Small molecule glucokinase activators disturb lipid homeostasis and induce fatty liver in rodents: a warning for therapeutic applications in humans

Abstract

Background and purpose: Small-molecule glucokinase activators (GKAs) are currently being investigated as therapeutic options for the treatment of type 2 diabetes (T2D). Because liver overexpression of glucokinase is thought to be associated with altered lipid profiles, this study aimed at assessing the potential lipogenic risks linked to oral GKA administration.

Experimental approach: Nine GKA candidates were qualified for their ability to activate recombinant glucokinase and to stimulate glycogen synthesis in rat hepatocytes and insulin secretion in rat INS-1E cells. In vivo activity was monitored by plasma glucose and HbA1c measurements after oral administration in rodents. Risk-associated effects were assessed by measuring hepatic and plasma triglycerides and free fatty acids, as well as plasma aminotransferases, and alkaline phosphatase.

Key results: GKAs, while efficiently decreasing glycaemia in acute conditions and HbA1c levels after chronic administration in hyperglycemic db/db mice, were potent inducers of hepatic steatosis. This adverse outcome appeared as soon as 4 days after daily oral administration at pharmacological doses and was not transient. GKA treatment similarly increased hepatic triglycerides in diabetic and normoglycaemic rats, together with a pattern of metabolic phenotypes including different combinations of increased plasma triglycerides, free fatty acids, alanine and aspartyl aminotransferases, and alkaline phosphatase. GKAs belonging to three distinct structural families induced hepatic steatosis in db/db mice, arguing in favour of a target-mediated, rather than a chemical class-mediated, effect.

Conclusion and implications: Given the risks associated with fatty liver disease in the general population and furthermore in patients with T2D, these findings represent a serious warning for the use of GKAs in humans.

Linked article: This article is commented on by Rees and Gloyn, pp. 335-338 of this issue. To view this commentary visit http://dx.doi.org/10.1111/j.1476-5381.2012.02201.x.

Figure 1

Chemical structure and pharmacokinetic properties of the nine compounds used throughout the study. (A and B) General structure of the N-pyridinyl-benzamide scaffold and of the seven synthesized N-pyridinyl-benzamide derivatives. (C) Chemical structure of piragliatin. (D) Chemical structure of N00268746. (E) Enzymatic and pharmacokinetic properties of the nine compounds. Enzyme assays were performed using human recombinant GK. NC, not calculable, no plateau at 10 μM and high compound absorbance at 100 μM. IN, inactive. Fabs, prediction of the intestinal absorption. Clint, in vitro intrinsic clearances using mouse and rat hepatic microsomes. MF, prediction of metabolic availability based on in vitro metabolic stability measurements with hepatic microsomes. Values are means of at least three experiments for enzymatic properties on GK, means of duplicate wells for Fabs and the disappearance in the presence of microsomes at five separate time points was performed with one measurement per time for Clint and MF.

Figure 2

In vitro and in vivo pharmacological profile of the nine compounds used throughout the study. (A) D-[U-14C] glucose incorporation reflecting glycogen synthesis in primary rat hepatocytes after 3 h of treatment with 10 μM GKA22 (reference compound). Data are means ± SEM of 30 independent experiments performed in quadruplicate wells. (B) Glycogen synthesis in primary rat hepatocytes after 3 h of treatment with 10 μM compounds, relatively to GKA22. Data are means ± SEM of three to eight independent experiments performed in quadruplicate wells. Significance for each compound was determined versus the N00234379 inactive compound by Student's t-test. (C) Insulin secretion in rat INS1-E cells after 2 h of treatment with 10 μM GKA22 used as a reference compound. Data are means ± SEM of 18 independent experiments performed in quadruplicate wells. (D) Insulin secretion in rat INS1-E cells after treatment with 10 μM compounds for 2 h, relative to GKA22. Data are means ± SEM of three to five independent experiments performed in quadruplicate wells. Significance for each compound was determined versus the N00234379 inactive compound by Student's t-test. (E) Anti-hyperglycaemic action of the nine compounds after acute oral administration in db/db mice. Data are means ± SEM of 8 to 10 animals. Significance was determined by post hoc analysis of dose effect at fixed time levels with Dunnett's test. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 3

Chemically unrelated GKAs increase the hepatic TG content after 4 days of daily oral treatment in db/db mice. (A–D) Hepatic TG content after 4 days. Each panel represents an independent experiment with GKA50 being used as positive control. All compounds were administered once per day at 130 μmol kg⁻¹. Data are means ± SEM of 10 animals for each experiment. Significance was determined by Student's t-test. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4

GKA50 and N00236460 decrease HbA1c levels and induce hepatic steatosis after chronic oral administration in db/db mice. (A–F) 6 week daily treatment with GKA50, N00236460 or metformin. The doses expressed as mg kg⁻¹ correspond to equimolar concentrations for GKA50 and N00236460 (80 and 130 μmol kg⁻¹ for the low dose and high dose groups respectively). (A) Differences in plasma HbA1c levels between day 1 and day 42. (B) Liver weights relative to body weight after 6 weeks. (C) Hepatic TG content after 6 weeks. (D) Plasma TG levels after 6 weeks. (E) Alanine aminotransferase levels and (F) aspartyl aminotransferase levels after 6 weeks. Data are means ± SEM of 11 to 12 animals. Significance was determined by anova and Dunnett's multiple comparison test to measure the dose-dependent effect of GKA50 and N00236460, and by Student's t-test for metformin. (G) Time-related effect of chronic oral treatment with N00236460 on the hepatic TG content in db/db mice. Independent experiments were performed for each time point of the treatment. Data are means ± SEM of 15 animals. Significance was determined by Student's t-test. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5

A 2 week daily oral treatment with GKA50 and N00236460 induces hepatic steatosis in db/db mice mainly by increasing trioleate and increases hepatic oleic acid levels without any substantial effect on other FFAs. On the day of killing, groups of 20 mice receiving similar treatment during the study were separated into two groups, with 10 mice fasted for 4 h and 10 mice fed ad lib. The doses expressed as mg kg⁻¹ correspond to equimolar concentrations (130 μmol kg⁻¹) for GKA50 and N00236460. (A) Differences in plasma HbA1c levels between day 1 and day 14. (B) Liver weights relative to body weight after 2 weeks. (C) Hepatic TG content measured after KOH/ethanol extraction after 2 weeks. (D–F) Detailed composition of hepatic TGs after chloroform/methanol extraction and analysis by UPLC-ESI-MS after 2 weeks. (D) Hepatic trioleate levels. (E) Hepatic tripalmitate levels. (F) Hepatic trilinoleate levels. (G) Plasma TG levels after 2 weeks. (H) Detailed composition of hepatic FFAs after chloroform/methanol extraction and analysis by UPLC-ESI-MS after 2 weeks. Data are means ± SEM of 10 animals. Significance was determined by Student's t test. *P < 0.05, **P < 0.01, ***P < 0.001 for comparison between GKA50- and N00236460-treated groups and control groups. ^§P < 0.05, ^§§P < 0.01, ^§§§P < 0.001 for comparison between control fed and fasted groups.

Figure 6

Hepatic steatosis and disturbed plasma lipid homeostasis after 4 weeks of daily oral treatment with GKA50 and N00236460 in ZDF rats and normoglycaemic Wistar rats. The doses expressed as mg kg⁻¹ correspond to equimolar concentrations for GKA50 and N00236460 (30 and 80 μmol kg⁻¹ for the low-dose and high-dose groups respectively). (A–G) ZDF rats, (H–N) Wistar rats. (A, H) Differences in plasma HbA1c levels between day 1 and day 28. (B,I) Liver weights relative to body weight after 4 weeks. (C, J) Hepatic TG content after 4 weeks. (D, K) Plasma TG levels after 4 weeks. (E,I) Plasma FFAs levels after 4 weeks. (F, M) Alanine aminotransferase levels after 4 weeks. (G, N) Aspartyl aminotransferase levels after 4 weeks. Data are means ± SEM of 10 animals. Significance was determined by anova and Dunnett's multiple comparison test to measure the dose-dependent effect of GKA50 and N00236460. *P < 0.05, **P < 0.01.

Figure 7

Plasma HbA1c variations versus hepatic TGs after chronic administration of GKAs in rodent models. Individual data obtained from chronic studies in (A) db/db mice, (B) ZDF rats and (C) Wistar rats were drawn in the original graphs. For a clearer visualization of results, areas encompassing all individual data obtained in a same group were drawn as shown in the inset of (B). Coloured dots within each area represent the means of HbA1c variations and hepatic TGs.