Low-dose infusions of leptin into the nucleus of the solitary tract increase sensitivity to third ventricle leptin

Abstract

Previous studies suggest that weight loss occurs when leptin receptors in both the forebrain and hindbrain are activated. Experiments described here tested whether this integration is mediated through a neural connection or by leptin diffusion through the subarachanoid space. If the hypothalamus and hindbrain communicated through a neural pathway, then a very low dose of leptin infused directly into the nucleus of the solitary tract (NTS) would enhance the response to third ventricle (3V) leptin but would have no effect if infused into the fourth ventricle (4V). A 12-day infusion of 10 ng/24 h into the 4V or the NTS reduced body fat. Leptin at 5 ng/24 h into the 4V or NTS had no effect on food intake or body composition, but infusion of 5 ng of leptin/24 h into the NTS combined with a 3V injection of 0.1 μg of leptin inhibited food intake between 6 and 12 h after injection. Cumulative intake was inhibited for up to 36 h. 3V leptin had no effect on food intake of rats receiving the 4V leptin infusion. Similar results were found using infusions of 5 ng leptin/24 h and a 3V injection of 0.025 μg leptin. These data suggest that activation of leptin receptors in the NTS lowers the threshold for response to leptin in the forebrain through a neural network.

INTRODUCTION

A majority of studies examining the central control of energy balance by leptin have focused on the arcuate nucleus of the hypothalamus (Arc) (33) and the nucleus of the solitary tract (NTS) in the hindbrain (10). Recently, we have used threshold and very low-dose injections and infusions of leptin to explore the possibility of an integrated central response to leptin. Our data suggest that leptin receptors in both the hypothalamus and the hindbrain have to be activated to elicit a suppression of food intake and weight loss (4, 12). In a majority of experiments in which exogenous leptin is injected into the brain, it is assumed that only the leptin receptors in the vicinity of the injection are activated, whereas we and others have shown that injection (26) or infusions (16) of leptin into the fourth ventricle (4V) or NTS increases phosphorylated signal transducer and activator of transcription 3 (pSTAT3) in hypothalamic nuclei. pSTAT3 is accepted as a marker of leptin receptor activation (30) and has also been shown to be essential for the effects of leptin on body weight (1). A study by Ruiter et al. (26) found that injection of 150 ng of leptin into the 4V or 50 ng injected into the NTS increased pSTAT3 in multiple hypothalamic nuclei. Subsequently, we found that subthreshold doses of leptin infused into the third ventricle (3V) or 4V independently had no effect on food intake, body weight, or body composition of rats. If, however, the 3V and 4V infusions occurred simultaneously, then there was a substantial inhibition of food intake and a significant loss of weight in the rats during a 12-day infusion (4, 12). Additionally, we have shown that weight loss in rats receiving 4V infusions of leptin occurs only when the hindbrain infusion is associated with an increase in hypothalamic pSTAT3 (16). Examination of pSTAT3 distribution in the hypothalamus showed a significant increase in the ventromedial, dorsomedial, and arcuate nuclei of the hypothalamus when rats received the simultaneous infusions (5). Quantification of ΔFosB, a marker of chronic neuronal activation, demonstrated that the only areas that showed evidence of chronic activation were those that also showed increased pSTAT3 and are known to express significant levels of leptin receptors (5).

A recent study tested whether activation of hindbrain leptin receptors lowered the threshold for response to leptin in the forebrain or whether leptin in the forebrain lowered the threshold for response in the hindbrain. Constant low-dose infusions of leptin into the 4V lowered the threshold for a response to an acute injection of leptin into the 3V, but constant infusion of leptin into the 3V did not exaggerate the response to acute injections into the 4V (17). The results from these studies suggested that activation of hindbrain leptin receptors lowers the threshold for activation of hypothalamic leptin receptors so that they respond to exogenous or endogenous concentrations of leptin that are ineffective if the hindbrain leptin receptors are not stimulated. The objective of the experiments described here was to initiate an exploration of whether the connection between hindbrain and forebrain leptin receptors is neural or humoral due to diffusions of leptin from the hindbrain through the subarachanoid space. Ruiter et al. (26) examined the diffusion of a gold thioglucose injection into the 4V and found that it had not reached the hypothalamus in the time it took for hypothalamic pSTAT3 to increase in rats receiving NTS injections of leptin, leading to the assumption that there is a direct neural connection between leptin-responsive cells in the NTS and leptin responsive cells in select hypothalamic. This assumption is further tested here by infusing extraordinarily low doses of leptin into either the 4V or the NTS of rats and then testing the response to a low-dose injection of leptin into the 3V. We hypothesized that if the lowered threshold for response to leptin was dependent on a neural connection between the hindbrain and hypothalamus, then there would be a low level of leptin infusion that would enhance the response to leptin injection when the infusion was into the NTS but not when the same dose of leptin was infused into the 4V.

METHODS

All animal procedures were approved by the Institutional Animal Care and Use Committee at Augusta University, and animals were maintained in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (23). The rats used in these studies were male Sprague-Dawley rats (Envigo, Prattville, AL) weighing 275–300 g at the start of a study. They were individually housed in wire mesh cages, except for those that were placed in calorimetry cages. They had continuous free access to chow and water and a Nylabone (Nylabone Products, Neptune, NJ) for enrichment. The room was maintained at 21–22°C (70–72°F) and ~50% humidity, with the lights on from 7:00 AM to 7:00 PM. Rats were weighed daily.

Experiment 1: NTS vs. 4V leptin infusion dose response.

Twenty-four rats were fitted with either a 4V or a unilateral NTS 26-gauge infusion cannula (Plastics One, Roanoke, VA). They were anesthetized with 90 mg/kg ketamine and 10 mg/kg xylazine given by intraperitoneal injection. Coordinates for 4V cannula placement on the midline were 2.5 mm anterior to the occipital suture and 6.5 mm ventrodorsal. Coordinates for the NTS were 13.1 mm posterior to the bregma, 0.5 mm lateral to the midline, and 8.3 mm ventrodorsal. Placement was according to the Paxinos and Watson brain atlas (24). Rats received a subcutaneous injection of analgesic (2 mg/kg ketoprofen; Fort Dodge Animal Health, Fort Dodge, IA) immediately before surgery and again the next day. Five days after surgery, the rats were housed in a calorimeter (TSE LabMaster, Metabolic Research Platform; TSE Systems International, Chesterfield, MO). Oxygen consumption, and carbon dioxide production were sampled for 3 min from each cage every 39 min. Values measured during the last min were used to calculate energy expenditure, expressed both as kilocalories per hour per rat and per unit metabolic body weight (kcal·h⁻¹·weight^0.75) and respiratory exchange ratio (RER) as an index of macronutrient oxidation. Food intake data were reported every 39 min, although it was collected every minute. Body weight was recorded manually at 7:30 AM each morning, when food hoppers and water bottles were refilled, and cage bedding was changed every second day. Calorimetry measures were initiated at 8:00 AM and stopped at 7:20 AM the next day so that only one cycle of measurement was lost each day. The experiment was completed with two cohorts of rats because we had only 12 calorimetry cages.

After 3 days of adaptation to the calorimeter, the rats were anesthetized with isoflurane, and an Alzet miniosmotic pump was connected to the infusion cannula. The pumps (Model 1004; DURECT, Cupertino, CA) delivered 0, 5, or 10 ng leptin/24 h in a volume of 0.11 μl/h (rat recombinant leptin; R&D Systems, Minneapolis, MN). On day 14 of infusion, food was removed from the cages at 7:00 AM, and starting at 10:00 AM rats were euthanized by decapitation. Because the brain was not fixed, it was not possible to confirm cannula placement in this experiment. Inguinal, epididymal, retroperitoneal, and mesenteric fat depots were dissected, weighed, and returned to the carcass. The carcass less the gastrointestinal tract was analyzed for composition as described previously (13).

Experiment 2: effect of NTS vs. 4V leptin infusion on the response to 3V leptin injection.

Forty male Sprague-Dawley rats were fitted with either a 4V or a unilateral NTS infusion cannula, as described above, and a 3V guide cannula, as described previously (3). Two rats were removed from the experiment due to a misplaced 3V cannula. Placement of the 3V cannula was tested 5 days after surgery, as described previously (3), and the rats were housed in the calorimeter for the same measures as described for experiment 1. The experiment was completed with four cohorts of rats, with all treatments represented in each cohort. After a 3-day adaptation to the calorimeter, the rats were anesthetized with isoflurane and an Alzet minosmotic pump (Model 1004) delivering either saline or 5 ng leptin/24 h attached to the infusion cannula. At the same time as the pump was attached, an iButton (Embedded Data Systems, Lawrenceburg, KY) was placed over the intrascapular brown fat (iBAT) of the rat and was programmed to report temperature every 30 min until the end of the experiment. Two NTS-infused rats stopped eating at the start of infusion and were removed from the experiment. Three rats lost a cannula during the infusion period and were removed from the experiment. At the end of the study, the placement of NTS cannula could not be confirmed by histology in two rats, and their data were removed from the experiment. It was not possible to confirm the location of 4V infusion cannula because brain sections were collected for free-floating immunohistochemistry and the cerebellum separated from the brain stem. Thirty-one rats completed the experiment. The distribution among groups was 7 4V saline, 7 4V leptin, 8 NTS saline, and 9 NTS leptin.

On days 4 and 9 of infusion, food was removed from the cages at 7:00 AM. Starting at 5:00 PM, rats received 2-μl 3V injections containing 0.1 μg of leptin or saline. The injections were delivered over a 1-min period using an infusion pump (PHD 2000 Infusion Pump; Harvard Apparatus, Holliston, MA). At 6:00 PM, food was returned to the cages. Body weights of the rats were recorded at 14, 24, 38, and 62 h after injection. Rats acted as their own controls so that those injected with leptin on day 4 were injected with saline on day 9 and vice versa. On day 14 of infusion, food was removed from the cages at 7:00 AM, and starting at 1:00 PM rats were anesthetized with ketamine-xylazine and perfused pericardially first with 300 ml of heparinized saline and then with 500 ml of 4% paraformaldehyde. The brains were removed and held overnight in paraformaldehyde at 4°C and then transferred to 25% sucrose solution containing 1% sodium azide. Thirty-micrometer sections were made through the hindbrain and hypothalamus, and pSTAT3 was detected in every fourth section by immunohistochemistry, as described previously (4) using ABC reagents and DAB (Vector Laboratories, Burlingame, CA). The primary antibody was from Cell Signaling Technology (no. 9145; pStat3 (Y705) XP Rabbit mAb), and the secondary antibody was anti-rabbit (no. BP9100; RTU biotinylated antibody anti-rabbit IgG (H⁺L), Vector Laboratories), and specificity was confirmed by performing immunohistochemistry in the absence of primary antibody (data not shown). The number of pSTAT3-positive nuclei within defined brain areas was counted manually.

Experiment 3: effect of NTS vs. 4V leptin infusion on hypothalamic STAT3 activation following 3V leptin injection.

There was no significant effect of either NTS or 4V infusions of 5 ng leptin/24 h on hypothalamic or NTS pSTAT3 expression in the rats from experiment 2. This tissue was collected without leptin stimulation of the forebrain; therefore, this third experiment was designed to test whether there was an increase in hypothalamic pSTAT3 levels in rats receiving 4V or NTS infusions of 5 ng leptin/24 h immediately after a leptin injection. A lower dose of leptin was injected than in experiment 2 to ensure that we would be able to detect whether the leptin infusions modified the response to 3V leptin.

Thirty-two rats were fitted with NTS or 4V infusion cannulae and 3V guide cannulae as described above. Five days after surgery, they were housed in the calorimeter, and placement of the 3V cannulae was tested. The rats were adapted to the calorimeter for 3 days and then miniosmotic pumps delivering either 0.9% saline or 5 ng leptin in saline per 24 h were connected to the cannula. Three rats were removed from the study because of misplaced 3V cannulae, and three were removed because they stopped eating as soon as the infusions started. The final groups sizes were five 4V saline, seven 4V leptin, seven NTS saline, and seven NTS leptin. On days 4 and 7 of infusion, the rats were food deprived from 8:00 AM to 5:00 PM, at which time they received a 3V injection of either saline or 0.025 μg of leptin. Food was returned to the cages at 6:00 PM. Each rat acted as its own control, as described above. On day 9 of infusion, food was removed from the cages at 8:00 AM. Starting at 11:30 AM, rats received a 3V injection of either saline or 0.025 μg of leptin. Exactly 30 min later, the rats were anesthetized with ketamine-xylazine. One inguinal and one epididymal fat pad were dissected and weighed. Exactly 40 min after the 3V injection, the rats were perfused and the brains collected for immunohistochemical detection of pSTAT3 as described above.

Data analysis.

Statistically significant differences in daily food intake, body weight, energy expenditure, RER, or leptin response were determined by repeated-measures ANOVA. Initially, all animals were included in the analysis, but then treatment effects within groups of 4V or NTS infused rats were determined. If a significant effect of infusion or injection was identified, then data for individual time points or days were analyzed by one-way ANOVA and post hoc Duncan’s multiple range test. Rat identification number was used as a covariant in the analysis of leptin response, where each animal acted as its own control. Data analysis was performed using Statistica v. 9.0 (StatSoft), and P < 0.05 was considered significant.

RESULTS

Experiment 1: NTS vs. 4V leptin infusion dose response.

In this study, small groups (n = 4) of rats were infused with 0, 5, or 10 ng leptin/day to identify a threshold dose that could be used in experiments 2 and 3. When food intake of all groups of rats was compared for all 13 days of infusion, there was a significant effect of leptin infusion (P < 0.002), of time (P < 0.0001) and an interaction between infusion and time (P < 0.02). Post hoc analysis showed that the differences occurred during the first few days of leptin infusion (Fig. 1A), and when total intake during the first 6 days of infusion were compared there was a significant effect of leptin infusion (P < 0.004), with no effect of site of infusion. Post hoc analysis indicated that infusion of 10 ng leptin/day at either site inhibited food intake compared with intake of rats infused with 5 ng leptin/day (Fig. 1B). There was no effect of leptin infusion into either the 4V or NTS on body weight (Fig. 1C), but at the end of the study, carcass fat was significantly lower in rats that had been infused with 10 ng leptin/day than in controls [Fig. 1D: site of infusion, not significant (NS); leptin infusion, P < 0.009; interaction, NS).

Fig. 1.

Food intake (A and B), body weight (C), carcass fat (D), and respiratory exchange ratio (RER) during the dark period (E and F) for experiment 1. Data are means ± SE for groups of 4 rats. Data were analyzed by repeated-measures ANOVA for all rats followed by one-way ANOVA and post hoc Duncan’s multiple range test. Differences were considered significant at P < 0.05. Asterisks indicate that rats infused with 10 ng leptin/day in the NTS were different from rats infused into either the 4th ventricle (4V) or nucleus of the solitary tract (NTS) with 5 ng leptin/day; # indicates that rats infused with 10 ng leptin/day into the 4V were different from rats infused with 5 ng leptin/day into the 4V or NTS or with saline into the NTS. Values in B, D, and F that do not share a common superscript are significantly different. In E, $ indicates that rats infused with 10 ng leptin/day into the 4V are different from those infused with 10 ng/day into the NTS; % indicates that rats infused with saline into the NTS are different from those infused with 5 ng leptin/day into the NTS; ϕ indicates a significant difference between rats infused with saline in the NTS and rats infused with 5 or 10 ng leptin/day into the NTS or 10 ng leptin/day into the 4V.

There was no effect of leptin infusion into either the 4V or NTS on energy expenditure during the day or night (data not shown). There was no effect of leptin infusion on RER during the light period (data not shown), but comparison of daily RER during the dark period indicated a significant effect of time (P < 0.0001) and an interaction between leptin infusion and time (P < 0.0002). Post hoc analysis showed that RER of leptin-infused rats tended to be higher than that of saline-infused rats during the last 6 days of infusion (Fig. 1E), even though there was no difference in food intake of the different groups during the dark period for those days (data not shown). Comparison of the average RER during the dark period for the last 7 days of infusion showed a significant effect of infusion (P < 0.03) that was not apparent during the first 7 days (data not shown). Post hoc analysis indicated that RER was significantly increased in rats infused with 5 ng leptin/day into the NTS or 10 ng leptin/day into the 4V compared with saline-infused controls (Fig. 1F).

Experiment 2: effect of NTS vs. 4V leptin infusion on the response to 3V leptin injection.

Infusion of 5 ng leptin/day into the 4V or the NTS had no significant effect on daily food intake of the rats during the experiment (Fig. 2, A and B). Food intake of all of the rats was reduced on the days of injection, because they were food deprived during the light period. Therefore, intake on these days represents intake during the 13 h after the 3V injection. Leptin infusion into either the 4V or NTS did increase energy expenditure during both the light and dark periods (Fig. 2, C and D: site, NS; leptin, P < 0.03; time, P < 0.001; interactions NS). The effect was significant for the first 5 days of infusion during the light period and for the first 8 days of infusion during the dark period. There was no effect of leptin infusion into either the 4V or NTS on daily RER (data not shown). There was a significant effect of day (P < 0.001) and an interaction between infusion site and day (P < 0.01) for iBAT temperature during the light period. During the dark period, there was a significant effect of the site of infusion (P < 0.02) and a significant effect of day (P < 0.001), but no interaction. Post hoc analysis identified several days during the second week of infusion, in which iBAT temperature was reduced in rats infused with either saline or leptin into the NTS compared with rats receiving 4V infusions of saline. This was apparent during both the light and dark periods (Fig. 2, E and F).

Fig. 2.

Daily food intake (A), body weight (B), energy expenditure during the light period (C) and dark period (D), and inguinal brown adipose tissue (iBAT) temperature during the light (E) and dark period (F) in experiment 2. Data are means ± SE for groups of 7–9 rats. Significant differences were determined by repeated-measures ANOVA followed by one-way ANOVA for each day and post hoc Duncan’s multiple range test. Differences were considered significant at P < 0.05. Asterisk indicates that both leptin-infused groups were different from both saline-infused groups; # indicates that rats receiving 4th ventricle (4V) saline infusions were different from both leptin groups of leptin-infused rats; $ indicates that rats receiving nucleus of the solitary tract (NTS) saline infusions were different from both groups of leptin-infused rats; @ indicates that rats receiving 4V saline infusions were different from rats receiving NTS saline infusions and that rats receiving 4V saline infusions were different from both groups of NTS-infused rats.

Because of the large amount of data collected from the calorimeter, the different parameters were either averaged or summed over 6-h intervals. 3V injections of 0.1 μg of leptin inhibited weight gain in rats receiving 4V infusions of saline (4V saline/leptin) and in rats receiving NTS infusions of leptin (NTS leptin/leptin, Fig. 4A; infusion site, P < 0.02; infusion, NS; injection, P < 0.003, time, P < 0.0001; site × infusion × injection, P < 0.03). The difference in weight gain in response to leptin was significant for 62 h in 4V-infused rats, whereas it was significant for only 38 h in the NTS-infused rats. The weight loss in rats receiving NTS infusions of leptin was associated with a significant inhibition of cumulative food intake from 6 to 36 h after injection (Fig. 3B, infusion site, NS; infusion, NS; injection, NS; time, P < 0.001; site × infusion × injection time, P < 0.04). There tended to be an inhibition of intake following 3V leptin injection in rats receiving 4V infusions of saline, but this did not reach significance for either cumulative intake or for intake during any specific 6-h time period compared with 4V saline/saline rats. During the first 6 h after injection, the intake of 4V saline/leptin rats was significantly lower than that of 4V leptin/leptin rats, but not 4V leptin/saline rats (Fig. 3, B and C). Food intake of NTS leptin/leptin rats was inhibited during the 6–12 h after injection (Fig. 3C; infusion site, NS; infusion, NS; injection, NS; time, P < 0.0001; site × injection × infusion, P < 0.05). After that, there was no difference in intake of the any of the NTS-infused groups, but there was no effort to compensate for the inhibition at the earlier time point, which resulted in the prolonged inhibition of cumulative intake.

Fig. 4.

Respiratory exchange ratio (RER; A) and energy expenditure (B) during the 60 h following an acute 3rd ventricle (3V) injection of 0.1 μg leptin in experiment 2. Data are means ± SE for groups of 7–9 rats. Significant differences were determined by repeated-measures ANOVA for all rats followed by one-way ANOVA for each day within 4th ventricle (4V)- or nucleus of the solitary tract (NTS)-infused groups and post hoc Duncan’s multiple range test. Differences were considered significant at P < 0.05. * indicates that NTS leptin/leptin rats were different from NTS leptin/saline rats; $ indicates that rats receiving NTS infusions of saline were different from those receiving NTS infusions of leptin; χ indicates that NTS saline/saline rats were different from 4V saline/saline rats.

Fig. 3.

Weight gain (A) and food intake (B and C) of rats in experiment 2 during the 62 h following an acute 3rd ventricle (3V) injection of 0.1 μg leptin. Data are means ± SE for 7–9 rats. Significant differences were determined by repeated-measures ANOVA for all rats followed by one-way ANOVA for each day and post hoc Duncan’s multiple range test. Differences were considered significant at P < 0.05. # indicates that 4th ventricle (4V) saline/saline rats were different from 4V saline/leptin rats; % indicates that nucleus of the solitary tract (NTS) leptin/leptin rats were different from NTS leptin/saline and NTS saline/saline rats; * indicates that NTS leptin/leptin rats were different from NTS leptin/saline rats. Values for intervaled food intake at a specific time point that do not share a common superscript are significantly different within the 4V-infused and NTS-infused groups of rats.

When iBAT temperatures of all treatment groups were compared at each time interval 4V saline rats tended to have higher temperatures than NTS leptin rats, and this difference was significant during the 6- to 12-h, 18- to 24-h, 24- to 30-h, and 30- to 36-h intervals (infusion site, P < 0.01; infusion, NS; injection, NS; time, P < 0.001; site × infusion × time, P < 0.05), but there was no effect of leptin injection within either 4V- or NTS-infused groups (data not shown). RER followed food intake and was significantly lower in NTS leptin/leptin than in NTS leptin/saline rats in the 6- to 12-h time interval after injection (Fig. 4A; infusion site, NS; infusion, NS; injection, P < 0.02; time, P < 0.0001; site × injection × time, P < 0.06). There were no other significant differences between groups, and the time effect was associated with diurnal variation. There were no significant differences in energy expenditure between 4V- and NTS-infused rats. Within the NTS-infused groups, saline-infused animals had lower energy expenditures than leptin-infused animals, and this reached significance at three time points during the first 36 h following injection (Fig. 4B; infusion site, NS; infusion, P < 0.03; injection, NS; time, P < 0.0001).

Experiment 3: effect of NTS vs. 4V leptin infusion on hypothalamic STAT3 activation following 3V leptin injection.

There was no effect of infusing 5 ng leptin/day into either the 4V or NTS on daily body weight, food intake, RER during either the light or dark period, or energy expenditure during the dark period. For each of these variables there was no significant effect of site of infusion or of infusion of leptin, but there was a significant effect of time (P < 0.001) with no interactions (data not shown). Energy expenditure of the 4V leptin rats was higher than that of any other group before infusions started, and this difference was maintained on all except days 2, 7, and 8 of infusion (data not shown, infusion site, NS; leptin, NS; time, P < 0.0001; site × leptin, P < 0.04). Despite minimal differences in these daily measures, there was a significant effect of leptin infusion on the weight of epididymal and inguinal fat pads at the end of the experiment (site of infusion, NS; leptin, P < 0.005; interaction, NS). Surprisingly, the fat pads from leptin-infused rats were larger than those from saline-infused rats irrespective of whether the infusion was into the 4V or the NTS (Fig. 5A).

Fig. 5.

Fat pad weight (A), food intake (B–E), and hypothalamic phosphorylated signal transducer and activator of transcription 3 (pSTAT3) at bregma −3.60 mm (F) of rats in experiment 3 following an acute 3rd ventricle (3V) injection of 0.025 μg leptin. Data are means of 5–7 rats. Significant differences in food intake were determined by repeated-measures ANOVA for all rats followed by one-way ANOVA for each day within 4th ventricle (4V)- or nucleus of the solitary tract (NTS)-infused groups and post hoc Duncan’s multiple range test. Data for a specific time point within either the 4V or NTS infusions groups that do not share a common superscript are significantly different at P < 0.05. Differences in fat depot weight were determined by one-way ANOVA and post hoc Duncan’s multiple range test within 4V- or NTS-infused groups. Differences in pSTAT3 were determined by 2-way ANOVA and post hoc Duncan’s multiple range test.

3V injections of 0.025 μg of leptin significantly inhibited cumulative food intake of rats receiving NTS infusions of 5 ng leptin/24 h (Fig. 5, B and C) but not of any other group (site of infusion, NS; infusion, NS; injection, NS; time, P < 0.0001; site × infusion, P < 0.04; site × time, P < 0.007; infusion x injection, P < 0.06; site × infusion × time, P < 0.009; infusion × injection × time: P < 0.013). Intake of the NTS leptin/leptin rats significantly inhibited 12 and 18 h after the injection compared with all other groups of NTS-infused rats (Fig. 5C).

Food intake during each time interval was also considered. Because the rats ate so much during the first 39 min of access to food, this was separated out from intake during the rest of the first 6 h. When all of the data were considered for time intervals up to 36 h after 3V injection, there was a significant effect of time (P < 0.0001) and a significant interaction between time, site of infusion, and infusion (P < 0.001). Analysis of data at each time interval indicated no difference between groups during the first 39 min of access to food but a significant interaction between site of infusion and infusion during the 39-min- to 6-h interval and the 6–12 h interval (P < 0.007). During the 6- to 12-h interval, there also was a significant interaction between infusion and injection (P < 0.016). Post hoc analysis indicated that 4V saline/saline rats ate less than 4V leptin/saline rats during the 39-min- to 6-h interval, possibly because they ate the most during the first 39 min following injection (Fig. 5D). Food intake of NTS leptin/leptin rats was lower than of any other NTS-infused group. This difference reached significance during the 6- to 12-h interval but not the 39-min- to 6-h interval (P < 0.08).

There was no effect of 3V leptin injection on energy expenditure of the rats. Analysis of RER data showed an effect of time (P < 0.0001) and an interaction between site of infusion, infusion, and time (P < 0.001). Analysis of each time interval showed a significant interaction between site of infusion and infusion (P < 0.02) and infusion by injection (P < 0.01) for both the 39-min- to 6-h and the 6- to 12-h intervals. Post hoc analysis showed no difference in RER between the NTS-infused groups, but RER of 4V saline/leptin- was lower than that of 4V leptin-infused groups and all of the NTS-infused groups (data not shown).

Measurement of pSTAT3 40 min after a 3V injection of 0.025μg of leptin showed no significant difference in pSTAT3 in the NTS of the different groups of rats (data not shown). In the hypothalamus, leptin injections tended to increase pSTAT3 in the Arc, ventromedial hypothalamus, and dorsomedial hypothalamus of all of the groups of rats. This only reached significance in the dorsomedial hypothalamus of NTS leptin/leptin rats (Fig. 5F; infusion, NS; injection, P < 0.0003; interaction, NS).

DISCUSSION

The experiments described herein were designed to compare the effects of very low-dose infusions of leptin into the 4V or the NTS on the response to a 3V injection of leptin. We (5, 16) previously proposed that activation of leptin receptors in the hindbrain lowers the threshold for response to leptin in the forebrain, and the studies described herein were intended to determine whether this integration of response was mediated by a neural circuit or by diffusion of leptin from the 4V through the subarachanoid space to the forebrain. If the integration was mediated by a neural network, then there would be an enhancement of the response to a 3V injection of leptin with a very low-dose leptin infusion into the NTS but not with the same infusion into the 4V. Because unusually low doses of leptin were used in these experiments some of the leptin responses, although significant, appear subtle compared with those reported for high-dose treatments; however, it was essential to use threshold doses of leptin to tease out a difference between intraventricular and parenchymal injections.

In the first experiment, it was surprising to find that as little as 10 ng leptin/24 h infused into either the 4V or the NTS was sufficient to have a transient inhibitory effect on food intake and to reduce body fat during the 14-day infusion. Loss of fat is usually associated with a decrease in RER, because lipids mobilized from fat stores are oxidized for energy (35). By contrast, in that experiment, we found that once food intake had returned to control levels RER was slightly but significantly higher in the leptin-infused rats, implying a greater reliance on oxidation of carbohydrate, which would have been derived from glycogen stores and dietary intake. Therefore, the fat loss must have occurred during the first week of infusion, when food intake was inhibited, and was not subsequently restored even when food intake had returned to control levels. The transient effect of leptin on food intake but sustained inhibition of body fat mass has been reported for other leptin infusion studies. Body weight will only normalize once the leptin infusion has stopped (27), or it may remain suppressed compared with controls if leptin has been infused for an extended period of time (25). Therefore, sustained increases in peripheral or central levels of leptin appear to result in a drive to change to a lean body composition, and the initial inhibition of food intake may be a mechanism that allows this new metabolic equilibrium to be achieved (11).

Infusion of 5 ng leptin/24 h into either the 4V or NTS did not significantly change any of the end points measured in experiment 1 and was therefore considered a subthreshold dose of leptin. The lack of effect of 5 ng leptin/day infusion on food intake or body weight was confirmed in experiment 2; however, this low-dose infusion did cause a significant increase in energy expenditure during the dark period irrespective of the site of infusion. The change in expenditure was not reflected by an increase in iBAT temperature or compensated for by an increase in food intake. Closer inspection of expenditure during the days before pump placement suggested that the two leptin-infused groups of rats had nonsignificantly higher expenditure than the saline-infused groups even before leptin infusions started, and it is possible that there would not have been a significant effect of leptin on expenditure if they had been better matched before the infusion started. At the end of experiment 3, it was surprising to find a significant increase in the weight of both inguinal and epididymal fat from rats that had received either 4V or NTS infusions of leptin, especially considering that the results from experiment 1 had shown no effect of 5 ng leptin/day on body composition. It is not clear why there would be this difference between the two experiments, which were conducted in the same room and using the same size rats from the same supplier and fed the same diet. In experiment 3, there were no detectable differences in daily food intake, energy expenditure, or brown fat thermogenesis that might have contributed to this difference in body composition. However, it is possible that small, nonsignificant differences were present and could have contributed to the small change in energy deposition required to increase body fat over a period of 9 days. We previously reported that peripheral infusion of leptin in chronically decerebrate rats (15) or leptin infused into the 4V increases body fat mass (14); however, that occurred with a higher-dose infusion than was used here and was interpreted as the hindbrain having the potential to protect against excessive weight loss in conditions of hyperleptinemia. Recently, I reported that infusions of 0.01 μg leptin/day into the 3V also caused a significant increase in weight gain with no change in food intake (17), which obviously would not fit with the previous interpretation of leptin-induced weight gain. However, when these results are taken together, it does appear that under certain circumstances there must be a condition in which activation of leptin receptors causes weight gain, but these circumstances have not been identified or deliberately imposed in any of the experiments in which the changes were observed.

3V leptin injections in rats receiving 4V infusions of saline or leptin had no significant effect on food intake. By contrast, rats receiving NTS infusions of leptin, but not saline, responded by reducing food intake. These data support the idea that the enhanced responsiveness to leptin in the forebrain associated with activation of hindbrain leptin receptors is mediated by a neural mechanism rather than by diffusion of leptin from the 4V to the hypothalamus. The effect of leptin on food intake was transient and significant only during the period of 6–12 h after injection. Although there were no treatment effects at later time points, the failure of the rats to compensate for the temporary hypophagia meant that cumulative intake of NTS leptin/leptin rats remained lower than that of the other animals for a sustained period of time. The objective of experiment 3 was primarily to test whether leptin infused directly into the NTS would increase the level of STAT3 activation in the hypothalamus, and extraordinarily low doses of leptin were used for both the infusions and injection. Therefore, it was surprising to find significant effects of 0.025 μg of leptin in the 3V on food intake in rats receiving NTS infusions of 5 ng leptin/24 h. These data confirmed the results of experiment 2 and supported the concept that activation of NTS leptin receptors lowers the threshold for activation of hypothalamic leptin receptors and that integration of leptin activity at both sites results in a suppression of food intake. Because this effect was only evident in rats receiving infusions directly into the NTS and not those receiving a 4V infusion, these data also support the notion that communication between the hindbrain and hypothalamus is mediated by a neural pathway. The inhibition of food intake in NTS infused rats was delayed by several hours after leptin injection into the 3V. This delay implies the involvement of a complex pathway or the requirement for protein synthesis. However, it also is possible that, because all of the rats in our experiments were food deprived for 11 h before the leptin injection, the resulting hunger signals overrode initial leptin-induced hypophagia.

An alternate explanation for chronic activation of NTS leptin receptors enhancing the response to acute 3V leptin is that leptin infusion “primes” the hindbrain to be more sensitive to afferent information from the forebrain. An example of priming by leptin is the substantial enhancement of iBAT thermogenesis when both leptin and thyroid-releasing hormone are applied simultaneously to the hindbrain (18). Inhibition of food intake of rats receiving both forebrain and hindbrain leptin is likely mediated by an amplified response to peripheral satiety signals. This is supported by evidence that 3V leptin exaggerates the satiety produced by peripheral injections of cholecystokinin (CCK)-8 (8), and Schwartz and Moran (31) reported that 3V leptin increases activity of NTS neurons that respond to gastric loads independently of changes in vagal afferent activity. However, there is simultaneous activation of the paraventricular nucleus of the hypothalamus (PVH) and NTS, indicated by cFos expression (8). Although the PVH does not express leptin receptors, leptin-responsive hypothalamic nuclei, including the arcuate (34) dorsomedial (9), and ventromedial nuclei (32) project to the PVH, which in turn projects to the NTS and modulates sensitivity to peripheral satiety signals. Oxytocin release from PVH neurons has been identified as a requirement for 3V leptin enhancement of the satiety effects of peripheral CCK (2). Our experiments do not exclude the possibility that leptin infusion makes the NTS more sensitive to afferent information from the PVH, but cFos suggests increased activation of the PVH (7, 8), which then controls the NTS. One set of observations that could not be explained by priming of the NTS by leptin is that 4V infusion (16) or injection (26) of leptin at concentrations that suppress food intake produces a simultaneous increase in hypothalamic leptin signaling. It would not be necessary for hypothalamic signaling to be enhanced in order for food intake to be suppressed if it were driven by NTS sensitization.

In previous studies (5, 16), we found that activation of leptin receptors in the hindbrain increases leptin-induced pSTAT3 in the hypothalamus. In experiment 3, we measured pSTAT3 in the hypothalamus and hindbrain of rats 40 min after they received a 3V injection of leptin. There was no clear exaggeration of leptin-induced pSTAT3 in the arcuate or ventromedial hypothalamic nuclei; however STAT3 activation was higher in the dorsomedial hypothalamus of NTS leptin-infused rats than in any of the other groups. There were only three or four rats per treatment group for the immunohistochemistry, which may have contributed to the lack of significant differences between groups. The time interval between leptin injection and collection of the brain was based on evidence that hypothalamic STAT3 activation peaks an hour after leptin injection (29). However, the time lag between leptin injection and inhibition of food intake in experiments 2 and 3 suggests that there is a substantial delay between activation of the leptin receptors and measurable behavioral responses. It is possible that more differences in hypothalamic pSTAT3 would have been detected if there had been an extended interval between leptin injection and collection of brain; however, it is also likely that the effect on food intake involves a more complex system, and it will be necessary to identify the downstream effectors to detect differences between 4V- and NTS leptin-injected groups.

It has been reported that peripherally (6, 20) or centrally injected leptin (19) reduces meal size, and this appears to result from a synergy between central leptin and peripheral satiety signals. A subthreshold (3.5 μg) 3V injection of leptin 1 h before a gastric preload of Ensure significantly inhibited the subsequent 30-min food intake of rats (7). The involvement of satiety signals was demonstrated by similar experiments in which a subthreshold dose of 3V leptin was given 1 h before a peripheral injection of CCK (8) or bombesin (21). There was a synergistic inhibitory effect on food intake measured during the 30 min immediately following administration of CCK or bombesin, which was associated with increased cFOS in the NTS (8, 21) and in the PVN of the hypothalamus (8). Further evidence for an interaction between hypothalamic leptin signaling and efficacy of satiety signals is provided by evidence that leptin receptor-deficient Kolestsky rats are relatively insensitive to CCK and consume larger meals than their controls. Selective replacement of leptin receptors in the Arc of these rats makes them more sensitive than their controls to CCK and also reduces spontaneous meal size during ad libitum feeding (22). In all of these studies the subthreshold dose of leptin was much higher (30- to 140-fold greater) than that used in the experiments described here. One reason may be a difference in bioactivity of leptin from different sources (PeproTech vs. R&D Systems); another may be that rats were eating chow in the experiments described here, but were drinking Ensure (Ross Products, Columbus, OH) in those testing the interaction between leptin and satiety signals.

The data presented here are interpreted as evidence that activation of hindbrain leptin receptors lowers the threshold for response to third ventricle leptin through a neural pathway. An alternate hypothesis is that the hypophagic response to leptin is the result of simultaneous activation of hypothalamic and hindbrain receptors. The experiments described here do not exclude the possibility of a simple additive effect; however, evidence that application of leptin to the NTS causes phosphorylation of STAT3 in select hypothalamic nuclei in the absence of administration of exogenous leptin to the forebrain (26) supports the idea that hindbrain leptin receptors lower the threshold for response in the hypothalamus. It is well established that there are afferent projections from the NTS to hypothalamic nuclei (28); however, to further elucidate the mechanism hypothesized here, it will be necessary to demonstrate a mono- or polysynaptic connection between cells in the hindbrain that express leptin receptors and those in hypothalamic nuclei that express leptin receptors. In addition, it will be necessary to identify the downstream effector that responds by modulating meal size and/or frequency. This will require a better understanding of the integration between long-term signals of energy balance, such as leptin, with peripheral satiety signals.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-053903, awarded to R. B. S. Harris.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author.

AUTHOR CONTRIBUTIONS

R.B.H. conceived and designed research; performed experiments; analyzed data; interpreted results of experiments; prepared figures; drafted manuscript; edited and revised manuscript; approved final version of manuscript.