B lymphocyte alterations accompany abatacept resistance in new-onset type 1 diabetes
Peter S. Linsley, Carla J. Greenbaum, Cate Speake, S. Alice Long, Matthew J. Dufort
Peter S. Linsley, Carla J. Greenbaum, Cate Speake, S. Alice Long, Matthew J. Dufort
View: Text | PDF
Research Article Therapeutics

B lymphocyte alterations accompany abatacept resistance in new-onset type 1 diabetes

Abstract

Costimulatory interactions control T cell activation at sites of activated antigen-presenting cells, including B cells. Blockade of the CD28/CD80/CD86 costimulatory axis with CTLA4Ig (abatacept) is widely used to treat certain autoimmune diseases. While transiently effective in subjects with new-onset type 1 diabetes (T1D), abatacept did not induce long-lasting immune tolerance. To elucidate mechanisms limiting immune tolerance in T1D, we performed unbiased analysis of whole blood transcriptomes and targeted measurements of cell subset levels in subjects from a clinical trial of abatacept in new-onset T1D. We showed that individual subjects displayed age-related immune phenotypes (“immunotypes”) at baseline, characterized by elevated levels of B cells or neutrophils, that accompanied rapid or slow progression, respectively, in both abatacept- and placebo-treated groups. A more pronounced immunotype was exhibited by a subset of subjects showing poor response (resistance) to abatacept. This resistance immunotype was characterized by a transient increase in activated B cells (one of the cell types that binds abatacept), reprogrammed costimulatory ligand gene expression, and reduced inhibition of anti-insulin antibodies. Our findings identify immunotypes in T1D subjects that are linked to the rate of disease progression, both in placebo- and abatacept-treated subjects. Furthermore, our results suggest therapeutic approaches to restore immune tolerance in T1D.

Authors

Peter S. Linsley, Carla J. Greenbaum, Cate Speake, S. Alice Long, Matthew J. Dufort

×

Figure 5

Faster progression and elevated B cells levels predict poor suppression of insulin antibody response.

Options: View larger image (or click on image) Download as PowerPoint
Faster progression and elevated B cells levels predict poor suppression ...
Antibodies targeting insulin (mIAA), ICA512 (IA-2), and GAD65 were measured at the indicated individual visits. Reported antibody levels were normalized using the R package bestNormalize before plotting. Shown are mean ± SEM values. (A–C) Treatment affects levels of mIAA, ICA512, and GAD65 antibodies. There were 55–71 abatacept-treated and 26–34 placebo-treated subjects at each visit for each of the indicated autoantibodies. The vertical dotted line denotes the day of the last does (day 700). P values were determined by a mixed-effect linear model on measurements made during the treatment period (i.e., <700 days): value~visit+ group + visit*group +(1|id). (D) This model contained fixed-effect terms for visit (days) and group; an interaction term for visit and group (i.e., changes in group over time); and a random effect for subject (id). P values are represented are shown as P values for group/P value for interaction. Treatment and age impact insulin antibody levels. As in A, except that samples from abatacept-treated subjects were further stratified according to age at diagnosis. Older, older than median age; younger, younger than or equal to median age (12.8 years). (E) Treatment and response to therapy impact insulin autoantibody levels. As in A, except that samples from abatacept-treated subject samples were further stratified according to responder status. (F) Differences by treatment group in the relationship between B cell gene expression and insulin antibody levels. Ellipses represent 95% confidence intervals. P values (p.overall) were calculated using the compareGroups R package, using a linear model: group ~ CD19.mod gene expression + antibody level model. P values represent the probability of group identity. P values are shown as P values for CD19.mod gene expression/P value for mIAA antibody levels. There were 12–19 abatacept-treated R subjects; 43–52 abatacept-treated NR subjects; and 26–34 placebo-treated subjects at each visit. *****P < 10–5; **P ≥ 10–5 and < 10–2; *P ≥ 10–2 and < 0.05; NS, P > 0.05.