Leptin Activates a Novel CNS Mechanism for Insulin-Independent Normalization of Severe Diabetic Hyperglycemia

Animals

Adult male Wistar rats (Harlan Laboratories, Inc., Indianapolis, IN) were housed individually under specific-pathogen-free conditions, in a temperature-controlled room with a 12-h light, 12-h dark cycle and ad libitum access to water and chow (Purina 5053, Purina Mills, St. Louis, MO) unless otherwise stated. All procedures were performed in accordance with National Institutes of Health Guidelines for the Care and Use of Animals and were approved by the Animal Care Committee at the University of Washington.

Effect of CNS leptin infusion on blood glucose levels in streptozotocin (STZ)-induced diabetes

Animals were implanted with a cannula to the lateral ventricle (Alzet; DURECT Corp., Cupertino, CA) under isoflurane anesthesia at stereotaxic coordinates: 1.5 mm lateral, 0.8 mm posterior to bregma, and 3.5 mm below the skull surface (19). At least 1 wk after cannula implantation, animals received either two consecutive daily sc injections of STZ (40 mg/kg · body weight) to induce uDM or veh (vehicle) [NaCit (pH 4.5)]. On the same day, under isoflurane anesthesia, animals were implanted with an osmotic minipump sc that was connected to the lateral ventricle to enable direct infusion of either veh [PBS (pH 7.9)] or leptin (1 μg/d; Dr. Parlow; National Hormone Peptide Program, Torrance, CA) to the brain for up to 11 d. This dose of leptin was selected based on the data generated from a leptin dose-response study. To control for the effects of leptin to reduce food intake, a third group of STZ-veh-treated animals was pair fed (PF) to the intake of STZ-leptin-treated animals as previously described (20). An additional control group was added, in which STZ-veh-treated animals were treated with leptin (1 μg/d) systemically, to determine whether the central effect of leptin was due to leakage into the periphery. Thus, a total of five groups of rats were studied: 1) veh-veh, 2) STZ-veh, 3) STZ-leptin, 4) STZ-veh-PF, and 5) STZ-leptin-sc (n = 7–9 per group). Food intake, body weight, water intake, and blood glucose levels were measured daily midlight cycle for a total of 11 d.

The experimental paradigm described above was repeated in separate groups of animals, with the exception that an indwelling catheter was inserted into the left carotid artery, as previously described (13), to permit blood sampling from conscious, unstressed rats for determination of plasma glucagon, corticosterone, and catecholamine levels.

Measures of glucose tolerance

Intraperitoneal glucose-tolerance tests (30% d-glucose; 2 g/kg) were conducted in 4-h fasted animals on d 9 after STZ administration. Blood glucose levels were measured at t = 0, 30, 60, 90, and 120 min, and a blood sample was obtained at t = 0 and 30 min for determination of plasma insulin concentrations.

Effect of CNS leptin infusion on HGP and glucose uptake in STZ-diabetic rats

Eight days after the administration of STZ, we used tracer dilution techniques to determine whether CNS leptin treatment reduces HGP, increases glucose uptake, or both in STZ-induced diabetic rats. Animals were fasted for 4 h, and a blood sample was obtained at t = 0 min for determination of fasting plasma glucose, insulin, leptin, glucagon, and corticosterone levels in appropriately treated tubes. A 24 μCi prime of [3-³H] glucose was given at t = 0 min for 3 min followed by a continuous 0.2 μCi/min infusion for 90 min. A bolus of 50 μCi 2[¹⁴C]-deoxyglucose (2-DG) was given at t = 48 min. Blood samples were taken at 10-min intervals from 60 to 90 min and processed to determine plasma [3-³H]glucose and 2-DG. At t = 90 min, animals were anesthetized with ketamine:xylazine and the liver, red gastrocnemius muscle, brown adipose tissue (BAT), heart, and brain were rapidly excised and stored at −80 C until analyzed. To more directly test whether the ability of intracerebroventricular (icv) leptin to enhance glucose utilization occurs through an insulin-independent mechanism, nondiabetic and STZ-treated animals that received icv leptin were prepared as described above. At t = 0 min, animals were administered a 24 μCi prime of [3-³H]glucose for 3 min followed by a continuous 0.2 μCi/min infusion for 90 min. At t = 90 min, a primed continuous variable infusion of 50% dextrose was administered and periodically adjusted to clamp the blood glucose concentration at approximately 210 mg/dl in veh-veh- and STZ-leptin-treated animals. A bolus of 50 μCi 2[¹⁴C]DG was given at t = 178 min. Blood samples were taken at 10-min intervals from 180 to 210 min and processed to determine plasma [3-³H]glucose, 2[¹⁴C]DG and insulin levels. At t = 210 min, animals were anesthetized with ketamine:xylazine, tissues rapidly excised and stored at −80 C for later analysis.

Blood and urinary collection and assay

Plasma for [3-³H]glucose determinations were deproteinized with Ba(OH)₂ and ZnSO₄, then dried overnight at 60 C. An aliquot of urine was collected into chilled tubes from animals that were housed briefly in metabolic cages in the fed state during the midlight cycle, 10 d after STZ administration, and were subsequently stored at −20 C. Plasma glucose and urinary glucose levels were measured using a GM9D glucose direct analyzer (Analox Instruments, London, UK). Daily blood glucose levels were measured using a hand-held glucometer (Accu-Chek, Indianapolis, IN) on blood obtained from tail capillary samples. Blood samples were collected for measurement of insulin, leptin, nonesterified free fatty acids (NEFAs), and ketone bodies from the tail vein in chilled EDTA-treated tubes. Blood for catecholamine and glucagon assays was collected from arterial blood samples collected on EGTA/glutathione tubes and tubes containing 10 μl of 1 m benzamidine (Sigma, St. Louis, MO) and 1 U heparin, respectively, whereas blood for corticosterone was collected in EDTA-treated tubes. Whole blood was centrifuged at 1500 rpm for 20 min, and plasma was removed, aliquoted, and stored at −20 C. Plasma immunoreactive insulin and leptin levels were determined by ELISA (Crystal Chem, Inc., Downers Grove, IL), NEFAs using a colorimetric assay kit (Wako Chemicals, Richmond, VA) and total ketone bodies using a colorimetric kit (Wako Chemicals). Catecholamine levels were measured using a sensitive and specific radioenzymatic assay (21), glucagon was assayed by a glucagon RIA kit (Linco Research, St. Charles, MO), whereas plasma corticosterone levels were measured using enzyme immunoassay (Diagnostics Systems Laboratories, Webster, TX).

Reverse transcription-polymerase chain reaction

Total RNA was extracted from liver and pancreas using TRIzol B according to manufacturers' instructions (MRC, Cincinnati, OH). RNA was quantitated by spectrophotometry at 260 nm (Nanodrop 1000; Thermo Scientific, Rockford, IL) and reverse-transcribed with avian myeloblastosis virus reverse transcriptase (1 μg; Promega, Madison, WI). Levels of mRNA for glucose-6-phosphatase (G6Pase), phosphoenolpyruvate kinase (Pepck), Igf binding protein (Igfbp)-2 and 18S RNA (internal control) were measured by semiquantitative real-time PCR on an ABI Prism 7900 HT (Applied Biosystems, Foster City, CA). The primer sequences for 18S RNA, G6Pase, Pepck, and Igfbp-2 were designed using Primer Express (version 2.0.0; Applied Biosystems) and are as follows: 18S RNA forward, CGGACAGGATTGACAGATTG and reverse, CAAATCGCTCCACCAACTAA; G6Pase forward, TCAACCTCGTCTTCAAGTGGATT and reverse, CTGCTTTATTATAGGCACGGAGCT; Pepck forward, GGCGGAGCATATGCTG-ATCC and reverse, CCACAGGCACTAGGGAAGGC; and Igfbp-2 forward, TCTACAACG-AGCAGCAGG and reverse, ATCCTCCACACCCATCAC. PCR data were analyzed using the Sequence Detection System software (SDS version 2.2; Applied Biosystems). Expression levels of each gene were normalized to a house-keeping gene (18S RNA) and expressed as a percentage of veh-veh controls. Nontemplate controls were incorporated into each PCR run.

Body composition analysis

Liver tissue triglyceride content and measures of body lean and fat mass were determined using quantitative magnetic resonance spectroscopy (Echo Medical Systems, Houston, TX). Liver glycogen levels were determined using a colorimetric assay (Biovision, Mountain View, CA). Both liver and glycogen were standardized to grams wet weight.

Glucose rates of appearance (Ra), disposal (Rd), and Rg

After deproteinization and drying, plasma [3-³H]glucose and 2[¹⁴C]DG radioactivity was determined by liquid scintillation counting (Packard TRI-CARB 2900TR) (22) with Ultima Gold (Packard) as scintillant (22). Ra, Rd, and endogenous Ra were determined as described previously (13). For 2[¹⁴C]DG measures, tissue samples were homogenized in 0.5% perchloric acid and neutralized with KOH. One sample was measured directly to count 2[¹⁴C]DGP and 2[¹⁴C]DG, whereas a second sample was deproteinized with Ba(OH)₂ and ZnSO₄ to measure only 2[¹⁴C]DG. Tissue glucose clearance was calculated by dividing the 2[¹⁴C]DGP dpm for each tissue by the plasma area under the curve 2[¹⁴C]DG as determined by the trapezoid method. Multiplying tissue glucose clearance by the average plasma glucose concentration gives the rate of Rg (23). In all experiments, the accumulation of 2[¹⁴C]DGP was normalized to tissue weight.

Statistical analysis

All results are expressed as mean ± sem. Statistical analyses were performed using Statistica (version 7.1; StatSoft, Inc., Tulsa, OK). A one-way ANOVA with a least significant difference post hoc test was used to compare mean values between multiple groups and a two-sample unpaired Student's t test was used for two-group comparisons. In all instances, probability values of less than 0.05 were considered significant.

CNS leptin infusion normalizes blood glucose levels in STZ-induced diabetes

Consistent with previous data (24, 25), plasma levels of both insulin and leptin (Fig. 1, A and B) were significantly reduced within 48 h in all animals that received STZ and remained very low relative to nondiabetic controls, confirming previous reports that uDM is a model of leptin and insulin deficiency (24). As expected (5, 6), food intake was significantly increased by d 5 after STZ administration in diabetic animals receiving icv veh, and this effect was completely blocked by icv leptin (Fig. 1C). To control for differences in food intake, we included a group of STZ-diabetic rats that were PF (STZ-veh-PF) to the mean daily intake of the STZ-leptin group.

Although blood glucose levels were markedly increased in diabetic animals receiving icv veh throughout the study, blood glucose levels normalized to nondiabetic levels by d 7 in STZ-diabetic rats receiving icv leptin (Fig. 1D). This effect was due neither to a leptin-mediated reversal of diabetic hyperphagia, because hyperglycemia was not attenuated in STZ-veh-PF animals, nor to a peripheral effect of leptin, because sc infusion of leptin at the same dose did not affect blood glucose in STZ-diabetic rats (Fig. 1D). After the induction of diabetes in STZ-treated rats receiving icv veh, body weight continuously decreased despite pronounced hyperphagia, because the excess calories cannot be stored as fat and are ultimately lost through the urine. Interestingly, leptin-treated STZ-diabetic rats exhibited similar weight loss to STZ-diabetic rats receiving icv veh but exhibited a greater loss of fat mass (Supplemental Fig. 1, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org). These data suggest that icv leptin treatment reduces food intake and stimulates energy expenditure, but these animals do not lose more weight compared with STZ-diabetic animals receiving veh, presumably because they lose less calories through the urine.

Potential mechanism(s) mediating the glucose-lowering action of CNS leptin

One potential mechanism that could contribute to the glucose-lowering effect of leptin is an increase in urinary glucose loss. However, urinary glucose was undetectable in diabetic animals that received icv leptin similar to nondiabetic controls, whereas STZ-veh-treated animals displayed the expected, marked glycosuria (Fig. 2A). Similarly, water intake was markedly increased in both groups of STZ-veh-treated animals, owing to the dehydrating effect of pronounced glycosuria in uDM (26), whereas water intake in diabetic animals receiving icv leptin was equivalent to that of nondiabetic controls (Fig. 2B). Thus, the effect of leptin to normalize blood glucose levels in uDM is not mediated by increased urinary glucose loss and in fact occurs despite the complete reversal of glycosuria.

Another potential mechanism by which leptin action in the CNS normalizes diabetic hyperglycemia is by restoring pancreatic β-cells. However, arguing against this hypothesis, basal plasma insulin levels remained very low in the fed state in STZ-treated animals, irrespective of icv treatment status, and pancreatic preproinsulin mRNA levels were severely reduced (by >95%) in all animals that received STZ (data not shown). Last, 30 min after a glucose challenge (2 g/kg ip), plasma insulin levels increased as expected in nondiabetic animals but remained extremely low in both veh and leptin-treated STZ-diabetic rats (Fig. 2C). Thus, the severe insulin deficiency (under both basal and glucose-stimulated conditions) caused by STZ-induced β-cell destruction was not ameliorated by icv leptin. However, despite the absence of any increase in circulating insulin, STZ-leptin-treated animals exhibited near normal glucose tolerance compared with nondiabetic controls, implying the existence of an insulin-independent mechanism for rapid glucose clearance that is activated by CNS leptin action (Fig. 2D).

Effect of icv leptin infusion on HGP and glucose uptake in STZ-induced diabetes

We next used tracer dilution techniques to determine whether icv leptin normalized Ra or Rd (or both) in rats with uDM. Consistent with the key role played by increased HGP in the hyperglycemia of uDM (27), basal Ra measured by tracer dilution were markedly elevated in STZ-veh-PF animals compared with nondiabetic controls (Fig. 3A). Central leptin administration dramatically reduced Ra to nondiabetic control values (Fig. 3A), strongly suggesting that normalization of diabetic hyperglycemia resulted, at least in part, from reduced HGP. Because this suppression of HGP occurs despite near undetectable plasma insulin levels, it differs from the previously documented effect of CNS leptin to increase hepatic insulin sensitivity (10–13) and implies the existence of an insulin-independent mechanism.

Because animals with uDM excrete large amounts of glucose in the urine, tracer-based estimates of the Rd are artificially increased in STZ-diabetic animals receiving veh, because Rd values encompass both whole-body glucose utilization and urinary glucose loss (Fig. 3B). In contrast, Rd values in STZ-diabetic animals that receive icv leptin are not confounded by this problem, because the animals are neither hyperglycemic nor glycosuric. We found that basal Rd values were similar between STZ-diabetic animals receiving leptin and nondiabetic controls but were unable to compare these values with veh-treated STZ-diabetic rats due to the confounding effects of glycosuria on Rd. We therefore measured Rg using labeled 2-DG and found that glucose uptake was significantly increased in red gastrocnemius muscle and brain in STZ-diabetic animals that received icv leptin relative to veh (Fig. 3C). In BAT, glucose uptake was significantly increased in STZ-diabetic rats receiving leptin compared with both those receiving veh and nondiabetic controls, whereas in heart, glucose uptake was significantly reduced in STZ-diabetic rats receiving veh compared with nondiabetic controls, and this effect was reversed with leptin treatment (Fig. 3D).

To investigate the role of the liver in the antidiabetic effects of leptin, we examined hepatic gene expression using real-time PCR as well as measures of liver triglyceride and glycogen content. As expected, hepatic expression of the key gluconeogenic genes G6Pase and Pepck were markedly increased in the liver of STZ-veh-treated animals relative to nondiabetic controls (Fig. 4, A and B). Because these enzymes are rate-limiting for gluconeogenesis, their increased expression likely contributes to elevated rates of HGP in this setting. By comparison, hepatic expression of these genes was significantly reduced in STZ-diabetic rats receiving icv leptin, suggesting that leptin action in the brain reduces HGP in rats with uDM by inhibiting gluconeogenic gene expression, independent of insulin signaling (Fig. 4, A and B).

Igfbp-2 was recently identified as a leptin-regulated hepatic gene implicated in the action of leptin to suppress HGP (28). Surprisingly, we found that hepatic expression of Igfbp2 mRNA was increased by mote than 25-fold in leptin-deficient STZ-diabetic rats compared with nondiabetic controls. Even though hepatic Igfbp2 mRNA levels were further increased by leptin, this effect seems unlikely to explain the effect of leptin to normalize HGP (P < 0.05) (Fig. 4C). Although hepatic triglyceride content was also reduced in STZ-leptin-treated animals compared with other groups, this effect was relatively small (Fig. 4D), whereas liver glycogen levels were dramatically reduced in STZ-treated animals receiving icv leptin compared with the other groups (Fig. 4E), a finding that may underlie the effect of icv leptin to reduce HGP.

To directly test the hypothesis that icv leptin suppresses HGP and stimulates glucose utilization in animals with uDM through an insulin-independent mechanism, we subjected both leptin-treated STZ-diabetic animals and veh-treated nondiabetic controls to a hyperglycemic clamp (Fig. 5). In nondiabetic animals, plasma insulin and C-peptide levels were significantly increased during the hyperglycemic clamp (P < 0.05) (Fig. 5, A and B), resulting in complete suppression of HGP and enhancement of glucose utilization by more than 250% (Fig. 5, E–H). Similarly, despite persistent, severe insulin deficiency, HGP was suppressed and glucose utilization was increased in STZ-diabetic rats receiving icv leptin, although less than that observed in nondiabetic animals (Fig. 5, E–H). Furthermore, the effect of hyperglycemia to increase Rg in skeletal muscle and heart was similar between STZ-leptin and nondiabetic controls, and icv leptin treatment further increased glucose uptake into BAT (Fig. 5, I and J).

CNS leptin infusion normalizes plasma glucagon and corticosterone levels in STZ-induced diabetes

Elevated plasma glucagon and corticosterone levels are implicated in the pathogenesis of hyperglycemia in uDM (29–31). As was observed in response to pharmacological hyperleptinemia in an earlier report (4), the elevation in plasma glucagon levels observed in STZ-treated animals was prevented by central leptin administration (Fig. 6A). Plasma corticosterone levels were similarly elevated in animals with STZ diabetes, a response also prevented by icv leptin but not PF (Fig. 6B). These findings raise the possibility that in uDM, reduced leptin action in the brain stimulates the secretion of counterregulatory hormones and that the antidiabetic effect of icv leptin is mediated in part via normalization of these hormone levels. However, we recently reported that in rats with STZ diabetes, physiological leptin replacement is sufficient to normalize circulating glucagon and corticosterone levels, yet does not reverse hyperglycemia (20). Although changes in plasma norepinephrine and epinephrine levels are unlikely to explain leptin's effects (Fig. 6, A and D), plasma levels of NEFAs and ketone bodies were also elevated in STZ-veh animals and normalized with icv leptin administration but only after blood glucose levels had returned to normal (Fig. 6, E and F).