Allosteric Activators of Glucokinase: Potential Role in Diabetes Therapy

See allHide authors and affiliations

Science  18 Jul 2003:
Vol. 301, Issue 5631, pp. 370-373
DOI: 10.1126/science.1084073


Glucokinase (GK) plays a key role in whole-body glucose homeostasis by catalyzing the phosphorylation of glucose in cells that express this enzyme, such as pancreatic β cells and hepatocytes. We describe a class of antidiabetic agents that act as nonessential, mixed-type GK activators (GKAs) that increase the glucose affinity and maximum velocity (Vmax) of GK. GKAs augment both hepatic glucose metabolism and glucose-induced insulin secretion from isolated rodent pancreatic islets, consistent with the expression and function of GK in both cell types. In several rodent models of type 2 diabetes mellitus, GKAs lowered blood glucose levels, improved the results of glucose tolerance tests, and increased hepatic glucose uptake. These findings may lead to the development of new drug therapies for diabetes.

Glucose homeostasis is lost in type 2 diabetes because of combined defects in both insulin secretion and insulin action (1, 2). The characterization of patients with abnormal glycemic control due to either gain- or loss-of-function mutations in GK has provided new insights into the pathogenesis of type 2 diabetes. Loss-of-function mutations in the gene encoding GK have been linked to maturity-onset diabetes of the young type 2 (MODY2), an autosomal dominant form of diabetes mellitus characterized by early onset and mild chronic fasting hyperglycemia (3, 4). MODY2 patients display impaired glucose responsiveness of β cells, decreased net accumulation of glycogen, and increased hepatic glucose production after meals (5, 6). The GK mutations found in MODY2 patients result in decreased activity of this enzyme as a result of reduction in its Vmax and/or reduced affinity toward its substrates, glucose and adenosine triphosphate (ATP) (711). In contrast, gain-of-function GK mutations, which increase the catalytic activity of GK, cause persistent hyperinsulinemic hypoglycemia of infancy as a result of lowering the threshold for glucose-stimulated insulin release (GSIR) (12, 13). These results illustrate the preeminent role of the GK glucose sensor in glucose homeostasis and suggest that a pharmacological activation of GK enzymatic activity in type 2 diabetes could have clinical benefits.

To search for small molecules that increase GK enzymatic activity, we screened a library of 120,000 structurally diverse synthetic compounds. One compound increased the enzymatic activity of GK, and chemical optimization of this initial molecule led to the synthesis of RO-28-0450 as a lead GK activator (GKA) (Fig. 1A, inset). Because RO-28-0450 is a racemic compound, we synthesized and tested the R- and S-enantiomers. Activation of GK was exquisitely sensitive to the chirality of the molecule: The R enantiomer, RO-28-1675, was a potent GKA, whereas the S enantiomer, RO-28-1674, was inactive (Fig. 1A). RO-28-1675 also reversed the inhibitory action of the human glucokinase regulatory protein (GKRP) (14) (Fig. 1A).

Fig. 1.

Chemical structure of RO-28-0450 (inset) and effects of enantiomers on GK activity. (A) Effects of RO-28-1675 (solid symbols) and RO-28-1674 (open symbols) enantiomers on activation of GK in the absence (squares) and presence (circles) of inhibitor (GKRP). Effects of test compounds are expressed as a percentage of untreated GK activity. (B) Rate (v) versus glucose (S) plot in the absence (solid circles) and presence of RO-28-1675 (open circles, 0.03 μM; solid squares, 0.1 μM; open squares, 0.3 μM; solid triangles, 1 μM; open triangles, 3 μM; solid diamonds, 10 μM; open diamonds, 20 μM). The data were fit with DataFit software (version 7.1; Oakdale Engineering, Oakdale, PA) to the velocity equation for a nonessential mixed-type activator (21) that was modified for an enzyme showing cooperative kinetics. The graph shows actual data points (symbols) and best fit line. Inset shows the 1/v versus 1/[S]1.558 plot in the presence of various fixed concentrations of RO-28-1675. (C and D) Slope1/S versus RO-28-1675 concentration (C) and 1/v-axis intercept versus RO-28-1675 concentration replots (D) using the slopes and y-axis intercepts, respectively, from the 1/v versus 1/[S]1.558 plot (see inset). GK activity was measured as described (14).

RO-28-1675 increased the enzymatic activity of recombinant human GK in a dose-dependent manner (Fig. 1B). At a concentration of 3 μM, RO-28-1675 increased the Vmax of GK by a factor of about 1.5 and decreased the substrate concentration at 0.5Vmax ([S]0.5) for glucose from 8.6 mM to 2.0 mM. The ability of RO-28-1675 to decrease the glucose [S]0.5 of GK (Fig. 1C) and increase the Vmax of the enzyme (Fig. 1D) was dose-dependent. The Hill coefficient (fig. S1) and the affinity for the second substrate, ATP (15), were unaffected by RO-28-1675. GKAs were specific for GK and did not activate hexokinase isozymes of brain (HK-I) or muscle (HK-II) (15).

We next evaluated the effects of RO-28-1675 on GSIR with freshly isolated perifused rat pancreatic islets under conditions in which glucose was increased progressively from 0 to 20 mM, with 1-mM increments over a 20-min period (Fig. 2). The threshold concentrations of glucose required to induce insulin release above basal rates decreased stepwise from 7 to 3 mM glucose when islets were treated with 0 to 10 μM RO-28-1675. This compound also increased the magnitude of the response at glucose concentrations known to cause near-maximal insulin release in untreated islets. Experiments with isolated islets confirmed that these effects were due to increased glucose usage, as determined by the usage of [2-3H]glucose and increase in intracellular free Ca2+ (15). The specificity of GKAs upon fuel-stimulated insulin release was tested by evaluating their effect on the metabolism of 2-keto[1-14C]isocaproic acid (α-KIC), a powerful glucose-independent fuel stimulant of β cells. Neither the metabolism of α-KIC nor its insulin-releasing action were influenced by GKAs (15). Thus, in contrast to sulfonylureas or meglitinides, which induce insulin release by inhibiting the KATP channel, GKAs enhance insulin release by activating glucose metabolism.

Fig. 2.

Effect of RO-28-1675 on GSIR in freshly isolated, perifused rat islets. Islets were isolated from 200- to 250-g fed male Wistar rats as in (22). After a 39-min preperifusion phase without substrate (14), a glucose ramp was applied from 0 to 20 mM (1 mM/min) in the presence of untreated (solid circles), 0.5% dimethyl sulfoxide (DMSO, open circles), and RO-28-1675 (solid squares, 0.3 μM; open squares, 1 μM; solid triangles, 3 μM; open triangles, 10 μM). For reasons of clarity, standard errors are shown for the DMSO control and for the 3-μM drug level but not for the other conditions. There were at least three perifusions for each condition.

Administration of a single oral dose of RO-28-1675 reduced blood glucose levels in wild-type C57BL/6J mice, whereas RO-28-1674, the inactive stereoisomer, had no effect (Fig. 3A). Similarly, a single oral dose given to diet-induced obese (DIO) C57BL/6J mice, an animal model of type 2 diabetes, caused a dose-dependent reduction in blood glucose (Fig. 3B). The drug's glucose-lowering effects were paralleled by an increase in plasma insulin levels that peaked at 45 min after administration (Fig. 3, C and D). RO-28-1675 was similarly effective in reducing basal blood glucose levels in other animal models of type 2 diabetes, including KK/Upj-Ay/J mice, ob/ob mice, and Goto-Kakizaki rats (fig. S2, A to C). However, in older db/db mice that have blood glucose levels near 300 mg/dl and are hypoinsulinemic, GKAs lost their effectiveness (15).

Fig. 3.

Glucose-lowering and insulin-releasing effects of RO-28-1675 in mice. (A) Blood glucose levels in male C57BL/6J mice (10 weeks old, Jackson Labs) treated with a single oral dose of vehicle (solid circles), RO-28-1675 (50 mg/kg, open circles), or RO-28-1674 (50 mg/kg, squares). (B) Glucose levels in male DIO C57BL/6J mice (19 weeks old) fed a high-fat diet (No. F1850, Bioserve, Frenchtown, NJ) for 12 weeks and then treated with a single oral dose of vehicle (solid circles) or RO-28-1675 (10 mg/kg, open circles; 30 mg/kg, solid squares; 50 mg/kg, open squares). (C and D) Simultaneous plasma glucose levels (C) and insulin levels (D) measured for each time point in C57BL/6J mice treated with vehicle (solid circles) or RO-28-1675 (50 mg/kg, open circles). All mice were fasted for 2 hours before oral administration of the test compound. Food was withheld until the end of the study. Mice had free access to water. Glucose levels were measured with a YSI Model 2700 Biochemistry Analyzer (YSI Inc., Yellow Springs, OH). Plasma insulin was determined with a radioimmunoassay kit (Linco, St. Louis, MO). Polyethylene glycol 400:Gelucire 44/14 (40:60) was used as the vehicle for all in vivo studies. Results are reported as means ± SEM (n = 6 to 10 per time point). A Student's t test was used to test for statistical significance (*P < 0.05, **P < 0.01, ***P < 0.005). All animal procedures were approved by the Institutional Animal Care and Use Committee.

Oral administration of RO-28-1675 also significantly improved postprandial glucose control, as assessed by an oral glucose tolerance test (OGTT) in ob/ob mice and DIO C57BL/6J mice (Fig. 4) as well as in C57BL/6J mice, Wistar rats, and Goto-Kakizaki rats (fig. S3, A and B). The improved response in the OGTT occurred without further enhancement of insulin levels, relative to the control group, after the glucose load (15). This is consistent with improved hepatic glucose disposal in the presence of GKAs, presumably due to extrapancreatic effects of this compound.

Fig. 4.

Acute effects of orally administered RO-28-1675 on an oral glucose tolerance test in DIO and ob/ob mice. (A) Male DIO C57BL/6J mice (19 weeks old) were administered a single oral dose of vehicle (solid circles) or RO-28-1675 (15 mg/kg, open circles). (B) Female ob/ob mice (9 weeks old, Jackson Labs) were administered a single oral dose of vehicle (solid circles) or RO-28-1675 (10 mg/kg, open circles) as described in Fig. 3. Mice (n = 10 per treatment group) were fasted overnight, administered test compounds, and given glucose by oral gavage (2 g/kg) 2 hours after dose. A Student's t test was used to test for statistical significance (**P < 0.01, ***P < 0.005).

To determine whether RO-28-1675 has a direct effect on the liver, in addition to its action on β cells, we studied 18-hour-fasted conscious Sprague-Dawley rats maintained on a pancreatic clamp (fig. S4). During the hyperglycemic phase of the study, in which blood glucose and plasma insulin levels were clamped at the same levels in both the vehicle and RO-28-1675 treatment groups, RO-28-1675 increased glucose disposal rates (table S1). As expected, hyperglycemia per se decreased net endogenous glucose production (EGP) in the vehicle group, and treatment with RO-28-1675 appeared to reverse EGP in favor of hepatic glucose uptake (table S1). Similar experiments in male ZDF-Gmi rats, where hyperglycemia did not suppress EGP, showed that GKAs significantly reduced EGP (15). We attribute the apparent switch from net EGP to net hepatic glucose uptake to the stimulatory effect of RO-28-1675 on hepatic GK. Increases in hepatic glucose-6-phosphate, fructose-6-phosphate, lactate, and glycogen levels were also observed in the RO-28-1675 group relative to the vehicle group (fig. S5, A to D). The effect of RO-28-1675 on glycogen synthesis during the clamp was consistent with observed increases in glycogen synthesis in rat primary hepatocytes treated with RO-28-1675 (15).

Taken together, our results suggest that RO-28-1675 has a dual mechanism of action: It enhances insulin release from the pancreas, and it stimulates glucose usage in the liver. These data are consistent with studies in which hepatic GK was overexpressed in diabetic mice (1619) and further support a key role for hepatic GK in maintaining glucose homeostasis.

A central issue in drug discovery programs to advance diabetes therapy is the identification of promising drug targets. A plausible rationale for narrowing the options is to focus on enzymes that have a high control strength in glucose homeostasis, such as GK (20). On the basis of this perspective, we anticipate that GKAs will influence glucose homeostasis in humans as they do in experimental animals. In support of this rationale, GKAs appear to mimic the effects of currently known GK-activating mutations Val455 → Met (12) and Ala456 → Val (13) associated with hyperinsulinism and hypoglycemia in humans. Interestingly, evidence suggests that GKAs interact with an allosteric activator site that coincides with a region of the enzyme where these mutations cluster (fig. S6).

Hallmarks of diabetes include impaired insulin secretion and enhanced hepatic glucose production. Our enthusiasm for further development of GKAs is based on the consideration that this class of molecules will have an impact on these two physiological endpoints in people with type 2 diabetes. Indeed, the use of combination therapy is thought to offer better glycemic control relative to monotherapy. Therefore, the ability of GKAs to influence both insulin release and hepatic glucose metabolism could provide greater efficacy as a monotherapy. Future studies of the compounds described here may lead to safe and effective GKAs for clinical trials and new pharmacological tools for elucidating the complex role of GK in glucose homeostasis.

Supporting Online Material

Materials and Methods

Figs. S1 to S6

Table S1


References and Notes

Stay Connected to Science

Navigate This Article