Enhanced Hypothalamic AMP-Activated Protein Kinase Activity Contributes to Hyperphagia in Diabetic Rats

Male Sprague-Dawley rats weighing 300–350 g were maintained at ambient temperature (22 ± 1°C) and with 12:12-h light-dark cycles (lights on 7:00 a.m.) with free access to water and rat diet unless otherwise indicated. All procedures were approved by the Institutional Animal Care and Use Committee at the Asan Institute for Life Sciences (Seoul, Korea).

Diabetes was induced by a single intraperitoneal injection of 65 mg/kg STZ (Sigma, St. Louis, MO), as previously described (15). Control rats received an intraperitoneal injection of an equal volume of citrate buffer. Successful induction of diabetes was defined as a blood glucose level of >11.1 mmol/l. Three days after STZ injection, the diabetic rats were randomly assigned to receive insulin treatment or no treatment (n = 10 each group). For the insulin-treated group, intermediate-acting insulin (Humulin-N; Eli Lilly, Indianapolis, IN) was administered each day just before the dark period (between 1600 and 1700) until the animals were killed. Blood glucose levels were measured each day just before insulin injection. Doses of insulin (5–7 units per rat) were adjusted to reach target blood glucose levels of 7.8 to 11.1 mmol/l, and food intake and body weight were monitored daily.

On the 21st day, the rats were killed by decapitation between 0900 and 1100 following an overnight fast. The medial part of the hypothalamus was dissected by the anterior border of the optic chiasm, the posterior border of the mammillary body, the upper border of the anterior commissure, and the lateral border half way from the lateral sulcus in the ventral side of brain. By microscopic examination, we confirmed that each medial hypothalamus included the arcuate nucleus, paraventricular nucleus, and ventromedial and dorsomedial nuclei. Liver and skeletal muscle tissues were obtained after freezing in situ by aluminum tongs precooled in liquid nitrogen (16). Trunk blood was obtained, and plasma was separated by centrifugation (400 g) for 15 min at 4°C. Blood and tissues were stored at −70°C until assayed.

Rats administered vehicle or STZ (n = 5–7) were killed 21 days later, and the whole brains were rapidly removed and frozen in isopentane on dry ice. Individual hypothalamic nuclei were dissected using a “micropunch” technique as described previously (17).

Twenty-three gauge stainless steel cannulae (Plastics One, Roanoke, VA) were implanted into the third cerebral ventricle (ICV) of rats as previously described (18). Following a 7-day recovery period, the correct positioning of each cannula was confirmed by a positive dipsogenic response to ICV administration of angiotensin II (150 ng/rat). Diabetes was induced by STZ as described above. Twenty-one days later and following a 24-h fast, rats (n = 6–7) were injected intracerebroventricularly with 10 μl of vehicle (saline or DMSO), leptin (0.5 nmol), insulin (300 nmol), or the AMPK inhibitor, compound C (100 and 300 nmol; Merck, Whitehouse Station, NJ) (19) in the early light phase. Food intake was monitored for 1 h postinjection, and the rats were killed by decapitation. The medial hypothalamus was harvested for determination of AMPK activity or mRNA expression of neuropeptides.

Plasmids encoding c-myc-tagged forms of dominant-negative α1-AMPK with a mutation altering Asp-157 to alanine (20) and of α2-AMPK with a mutation altering Lys-45 to arginine (21) were a gift from Dr. J. Ha (Department of Molecular Biology, Kyunghee University College of Medicine, Seoul, Korea). On the 21st day after diabetes induction, rats were randomly divided into two groups (n = 9–10). Rats were injected with either 1 μl of adenoviruses expressing β-gal (10¹¹ pfu/ml) or a mixture (1:1 vol) of adenoviruses expressing dominant-negative α1-AMPK and dominant-negative α2-AMPK (10¹¹ pfu/ml) into the bilateral mediobasal hypothalamus as previously described (12). Injection site was confirmed by β-gal staining as previously described (22). Food intake and body weight were monitored daily. In the early light phase (0900 to 1000) of the 5th day, rats were killed by decapitation, and the medial hypothalamus was collected.

Immunoblot analysis was conducted as previously described (23) using antibodies directed against the phosphorylated (Thr-172) and total forms of AMPK (Cell Signaling, Beverly, MA) or phosphorylated (Ser-79) acetyl-CoA carboxylase (ACC) (Upstate Biotech, Waltham, MA). To measure total ACC protein expression, membrane was incubated with streptavidin instead of primary antibody. Band density was corrected by the density of β-actin.

For measurement of isoform-specific AMPK activity, 40 μg of tissue lysates was immunoprecipitated by incubation with specific antibodies against the α1- and α2-AMPK catalytic subunits (Upstate) and 15 μl of 25% (wt/vol) protein G-sepharose beads (Santa Cruz Biotechnology, Santa Cruz, CA) overnight at 4°C. AMPK activity was determined using the modified method of Davies et al. (24) with these fractionated proteins in kinase assay buffer (62.5 mmol/l HEPES, pH 7.0, 62.5 mmol/l NaCl, 62.5 mmol/l NaF, 6.25 mmol/l sodium pyrophosphate, 1.25 mmol/l EDTA, 1.25 mmol/l EGTA, and 1 mmol/l dithiothreitol) containing 200 μmol/l AMP, ATP mixture (200 μmol/l ATP and 1.5 μCi of [-³²P]ATP), with or without 200 μmol/l SAMS peptide (HMRSAMSGLHLVKRR) at 30°C for 10 min. The reaction was terminated by spotting the reaction mixture on phosphocellulose paper (P81), and the paper was extensively washed with 150 mmol/l phosphoric acid. The radioactivity was measured with a scintillation counter.

Hypothalamic carnitine palmitoyltransferase-1 (CPT-1) activity was measured as described (25).

NPY and POMC mRNA expression was quantified by semiquantitative RT-PCR using primer sets for NPY (5′-TAG GTA ACA AAC GAA TGG GG-3′ and 5′-GTC TTC AAG CCT TGT TCT GG-3′), POMC (5′-ATG CCG AGA TTC TGC TAC AG-3′ and 5′-ATG ATG GCG TTC TTG AAG AG-3′), or glyceraldehyde-3-phosphate dehydrogenase (5′-ACC ACA GTC CAT GCC ATC AC-3′ and 5′-TCC ACC ACC CTG TTG CTG TA-3′). The amplification protocol consisted of 27 cycles of denaturation at 94°C for 30 s, annealing at 56°C for 30 s, and extension at 72°C for 2 min.

Plasma glucose was determined using a glucose analyzer (YSI 2300; Yellow Springs Instruments, Yellow Springs, OH). Plasma insulin and leptin concentrations were determined by radioimmunoassay (Linco, St. Louis, MO).

All data are presented as mean ± SE. Comparisons between groups were by unpaired Student’s t test or ANOVA followed by the post hoc least significance difference test. Significance was defined as P < 0.05.

Weight gain during the study period was significantly less in diabetic rats than in control rats (Table 1). Epididymal fat mass at the end of the study was also less in the diabetic rats (Table 1). Average food intake was greater in the diabetic rats than in control rats (Table 1). Plasma insulin and leptin levels at the end of the study period were significantly lower in the diabetic rats than in the control rats (Table 1).

Hypothalamic AMPK phosphorylation, an indicator of AMPK activation, was about two times higher in diabetic rats compared with control rats (Fig. 1A). Hypothalamic total AMPK protein levels tended to increase in diabetic rats, but the difference did not reach statistical significance (Fig. 1B). α2-AMPK activity was about 1.5-fold higher in the hypothalamus of diabetic rats, while there was no significant difference in hypothalamic α1-AMPK activity (Fig. 1C and D). α2-AMPK protein levels in the hypothalamus were not different between control and diabetic rats (data not shown).

When we analyzed AMPK activities in discrete hypothalamic nuclei, α2-AMPK activity was significantly higher in the medial part of the hypothalamus, including the arcuate nucleus, paraventricular nucleus, and ventromedial and dorsomedial nuclei of diabetic rats compared with similar structures in control rats (Fig. 1E). In contrast to the hypothalamus, α2-AMPK activities were lower in the livers of diabetic rats compared with control rats, but there were no differences in skeletal muscle α2-AMPK activities (Fig. 1F).

AMPK inhibits ACC activity through phosphorylation (9,10). In the hypothalamus of diabetic rats, ACC phosphorylation was higher (i.e., ACC activity was lower) compared with control rats (Fig. 2A). Hypothalamic α-ACC phosphorylation was 2.5-fold higher and hypothalamic β-ACC phosphorylation was 5.8-fold higher in diabetic rats than in control rats (Fig. 2B and C). Total ACC protein expression was not changed by induction of diabetes (Fig. 2A).

Reduced ACC activity has been shown to increase fat oxidation in skeletal muscle by decreasing the intracellular level of malonyl CoA, a potent inhibitor of CPT-1 (9,10). Consistent with the changes in hypothalamic ACC activity, we found that hypothalamic CPT-1 activity was higher in diabetic rats than in control rats (Fig. 2D).

Consistent with previous reports (6–8), chronic insulin treatment completely reversed diabetes-induced changes in body weight, fat weight, and food intake, as well as in plasma leptin and insulin levels (Table 1). In contrast, plasma glucose levels were still higher in the insulin-treated diabetic rats than in control rats.

In insulin-treated diabetic rats, hypothalamic AMPK phosphorylation decreased to the level observed in control rats (Fig. 1A and B). Although insulin treatment reduced hypothalamic α2-AMPK activity, it had no effect on α1-AMPK activity (Fig. 1C and D). Insulin treatment decreased hypothalamic α-ACC and β-ACC phosphorylation to a level below that observed in control rats (Fig. 2A–C). Increased hypothalamic CPT-1 activity in diabetic rats was also reduced by chronic treatment of insulin (Fig. 2D).

To investigate whether enhanced hypothalamic AMPK activity contributes to diabetic hyperphagia, compound C, an AMPK inhibitor (19), was administered intracerebroventricularly in diabetic rats. While diabetic rats consumed larger amounts of food than control rats after a 24-h fast, ICV administration of compound C (100 and 300 nmol) reduced food intake to the level of nondiabetic control rats (Fig. 3A). Hypothalamic α2-AMPK activity was increased in diabetic rats and reduced by ICV administration of compound C (Fig. 3B). Expression of mRNA encoding the orexigenic NPY, which was higher in the hypothalamus of diabetic rats, was decreased by ICV administration of compound C (Fig. 3C). In contrast, anorexigenic POMC mRNA was increased by ICV compound C (Fig. 3D).

To ensure the effect of AMPK inhibition on diabetic hyperphagia, adenovirus-containing dominant-negative AMPK was injected into the bilateral mediobasal hypothalamus of diabetic rats. Overexpression of dominant-negative AMPK decreased total hypothalamic AMPK activity by ∼30% (Fig. 3E). Food intake was significantly reduced in rats expressing dominant-negative AMPK in their hypothalamus (Fig. 3F).

Plasma leptin and insulin levels were significantly lower in diabetic rats compared with control rats (Table 1). It has been shown that administration of leptin or insulin reduces hypothalamic AMPK activities (13,14). To test if the deficiency of these two hormones and the resultant hypothalamic AMPK activation may contribute to diabetic hyperphagia, we administered leptin and insulin into the third ventricle of diabetic rats. ICV administration of leptin and insulin reduced food intake as well as α2-AMPK activity in diabetic rats (Fig. 4A and B). Similarly, ICV leptin and insulin reversed diabetes-induced changes in NPY and POMC mRNA expression (Fig. 4, C and D).

We have shown here that AMPK phosphorylation and α2-AMPK activities are increased in the hypothalamus of STZ-induced diabetic rats. Because food intake is increased by AMPK activation in the hypothalamus but reduced by AMPK inhibition (12–14), our results suggest that the enhanced hypothalamic AMPK activity observed in diabetic rats can cause this increase in food intake. This notion is supported by our finding that hypothalamic AMPK inhibition reduced diabetic hyperphagia.

The mechanism by which hypothalamic AMPK regulates food intake, however, remains unclear. AMPK activation is known to decrease ACC activity and intracellular levels of malonyl CoA. Malonyl CoA has been hypothesized to act as an inhibitor of food intake (26). Thus, a reduction in malonyl CoA levels could increase food intake. Alternatively, reduced malonyl CoA levels stimulate CPT-1 activity (9,10). Chemical or molecular inhibition of hypothalamic CPT-1 activity has been shown to decrease food intake (17). Thus, enhanced hypothalamic CPT-1 activity in diabetic rats may cause hyperphagia. In addition, we have shown that ICV administration of the AMPK inhibitor compound C reduced orexigenic NPY mRNA levels and increased anorexigenic POMC mRNA levels. Thus, changes in hypothalamic AMPK activity may also affect the expression of mRNA-encoding hypothalamic neuropeptides through as yet unknown mechanisms.

A number of factors may be involved in the diabetes-induced changes in hypothalamic AMPK. In our study, plasma insulin and leptin concentration was profoundly reduced in diabetic animals. Moreover, ICV administration of insulin and leptin inhibited the increase in hypothalamic AMPK activity and food intake. Thus, deficiencies of leptin and insulin may be a major contributor to the increase in hypothalamic AMPK activity observed in diabetic rats. Previous studies have shown that hypothalamic AMPK activity was decreased by glucose (12,13). Taken together, inhibition of hypothalamic AMPK by high glucose in chronic diabetic state may be insufficient to overcome AMPK activation by insulin and leptin deficiency.

Interestingly, diabetes-induced changes in AMPK activity were tissue specific. We found that AMPK activity was increased in the hypothalamus of diabetic rats but decreased in the liver and unchanged in skeletal muscle. Similarly, administration of leptin (27) and α–lipoic acid (12) (K.H. Song, J.Y.P., J.M. Koh, H.S. Kim, H.S.P., H.J. Park, M.S.K., J. Ha, J.H. Youn, K.U.L., unpublished data) stimulated AMPK activity in skeletal muscle but suppressed this activity in the hypothalamus. Additional studies are needed to clarify the molecular mechanism by which similar metabolic conditions or agents cause differential changes in AMPK activity in different tissues.

In summary, we have shown here that hypothalamic AMPK phosphorylation and activity were increased in diabetic rats. Inhibition of hypothalamic AMPK activity decreased food intake and normalized hypothalamic neuropeptide expression in diabetic rats. These results demonstrate that increased hypothalamic AMPK activity contributes to the development of diabetic hyperphagia.

This work was supported by a grant from the Korean Ministry of Health and Welfare (03-PJ1-PG1-CH05-0005), National Research Laboratory grant from the Korean Ministry of Science and Technology (M1040000000804J000000810) and Asan Institute of Life Sciences (04-326).

The authors thank Dr. Juhun Ha (Kyunghee University) for providing dominant-negative AMPK-expressing plasmids.