MLR-1023 [Tolimidone; CP-26154; 2(1H)-pyrimidinone, 5-(3-methylphenoxy)] is an allosteric Lyn kinase activator that reduces blood glucose levels in mice subjected to an oral glucose tolerance test (J Pharmacol Exp Ther 342:15–22, 2012). The current studies were designed to define the role of insulin in MLR-1023-mediated blood glucose lowering, to evaluate it in animal models of type 2 diabetes, and to compare it to the activities of selected existing diabetes therapeutics. Results from these studies show that in an acute oral glucose tolerance test MLR-1023 evoked a dose-dependent blood glucose-lowering response that was equivalent in magnitude to that of metformin without eliciting a hypoglycemic response. In streptozotocin-treated, insulin-depleted mice, MLR-1023 administration did not affect blood glucose levels. However, MLR-1023 potentiated the glucose-lowering activity of exogenously administered insulin, showing that MLR-1023-mediated blood glucose lowering was insulin-dependent. In a hyperinsulinemic/euglycemic clamp study, orally administered MLR-1023 increased the glucose infusion rate required to sustain blood glucose levels, demonstrating that MLR-1023 increased insulin receptor sensitivity. In chronically treated db/db mice, MLR-1023 elicited a dose-dependent and durable glucose-lowering effect, reduction in HbA1c levels and preservation of pancreatic β-cells. The magnitude of effect was equivalent to that seen with rosiglitazone but with a faster onset of action and without causing weight gain. These studies show that MLR-1023 is an insulin receptor-potentiating agent that produces a rapid-onset and durable blood glucose-lowering activity in diabetic animals.
Type 2 diabetes (T2D) is a metabolic disease that is driven primarily by insulin receptor insensitivity (Petersen and Shulman, 2006). This receptor insensitivity leads to steadily increasing blood glucose levels, compensatory elevations in serum insulin followed by a degeneration of insulin-secreting pancreatic β-cells, a precipitous drop in serum insulin levels, and elevated HbA1c levels (Kahn, 2003). Thiazolidinediones (TZDs) such as rosiglitazone increase insulin receptor sensitivity and promote glycemic control through the activation of the peroxisome proliferator-activated receptor γ (PPARγ) nuclear hormone receptor (Rosen and Spiegelman, 2001; Sears et al., 2009). Although TZDs increase insulin receptor sensitivity, treatment with these drugs and chronic activation of PPARγ lead to edema and weight gain and produce cardiovascular risks (Ajjan and Grant, 2008; Bergenstal et al., 2010). The clinical limitations associated with the current generation of insulin-sensitizing agents have driven drug discovery research toward the identification of insulin receptor-sensitizing agents that act independently of PPARs (DeFronzo, 2010).
MLR-1023 [tolimidone; CP-26154; 2(1H)-pyrimidinone, 5-(3-methylphenoxy)] is a clinical stage drug candidate that was previously developed through phase 2 clinical studies for the treatment of gastric ulcer. This compound was identified as a candidate for T2D using an in vivo phenotypic screen (Saporito et al., 2012a). In the companion article, we show that MLR-1023 is a direct, selective, and potent activator of the nonreceptor Src-related Lyn kinase (Saporito et al., 2012b). In adipocytes, activated Lyn kinase phosphorylates insulin receptor substrate-1 (IRS-1) and increases insulin receptor signaling (Müller et al., 2000, 2005). In addition, insulin receptor activation leads to phosphorylation of Lyn kinase, suggesting a potential feedback regulatory loop between insulin receptor activation and Lyn kinase (Anderwald et al., 2002). In vivo, MLR-1023 administration reduced blood glucose levels in mice subjected to an oral glucose tolerance test (OGTT) without increasing insulin secretion (Saporito et al., 2012b). The known involvement of Lyn kinase in modulating insulin signaling and the effects of MLR-1023 on blood glucose levels suggested that MLR-1023 regulated glucose levels through the modification of insulin-signaling.
The aims of the current studies were to understand the role of MLR-1023 in modulation of insulin-signaling events in vivo and to evaluate it in preclinical models of T2D. Specifically, MLR-1023 was investigated for the requirement of circulating insulin, its effects on insulin receptor sensitivity, and its antidiabetic properties in a clinically translatable model of T2D. Results from these studies demonstrate that MLR-1023 is a unique insulin-sensitizing agent that is clearly distinct from activators of PPARs.
Materials and Methods
MLR-1023 was synthesized according to previously developed synthetic methods (Lipinski et al., 1980) at Advanced Synthesis Group (Newark, DE). Rosiglitazone and metformin were obtained from Sigma-Aldrich (St. Louis, MO). Insulin enzyme-linked immunosorbent assay (ELISA) kits were from Crystal Chem Inc. (Downers Grove, IL). Aviva glucose test strips and monitors were from Roche Diagnostics (Nutley, NJ). The A1CNow Multi-test A1c system was from Bayer Healthcare (Sunnyvale, CA). All other compounds and reagents were obtained from Sigma-Aldrich.
MLR-1023 was formulated in phosphate-buffered saline (PBS; pH 4.5). In these formulations, MLR-1023 was soluble to approximately 5 mg/ml. For acute and chronic studies in mice (OGTT, db/db), MLR-1023 was administered intraperitoneally. For the glucose clamp studies, MLR-1023 was formulated as a suspension in 0.5% Cm-cellulose (0.5%) and Tween 80 (0.025%) in PBS at a concentration of 50 mg/ml and was administered via oral gavage. Rosiglitazone and metformin were formulated in PBS (pH 7.2).
All experiments were conducted in accordance with the National Institutes of Health regulations of animal care covered in Principles of Laboratory Animal Care (National Institutes of Health, 1985) publication 85-23 and were approved by the Institutional Animal Care and Use Committee. CD1-ICR male mice 8 to 12 weeks of age (Ace Animals, Boyertown, PA) were used in studies of baseline glucose, OGTT, and streptozotocin (STZ) studies. Six-week-old (at the start of the study) db/db male mice (BKS.Cg− + Leprdb/+ Leprdb/OlaHsd) were used for chronic studies (Harlan Laboratories, Indianapolis, IN). Mice were given free access to food and water unless otherwise noted. Zuckerfa/fa male rats were acquired from Charles River Laboratories, Inc. (Wilmington, MA) at 6 weeks of age and were used for hyperinsulinemic/euglycemic glucose clamp studies. Animals were kept on a 12:12 h light/dark cycle and were maintained on standard diet. Standard diet was Harlan Chow (2016 Teklad Global 16% Protein Rodent Diet; Harlan Laboratories).
Blood glucose levels were measured using Accu-Chek Aviva Glucometers (Roche Diagnostics) and are reported in milligrams per deciliter. Glucometers were calibrated before each study. Blood (5 μl) was acquired from a tail snip and directly applied to a glucose test strip.
Glucose Tolerance Tests.
For glucose tolerance tests, mice were fasted for 18 h and were treated with drug or vehicle and then a glucose solution (1.5 g/kg) via oral gavage. Blood glucose levels were measured from a drop of blood (5 μl) that was collected from the tail. Blood glucose levels were measured before drug administration, before glucose challenge, and at various times after glucose challenge. MLR-1023, metformin, and vehicle were administered intraperitoneally with a single administration 15 to 30 min before the glucose challenge.
Diabetic mice were generated by administration of STZ over an 8-day period. STZ was formulated in 0.1 M sodium citrate buffer (pH 4.5). STZ was administered at a dose level of 70 mg/kg i.p. three times (days 1, 3, and 8). Mice were excluded from the study if fasting blood glucose levels were below 300 mg/dl on day 10. For determining effects of MLR-1023 on insulin-mediated blood glucose lowering, insulin was administered at a dose of 0.45 U, and MLR-1023 or vehicle was administered 30 min later. Blood glucose levels were measured before insulin treatment (baseline), before MLR-1023 treatment, and then periodically up to 8 h after treatment.
Hyperinsulinemic/Euglycemic Glucose Clamp Studies.
Hyperinsulinemic/euglycemic glucose clamp studies were conducted according to methods described previously (Kramer et al., 2001). In brief, Zuckerfa/fa rats were surgically implanted with two jugular vein catheters. Three days after surgery, rats were subjected to the clamp procedure. To start the clamp study, rats were infused with insulin (1 U/ml in 0.1% bovine serum albumin, 0.9% NaCl; Humulin-N; Eli Lilly & Co, Indianapolis, IN) at a steady rate of 20 mU · kg−1 · min−1. Five minutes after initiating insulin infusion, glucose (50% d-glucose) was infused through the other catheter at a starting rate of 10 μl/min. The glucose infusion rate (GIR) was varied to maintain the rats blood glucose levels at euglycemic levels (100 ± 10 mg/dl). The criteria for steady-state GIR was set as six consecutive blood glucose readings that fell within this euglycemic range. Blood glucose levels were measured using Accu-Chek Aviva Glucometers (Roche Diagnostics) every 5 min beginning 5 min before insulin infusion. MLR-1023, rosiglitazone, and vehicle were administered via oral gavage once per day beginning 7 days before the clamp study. On the day of the study, compounds and vehicle were administered 30 min before the start of the clamp.
Six-week-old db/db mice were treated with MLR-1023, rosiglitazone, or vehicle (once per day intraperitoneally) for 56 days. Mice were weighed three times per week during the course of the study. Glucose levels were measured once or twice per week beginning before the first administration of test compound. After the initiation of treatment, blood glucose levels were measured before that day's compound administration (24 h after the previous administration). For insulin measurements, blood was collected every 2 weeks during the course of the study, beginning before the first administration of vehicle or drug. For HbA1c levels, blood was collected at the end of the study, serum was prepared, and insulin levels were measured by ELISA.
Fasted mice were anesthetized under isoflurane, and their pancreases were dissected free and split into two equal parts. One half of the pancreas was used for determining insulin levels, and the other half was used for insulin immunostaining. For insulin content, insulin was extracted by an acid-ethanol extraction method. Each pancreas was homogenized in ethanol (75%), and HCl (1.5%) supernatant from centrifuged samples was analyzed to determine insulin levels by ELISA.
For insulin immunostaining, each pancreas was paraffin-embedded, sectioned at 4-μm thickness, and stained with rabbit anti-insulin antibody (ab63820; Abcam, Inc., Cambridge, MA). Insulin-containing cells were visualized by a horseradish peroxidase-linked secondary antibody. Sections were counterstained with hematoxylin. Insulin-stained sections were scored on a one to five scale based on the presence and intensity of insulin-stained cells.
Serum Hormone and HbA1c Measurements.
Serum was prepared from blood that was collected by retro-orbital eye bleed. Insulin levels were measured from mouse serum using an Ultra Sensitive Mouse/Rat Insulin ELISA Kit (Crystal Chem Inc.) according to the manufacturer's directions. For pancreatic insulin measurements, samples were normalized to total protein concentration assessed using a Pierce BCA Assay kit (Thermo Fisher Scientific, Waltham, MA). HbA1c levels were measured using an A1CNow Multi-test A1c system (Bayer Healthcare).
All data were analyzed by one- or two-way analysis of variance. A post hoc Dunnett's or Bonferroni test was used to determine statistical differences between treatment groups. For the insulin-staining intensity analysis, data were analyzed by nonparametric Kruskal-Wallis test followed by a post hoc Dunn's test. Data are expressed as the mean ± S.E.M., unless otherwise indicated. Group sizes ranged from four to eight mice per group.
Effects of MLR-1023 Administration in an OGTT.
MLR-1023 was tested at doses of 10 to 100 mg/kg i.p. for its effects on glucose levels in an OGTT. In this study, oral glucose administration to fasted, vehicle-treated mice evoked a 3.5-fold increase in blood glucose levels (70–248 mg/dl) with a peak increase observed 45 min after glucose challenge (Fig. 1A). A single administration of MLR-1023 elicited a reduction in blood glucose levels, with a maximal response produced by a dose level of 30 mg/kg. This dose reduced the peak blood glucose levels to 137 mg/dl. In another study, MLR-1023 (30 mg/kg) was compared with the maximal effective dose of metformin (150 mg/kg). The degree of reduction in blood glucose after administration of MLR-1023 was equivalent to that produced by metformin (Fig. 1B). Metformin elicited a small reduction in glucose before oral glucose administration that was not observed with MLR-1023 administration.
Effects of MLR-1023 Administration on Blood Glucose Levels in Fed and Fasted Mice.
MLR-1023 was tested for its effects on blood glucose levels in mice subjected to increasing fasting times to determine whether MLR-1023 produced hypoglycemia (Fig. 2). Increasing the fasting duration from 0 to 6 and 18 h reduced baseline blood glucose levels from 161 ± 8 to 123 ± 7 and 62 ± 3 mg/dl, respectively. Ninety minutes after an acute administration of MLR-1023 (30 mg/kg) to freely fed mice, blood glucose levels were significantly reduced from 161 ± 8 to 133 ± 8 mg/dl (Fig. 2A). With fasting times of 6 (Fig. 2B) or 18 h (Fig. 2C), MLR-1023 administration did not significantly affect blood glucose levels. In contrast, metformin (150 mg/kg) significantly reduced blood glucose levels regardless of fasting times.
Effects of MLR-1023 in Combination with Insulin in STZ Insulin-Depleted Mice.
MLR-1023 was tested for blood glucose lowering in STZ-treated mice (Fig. 3). Baseline blood glucose levels were 530 ± 8 mg/dl. Serum insulin levels at the time of the study were below the level of detectability (<0.5 ng/ml; data not shown). MLR-1023 administration (100 mg/kg i.p.) did not affect blood glucose levels up to 6.5 h after administration (Fig. 3A). Administration of insulin alone (0.45 U) reduced blood glucose levels to 199 ± 29 mg/dl with a Tmax of 60 min (Fig. 3B). Blood glucose levels returned to starting levels by 90 min. When MLR-1023 was administered 30 min after insulin administration, there was a significant potentiation of the insulin-mediated blood glucose-lowering response. Maximal reductions in blood glucose levels were increased to 147 ± 15 mg/dl. In addition, the Tmax was extended to 90 min. The most significant effect of MLR-1023 administration with insulin versus insulin alone was in the extension of time for blood glucose reverting to starting levels (240 versus 90 min, respectively). MLR-1023 elicited a dose-dependent potentiation of the insulin response, with a maximal effect observed with a dose level of 30 mg/kg (see Supplemental Fig. 1).
Effects of MLR-1023 in a Hyperinsulinemic/Euglycemic Clamp Model.
In the glucose clamp study, Zucker rats were administered vehicle, rosiglitazone, or MLR-1023 daily for 7 days before the clamp study. All compound administrations were via oral gavage. Baseline blood glucose levels (measured 5 min before insulin infusion) in vehicle-, MLR-1023-, or rosiglitazone-treated rats were 167 ± 6, 118 ± 6, and 108 ± 12 mg/dl, respectively (Fig. 4A). These reductions in MLR-1023 and rosiglitazone treatments, compared with vehicle, were significant (p < 0.05) and consistent with the effects of these compounds on baseline blood glucose levels in other diabetic models.
Rats were clamped at 100 ± 10 mg/dl by 50 min after the start of infusion of insulin and glucose. The GIR required to maintain rats at this blood glucose level was 7.4 ± 1.0 mg · kg−1 · min−1 in the vehicle group (Fig. 4B). Compared with vehicle controls, rosiglitazone administration increased the GIR to 34.6 ± 1.3 mg · kg−1 · min−1 (4.7-fold; p < 0.001 compared with vehicle-treated). This rosiglitazone-mediated increase in glucose uptake rate was consistent with previous studies (Kramer et al., 2001). MLR-1023 administration increased the GIR to 33.4 ± 1.8 mg · kg−1 · min−1 (4.5-fold compared with vehicle; p < 0.001 compared with vehicle).
Effects of MLR-1023 in db/db Mice.
MLR-1023 was tested in db/db mice with acute and chronic administration. Three separate MLR-1023 db/db studies were conducted. The first study measured the effects after a single administration of MLR-1023. In this study, MLR-1023 administration elicited a dose-dependent reduction in blood glucose levels. At dose levels of 15 and 50 mg/kg, MLR-1023 reduced blood glucose from 244 ± 12 to 188 ± 19 and 165 ± 17 mg/dl, respectively (Fig. 5).
In the second db/db study, chronically administered MLR-1023 was tested in a dose-response paradigm with measurement of multiple, clinically translatable T2D parameters (Fig. 6). In this study, MLR-1023 was administered daily with blood glucose measurements every 14 days. These blood glucose levels were determined after a 6-h fast and 24 h after the last administration of compound. Vehicle-treated db/db mice exhibited a progressive increase in blood glucose levels from 122 at 6 weeks of age to 543 mg/dl at 14 weeks of age. MLR-1023 administration elicited a dose-dependent reduction in blood glucose levels compared with the vehicle-treated db/db mice (Fig. 6A) with significant (p < 0.01) blood glucose lowering detected at dose levels of 30 and 100 mg · kg−1 · day−1. Significant effects were detected at both doses at each glucose measurement time point. Administration of MLR-1023 did not affect weight gain in these mice (Fig. 6B).
Consistent with previously reported studies with db/db mice (Baribault, 2010), serum insulin levels in the vehicle-treated mice showed an early peak (14 days after the first dose) followed by a progressive decline over time (Fig. 6C). MLR-1023 administration at a dose level of 30 mg/kg slowed and at a dose level of 100 mg/kg blocked the decline in serum insulin. The effects at both dose levels were statistically significant (p < 0.05).
MLR-1023 administration produced dose-dependent reductions in HbA1c levels as measured at the end of the study (Fig. 6D). Statistically significant (p < 0.01) reductions of 1.6 and 2.7% were achieved at dose levels of 30 and 100 mg/kg, respectively.
Effects of MLR-1023 on Insulin-Secreting Pancreatic β-Cells in db/db Mice.
Pancreatic insulin-staining intensity and pancreatic insulin levels were measured at the end of the chronic db/db study in an effort to assess pancreatic β-cell function. Pancreatic β-cells were visualized by immunohistochemical staining with an insulin antibody. Figure 7A shows representative photomicrographs of insulin-stained pancreatic tissue from vehicle-treated, MLR-1023-treated, and db/lean control mice. Tissues from MLR-1023-treated mice exhibited a higher staining intensity than those from vehicle-treated mice (Fig. 7B). This effect was dose-dependent and paralleled the MLR-1023 dose responses for glucose lowering and serum insulin levels, with a dose of 30 mg/kg eliciting a clearly discernable change in insulin-staining intensity. Tissues from animals treated with dose levels of 30 and 100 mg/kg showed comparable staining intensity to that seen in lean/db mice.
Pancreatic insulin levels in vehicle-treated db/db mice were significantly less (p < 0.05) than those found in db/lean mice (Fig. 7C). This decreased level is consistent with a loss of pancreatic β-cell function in db/db mice of this age. Pancreatic insulin-staining intensity and insulin levels from MLR-1023-treated mice (30 and 100 mg/kg) were significantly higher than those in vehicle-treated mice.
Comparison of MLR-1023 to Rosiglitazone in db/db Mice.
A subsequent db/db study was conducted to compare the efficacy of MLR-1023 with that of rosiglitazone (Fig. 8). In this study, vehicle, MLR-1023 (30 mg/kg), and rosiglitazone (10 mg/kg) were administered daily, and blood glucose levels were measured once per week beginning 3 days after the initial administration. Blood glucose levels were measured 24 h after administration of the test agent. Under these conditions, MLR-1023 produced a statistically significant (p < 0.01) decrease in blood glucose levels that was apparent at the first glucose measurement (Fig. 8B). This glucose lowering was sustained for the study duration. Blood glucose levels in the MLR-1023-treated db/db mice did not differ from those of untreated db/lean mice at any time after the initial administration. The magnitude of MLR-1023-mediated blood glucose lowering was equivalent to that of rosiglitazone. However, rosiglitazone treatment did not significantly lower blood glucose levels until 10 days after the initial administration. This delayed rosiglitazone-mediated effect is consistent with reports of delayed onset of rosiglitazone-mediated blood glucose lowering in db/db mice (Zhang et al., 1996; Connor et al., 1997). Rosiglitazone administration produced a characteristic acceleration of weight gain that was first apparent after 10 days of dosing and was sustained for the duration of the study. MLR-1023 administration did not accelerate weight gain in these db/db mice (Fig. 8C).
The results from the current studies show that MLR-1023 is an orally active insulin-sensitizing agent that produces a rapid-onset and durable decrease in blood glucose levels in diabetic animals. Unlike TZDs, MLR-1023 did not activate PPARs (Saporito et al., 2012b), produced a rapid onset of action, and did not cause weight gain. With chronic administration, MLR-1023 administration produced a sustainable decrease in blood glucose levels and decreased HbA1c levels and prevented the loss of pancreatic β-cells. These results, coupled with data from the companion article (Saporito et al., 2012b), show that MLR-1023 represents a novel class of insulin sensitizer distinct from the TZDs. These results indicate that MLR-1023 is a potential therapeutic for the treatment of T2D.
In the current studies, acute MLR-1023 administration produced a dose-dependent blood glucose lowering in mice subjected to an OGTT. The blood glucose-lowering effect of MLR-1023 was similar in magnitude to that of metformin, but with a different pharmacological profile. For example, both MLR-1023 and metformin significantly reduced blood glucose levels in fed euglycemic mice; however, unlike metformin, MLR-1023 did not reduce blood glucose levels in fasted mice. This difference indicated that MLR-1023 and metformin acted by distinct mechanisms. Metformin is a gluconeogenesis inhibitor that lowers blood glucose levels in an insulin-independent manner (Hundal et al., 2000; Shaw et al., 2005). The absence of MLR-1023 activity in fasted mice, which are shifted into a highly gluconeogenic state, suggested that MLR-1023 was likely not to be a gluconeogenesis inhibitor (Moore et al., 1998). In addition, fasted mice have very low levels of circulating insulin, consistent with the notion that insulin is required for MLR-1023-mediated blood glucose lowering (Zhang et al., 2012).
To investigate the hypothesis that insulin was required for MLR-1023 activity, MLR-1023 was tested in mice that were pretreated with STZ. STZ destroys pancreatic β-cells and depletes animals of circulating insulin (Kolb, 1987). In these insulin-depleted hyperglycemic mice, MLR-1023 administration did not lower blood glucose levels. However, in insulin-depleted mice treated with exogenous insulin, MLR-1023 administration potentiated and prolonged the effects of insulin. These results clearly demonstrated that circulating insulin was an absolute requirement for MLR-1023-mediated blood glucose-lowering activity in vivo. The insulin requirement for MLR-1023-mediated blood glucose-lowering activity indicated that MLR-1023 did not act as an insulin mimetic. In addition, these data indicated that MLR-1023 was affecting insulin receptor responsiveness, insulin signaling, or insulin-mediated glucose utilization.
To further test the relationship between MLR-1023 and insulin responsiveness, MLR-1023 was evaluated in a hyperinsulinemic/euglycemic glucose clamp model in diabetic Zucker rats. This clamp method is a technique for evaluating the ability of a compound to affect insulin receptor responsiveness in experimental animals and in the clinic (Bloomgarden, 2006). The GIR required to maintain euglycemia in hyperinsulinemic animals directly correlates to tissue glucose uptake, which in turn directly correlates to insulin receptor responsiveness. In the current studies, oral administration of MLR-1023 increased the GIR in clamped animals to a degree equivalent to that of rosiglitazone. These data demonstrated that MLR-1023 administration increased insulin sensitivity in vivo. MLR-1023 is not a PPAR agonist (Saporito et al., 2012b), indicating that the increase in insulin receptor sensitization occurs through a unique mechanism.
MLR-1023 is a direct, selective, and potent Lyn kinase activator, and the presence of Lyn kinase is required for MLR-1023-mediated blood glucose lowering (Saporito et al., 2012b). In adipocytes, Lyn kinase directly phosphorylates IRS-1 and promotes Glut-4 translocation (Müller et al., 2000, 2005). The known interaction of Lyn kinase with IRS-1 suggests that MLR-1023-mediated increases in insulin pathway sensitization occur immediately downstream from the insulin receptor. Data presented here and in the companion article are consistent with a model in which MLR-1023 activates Lyn kinase, which in turn promotes the tyrosine phosphorylation of IRS-1. This tyrosine phosphorylation of IRS-1 may amplify and prolong the insulin-signaling response. Further studies are required to define the mechanistic processes surrounding MLR-1023-mediated insulin-signaling sensitization.
The anti-diabetic activity of MLR-1023 was studied extensively in db/db mice. The db/db mouse model exhibits many features of clinical stage T2D including a progressive decrease in insulin receptor sensitivity, steadily increasing blood glucose levels and HbA1c levels, compensatory elevations in serum insulin followed by a degeneration of insulin-secreting pancreatic β-cells, and a concomitant drop in serum insulin levels (Baribault, 2010). These mice are responsive to a broad range of clinically approved pharmacological agents, and the blood glucose-lowering effects that these agents evoke in these mice are translatable to T2D patients. For example, metformin (an inhibitor of gluconeogenesis), rosiglitazone (an activator of PPARγ), and exendin-4 (a glucagon-like peptide 1 receptor agonist) elicit dramatic improvements in metabolic function, including sustained reductions in blood glucose (Connor et al., 1997; Young et al., 1999; Fujita et al., 2005). In the current studies, chronic administration of MLR-1023 to db/db mice elicited a dose-dependent, durable blood glucose-lowering response when administration was initiated at an early stage of diabetic development and at a stage when pancreatic β-cells were functional. The effective dose range in the db/db study was similar to that in the single-administration OGTT studies. However, in the db/db studies, blood glucose levels were measured 24 h after the previous administration of MLR-1023. This long-lasting glycemic control was seen at the first time point (3 days) and was maintained over the course of the study (8 weeks). MLR-1023 also reduced HbA1c levels by 2.7%. The decreased HbA1c levels paralleled the decreased blood glucose levels and were consistent with durable MLR-1023-mediated glycemic control. In addition, the long-lasting blood glucose-lowering effects indicate that MLR-1023 produces a durable insulin sensitization.
Insulin levels in MLR-1023-treated mice were sustained at an elevated level during the course of the study. Because MLR-1023 did not increase insulin secretion (Saporito et al., 2012b), these data suggested that it preserved the function of insulin-secreting pancreatic β-cells. In fact, pancreatic insulin levels and insulin-staining intensity in β-cells from MLR-1023-treated mice were significantly higher than those found in vehicle-treated mice. These data indicate that chronically administered MLR-1023 preserved the levels of insulin in pancreatic β-cells and that this preservation led to sustained levels of circulating insulin in db/db mice. This preservation may be due to a direct protective activity of MLR-1023 on these cells, or the sustained blood glucose-lowering effect and subsequent reduced stress on these insulin-secreting cells may have slowed and blocked degeneration of these cells. Similar β-cell preservation effects have been observed with sitagliptin, glucagon-like peptide 1, and TZDs in diabetic animals (Stoffers et al., 2000; Kawasaki et al., 2005; Mu et al., 2006, 2009; DeFronzo and Abdul-Ghani, 2011b). In addition, T2D can be prevented when administration of these anti-diabetic compounds is initiated at a prediabetic stage, suggesting a relationship between β-cell preservation in db/db mice and delay of disease onset in the clinic (DeFronzo and Abdul-Ghani, 2011a).
Compared with rosiglitazone, MLR-1023 produced an equivalent, long-lasting glycemic control response. However, the pharmacological profile of MLR-1023 differed from that of rosiglitazone in two ways. First, MLR-1023 elicited a blood glucose-lowering response with a faster onset than rosiglitazone. In the acute study, a single administration of MLR-1023 lowered blood glucose within 3 h. In the chronic study, the onset of MLR-1023-mediated glucose lowering was significantly faster than the onset produced by rosiglitazone. The acute actions of MLR-1023 suggest that MLR-1023 regulated blood glucose levels by mechanisms that were independent of the transcriptional events required for rosiglitazone activity. Second, in contrast to rosiglitazone and other PPAR agonists, MLR-1023 did not produce an acceleration of weight gain (Harrington et al., 2007). The absence of an MLR-1023-mediated increase in weight gain is consistent with MLR-1023 not affecting PPARs. Moreover, this PPAR-mediated weight gain effect seen in rodents translates to the clinic and represents a serious liability of that drug class. This lack-of-weight-gain feature for MLR-1023 represents an important differentiating characteristic from all other previously described insulin sensitizers (TZDs), insulin secretagogs, and insulin-like drugs (Palmer et al., 2004; Harrington et al., 2007; Phung et al., 2010).
As a repositioned drug candidate, there is good evidence indicating that MLR-1023 will be a safe and well tolerated drug. The existing preclinical and clinical data packages generated when the drug was previously developed show that the compound has a favorable chronic safety and tolerability profile (Saporito et al., 2012a). In addition, the function of Lyn kinase, as elucidated in genetically modified rodents, further supports this view that chronic activation of this kinase will be well tolerated (Harder et al., 2001). Most importantly, given the U.S. Food and Drug Administration's increased scrutiny on cardiovascular safety for new diabetic candidates (Diabetes Mellitus—Evaluating Cardiovascular Risk in New Antidiabetic Therapies to Treat Type 2 Diabetes, http://www.fda.gov/downloads/Drugs/Guidance ComplianceRegulatoryInformation/Guidances/ucm071627.pdf, December 2008), there is no evidence from prior preclinical toxicology studies, clinical toxicology data, or inherent in the mechanism that suggest any potential cardiovascular liability with this drug candidate.
In summary, MLR-1023 is a mechanistically novel insulin-sensitizing agent. The results from these studies show that MLR-1023 achieves the glycemic control activity of existing insulin sensitizers but lacks the hallmark liabilities of that class. For these reasons MLR-1023 is a compelling therapeutic candidate for the treatment of T2D.
Participated in research design: Ochman, Lipinski, Handler, Reaume, and Saporito.
Conducted experiments: Ochman and Saporito.
Contributed new reagents or analytic tools: Lipinski.
Performed data analysis: Ochman, Reaume, and Saporito.
Wrote or contributed to the writing of the manuscript: Reaume and Saporito.
We thank John Amatruda, Perry Molinoff, and John Ciallella for helpful advice and comments on the manuscript.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- type 2 diabetes
- peroxisome proliferator-activated receptor
- 2(1H)-pyrimidinone, 5-(3-methylphenoxy)
- insulin receptor substrate-1
- oral glucose tolerance test
- enzyme-linked immunosorbent assay
- phosphate-buffered saline
- glucose infusion rate
- area under the curve
- dipeptidyl peptidase-4
- glucagon-like peptide
- Received January 18, 2012.
- Accepted March 16, 2012.
- Copyright © 2012 by The American Society for Pharmacology and Experimental Therapeutics