Glucokinase Inhibition: A Novel Treatment for Diabetes?

Failure of pancreatic β-cell function and loss of β-cell mass are key events in the development and progression of multiple forms of diabetes (1–4). Autopsy studies reveal ∼40–60% reduction of β-cell mass in people with type 2 diabetes (T2D) compared with age-, sex-, and BMI-matched individuals without diabetes (2,5,6). In T2D, insulin resistance is initially compensated by an increase in insulin secretion, followed by a substantial loss of functional β-cell mass closer to the clinical manifestation (7,8). Hyperinsulinemia has been shown to be a strong predictor of diabetes in numerous populations (9,10). Since current diabetes therapies are not initiated until individuals are prediabetic, at which time islets may have lost 50–80% of endogenous function (2,11), approaches aimed at preserving β-cell function and mass may be key not only for treating diabetes, but also for preventing its development and progression.

Extracellular glucose can have multiple effects on the β-cell depending on stimulation intensity and duration. While short-term or slightly elevated glucose can induce β-cell adaptation, sensitization of glucose-stimulated insulin secretion, and upregulation of glycolytic enzymes and accelerated glucose oxidation (4,12–14), chronic exposure to high glucose can induce reduction of functional β-cell mass and development of frank diabetes (6,15,16).

Glucokinase (GCK) (hexokinase type IV) is the first enzyme in β-cell glycolysis, and it phosphorylates glucose to glucose-6-phospate (Fig. 1). Mainly expressed in the pancreas and liver, but also in the hypothalamus and gastrointestinal tract (17), GCK has a uniquely low affinity for glucose, narrow substrate specificity, and lack of product inhibition. In pancreatic β-cells, GCK plays a critical rate-limiting role in glucose metabolism, and hence in glucose-stimulated insulin secretion (Fig. 1), and in the liver it initiates glucose uptake to storage as glycogen and suppresses gluconeogenesis (18). The central role of GCK in glucose regulation of insulin secretion is highlighted by the demonstration that heterozygous gain-of-function (GOF) mutations cause permanent hyperinsulinemia hypoglycemia of infancy with excessive insulin secretion (19,20), and heterozygous and homozygous loss-of-function (LOF) mutations cause maturity-onset diabetes of the young 2 (MODY2) and neonatal diabetes (ND), respectively, with decreased insulin secretion (19–21) (Fig. 1A). Consistent with absence or reduction of GCK being expected to decrease glucose metabolism and insulin secretion (22), mice lacking GCK, either globally or specifically in pancreatic β-cells, die of severe diabetes within a few days after birth, whereas heterozygous GCK knockout mice demonstrate mild hyperglycemia and reiterate features of human MODY2 (23–25). GCK knockout islets show defective insulin secretion in response to glucose but an almost normal response to other secretagogues (24). Mice lacking GCK only in the liver (but normal GCK in β-cells) display pronounced defects in both glycogen synthesis and glucose turnover rates during hyperglycemic clamps, but are only mildly hyperglycemic (25).

Increased activation of GCK has been proposed as a potential cause for the compensatory increase in insulin secretion in early stages of T2D, and in pursuit of further benefit, GCK activators (GCKAs) have been developed as potential antidiabetic agents (reviewed in Nakamura et al. [26]). Their ability to acutely lower blood glucose through increasing insulin secretion and β-cell proliferation, as well as to promote glucose uptake in the liver, has been shown in several murine models of diabetes (27,28). However, clinical application of GCKAs has been hampered by negative side effects, such as hypoglycemia and dyslipidemia, and by unexplained lack of long-term glycemic control (26,29–31). Interestingly, mice with genetic activation of GCK in β-cells demonstrate an initial increase in insulin secretion and β-cell proliferation, but within a week, they progress to β-cell damage associated with DNA double-strand breaks, increased p53 activity, and insulin secretion failure (32). Recent work has suggested that normal islets chronically exposed to high glucose in vitro develop a “left shift” in glucose sensitivity, with lower levels of glucose triggering insulin secretion (33–35). The paradoxical increase in insulin secretion in response to lowering glucose in insulin-resistant Zucker fatty and Zucker diabetic fatty rats was associated with enhanced fuel metabolism through glycolysis, an effect that was suppressed by inhibition of glycolysis with d-mannuheptulose (36). These results are in agreement with those from mice with 90% pancreatectomy (37). In these animals, the consequent markedly increased requirement of secretion by the remaining pancreatic islets is initially compensated by β-cell glucose hypersensitivity and insulin hypersecretion before secretory failure (37). The initial β-cell adaptation to insulin resistance, GCK activation, and subsequent β-cell failure appear to correspond to the natural progression of T2D (4,38), and overstimulation of glucose metabolism through β-cell GCK has thus been proposed as a culprit in subsequent loss of functional β-cell mass (39,40) (Fig. 1B).

Although β-cells have shown the ability to regain function when rested from glucose-driven hyperexcitability or hypersecretion or when they are removed from the toxic unfavorable environment (40–44), the durability of this recovery remains unclear. Preclinical and clinical studies lead us to counterintuitively suggest that decreasing glucose metabolism by reducing GCK activity in β-cells could actually be a successful approach to long-term preservation of functional β-cell mass in diabetes. First, individuals carrying GCK-LOF mutations (MODY2) show nonprogressive, mild-fasting hyperglycemia, and a glucose tolerance and counterregulatory response to hypoglycemia similar to those of individuals without diabetes. More importantly, they show no decrease in β-cell function over time (assessed by glucose tolerance test) (45,46). Islets from obese T2D db/db mice treated with the hexokinase selective inhibitor d-mannoheptulose demonstrated reinstated glucose sensing, ATP content, and NAD(P)H flux as well as restored calcium oscillations, pulsatility, and insulin secretion (47). Moreover, db/db mice with GCK haploinsufficiency specifically in β-cells (GCK^+/−db/db mice) demonstrate increased β-cell mass, enhanced glucose intolerance, and improved insulin secretion compared with db/db mice (48). Islets from GCK^+/−db/db mice show lower expression of endoplasmic reticulum stress genes and higher expression of mature β-cell identity markers, less mitochondrial damage, and increased β-cell area/mass compared with db/db mice (48). Although there were no differences in fed blood glucose and glucose tolerance between GCK^+/−db/db and db/db mice aged 10 weeks, insulin content and β-cell proliferation were significantly higher in GCK^+/−db/db mice, with no differences in apoptosis or progenitor markers, such as neurogenin 3 and aldehyde dehydrogenase 1a3 (48). Strikingly, in our K_ATP GOF mouse model of monogenic neonatal diabetes (43,49), heterozygous knockout of GCK in β-cells markedly delayed development of diabetes, with an overall reduction of blood glucose over time and improved glucose tolerance (50). Most importantly, β-cell insulin content, proinsulin-to-insulin ratio, and β-cell mass and identity were all preserved relative to K_ATP-GOF mice with normal GCK activity (50). Thus, although paradoxical and counterintuitive, decreasing glycolytic rate by partial GCK inactivation can preserve insulin secretion and prevent glucotoxicity-induced loss of insulin content and β-cell failure (Fig. 1B).

To develop effective strategies for diabetes prevention, we propose to move the focus from interventions that reduce glucose levels toward those that protect functional β-cell mass, essentially shifting the paradigm from a glucose-centric to a β-cell approach. It has been inferred from clinical studies that decline in functional β-cell mass begins before the onset of T2D and proceeds thereafter, leading to worsening glycemic control and requiring progressive intensification of diabetes therapy, often culminating in the need for exogenous insulin therapy (2,51). In recent years it has become clear that antidiabetic drugs that initially promote insulin secretion by increasing metabolic or electrical signaling can lower blood glucose, but they may not be optimal for long-term preservation of functional β-cell mass (2,52). Instead, as discussed above, evidence from preclinical studies that genetic GCK haploinsufficiency in animal models of obesity/T2D and in K_ATP-induced monogenic neonatal diabetes improves glucose tolerance and prevents loss of functional β-cell mass (26,39,47,48,50) points to reduction of glucose metabolism as a potentially highly effective treatment for diabetes (Fig. 1B). The proposed paradoxical use of GCK inhibitors (GCKIs) is supported by the demonstration that individuals with MODY2 have glucose tolerance similar to that of nondiabetic subjects, and that their β-cell function does not decrease over time (45,46).

A clear challenge to the use of GCKIs as a treatment for diabetes will be titration of the degree of GCK reduction, since complete GCK inactivation will cause severe hyperglycemia, evidenced by the permanent ND caused by homozygous GCK-LOF mutations. A second challenge will be selection of appropriate agents to achieve the desired GCK reduction as well as the timing of treatment initiation. We propose that GCKIs be introduced during the compensatory hyperinsulinemic phase, in which there is, presumably, a high β-cell glycolytic rate (Fig. 1), before β-cell failure. However, better understanding of metabolic stress in prediabetes, and the mechanisms underlying progression from hypersecretion to decline in β-cell function, will be needed to guide clinical action. Even though d-mannuheptulose has been used in vitro to inhibit GCK in islets from db/db obese/diabetic mice, with beneficial effects on β-cell function (47), its in vivo use will be problematic since it is a nonselective hexokinase inhibitor with activity in the millimolar range (39). Thus, a third challenge will be to find inhibitory drugs that specifically target β-cell GCK. Besides expression in β-cells, GCK is also present in the liver, where it also plays an important role in glucose homeostasis by increasing or decreasing glucose production and storage. While glucose is the primary regulator of pancreatic GCK expression, insulin signaling regulates GCK in the liver (53). Thus, through their direct action on liver GCK, GCKIs could affect glycogen synthesis and glucose turnover. However, indirect effects through crosstalk with signaling coming from the action of GCKIs on β-cells, presumable reduction of glycolytic activity in overactive β-cells, and therefore preservation of β-cell function-insulin secretion, could add complexity (39).

GCK is also present in the brain and gastrointestinal tract and in pancreatic glucagon-producing α-cells (17,39,54,55). In α-cells, GCK plays an important role in the response to glucose sensing and glucagon secretion: high-fat diet–fed mice with genetic activation of α-cell GCK demonstrated reduced glucagonemia (54), and mice lacking α-cell GCK displayed hyperglucagonemia and loss of glucose control of glucagon secretion (55). Studies in the brain suggest that ventromedial hypothalamus GCK activity is an important regulator of the counterregulatory response to insulin-induced hypoglycemia and inhibition of hypothalamic GCK activity upon insulin-induced hypoglycemia might improve the dampened counterregulatory response observed in tightly controlled diabetic subjects (56). A further potential issue is whether GCKI monotherapy will be sufficient for normalization of glucose sensing, avoidance of cellular stress, restoration of pulsatility, and preservation of functional β-cell mass during development and progression of diabetes.

Despite these challenges, the striking effects of GCK inhibition in protection from diabetes development in animal models leads us to propose that inhibition of glucose metabolism be considered as a potential treatment approach in the early hyperinsulinemic stage to prevent or delay the subsequent β-cell failure that comes with diabetes progression (Fig. 1B). Given the experimental evidence, we have centered our consideration on the potential use of GCKIs specifically to halt progression to β-cell dysfunction. Other potential pathway targets (e.g., G6PC2, which is highly expressed in human and mouse β-cells) could provide alternatives for reducing metabolic flux and preserving β-cell function: although new targets willraise additional issues regarding their expression and effects in other tissues.

The critical role of GCK in β-cell glucose metabolism and insulin secretion is highlighted by the fact that inactivating and activating mutations cause diabetes and hyperinsulinism, respectively. Given this paradigm, GCK activation has been tried, unsuccessfully, as a therapeutic approach to maintain insulin secretion in diabetes. Paradoxically, GCK inhibition, by reducing glycolytic flux, has been shown to preserve β-cell mass, reduce endoplasmic reticulum stress markers, conserve mitochondrial morphology and function, and maintain mature β-cell identity in preclinical mouse models of monogenic diabetes and T2D. Such studies 1) provide evidence for β-cell glucotoxicity being induced by increased glucose metabolism, rather than by exhaustion due to β-cell overwork and insulin hypersecretion: 2) delineate a common pathway for the progressive β-cell failure with loss of β-cell mass and function in different forms of diabetes: and 3) identify a counterintuitive novel therapeutic approach to potentially halt diabetes progression by preserving functional β-cell mass. Thus, GCKIs provide an attractive target for normalization of glucose sensing, avoidance of cellular stress, restoration of pulsatility, and preservation of functional β-cell mass during development and progression of diabetes.

Funding. M.S.R. is supported by National Institutes of Health grant DK123163, and C.G.N. is supported by National Institutes of Health grant HL140024.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.