Leptin Action in the Ventromedial Hypothalamic Nucleus Is Sufficient, But Not Necessary, to Normalize Diabetic Hyperglycemia

Abstract

In rodent models of type 1 diabetes, leptin administration into brain ventricles normalizes blood glucose at doses that have no effect when given peripherally. The ventromedial nucleus of the hypothalamus (VMN) is a potential target for leptin's antidiabetic effects because leptin-sensitive neurons in this brain area are implicated in glucose homeostasis. To test this hypothesis, we injected leptin directly into the bilateral VMN of rats with streptozotocin-induced uncontrolled diabetes mellitus. This intervention completely normalized both hyperglycemia and the elevated rates of hepatic glucose production and plasma glucagon levels but had no effect on tissue glucose uptake in the skeletal muscle or brown adipose tissue as measured using tracer dilution techniques during a basal clamp. To determine whether VMN leptin signaling is required for leptin-mediated normalization of diabetic hyperglycemia, we studied mice in which the leptin receptor gene was deleted in VMN steroidogenic factor 1 neurons using cre-loxP technology. Our findings indicate leptin action within these neurons is not required for the correction of diabetic hyperglycemia by central leptin infusion. We conclude that leptin signaling in the VMN is sufficient to mediate leptin's antidiabetic action but may not be necessary for this effect. Leptin action within a distributed neuronal network may mediate its effects on glucose homeostasis.

Recent work indicates that the administration of the adiposity hormone leptin directly into the brain normalizes blood glucose levels in rodent models of uncontrolled insulin-deficient diabetes (uDM), despite persistent, severe insulin deficiency (1–6). Moreover, the glucose-lowering effects of intracerebroventricular (icv) leptin are characterized by normalization of increased hepatic glucose production (HGP) and by increased tissue glucose uptake (3). Under the influence of leptin, therefore, the brain is capable of achieving insulin-independent normalization of blood glucose levels in animals with uDM. The current studies were undertaken to identify the neurocircuits responsible for this effect.

The ventromedial nucleus of the hypothalamus (VMN) is of interest as a target for leptin's antidiabetic effects for the following reasons: 1) leptin receptors are expressed in this brain area (7–9), 2) leptin activates neurons in this brain area (10), 3) deletion of leptin receptors from VMN neurons causes a mild obesity and insulin-resistant phenotype in mice (11, 12) and 4) the VMN is implicated in the control of glucagon secretion (13). To investigate this hypothesis, we determined whether leptin action limited to the VMN is either sufficient or necessary to explain its glucose-lowering effects in uDM.

We report that the microinjection of leptin into the VMN is sufficient to normalize diabetic hyperglycemia in streptozotocin (STZ)-induced diabetic (DM) rats via a mechanism that is independent of changes in food intake and involves the suppression of HGP without effects on tissue glucose uptake in skeletal muscle or brown adipose tissue (BAT). Because the antidiabetic effects of leptin do not differ between mice in which leptin receptors have been deleted selectively from VMN neurons [LepR knockout (KO)^VMN] and controls, these findings indicate that although leptin action within the VMN is sufficient to ameliorate hyperglycemia, it is not necessary for this effect. Thus, the VMN may be part of a distributed network through which leptin action in the brain promotes glucose homeostasis independently of the action of insulin.

Materials and Methods

Animals

All procedures were performed in accordance with National Institutes of Health Guidelines for Care and Use of Animals and were approved by the Animal Care Committee at the University of Washington. All studied animals were individually housed in a temperature-controlled room with a 12-hour light, 12-hour dark cycle under specific-pathogen free conditions and provided with ad libitum access to water and chow unless otherwise stated (PMI Nutrition, Brentwood, Missouri). Adult male Wistar rats were obtained from Harlan Laboratories (Indianapolis, Indiana).

To determine whether the effects of leptin to lower blood sugars in uDM is dependent on leptin receptor expression in the VMN, we generated mice with the deletion of the leptin receptor from the VMN (LepR KO^VMN). To accomplish this, we used steroidogenic factor-1 (SF1), a transcription factor expressed in the ventromedial hypothalamic nucleus, to drive Cre recombinase expression and bred Cre-positive mice with leptin receptor^flox/flox animals as described previously (11, 12, 14). Both Sf1-Cre mice and Lepr^flox/flox mice were kindly provided to us by Dr Streamson Chua Jr (Albert Einstein College of Medicine, New York, New York). Mice were genotyped for Sf1-Cre and the Lepr^flox allele as described previously (11).

Surgery

Rats underwent implantation of a bilateral cannula directed to the VMN (Plastics One, Roanoke, Virginia) under isoflurane anesthesia at stereotaxic coordinates (2.8 mm posterior to bregma, ±0.6 mm lateral, and 8.5 mm below the skull surface) or to the lateral ventricle as previously described (15). Bilateral intraparenchymal injections were administered using a microinjector needle that extended 1 mm beyond the tip of the cannula over 60 seconds in a final volume of 0.5 μL. As a negative control for effects of leptin at a site adjacent to the VMN, an additional cohort of animals underwent bilateral cannulation of the anterior hypothalamic area (AHA) (stereotaxic coordinates 1.6 mm posterior to bregma; ±0.6 mm lateral, and 8.0 mm below the skull surface) for microinfusion of leptin into this brain area, which does not express leptin receptors (9). The AHA lies within 0.5 mm of the border of the VMN, a distance equal to that between the center of the VMN and the center of the arcuate nucleus (ARC) of the hypothalamus. As a further control, another additional cohort of rats was implanted with a single cannula to the lateral ventricle as previous described (3).

For mouse cannulations, animals were implanted with a single cannula to the lateral ventricle (Alzet; DURECT Corp, Cupertino, California) at the following stereotaxic coordinates: 0.7 mm posterior to bregma, 1.3 mm lateral, and 2.3 mm below the skull surface. After the induction of STZ-DM, animals were subsequently implanted with a sc osmotic minipump that was connected to the lateral ventricle cannula to enable direct infusion of either vehicle (veh) [PBS (pH 7.9)] or leptin (1 μg/d) directly to the brain.

Experimental protocol

Effect of intra-VMN leptin on food intake, body weight, and blood glucose in STZ-induced diabetes

Cannulated adult male Wistar rats were either made diabetic with 2 consecutive daily sc injections of STZ (40 mg/kg body weight) or received vehicle (NaCit, pH 4.5) and remained nondiabetic, as previously described (3, 16, 17). Three days later, STZ-diabetic animals received daily bilateral injections of either vehicle or leptin (A. F. Parlow, National Hormone Peptide Program, Torrance, California) at a dose (0.05 μg per side) for 12 consecutive days. This dose of leptin was selected based on the data generated from a leptin dose-response study. To control for the effect of leptin to reduce food intake, an additional group STZ-veh-treated animals was pair fed to the intake of the STZ-leptin group, as previously described (3, 16). In total this yielded 6 groups: 1) nondiabetic (veh)-VMN PBS infusion (veh) (n = 6), 2) STZ-veh ad libitum (STZ-veh) (n = 9), 3) STZ-VMN leptin (n = 6), 4) STZ-veh pair fed (STZ-veh-PF) (n = 7), 5) STZ-icv leptin (n = 7), and 6) STZ-AHA leptin (n = 4). Food intake, body weight, and blood glucose levels were measured daily.

Effect of intra-VMN leptin on HGP and glucose uptake in STZ-induced diabetes

Adult male Wistar rats bearing catheters to the right jugular vein and left carotid artery (Harlan) were studied using the same protocol as described above. Ten days after the administration of STZ or vehicle, tracer dilution techniques were used to determine the effect of intraparenchymal leptin on HGP and tissue glucose uptake. Measures of glucose appearance using [3-³H]glucose and tissue glucose uptake using 2[¹⁴C]-deoxyglucose were obtained as previously described (3, 15) in nondiabetic (veh-veh) (n = 5), STZ-veh-PF (n = 6), and STZ-VMN leptin (n = 8) rats.

SF1 Cre × leptin receptor^flox/flox mice

To determine whether the effects of leptin to lower blood sugars in uDM is dependent on leptin receptor expression in the VMN, we generated mice with deletion of leptin receptor from the VMN (LepR KO^VMN) as described above. Both male Cre⁺ (LepR KO^VMN) and Cre⁻ LepR^flox/flox [Lepr wild type (WT)] mice were implanted with a single cannula to the lateral ventricle and after at least a 7-day recovery period, animals received either 2 sc injections of STZ spaced 3 days apart (150 mg/kg body weight) to induce uDM or vehicle [NaCit (pH 4.5)]. Seven days later, at a time when the mice had become diabetic (defined as fed blood glucose level > 250 mg/dL for 2 consecutive days), animals were implanted with a sc osmotic minipump that was connected to the lateral ventricle cannula to enable direct infusion of either veh [PBS (pH 7.9)] or leptin (1 μg/d) to the brain for 13 days. Previous studies have demonstrated that in addition to other brain areas, icv administration of leptin activates the phosphorylation of STAT3 (pSTAT3), a downstream marker of leptin signaling in the VMN (18, 19), suggesting that icv leptin is capable of reaching this particular brain region. Overall, we generated 6 groups of animals: 1) nondiabetic (veh)-icv PBS infusion (veh) Lepr WT (n = 4); 2) veh-veh LepR KO^VMN (n = 6); 3) STZ-veh Lepr WT (n = 3); 4) STZ-veh LepR KO^VMN (n = 7); 5) STZ-leptin Lepr WT (n = 4); and 6) STZ-leptin LepR KO^VMN (n = 6). Our primary focus was on how leptin action in the brain restores euglycemia in diabetic animals, rather than how nondiabetic and STZ-diabetic animals differ in their response to leptin treatment; thus, we chose to omit leptin-treated nondiabetic LepR KO^VMN and WT littermate groups in the present study. After the placement of the osmotic minipump, body weight, food intake, and blood glucose were monitored daily.

Blood collection and tissue processing

At study completion, liver and BAT samples were harvested, frozen on dry ice, and stored at −80°C. Blood samples for plasma hormonal measures were collected in appropriately treated tubes (3, 16) and centrifuged, and the plasma was removed, aliquoted, and stored at −20°C for subsequent assay. Plasma insulin and leptin levels were determined by ELISA (Crystal Chem, Chicago, Illinois), ketone bodies using a colorimetric kit (Wako Chemicals, Richmond, Virginia), and glucagon assayed by a glucagon RIA kit (Linco Research, St Charles, Missouri). For immunohistochemical studies to verify cannula placement and injectate spread in rats and to verify the selective deletion of leptin receptors in the VMN of mice, anesthetized animals were perfused with PBS followed by 4% paraformaldehyde in 0.1 M PBS. Brains were removed, postfixed in 4% paraformaldehyde, sucrose (25%) embedded, and snap frozen in isopentane cooled with dry ice. Brains were sectioned at 14 μM in the coronal plane throughout the hypothalamus, slide mounted, and stored at −80°C for immunohistochemical staining and cannula verification.

Immunohistochemical staining

For both rat and mouse studies, pSTAT3 immunostaining was carried out on perfused-fixed, anatomically matched sections after washing in Tris phosphate-buffered saline at room temperature. Sections were subsequently blocked in freshly prepared 5% nonfat milk for 30 minutes and then blocked in 0.1% hydrogen peroxide for 1 minute. Sections were later blocked for 1 hour with 5% normal goat serum (Jackson ImmunoResearch Laboratories, West Grove, Pennsylvania) in PBS, containing 0.01% Triton X-100, and subsequently incubated in this blocking solution plus rabbit anti-pSTAT3 antibody (1:1000; Sigma-Aldrich, St Louis, Missouri) overnight at 4°C. Sections were finally incubated for 2 hours in blocking solution containing biotinylated goat antirabbit antibody [Vectastain Elite avidin-biotin complex kit; Vector Laboratories, Burlingame, California], and incubated with Vector avidin-biotin complex reagent (Vector Laboratories) and developed with diaminobenzidine. Stained sections were mounted onto slides, dehydrated, cleared in xylene, and coverslipped with Permount, using standard procedures (19).

Immunohistochemical verification of cannula and injection site

In addition to the verification of cannula placement by histological analysis (20), the spread of injection for intra-VMN surgeries was verified by two independent methods. First, we performed bilateral microinjection of a Cy3-labeled recombinant peptide (leptin) (Phoenix Pharmaceuticals, Belmont, California; 0.5 μL per side) into the VMN and assessed distribution of the Cy3 label postmortem, as previously described (20). Additionally, pSTAT3-positive cell nuclei were visualized within the ARC, VMN, and dorsomedial nucleus of the hypothalamus using a Nikon Eclipse E800 microscope fitted with a grid reticule with the investigator blinded to study conditions (Nikon, Tokyo, Japan). Representative photomicrographs of cannula placement and the injection site are shown in Figure 1. Data from rats with evidence of spread beyond the VMN by any of the above criteria were excluded from further analysis (n = 4).

Figure 1.

Verification of cannula placement and spread of injectate after the microinjection into the VMN. A, Schematic of mediobasal hypothalamus. B, Cannula placement from a coronal section at the level of the VMN. C, A fluorescent micrograph showing cy3-labeled leptin within the VMN. D, Representative image of immunohistochemical staining of pSTAT3, a downstream marker of leptin receptor activation. For 4 animals, cy3-labeled leptin and pSTAT3 signaling overlapped with the borders of other hypothalamic areas and were excluded from the analysis. DMN, dorsomedial nucleus; 3V, third ventricle.

To verify and confirm the selective deletion of leptin signaling in the VMN of our genetic mouse model, 15-week-old male LepR KO^VMN and LepR WT littermate controls received an ip injection of 100 μg recombinant mouse leptin or PBS as a control. One hour later, mice were anesthetized and perfused fixed, and coronal brain sections were subsequently processed for immunohistochemical detection of pSTAT3, a downstream marker of leptin signaling as described above.

Reverse transcription-polymerase chain reaction

Total RNA was extracted from BAT and liver using TRIzol B according to the manufacturers' instructions (Molecular Research Center, Cincinnati, Ohio). RNA was quantitated by spectrophotometry at 260 nm (NanoDrop 1000, Thermo Scientific, Rockford, Illinois), reverse transcribed with avian myeloblastosis virus reverse transcriptase (Promega, Madison, Wisconsin) and real-time PCR performed on an ABI Prism 7900 HT (Applied Biosystems, Foster City, California) as described previously (3, 16). Expression levels of each gene were normalized to a housekeeping gene (18S RNA) and expressed as a percentage of veh-veh controls.

Statistical analyses

All results are expressed as mean ± SEM. Comparisons between multiple groups were initially made using a repeated-measures ANOVA, but because not all experimental group permutations were studied, full interactive models were not tested. Instead, statistical results from 1-way ANOVA and Tukey's post hoc tests for comparisons between groups are presented. For 2-group comparisons, a 2-sample, unpaired Student's t test was used. Statistical analyses were performed using PASW Statistics (version 18, Chicago, Illinois). Probability values of less than 0.05 were considered significant.

Results

Immunohistochemical verification of cannula placement and injection spread

As a first step, we histologically verified both the anatomic placement of cannulae (Figure 1B) directed to the VMN and the spread of injection after a bilateral VMN microinjection of cy3-labeled leptin (Figure 1C). As a secondary method of validation, we also stained for pSTAT3, a marker of leptin receptor activation (Figure 1D), after a VMN leptin microinjection. Of the total of 32 rats undergoing the VMN cannulation, 4 animals were excluded on the basis of cannula placement outside the VMN, confirmed by analysis of cy3-labeled leptin and/or pSTAT3 staining.

Effect of intra-VMN leptin on food intake, body weight, and blood glucose levels in STZ-induced diabetes

As expected, plasma insulin and leptin levels were markedly reduced in all STZ-diabetic animals relative to nondiabetic controls (Figure 2, A and B). In addition, consistent with previous reports (17), STZ-veh-treated animals were characterized by marked hyperphagia relative to nondiabetic controls. This diabetic hyperphagia was prevented in STZ-DM animals that received a low dose of leptin directly into the VMN but not when injected into the lateral ventricle (Figure 2C). Moreover, the marked hyperglycemia characteristic of STZ-DM was also normalized in animals receiving daily intra-VMN leptin injections, an effect that cannot be explained by reduced food intake because hyperglycemia was not markedly attenuated in pair-fed controls that received intra-VMN vehicle (Figure 2D). The glucose-lowering effect of intra-VMN leptin was also not observed in STZ-DM animals that received the same low dose of leptin either icv or as a microinjection into the AHA, a brain area that does not express leptin receptors but lies within a distance equal to the center of the VMN and ARC (mean blood glucose 358 ± 11 mg/dL for STZ lep-AHA vs 395 ± 13 mg/dL for STZ-veh; P = ns). Collectively these findings suggest that leptin action limited to the VMN is sufficient to reverse both the hyperphagia and hyperglycemia of uDM.

Figure 2.

Intra-VMN leptin normalizes diabetic hyperphagia and hyperglycemia in STZ-diabetic rats. Shown are the plasma insulin (A), plasma leptin (B), mean daily food intake (C), and fed blood glucose levels (D) in STZ-induced diabetic animals receiving either vehicle (STZ-veh) and pair-fed (STZ-veh-PF) or leptin given icv (STZ-lep ICV) or directed to the VMN (STZ-lep VMN) (n = 6–9 per group). The arrow represents the start of daily leptin injections. Data represent mean ± SEM. *, P < .05 vs veh-veh; #, P < .05 vs STZ-veh-PF. The asterisk at the top of panels C and D indicates statistical differences between STZ-diabetic rats relative to veh-veh-treated controls.

After the induction of diabetes in STZ-treated rats, there was a significant fall in body weight despite elevated food consumption because the excess calories cannot be stored as fat due to the insulin deficiency and are ultimately lost through the urine. Consistent with our previous observations (3), we found that leptin-treated STZ-diabetic rats exhibited weight loss similar to STZ-diabetic rats receiving vehicle (−65.3 ± 6.3 g for STZ-veh vs −61.3 ± 2.5 g vs STZ-lep; P = ns), whereas the STZ-veh pair-fed group lost significantly more weight than those STZ-DM rats fed ad libitum (−65.3 ± 6.3 g for STZ-veh vs 80.8 g ± 2.9 g for STZ-veh-PF; P < .05). Together these data suggest that although leptin treatment reduces food intake and may increase energy expenditure, leptin-treated STZ-diabetic rats do not lose more weight compared with STZ-veh-treated animals, presumably because fewer calories are lost through the urine (3, 16).

Effect of intra-VMN leptin on HGP and glucose uptake in STZ-induced diabetes

To identify the mechanism whereby intra-VMN leptin normalizes diabetic hyperglycemia, we used tracer dilution techniques during a basal period (ie, without insulin or glucose infusion) to measure HGP and tissue glucose uptake. Because intra-VMN leptin administration blocked diabetic hyperphagia and this reduction of food intake by itself attenuated diabetic hyperglycemia (Figure 2D), there was a critical need to control for differences in food intake. For the tracer studies, therefore, the key comparison was between STZ-leptin-treated and STZ-veh-treated pair-fed animals, rather than STZ-DM animals fed ad libitum. A veh-veh group was also included to assess the extent to which glucose production and/or uptake was normalized by leptin action in the VMN vs reduced food intake per se, ie, this strategy allowed us to assess whether intra-VMN leptin exerted effects in STZ-DM rats independently of its effects on food intake.

As expected, the rate of endogenous glucose appearance was significantly increased in STZ-veh-PF animals relative to nondiabetic controls (Figure 3B). Importantly, this increase of glucose appearance was normalized to nondiabetic values in STZ-diabetic rats that received intra-VMN-leptin (Figure 3B). Similarly, hepatic expression of the gene encoding the gluconeogenic enzyme phosphoenolpyruvate carboxykinase (Pepck) was elevated in STZ-diabetic animals that received intra-VMN vehicle relative to nondiabetic controls, and this effect was reversed by intra-VMN leptin (Figure 3C). Although hepatic gene expression of glucose-6 phosphatase (G6Pase) was also higher in STZ-diabetic animals that received intra-VMN vehicle than leptin, this difference failed to achieve statistical significance (Figure 3D). Moreover, we found that the elevated plasma glucagon levels in STZ-DM vehicle-treated rats were normalized in STZ-DM rats that received leptin to the VMN (Figure 3E). In addition, intra-VMN leptin also normalized the increase in ketone body levels characteristic of STZ-DM (Figure 3F). Because increased HGP contributes importantly to diabetic hyperglycemia (21), these data suggest that leptin action in the VMN lowers blood glucose levels in uDM in part by normalizing HGP, an affect accompanied by a normalization of plasma glucagon levels.

Figure 3.

Intra-VMN leptin suppresses HGP in STZ-diabetic rats. Three-hour fasted plasma glucose levels (A), the rate of glucose appearance (Ra) as determined from [3-³H] glucose tracer studies (B), and hepatic expression of phosphoenolpyruvate carboxykinase (Pepck) (C) and glucose 6 phosphatase (G6Pase) (D) using real-time PCR and plasma levels of glucagon (E) and ketone bodies (F) in STZ-induced diabetic animals receiving either icv vehicle and pair fed or intra-VMN leptin relative to nondiabetic controls (n = 5–8 per group) are shown. Data represent mean ± SEM. *, P < .05 vs veh-veh; #, P < .05 vs STZ-veh-PF.

Due to the confounding effect of glycosuria in uDM (3), we measured tissue glucose uptake using labeled 2[¹⁴C]-deoxyglucose rather than tracer-based estimates of glucose disappearance. In contrast to the effect of VMN-leptin to lower HGP in STZ-diabetic rats, glucose uptake in skeletal muscle and BAT was not increased in STZ-diabetic animals receiving intra-VMN leptin; however, glucose uptake was 3-fold elevated in the heart of VMN leptin rats (data not shown). Additionally, the effect of uDM to dramatically reduce activity of BAT, as judged by reduced BAT Ucp1 and Pgc-1α mRNA levels, was unaffected by VMN leptin treatment (data not shown). These findings suggest that VMN-leptin is unlikely to ameliorate diabetic hyperglycemia via a mechanism primarily involving increased tissue glucose uptake in skeletal muscle or BAT activation.

Role of leptin receptors within the VMN on food intake, body weight, and blood glucose in STZ-induced diabetes

To determine whether leptin signaling in the VMN is necessary for the glucose-lowering effects of leptin in uDM, we studied mice with selective deletion of leptin receptors from SF1 neurons, the principal leptin-responsive neuronal subtype in the VMN (11, 12, 22). To verify the deletion of the leptin receptor in the VMN, we examined leptin-induced activation of pSTAT3, a downstream marker of leptin signaling in both LepR KO^VMN and LepR WT animals. Our findings show that leptin markedly increases pSTAT3 in the ARC of leptin-treated LepR KO^VMN and LepR WT animals relative to vehicle-treated controls. In contrast, pSTAT3 was markedly decreased in the VMN of LepR KO^VMN relative to LepR WT animals (Figure 4). These studies demonstrate the selective deletion of leptin in the VMN but not closely adjacent hypothalamic nuclei.

Figure 4.

Leptin-induced pSTAT3 activation in LepR WT and LepR KO^VMN mice. pSTAT3 immunohistochemistry on brain sections from representative LepR WT mice injected with saline (A) or leptin (C) and a LepR KO^VMN mouse injected with leptin (D) is shown. B, Schematic of the mediobasal hypothalamus. DMN, dorsomedial nucleus; 3V, third ventricle.

After the administration of STZ, both LepR WT and LepR KO^VMN animals exhibited similarly elevated levels of blood glucose, food intake, and weight loss relative to nondiabetic controls (Figure 5, A–D). As expected, blood glucose levels were normalized to nondiabetic control values in STZ-DM Lepr WT animals that received continuous icv leptin infusion (1 μg/d) (Figure 5B). However, this potent antidiabetic effect of leptin was fully preserved in STZ-DM LepR KO^VMN animals (Figure 5B). Thus, leptin action on SF1 neurons in the VMN does not appear to be required for leptin's glucose-lowering effects in uDM. Similarly, the effects of leptin on food intake and body weight did not differ between diabetic Lepr WT and LepR KO^VMN animals (Figure 5, C and D).

Figure 5.

Antidiabetic effects of leptin do not require leptin signaling through SF1 neurons. Plasma insulin concentrations (A), blood glucose (B), food intake (C), and body weight (D) in STZ-induced diabetic animals receiving either icv vehicle or leptin relative to nondiabetic controls for both genotypes (LepR WT, and LepR KO^VMN) (n = 3–7 per group) are shown. Data represent mean ± SEM. *, P < .05 vs veh-veh; #, P < .05 vs STZ-veh. The asterisk at the top of panel B indicates differences between STZ-veh- and veh-veh-treated mice for both genotypes.

Discussion

Although the treatment of type 1 diabetes has traditionally been based on insulin replacement, recent studies demonstrate that leptin can also normalize blood glucose levels in rodent models of uDM via an action in the central nervous system (CNS) (1–6). Because leptin-sensitive VMN neurons are implicated in the control of energy and glucose homeostasis (11, 12, 23), we hypothesized that leptin signaling in this brain area contributes to its potent glucose-lowering effects in diabetic animals. Our data show that intra-VMN leptin action is sufficient to correct diabetic hyperglycemia in STZ-DM rats and that the underlying mechanism appears to involve a suppression of HGP that is accompanied by a normalization of elevated plasma glucagon levels. Somewhat surprisingly, we did not detect the increase of peripheral tissue glucose uptake in skeletal muscle or the activation of BAT previously reported when leptin is given icv to rats with uDM (3, 24). Because these effects occurred despite severe insulin deficiency, our findings suggest that activation of leptin-sensitive VMN neurons inhibits HGP via an insulin-independent mechanism. Yet the antidiabetic effects of leptin were not attenuated in mice in which the leptin receptor was deleted from VMN SF1 neurons (LepR KO^VMN). Collectively these findings suggest that leptin action in the VMN is sufficient, but not necessary, to explain its glucose-lowering effect in animals with diabetes. These data are compatible with the existence of a distributed neuronal network on which leptin acts to promote glucose homeostasis.

Several caveats should be considered in the interpretation of these results. First, despite excluding animals in which either fluorescently labeled leptin or leptin-induced pSTAT3 staining were detected outside the VMN and despite evidence that microinjection of leptin to the adjacent AHA was devoid of glucose-lowering effects, it is not possible to definitively rule out the possibility that leptin diffused out of the VMN to mediate its biological actions at some other location. Second, because recombination is unlikely to be 100% efficient and because a small subset of leptin receptor neurons within the VMN may not express SF1 and therefore would not be affected in LepR KO^VMN mice (12), these few remaining leptin receptor-positive neurons could have mediated leptin's antidiabetic effects. This is unlikely based on previous immunohistochemical and electrophysiological evidence that suggests leptin-induced activation of pSTAT3 in the VMN of LepR KO^VMN mice is located exclusively in SF1-negative, rather than SF1-positive, neurons (12). Moreover, although leptin depolarizes and increases the firing rate of SF1 neurons in the VMN, an effect absent in LepR KO^VMN mice, SF1-negative neurons are heterogeneous in their response to leptin (12). Instead, however, we support the hypothesis that leptin signaling in the VMN is sufficient, but not necessary, for its antidiabetic effects in uDM, based on a model whereby leptin acts on multiple nodes of a distributed network of leptin-sensitive neurons to mediate its biological actions. These findings support previous evidence that leptin action on food intake is distributed across several different neuronal populations and brain regions (25, 26). This model is also consistent with the comparatively mild phenotype of mice with selective deletion of leptin receptors from any of several neuronal populations (ie, agouti-related peptide or proopiomelanocortin neurons) (27, 28) or brain regions (VMN) (11, 12) known to be critical for normal energy and glucose homeostasis. Further studies are therefore warranted to identify other neuronal populations and brain regions in the CNS that are sufficient and necessary for leptin's glucose-lowering effects, if these regulate HGP, peripheral glucose uptake, or both and if nutritional- (eg, free fatty acids) or hormonal-related inputs (eg, insulin) in addition to leptin impinge on this neurocircuit. One possible pathway that may be involved in this neurocircuit is the melanocortin system because leptin's antidiabetic effects in uDM require melanocortin signaling (29); however, the activation of melanocortin signaling alone with a melanocortin 3/4 receptor agonist is not sufficient to mimic the glucose-lowering effects of leptin (29).

Our previous findings suggest that leptin action in the brain normalizes blood glucose levels in uDM in part by suppressing HGP (3). Here we show that leptin action in the VMN reverses the effect of STZ-DM to increase both HGP and gluconeogenic gene expression. These observations implicate the VMN in leptin-mediated inhibition of HGP. Although our data do not provide information as to what extent reduced food intake per se affects HGP in animals with uncontrolled diabetes, our results demonstrate intra-VMN leptin normalized HGP in rats with STZ-DM compared with diabetic animals consuming the same amount of food. Consistent with a role for this brain area in the control of glucagon secretion (13), our findings demonstrate that the glucose-lowering effects of VMN leptin are accompanied by suppressing hyperglucagonemia. However, because plasma glucagon levels in uDM are normalized with either a physiological replacement dose of leptin (16) or central administration of brain-derived neurotrophic factor (15), without normalizing blood glucose levels in uDM, it suggests that mechanisms in addition to a suppressing glucagon are likely to contribute to the suppression of HGP.

In contrast to its effects on HGP, intra-VMN leptin failed to increase glucose uptake in skeletal muscle and BAT, although the effect of leptin to increase glucose uptake in the heart remained intact (3). This finding comes as a surprise because in nondiabetic rats, microinjection of leptin into the VMN promotes peripheral tissue glucose uptake in BAT and muscle via the sympathetic nervous system (30, 31), suggesting that a fundamental difference exists between diabetic and nondiabetic animals in this type of VMN action of leptin. One possible explanation for these findings is that insulin is required for the effect of intra-VMN leptin to promote peripheral tissue glucose uptake. Available data do, however, indicate that this effect of leptin involves brain areas outside the VMN. Consistent with this, leptin action in the hypothalamic ARC stimulates glucose uptake in BAT (32) and leptin receptor-containing neurons in the adjacent retrochiasmatic area project transsynaptically to skeletal muscle (33).

In summary, our data suggest that in diabetic animals, leptin signaling limited to the VMN is sufficient to ameliorate diabetic hyperglycemia by normalizing elevated rates of HGP and plasma glucagon levels. However, because leptin signaling in the VMN does not appear to be necessary for its glucose-lowering actions, we hypothesize that leptin acts on a distributed network of leptin-sensitive neurons to regulate glucose metabolism.