Orexin-A Hyperphagia: Hindbrain Participation in Consummatory Feeding Responses

Animals

Adult albino male Sprague Dawley rats (Charles River, Wilmington, MA) weighing 184–368 g on the first day of the main experiments were used. Test groups were more closely age and weight matched within each experiment. The lesion groups required more preparation before the experiment, resulting in greater body weight at testing. Rats were maintained individually in plastic tubs (48 × 25 × 15 cm) with wire lids on a 12-h light, 12-h dark schedule in a temperature-controlled room. Food (Purina rat chow 5001; Lab Diets, St. Louis, MO) and tap water were available ad libitum in the home cage, except where noted below. Rats were tested at the same time each day, between 4 and 8 h after lights on (0700 h), in a separate test cage.

Surgery

All procedures were approved by the Amherst College Institutional Animal Care and Use Committee. Rats were anesthetized (ip) with a mixture of ketamine HCl (66 mg/kg) and xylazine HCl (6 mg/kg). A 22-gauge guide cannula (Plastics One, Roanoke, VA) was stereotaxically implanted into either the 3V (from bregma: anteroposterior −2.3 mm, mediolateral 0 mm, and dorsoventral −8.5 mm from skull surface) or 4V (from lambda: anteroposterior −3.2 mm, mediolateral 0 mm, and dorsoventral −7.2 mm from skull surface) and fastened with dental acrylic and skull screws. The 28-gauge injection cannula extended 1 mm below the tip of the guide cannula, and a dummy cannula cut flush to the guide tip was maintained in the guide cannula at all other times. After surgical recovery, correct placement of the 4V cannula was confirmed by an increase in blood glucose after a cannula infusion of the antimetabolite, 5-thio-d-glucose (120 μg/2 μl; 2 μl/min) (39). Tail blood was collected every 10 min starting 20 min before the 5-thio-d-glucose injection, using a commercially available glucometer and test strips. Rats that did not exhibit at least a 50% increase in blood glucose within 30 min of injection were removed from the study. Correct placement for the 3V cannula was confirmed by a minimum of 5 ml water consumption within 30 min of a 50 ng/5 μl cannula injection of angiotensin II (32,40,41).

At least 2 d after 3V cannula placement confirmation, rats were anesthetized a second time and received either an AP/NST lesion (APX group) or sham procedures. The surgery was done after cannula patency confirmation to minimize unnecessary surgery in rats that could not be tested and to maximize their viability to provide pilot data for unrelated experiments. The anesthetized rat was placed in the stereotaxic apparatus with the head flexed ventral to maximize exposure of the brainstem. The dorsal medullary surface was exposed by retracting the overlying neck muscles and removing the posterior atlanto-occipital membrane. Using a surgical microscope (Seiler M900), the floor of the 4V and the region containing the AP was identified. The AP was carefully aspirated using a PE-10 tubing nozzle attached via larger PE and Tygon tubing segments connected to a vacuum pump. The access hole through the meninges and membrane was then carefully filled with gel foam, the neck muscles and scalp were sutured, and the rat was allowed to recover from the operation. Sham-lesioned rats were treated identically with the exception that the aspirator was not introduced near the AP.

Concluding behavioral experiments, cannula placements were again confirmed; India ink (2 μl) was injected immediately after a lethal overdose of sodium pentobarbital (100 mg/kg). Rats were then transcardially perfused with isotonic saline followed by 10% formalin. The brain was removed, and the whole brain was bisected midsagittally in 4V cannula fitted rats, whereas the forebrain was blocked and bisected in 3V cannula fitted rats, and both sets of brains were inspected for ventricular ink perfusion. Data for rats with no ink perfusion of the 3V or 4V were discarded. Ink was never observed to have reached the third or lateral ventricles after a 4V infusion. The hindbrain sections of the 3V fitted rats (which had also received APX/SHAM surgery) were then preserved 1–3 d in a 20% sucrose-formalin solution and cut in to 50-μm serial sections, mounted on microscope slides, and then stained using neutral red for analysis of the extent of lesion using bright-field microscopy at ×40 and ×100 magnification.

Apparatus

Rats were taken from their home cages and tested in individual plastic tubs (48 × 25 × 15 cm). A drinking spout (3-mm orifice; Girton Inc., Millville, PA) was introduced to the test chamber with the spout orifice positioned 4 cm from the floor and 0–1 mm behind a slit (8 × 28 mm) in a metal plate attached to the front of the cage. A lickometer (MS-108; DiLog Instruments, Tallahassee, FL) and PC were used to record licking; tongue contacts with the spout completed a circuit, which allowed the computer to record the time of each lick with 1-msec resolution. Files for each test session for each rat were saved for offline analysis.

Procedures

Before testing, rats in all experiments were habituated in the test cage daily where they were free to ingest 0.5 m sucrose solution for 90 min. Habituation training continued until session intakes stabilized and exceeded 5 ml per session (two to five sessions).

Data analysis

Data were analyzed according to previously established analysis parameters, as follows (see Refs. 40,41, and 44 for details). Microstructure analyses were limited to the first meal in the test session.

Meal size (lick count) was calculated as the number of licks in the meal (first lick of the first burst to last lick of the last burst) (40,45). The end of the meal was defined by a pause in licking greater than or equal to 10 min (44). Meal duration (minutes) was defined as the session time of the last lick in the meal minus the session time of the first lick in the meal. Average ingestion rate (licks per minute) was calculated as the number of licks in the meal divided by meal duration. Lick volume (microliters) was calculated by dividing the difference between pre- and posttest weights of the spout bottle by the total number of licks in the session. Meal size (milliliters) was then calculated by multiplying the meal lick count by the lick volume.

The temporal distribution of licking was analyzed using a variety of custom-made programs (40,46,47). A licking burst was defined as two or more consecutive licks with no inter-lick interval (ILI) exceeding 1 sec. Thus, pauses greater than 1 sec determined burst termination (45). Burst duration (seconds) was calculated by subtracting the session time of the first lick in the burst from the time of the last lick in that burst. Mean burst size (lick count) was calculated as the cumulative number of licks in all bursts in the meal divided by the number of bursts in the meal. To minimize artifacts due to nonlingual spout contacts, meal onset was defined as the first lick of the first burst containing at least three licks. Latency (seconds) was defined as the time between placement of the rat into the test cage and meal onset. Initial lick rate was the number of licks in the first minute of the meal.

ILIs were analyzed in several ways. The average within-burst ILI (milliseconds) was determined by averaging all ILIs less than 1 sec. Because more than 95% of all ILIs in a meal are less than 250 msec (in rats) and are normally distributed below this cutoff (44,48), the average duration of ILIs less than 250 msec was also determined. We also evaluated the relative proportion of ILIs less than 250 msec as a percentage of all ILIs within bursts (ILIs less than 1 sec). Any significant variation in this proportion indicates a reciprocal variation in the proportion of ILIs ranging from 250–999 msec.

The mean pause duration (seconds) was defined as the meal duration minus the cumulative duration of bursts in the meal, divided by the number of meal pauses (number of bursts minus one). Percent pause duration for the meal was the cumulative time of all pauses divided by the meal duration multiplied by 100.

For experiment 1, results for the five drug levels (aCSF, 0.1, 0.4, 1, and 10 nm) were analyzed with a one-way repeated-measures ANOVA using SPSS 16 software. Responses for each measure in experiments 2 and 3 across the two drug days (ORXA/aCSF) for each of the three taste solutions (water and 0.1 and 1 m sucrose) were compared with two-way (drug × tastant) repeated-measures ANOVA. Planned t tests were used to compare drug responses for each solution, to inform any significant main effects or interactions. A mixed-factors two-way ANOVA was used to assure comparable baseline conditions between lesion groups (group aCSF condition × tastant) in experiment 3 and to compare the sham group results with those obtained in experiment 2, to allow comparison of forebrain (3V) vs. hindbrain (4V) icv infusions of ORXA in intact rats. Difference scores (ORXA minus aCSF) for each rat in the 3V and 4V groups were also compared to evaluate the effect magnitudes of ORXA-induced responses. In experiment 4, intake for both treatments was compared with that on preceding baseline days or aCSF conditions using paired t tests. The criterion for a statistically significant difference was P < 0.05.

Experiment 1: 4V ORXA dose-response effects

In the seven rats in this experiment, ORXA doubled intake at the two strongest doses tested, 1 and 10 nm [F_(4,24) = 23.29; P = 0.003; Fig. 1A]. Interestingly, intake was increased by different underlying processes at each dose. The increase at the 1 nm dose was due primarily to an increase in first meal size, as indicated by significant ANOVA [F_(4,24) = 2.83; P < 0.05] and paired t tests (Fig. 1B); however, after the 10 nm infusion, there was a stronger effect on meal frequency rather than meal size [F_(4,24) = 4.08; P < 0.01; Fig. 1C].

The microstructure changes after ORXA suggested that 1 nm ORXA increased meal size by suppressing inhibitory postingestive feedback. First, ORXA did not significantly influence either the initial lick rate or the mean burst size, the two measures commonly associated with gustatory evaluation (Table 1). Second, neither lick volume nor any of the ILI measures were affected (Table 1), indicating that ORXA did not influence oromotor coordination or the intrinsic rate of licking within bursts. ORXA also had no influence on latency to begin feeding (Table 1). Rather, ORXA significantly prolonged the meal as indicated by a statistically significant increase in meal duration (Fig. 2A). The mean burst count at 1 nm was also significantly increased relative to aCSF (Fig. 2B).

Experiment 2: 4V ORXA effects on licking microstructure for a range of taste solutions

For the 16 rats in this experiment, meal size under baseline conditions varied across tastants, as anticipated (32,41,42,44) (Fig. 3A). As observed in experiment 1, ORXA significantly increased meal size; however, this result was observed only for the 0.1 m sucrose solution (Fig. 3A). Interestingly, ORXA influenced appetitive feeding (meal frequency) in a manner dissociated from its influence on consummatory feeding (meal size), because the meal frequency was increased for all three taste solutions (Fig. 3B). This appetitive effect resulted in significant increases for overall intake for the total test session for both sucrose solutions, although the effect was most pronounced for 0.1 m sucrose (Table 2), where the increases in consummatory and appetitive feeding synergized.

Analysis of the licking microstructure data (detailed as follows) for the ORXA-induced increase in 0.1 m sucrose meal size indicated that the results generally replicated the findings of experiment 1.

There was a range of gustatory hedonic responses to the three taste solutions. The mean burst size and number of licks in the first minute increased significantly as sucrose concentrations were increased under aCSF conditions (Table 2). However, there was no significant effect of ORXA on these measures (Table 2), suggesting no influence on orosensory tastant evaluation.

As with the gustatory evaluation measures, measures reflecting caloric intensity of the tastants also varied across the test solutions under baseline conditions. Meal duration varied significantly across taste solution conditions (Fig. 4A), as did the number of bursts (Fig. 4B) and the average ingestion rate, with the fastest consumption rate expressed for 1 m sucrose (Table 2). After 4V ORXA, meal duration was prolonged by 66% for the 0.1 m sucrose solution (Fig. 4A). ORXA prolonged the meal principally by adding more bursts to the meal (Fig. 4B), because mean burst size, mean pause duration, and pause time as a percentage of meal duration measures were not increased (Table 2).

Experiment 3: 3V ORXA infusions in rats with AP/NST lesions or sham procedures

The lesion status could be not be verified in four rats due to difficulties with histological preparation; thus, the data for these rats were discarded, resulting in 17 APX and eight sham rats for which the data were analyzed. Figure 5 shows example photomicrographs depicting the range of lesions observed. Lesions ranged from considerable but incomplete destruction of the AP to complete ablation of the AP with collateral damage to the subjacent dorsal medial and commissural subnuclei of the NST, consistent with the range of damage reported in previous studies using this preparation (49,50,51).

Experiment 4: responses to other stimuli in AP/NST lesioned rats

The subgroup of APX rats tested in experiment 3 that received a 3V MTII infusion after the main experiment (n = 10) exhibited a significant reduction of 0.1 m sucrose intake after infusion (Fig. 8A). Meal size responses in this subgroup also were not increased after ORXA infusion.

The second group of APX rats with confirmed lesions (n = 10) exhibited a significant increase in consumption of 0.1 m sucrose after 22.5 h food and water deprivation and a marginally significant increase after 3V injection of angiotensin II (Fig. 8B).

ORXA differentially influences appetitive and consummatory feeding

In experiment 1, the 4V 1 and 10 nm ORXA doses each increased 0.1 m sucrose intake, but the 1 nm dose increased only meal size (consummatory feeding), whereas the 10 nm dose increased meal frequency (appetitive feeding). In experiment 2, the 4V 1 nm ORXA dose increased meal size only for 0.1 m sucrose, whereas it increased meal frequency for all tastants. The same dissociation was observed in sham lesioned rats after 3V ORXA (1 nm). The results are consistent with a previous report that 3V ORXA (15) increased intraoral intake (a model of consummatory feeding (52) of 0.1 m sucrose. Finally, AP lesions abolished the increases in meal size after 3V ORXA, but increases in meal frequency were still observed.

The meal frequency results described here are consistent with the well known activating effects of ORXA (15), which has been reported to increase locomotor and exploratory activity (4), reward learning (7,12), feeding frequency (14), response to caloric deprivation, and willingness to work for food (6). ORXA neurons are moderately active during feeding and maximally active during exploratory behavior (53), and they may mediate fasting-induced waking processes. Collectively, these reports implicate ORXA in foraging behavior (see Ref. 4 for discussion), and our results are consistent with this interpretation. However, our results also indicate that ORXA separately influenced consummatory feeding through at least partially nonoverlapping sites of action relative to those for meal frequency effects.

ORXA suppresses inhibitory postingestive feedback (satiation)

We analyzed the licking microstructure of intact rats after 3V or 4V ORXA administration. Numerous studies show that under controlled laboratory conditions, licking patterns vary reliably with the orosensory (gustatory) and postingestive (satiating) stimulation properties of the tastant. Because very little is ingested early in the meal, orosensory evaluation of the tastant is the primary influence on behavior in this phase (41,42,43,44,54). The initial rate of licking in the meal varies systematically with increases in the concentration of palatable or aversive tastants (45,46,55). The mean lick burst size also increases or decreases in accordance with the palatability of the stimulus (32,45,46,55). As the meal progresses, the increasing size of the gastrointestinal load produces inhibitory feedback that leads to an increasingly rapid decline in ingestion rate. Increasing the caloric or volumetric properties of food produces more rapid reductions in lick rate, burst number, and meal duration (but not mean burst size or lick rate within bursts), suggesting that these measures reflect postingestive satiation (42,43,56,57,58,59,60). These response patterns were replicated across taste solutions in the control conditions of experiments 2 and 3.

ORXA did not affect initial lick rate or mean burst size in any experiment. This suggests that previous findings of ORXA increases in consumption of palatable foods such as sweetened chow or high-fat diet were not due to an ORXA influence on hedonic gustatory evaluation. This conclusion is also supported by reports that behavioral responses to the orexin 1 receptor antagonist SB-334867 did not overlap those to bitter quinine hydrochloride (61,62). Our results also suggest that previous reports of increases in water consumption (1,11) may be due to the action of ORXA to increase appetitive feeding (foraging) rather than a specific action on thirst-related systems (1,63), because neither meal size nor microstructure for water licking was affected by ORXA.

The current findings suggest that icv ORXA suppressed inhibitory postingestive feedback. After ORXA administration, rats prolonged the meal in all experiments as much as 2-fold, and this was achieved by increasing the number of licking bursts in the meal. Similar findings were reported by Ida et al. (10) who described increased total feeding duration after forebrain ORXA administration. Furthermore, a videographic analysis of feeding behavior after forebrain icv infusions of either ORXA or SB-334867 indicated that ORXA and the antagonist respectively delayed or hastened the onset of behavioral satiety (the transition from eating to resting) by 12–15 min, with no effects on locomotor behavior at low to moderate doses (8).

Our results are also consistent with studies that implicate ORXA in gastrointestinal processing. Orexin 1 receptors have been localized to neurons in the DMX that innervate the gastrointestinal tract (64), and ORXA stimulation enhances inhibitory post-synaptic currents in DMX neurons (65) and influences gastric acid and pancreatic secretions (3,66,67). ORXA applied to the DMX at the level of the AP was reported to increase gastric contractions (2). Intracerebroventricular ORXA has also been shown to alter gastric motility (9).

Finally, it is important to note that ORXA-induced increases in meal size were not due to nonspecific oromotor activation. Taste-related differences in response to the taste solutions were sustained after ORXA, consummatory responses to water were unaffected by ORXA, and ORXA did not increase the rate of licking within bursts.

Hindbrain structures may mediate the consummatory feeding responses to ORXA

We found that AP/NST lesions abolished meal size responses to 3V ORXA without influencing sensitivity to the gustatory and caloric properties of the test solutions or to the hyperphagic/dipsic effects of deprivation and angiotensin II, or the anorectic effects of MTII. The 3V and 4V ORXA effects on meal size were nearly identical in magnitude in intact rats, as was the qualitative pattern of the behavioral processes underlying meal size increases, with microstructure measures of postingestive feedback inhibition similarly affected in the 3V sham and the 4V groups. These results suggest that hindbrain sites involving the AP/NST may be important for consummatory feeding responses to icv ORXA stimulation.

It is worth noting that the AP/NST lesions did not eliminate appetitive (meal frequency) responses to 3V ORXA. Previous reports note that intraparenchymal infusions of low doses of ORXA to the paraventricular, dorsomedial, and perifornical/LH areas increase chow and/or water intake. These or other sites of ORXA action may mediate these intake increases through increases in appetitive feeding behaviors, independent of the neural systems that mediate ORXA consummatory feeding effects.

The precise brainstem sites of ORXA feeding action remain to be determined. Although there is evidence of ORXA ligand and orexin-1 receptors throughout structures of the dorsal vagal complex (see introductory section), previous studies showed that intra-NST ORXA infusions only minimally influenced feeding responses (11,31); interestingly, the extent of spread to the AP in those studies may have been limited (see Fig. 14F in Ref. 11). Our data suggest that the AP may be an important structure mediating consummatory feeding responses to icv ORXA, which could clarify the discrepancies between the reported 4V and intra-NST ORXA feeding responses. It is also possible that the resolution of our behavioral analysis, or our choice of foodstuffs, permitted us to observe significant ORXA responses that were not available to previous investigators, considering that we observed meal size effects only for the moderately palatable/caloric sucrose solution. The role of the AP in ORXA hyperphagia could be further clarified by analysis of 4V ORXA infusions in AP-lesioned rats and direct AP injections of orexin receptor ligands.

Finally, although we show that the AP/NST is necessary for consummatory ORXA responses, this does not prove that direct LH-AP ORXA projections are of central importance in mediating the ORXA consummatory feeding response. It is possible that second- and n-order mechanisms sensitive to ORXA stimulation convey ORXA-related signals to the hindbrain. However, our results do indicate that the AP/NST is an obligatory relay within any such ORXA sensitive networks.

Limitations

It is unclear why ORXA modified meal structure specifically for the 0.1 m sucrose solution without affecting the responses for 1 m sucrose. Relative to 1 m sucrose, the 0.1 m sucrose solution was both less palatable and less caloric. Because we observed no effect of ORXA on measures of gustatory evaluation (initial lick rate and mean burst size), the moderate caloric basis of the 0.1 m sucrose solution appears to be principal to the meal size effect of ORXA in this study. It is possible that the caloric density of 1 m sucrose produced stronger satiation signals that were not sufficiently subdued by our near-threshold dose of ORXA (a ceiling effect); however, others have reported that ORXA increased consumption after allowing animals to preingest food (12). Furudono et al. (9) reported that lateral ventricle ORXA infusions increased consumption of noncaloric 0.1% saccharin solution, but it remains unclear whether this effect (or the aforementioned) was due to appetitive responses to ORXA, which we observed for all solutions including water and 1 M sucrose. It would be worthwhile to evaluate whether ORXA influences consummatory responses for noncaloric or isopalatable solutions of different caloric density.

We also found that 4V ORXA dose-dependently increased the number of meals initiated (Fig. 1), a result we did not observe after 4V NPY infusions (32). However, our result is consistent with the reports that ghrelin and strong MTII doses influenced meal frequency after hindbrain delivery (69,70). Initially, these data appear at odds with the well known finding that chronic decerebrate rats, in which supracollicular structures are severed from the hindbrain, do not express appetitive feeding behaviors (including meal initiation), implying that hindbrain sites do not mediate appetitive functions (71). In attempting to reconcile these apparent contradictions, there are several possibilities to consider. It is possible that hindbrain ORXA stimulation in nonlesioned rats influenced appetitive feeding control systems in the forebrain through intact hindbrain-forebrain relays; there are, for example, direct ascending projections from the NST to the paraventricular hypothalamus and LH (72,73). Alternatively, it may be that appetitive behavioral responses are controlled via hindbrain structures, but these systems require permissive input from forebrain structures (74,75). Lastly, it is possible that there was back diffusion of the 4V ORXA injection to the forebrain ventricles (against the caudal flow of CSF), although several factors call this into question. First, previous studies have not observed ink flow to the forebrain after a 4V injection (25,32,40,41). Second, dipsogenic responses to angiotensin II are not elicited after 4V application. Third, 4V injections of MCH failed to influence feeding even when delivered at maximal doses (32). Regardless, our findings after APX indicate that the AP/NST is necessary for the expression of hyperphagic consummatory responses, even when ORXA is delivered to the 3V.

Comparisons with the behavioral effects of other hypothalamic peptides

Although MCH and ORXA have a common origin (the LH area) and common effects on food intake, the two peptides influence feeding behavior in profoundly different ways, and through different sites of action. The 3V MCH increased meal size by increasing measures of gustatory evaluation (mean burst size and initial lick rate), without influencing the postingestive feedback-related measures. ORXA produced the opposite pattern after either 3V or 4V infusions. In further contrast, MCH injection to the 4V was entirely without effect (32), suggesting that unlike ORXA and NPY, the hyperphagic effects of MCH require forebrain stimulation. This conclusion dovetails with other studies reporting that ORXA but not MCH feeding responses were blocked by the opioid antagonist naloxone (77), that subthreshold dose coinfusions of 3V ORXA/NPY and MCH did not synergistically increase food intake (78), and that forebrain icv NPY and ORXA modulated gastric motility, whereas forebrain icv MCH infusions did not under the same conditions (9).

Overall, ORXA actions were more similar to those of NPY and POMC receptor ligands. Like ORXA, NPY delivered to the 3V increased meal frequency for water, sucrose, and saccharin solutions, but meal size was increased only for caloric sucrose solutions (32,40). Furthermore, 3V and 4V NPY and ORXA, and 4V MTII infusions, affected licking measures associated with satiating postingestive feedback, not those associated with gustatory evaluation (32,40,41,76). In addition, subthreshold 3V NPY and ORXA doses did synergistically increase food intake (78).

Future studies exploring the effects of ORXA antagonists in the 4V or hindbrain nuclei, applied separately and conjointly with NPY or POMC receptor ligands, could provide additional support for the hypothesis that NPY, ORXA and POMC receptor ligands influence common mechanisms of meal size control through sites of action in the caudal brain stem. It is worth noting in this regard that the hindbrain mechanisms mediating responses to NPY, ORXA, and POMC are likely to be partially rather than completely overlapping, because rats with AP/NST lesions retained consummatory responses to forebrain-delivered MTII. Although it is true that icv administration of any compound can never truly mimic the quantitative, spatial, and temporal dynamics of the natural system, it is clear that the consummatory hyperphagia elicited here by forebrain icv ORXA stimulation required the integrity of the AP/NST. Overall, these findings further clarify the functional neuroanatomy of the central ORXA system, and they may shed further light on the organization of hypothalamic-hindbrain controls of feeding behavior.