Small-molecule inhibition of inflammatory β-cell death

Abstract

With the worldwide increase in diabetes prevalence there is a pressing unmet need for novel antidiabetic therapies. Insufficient insulin production due to impaired β-cell function and apoptotic reduction of β-cell mass is a common denominator in the pathogenesis of diabetes. Current treatments are directed at improving insulin sensitivity, and stimulating insulin secretion or replacing the hormone, but do not target progressive apoptotic β-cell loss.

Here we review the current development of small-molecule inhibitors designed to rescue β-cells from apoptosis. Several distinct classes of small-molecules have been identified that protect β-cells from inflammatory, oxidative and/or metabolically-induced apoptosis. Although none of these have yet reached the clinic, β-cell protective small-molecules alone or in combination with current therapies provide exciting opportunities for the development of novel treatments for diabetes.

Introduction

In both type 1 and type 2 diabetes mellitus, functional β-cell mass is reduced. A reduction of 40% β-cell mass in young adults is sufficient to induce hyperglycemia, a hallmark of diabetes [1]. The etiology of type 1 diabetes is driven by autoimmune inflammatory processes that result in the progressive death of β-cells. Although type 2 diabetes is often characterized by insulin resistance, disease progression depends on insufficient insulin production, likely as a consequence of reduced β-cell mass by apoptosis [2]. Thus, in both diseases, a loss of β-cell viability contributes to the development of absolute or relative insulin insufficiency observed in type 1 and type 2 diabetic patients, respectively.

Current drugs for diabetes, in which treatment strategies focus on insulin replacement therapy (type 1) and reducing insulin resistance (type 2), do not directly target the underlying pathophysiology of the loss of β-cells. Small molecules that inhibit β-cell death may have therapeutic potential as an adjunct to traditional diabetes therapies. In fact, several small molecules, such as metformin and the thiazolidinedione class of diabetes drugs, have recently been shown to prevent β-cell apoptosis through different mechanisms of action. In this review, we will highlight these and other small molecules that inhibit β-cell apoptosis in the setting of islet inflammation in both type 1 and type 2 diabetes.

Cytokine-Induced Apoptosis

Proinflammatory cytokines such as interleukin-1β (IL-1β), interferon-γ (IFN-γ), and tumor necrosis factor-α (TNF-α) are produced by immune cells infiltrating the endocrine pancreas and play an important role in the autoimmune-mediated killing of β-cells observed in type 1 diabetes. Low-grade inflammation is hypothesized to contribute to the pathophysiology of type 2 diabetes [3], supporting a role for proinflammatory cytokines as well. The deleterious effects of these cytokines on β-cells were discovered in the mid-1980s and have since been extensively reviewed in the context of type 1 and 2 diabetes [4,5]. These cytokines activate intracellular signaling cascades that lead to apoptosis in β-cells. Perturbation of these pathways with small molecules represents a strategy to inhibit cytokine-induced apoptosis relevant for both types of diabetes [6,7].

The cytokine IL-1β induces phosphorylation and activation of c-Jun N-terminal kinase (JNK), which contributes to cytokine-induced β-cell apoptosis. Importantly, cell-permeable peptides functioning as JNK ligands block the deleterious effects of IL-1β in the βTC-3 β-cell line [8]. Selective peptide-mediated JNK inhibition reduces ischemia in mice and rats, providing evidence that blockade of JNK is possible in vivo [9]. The interleukin-1 receptor antagonist (IL-1Ra) anakinra may also have potential to inhibit autoimmune destruction of β-cells in type 1 diabetic patients by blocking the effects of IL-1β. This recombinant 153 amino-acid protein is currently approved to treat rheumatoid arthritis. Anakinra has been shown to possess anti-diabetic properties in both the STZ and the NOD mouse models of type 1 diabetes [10,11]. In the latter case, anakinra was effective in combination with anti-CD3 antibodies, indicating that combination therapy may be a viable treatment strategy. Although anakinra or the monoclonal anti IL-1β antibody, canakinumab, did not improve β-cell function in type 1 diabetes [12]. IL-1 antagonists did show promise in several clinical trials for the treatment of type 2 diabetes [7,13,14]. As an example administering anakinra for thirteen weeks led to improved glycated hemoglobin by increasing β-cell function [7]. Furthermore, a 39-week follow-up study showed that most effects were sustained in responders [15]. These studies warrant further clinical trials of anti-IL-1 antagonists in type 1 diabetic patients, reasonably in combination with anti-CD3 therapies, and hold promise for the development of small molecules that perturb the IL-1β signaling pathway. In this respect, other IL-1 antagonists have been developed but not yet tested in clinical trials [16,17].

Importantly, high levels of glucose induce IL-1β production and apoptosis in human islets, which is abrogated by IL-1Ra [18], providing another mechanistic explanation for the beneficial effects of anakinra on β-cells. Thus, small molecules perturbing IL-1β induction/processing provide another feasible way to protect β-cells. The sulfonylurea drug glyburide blocks islet amyloid popypeptide (IAPP)-induced IL-1β production [19], arguing for an additional anti-diabetic effect of glyburide besides its role as an insulin secretagogue.

The transcription factor NFκB is a key mediator of cytokine-induced apoptosis [20], with both pro- and anti-inflammatory properties [21,22]. The active transcriptional subunit p65 is consolidated in the cytoplasm by the inhibitor of NFκB (IκB) under resting conditions. However, IL-1β induces phosphorylation-dependent degradation of IκB via activation of the IκB kinase (IKKβ), which results in translocation of p65 to the nucleus and subsequent gene expression. Several small molecules target different enzymes in this pathway [23–26] but have not, to our knowledge, been tested in β-cells. For example, the β-carboline class of small molecules that inhibit the activity of IKKβ are effective in reducing NFκB signaling in a subgroup of diffuse large B-cell lymphoma [23,24]. In addition, high-throughput screening (HTS) has identified several activators and inhibitors of NFκB signaling [25,26]. Such small molecules may have promise in future studies to modulate pathways involved in cytokine-induced apoptosis.

Interestingly, the histone deacetylase (HDAC) inhibitors suberoylanilide hydroxamic acid (SAHA) and trichostatin A (TSA) (Figure 1) have been found to reduce cytokine-induced β-cell apoptosis in vitro [27]. The anti-inflammatory effects of HDAC inhibition in β-cells were observed with either IL-1β or IFN-γ signaling alone [28], indicating either overlapping HDAC targets in these signaling pathways or several targets of HDAC activity. In addition, TSA demonstrates protective effects against the development of type 1 diabetes in the NOD mouse [29]. Recently, a protective role of SAHA alone could not be confirmed in vivo; however, in combination with the dipeptidyl peptidase-IV inhibitor MK-626, an increased β-cell mass and decreased insulitis were observed [30]. However, in contrast to the TSA study, in which the HDAC inhibitor was administered for six weeks between week 18 and 24, the animals only received treatment for four weeks. Differences in concentration, delivery method (oral vs. intraperitoneal injections), and potency of the two different HDAC inhibitors may also explain the different outcomes.

Figure 1.

Small-molecule inhibitors of cytokine-induced β-cell death.

Of the eleven HDACs expressed in β-cells [31], HDACs 1 and 3 were identified as the mediators of β-cell cytotoxicity using small-molecule dissection and knock-down studies [32,33]. Although the mechanism has not been fully elucidated, HDAC3-selective knock-down reduced binding of the NFκB subunit p65 to DNA and subsequently reduced inflammatory gene transcription [33]. In 293T and HeLa cells, HDAC3 has been shown to affect NFκB activity through deacetylation of the subunit p65 [34,35]. A similar mechanism may be occurring in β-cells, in which NFκB-signaling may be targeted indirectly by HDAC inhibition.

Glycogen synthase kinase 3β (GSK3β) is another well-described target of small molecules, and has been studied for the treatment of diabetes. A phenotypic assay for HTS developed in our group identified the GSK3β inhibitor alsterpaullone (Figure 1) as a suppressor of cytokine-induced β-cell apoptosis [36]. The role that GSK3β plays in cytokine signaling in β-cells is largely unknown. GSKβ is negatively regulated by Akt [37], and proinflammatory cytokines have been found to increase β-cell apoptosis by down-regulating Akt activity, causing a decrease in expression of anti-apoptotic Bcl-2 proteins [38,39]. Thus, it could be speculated that an increase in GSK3β activity, presumably through cytokine-mediated inhibition of Akt, down-regulates anti-apoptotic proteins, and may explain the protective effects of GSK3β inhibition in cytokine-induced β-cells apoptosis.

In addition to alsterpaullone, several other screening hits increased β-cell viability in the presence of cytokines. These hits belonged to a diversity-oriented synthesis (DOS) library, in which novel small molecules were synthesized with varying structural complexities that resemble naturally occurring small molecules [40]. Structure-activity relationship (SAR) studies of one of these hits led to the discovery of BRD0476 (Figure 1), which has low-micromolar activity in inhibiting apoptosis [41]. Importantly, the compound also hampered cytokine-mediated loss of glucose-stimulated insulin secretion in the INS-1E β-cell line.

Small-molecule tyrosine-kinase inhibitors (TKIs) have been shown to not only prevent, but also reverse the progression of type 1 diabetes in vivo and in vitro in human islets [42,43]. (The therapeutic potential of TKIs for the treatment of diabetes has previously been reviewed by Mokhtari and Welsh [44]). These TKIs include the FDA-approved drug imatinib (Figure 1), which is used to treat chronic leukemia. Β-cell protection is likely through converged inhibition of the tyrosine kinase activities of platelet growth factor receptor (PDGFR) and c-Abl. The latter is known to phosphorylate and activate JNK, p38 MAP kinase, and, interestingly, NFκB signaling [42]. Of note, in non-β-cells, c-Abl causes cytochrome c release and apoptosis under tunicamycin-induced endoplasmic reticulum (ER) stress [45]. This mechanism may indicate a general role for c-Abl in ER stressed β-cells. Along these lines, the loss of β-cells observed in type 1 diabetic patients has been suggested to be caused by ER stress, as proinflammatory cytokines downregulate the Ca²⁺ pump sarcoendoplasmic reticulum Ca²⁺-ATPase (SERCA) and upregulate the C/EBP homologous protein (CHOP) [46]. CHOP is suggested to be a pro-apoptotic marker of ER stress and has been found to be over-expressed in type 1 diabetic patients [47]; thus, reducing cytokine-mediated ER stress in the β-cell may provide additional therapeutic opportunities. However, it must be noted that CHOP was expressed in β-cells in islet sections from obese type 2 diabetic patients, but not in newly diagnosed type 1 diabetics [48], calling into question the relevance of CHOP as a marker of ER stress.

Reactive Oxygen Species (ROS)-Induced β-Cell Death

β-Cells have a high metabolic activity due to glucose oxidation-dependent stimulus-secretion coupling, which results in formation of reactive oxygen species (ROS). With uniquely low expression of anti-oxidative enzymes such as catalase, superoxide dismutase and glutathione peroxidase, β-cells are particularly susceptible to ROS [49], and in both type 1 and type 2 diabetes, increased ROS formation leads to β-cell demise [50]. Several distinct features contribute to the formation of ROS. For example, proinflammatory cytokines induce the expression of inducible nitric oxide synthase (iNOS) which catalyzes the production of toxic nitric oxide (NO•) from L-arginine [51]. In addition, β-oxidation of saturated non-esterified fatty acids (NEFAs) in peroxisomes and mitochondria generate hydrogen peroxide (H₂O₂) (reviewed in [52]). Lastly, sustained levels of high glucose, likely through alternative catabolic pathways, lead to increased formation of hydroxyl radicals (•OH) [53].

Thus, formation of ROS seems to be a common denominator of various inducers of β-cell damage, explaining why antioxidants have been tested to preserve β-cells in different models of both type 1 and type 2 diabetes. The anti-oxidant small molecules N-acetyl-L-cysteine (NAC) and aminoguandine (AG) (Figure 2) reduced hyperglycemia and improved insulin responsiveness both in vivo (glucose tolerance test) and ex vivo (isolated islets assessed in the glucose-stimulated insulin secretion assay) in two different animal models for type 2 diabetes [54,55]. NAC is known to scavenge ROS directly, by reducing hydroxyl radicals (•OH) and hydrogen peroxide (H₂O₂), and indirectly, by promoting the biosynthesis of the endogenous antioxidant glutathione [56]. AG sequesters ROS in vivo in a similar manner to NAC [57]. In addition, AG inhibits formation of advanced glycosylation end (AGE) products by reacting with ROS-generated advanced dicarbonyl intermediates [58].

Figure 2.

Small-molecule inhibitors of ROS-induced β-cell death.

In cell-based assays, several other small-molecule antioxidants have been found to protect β-cells against ROS-induced damage (for a review, see [59]), arguing for beneficial roles of antioxidants per se on the β-cell. The rescue potential of antioxidants has also been shown in more relevant physiological settings in which β-cell death was induced by lipotoxicity [60] or glucotoxicity [61]. Importantly, Xiao et al. [62] did show that the antioxidant taurine (Figure 2), but not NAC, protected against a NEFA-mediated decrease in β-cell function in human subjects. Despite several clinical trials having failed to prove the ability of the lipophilic antioxidant vitamin E to improve glycemic control [63], a few trials have shown an improvement in glycosylated HbA_1c and fasting blood glucose by herbs with antioxidant characteristics [64].

Iron is a redox-reactive metal that can catalyze the generation of ROS. Hansen et al. [65] showed that knock-down of the divalent metal transporter (DMT)-1 in animal models for both type 1 and type 2 diabetes led to a reduction in blood glucose and an increase in circulating insulin levels. Furthermore, the authors demonstrated that the small-molecule iron chelators desferoxamine and deferasirox (Figure 2) reduced cytokine-induced ROS formation and β-cell apoptosis in vitro. Both desferoxamine and deferasirox are FDA-approved drugs for the treatment of acute iron toxicity and transfusion-induced iron overload, which may warrant their testing for β-cell protective effects in a clinical setting, with careful dose titration to avoid anemia. Intriguingly, increased levels of the iron-carrying plasma protein transferrin is associated with an increased risk of developing both type 1 and type 2 diabetes [66]. This finding provides epidemiological evidence for an association between iron levels and diabetes, which may be explained by increased oxidative stress and, as a consequence, apoptosis in the pancreatic β-cell.

Glucotoxicity-, Lipotoxicity- and Glucolipotoxicity-Induced β-Cell Death

In vitro, it is well established that sustained high levels of glucose and NEFAs such as palmitate (in combination or alone) induce apoptosis in the β-cell, which has given rise to the popular terms of glucotoxicity, lipotoxicity, and glucolipotoxicity. The combination of high glucose and some, but not all, free fatty acids like palmitate, linoleate and stearate has been reported to exhibit synergistic cytotoxic effects [67]. These toxicities are hypothesized to reflect what occurs in patients consuming diets rich in fats and who have inactive lifestyles that lead to insulin resistance, two important risk factors for developing type 2 diabetes. This gives rise to continuously increased circulating levels of glucose (hyperglycemia) and NEFA (hyperlipidemia), which eventually become toxic to the β-cell. Furthermore, it has been shown that increased levels of NEFA in human subjects impair β-cell function [68,69]. The mechanism of action is not completely understood, but substantial evidence suggests several contributors to apoptosis, such as ROS formation, oxidative stress, and induction of ER stress (reviewed in [70,71]). Several small molecules have been shown to protect against glucotoxicity, lipotoxicity, and/or glucolipotoxicity in the β-cell, including antioxidants, GSK3β inhibitors, a Ca⁺²-channel blocker, and FDA-approved type 2 diabetic therapeutics.

As described earlier in this review, alsterpaullone (Figures 1 and 3) is a GSK3β inhibitor that protects β-cells against proinflammatory cytokines. In addition to alsterpaullone, 1-azakenpaullone, 6-bromoindirubin-3’-oxime (BIO), and CHIR99021 (Figure 3) are increasingly specific GSK3β inhibitors that protect INS-1E β-cells against glucolipotoxicity [72]. In subsequent studies, the potency of 1-azakenpaullone was increased by replacing the bromide with a cyano functional group at the 9-position to generate 9-cyano-1-azapaullone (cazpaullone; Figure 3) [73]. Cazpaullone was designed using a rational structure-based strategy, in which docking studies revealed a favorable hydrogen bond between the cyano group of cazpaullone and Lys85 of GSK3β (Figure 4). GSK3β inhibitors were also shown to induce β-cell replication in rat islets [72], likely by decreasing the cyclin-dependent kinase inhibitor p27 and causing increased cyclin expression [74]. Importantly, CT 20026 (Figure 3), a structurally related analog of CHIR99021 with increased solubility and stability, improved the diabetic state in ZDF rats [75], providing further optimism for GSK3β as a drug target for preventing β-cell apoptosis and/or inducing β-cell replication.

Figure 3.

GSKβ inhibitors that suppress glucotoxicity, lipotoxicity and glucolipotoxicity-induced β-cell death.

Figure 4.

Overlay of alsterpaullone (light-brown) and cazpaullone (atom color code) bound to the ATP binding site of GSK3. Hydrogen bonds involving cazpaullone are indicated as green dashed lines. Reprinted with permission from Stukenbrock et al. J. Med. Chem. 2008, 51 (7). Copyright 2008 American Chemical Society.

The Ca⁺²-channel blocker verapamil (Figure 5) not only reduced but also reversed STZ-induced diabetes, a model for type 1 diabetes, and ameliorated blood glucose levels in the ob/ob mouse, a genetic model of type 2 diabetes [76]. In both instances, insulin levels were improved and β-cell apoptosis was reduced, indicating that Ca⁺²-channel blockers preserve β-cell viability and function. The mechanism of action of verapamil may involve inactivation of the Ca⁺²-dependent phosphatase calcineurin, causing sustained phosphorylation and inactivation of the transcriptional regulator carbohydrate-responsive element-binding protein (ChREBP) and, thus, lowering transcription of the pro-apoptotic transcription factor thioredoxin-interacting protein (TXNIP) [76]. Of note, TXNIP has been identified as a critical inducer of β-cell apoptosis in models for type 1 and type 2 diabetes [77,78]. In vitro, TXNIP is associated with glucotoxicity [77] but not lipotoxocity [79]. Its role in proinflammatory cytokine-induced β-cell apoptosis has yet to be described, but this is of particular importance based on the promising results from the STZ-model. Recently, TXNIP was linked to ER stress [80,81] and ROS generation as well as activation of the inflammasome [82], suggesting a role for low-grade inflammation in both type 1 and type 2 diabetes.

Figure 5.

Additional small-molecule inhibitors of glucotoxicity-, lipotoxicity- and glucolipotoxicity in β-cells.

Several drugs currently used for the treatment of type 2 diabetes have also been reported to show some direct effects on the β-cell. Metformin (Figure 5), the first-in-line drug for obese type 2 diabetic patients, has been shown in vitro to protect rat and human islets from glucotoxicity and lipotoxicity [83,84]. In the β-cell, metformin has been proposed to change the catabolism of fatty acids from esterification towards oxidation through activation of AMPK, thereby reducing lipotoxicity [67]. However, a detrimental role of metformin, through activation of AMPK, has also been reported [85], challenging the β-cell protective effects of this drug.

Recently, metformin was shown to prevent glucotoxicity-induced opening of permeability transition pores (PTP) in INS-1 cells [86]. Extensive opening of PTPs causes cell death via mitochondrial depolarization, which feasibly is induced in the β-cell by high glucose through ROS formation or induction of mitochondrial-permeable members of the Bcl-2 family [86]. It could be speculated that the effects of metformin on PTP opening is a secondary effect of reduction in ROS formation through induction of AMPK activity [87], but this hypothesis has yet to be confirmed.

Thiazolidinediones are small molecules that activate the nuclear receptor PPARγ to increase insulin sensitivity [88]. This class includes the anti-diabetic drugs rosiglitazone and troglitazone (Figure 5), which have beneficial effects on β-cell function in type 2 diabetic patients [89]. Both rosiglitazone and troglitazone have been shown to preserve β-cell mass and function in the ZDF rat model of type 2 diabetes [90,91]. Furthermore, rosiglitazone reduces glucolipotixicity in vitro in early, but not later, phases of glucolipotoxicity-induced apoptosis [92]. However, troglitazone failed to protect β-cells against lipotoxicity in vitro [93], despite other reports showing protective effects on β-cell function [94]. Extensive future studies will be needed to explain these different outcomes. Interestingly, troglitazone has been shown to protect β-cells against proinflammatory cytokines in vitro [95], raising another explanation for the beneficial outcomes of troglitazone in human subjects, i.e. reducing the damage of low-grade inflammatory assault on β-cells.

Lastly, cyclosporine, which is used in transplantation to prevent graft-versus-host disease, occasionally causes post-transplantation diabetes by reducing β-cell mass. In vitro and in vivo, rosiglitazone has been shown to inhibit cyclosporine-induced β-cell apoptosis, likely by reducing ER stress [96]. Taken together, these studies provide substantial evidence to support the β-cell protective effects of rosiglitazone in vivo and, to some extent, in vitro.

Future Perspectives

In 1922, Leonard Thompson was the first diabetic to receive successful treatment in the form of an exogenous agent, bovine insulin. Since then, the hunt for better and safer drugs has led to improved therapeutics for diabetes, such as insulin analogues, PPARγ agonists, and GLP-1 analogues. However, life-long insulin therapy is still the preferred treatment for type 1 diabetes. Due to the immune-mediated etiology of type 1 diabetes, the discovery of small molecules that preserve functional β-cell mass, in combination with suitable biomarkers for early β-cell loss, could potentially prevent life-long insulin injections and complications in diabetes-prone patients. As described in this review, several classes of small molecules show potential to protect β-cells against proinflammatory cytokines in vitro and in vivo, although human clinical trials will be needed to determine the utility of this approach as a therapeutic strategy.

The development of HTS methods allows the screening of thousands of small molecules in isolated cell systems modeling different modes of β-cell death. These HTS assays have led to increasing rates of discovery of protective small molecules, and hits from such screens will need to travel the long road from bench to bedside, in which these small molecules will need to be evaluated for pharmacokinetic properties, toxicity, and efficacy. Nevertheless, a variety of small molecules with diverse chemical structures, as illustrated in this review, have been validated both in vitro and in vivo to potentially treat type 1 and 2 diabetes by inhibiting inflammatory β-cell death.

Although a considerable number of small molecules show β-cell protective characteristics both in vitro and in vivo, no approved anti-diabetic treatment has confirmed β-cell protective properties in humans. A functional β-cell mass is essential in preventing the development of type 1 and type 2 diabetes, hence it is of critical importance to identify novel β-cell-protective small molecules. Several questions need to be answered to reach this goal. First, a clear role of existing diabetic therapies should be understood at the level of the β-cell, exemplified by metformin, in which contradicting evidence has been reviewed here. Moreover, will a β-cell-specific therapeutic be sufficient to encounter the pathogenic consequences of increased insulin resistance, or should the aim be combination therapy? Theoretically, with the insulin-resistant but non-diabetic population in mind, a β-cell-exclusive therapy seems adequate, but with several targets in combination therapies, the window of opportunity is larger in terms of dose, duration, and onset of treatment(s), suggesting that this mode should be the preferred aim.

A general concern in small molecule-based therapeutic strategies is the high risk of severe side effects, by inhibiting ubiquitously expressed targets and targets with pleiotropic properties. However, as exemplified by the use of tyrosine-kinase inhibitors and HDAC inhibitors in cancer treatment, it is a feasible approach, further supported by the β-cell protective effects of several small molecules in animal models reviewed here.

Another important aspect to consider is how deleterious actions should be hampered. Although multiple clinical studies using anti-oxidants have been conducted, a clear effect is lacking. Instead of targeting the already present ROS, approaches that interfere with the production of it, or increase the endogenous anti-oxidant system, might provide better outcomes. Another example of a different approach is the targeting of NFκB. Though several small-molecule inhibitors exist, none has so far been used in the setting of the β-cell, probably due to both pro- and anti-apoptotic outcomes of NFκB signaling. However, as described here, both HDACs and tyrosine-kinases interfere with NFκB-signaling, and inhibitors of both classes elicit β-cell protective properties. Hence, indirect inhibition of pro-apoptotic, but not anti-apoptotic, NFκB signaling by these inhibitors provides a novel approach to selectively target pro-apoptotic NFκB-mediated signal cascades.

Another important issue to consider is the translation of substantial pre-clinical evidence into clinical trials, which may give rise to negative results, exemplified by the IL-1β antagonist clinical trials described here. Is the dose and/or the durability of the treatment correct? Perhaps combinatorial therapy will provide β-cell protection? The specificity and selectivity of small molecules in a physiologic system is another largely unknown field. The majority of reported IC₅₀ values are determined in either cell-free or cell line-based settings, but are these values relevant in a primary cell or whole organism in a clinical setting? Unknown off-targets, complex formation, and differences in ion concentrations are also likely to interfere. Along these lines, SAR analysis is a valuable tool to increase selectivity and thereby reduce off-target effects. In conclusion, exciting advances are seen in the field of pancreatic β-cell-active small molecules, but a clear clinical proof-of-concept has yet to be produced.