Targeting proinsulin-reactive CD8+ T cells: a new direction for Type 1 diabetes treatment

Keywords::

• antigen-specific immune therapy
• autoimmune disease
• CD8 T cells
• DNA vaccine
• proinsulin
• Type 1 diabetes

Type 1 diabetes (T1D) is a chronic autoimmune disease in which the β-cells in the pancreatic islets of Langerhans are destroyed by autoreactive T cells. Currently, there is no cure for T1D. Therefore, patients need life-long insulin replacement therapy to control blood glucose levels. In the past, various efforts have been undertaken to develop and assess strategies that halt disease progression [1]. Immunotherapeutic strategies have aimed either at systemic down-modulation of the immune system or targeting specific autoimmune mechanisms involved in the disease. Non-specific immune suppressive approaches broadly inhibit the immune system and may therefore have undesirable risks and side effects. The calcineurin inhibitor cyclosporine A was among the first non-specific immune suppressants to be used in T1D. Continuous treatment initiated soon after disease onset demonstrated that it was possible to preserve β-cell function, with patients experiencing insulin independency [1]. However, this effect was short lasting and associated with accelerated renal dysfunction [2]. The enthusiasm for the use of this and other non-specific immune suppressants faded with trials using azathioprine, prednisolone and anti-thymocyte globulin showing no lasting clinical efficacy [1]. Clearly, the development of an antigen-specific therapy selectively targeting pathogenic autoreactive T cells, while sparing the ability of the immune system to respond to pathogens would be highly desirable.

In T1D, attention was focused mainly on CD4 T cells. However, it has become evident that CD8 T cells also play a crucial role in β-cell destruction. Studies on insulitic lesions showed large infiltrates of CD8 T cells and increased MHC-I expression in pancreata from recent onset T1D patients [3,4]. Recently, the presence of islet autoreactive CD8 T cells in human insulitic lesions could be demonstrated, providing the ultimate indication that T1D is indeed an autoimmune disease where islet autoreactive CD8 T cells selectively infiltrate pancreatic islets and cause β-cell destruction [5]. Importantly, this study also demonstrated that patients with disease for many years still possess insulin-positive β-cells, in spite of undetectable levels of C-peptide, as well as islet infiltration and autoimmunity. These findings have led to new opportunities: as long as β-cell mass is preserved and chronic islet inflammation lasts, interventional therapy is warranted. Indeed, recently diagnosed T1D patients treated with high-dose immunosuppression followed by autologous hematopoietic stem cell transplantation [6] became insulin independent, often for several years, without therapeutically targeting β-cell function directly. Yet, the in situ studies on insulitic lesions also revealed large heterogeneity between individual patients, even within a single pancreas, preparing the diabetes community for the need for fine diagnosis using genetic, immunological and metabolic biomarkers, followed by optimized therapy, that is, personalized medicine [5,7].

Among the immune intervention therapies targeting T cells, none was designed to specifically target pathogenic autoreactive CD8 T cells. Thymoglobulin proved to preferentially hit on CD4 T cells (changing the CD4/CD8 ratio in blood), while autoimmune memory could hardly be dealt with [8,9]. In the context of islet transplantation, thymoglobulin induction therapy did not change the rate of islet autoimmunity, with subsequent consequences for clinical outcome. Other candidate immune suppressive therapies that may hit more efficaciously on CD8 T cells and immune memory include alemtuzumab (anti-CD52 monoclonal antibody). Induction therapy with alemtuzumab shows great promise to change the prospect of islet replacement therapy in preimmunized islet recipients by affecting baseline CD4 and CD8 islet autoimmunity.

Clinical trials using monoclonal antibodies against CD3 were designed with the goal of reducing T-cell activation, while preserving immune regulation via suboptimal T-cell receptor signals and/or induction of regulatory T cells (Tregs). These efforts have shown promising results, with preserved stimulated C-peptide responses and reduced need for exogenous insulin [10]. It has been suggested that treatment with anti-CD3 induces adaptive CD8 Tregs. Some patients responding to therapy were shown to possess an increased proportion of CD8 versus CD4 T cells after treatment, and once cultured in vitro these cells displayed inhibitory properties [11]. The induction of Tregs in T1D is an attractive goal, and studies have investigated the possibility of inducting islet-specific CD8 Tregs. Dendritic cells (DCs) rendered tolerogenic by vitamin D3 and dexamethasone treatment during monocyte differentiation have been shown to efficiently induce islet-specific CD4 Tregs in vitro [12]. This approach has been modified to test for induction of CD8 Tregs by loading modulated DCs with MHC-I binding islet peptides [13]. Indeed, priming of islet autoreactivity was blunted, while existent islet autoreactive CD8 T cells were eliminated, presumably through coligation with the programmed death ligand-1. However, tolerogenic DCs loaded with HLA class I binding islet epitopes also became targets of islet-specific CD8 T-cell destruction, even though epitopes of islet autoantigens often have low binding affinity to HLA class I [14,15]. Therefore, this strategy to target CD8 islet autoimmunity appears to be inefficient to treat T1D patients.

Antigen-specific therapies in T1D are in their early stage, but they show promise. Oral administration of insulin and injection of islet antigens or peptides thereof have led to delay of disease onset (oral insulin), progression (Diamyd-GAD, DiaPep277) or induction of tissue-specific immune modulation (proinsulin peptide), but these results were not always consistent (GAD65, DiaPep277). Defining the optimal dose, frequency of injections, choice of adjuvants and suitable patients are important and need further studies. Yet, collectively, these antigen- and tissue-specific immunotherapies have consistently shown a high degree of safety, in spite of the theoretical risk that injecting the autoimmune target could accelerate disease. A recent study has offered a new glimpse of hope in the development of antigen-specific strategies [16]. Eighty recently diagnosed T1D patients were randomized to receive various doses of an engineered DNA vaccine, termed BHT-3021, encoding proinsulin. The choice of targeting proinsulin-specific CD8 T cells stems from the fact that insulin is a known major autoantigen in T1D. Moreover, CD8 T cells reactive to proinsulin have been found in increased frequencies in the peripheral blood of T1D patients, and preproinsulin-specific CD8 T cells isolated from a T1D patient were shown to kill human β-cells in vitro [14,17,18]. In preclinical studies, the DNA vaccine was capable of preventing and reversing active insulitis in hyperglycemic non-obese diabetic mice. It was therefore hypothesized that BHT-3021 would preserve β-cell function in T1D patients by decreasing the antigen-specific autoimmune response against proinsulin, while leaving the remainder of the immune system undisturbed. T1D patients received intramuscular injections of BHT-3021 or BHT-placebo weekly for 12 weeks in four different dosages (0.3, 1.0, 3.0 and 6.0 mg) and C-peptide levels were measured to assess β-cell function and therapeutic efficacy and safety of the vaccine. Notably, an improvement in C-peptide levels was observed in patients treated with BHT-3021, particularly with 1 mg at 15 weeks. The results from this study point to a preservation of β-cell function during the dosing period that diminishes after discontinuation of therapy.

The mechanism by which this new type of gene therapy affects islet autoimmunity is still unclear. Indeed, a major challenge in T1D is the identification and implementation of suitable immune biomarkers that point to a mechanism of action of disease or immunotherapy [19,20]. Immune correlates of insulitis, disease progression and clinical efficacy of therapeutic interventions could allow prediction of clinical efficacy at early stages and guide choice of therapy, ultimately leading to shorter trials and higher success rates after therapy. In the proinsulin gene therapy trial, islet- and vaccine-specific modulation of the immune system was measured using HLA class I multimers loaded with islet peptides of a range of distinct islet antigens (preproinsulin, GAD, IA2, IAPP and IGRP) to detect changes in islet-specific CD8 T cells or a mix of virus epitopes to define specificity of the intervention [16]. Excitingly, clinical efficacy was inversely associated with changes in CD8 T-cell autoimmunity, pointing to the possibility that this gene therapy affects CD8 T cells selectively. Indeed, frequencies of T cells to islet epitopes absent in the gene vaccine were unchanged, as were those against pathogens. This study demonstrates that changes in the frequencies of antigen-specific CD8 T cells can act as biomarkers of clinical efficacy, as changes in autoreactive CD8 T cells were inversely correlated with changes in β-cell function.

It is intriguing why administration of a DNA vaccine encoding the islet antigen would be more efficient in achieving clinical benefit than other approaches administering the antigen as protein. First, injection of the proinsulin DNA vaccine directly into the muscle allows bypassing denaturation and degradation that occurs in the gastrointestinal tract upon oral administration of the antigen. To avoid potential immunogenicity of the DNA plasmid, this vector was engineered with reduced numbers of proinflammatory CpG motifs that are known to activate innate immune responses through binding to Toll-like receptor 9 and other DNA sensors. Furthermore, since muscle cells are not professional antigen-presenting cells, they will present peptides in their MHC-I molecules without costimulation. Proinsulin-specific CD8 T cells may either be deleted by apoptosis, by T-cell receptor signaling in the absence of costimulation, or be suppressed by Tregs and sequestered from the pancreas and peripheral circulation.

Are we there yet? Not quite. Results from gene therapy are encouraging, demonstrating the safety and potential efficacy of C-peptide preservation during the dosing period. However, this effect was eventually lost when DNA vaccination was stopped and it remains unclear whether long-term benefit can be achieved with BHT-3021. Yet, it was not quite conceivable that treatment of patients with up to 5 years of disease and reduced β-cell mass and function would even allow to assess clinical efficacy (aside from safety), as β-cell function is believed to be low in patients with longer disease duration. Follow-up studies including longer duration of treatment and patients with recent onset will tell us whether we have finally grasped the holy grail of targeting pathogenic autoimmune responses while leaving the rest of the immune system intact.

Financial & competing interests disclosure

The authors are supported by grants from the Juvenile Diabetes Research Foundation, the Dutch Diabetes Research Foundation, The Netherlands Organization for Health Research and Development (ZonMW: VICI award) and the European Union (EU-FP7 Framework, NAIMIT, BetaCellTherapy & EE-ASI). BO Roep has consulted for Diamyd Medical, Andromeda Biotech, Novo Nordisk AS, GSK and Lilly and received financial support for mechanistic studies from Bayhill Therapeutics. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.