Hypothalamic resistin induces hepatic insulin resistance (original) (raw)
To test our hypothesis, we first examined whether a primary increase in the central or hypothalamic availability of resistin per se is sufficient to stimulate GP and generate hepatic insulin resistance. We next examined the systemic effects of hypothalamic resistin action by inducing a physiological increase in the circulating levels of resistin while selectively negating resistin’s central effects through the infusion of a specific anti-resistin Ab.
Central administration of recombinant resistin induces hepatic insulin resistance. To investigate the central effects of resistin independent of its pleiotropic systemic actions, we examined whether the central administration of recombinant resistin acutely regulates hepatic insulin action. To this end, 3 groups of conscious rats received icv infusions of recombinant mouse resistin, a biologically active cysteine mutant of resistin (cys), or vehicle (Figure 1B and Supplemental Table 1; supplemental material available online with this article; doi:10.1172/JCI30440DS1). The cys used in these studies has been shown to potently stimulate fluxes in liver glucose (42). The icv infusion of resistin or cys moderately increased the plasma glucose and insulin levels (Supplemental Table 1). Insulin action was assessed by a combination of icv infusions using the hyperinsulinemic-euglycemic clamp technique (Figure 1B). As expected, in the presence of the approximately 3-fold increase in the circulating insulin levels (using the clamp technique; Supplemental Table 1), glucose had to be infused in order to maintain euglycemia in icv vehicle–infused animals (15.6 ± 0.4 mg/kg/min). By contrast, following icv infusions of resistin or cys, glucose was infused at a much lower rate of approximately 10 mg/kg/min in order to prevent hypoglycemia (Figure 1C). It should be noted that the plasma insulin and glucagon concentrations (data not shown) were similar in the 3 groups (Supplemental Table 1). Thus, central administration of resistin diminished hepatic insulin action on glucose homeostasis.
We next examined the potential mechanism(s) by which icv resistin impaired whole-body insulin action. We assessed glucose kinetics by tracer dilution methodology in order to establish whether the decreased rate of glucose infusion induced by icv resistin or cys was due to either a decrease in glucose uptake or a stimulation of endogenous GP. The rates of glucose utilization (Figure 1D) were not significantly affected by either icv resistin (19.6 ± 0.5 mg/kg/min) or cys (20.4 ± 1.0 mg/kg/min) compared with vehicle (19.8 ± 0.3 mg/kg/min). Conversely, icv resistin and cys markedly increased the rate of GP and therefore impaired the suppressive effect of circulating hyperinsulinemia on the rate of GP (Figure 1, E and F). icv vehicle infusion was associated with a marked (~75%) decline in the rate of GP from basal levels (Figure 1F). However, icv resistin and cys lead to a modest (~40%) decline in the rate of hepatic GP compared with basal rates (Figure 1F). These changes in glucose output occurred in the presence of similar plasma insulin levels (~75 μU/ml; Supplemental Table 1) and completely accounted for the effect of icv resistin or cys on whole-body glucose metabolism. Thus, a primary increase in central resistin action can stimulate GP in the presence of physiological levels of circulating insulin.
These potent metabolic effects of resistin could be mediated by its action anywhere within the CNS. To gain insight into the anatomical localization of these effects, we next infused a 20-fold lower dose of resistin bilaterally within the MBH. MBH infusion of resistin entirely reproduced the effects of icv resistin on glucose infusion and GP (Figure 1, C–F). Thus, increased MBH availability of resistin is sufficient to increase GP.
GP represents the net contribution of glucosyl units derived from gluconeogenesis and glycogenolysis. However, a portion of glucose entering the liver via the direct phosphorylation of glucose is also a substrate for dephosphorylation via glucose-6-phosphatase (G6Pase) creating a futile cycle, termed “glucose cycling”. In order to further delineate the mechanisms by which central administration of resistin modulates glucose homeostasis, we estimated the in vivo flux of glucose through G6Pase and the contribution of gluconeogenesis and glycogenolysis relative to glucose output. icv resistin, icv cys, and MBH resistin markedly increased the glucose flux through G6Pase (Figure 2A) and glucose cycling (Figure 2B) in parallel with their effects on GP (Figure 1E). Importantly, the increase in GP was largely accounted for by a marked induction of glycogenolysis, while the rate of gluconeogenesis was only modestly increased (Figure 2, C and D). Real-time PCR analyses revealed that icv resistin, icv cys, and MBH resistin did not alter the hepatic mRNA levels for the rate-limiting enzymes for gluconeogenesis, G6Pase, and phosphoenolpyruvate (PEP) carboxykinase (PEPCK) (Figure 2E) On the other hand, both icv and MBH each resistin moderately decreased the hepatic levels of phosphorylated AMP-activated protein kinase α (p-AMPKα) as analyzed by immunoblot (Figure 2F). Thus, direct enhancement of resistin action within the third cerebral ventricle or MBH per se was sufficient to recapitulate the action of systemic resistin on the in vivo fluxes through G6Pase, rates of glycogenolysis, and on the hepatic phosphorylation of AMPK.
Central resistin increased hepatic glucoses fluxes predominantly via glycogenolytic pathways. (A and B) icv or IH infusion of resistin (black bars) resulted in 2- to 3-fold increases in hepatic flux through G6Pase (A) and glucose cycling (B) compared with vehicle (white bars). (C and D) Increased hepatic glucose fluxes were mainly accounted for by an increased rate of glycogenolysis (D) rather than gluconeogenesis (C). (E) Quantitative real-time RT-PCR revealed that central resistin had no effect on the hepatic expression of the key gluconeogenic enzymes PEPCK and G6Pase. (F) Decreased levels of p-AMPKα were apparent in the livers of resistin-treated animals as analyzed by immunoblot. *P < 0.05 compared with vehicle.
As it is now widely accepted that increases in adiposity lead to an increased physiological inflammatory milieu, we examined the effect of central resistin on the expression of key inflammatory regulators in the liver (43–46). Strikingly, as determined by quantitative real-time RT-PCR, central resistin administration leads to an approximately 3- to 4-fold increase in the hepatic expression of SOCS-3 and TNF-α mRNA and an even greater increase in the expression of IL-6 at the end of the clamps (Figure 3A). However, resistin infusion did not alter the expression of IKK-β or the other hepatic gluco- and lipo-regulatory genes, PPARγ coactivator 1α (PGC-1α), fatty acid synthase (FAS), acetyl-CoA carboxylase 1 (ACC1), and steroyl-CoA decarboxylase 1 (SCD1) (Figure 3A). Mirroring the changes in gene expression and hepatic insulin action on glucose fluxes, immunoblot analysis revealed that hepatic SOCS-3 protein levels were similarly elevated. Phosphorylated glycogen synthase kinase 3β (p-GSK3β) was decreased, and the phosphorylation state of IκB-α, a downstream target of IKK-β, was unchanged following central resistin administration (Figure 3, B and C). Finally, hepatic levels of STAT3 protein were dramatically lower in animals infused with resistin either icv or in the MBH (Figure 3, B and C).
Real-time RT-PCR and Western immunoblot analysis of hepatic inflammation and insulin signaling. (A) icv resistin administration (black bars) increased hepatic gene expression of proinflammatory mediators SOCS-3, IL-6, and TNF-α compared with vehicle-treated animals (white bars) but had no change on IKK-β, FAS, ACC1, SCD1, and PPARγ. (B and C) Significant decreases in total Stat3 protein and p-GSK3β were detected in the livers following central resistin administration (black bars), with a reciprocal but converse elevation in SOCS-3 when compared with controls (white bars). The levels of p–IκB-α remained unchanged (C). *P < 0.05 compared with vehicle.
These gain-of-function experiments suggest that modulation of resistin action within the MBH is sufficient to markedly affect liver glucose homeostasis and selective inflammatory markers. However, the physiological role of hypothalamic resistin in the regulation of GP cannot be determined by these gain-of-function studies and thus requires loss-of-function experiments in the presence of physiological increases in the circulating resistin levels.
Antagonism of hypothalamic resistin action. Is the action of resistin within the hypothalamus required for the effect of circulating hyperresistinemia on GP? To address this question, we developed an experimental strategy aimed at antagonizing the central action of resistin in the presence of systemic elevations in plasma resistin concentrations. We reasoned that if hypothalamic resistin action is required for stimulation of GP by hyperresistinemia, negating the MBH effects of resistin should diminish the ability of resistin to enhance GP in the presence of a physiological elevation in the plasma resistin levels. To this end, we used Rs Abs to selectively impede the action of the infused recombinant mouse resistin within the MBH (Figure 4A). A subset of i.v. resistin–infused animals also received a MBH infusion of a control anti-human resistin Ab (Con Ab), with which mouse resistin fails to maintain antigenicity. This group served as a control for nonspecific effects that central infusion of Ab may have had on GP. We performed pancreatic-hyperinsulinemic clamp studies (Figure 4A) designed to increase the plasma insulin concentrations by approximately 3-fold above basal levels. Such an increase is normally associated with a significant stimulation of peripheral glucose disposal and a marked decrease in GP and therefore allows one to examine the effects of insulin on both the liver and peripheral tissues. The rate of glucose utilization during the clamp studies was not significantly affected by peripheral (i.v.) infusion of resistin regardless of the MBH administration of resistin antibodies (Figure 4B). However, the inhibition of GP during the insulin clamp studies was improved by the MBH administration of Rs Ab but not Con Ab (Figure 4C). Thus, antagonism of MBH resistin action diminishes the stimulatory effect of systemic resistin on hepatic GP in the presence of physiological hyperinsulinemia.
The effects of increased levels of circulating plasma resistin are attenuated in animals treated with IH anti-resistin antibodies. (A) Experimental design for hyperinsulinemic-euglycemic clamp studies. IH infusion of Con Ab or Rs Ab (0.33 μl bolus followed 0.5 μl/h infusion) was initiated at 0 minutes and continued throughout the 360-minute course of study. Resistin (30 μg total dose) was infused i.v. at a constant rate starting at 60 minutes. The remainder of the study was completed as described in Figure 1. (B) Rates of glucose uptake during the insulin clamp studies were unaffected in all groups. (C) The rate of endogenous GP for i.v. resistin–infused animals also receiving an IH infusion of Con Ab was greatly increased compared with controls but was attenuated in animals receiving the IH infusion of Rs Ab. (D) Changes in the percentage suppression of endogenous GP in animals receiving i.v. resistin and IH Ab infusions. *P < 0.05 compared with vehicle; #P < 0.05 compared with Con Ab.
Similarly, MBH infusion of Rs Ab attenuated the marked elevations in glucose flux through G6Pase and total glucose cycling that were induced by peripheral resistin infusion (Figure 5, A and B). Again these changes in hepatic glucose fluxes were accounted for mainly by changes in the glycogenolysis, although the peripheral infusion of resistin in the Con Ab cohort also led to a modest yet nonsignificant increase in gluconeogenesis (Figure 5, C and D). That a hypothalamus-mediated effect of resistin on hepatic glucose fluxes predominantly occurs via glycogenolytic pathways (Figure 2D) is reinforced by the attenuation of this effect upon central blockade of resistin action (Figure 5D). In contrast to the absence of a centrally mediated effect of resistin on the hepatic expression of key glucoregulatory genes, the approximately 3- to 4-fold increase in G6Pase expression brought about by peripheral (i.v.) resistin infusion appeared to be a direct result of resistin in the liver, as there was no attenuation by MBH Rs Ab (Figure 5E). Hepatic PEPCK mRNA expression remained unchanged. In accordance with the changes in GP, MBH Rs Ab attenuated the decrease in p-AMPKα associated with i.v. resistin infusion (Figure 5F).
Central blockade of resistin after peripheral (i.v.) resistin infusion attenuates increases in hepatic glucose fluxes. (A and B) IH administration of Rs Ab decreased hepatic flux through G6Pase (A) and glucose cycling (B) in animals receiving a constant i.v. infusion of resistin, compared with animals receiving Con Ab, although it failed to completely normalize these to vehicle-treated levels. (C and D) Alterations of hepatic glucose fluxes in these animals were mirrored by concomitant changes in glycogenolysis (D) but not gluconeogenesis (C). (E) Peripheral (i.v.) infusion of resistin increased the hepatic expression of G6Pase message levels, which were not attenuated by central Rs Ab administration as assayed by real-time RT-PCR. Parallel changes in PEPCK expression levels were absent. (F) Depressed levels of p-AMPKα and subsequent, albeit minor, attenuation of this effect by central Rs Ab were apparent in the livers of i.v. resistin–treated animals compared with vehicle as analyzed by immunoblot. *P < 0.05 compared with vehicle; #P < 0.05 compared with Con Ab.
Strikingly, hepatic gene expression analysis revealed that the increases in IL-6 and TNF-α mRNAs following resistin infusion were entirely due to centrally mediated effects of resistin, as MBH Rs Ab completely blocked the induction of these transcripts despite the ongoing peripheral hyperresistinemia (Figure 6A). The elevation of liver SOCS-3 expression, however, appears to be mediated both directly and indirectly, as intrahypothalamic (IH) Rs Ab was unable to decrease the elevations in mRNA or protein levels in the presence of i.v. resistin infusion (Figure 6, A–C). Resistin failed to demonstrate any direct or central action on the regulation of hepatic IKK-β, PGC-1α, FAS, ACC1, and SCD1 (Figure 6A). Blockade of central resistin action produced no alteration of GSK3β phosphorylation or total STAT3 levels, despite the increase in hepatic insulin action, thus suggesting potential effects regulated by both direct and indirect centrally mediated resistin action on these parameters (Figure 6, B and C).
The increased expression of inflammatory genes in the liver following peripheral (i.v.) resistin infusion is ameliorated by central Rs Ab. (A) The increase in hepatic gene expression of key inflammatory mediators TNF-α and IL-6 in animals receiving a peripheral resistin infusion (30 μg total over 5 h) with central (IH) Con Ab (black bar) compared with vehicle-infused animals (white bars) was attenuated in animals receiving the same peripheral resistin infusion (30 μg total over 5 h) but with IH Rs Ab treatment (gray bars), blockade of central resistin signaling was unable to repress SOCS-3 expression. No change on IKK-β, FAS, ACCα, SCD1, or PPARγ gene expression was noticed following peripheral resistin infusion in either group. (B and C) Parallel to changes in hepatic gene expression, SOCS-3 protein levels as analyzed by Western blot were elevated following i.v. resistin infusion, with no amelioration of this effect by central Rs Ab. The decrease of hepatic STAT3 protein and p-GSK3β levels also remained unaffected by central resistin blockade. No changes in p–IκB-α were detected following peripheral resistin administration in either cohort (black and gray bars) compared with vehicle (white bars). *P < 0.05 compared with vehicle; #P < 0.05 compared with Con Ab.