Co-stimulation modulation with abatacept in patients with recent-onset type 1 diabetes: a randomised, double-blind, placebo-controlled trial

Summary

Background

The immunopathogenesis of type 1 diabetes mellitus is associated with T-cell autoimmunity. To be fully active, immune T cells need a co-stimulatory signal in addition to the main antigen-driven signal. Abatacept modulates co-stimulation and prevents full T-cell activation. We evaluated the effect of abatacept in recent-onset type 1 diabetes.

Methods

In this multicentre, double-blind, randomised controlled trial, patients aged 6–45 years recently diagnosed with type 1 diabetes were randomly assigned (2:1) to receive abatacept (10 mg/kg, maximum 1000 mg per dose) or placebo infusions intravenously on days 1, 14, 28, and monthly for a total of 27 infusions over 2 years. Computer-generated permuted block randomisation was used, with a block size of 3 and stratified by participating site. Neither patients nor research personnel were aware of treatment assignments. The primary outcome was baseline-adjusted geometric mean 2-h area-under-the-curve (AUC) serum C-peptide concentration after a mixed-meal tolerance test at 2 years' follow-up. Analysis was by intention to treat for all patients for whom data were available. This trial is registered at ClinicalTrials.gov, NCT00505375.

Findings

112 patients were assigned to treatment groups (77 abatacept, 35 placebo). Adjusted C-peptide AUC was 59% (95% CI 6·1–112) higher at 2 years with abatacept (n=73, 0·378 nmol/L) than with placebo (n=30, 0·238 nmol/L; p=0·0029). The difference between groups was present throughout the trial, with an estimated 9·6 months' delay (95% CI 3·47–15·6) in C-peptide reduction with abatacept. There were few infusion-related adverse events (36 reactions occurred in 17 [22%] patients on abatacept and 11 reactions in six [17%] on placebo). There was no increase in infections (32 [42%] patients on abatacept vs 15 [43%] on placebo) or neutropenia (seven [9%] vs five [14%]).

Interpretation

Co-stimulation modulation with abatacept slowed reduction in β-cell function over 2 years. The beneficial effect suggests that T-cell activation still occurs around the time of clinical diagnosis of type 1 diabetes. Yet, despite continued administration of abatacept over 24 months, the decrease in β-cell function with abatacept was parallel to that with placebo after 6 months of treatment, causing us to speculate that T-cell activation lessens with time. Further observation will establish whether the beneficial effect continues after cessation of abatacept infusions.

Funding

US National Institutes of Health.

Introduction

The most common form of type 1 diabetes mellitus is immune-mediated, in which insulin-producing β cells are destroyed.¹ Yet, at the time of diagnosis, most patients still have appreciable amounts of insulin production as measured by C-peptide concentrations.² Preservation of residual β-cell function (as measured by peak C-peptide ≥0·2 nmol/L) is highly desirable because it can reduce short-term and long-term complications of the disease.3, 4 Several clinical trials have been undertaken in an attempt to arrest autoimmunity in type 1 diabetes with immunomodulatory agents or antigen-based treatments. Most notably, trials of anti-CD3,5, 6 anti-CD20,⁷ and a GAD-65 antigen vaccine⁸ have shown some efficacy in preservation of β-cell function as evidenced by stimulated C-peptide secretion.

T cells play a central part in autoimmunity associated with type 1 diabetes. To become fully activated, these cells need two crucial signals.⁹ The first signal is an interaction between an antigen in the groove of the MHC molecule on antigen-presenting cells with the T-cell receptor. The most important second signal is the interaction between CD80 and CD86 on the antigen-presenting cells and CD28 on the T cells. This co-stimulatory second signal is needed for full activation of T cells, and without it they cannot function. Therefore, co-stimulation blockade has been proposed as a therapeutic modality for autoimmunity and transplantation.¹⁰ Abatacept (CTLA4–immunoglobulin fusion protein) selectively binds to CD80 and CD86, thereby blocking the interaction with CD28 and interfering with the early phases of T-cell activation, proliferation, and survival. It inhibits naive T-cell activation, thus having the potential to selectively inhibit T-cell response to specific antigens instead of broad immunosuppression. Effector-memory T-cell responses are less dependent on CD28 co-stimulation and, presumably, are less inhibited by co-stimulation blockade.¹¹ Thus, abatacept is expected to be mildly immunomodulatory, and most likely to affect disease at early stages of pathogenesis. Studies in both animals and human beings have shown that interruption of the co-stimulatory second signal beneficially affects autoimmunity. Lenschow and coworkers¹² showed that co-stimulatory blockade with CTLA4–immunoglobulin fusion protein prevented diabetes in the NOD mice model of type 1 diabetes, when administered after development of insulitis but before frank diabetes ensues. Overall, co-stimulation blockade with abatacept has been shown to have clinical effectiveness in psoriasis¹³ and psoriatic arthritis¹⁴ and is approved for treatment of rheumatoid arthritis,¹⁵ including juvenile rheumatoid arthritis.¹⁶ Additionally, co-stimulation blockade has been effective in control of allograft rejection.¹⁷

Overall, the clinical success of co-stimulation blockade together with its good safety and tolerability profile made this approach an attractive candidate for assessment in recent-onset type 1 diabetes. We postulated that co-stimulation modulation with abatacept, by blocking the generation of autoaggressive T cells, would halt or slow autoimmune β-cell destruction leading to preservation of C-peptide secretion in recently diagnosed patients with type 1 diabetes.