₀₅ ± SEM at the indicated serum dilutions (control vs. FOB depleted: P = 0.0001 by ANOVA). MZB, MZ B cells; FOB, FO B cells; Aby, antibody.

FIG. 4.

B cells entering pancreatic islets in NOD mice become CD20-negative plasma cells. A: Proportion of total B cells in pancreatic islet–associated leukocytes from 7- or 13-week-old untreated NOD female mice. B: Proportion of total B cells in pancreatic islet–associated leukocytes from 7- or 13-week-old NOD female mice treated 2 weeks earlier with anti-CD20. Noted statistical significance values in A and B were determined by ANOVA analyses. C: Pancreatic islet–infiltrating B cells in 13-week-old untreated NOD control mice acquire a CD20-negative phenotype (right panel). B cells in PLNs were assessed as a CD20-staining positive control (left panel). Flow cytometric profiles are shown for B220 gated cells. D: CD20-negative pancreatic islet–associated B cells in untreated NOD mice convert to a CD138-positive plasma cell phenotype (right panel). B cells in PLNs were analyzed as a control (left panel). Flow cytometric profiles are shown for B220 gated cells.

FIG. 5.

Both Tregs and diabetogenic BDC2.5 clonotypic CD4 T cells expand within PLNs of NOD mice depleted of FO B cells by anti-CD20 treatment. Beginning at 4 weeks of age, NOD females received two injections at a 21-day interval of the CD20-specific or control antibody and at 4–7 days after the second treatment, were assessed for numbers of (A) total CD4 T cells, (B) BDC2.5 clonotypic cells by specific tetramer staining, and (C) Tregs within PLNs. Noted statistical significance values were determined by ANOVA analyses. Aby, antibody.

FIG. 6.

Coablation of Tregs abrogates the diabetes protective effects of selectively eliminating FO B cells in pre-IAA onset NOD mice. Anti-CD20 or control antibody treatment was initiated in 10-week-old NOD female mice typed to be IAA negative, with a subset also injected at 2-week intervals with a CD25-specific antibody to deplete Tregs. Compared with controls, diabetes development was significantly decreased only in NOD mice treated with anti-CD20 alone (P = 0.0175, Kaplan-Meier analyses). T1D, type 1 diabetes.