Research Article: Tissue-Specific Analysis of Glycogen Synthase Kinase-3α (GSK-3α) in Glucose Metabolism: Effect of Strain Variation

Date Published: January 6, 2011

Publisher: Public Library of Science

Author(s): Satish Patel, Katrina Macaulay, James R. Woodgett, Jose A. L. Calbet.

Abstract: Over-activity and elevated expression of glycogen synthase kinase-3 (GSK-3) has been implicated in the etiology of insulin resistance and Type 2 diabetes. Administration of specific GSK-3 inhibitors to diabetic or obese rodent models improves glycaemic control and insulin sensitivity. However, due to the indiscriminatory nature of these inhibitors, the relative contribution of the two isoforms of GSK-3 (GSK-3α and GSK-3β) is not known. Recently, we demonstrated that an out-bred strain of mice (ICR) lacking expression of GSK-3α in all tissues displayed improved insulin sensitivity and enhanced hepatic glucose metabolism. We also found that muscle (but not liver) inactivation of GSK-3β conferred insulin and glucose sensitization in an in-bred strain of mice (C57BL/6).

Here, we have employed tissue-specific deletion of GSK-3α, to examine the relative contribution of two insulin-sensitive tissues, muscle and liver, towards the insulin sensitization phenotype originally observed in the global GSK-3α KO animals. We found that mice in which GSK-3α has been inactivated in either skeletal-muscle or liver displayed no differences in glucose tolerance or insulin sensitivity compared to wild type littermates. Given the strain differences in our original analyses, we examined the insulin and glucose sensitivity of global GSK-3α KO animals bred onto a C57BL/6 background. These animals also revealed no significant differences in glucose metabolism/insulin sensitivity compared to their wild type littermates. Furthermore, deletion of hepatic GSK-3α on the out-bred, ICR background failed to reproduce the insulin sensitivity manifested by the global deletion of this isoform.

From these data we conclude that the improved insulin sensitivity and hepatic glucose homeostasis phenotype observed upon global inactivation of GSK-3α is strain-specific. We surmise that the insulin-sensitization observed in the out-bred strain of mice lacking GSK-3α is mediated by indirect means that do not require intrinsic function of GSK-3α in skeletal muscle and liver tissues.

Partial Text: Glycogen synthase kinase-3 (GSK-3) is a ubiquitously expressed serine/threonine protein kinase that is encoded by two distinct genes, GSK-3α (52 kDa) and GSK-3β (47 kDa). These two isoforms are highly conserved and share ∼98% sequence similarity in their catalytic domains [1]. GSK-3 is a constitutively active kinase in resting cells that becomes rapidly inactivated by phosphorylation at Ser 21 (GSK-3α) and Ser 9 (GSK-3β) in response to insulin through a phosphatidylinositol 3 (PI-3) kinase/protein kinase B (PKB, also termed Akt)-dependent manner.

Conditional GSK-3α mice that express a modified allele of GSK-3α, where exon 2 is flanked by LoxP sequences, have been previously described [11]. Skeletal muscle-specific GSK-3α knockout (KO) animals were generated by breeding GSK-3α floxed mice (flx/flx –C57BL/6/129 background) with mice expressing Cre under the control of the myosin light chain 1f (MLC1f) promoter (MLC Cre- C57BL/6B6/129 background), a gene that is predominantly expressed in fast twitch skeletal muscle fibres. The resultant skeletal muscle GSK-3α KO mice (MLC Cre +) are viable, fertile and born to the expected Mendelian frequency. As expected, GSK-3α expression is completely lost in quadricep (quad), gastrocnemius (gastroc) and extensor digitorum longus (EDL) muscles, whereas there is only partial reduction of GSK-3α expression in soleus, which is primarily composed of slow twitch fibres (Figure 1A). Importantly in the GSK-3α flx/flx MLC Cre + (MLC Cre +) mice, loss of GSK-3α expression is restricted to skeletal muscle tissue only and remains unaltered in brain, heart, liver and testes of mice when compared to GSK-3α flx/flx MLC Cre – (MLC Cre -) (Figure 1A). Weekly weight analysis from age 4–22 weeks revealed no significant differences in whole body weight between MLC Cre + and MLC Cre – littermate control animals (Figure 1B).

Previous data have demonstrated a beneficial role of acute GSK-3 inhibition in the treatment of diabetes and obesity. Treatment of diabetic human skeletal muscle cultures with GSK-3 inhibitors stimulates glycogen synthase activity and potentiates insulin-stimulated glucose transport and glycogen accumulation [3]. Moreover, administration of these compounds to diabetic rodents improves glucose tolerance by increasing hepatic glycogen synthesis and coordinately reducing glucose output by inhibiting gluconeogenesis [19]. Whilst GSK-3 inhibitors are unable to discriminate between GSK-3α and GSK-3β, we and others have shown that there are indeed isoform- and tissue-specific roles for GSK-3 in the regulation of glucose metabolism. Selective loss of GSK-3β in skeletal muscle improves glucose tolerance and insulin signaling, while removal of this isoform in β-islet cells can reduce hyperglycaemia in diabetic IRS-2 KO mice [12], [13]. Furthermore, genetic ablation of GSK-3α results in improved whole-body glucose tolerance that is accompanied with enhanced hepatic insulin signaling and elevated glycogen accumulation [11]. These observations led us to investigate which tissues are responsible for the phenotype observed in the GSK-3α global KO mouse. In the present study we found that mice harbouring specific deletion of GSK-3α in skeletal muscle or liver displayed normal sensitivity to insulin and normal insulin signaling. This unexpected result led us to initially infer that removal of hepatic GSK-3α is insufficient to reproduce the whole body improvement of glucose metabolism that is observed with the global GSK-3α KO mouse. However, we noted that the characterization of the global GSK-3α KO had been performed on the out-bred strain, ICR but our tissue-specific knockout models were on the C57BL/6 in-bred strain (with some 129 strain content from the deletor strains). To test for possible effects of genetic background, we back-crossed the global GSK-3α KO onto the C57BL/6/129 background. Unlike the ICR strain, we observed that global inactivation of GSK-3α on the C57BL/6/129 background did not result in improved glucose/insulin tolerance or enhanced hepatic glycogen accumulation (Figure 7) or insulin sensitivity (data not shown). This result indicated strain-specific effects of genetic ablation of GSK-3α on glucose metabolism and led us to generate tissue-specific deletions of GSK-3α on the ICR background. We speculated that hepatic deletion of GSK-3α on the ICR background would reproduce the liver-specific phenotype of the ICR-GSK-3α global KO. However, this was not found to be the case, as the ICR-GSK-3α Alb Cre + mice displayed relatively normal insulin sensitivity. These experiments demonstrate that the role of GSK-3α on glucose metabolism is complex and is sensitive to genetic background (through as yet unidentified modifiers) and depends on tissues other than, or in addition to, skeletal muscle and liver. There are numerous examples demonstrating that background strain as well as environmental factors can affect the metabolic phenotypes seen in transgenic mouse models. For instance, mice that lack IRS-2 exhibit mild to severe diabetes dependent on the background strain, whereas mice lacking the muscle glycogen subunit of protein phosphatase 1 (GM) can result in obesity and insulin resistance in a 129/Ola donor strain, but remain lean and glucose tolerant in a 129/SvJ background. In the present case, we find that GSK-3α-induced improvement of whole-body glucose tolerance and hepatic insulin sensitivity is apparent only within the out-bred ICR strain and not within the in-bred C57BL/6 strain. It has been reported that compared to the C57BL/6, the ICR strain is more susceptible to experimental forms of insulin resistance [20]. We speculate that the beneficial effect of GSK-3α inhibition is associated with the genetic/environmental susceptibility of insulin resistance within these rodent strains. Given the beneficial effects of GSK-3 inhibitors in obese and insulin resistant rodents [4], [6], [7], it would be interesting to determine the effects of GSK-3α KO on these models.