Cellular Mechanism of Action of Metformin - Diabetes Care

10 downloads 312 Views 1MB Size Report
Action of Metformin. Amira Klip, PhD. Lawrence A. Leiter, MD. Metformin is a hypoglycemic drug effective in the treatment of non-insulin-dependent diabetes ...
Amira Klip, PhD Lawrence A. Leiter, MD

Cellular Mechanism of Action of Metformin

Metformin is a hypoglycemic drug effective in the treatment of non-insulin-dependent diabetes mellitus and increasingly used in Canada and Europe. Effects on intestinal glucose absorption, insulin secretion, and hepatic glucose production are insufficient to explain its hypoglycemic action, with most evidence suggesting that the major effect of the drug is on glucose utilization. In vivo and in vitro studies have demonstrated that metformin stimulates the insulininduced component of glucose uptake into skeletal muscle and adipocytes in both diabetic individuals and animal models. This increase is more significant in diabetic than in nondiabetic animals, suggesting an enhanced action of the drug in the hyperglycemic state. The increase in glucose uptake is also reflected in an increase in the insulin-dependent portion of glucose oxidation. Potential sites of action of metformin are the insulin receptor and the glucose transporters. Although metformin increases insulin binding in various cell types, this effect is not universal and does not correlate with stimulation of glucose utilization. In contrast, direct effects of the drug on the glucose-transport system have been demonstrated. Metformin elevates the uptake of nonmetabolizable analogues of glucose in both nondiabetic rat adipocytes and diabetic mouse muscle. In the latter, the stimulatory effect of the drug is additive to that of insulin. In human and rat muscle cells in culture, metformin increases glucose-analogue transport independently of and additive to insulin, suggesting an insulin-independent action. Most of these results suggest that the basis for the hypoglycemic effect of this biguanide is probably at the level of skeletal muscle by increasing glucose transport

From the Division of Cell Biology, The Hospital for Sick Children, and the Department of Medicine, St. Michael's Hospital, Toronto, Ontario, Canada. Address correspondence and reprint requests to Dr. Amira Klip, Division of Cell Biology, The Hospital for Sick Children, Toronto, Ontario M5C 1X8, Canada.

696

across the cell membrane. Diabetes Care 13:696-704, 1990

he disubstituted biguanide metformin is an effective oral hypoglycemic agent (1-3; Fig. 1). It is approved for the management of non-insulindependent diabetes mellitus (NIDDM) in several countries, including Canada, although it is still an investigational drug in the United States. Use of this drug has been associated with only minimal toxicity. Lactic acidosis, although a problem with the biguanide phenformin, has rarely been reported with metformin and only in situations in which the drug should not have been used, e.g., renal impairment (4). Major clinical differences between metformin and the drugs of the other major class of oral hypoglycemic agents (sulfonylureas) are that 7) therapeutic dosages of metformin do not cause hypoglycemia, 2) metformin does not lower blood glucose in nondiabetic individuals (5), and 3) only metformin has direct beneficial effects on serum lipids and lipoproteins (6-8). Various beneficial vascular effects of metformin have also been reported, including reduction of microvascular permeability (9) and vascular cell proliferation (10) in diabetic animal models. The selectivity toward the hyperglycemic state is of great interest and suggests that the effect of metformin may be closely linked to the defective pathways inherent to NIDDM. Insulin resistance is a major feature of NIDDM and is perhaps the primary defect in the disease (11). Skeletal muscle is the major tissue responsible for insulin-dependent glucose utilization, and therefore the resistance in muscle is a major cause of hyperglycemia

T

DIABETES CARE, VOL. 13, NO. 6, JUNE 1990

A. KLIP AND L.A. LEITER

NH CH C

H

dent) and/or insulin stimulated. These include insulinreceptor binding and signals derived from it, glucose transport across the cell membrane, glucose oxidation at the level of pyruvate dehydrogenase, and glycogen synthesis at the level of the glycogen-synthase complex. Thus, metformin could potentially increase glucose transport into the cell, its oxidation to CO2, or its conversion to glycogen.

FIG. 1. Structure of metformin (ty/V-dimethylbiguanide). INSULIN ACTION

(11,12). The molecular basis of muscle insulin resistance is unknown, but defects have been proposed to occur at the levels of the insulin receptor, messenger signals, and glucose-transport system (13). Although metformin has been available since the 1950s, its precise mechanism of action is still unknown. This article reviews the potential sites of action of the drug. We suggest that metformin mimics and/or potentiates insulin action in muscle, specifically the stimulation of glucose uptake.

POSSIBLE SITES OF ACTION OF METFORMIN: PANCREAS, GUT, LIVER, OR MUSCLE? Metformin is not metabolized in humans or rats, and its circulating levels range from 12 to 100 |xM in humans after the standard oral dose of 1.5 g/day (14,15). Theoretically, the drug could lower glycemia by reducing glucose absorption in the intestine, stimulating insulin secretion from the pancreatic 3-cells, decreasing glucose output in the liver, and/or increasing glucose uptake in peripheral tissues (i.e., muscle and fat). Metformin inhibits intestinal glucose absorption in nondiabetic and diabetic rats, but this action is insufficient to account for the full ability of the drug to reduce glycemia (1,16). Moreover, one study could not detect impaired glucose absorption in metformin-treated patients (17). Stimulation of insulin secretion from the pancreas has been ruled out, because the drug does not increase the circulating levels of insulin or C-peptide in individuals with NIDDM (1,2,5,17,18). Although metformin was reported to inhibit hepatic gluconeogenesis in fasted guinea pigs and rats (19) and to increase hepatic insulin sensitivity in lean NIDDM patients (17), other studies showed no effect of metformin on hepatic glucose production in obese NIDDM patients (20). In contrast to these inconclusive results, there is evidence indicating that metformin increases glucose uptake in peripheral tissues (1,2,20). Effects in both adipocytes and muscle cells have been described, and some of these studies have suggested that the main effect of metformin in obese NIDDM patients is not in fat but rather in muscle (21). The cellular mechanism of action of metformin in muscle could involve any of the rate-limiting steps in glucose metabolism, either basal (i.e., insulin indepen-

DIABETES CARE, VOL. 13, NO. 6, JUNE 1990

In the past few years, understanding of the mechanism of insulin-stimulated glucose uptake into cells has increased considerably. The subject has been reviewed extensively elsewhere (22). The first event is the binding of the hormone to specific receptors on the cell surface. These insulin receptors are ap-heterodimers that bind insulin at the extracellular a-subunit, and this results in transmission of a signal to the intracellular (3-subunits, which then become self-phosphorylated on tyrosine residues unmasking a kinase activity that phosphorylates other cellular substrates (23). None of these phosphorylated products has been directly related to a metabolic action of insulin. Glucose transport into cells is mediated by specific-membrane proteins, i.e., the glucose transporters. Although stimulation of glucose transport appears to require the activity of the insulin-activated receptor kinase, glucose transporters are not phosphorylated by insulin in intact cells, suggesting that other substrates of the kinase regulate glucose transport (24, 25). Water-soluble inositol phosphoglycans have been claimed to mediate insulin actions such as stimulation of glycogen synthesis and glucose oxidation but conspicuously not stimulation of glucose transport (26,27).

PERIPHERAL ACTION OF METFORMIN Effects of metformin on peripheral glucose utiliza-

tion. Glucose transport in vivo cannot be assessed directly but rather is inferred from measurements of glucose utilization. The observation that in metform in-treated NIDDM patients oral glucose tolerance was improved significantly relative to placebo-treated patients triggered more in-depth studies of metformin action on glucose utilization (28). Glucose utilization (disposal) in vivo is often measured by the euglycemic clamp technique. With this procedure, it has been shown that metformin increases glucose utilization of peripheral tissues by 50% at a high insulin infusion rate and by 25% at a low infusion rate (29,30). These increases are not trivial, because diabetic patients have a decrease in peripheral glucose utilization of —20-40% relative to control subjects (11). Intriguingly, a recent study failed to detect increased peripheral glucose utilization under hyperinsulinemic conditions (18). Because clamp studies use maximal insulin concentrations, the effect of metformin in the absence of the hormone could not be assessed in

697

MECHANISM OF ACTION OF METFORMIN

these studies. Moreover, no studies of metformin action on glucose uptake have been reported in nondiabetic individuals. An interesting observation that further emphasizes the effect of metformin on glucose utilization derives from studies of insulin resistance after injury in the rat, a condition that leads to hyperglycemia (31). Administration of metformin at the time of injury relieved the hyperglycemia and increased the rate coefficient of glucose utilization. This effect of metformin was not observed in uninjured rats, further substantiating the observations made in humans that the effectiveness of metformin in decreasing glycemia is specific for the hyperglycemic state. Effects of metformin on insulin receptor. Early studies of metformin action found that the drug increased by - 2 0 % the number of insulin receptors in circulating cells (erythrocytes and monocytes) of patients with either insulin-dependent diabetes mellitus or NIDDM (32-34). These observations led to the belief that the effect of the drug on glycemia could be explained by an increase in insulin binding. This conclusion was further supported by the lack of increase in insulin binding in cells from metformin-treated nondiabetic individuals in whom the drug had no effect on glycemia (33,35). However, the ability to increase insulin binding exclusively in cells from diabetic patients was contested by observations of metformin-dependent increased insulin binding to cells from obese nondiabetic individuals as well (33). Furthermore, other studies reported that metformin was without effect (28,30) or even decreased insulin binding in monocytes from NIDDM patients (29). Adipocytes from metform in-treated diabetic patients also showed no increase in insulin binding, and the drug did not increase hormone binding to isolated human adipocytes in vitro (21). In animal models, the results have been equally inconsistent. Hepatocytes isolated from metformin-treated lean and obese mice had increased insulin

binding, but soleus muscle cells and hepatocytes isolated from streptozocin-induced diabetic mice showed no change in insulin binding associated with in vivo metformin treatment (36,37). In vitro addition of metformin to isolated cells resulted in an increase in insulin receptors in cells as diverse as normal human erythrocytes, the mammalian lymphoid cell line IM-9, and human breast cancer cells (38-41). Thus, the ability of metformin to increase insulin receptors and insulin binding cannot be consistently demonstrated (Table 1). Even in instances when metformin was shown to increase insulin binding to cells, this effect did not appear to be directly related to the subsequent metabolic and clinical effects of the drug. Prager et al. (30) and Fantus and Brosseau (42) described groups of NIDDM individuals whose glycemic control was improved by metformin treatment without significant changes in insulin binding to monocytes. Similarly, in mouse skeletal muscle, increased glycogen formation was demonstrated in the absence of increased insulin binding (37). Additional evidence dissociating the effects of metformin on insulin binding from metabolic actions include the observations of increased binding of monocytes from obese patients (33) in the absence of changes in glycemia and increased binding without hypoglycemic response in hepatocytes from ob/ob mice (36). The best evidence against a causal relationship between effects on binding and metabolism is that in isolated nondiabetic rat adipocytes treated in vitro with metformin, the increase in insulin binding was secondary (seen after 20 h) to the metabolic effects on glucose utilization (seen after only 2 h; 42). Moreover, NIDDM patients could be grouped into those who had increased insulin binding to monocytes and those who did not, yet glycemia was lowered in both groups (42). From these studies in human and rat adipocytes and in human monocytes, it appears that although metformin may increase insulin binding, its direct action occurs at a post-insulin-binding level.

TABLE 1 Effects of metformin on insulin binding and receptor kinase Species Human Nondiabetic Obese NIDDM NIDDM NIDDM NIDDM Rat Nondiabetic STZ-D Mouse Nondiabetic ob/ob STZ-D STZ-D

Cell type

Metformin administration

vivo vivo vivo vivo vivo vitro

Insulin binding

In In In In In In

Fat Muscle

In vitro In vivo

t

Hepatocyte Hepatocyte Hepatocyte Muscle

In In In In

T t

t t 4 NE NE NE

NE NE

Refs.

33, 35 33 29, 33, 34 28, 30, 42 21 21

NE

Blood Blood Blood Blood Fat Fat

vivo vivo vivo vivo

Insulin-receptor kinase

NE

T

44 45 36 36 36 37

NE, no effect; f , increase; j , decrease; NIDDM, non-insulin-dependent diabetes mellitus; STZ-D, streptozocin-induced diabetes.

698

DIABETES CARE, VOL. 13, NO. 6, JUNE 1990

A. KLIP AND L.A. LEITER

Effect of metformin on insulin-receptor kinase activity and receptor internalization. Circulating cells and cultured fibroblasts from insulin-resistant patients have abnormal insulin-receptor kinase activity, as do muscle cells from diverse diabetic animal models (23,43). Insulin-receptor kinase has been postulated to be required for insulin signaling, including stimulation of glucose transport. The relevance of kinase activity has been contested, however, in studies in which signaling is elicited with anti-insulin-receptor antibodies in the absence of stimulation of the kinase (23). It is conceivable that metformin could increase insulin-receptor kinase activity and in this way improve insulin action. Two studies have investigated this question in nondiabetic rat adipocytes (44) and diabetic rat skeletal muscle (45). In the former study, there was no effect of metformin on either receptor autophosphorylation or its exogenous kinase activity. In contrast, in the latter study, the authors observed a decrease in kinase activity of streptozocin-induced diabetic rat muscle that increased to supranormal levels in response to metformin treatment. Clearly, the effect of metformin on insulin-receptor kinase requires further clarification (Table 1). Another aspect of insulin-receptor function, i.e., its internalization after ligand binding, has also been found to be diminished in monocytes from obese NIDDM patients relative to control subjects, and this reduction was corrected by in vivo metformin treatment (46). It is important to consider that the recovery of insulinreceptor kinase activity in muscle and of receptor internalization in monocytes may be an indirect rather than direct result of metformin therapy as a consequence of

improved glycemia and body mass index. It was recently reported that decreasing glycemia through weight loss in diabetic patients or weight loss alone in obese patients can restore normal insulin-receptor kinase activity (47,48). A direct cause-effect relationship between metformin action and the recovery of normal receptor kinase and internalization thus remains to be proved. Effects of metformin on glucose uptake, oxidation, and glycogen synthesis in isolated tissues and cells. Most studies of metformin action on glucose uptake in isolated tissues or cells do not differentiate between transport and subsequent metabolism. In most cases, there are no good measurements of initial rates of glucose transport, because for the most part glucose rather than nonmetabolizable analogues has been used. The effects of metformin on glucose uptake have been studied in numerous cell and tissue systems, either after in vivo treatment with the drug or by direct exposure in vitro, as described below (Table 2). Studies of muscle glucose metabolism after pretreatment with metformin are few and inconclusive. An early study showed that in food-restricted rats that become hypoglycemic, glucose uptake into the perfused hindquarter muscle (whether basal or insulin stimulated) was not different in metformin-treated and untreated animals (49). This lack of action may be due to the prevailing hypoglycemia. Conversely, glucose oxidation was elevated in soleus muscle isolated from metform in-pretreated streptozocin-induced diabetic mice (50). This effect was noticed only in the presence of insulin. In contrast, the only human study to measure in vivo glucose uptake into muscle demonstrated that, in sulfo-

TABLE 2 Effects of metformin on glucose metabolism Species Human Nondiabetic Nondiabetic NIDDM Sulfony I urea-treated NIDDM Rat Nondiabetic Nondiabetic Hyperglycemic due to injury Hypoglycemic due to food restriction L6 cells Alloxan-induced diabetic Guinea pig Nondiabetic Mouse Streptozocin-induced diabetic

Cell type

Metformin administration

vitro vitro vivo vivo

Glucose uptake

Fat Muscle Peripheral Muscle

In In In In

Fat Muscle Peripheral Muscle Muscle line Muscle

In vitro In vitro In vivo In vitro In vitro

NE B,

Fat

In vitro

|B

Muscle

In vivo

NE B, t i

Glucose oxidation

Glycogen synthesis

NE B; f i(r)

34 35 18, 29, 30 17

|B

42 52 31 49 57 52

ti t i, NE i

NE B,

Refs.

fit

|i NE B, NE i

|B |i

51 | B , fi(r)

50

NE, no effect; B, basal (i.e., in absence of insulin); | , increase; i(r), insulin responsiveness; i, in presence of insulin; | , decrease; NIDDM, non-insulin-dependent diabetes mellitus. *Relative to sulfonylurea-treated NIDDM without metformin therapy, which had higher glycemia. tin presence of sodium butyrate only.

DIABETES CARE, VOL. 13, NO. 6, JUNE 1990

699

MECHANISM OF ACTION OF METFORMIN

nylurea-treated patients, the addition of metformin therapy actually decreased forearm glucose uptake after a glucose load and did not change uptake in the postabsorptive state (17). The authors pointed out, however, that these results do not conflict with the observed increase in peripheral glucose utilization under clamp conditions observed in other studies, because the ambient glucose concentration also dictates glucose uptake, and during the forearm glucose uptake measurements, the metformin-treated patients had lower glycemia than the diabetic control subjects. In nondiabetic rat adipocytes and guinea pig fat pads, metformin added in vitro increased basal glucose oxidation in a concentration-dependent manner, with maximal effects observed at 5 mM and half-maximal effects at 0.3 mM (42,51). The maximum increase was about a doubling in glucose oxidation, and the full response was observed within 2 h of addition of the drug. Interestingly, there was no effect of the drug in the presence of maximal insulin concentrations (42). These results in nondiabetic rat adipocytes suggest that metformin can stimulate glucose oxidation independently of insulin and that the biguanide uses some finite component in the insulin pathway. No similar studies have been performed in adipocytes of diabetic animals. In contrast to rodent adipocytes, addition of metformin in vitro to isolated fat tissue fragments from nondiabetic humans did not affect glucose oxidation in the absence of insulin, but it increased insulin responsiveness by 20% (34). These results suggested that, in cells of nondiabetic individuals, the drug potentiates insulin action rather than mimicking the effect of the hormone and that it does not exhaust any component of the insulin pathway. As was the case for studies with rodents, there are no reports of effects of metformin on glucose oxidation in adipocytes from diabetic individuals. Studies in rat skeletal muscle have indicated that in vitro treatment with metformin also increases glucose uptake in this tissue. Frayn and Adnitt (52) observed a 15-35% increase in insulin-dependent glucose uptake in the presence of 10 (xg/ml metformin in vitro (measured by the increase in concentration of intracellular free glucose) in diaphragm muscle from alloxan-induced diabetic rats. The increase was specific for the insulindependent uptake and for diabetic animals, because it was not seen in diaphragms from control rats unless incubated in the presence of sodium butyrate, which lowered basal uptake. Along the same lines is the only study that has investigated the effect of metformin on human muscle in vitro (53). Fragments of nondiabetic human rectus and gluteus muscles removed at surgery were used. Glucose disappearance from the medium was measured after 90 min of exposure to metformin (60 fiM). Unfortunately, these measurements do not reflect the initial rate of glucose transport but rather an averaged rate of glucose uptake. As with the rat muscle, the drug did not affect basal or insulin-stimulated transport unless hexokinase was first depressed by sodium butyrate. Under these

700

conditions, metformin increased uptake only in the presence of insulin by —30%. This suggested that the drug increases glucose uptake in nondiabetic human muscle, preferentially in the insulin-stimulated state. Regrettably, the effect in diabetic human muscle was not investigated. In contrast to the above studies on glucose uptake and oxidation, the effect of metformin on glycogen synthesis has been less explored. Earlier studies indicated that phenformin, a related biguanide, increased glucose uptake into diabetic muscle undergoing work load by increasing glycogen synthesis (54). Metformin enhanced basal and insulin-stimulated glucose incorporation into glycogen in cultured rat hepatoma cells (55). In contrast, in rat hepatocyte monolayers, metformin actually reduced basal and insulin-stimulated glycogen synthesis (56). In a preliminary communication, Molnar and Davidson (57) described a metformin-dependent increase in basal glycogen synthesis in L6 muscle cells after 24 h. The results summarized above and in Table 2 describe effects of metformin on glucose utilization (measured as glucose uptake, oxidation, and glycogen synthesis). The regulatory steps in these processes are the transport of glucose across the cell membrane, the activity of pyruvate dehydrogenase, and the activity of the glycogensynthase system. Although the effect of metformin on the specific activities of these enzymes has not been studied directly, several reports have emerged on the effect of metformin on glucose transport across the membrane. Under physiological conditions, this process is the rate-limiting step in glucose utilization. Effects of metformin on glucose transport in isolated tissues and cells. Glucose transport is mediated by specific glucose transporters embedded in the cell membrane. The transporters of muscle and fat cells belong to a family of related transmembrane polypeptides of —45,000 Mr (58-62). Two isoforms of transporters have been detected in these tissues, denominated GLUT-1 and GLUT-4, which differ in their subcellular localization. In adipocytes (61,63) and skeletal muscle (64), GLUT-1 is found predominantly in the plasma membrane and GLUT-4 primarily in intracellular membranes. Exposing adipocytes or skeletal muscle to insulin appears to cause migration of glucose transporters to the plasma membrane, predominantly the GLUT-4 isoform, based on in vitro observations of disappearance of GLUT-4 molecules from intracellular membranes and their recovery in plasma membranes (61,63-65). Such recruitment of glucose transporters constitutes a primary mechanism of action of insulin that results in stimulation of glucose-transport activity. In addition, it has been proposed that the hormone further increases the intrinsic activity (i.e., kinetic turnover rate) of the transporters (GLUT-1 and/or GLUT-4) in the plasma membrane of fat and muscle cells (66,67). Strategies to study glucose transport in fat and muscle include the use of nonmetabolizable analogues of glucose. Glucose-transporter number is determined through the binding of the spe-

DIABETES CARE, VOL. 13, NO. 6, JUNE 1990

A. KLIP AND LA. LEITER

cific inhibitor cytochalasin B or of anti-transporter antibodies (59,61,63-65,68). These techniques have been applied to study the effect of metformin on glucose transport, and the results are reviewed below (Table 3). Jacobs et al. (44) measured 2-deoxyglucose uptake in nondiabetic rat adipocytes. Under defined experimental conditions (low extracellular substrate and short times), uptake of this glucose analogue is a measurement of transport across the cell membrane. Insulin-stimulated uptake was increased by —20% after 20 h of exposure to 0.1 mM metformin in vitro (44). This effect was on responsiveness rather than on sensitivity, indicating that the drug facilitates a postreceptor process, possibly glucose transport (44). Effects on basal transport were not reported. Similarly, in a preliminary study, Matthaei et al. (69) found that metformin increases insulin-stimulated transport of the nonmetabolizable glucose analogue 3-O-methylglucose but not basal transport in rat adipocytes. These authors further detected increased cytochalasin B binding (i.e., glucose-transporter number) in plasma membranes derived from metformin plus insulin-treated adipocytes relative to insulin-treated cells, suggesting that metformin potentiates the insulin-dependent recruitment of glucose transporters. Because there was no effect on glucose-transporter recruitment by metformin alone, it was proposed that in rat adipocytes, the drug acts through potentiation of the insulin-stimulated pathway. The only study that covered metformin's effects on glucose transport in skeletal muscle is that of Bailey and Puah (50), who measured transport of 3-O-methylglucose in soleus muscle of streptozocin-induced diabetic mice. The rate of uptake (i.e., transport) of this hexose increased by 20% at both submaximal and maximally stimulating insulin concentrations. The effect of metformin was tested after in vivo administration. In contrast to the lack of effect of the drug on basal glucose oxidation, basal 3-O-methylglucose transport (i.e., in the absence of insulin) was also stimulated by 20% in muscle from metformin-treated mice. These observations suggested that the drug elicits a direct effect on the glucose-transport system. Thus, in contrast to the effect of metformin in adipocytes, which is insulin dependent, the effect of metformin on glucose transport in muscle appears to be insulin independent.

Muscle cells in culture offer a distinct advantage for the study of glucose transport and metabolism because initial rates of transport can be measured. Furthermore, prolonged incubations required to test the chronic effect of hormones and drugs can be performed. Muscle cell lines and primary cultures of skeletal muscle have been used to define cellular mechanisms of action of metformin, and the results of these studies are summarized below. The best-characterized muscle cell line is the L6 line of rat skeletal muscle origin, which differentiates in vitro mimicking the myoblast to myotube maturation process that occurs during muscle ontogenesis. The characteristics of glucose transport and its regulation by hormones have been well documented (70,71). In these cells, 2deoxyglucose and 3-O-methylglucose transport are fully inhibited by cytochalasin B, and transport is stimulated by insulin (71). We have recently observed that micromolar concentrations of metformin stimulate the uptake of 2-deoxyglucose and 3-O-methylglucose in L6 muscle cells incubated in high-glucose medium (72). This response did not require the presence of insulin and was additive to the effect of the hormone. The effect was time dependent, becoming noticeable by 4 h and peaking after —24 h. Despite the long time required for this effect, the stimulation produced by metformin after 24 h did not require ongoing protein synthesis. Preliminary data indicate that metformin increased the amount of CLUT-4 glucose transporters in the plasma membrane. In this fashion, the chronic effect of metformin administration is reminiscent of the response to acute exposure to insulin previously observed in these cells (73) and in rat skeletal muscle (65). Interestingly, this response differs from that to chronic (24 h) exposure to insulin, which is characterized by increased glucose transport resulting from increased glucose-transporter synthesis (74,75). Primary cell cultures of skeletal muscle have been shown recently in our laboratory to express glucosetransport activity of general properties similar to those of adult skeletal muscle such as stereospecificity, inhibition by cytochalasin B, and stimulation by insulin (76). In preliminary experiments with metformin, we observed that drug concentrations in the therapeutic range stimulate 2-deoxyglucose and 3-O-methylglucose up-

TABLE 3 Effects of metformin on glucose transport and transporters Species Human Nondiabetic Rat Nondiabetic L6 cells Mouse Streptozocin-induced diabetic

Cell type

Metformin administration

Glucose transport

Primary muscle

In vitro

t B, ti(r)

Fat Muscle line

In vitro In vitro

f i(r) t B, j i(r)

Muscle

In vivo

t B,

T i(r)

Transporter recruitment

Refs.

72 | i, Cytochalasin B binding t B, GLUT-4 isoform

44, 69 72 50

, Increase; B, basal (i.e., in absence of insulin); i(r), insulin responsiveness; i, in presence of insulin.

DIABETES CARE, VOL. 13, NO. 6, JUNE 1990

701

MECHANISM OF ACTION OF METFORMIN

take in these cultures, even in the absence of insulin, in high-glucose medium (72). These results, together with the observations made in L6 muscle cells, suggest that biguanide stimulates glucose transport in muscle cells through a mechanism similar to that of insulin but independent of the presence of the hormone. This becomes especially important in view of recent findings of diminished glucose transporters in muscle (77) and fat cells (78,79) of diabetic animals. The mechanism of action of metformin in inducing recruitment of GLUT-4 glucose transporter remains to be determined. It is an exciting possibility that this drug could become instrumental in studying the molecular basis of the recruitment mechanism and perhaps of insulin action.

increases the GLUT-4 glucose transporters present in the plasma membrane. In summary, it has been suggested for several years that the major effect of metformin is to increase the peripheral utilization of glucose, although the precise cellular mechanism was unknown. Additional evidence suggests that the drug exerts its major action at the postreceptor level to increase glucose uptake, specifically glucose transport. Preliminary studies from our laboratory performed on muscle cells (the tissue primarily responsible for the insulin resistance of NIDDM) in culture suggest that metformin may stimulate glucose transport by increasing GLUT-4 glucose transporters in the plasma membrane. This effect is similar to that caused by acute administration of insulin, although the hormone is not required to observe this effect in culture. Thus, it appears that the basis for the hypoglycemic effect of metformin is at the level of skeletal muscle by increasing glucose transport across the cell membrane.

CONCLUSION etformin is a hypoglycemic drug effective in the treatment of NIDDM and increasingly used in Canada and Europe. Effects of intestinal glucose absorption, insulin secretion, or hepatic glucose production are insufficient to explain its hypoglycemic action, with most evidence suggesting promotion of glucose utilization. Metformin therapy stimulates the insulin-induced component of whole-body glucose utilization in diabetic individuals. In vivo treatment with metformin and in vitro metformin administration increase insulin binding; however, this effect is not universal and does not correlate with induction of metabolic responses. Studies of metformin on the insulin-receptor kinase are scarce and contradictory. Muscle glucose oxidation (insulin dependent) is elevated after in vivo metformin administration to diabetic mice. Adipose glucose oxidation increases after in vitro metformin administration to nondiabetic rats and humans. In rats, the stimulation requires submaximal concentrations of insulin; in humans, the metformin effect is additive to that of insulin. In diabetic rat muscle, metformin addition in vitro increases the insulin-dependent portion of glucose uptake; in nondiabetic rat and human muscle, this effect was observed only in sodium butyrate-pretreated muscles. Metformin increases basal and insulindependent glycogen synthesis in cultured hepatoma cells but decreases it in hepatocyte monolayers. In L6 muscle cells, it increases basal glycogen synthesis. In nondiabetic rat adipocytes, metformin added in vitro increases the insulin responsiveness of 2-deoxyglucose and 3-O-methylglucose transport. In diabetic mouse muscle, both basal and insulin-stimulated 3-O-methyglucose transport increase after in vivo metformin treatment; the drug increases insulin responsiveness. In both human and rat muscle cells in culture, metformin increases 2deoxyglucose and 3-O-methyglucose transport independently of and additive with insulin, suggesting insulin-independent action. In L6 muscle cells, metformin

M

702

ACKNOWLEDGMENTS This work was supported by a University-Industry grant from the Medical Research Council of Canada and Nordic Laboratories, Laval, Quebec, Canada. We thank Toolsie Ramlal, Vivian Sarabia, and Philip J. Bilan for participation in the experiments performed in muscle cells in culture.

REFERENCES 1. Bailey CJ: Metformin revisited: its actions and indications for use. Diabetic Med 5:315-20, 1988 2. Schafer G: Biguanides: a review of history, pharmacodynamics and therapy. Diabete Metab 9:148-63, 1983 3. Steiner DF, Williams RH: Actions of phenethylbiguanide and related compounds. Diabetes 8:154-57, 1959 4. Cohen RD: The relative risks of different biguanides in the causation of lactic acidosis. Res Clin Forums 1:12530, 1979 5. Hermann LS: Metformin: a review of its pharmacological properties and therapeutic use. Diabete Metab 5:233-45, 1979 6. Rains SGH, Wilson GA, Richmond W, Elkeles RS: The effect of glibenclamide and metformin on serum lipoproteins in type 2 diabetes. Diabetic Med 5:653-58, 1988 7. Sirtori CR, Tremoli E, Sirtori M, Conti F, Paoletti R: Treatment of hypertriglyceridemia with metformin: effectiveness and analysis of results. Atherosclerosis 26:583-92, 1977 8. Descovich G, Montaguti U, Ceredi C, Cocuzzsa E, Sirtori CR: Long-term treatment with metformin in a large cohort of hyperlipidemic patients. Artery 4:348-59, 1978 9. Colantuoni A, Bertuglia S, Donato L: Effects of metformin on microvascular permeability in diabetic Syrian hamsters. Diabete Metab 14:549-53, 1988 10. Koschinsky T, Bunting CE, Rutter R, Gries FA: Influence of metformin on vascular cell proliferation. Diabete Metab 14:566-70, 1988

DIABETES CARE, VOL. 13, NO. 6, JUNE 1990

A. KLIP AND LA. LEITER

29. Nosadini R, Avogaro A, Trevisan R, Valerio A, Tessari P, 11. DeFronzo RA: Lilly Lecture 1987: the triumvirate: p-cell, Duner E, Tiengo A, Velussi M, Del Prato S, De Kreutzmuscle, liver: a collusion responsible for NIDDM. Diaenberg S, Muggeo M, Crepaldi G: Effect of metformin on betes 37:667-87, 1988 insulin-stimulated glucose turnover and insulin binding to 12. Baron AD, Brechtel G, Wallace P, Edelman SV: Rates and receptors in type II diabetes. Diabetes Care 10:62-67, tissues of non-insulin- and insulin-mediated glucose up1987 take in humans. Am ) Physiol 255:E769-74, 1988 30. Prager R, Schernthaner G, Graf H: Effect of metformin on 13. Ciaraldi TP, Kolterman OG, Scarlett JA, OlefskyJM: Role peripheral insulin sensitivity in non insulin dependent diof glucose transport in the postreceptor defect of nonabetes mellitus. Diabete Metab 12:346-50, 1986 insulin-dependent diabetes mellitus. Diabetes 31:101622, 1982 31. Frayn KN: Effects of metformin on insulin resistance after injury in the rat. Diabetologia 12:53-60, 1976 14. Pentikainen PJ, Neuvonen PJ, Penttila A: Pharmacokinetics of metformin after intravenous and oral administration 32. Pagano G, Tafliaferro V, Carta Q, Caselle MT, Bozzo C, to man. Eur j Clin Pharmacol 16:195-202, 1979 Vitelli F, Trovati M, Cocuzza E: Metformin reduces insulin requirement in type 1 (insulin-dependent) diabetes. Dia15. Sirtori CR, Franceschini G, Galli-Klenle M, Cighetti G, betologia 24:351-54, 1983 Gallik G, Bondioli A, Conti F: Disposition of metformin (N,N-dimethylbiguanide) in man. Clin Pharmacol Ther 33. Trischitta V, Gullo D, Pezzino V, Vigneri R: Metformin 24:683-93, 1978 normalizes insulin binding to monocytes from obese nondiabetic subjects and obese type II patients. I Clin Endo16. Lorch E: Inhibition of intestinal absorption and improvecrinol Metab 57:713-18, 1983 ment of oral glucose tolerance by biguanides in the nor34. Lord JM, White SI, Bailey CJ, Atkins TW, Fletcher RF, mal and in the streptozotocin-diabetic rat. Diabetologia Taylor KG: Effect of metformin on insulin receptor binding 7:195-203, 1971 and glycaemic control in type II diabetes. Br Med I 17. Jackson RA, Hawa Ml, Jaspan JB, Sim BM, DiSilvio L, 286:830-31, 1983 Featherbe D, Kurtz AB: Mechanism of metformin action in non-insulin-dependent diabetes. Diabetes 36:632-40, 35. Cigolini M, Bosello O, Zancaro C, Orlandi PG, Fezzi O, 1987 Smith U: Influence of metformin on metabolic effect of 18. Wu MS, Johnston P, Sheu WH-H, Hollenbeck CB, Jeng insulin in human adipocyte tissue in vitro. Diabete Metab C-Y, Goldfine ID, Chen Y-DI, Reaven GM: Effect of met10:311-15, 1984 formin on carbohydrate and lipoprotein metabolism in 36. Lord JM, Atkins TW, Bailey CJ: Effect of metformin on NIDDM patients. Diabetes Care 13:1-8, 1990 hepatocyte insulin receptor binding in normal, strepto19. Meyer F, Ipaktchi M, Clauser H: Specific inhibition of zotocin diabetic and genetically obese diabetic (ob/ob) gluconeogenesis by biguanides. Nature (Lond) 213:203— mice. Diabetologia 25:108-13, 1983 204, 1967 37. Lord JM, Puah JA, Atkins TW, Bailey CJ: Post-receptor 20. Hother-Nielsen O, Schmitz O, Andersen PH, Beck-Nieleffect of metformin on insulin action in mice. / Pharm sen H, Pedersen O: Metformin improves peripheral but Pharmacol 37:821-23, 1985 not hepatic insulin action in obese patients with type II 38. Goldfine ID, Iwamoto Y, Pezzino V, Trischitta V, Purrello diabetes. Acta Endocrinol 120:257-65, 1989 F, Vigneri R: Effects of biguanides and sulfonylureas on 21. Pedersen O, Nielsen OH, Bak J, Richelsen B, Beck-Nielinsulin receptors in cultured cells. Diabetes Care 7:54sen H, Schwartz Sorensen N: The effects of metformin on 58, 1984 adipocyte insulin action and metabolic control in obese 39. Holle A, Mangels W, Dreyer M, Kuhnau J, Rudiger HW: subjects with type 2 diabetes. Diabetic Med 6:249-56, Biguanide treatment increases the number of insulin-re1989 ceptor sites on human erythrocytes. N Engl J Med 305: 22. Czech MP: Molecular Basis of Insulin Action. New York, 563-66, 1981 Plenum, 1985 40. Vigneri R, Gullo D, Pezzino V: Metformin and insulin 23. Klip A, Douen AG: Role of kinases in insulin stimulation receptors. Diabetes Care 7 (Suppl. 1):113-17, 1984 of glucose transport. / Membr Biol 111:1-23, 1989 41. Vigneri R, Pezzino V, Wong KY, Goldfine ID: Comparison 24. Gibbs EM, Allard JW, Lienhard GE: The glucose transof the in vitro effect of biguanides and sulfonylureas on porter in 3T3-L1 adipocytes is phosphorylated in response insulin binding to its receptors in target cells. I Clin Ento phorbol ester but not in response to insulin.) Biol Chem docrinol Metab 54:95-100, 1982 261:16597-601, 1986 42. Fantus IG, Brosseau R: Mechanism of action of metfor25. James DE, Hiken F, Lawrence JC Jr: Isoproterenol stimumin: insulin receptor and post-receptor effects in vitro and lates phosphorylation of the insulin-regulatable glucose in vivo, j Clin Endocrinol Metab 63:898-905, 1986 transporter in rat adipocytes. Proc Natl Acad Sci USA 43. Grunberger G, Comi RJ, Carpentier J-L, Podskalny JM, 86:8368-72, 1989 McElduff A, Taylor SI, Gorden P: Insulin receptor tyrosine kinase activity is abnormal in circulating cells and cul26. Saltiel AR, Cuatrecasas P: Insulin stimulates the generatured fibroblasts but normal in transformed lymphocytes tion from hepatic plasma membranes of modulators defrom a type A insulin-resistant patient. I Lab Clin Med rived from an inositol glycolipid. Proc Natl Acad Sci USA 112:122-32, 1988 83:5793-97, 1986 27. Kelly KL, Mato JM, Merida I, Jarett L: Glucose transport 44. Jacobs DB, Hayes GR, Truglia JA, Lockwood DH: Effects and antilipolysis are differentially regulated by the polar of metformin on insulin receptor tyrosine kinase activity head group of an insulin sensitive glycophospholipid. Proc in rat adipocytes. Diabetologia 29:798-801, 1986 Natl Acad Sci USA 84:6404-407, 1987 45. Rossetti L, DeFronzo RA, Gherzi R, Stein P, Andraghetti 28. Prager R, Schernthaner G: Insulin receptor binding to G, Falzetti G, Klein-Robbennaar E, Cordera R: Insulin monocytes, insulin secretion, and glucose tolerance folreceptor tyrosine kinase activity and in vivo insulin action lowing metformin treatment: results of a double-blind crossin diabetic rats: effect of metformin treatment. Metaboover study in type II diabetics. Diabetes 32:1083-86, 1983 lism. In press

DIABETES CARE, VOL. 13, NO. 6, JUNE 1990

703

MECHANISM OF ACTION OF METFORMIN

46. Benzi L, Trischitta E, Ciccarone AM, Cecchetti P, Brunetti A, Cai F, Vigneri R, Navalesi R: Metformin improves the reduced insulin internalisation and intracellular processing in obese diabetic patients but not in subjects with simple obesity (Abstract). Diabetologia 31:469A, 1988 47. Brillon DJ, Freidenberg GR, Henry RR, Olefsky JM: Mechanism of defective insulin-receptor kinase activity in NIDDM: evidence for two receptor populations. Diabetes 38:397-403, 1989 48. Caro JF, Dohm LG, Pories WJ, Sinha MK: Cellular alterations in liver, skeletal muscle, and adipose tissue responsible for insulin resistance in obesity and type II diabetes. Diabetes Metab Rev 5:665-89, 1989 49. Kemmer FW, Berger M, Herberg L, Gries FA: Effects of metformin on glucose metabolism of isolated perfused rat skeletal muscle. Arzneim Forsch 27:1573-76, 1977 50. Bailey CJ, Puah JA: Effect of metformin on glucose metabolism in mouse soleus muscle. Diabete Metab 12:212— 18, 1986 51. Jangaard NO, Pereira JN, Pinson R: Metabolic effects of the biguanides and possible mechanism of action. Diabetes 17:96-104, 1968 52. Frayn KN, Adnitt PI: Effects of metformin on glucose uptake by isolated diaphragm from normal and diabetic rats. Biochem Pharmacol 21:3153-62, 1972 53. Frayn KN, Adnitt PI, Turner P: The use of human skeletal muscle in vitro for biochemical and pharmacological studies of glucose uptake. Clin Sci 44:55-62, 1973 54. WicklmayrM, Dietze G, MehnertH: Effect of phenformin on substrate metabolism of working muscle in maturity onset diabetes. Diabetologia 15:99-104, 1978 55. Purrello F, Gullo D, Buscema M, Pezzino V, Vigneri R, Goldfine ID: Metformin enhances certain insulin actions in cultured rat hepatoma cells. Diabetologia 31:385-89, 1988 56. Alengrin F, Grossi G, Canivet B, Dolais-Kitabgi J: Inhibitory effects of metformin on insulin and glucagon action in rat hepatocytes involve post-receptor alterations. Diabete Metab 13:591-97, 1987 57. Molnar IG, Davidson MB: Metformin has both insulinlike and anti-insulin effects in cultured muscle cells (Abstract). Clin Res 36:157A, 1988 58. Klip A, Walker D, Ransome KJ, Schroer DW, Lienhard GE: Identification of the glucose transporter in rat skeletal muscle. Arch Biochem Biophys 226:198-205, 1983 59. Birnbaum MJ: Identification of a novel gene encoding an insulin-responsive glucose transporter protein. Cell 57:30515, 1989 60. Charron MJ, Brosius FC, Alper SL, Lodish HF: A novel glucose transport protein expressed predominantly in insulin-responsive tissues. Proc Natl Acad Sci USA 86:253539, 1989 61. James DE, Strube M, Mueckler M: Molecular cloning and characterization of an insulin-regulatable glucose transporter. Nature (Lond) 338:83-87, 1989 62. Mueckler M, Caruso C, Baldwin SA, Panico M, Blech I, Morris HR, Allard WJ, Leinhard GE, Lodish HF: Sequence and structure of a human glucose transporter. Science 229:941-45, 1985 63. James DE, Brown R, Navarro J, Pilch P: Insulin-regulatable tissues express a unique insulin-sensitive transporter protein. Nature (Lond) 333:183-86, 1988 64. Douen AG, Cartee G, Ramlal T, Klip A: Exercise modu-

704

65.

66.

67.

68. 69.

70.

71.

72.

73.

74.

75.

76. 77.

78.

79.

lates the insulin-induced translocation of glucose transporters in rat skeletal muscle. FEBS Lett 261:256-60, 1990 Klip A, Ramlal T, Young DA, Holloszy JO: Insulin induced translocation of glucose transporters in rat hindlimb muscles. FEBS Lett 224:224-30, 1987 Simpson IA, Yver DR, Hissin PJ, Wardzala LJ, Karnieli I, Salans LB, Cushman SW: Insulin-stimulated translocation of glucose transporters in the isolated rat adipose cells: characterization of subcellular fractions. Biochim Biophys Acta 763:393-407, 1983 Sternlicht E, Barnard RJ, Grimditch GK: Mechanism of insulin action on glucose transport in rat skeletal muscle. Am J Physiol 254:E633-38, 1988 Cushman SW, Wardzala LJ: Potential mechanism of insulin action on glucose transport in the isolated rat adipose cell, j Biol Chem 225:4758-62, 1980 Matthaei S, Klein HH, Hamann A, Benecke H, Heni LJ, Greten H: Metformin (M) potentiates insulin (l)-induced translocation of glucose transporters from the low-density microsomes (LDM) to the plasma membrane in rat adipocytes (RA) (Abstract). Diabetes 37 (Suppl. 1):37A, 1988 Klip A, Li G, Logan WJ: Induction of sugar uptake response to insulin by serum depletion in fusing L6 myoblasts. Am ) Physiol 247:E291-96, 1984 Klip A, Logan WJ, Li G: Hexose transport in L6 muscle cells: kinetic properties and the number of 3H-cytochalasin B binding sites. Biochim Biophys Acta 687:265-80, 1982 Klip A, Sarabia V, Ramlal T, Bilan PJ, Leiter LA: Glucose transport in human muscle cells in culture: stimulation by insulin and metformin (Abstract). Diabetes 39 (Suppl. 1):82A, 1990 Ramlal T, Sarabia V, Bilan P, Klip A: Insulin-mediated translocation of glucose transporters from intracellular membranes to plasma membranes: sole mechanism of stimulation of glucose transport in L6 muscle cells. Biochim Biophys Res Commun 157:1329-35, 1988 Walker P, Ramlal T, Donovan JA, Doering TP, Sandra A, Klip A, Pessin JE: Molecular regulation by insulin and glucose of the glucose transport system in the L6 rat muscle cell line. / Biol Chem 246:6587-95, 1989 Walker P, Ramlal T, Sarabia V, Koivisto U-M, Bilan PJ, Pessin JE, Klip A: Glucose transport activity in L6 muscle cells is regulated by the coordinate control of subcellular glucose transporter distribution, biosynthesis, and mRNA transcription. / 6/0/ Chem. In press Sarabia V, Ramlal T, Klip A: Glucose uptake in human and animal muscle cells in culture. Biochem Cell Biol. In press Ramlal T, Rastogi S, Vranic M, Klip A: Decrease in glucose transporter number in skeletal muscle of mild diabetic (streptozotocin-treated) rats. Endocrinology 125:89097, 1989 Kahn BB, Cushman SW, Flier JS: Regulation of glucose transporter-specific mRNA levels in rat adipose cells with fasting and refeeding: implications for in vivo control of glucose transporter number. / Clin Invest 83:199-204, 1989 Garvey WT, Huecksteadt T, Birnbaum MJ: Pretranslational suppression of an insulin-responsive glucose transporter in rats with diabetes mellitus. Science 245:60-63, 1989

DIABETES CARE, VOL. 13, NO. 6, JUNE 1990