In patients with type 2 diabetes, a strong correlation between accumulation of intramuscular triclycerides (TGs) and insulin resistance has been found. The aim of the present study was to determine whether there is a causal relation between intramuscular TG accumulation and insulin sensitivity. Therefore, in mice with muscle-specific overexpression of human lipoprotein lipase (LPL) and control mice, muscle TG content was measured in combination with glucose uptake in vivo, under hyperinsulinemic-euglycemic conditions. Overexpression of LPL in muscle resulted in accumulation of TGs in skeletal muscle (85.5 ± 33.3 vs. 25.7 ± 23.1 μmol/g tissue in LPL and control mice, respectively; P < 0.05). During the hyperinsulinemic clamp study, there were no differences in plasma glucose, insulin, and FFA concentrations between the two groups. Moreover, whole-body, as well as skeletal muscle, insulin-mediated glucose uptake did not differ between LPL-overexpressing and wild-type mice. Surprisingly, whole-body glucose oxidation was decreased by ∼60% (P < 0.05), whereas nonoxidative glucose disposal was increased by ∼50% (P < 0.05) in LPL-overexpressing versus control mice. In conclusion, overexpression of human LPL in muscle increases intramuscular TG accumulation, but does not affect whole-body or muscle-specific insulin-mediated uptake, findings that argue against a simple causal relation between intramuscular TG content and insulin resistance.
Altered partitioning of triglycerides (TGs) between adipose tissue and nonadipose tissues, like muscle or liver, is hypothesized to be involved in the development of insulin resistance associated with obesity and type 2 diabetes (1,2,3). In patients with type 2 diabetes, there is a strong correlation between accumulation of intramuscular TGs and insulin resistance (4,5). A causal relation between muscle TG content and insulin resistance is supported by observations in mice (6) and rats (7). In lipoatrophic mice, hyperglycemia and hyperinsulinemia develop in association with excessive accumulation of TGs in muscle and liver (6). Reversal of this muscle TG accumulation by adipose tissue transplantation reverses the biochemical features of type 2 diabetes in these mice (8). Finally, these changes in TG content in lipoatrophic mice induce opposite changes in phosphatidylinositol 3-kinase activity in muscle, indicating a possible causal relationship between muscle TG content and insulin resistance (9). Rats fed a diet high in saturated fats and/or subjected to prolonged inhibition of fatty acid oxidation have also shown a correlation between insulin action and intramuscular TG content (7). Therefore, TG accumulation in tissues other than adipose tissue might provide a causal link in the association between obesity and diabetes (10,11).
Lipoprotein lipase (LPL) is most abundant in adipose tissue and muscle (12,13). LPL mediates lipolysis of TGs from TG-rich lipoproteins, resulting in the release of fatty acids from these lipoproteins (12,13). In this way, LPL controls entry of fatty acids into underlying tissues (adipose, muscle) (1,14), where the fatty acids can be stored in TGs (adipose tissue) or oxidized (muscle tissue) (1,15). It has been hypothesized that the relative levels of LPL within adipose and muscle tissue determine the partitioning of plasma TGs between these tissues (14,16,17). Consequently, modulation of muscle-specific LPL expression is a unique tool for manipulating muscle fatty acid uptake and, consequently, muscle TG content.
The aim of the present study was to evaluate the causality of the relation between muscle TG content and insulin sensitivity in a nonlipoatrophic mouse model. Therefore, we studied mice that overexpress LPL specifically in muscle (18,19) as a model for altered partitioning of TGs toward muscle tissue (20,21). We found that in these mice, muscle TG content was indeed significantly increased. However, the basal and insulin-mediated glucose uptake in these mice was not affected as compared with control mice, indicating the absence of a simple causal relationship between muscle TG stores and insulin sensitivity.
RESEARCH DESIGN AND METHODS
Male, muscle-specific human LPL-overexpressing C6/CBA mice on an endogenous murine LPL (LPL-TG or L2-MCK) background (18) and wild-type mice, ages 4–6 months, were taken from the breeding colony at the University of Graz, Austria. Mice were kept in a temperature- and humidity-controlled environment and had free access to standard laboratory diet and water. All animal experiments were approved by the Animal Ethics Committee of TNO-Prevention and Health, Leiden, the Netherlands.
Fasted plasma parameters.
After mice were fasted overnight, 150-μl blood samples were taken in paraoxon-coated capillaries via tail bleeding from the different genotypes. Plasma was collected by centrifugation, and total plasma glucose, free fatty acids (FFAs), TGs, and total cholesterol were determined via commercially available kits (Sigma, St. Louis, MO; Boehringer Mannheim, Mannheim, Germany; and Wako Chemicals, Neuss, Germany) according to the manufacturers’ instructions. Plasma insulin was measured by radioimmunoassay, using rat insulin standards (Sensitive Rat Insulin Assay; Linco Research, St. Charles, MO).
Glucose turnover studies.
After an overnight fast, animals were anesthetized (0.5 ml/kg Hypnorm [Janssen Pharmaceutica, Beerse, Belgium] and 12.5 mg/g midazolam [Genthon, Nijmegen, the Netherlands]), and an infusion needle was placed in one of the tail veins. Then basal glucose parameters were determined, during a 30-min period, by infusion of d-[3H]glucose alone (Amersham, Little Chalfont, U.K.), after a 2-h infusion to achieve steady-state levels. In other mice, a bolus of insulin (100 mU/kg Actrapid; Novo Nordisk, Chartres, France) and d-[3-3H]glucose (60 μCi/kg; Amersham) was given, and a hyperinsulinemic-euglycemic clamp was started with continuous infusion of insulin (3.5 mU · kg−1 · min−1) and d-[3-3H]glucose (0.8 μCi · kg−1 · min−1). Blood samples were taken every 15 min (tail bleeding) to monitor plasma glucose levels. A variable infusion of 12.5% d-glucose (in phosphate-buffered saline [PBS]) solution was started at time 0 and adjusted to maintain blood glucose at ∼5.0 mmol/l. During the last hour of the experiment, blood samples were taken every 20 min to determine [3H]glucose and 3H2O. After the last blood sample, mice were killed and their liver, cardiac muscle, skeletal muscle, and adipose tissue were taken for analysis. To estimate the insulin-stimulated glucose transport activity in individual tissues, 2-deoxy-d-[1-14C]glucose (2-DG; Amersham) was administered as a bolus (2 μCi) 45 min before the end of the clamps. Blood samples were taken 5, 25, and 45 min after bolus injection to determine plasma [3H]glucose, 3H2O, and 2-DG specific activities. After the last blood sample, mice were killed, and their liver, cardiac muscle, skeletal muscle, and adipose tissue were taken, directly frozen in liquid nitrogen, and kept at –80°C for analysis.
Tissue samples were homogenized (∼10% wet wt/vol) in PBS. To determine radioactivity of 3H in the lipid fraction of cardiac and skeletal muscle and liver, lipids were extracted by a modification of Bligh and Dyer’s method (22). In short, 200 μl of tissue homogenate was added to 1 ml of demi-water; then 4.5 ml of chloroform:methanol (1:2) was added and mixed thoroughly, after which 1.5 ml of demi-chloroform was added and mixed; and finally, 1.5 ml demi-water was added and mixed. After centrifugation, the chloroform layer was collected and used for determination of 3H radioactivity. Total TG and cholesterol content in these homogenates was also determined in these lipids, extracted as previously described (23). The radioactivity of 3H in muscle and liver glycogen was determined by digesting tissue samples in KOH and precipitating glycogen with ethanol, as described previously (24,25). For determination of tissue 2-DG uptake, tissues were homogenized (∼10%) in water, boiled, and subjected to an ion-exchange column to separate 2-DG-6-phosphate (2-DG-P) from 2-DG, as previously described (26,27).
Plasma glucose was measured by the glucose oxidase method (Beckman Glucose Analyzer II; Beckman Instruments, Fullerton, CA). Plasma insulin was measured by radioimmunoassay, using rat insulin standards (Sensitive Rat Insulin Assay, Linco). Total plasma [3H]glucose was determined in 10 μl plasma and in supernatants after trichloroacetic acid (20%) precipitation and water evaporation to eliminate tritiated water. The rates of glucose oxidation were determined as previously described by Koopmans et al. (28). Total plasma glucose, FFAs, TGs, and total cholesterol were determined via commercially available kits (Sigma, Boehringer Mannheim, and Wako), according to the manufacturers’ instructions.
Under steady-state conditions for plasma glucose concentrations, the rate of glucose disappearance equals the rate of glucose appearance. The latter was calculated as the ratio of the rate of infusion of [3-3H]glucose (dpm) and the steady-state plasma [3H]glucose specific activity (dpm/μmol glucose). The endogenous glucose production (EGP) was calculated as the difference between the rate of glucose disappearance and the infusion rate of d-glucose. Muscle- and adipose tissue−specific glucose uptake was calculated from tissue 2-DG-P content, which was expressed as percent of 2-DG of the dosage per gram of tissue, as previously described (25,29). Cardiac, skeletal, and hepatic muscle glycogen synthesis were calculated from 3H-label incorporation into tissue glycogen, as previously described (25,29). Adipose tissue and hepatic and muscle tissue lipid synthesis from [3H]glucose was calculated from 3H-label incorporation into the lipid extractable fraction, following the method of Bligh and Dyer (22), as previously described (29).
Results are presented as means ± SD for the number of animals indicated. Differences between experimental groups were determined by the Mann-Whitney U test (30). The level of statistical significance of the differences was set at P < 0.05. Analyses were performed using SPSS 10.0 for Windows software (SPSS, Chicago).
Plasma and tissue parameters.
The body weight of LPL-overexpressing mice and control animals did not differ significantly (28.6 ± 2.7 vs. 30.6 ± 2.1 g, respectively) (Table 1). Plasma TG and cholesterol concentrations were reduced by ∼85 and 50%, respectively (P < 0.05) (Table 1). FFA concentrations were found to be similar between the two genotypes in the fasted state. Fasted insulin levels were slightly increased in the LPL-overexpressing mice in comparison to wild-type mice, whereas no differences were detected in plasma glucose levels. TG content of cardiac and skeletal muscle of LPL-overexpressing mice was significantly (P < 0.05) increased, whereas hepatic TG content was significantly decreased (P < 0.05) compared with control mice (Fig. 1). There were no differences in tissue cholesterol content between the different genotypes (data not shown).
Glucose turnover studies.
As measured by continuous [3H]glucose infusions, there were no differences in basal whole-body glucose uptake, whole-body glucose oxidation, whole-body nonoxidative glucose disposal, or EGP between LPL-overexpressing and control mice (Fig. 2A). We also performed hyperinsulinemic-euglycemic clamp studies in both mouse lines after an overnight fast. We established stable glucose levels at ∼5 mmol/l and insulin levels at ∼10-fold higher than under fasted conditions in wild-type mice, with no differences in FFA concentrations during the clamp between both genotypes (Table 2). Figure 2B shows that there were no differences in whole-body glucose uptake between both genotypes. In contrast, whole-body glucose oxidation was decreased in the LPL-overexpressing mice compared with control mice (P < 0.05); as a consequence, whole-body nonoxidative glucose disposal was increased in LPL-overexpressing mice. The ability of insulin to suppress EGP was enhanced in LPL-overexpressing mice compared with wild-type mice (75 vs. 40%, respectively; P < 0.05) (Fig. 2B). Thus the liver seemed have increased insulin sensitivity because of the lower TG content (Fig. 1).
Tissue-specific glucose uptake and glycogen synthesis.
Glucose uptake in muscle and adipose tissue was determined under hyperinsulinemic-euglycemic conditions, 45 min after a bolus of 2-DG (Fig. 3). Glucose uptake determined by 2-DG uptake was significantly increased in cardiac muscle and adipose tissue of LPL-overexpressing mice compared with controls. However, there were no differences in glucose uptake in skeletal muscle between the different genotypes. Glycogen synthesis determined by [3H]glucose incorporation into glycogen during the hyperinsulinemic clamp was not significantly different in cardiac and skeletal muscle or liver between the two genotypes. In contrast, incorporation of 3H-label into the lipid extractable fraction was significantly higher in liver of LPL-overexpressing mice compared with controls (P < 0.05), whereas skeletal and cardiac muscle and adipose tissue [3H]glucose incorporation into the lipid fraction did not differ between the two genotypes.
Type 2 diabetes is a complex disease that is associated with strong impairment of insulin action with respect to glucose uptake—that is, insulin resistance. Increased intramuscular TG content has been found to be inversely associated with insulin action in type 2 diabetic patients (4,5,9). In lipoatrophic mice, a causal relationship between muscle TG accumulation and insulin resistance has also been found (6). The present study was performed to address the question of whether increased muscular TG storage is also causally related to insulin resistance in a nonobese, nonlipoatrophic mouse model. To increase TG content in skeletal muscle, we used mice with muscle-specific LPL overexpression, with 2.5-fold higher total postheparin lipolytic activity and twofold higher skeletal muscle lipoprotein mass (18). Although these mice exhibited increased TG content in skeletal muscle, neither whole-body nor muscle-specific insulin-mediated glucose uptake was altered.
LPL-overexpressing mice had normal body mass characteristics compared with control mice and were not obese (16,18); however, muscle TG content was significantly higher in LPL-overexpressing mice compared with controls, a finding that agrees with data published by Weinstock et al. (16). Together with the observation that hepatic TG content was strongly decreased, these data support the role of LPL in tissue-specific partitioning of TG from TG-rich lipoproteins toward underlying tissues (20,21). Although intramuscular TG content is usually increased in the presence of obesity, high-fat feeding, and/or lipoatrophy (4,6), the muscle-specific LPL-overexpressing mouse represents a useful model for examining the pathophysiological implication of increased intramuscular TG content per se without a background of obesity or lipoatrophy.
Under basal, fasted conditions, there were only minor differences in plasma insulin levels between the LPL-overexpressing mice and controls, with no effects on plasma glucose levels. Furthermore, whole-body glucose uptake, oxidation and nonoxidative glucose disposal, and EGP were not different in the LPL-overexpressing mice compared with control mice. These results indicate that under fasted conditions, increased intramuscular TG content has no effect on glucose homeostasis. In addition, under hyperinsulinemic-euglycemic conditions, no differences were observed between the two genotypes on insulin-stimulated whole-body glucose uptake. These results are in line with those of Brüning et al. (31), who found no alterations in glucose uptake in muscle-specific insulin receptor knockout mice, which represent the ultimate model of extreme muscle insulin resistance.
On the other hand, under hyperinsulinemic-euglycemic conditions, EGP was found to be lower in LPL-overexpressing mice compared with controls. This observation suggests increased hepatic sensitivity with respect to the suppressive effects of insulin on hepatic glucose production. We postulate that the decrease in hepatic TG content in these mice might upregulate insulin sensitivity and thereby reduce hepatic glucose production under hyperinsulinemic conditions. Hepatic fat content correlates with impaired suppression of EGP by insulin, as described by Gupta et al. (32) and recently shown by Kim et al. (33). Liver-specific LPL overexpression leads to TG accumulation in the liver, with no obvious effects on total body insulin-stimulated glucose uptake. Furthermore, suppression of hepatic glucose production by insulin was markedly impaired in the liver LPL-overexpressing mice, as was the amount of phosphatidylinositol 3-kinase activity.
Although insulin-mediated glucose uptake was not affected in LPL-overexpressing mice, intracellular handling of glucose was altered. Whole-body glucose oxidation was strongly decreased, whereas nonoxidative glucose disposal (glycogen and lipid synthesis) was increased in LPL-overexpressing mice compared with controls. Interestingly, these effects of muscle-specific LPL overexpression on insulin-stimulated glucose metabolism are comparable to the effects of short-term, high-fat feeding in humans (34), in whom whole-body glucose uptake was not altered, but glucose oxidation was depressed and nonoxidative glucose disposal was significantly increased.
Increased plasma FFA concentrations are inversely correlated with insulin action (35,36). Lowering of plasma FFA levels in normal and obese subjects has been shown to improve insulin action (37). In our study, no differences in plasma FFA levels were measured, neither under basal, fasted conditions nor during the hyperinsulinemic clamp study. Therefore, our study was well controlled for plasma FFA levels, and the results do not suggest impaired insulin action on lipolysis in LPL-transgenic mice.
The findings of the current study seem to contradict the inverse correlation between muscle TG content and insulin-stimulated whole-body glucose uptake found in humans (4,5,9), rats (7), and mice (6,8). However, there are distinct differences between the metabolic conditions present in our study and those of other investigators. In the human studies, the correlations were found in obese type 2 patients (4,5,9). In the rat study, the animals were treated with either a high–saturated fat diet or prolonged inhibition of fatty acid oxidation (7), and in the mouse studies, the animals were lipoatrophic (6,8). Our model represents a mouse model with increased fatty acid uptake specific to muscle tissue but not subjected to long-term high-fat feeding, inhibition of fatty acid oxidation, or altered body fat content. Consequently, it must be considered that muscle TG content per se—in the absence of other alterations—does not control insulin sensitivity, in accordance with the observations in muscle-specific insulin receptor knockout mice (31).
We found that increased intramuscular TG content as a result of muscle-specific overexpression of LPL does not seem to have an effect on insulin-stimulated uptake of glucose in skeletal muscle, as determined by 2-DG uptake. On the other hand, cardiac muscle and adipose tissue glucose uptake stimulated by insulin was significantly higher in LPL-overexpressing mice compared with controls. Thus cardiac muscle and adipose tissue seem to be more sensitive to insulin with respect to glucose uptake in LPL-overexpressing mice compared with wild-type mice. An explanation for the increased insulin sensitivity of adipose tissue may be related to the small size of adipocytes found in these LPL-overexpressing mice (16), as smaller adipose cell size is related to improved insulin sensitivity (38). Additional support for our observations is the fact that muscle-specific LPL-overexpressing mice are protected from diet-induced obesity (16,39). Furthermore, increased de novo TG synthesis in these mice (16) could account for increased glucose uptake, as was found in vitro by Ruan and Pownall (40) in 3T3-L1 adipocytes overexpressing acylglycerol-3-phosphate acyltransferase (AGAT; second step in TG synthesis). Overexpression of AGAT leads to increased glucose uptake and conversion into cellular lipids in 3T3-L1 adipocytes. Because adipose tissue does not contain glycerol kinase, it depends on glucose uptake and conversion into glycerol-3-phosphate. We cannot explain the mechanism of increased cardiac glucose uptake by experimental evidence. We hypothesize, however, that in muscle-specific LPL-transgenic mice, increased amounts of fatty acids need increased glycerol for greater TG synthesis. Because cardiac muscle does not contains glycerol kinase, glycerol must be provided via glucose uptake and subsequent conversion into glycerol-3-phosphate. In vitro evidence in isolated myotubes has shown that increased TG synthesis by overexpressing AGAT increases glucose conversion to cellular lipids, with a concomitant decrease in glycogen formation (40). Our data seem to point in the same direction, although increased conversion in lipids and decreased glycogen formation (Fig. 4) were not statistically different in our study because of the relative small numbers. Why this phenomenon is not found in skeletal muscle needs to be elucidated, but the reason may be related to the relatively low but functional glycerol kinase activity found in skeletal muscle (41).
To investigate the intracellular handling of glucose, we examined the incorporation of glucose in glycogen and lipid in the respective tissue homogenates. No significant differences were seen in skeletal and cardiac muscle and hepatic glycogen synthesis from [3H]glucose. Youn and Buchanan (24) found that decreased glucose oxidation is linked to increased glycogen synthesis, without any effect of tissue-specific glucose uptake in rats during fasting. The overnight fasting period, which induced complete depletion of the glycogen pool (24), and/or the physiological hyperinsulinemia during the clamp analyses were sufficient to stimulate glycogen synthesis in both genotypes, showing no differences between the LPL-overexpressing and control mice. On the other hand, lipid synthesis from [3H]glucose in liver, but not in skeletal and cardiac muscle and adipose tissue, increased in LPL-overexpressing mice. The observed increased lipid synthesis from glucose in the liver may have been related to the decreased hepatic TG content in which the liver increased de novo synthesis of TGs.
The question arises how our results can be compared with recent reports by Ferreira et al. (42) and Kim et al. (33). Various obvious differences in experimental conditions can be found between their studies and ours, such as genetic background, dietary fat content, body weight, muscle and liver TG content, and insulin levels during the hyperinsulinemic-euglycemic clamp. Differences in the genetic background of the mouse model could be responsible for several metabolic and physiological differences between different mouse strains. However, it remains to be elucidated whether this explains the observed differences among the three studies. One striking difference with our study is that Kim et al. (33) did not find a decreased hepatic TG content in LPL-overexpressing mice; they did find that liver fat content is inversely correlated with the inhibitory action of insulin on hepatic glucose production (33). Ferreira et al. (42) did not describe any data on liver TG content. Therefore, additional studies are required to resolve whether differences in hepatic TG accumulation contribute to the observed differences among the three studies.
In summary, specific overexpression of LPL in skeletal muscle alters TG disposal in mice. Muscle TG content was increased in these LPL-transgenic mice, whereas liver TG content was decreased. This altered TG partitioning did not affect basal or insulin-stimulated whole-body or muscle-specific glucose uptake. Therefore, the results of this study argue against a simple causal relation between intramuscular TG accumulation and insulin resistance.
The research described in this paper is supported by the Netherlands Organization for Scientific Research (NWO Grant 903-39-194), the Netherlands Heart Foundation (NHS Grant 97.067), and the Netherlands Diabetes Foundation.
Address correspondence and reprint requests to Dr. Peter J. Voshol, TNO-Prevention and Health, Division VBO, Zernikedreef 9, NL-2333 CK Leiden, The Netherlands. E-mail: email@example.com.
Received for publication 14 March 2001 and accepted in revised form 10 August 2001.
AGAT, acylglycerol-3-phosphate acyltransferase; 2-DG, 2-deoxy-d-[1-14C]glucose; 2-DG-P, 2-DG-6-phosphate; EGP, endogenous glucose production; FFA, free fatty acid; LPL, lipoprotein lipase; PBS, phosphate-buffered saline; TG, triglyceride.