Adipose-derived resistin and gut-derived resistin-like molecule–β selectively impair insulin action on glucose production (original) (raw)
Conscious rats were given an initial intra-arterial bolus of either resistin, RELMβ, or vehicle, followed by an intra-arterial infusion of the same material at a constant rate (Figure 1a). Resistin was infused at either 5 μg/h (low) or 25 μg/h (high), which promptly increased its serum concentration approximately two- to fivefold or ten- to 15-fold, respectively (Figure 1b). RELMβ or vehicle was infused at 20 μg/h. Prior to the start of the infusion protocols, there were no differences in mean body weight, plasma insulin, or glucose concentration among the four groups of rats (Table 1).
General characteristics of rats receiving resistin, RELMβ, or vehicle before and during pancreatic insulin clamp studies
We used a modification of the pancreatic clamp technique designed to accurately estimate the impact of recombinant resistin and RELMβ on glucose fluxes in the presence of either low or physiologically high insulin concentrations (Figure 1a, Table 1). To prevent the endogenous secretion of glucoregulatory hormones such as insulin and glucagon, we administered somatostatin, which effectively blunts secretion of these hormones. Thus, without endogenous insulin, a known infusion of insulin could be administered at a constant rate, periodically adjusted to maintain plasma glucose concentrations at a physiologic level (∼8.5 mM). This technique effectively minimized the confounding effects of hypoglycemia and its attendant counterregulatory responses on metabolic parameters. In fact, the plasma glucagon (53 ± 7, 56 ± 5, 53 ± 2, and 48 ± 2 ng/ml) and corticosterone (178 ± 13, 196 ± 36, 156 ± 40, and 142 ± 29 ng/ml) concentrations during the insulin clamp studies were similar in rats receiving vehicle, low- and high-dose resistin, and RELMβ, respectively. Additionally, the use of radioactive glucose tracers (16, 17) allowed us to measure glucose fluxes in vivo.
The effects of resistin, RELMβ, and vehicle were first compared in the presence of similar steady-state normoglycemia and individualized plasma insulin concentrations (Table 1) in conscious rats. Under these conditions, the levels of circulating insulin necessary to maintain the plasma glucose concentrations at basal levels provide a measure of whole-body insulin effectiveness (Table 1). During the intravenous infusion of resistin and RELMβ, insulin had to be infused at a higher rate than during vehicle infusions (Figure 2a). Furthermore, the rate of glucose production was markedly elevated in both resistin- and RELMβ-infused rats compared with vehicle-infused rats (Figure 2b). Of note, this occurred despite higher circulating insulin levels (Table 1). Therefore, the presence of increased glucose production in the face of elevated insulin concentrations provides a first demonstration of the potent inhibitory effects of these hormones on hepatic insulin sensitivity (Figure 2c). By contrast, neither resistin nor RELMβ significantly altered the effect of insulin on peripheral glucose uptake (Figure 2d).
Effect of resistin and RELMβ during pancreatic clamp studies. The measurements were obtained while plasma glucose concentration was maintained at approximately 8.5 mM under steady-state conditions. (a) Rates of insulin infusion (IIR; mU/kg/min) required to maintain the plasma glucose concentrations at basal levels. (b) Effect of resistin and RELMβ on the rate of glucose production (GP; mg/kg/min). (c) Effect of resistin and RELMβ on hepatic insulin sensitivity (HIS). To take into account the differences in circulating plasma insulin levels during the pancreatic clamp studies, an index of HIS was calculated as GP–1/insulin (μU/ml) × 100. (d) Effect of resistin and RELMβ on peripheral insulin sensitivity (PIS). To take into account the differences in circulating plasma insulin levels during the pancreatic clamp studies, an index of peripheral insulin sensitivity was calculated as Rd (mg/kg/min)/insulin (μU/ml). Rd, rate of glucose disappearance. *P < 0.05 vs. vehicle.
We next examined the effect of increased circulating insulin concentrations on the rates of glucose infusion and tissue glucose uptake (Figure 3, a and b). During this period, insulin levels were elevated in all groups by the same amount (initial infusion rate + 3 mU/kg/min). Importantly, all measurements were performed during the final 2 hours of the 3-hour hyperinsulinemic clamp study (14–17), a time after steady-state conditions for plasma glucose and insulin concentrations, glucose specific activity, and rates of glucose infusion were achieved. The rates of exogenous glucose infusion (GIR in Figure 3a) required to maintain the target plasma glucose concentration during physiologically high insulin were decreased by 32%, 26%, and 25%, respectively, during the infusions of low-dose resistin, high-dose resistin, and RELMβ compared with vehicle-infused rats. However, the infusion of recombinant resistin (low and high dose) or RELMβ did not significantly alter the rates of glucose uptake (respectively, 22.1 ± 1.3, 23.9 ± 1.2, and 23.9 ± 1.4 mg/kg/min) compared with vehicle (23.7 ± 1.6 mg/kg/min) (Figure 3b). By contrast, in the presence of similar plasma insulin concentrations (∼100 μU/ml), glucose production was higher during the infusion of resistin and RELMβ than during vehicle infusion (Figure 3c). In fact, this marked impairment in the inhibitory action of insulin on glucose production (Figure 3d) completely accounted for the decreased rate of glucose infusion during the clamp studies. These results indicate that short-term intra-arterial infusion of resistin stimulates glucose production in the presence of either basal replacement insulin concentrations or moderately elevated hyperinsulinemia. Thus, regulation of hepatic glucose production is the main mechanism by which resistin acutely regulates glucose tolerance. Of interest, the infusion of recombinant RELMβ generated metabolic effects qualitatively similar to those of resistin. Glucose production represents the net contribution of glucosyl units derived from gluconeogenesis and glycogenolysis. However, a portion of glucose entering the liver via phosphorylation of plasma glucose is also a substrate for dephosphorylation via glucose-6-phosphatase. This futile cycle between glucokinase and glucose-6-phosphatase is commonly named glucose cycling and accounts for the difference between the total glucose output (flux through glucose-6-phosphatase) and glucose production.
Effect of resistin and RELMβ on glucose disposal and production during insulin clamp studies. The measurements were obtained while plasma glucose concentration was maintained at approximately 7 mM under steady-state conditions. (a) Effect of resistin and RELMβ on the rate of glucose infusion (GIR; mg/kg/min). (b) Effect of resistin and RELMβ on the rate of glucose disappearance (Rd; mg/kg/min). (c) Effect of resistin and RELMβ on the rate of glucose production (GP; mg/kg/min). (d) Effect of resistin and RELMβ on hepatic insulin sensitivity (HIS; μU/ml). To take into account differences between animals in circulating plasma insulin levels during the insulin clamp studies, the same index of hepatic insulin sensitivity described in the Figure 2c legend was used. *P < 0.05 vs. vehicle.
To further delineate the mechanisms responsible for the effect of recombinant resistin and RELMβ infusion on hepatic glucose production, we estimated the in vivo flux through glucose-6-phosphatase and the relative contributions of glucose cycling, gluconeogenesis, and glycogenolysis to glucose output (Table 2 and Figure 4). Figure 4 depicts the effect of resistin and RELMβ on the rates of hepatic glucose fluxes during the pancreatic insulin clamp procedure. In the presence of similar plasma insulin concentrations, the rate of glucose production (shown in Figure 3) was increased by resistin and by RELMβ. Table 2 displays the [14C]PEP, [3H]UDP-glucose, [14C]UDP-glucose, and [3H-3]glucose specific activities used to calculate the contribution of PEP and plasma glucose (% indirect and % direct in Table 2) to the hepatic glucose-6-phosphate pool. Data obtained using the specific activity of UDP-galactose (not shown) confirmed the results obtained with UDP-glucose. These data allowed us to estimate the in vivo fluxes through glucose-6-phosphatase and the rates of glucose cycling, PEP gluconeogenesis, and glycogenolysis in all groups. As shown in Figure 4, the flux through glucose-6-phosphatase was markedly increased by resistin and RELMβ in parallel to the effect on glucose production. Consistent with this marked increase in overall glucose output, the rate of glucose cycling was also increased in resistin- and RELMβ-infused rats compared with vehicle-infused animals. Thus short-term infusion of the adipose-derived protein resistin and of the gut-derived protein RELMβ lead to marked stimulation of in vivo glucose-6-phosphatase flux. The latter increase was accounted for by marked increases in both gluconeogenesis and glycogen breakdown.
Effect of resistin and RELMβ on hepatic glucose fluxes (mg/kg/min) during insulin clamp studies. (a) Effect of resistin and RELMβ on total glucose output (in vivo flux through glucose-6-phosphatase). (b) Effect of resistin and RELMβ on the rate of glucose cycling. (c) Effect of resistin and RELMβ on the rate of PEP gluconeogenesis. (d) Effect of resistin and RELMβ on the rate of glycogenolysis. *P < 0.05 vs. vehicle.
Effect of resistin and RELMβ on the direct and “indirect” pathways of hepatic UDP-glucose formation