INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

LEPTIN, THE PROTEIN PRODUCT of the ob/ob gene (33), is produced by adipocytes (7) in direct proportion to adiposity in rats and humans (6, 20). Once produced, it is free to circulate in the bloodstream in both free and protein-bound form (27). When injected intravenously or intraperitoneally into normal and ob/ob rodents, leptin reduces feeding and adiposity and increases lean body mass (4, 12, 22). Conversely, the absence of leptin (33), or a defective receptor for leptin (16), leads to hyperphagia and grossly overweight animals. The above studies suggest that leptin is a peripheral signal of adiposity that may be involved in a negative feedback loop regulating body mass and appetite.

The results from intracerebroventricular injections of leptin are consistent with intravenous and intraperitoneal injection studies (4) and provide evidence in favor of a central nervous system (CNS) center for leptin negative feedback. In order for peripherally produced leptin to act within the brain, it must cross the blood-brain barrier, a process that has been demonstrated in both humans (10) and mice (3). Leptin receptor mRNA has been localized extensively to the CNS (11, 14), and although leptin has been shown to activate neurons (as measured by an increase in c-fos), in many nuclei of the hypothalamus, including the paraventricular nucleus (PVN) (5, 30, 31), a definitive effect of this peptide on the excitability of neurons remains to be determined.

The PVN was first implicated in the regulation of feeding and metabolism 50 years ago when lesions of this structure led to hyperphagia and obesity in dogs (13). Since that time numerous studies in which the PVN was lesioned or disconnected from surrounding tissues have produced a similar syndrome in rodents (19). These isolation studies (1, 9, 26) indicate that this nucleus is essential in the normal control of body weight homeostasis. Numerous substances, including neuropeptide Y (28), morphine (32), and GABA receptor agonists (15), stimulate feeding when microinjected into the PVN, whereas a number of gut peptides and monoamines act at the PVN, inhibiting feeding (18). The PVN is also the primary site mediating the norepinephrine-induced stimulation of feeding (17).

This study, using whole cell recording techniques, examines the effects of leptin on PVN neurons in a coronal hypothalamic slice. The responsiveness of PVN neurons to leptin will help define more clearly the neuronal mechanisms underlying the physiological control of body weight homeostasis.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Slice preparation. Male Sprague-Dawley rats (150-250 g; Charles River) were killed by decapitation, and the brain was quickly removed from the skull and immersed in cold (1-4°C) artificial cerebrospinal fluid (aCSF). The brain was blocked, and 400-µm hypothalamic slices, including the PVN, were prepared as described previously (2).

Electrophysiological techniques. Whole cell recordings were obtained using pipettes (resistance of 4-6 M) filled with a solution containing (in mM) 140 K-gluconate, 0.1 CaCl₂, 2 MgCl₂, 1.1 EGTA, 10 HEPES, 2 Na₂ATP, and adjusted at pH 7.25 with KOH. An Ag-AgCl electrode connected to the bath solution via a KCl-agar bridge served as reference. All signals were processed with an Axoclamp-2A amplifier, digitized using the CED 1401 plus interface, and stored on computer for offline analysis.

Solutions. The aCSF composition was (in mM) 124 NaCl, 2 KCl, 1.25 KPO₄, 2.0 CaCl₂, 1.3 MgSO₄, 20 NaHCO₃, and 10 glucose. Osmolarity was maintained between 285 and 300 mosM and pH between 7.3 and 7.4. Leptin fragment 2256 (Phoenix Pharmaceuticals) was dissolved in aCSF to a stock concentration of 1 µM and then frozen until use. The leptin fragment 5792 (Phoenix Pharmaceuticals) was first dissolved in 50-100 µl of acetonitrile, then diluted to a 1 µM stock concentration with aCSF and frozen until use. The 2256 leptin fragment has been shown to induce significant satiety effects in the rat, whereas leptin fragment 5792 yielded no significant alteration in feeding behavior (25).

All drugs were bath applied at the concentrations specified in the text and legends to all figures. The timing of drug application (2-8 s, resulting in a volume of 200-800 µl), is indicated by the bars in Fig. 1, which represent the time period for which the peptide in aCSF delivers fluid to the bath entry line. There is a delay of 30-45 s between this time and access on the peptide into the bath.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Whole cell recordings were obtained from 59 PVN neurons. These cells had resting membrane potentials negative to 55 mV and input resistances greater than 750 M. Before leptin application, all cells were electrophysiologically classified according to previously established criteria (29). Thirty-two cells exhibited properties (delay to first spike, linear current-voltage relationship) that are characteristic of type I PVN neurons. Twenty cells exhibited properties (low-threshold Ca²⁺ spike, inward rectification of current-voltage relationship at negative potentials) similar to type II cells. Seven cells could not be clearly differentiated into either type I or type II.

Effects of leptin on membrane potential. Bath application of 10 nM leptin to a slice previously unexposed to leptin depolarized 82% of the type I cells tested (n = 17). In addition, 67% of the type II cells tested (n = 9) were also depolarized by the active leptin fragment. After a minimum of 30 min, some slices were retested with a 10 nM dose of leptin (n = 9). Six of the cells tested in previously exposed slices did not depolarize, suggesting that there may be a slow desensitization to leptin. According to our calculations, this dose and application protocol resulted in the delivery of 2-8 pmol (200-800 µl of 10 nM solution) of leptin to the slice. Increasing the dose of leptin produced more pronounced depolarizations (n = 5) (Fig. 1). In some of the cells tested with the 100 nM dose, there was no recovery of the membrane potential even after 15 min (n = 2). The minimum dose required to elicit an effect on membrane potential was 0.1 nM (n = 5). The dose-response curve generated from testing doses from 0.1 to 100 nM is depicted in Fig. 2. From this curve, the EC₅₀ was determined to be 29 nM. Similar depolarizations were observed in response to bath application of leptin during synaptic isolation of the cells by bath application of the sodium channel blocker tetrodotoxin (2 µM, n = 5). To rule out the possibility of nonspecific actions of this peptide on the membrane potential of PVN neurons, seven cells were tested with the physiologically inactive 5792 leptin fragment. This fragment has previously been shown to have no significant effect on feeding behavior following intracerebroventricular application (25). In contrast to the depolarization induced by leptin-(2256), bath application of 10 nM of leptin-(5792) had no discernible effect on the membrane potential (Fig. 1). In three cells that did not depolarize in response to the inactive leptin fragment, reversible depolarizations were observed following application of the active leptin fragment.

View larger version (39K): [in this window] [in a new window]   Fig. 1.   Current clamp recordings from paraventricular nucleus (PVN) neurons demonstrate that leptin depolarizes cells. A: depolarization (18 mV) in response to bath application of 100 nM leptin in a type II cell. Downward deflections represent voltage responses to hyperpolarizing current pulses. B: 9-mV depolarization in a type I neuron following 10 nM leptin. This cell was previously shown to be unresponsive to the inactive leptin fragment 5792 (C). Bar indicates time of bath application of leptin.

View larger version (9K): [in this window] [in a new window]   Fig. 2.   Dose-response relationship of leptin-induced depolarizations. Membrane responses of all cells tested with leptin-(2256) are summarized in the dose-response curve shown above. Each point () represents mean of all depolarizations produced at the representative dose, and error bars indicate ±SE; 0.1 nM, n = 5; 1.0 nM, n = 4; 10 nM, n = 12; 100 nM, n = 5. Effects of inactive leptin fragment are also shown ().

Effects of leptin on membrane conductance. Current-voltage relationships were examined in six cells using slow voltage ramps (5 mV/s) under whole cell voltage clamp conditions. Under control conditions, these ramps elicited voltage-dependent currents (indicated by nonlinearity over voltage range) with a reversal potential (zero current) between 25 and 30 mV (see Fig. 3), indicating this current to be carried by a nonspecific mix of cations. Bath application of the active leptin fragment increased the amplitude of the inward current elicited by these ramps. Subtraction of control current from leptin current revealed that the difference (i.e., leptin induced) current was linear (Fig. 3B), indicating an effect of the peptide on a voltage-independent conductance. In addition, the observation that large inward currents were induced by leptin at voltages around the resting membrane potential of the cells (55 mV) are consistent with the depolarizing actions of this peptide observed in current clamp recordings. This difference current (Fig. 3B) exhibited a reversal potential of 28.2 ± 1.5 mV (n = 6), which is again consistent with the activation of a conductance carried by nonspecific mix of cations.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Current-voltage relationship in PVN neurons of leptin response in PVN neurons. A: traces shown are current response to voltage ramps applied to a PVN neuron in whole cell voltage clamp mode. Cell was ramped in artificial cerebrospinal fluid (control), and then leptin was bath applied. After a 60-s waiting period, voltage ramps were applied (leptin). B: subtraction of control current from current following peptide application reveals leptin-induced nonrectifying whole cell current (linear regression analysis r2 = 0.98, conductance 5 nS) that reverses at 28 mV.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The results of this study demonstrate that the physiologically active leptin fragment 2256 depolarizes neurons in the PVN. The effects of leptin likely result from the activation of a nonspecific cationic conductance as indicated by our observation that the difference current (leptin  control) in response to slow voltage ramps reverses at about 30 mV. These effects were observed in both type I and type II neurons, suggesting that the initiation of a cellular signal by this peptide may have diverse physiological consequences. The lack of effect following application of the inactive fragment suggests the response observed in response to active leptin is likely transduced by receptors that are specific for this fragment. This effect of leptin, which brings the membrane potential of PVN cells closer to the threshold for spike activation, is consistent with studies demonstrating an increase in the expression of c-fos in PVN following leptin application (5, 30, 31). The dose-related effects observed in the present study are consistent with previous work demonstrating a dose dependency of leptin effects on feeding (4, 22, 25).

The mechanisms responsible for the requisite decrease in feeding behavior following this activation of PVN neurons are unclear, but when examined in the context of previously defined effects of other substances that influence feeding behavior, a clear picture regarding the role of PVN as a satiety center begins to emerge. Substances that increase feeding behavior, such as neuropeptide Y, norepinephrine, GABA, morphine, and other opiates, inhibit neuronal firing in the PVN (21, 23, 24). In agreement with this, we report that leptin, a peptide that acts as a satiety signal and inhibits feeding, depolarizes PVN neurons, thereby increasing the likelihood they will discharge action potentials. This provides further evidence that regulating the excitability of cells within this nucleus may be a critical step in controlling feeding behavior.

In a previous study, leptin decreased input resistance and inhibited evoked excitatory postsynaptic currents in the arcuate nucleus (8). Although the doses of leptin used in the two studies, 25 nM in the arcuate nucleus versus 10 nM for a consistent depolarization in PVN, are comparable, the durations of leptin application (1-2 min vs. 2-8 s) are very different. Consequently, the neurons in the arcuate nucleus were exposed to a much larger effective dose than those in the PVN. According to our calculations, PVN neurons were exposed to between 2 and 8 pmol of leptin. These findings suggest that neurons in PVN may have a greater sensitivity to leptin than neurons in the arcuate nucleus. We would propose that PVN, although not the exclusive site of leptin actions, may be the primary central site coordinating the physiological response to leptin.