Tissue glycogen content and glucose intolerance

Insulin stimulates glycogen synthesis in the the liver and skeletal muscle. After a mixed meal, the secretion of insulin from pancreatic β cells thus results in about 20% and 30% of the carbohydrate intake being stored in the form of glycogen in the liver and skeletal muscle, respectively (1, 2). Defects in this process can therefore be a major contributor to postprandial hyperglycemia. Indeed, the glycogen contents of the liver and skeletal muscle are reduced in individuals with type 2 diabetes (3, 4).

Glycogen metabolism is controlled predominantly by the coordinated action of two enzymes, glycogen synthase and glycogen phosphorylase, both of which are regulated by phosphorylation and allosteric modulators. Insulin promotes the net dephosphorylation of both glycogen synthase and glycogen phosphorylase through the inhibition of protein kinases and the activation of protein phosphatases. Among the protein kinases, glycogen synthase kinase–3 (GSK-3) is thought to be an important target for insulin in its stimulation of glycogen synthase activity (5, 6). Among the protein phosphatases, protein phosphatase 1 (PP1) has been implicated in this action of insulin (6).

PP1 is an abundant protein serine-threonine phosphatase that is expressed in all compartments of eukaryotic cells. The catalytic subunit of PP1 thus interacts with a wide variety of targeting subunits that localize it to specific sites within the cell. A family of proteins that target PP1 to glycogen and thereby regulate its activity has been identified. These proteins include G_M (PPP1R3), G_L (PPP1R4), PTG (protein targeting to glycogen or PPP1R5), and PPP1R6.

Deletion of glycogen-targeting subunits of PP1 (PTG and G_M) in mice

PTG was cloned as a binding protein of the catalytic subunit of PP1 by Saltiel and coworkers in 1997 (7). This protein localizes PP1 to glycogen and also binds to glycogen synthase, glycogen phosphorylase, and phosphorylase kinase within cells. Overexpression of PTG resulted in a marked increase in both basal and insulin-stimulated glycogen synthesis in Chinese hamster ovary cells expressing insulin receptors. In this issue of the JCI, Saltiel and colleagues now describe the generation of mice in which the PTG gene has been deleted (8). Whereas homozygous deletion of the PTG gene resulted in embryonic death, mice missing only one copy of the gene were viable but found to have a reduced glycogen content in adipose tissue, the liver, the heart, and skeletal muscle. Although young (one to two months of age) PTG hemizygous mice exhibited normal glucose tolerance, the animals developed glucose intolerance and insulin resistance (three to four months) and a consequent increase in muscle triglyceride content (18 months) with age.

G_M, also known as PPP1R3 or R_GL, was the first glycogen-binding subunit of PP1 identified and is expressed exclusively in skeletal and cardiac muscle. Homozygous deletion of the G_M gene was shown not to result in any obvious defect (9). The homozygous mutant animals manifested normal glucose tolerance and insulin sensitivity at 12 to 24 weeks of age and a weight gain similar to that of their wild-type littermates at up to 12 months of age. Furthermore, these mice exhibited a glucose tolerance similar to that of wild-type animals even after 20 weeks of feeding with a high-fat diet. More recently, however, another G_M null mouse line was generated, and the homozygous mutants developed obesity, impaired glucose tolerance, and insulin resistance with age (after 44 to 52 weeks) (10).

Given that the nonfasting glycogen stores of PTG hemizygous mice were reduced by about 50% in adipose tissue, the liver, and the heart and by about 25% in skeletal muscle (white fibers), and that both types of G_M null mice manifested an approximately 90% reduction in the glycogen content of skeletal muscle, both of these glycogen-targeting subunits of PP1 (PTG and G_M) appear to play an important role in glycogen synthesis in vivo. There are, however, interesting differences between mice deficient in PTG and those lacking G_M. The most obvious such difference is the failure of PTG null mice to develop to term. Determination of the precise time and cause of embryonic death in PTG knockout mice will be important, given that this information may reveal a new and unexpected role for PTG in embryogenesis.

Partitioning of fuel substrates between glycogen and lipid

Another difference between the effects of PTG and G_M deficiency relates to glucose tolerance. Although there are some differences in phenotype between the two lines of G_M null mice (9, 10), glucose intolerance was not detected until at least 11 months of age in either model. In contrast, glucose intolerance was detected at three to four months of age and fasting hyperinsulinemia was already apparent at one to two months of age in PTG hemizygous mice. These differences between PTG and G_M mutant animals are suggestive of a difference in the fate of ingested glucose that is not utilized for glycogen synthesis. The second line of G_M null mice, in which glucose intolerance develops at 11 months of age, manifests an increased deposition of fat in the abdomen and at other sites at 12 months of age, suggesting that glucose not converted to glycogen in skeletal muscle accumulates as triglyceride in adipose tissue. The fact that PTG hemizygous mice showed no difference from wild-type littermates in body weight or in the weight of the liver or fat pads suggests that glucose not converted to glycogen in the liver, adipose tissue, the heart, or skeletal muscle does not accumulate as triglyceride in adipose tissue or the liver. The triglyceride content of hepatocytes and myocytes is thought to correlate negatively with insulin sensitivity (11, 12). The triglyceride content of muscle, but not that of the liver, was significantly increased (+130%) at 18 months of age in PTG hemizygous mice.

Comparison of the tissues in which glycogen stores are reduced between PTG and G_M mutant mice suggests that the deficiency of glycogen in the liver may be responsible for the more profound glucose intolerance of the PTG hemizygous animals. The effect on glucose tolerance of restoring the normal level of PTG gene expression specifically in the liver of the PTGhemizygous mice with the use of an adenovirus vector should help to verify this hypothesis.

Furthermore, PTG hemizygous mice will likely prove to be a good model with which to examine the relation between triglyceride accumulation in hepatocytes or myocytes and insulin resistance. It has not yet been established whether triglyceride accumulation in these cells is the cause or the result of insulin resistance (13, 14). Given that fasting hyperinsulinemia was already apparent at one to two months of age in the PTG hemizygotes, it will be interesting to examine both the triglyceride content and insulin signaling in muscle and the liver in animals at this age and at ages up to 18 months. The fact that these mice showed no change in the serum concentration of nonesterified free fatty acids but did exhibit a reduced glucose uptake in white fibers of muscle (at three to four months of age) suggests that they might also provide a unique tool with which to determine the mechanism of triglyceride accumulation in muscle associated with insulin resistance.

Glycogen metabolism and physiology: mouse-human differences

Finally, we have to consider the possibility that mice may not be an accurate model of human metabolism and physiology. This may be the case especially with regard to glycogen metabolism, given that the patterns of glycogen storage differ between the two species. Thus, although hepatic glycogen content is similar in humans and mice, the glycogen content of mouse muscle is only about 10% of that of human muscle, when expressed as a percentage of total body glycogen (15). The idea that a reduction in glycogen content of the liver might have a greater effect on glucose tolerance than does a decrease in the glycogen content of muscle, based on observations of mice, should therefore not automatically be assumed to apply to humans.

The phenotype of mice deficient in PTG suggests that the PTG gene is a candidate gene for type 2 diabetes and insulin resistance in humans. However, previous studies have indicated that PTG gene polymorphism does not contribute to insulin resistance or glucose intolerance (16, 17). Given that G_M gene polymorphism has been associated with insulin resistance in some human populations (18–20), it will be important to reexamine the possible relation between the PTG gene and insulin resistance in humans.