Introduction

Autoimmune diabetes, including type 1 diabetes mellitus (T1DM) and latent autoimmune diabetes in adults (LADA), are caused mainly by autoimmune destruction of β-cells. Globally, LADA accounts for 4–12% of diabetes cases [1]. The presence of islet autoantibodies, such as glutamic acid decarboxylase antibody 65 (GADA), insulin autoantibodies (IAAs), islet cell antibodies (ICAs), tyrosine phosphatase-related molecule-2 autoantibodies (IA2As), zinc transporter-8 autoantibodies (ZnT8As), and the newly identified tetraspanin 7 autoantibodies(TSPN7As) [2,3,4,5,6,7,8], strongly predicts the progression of overt autoimmune diabetes. Among them, GADA is undoubtedly the most critical autoantibody in autoimmune diabetes.

Insulin is the most frequent treatment choice to compensate for the insufficiency in insulin secretion caused by β-cell dysfunction [9]. In addition to insulin, a variety of oral agents are prescribed for autoimmune diabetes, such as insulin sensitizers (metformin and thiazolidinediones), incretins, and sodium-glucose cotransporter inhibitors [10,11,12]. Dipeptidyl peptidase-4 (DPP-4) inhibitors, a new class of hypoglycemic agents, increase meal-time secretion of endogenous incretins (including glucagon-like peptide-1 and gastric inhibitory polypeptide) by blocking their degradation, decrease postprandial release of glucagon, and stimulate the secretion of insulin, thereby activating multiple hypoglycemic mechanisms. Three large-scale clinical trials, EXAMINE [13], SAVOR-TIMI 53 [14], and TECOS [15], have been completed since 2008 to assess the cardiovascular outcomes of DPP-4 inhibitors in type 2 diabetes. Although SAVOR-TIMI 53 found that the DPP-4 inhibitor saxagliptin increased hospitalizations for acute heart failure, several meta-analyses [16,17,18] have confirmed the cardiovascular (CV) safety of DPP-4 inhibitors. In addition to safety against adverse CV events, DPP-4 inhibitors demonstrate weight neutrality and a low incidence of hypoglycemia, and thus have gained broad application in clinical practice [19].

Apart from specific hypoglycemic mechanisms, DPP-4 inhibitors possess an immune modulation profile, offering the potential for extendibility to autoimmune diabetes. Autoimmune diabetes is associated with marked alterations in the frequencies of circulating immune cells, including T cells [20], B cells [21], natural killer cells [22], and dendritic cells [23]. Among these, effector T cells (Teff) are considered crucial for the progression of autoimmune diabetes. In non-obese diabetic (NOD) mice, DPP-4 inhibitors decreased the type 1 helper T cell (Th1) immune response, upregulated secretion of Th2 anti-inflammatory cytokines, activated CD4+CD25+FoxP3+ regulatory T cells, and prevented production of the pro-inflammatory cytokine IL-17 [24]. Accordingly, we speculate that DPP-4 inhibitors may also attenuate the autoimmune processes underlying autoimmune diabetes. In this review, we summarize the physiological and immunomodulatory effects of DPP-4 inhibitors, and discuss potential applications for the treatment of autoimmune diabetes.

Overview of DPP-4 biology

Dipeptidyl peptidase-4, also known as cluster of differentiation antigen 26 (CD26), is a widely expressed and highly conserved membrane glycoprotein that cleaves N-terminal dipeptides from proteins with proline or alanine in the penultimate position. It is strongly expressed on the membranes of specialized cell types within liver, kidney, lung, spleen, and pancreas, including adipocytes, hepatocytes, endothelial cells, and epithelial cells, and is implicated in their physiological functions. It is also broadly expressed on various immune cells including T, B, and NK cells, macrophages, and dendritic cells [25]. Many regulatory peptides containing the CD26 target sequence are cleaved and inactivated by this enzyme, including incretin peptides such as glucagon-like peptide-1 (GLP-1), GLP-2, and gastric inhibitory polypeptide (GIP), as well as brain natriuretic peptide (BNP), peptide YY, stromal cell-derived factor-1 (SDF-1), and substance P. The diversity of substrates degraded by DPP-4 expands its actions, both physiological and pathological, and indeed DPP-4 is implicated in diabetes, solid tumor development, autoimmune diseases, and obesity [26].

CD26 also regulates multiple aspects of lymphocyte function independent of enzymatic activity [27]. It is the main cellular binding protein for ecto-adenosine deaminase (eADA), which is responsible for the degradation of adenosine. It also binds extracellular matrix (ECM) components such as collagen and fibronectin, thereby regulating the interactions of various cells with the ECM, including cancer cells. More germane to autoimmune diabetes, CD26 regulates T cell activation and immune signaling pathways by associating with the chemokine receptor CXCR4 and the serine protease fibroblast activated protein-alpha (FAP-α). In addition to its membrane-anchored form, DPP-4 also has a soluble form in plasma that is observed in certain cancers, autoimmune diseases, diabetes, and obesity.

While DPP4 inhibitors share common modes of action, there is still prominent heterogeneity in pharmacokinetic properties among compounds, such as differences in half-life, biological activity, metabolism, and excretion. Some DPP-4 inhibitors are substrate-enzymatic blockers (saxagliptin and vildagliptin), while others are competitive inhibitors (sitagliptin and alogliptin) [28]. Inhibitors of DPP-4 also differ in the route of elimination. For instance, most DPP-4 inhibitions are eliminated mainly through the kidney while less than 5% of linagliptin is eliminated through this route.

Functions of T cell subsets Th1, Th2, Treg, Th17, and CD8+ in immune modulation

Helper T (Th) cell subsets are distinguished by the different cytokines produced and major transcription factor expressed. Type 1 Th cells contribute to cell-mediated immune modulation, while Th2 cells promote humoral immunity. These two subsets exert reciprocal control over the function of the other. Thus, Th1/Th2 balance is a critical factor for induction and modulation of immune pathology in autoimmune diabetes. The helper T cell subset Th17 producing interleukin-17 (IL-17) has emerged as an important contributor to pathological conditions that were previously attributed to Th1 cells. Moreover, CD4+CD25+T regulatory cells (Tregs) appear critical for mediating Th17 activity and Th17/Treg balance [29].

Type 1 Th populations secreting pro-inflammatory cytokines such as IL-2 and interferon-γ (IFN-γ) are crucial mediators of β-cell autoreactivity. Higher levels of Th1-related cytokines were observed in children at greater risk of developing T1DM [30]. On the contrary, Th2 populations producing cytokines such as IL-4, IL-5, and IL-13 possess a crucial protective profile against autoimmunity in T1DM. The most important immune regulators are Tregs [31]. They can suppress the proliferation and function of Teff cells, and maintain immune balance. Forkhead transcription factor P3 (FoxP3), a specific marker of Treg cells, controls their differentiation and function. Patients with T1DM exhibit decreased Treg frequency [32], and this defect in Treg number and/or function is a major factor in disease progression [33,34,35]. Dysfunction of Tregs is also observed in LADA [36] as evidenced by significantly reduced expression of FoxP3 and a hypermethylated FoxP3 promoter region in CD4+T cells [37]. Treatment with Tregs is considered safe and effective for preserving β-cell function in T1DM [38, 39].

The IL-17-producing Th17 cell, a novel class of CD4+T cell with specificity for self-antigens [40], are extremely pathogenic and result in multiple severe inflammatory and autoimmune disorders [41,42,43]. Several studies have found substantially elevated IL-17 in T1DM patients [44,45,46]. In rodent models, plasticity of the Th1/Th17 cell balance has been implicated as a potential contributor to the progression of autoimmune diabetes [47].

In addition to CD4+T cells, there is growing evidence for the involvement of CD8+T cells in T1DM pathogenesis [48,49,50]. These autoreactive CD8+T cells can be found in the peripheral blood of T1DM patients [51], and reflect a novel and distinctive feature in immunopathogenesis. TGF-β can reduce apoptosis of differentiated autoreactive CD8+ T-cells, thereby promoting their expansion and infiltration into islets [9]. The use of non-depleting antibodies specific for CD4+ and CD8+ resulted in long-term diabetes remission in NOD transgenic mice [52], while immunotargeting of insulin-reactive CD8+ T cells in young NOD mice reduced disease incidence and slowed progression [53], thus providing a possible immunotherapy for T1DM.

In NOD mice, sitagliptin [54] decreased migration of splenic and lymph node CD4+ T cells through incretin-dependent and -independent pathways, suggesting that the anti-diabetic response may be mediated through selective effects on T cell subsets associated with autoimmunity. In CD26 knockout mice, a DPP-4 inhibitor upregulated Th2 and Treg cells, and suppressed the pathogenic effects of Th1 and Th17 cells [55]. In an in vivo study [24], a DPP-4 inhibitor demonstrated an immunosuppressive effect on Th1 and Th17 lymphocyte differentiation, and led to the generation of regulatory transforming growth factor (TGF-β1) and reduced CD26 gene expression, which help maintain the survival of pancreatic β-cells. In addition, CD26 is highly expressed on Th17 cells [56]; moreover, DPP-4 inhibition can prevent the production of IL-17 in NOD mice [57, 58], underscoring the potential of DPP-4 inhibitors as therapeutic agents to suppress the pathogenic effects of Th17 cells in autoimmune diabetes. Further, the frequency of the CD8+T lymphocyte subset was enhanced, the inflammatory response suppressed, and diabetes onset delayed or prevented by a DPP-4 inhibitor in NOD mice [59].

DPP-4 is highly expressed on T cells and regulates their biological function. The immune responses of Th1 cells were deregulated, IFN-γ was suppressed, and Th2 cytokines such as IL-4, IL-5, and IL-13 were upregulated in CD26 knockout mice [55]. Suppression of DPP-4 also reduced production of IL-2, IL-12, and IFN-γ by T cells and peripheral blood mononuclear cells (PBMC) [60, 61]. The expression of TGF-β was shown to be downregulated in T1DM patients [62]. Alternatively, TGF-β-activated kinase-1 (TAK1, Map3k7) can increase the production of TGF-β and delay the onset of autoimmune diabetes in NOD mice [63, 64]. DPP-4 inhibition can upregulate the level of TGF-β [24, 65] and stromal cell-derived factor-1 (SDF-1) in vitro [66, 67]. Thus, DPP-4 inhibition may slow the progression of diabetes by shifting the balance toward anti-inflammatory T cell subsets and cytokines (Table 1).

Table 1 Effect of DPP4 inhibition on T cell subsets and cytokines

Preclinical effects of DPP4 inhibitor in autoimmune diabetes

Administration of DPP-4 inhibitors can inhibit inflammatory signaling pathways [70, 71] mediating vascular smooth cell proliferation and oxidative stress in various cells types [72], and also appears to improve endothelial function, blood pressure, and lipid metabolism in T2DM [73]. Of special note, emerging evidence suggests that DPP-4 inhibition can protect against myocardial infarction and atherosclerosis [74,75,76,77]. In conclusion, these studies consistently demonstrate that DPP4 inhibition provides endothelial protection through suppression of inflammation. Thus, aside from improving metabolic control [78], DPP4 inhibitors can suppress autoimmune processes leading to diabetic vasculopathy.

Inhibition of DPP-4 induced islet neogenesis, β-cell regeneration, and insulin synthesis in streptozotocin (STZ) model diabetic mice [79]. DPP-4 inhibitors can also preserve pancreatic β-cell mass and protect β-cells not only in diabetes rodent models [80, 81] but also in impaired fasting glucose patients and diabetes patient [82, 83]. DPP-4 inhibitors were also found to improve islet graft survival [84, 85] and even delay the onset of autoimmune diabetes in NOD mice [65, 86].

DPP-4 inhibitors may also act synergistically with other anti-diabetes treatments. Combination therapy with a DPP-4 inhibitor and a proton-pump inhibitor (PPI) raised blood concentrations of gastrin, promoted β-cell neogenesis, attenuated insulitis, and re-establish normoglycemia in NOD mice [87, 88]. The combination of a DPP-4 inhibitor and Toll-like receptor 2 agonist also increased β-cell mass in NOD mice with new onset diabetes [89]. The combination of a DPP-4 inhibitor and angiotensin II receptor blocker promoted islet regeneration [90]. Moreover, the combination of a DPP-4 inhibitor with a histone deacetylase inhibitor [91], VitD3 [60], or low-dose monoclonal CD3 antibody [69] has shown benefits as adjunct therapy for T1DM prevention in NOD mice. Thus, we postulate that combination therapy including a DPP-4 inhibitor can directly promote the growth and survival of β-cells in patients with autoimmune diabetes.

Clinical effects of DPP-4 inhibitor on autoimmune diabetes

Several case reports [92,93,94,95] on autoimmune diabetes patients have found that DPP-4 inhibitor treatment, alone or in combination with other drugs, significantly improved glucose control and reduced insulin requirements, with a good tolerance profile. Therefore, DPP-4 inhibitors may also benefit autoimmune diabetes. Nevertheless, the current level of evidence is of limited quality. There have been several small-scale randomized, controlled clinical pilot trials on DPP-4 inhibitors for autoimmune diabetes [96,97,98,99,100,101] (Table 2). Incretin-based therapy in T1DM patients improved glycemic control and reduced both hypoglycemia incidence and insulin requirements. However, there are discrepancies among these trials regarding efficacy of glycemic control, attenuation of β-cell function, and hypoglycemia frequency. Four studies [96, 98, 100, 101] found that DPP-4 inhibition can decrease required insulin dose, improve overall glucose control, and attenuate the decline in C-peptide levels. In contrast, another study [102] concluded that co-therapy with a DPP-4 inhibitor does not reduce the frequency of hypoglycemia in C-peptide-negative T1DM. However, a post hoc analysis from 5 pooled RCTs [103] found that saxagliptin was effective in decreasing blood glucose levels in GADA-positive patients and tended to improve β-cell function during a 24-week follow-up. Furthermore, an associated meta-analysis [104] concluded that DPP-4 inhibition significantly reduces daily insulin dose in T1DM.

Table 2 Clinical trials of dipeptidyl peptidase-4 inhibitors for autoimmune diabetes

There are several limitations to these clinical trials. First, due to the low incidence of T1DM, the enrolled sample sizes were small, most of them were pilot studies, and long-term large-scale prospective study is still lacking. Second, diabetes duration differed obviously among these studies, with some restricted to newly diagnosed T1DM while others included long-term autoimmune diabetes with exhausted C-peptide function, which would reduce the potential benefits of DPP-4 inhibition. Thus, the residual β cell function as a criteria for autoimmune diabetes patients selection seems to be the key point. Third, the follow-up period varied from 8 weeks to 2 years, insufficient to observe possible long-term efficacy. A four-year pilot study suggested that sitagliptin may be more effective for preserving β-cell function through immune modulation than insulin replacement [105], highlighting the promise of DPP-4 inhibitors for prevention or slowing the progression of autoimmune diabetes.

In all of these pilot trials, however, the number of cases with autoimmune diabetes was small, so large-scale and multi-center RCTs are needed to confirm the efficacy of DPP-4 inhibitors. Three large-scale RTCs on DPP-4 inhibitors for LADA or T1D are currently ongoing (Table 2). Two of these studies (NCT02307695 and NCT02407899) are being conducted in China and currently recruiting patients, while the other in Norway (NCT01140438) is active but not yet recruiting (Table 3).

Table 3 Ongoing clinical trials of dipeptidyl peptidase-4 inhibitors for autoimmune diabetes

Safety and efficacy of DPP-4 inhibitors in autoimmune diabetes

Several completed clinical trials [26, 84, 103, 106] have evaluated the safety and efficacy of DPP-4 inhibitors in T1DM. Neither severe hypoglycemic events nor serious side effects were observed. Minor side effects included gastrointestinal reactions (vomiting, diarrhea, nausea, etc.), flulike symptoms, and skin reactions [101]. A meta-analysis [104] including 228 T1DM patients concluded that DPP-4 inhibitors combined with insulin do not increase or decrease the risk of hypoglycemia. The RCTs in progress should provide additional information on potential adverse effects and other safety concerns.

Conclusions

In summary, DPP-4 inhibitors act as potent immune modulators through regulation of T cell phenotype and cytokine secretion. They are demonstrated to improve blood glucose control and attenuate β-cell destruction in animal models of autoimmune diabetes. In addition, DPP-4 inhibitors have shown good safe profiles not only in preclinical studies but also in randomized controlled clinical trials for autoimmune diabetes treatment. Collectively, these observations indicate that DPP-4 inhibitors are promising therapeutics for autoimmune diabetes. Though we are cautious on the use of DPP4 inhibitors in adults with autoimmune diabetes and no recommendations can be given at present time, further large-scale cohort studies are warranted to validate the therapeutic efficacy and safety of DPP-4 inhibitors in the near future.