Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The role of the opioid peptides and their receptors in modulating ingestive behavior has been a source of intense study over the past quarter-century (for review, see Gosnell and Levine, 1996). In addition to the stimulation of feeding by opiate drugs acting at µ- (e.g., morphine; Sanger and McCarthy, 1980) and - (e.g., ketocyclazocine and butorphanol; Morley et al., 1982; Levine and Morley, 1983) receptors, feeding was observed following direct microinjections of opioid peptides themselves, including END (Grandison and Guidotti, 1977), enkephalin analogs (Jackson and Sewell, 1985; Stanley et al., 1989) as well as dynorphin (Morley et al., 1982; Morley and Levine, 1983). However, subsequent studies investigating opioid modulation of feeding elucidated specific receptor mechanisms through the use of selective opioid receptor subtype agonists and antagonists (for review, see Bodnar, 1996; Gosnell and Levine, 1996), and through the recent use of AS ODN probes directed against specific exons of opioid receptor genes (Leventhal et al., 1997, 1998b). Therefore, full characterization of the mechanisms by which opioid peptides act to stimulate feeding has not been examined in the rat, the species most used for opioid-induced feeding studies.

END (END_1-31) is derived from -lipotropin, which in turn is derived from its precursor peptide pro-opiomelanocortin (Mains et al., 1977). The two main cell groups of neuronal END are the hypothalamic arcuate nucleus (Watson et al., 1978) and the caudal nucleus tractus solitarius (Khachaturian et al., 1985). These cells innervate the preoptic area, septum, bed nucleus of the stria terminalis, other hypothalamic nuclei, temporal cortex, amygdala, periventricular thalamus, periaqueductal gray, nucleus raphe magnus, nucleus reticularis gigantocellularis, locus coeruleus, nucleus tractus solitarius, dorsal nucleus of the vagus, and lateral reticular nucleus. Biochemical END binding occurs at multiple opioid receptors (µ, , , and proposed ; Chang et al., 1979; Schulz et al., 1979; Akil et al., 1980). END stimulates food intake following microinjection into the hypothalamic ventromedial (Grandison and Guidotti, 1977) and paraventricular (Leibowitz and Hor, 1982) nuclei and nucleus accumbens (Majeed et al., 1986). Basal levels of pituitary and plasma END are elevated in genetically obese mice and rats (Margules et al., 1978). In contrast, hypothalamic END is decreased in streptozotocin-treated diabetic (Locatelli et al., 1986; Kim et al., 1999) and chronically food-restricted rats (Kim et al., 1996). Although these studies suggest a role for END in the mediation of ingestive behavior, they do not specify which opioid receptors participate in this feeding response. Opioid antagonist analyses of END feeding have been limited to goldfish and observe reductions induced by general (naloxone) and µ-selective (FNA), but not by - (NBNI) or  (7-benzidilidendenaltrexone and naltriben)-opioid receptor antagonists (DePedro et al., 1995, 1996).

The present study used two techniques to determine which opioid receptor subtypes participate in END-induced feeding in rats: general and selective opioid antagonists, and AS ODN probes directed against opioid receptor genes. First, a dose-response curve for END-induced feeding was determined, and potential reductions were examined following pretreatment with equimolar doses (5-40 nmol) of general (Ntx), µ- (FNA), - (Nti), and  (NBNI)-opioid antagonists. A second in vivo technique used AS ODN probes to establish the relationship of the cloned receptors to opioid actions using sequences complementary to regions of specific exons of mRNA to down-regulate opioid receptor proteins (Pasternak and Standifer, 1995). This technique has been used very effectively to provide converging evidence for antagonist effects for feeding responses particularly following opioid agonist treatment (Leventhal et al., 1997, 1998a,b). The present study used AS ODN probes directed against specific exons of the MOR-1, DOR-1, KOR-1, and KOR-3/ORL-1 opioid receptor clones to analyze their effects upon END-induced feeding. Specificity of AS ODN effects was confirmed using an MS ODN probe that was identical to a particular effective AS ODN except that the order of two pairs of bases was reversed.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Subjects and Surgery. Adult male albino Sprague-Dawley rats (Charles River Laboratories, 275-300 g, Wilmington, MA) were individually housed in suspended wire mesh cages, and maintained on a 12-h light/12 h dark cycle with Purina rat chow in food bins and water available ad libitum. All animals were pretreated with chlorpromazine (3 mg/kg i.p.) and were anesthetized with Ketamine HCl (100 mg/kg i.m.). A stainless steel guide cannula (22-gauge; Plastics One, Roanoke, VA) was implanted stereotaxically (Kopf Instruments, Tujunga, CA) into the left lateral ventricle using the following coordinates: incisor bar (+5 mm), 0.5 mm anterior to the bregma suture, 1.3 mm lateral to the sagittal suture, and 3.6 mm from the top of the skull. Each cannula was secured to the skull by three anchor screws with dental acrylic. All animals were allowed at least 2 weeks to recover from stereotaxic surgery before behavioral testing began. After completion of behavioral testing, which took approximately 6 to 8 weeks for each animal, all rats were sacrificed with an overdose of anesthetic, and cannula placements were verified visually by cutting coronal sections through the cannula placements. All animals had correct cannula placements in the lateral ventricle.

END Dose-Response Curve. All behavioral testing was conducted in the home cage between 2 to 8 h following the onset of the light cycle to minimize circadian effects on food intake. Rats were adapted to at least 4 days of baseline testing to eliminate any novelty-induced feeding responses elicited by placement of the pellets on the floor of the cage. It should be noted that intake during this phase of the light cycle is minimal as reflected by the low control values. In this and all subsequent protocols, before any experimental conditions, the food bins were removed from each cage and replaced with preweighed food pellets. Each intake value was measured by the weight of the food pellets in grams and adjusted for spillage that was collected on paper towels placed below the wire mesh cage. Following baseline, the first group of 12 cannulated rats was assessed for food intake after 1, 2, and 4 h following microinjection of END at doses of 0, 0.25, 1, 5, and 10 µg in counterbalanced order at weekly intervals. END was administered in a 5-µl volume of distilled water over 30 s through a stainless steel internal cannula (28-gauge; Plastics One), which extended 0.5 to 1.0 mm beyond the tip of the guide cannula, and which was connected to a Hamilton microsyringe by polyethylene tubing. Following infusion, the internal cannula was removed and immediately replaced with a stainless steel dummy cannula (28-gauge; Plastics One) to prevent any effusion, and to ensure cannula patency between microinjection conditions.

General and Selective Opioid Antagonists, END, and Food Intake. All antagonists were administered in 5-µl volumes of distilled water to guarantee solubility of the compounds. All groups of cannulated rats in the antagonist studies were initially assessed for food intake as previously described after 1, 2, and 4 h following vehicle and END at a dose of 10 µg, which was the most effective dose to produce the most marked and shortest latency feeding responses (under Results). The animals used in the END dose-response determination as well as new animals were divided into subgroups and included in the antagonist conditions. Therefore, rats were tested under control and END (10 µg) conditions first to ensure that each rat displayed significant feeding responses following END before being tested with specific antagonists. A total of 34 animals was used in the four antagonist paradigms. To allow for direct comparisons of antagonist effects, an identical equimolar dose range (5, 20, 40 nmol) was used. The first subgroup of eight rats received the general opioid antagonist Ntx (Sigma Chemical Co., St. Louis, MO) at doses of either 1.89, 7.56, or 15.12 µg 1 h before END and was tested for food intake 1, 2, and 4 h following the second injection. The order of antagonist dose treatments in this and subsequent central antagonist protocols was counterbalanced across animals, and a 1-week interval elapsed between treatments. The time interval between injections within each condition was based upon the respective peak and selective actions of the opioid antagonists in this and subsequent antagonist protocols. The second subgroup of nine rats received the µ-opioid receptor antagonist FNA (Research Biochemicals International, Natick, MA) at doses of either 0.245, 1.225, 2.45, 9.8, or 19.6 µg 24 h before END and was tested for food intake 1, 2, and 4 h following the second injection. The third subgroup of eight rats received the -opioid receptor antagonist Nti (Research Biochemicals International) at doses of either 2.55, 10.2, and 20.4 µg 1 h before END and was tested for food intake at 1, 2, and 4 h following the second injection. The fourth subgroup of nine rats received the ₁-opioid receptor antagonist NBNI (Research Biochemicals International) at doses of either 3.65, 14.6, or 29.2 µg 0.5 h before END and was tested for food intake at 1, 2, and 4 h following the second injection.

AS ODN Probes, END, and Food Intake. As described previously, all groups of cannulated rats in the AS ODN studies were initially assessed for food intake after 1, 2, and 4 h following vehicle and END at a dose of 10 µg and displayed significant feeding responses following END. All AS ODN probes were administered at 10-µg doses dissolved in 2-µl volumes of 0.9% normal saline based upon their previously determined effectiveness in feeding studies (Leventhal et al., 1997, 1998a,b) without producing nonspecific effects (for review, Rossi and Pasternak, 1997). All phosphodiester oligodeoxynucleotides (Midland Certified Reagent Company, Midland, TX) were purified in our (G.W.P., G.C.R.) laboratory, and the identified locations of the AS ODN probes were based on the different opioid receptor gene sequences listed in GenBank (Table 1). Each AS ODN probe was directed against the individual exons of either the MOR-1, DOR-1, KOR-1, or KOR-3/ORL-1 opioid receptor genes. During each 6-day test phase, rats received microinjections of their particular AS ODN probes on days 1, 3, and 5 as previously described (Leventhal et al., 1997); this time course of treatment both down-regulates the synthesis of new receptors and permits turnover of existing receptors (for review, see Pasternak and Standifer, 1995). Two weeks following a given AS ODN treatment, rats with patent cannulae were retested for END-induced feeding, and 1 week thereafter, were retested with a second AS ODN probe. Subgroups of rats were assigned to the following conditions by matching food intake following END: AS ODN probes directed against either exons 1, 2, 3, or 4 of the MOR-1 gene (n = 8/condition); directed against either exons 1, 2, or 3 of the DOR-1 gene (n = 8/condition); directed against either exons 1, 2, or 3 of the KOR-1 gene (n = 7-8/condition); directed against either exons 1, 2, or 3 of the KOR-3/ORL-1 gene (n = 7-8/condition); or a MS ODN probe directed against exon 1 of the MOR-1 gene (n = 6), which differed from its corresponding AS ODN probe by the sequence reversal of pairs of bases (Table 1). Twenty-four hours after the last AS or MS ODN treatment (day 6), all rats received END (10 µg), and food intake was assessed after 1, 2, and 4 h.

Statisitics. To determine significant agonist effects, separate one-way repeated-measures analyses of variance were performed on cumulative food intakes after 1, 2, and 4 h. Time was not considered as a separate variable because intakes are typically larger in the 1st h, and decline thereafter. Moreover, the time course of testing was discontinuous with 1-h intervals for the first two measures (1 and 2 h), and a 2-h interval for the last measure (4 h). Therefore, cumulative intakes were assessed as described previously (Leventhal et al., 1997, 1998b). Tukey comparisons (p < 0.05) were used to determine individual significant agonist effects relative to vehicle treatment. To determine significant antagonist or AS ODN effects upon END-induced feeding, difference scores were determined for each condition in each animal by subtracting a corresponding vehicle intake from the experimental score. Separate one-way analyses of variance were performed on these food intake difference scores after 1, 2, and 4 h. Baseline and END values were compiled individually for both the antagonist and AS ODN paradigms. Control data for the antagonist paradigms included baseline and END intake values from all 34 animals in these conditions. In the AS ODN paradigms, control values were determined by dividing baseline and END intake values into five subgroups (n = 19-21) based on opioid receptor gene antisense treatments. Dunnett's comparisons (P < 0.05) were used to determine individual significant antagonist or AS ODN effects relative to its corresponding END-induced feeding condition.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

END-Induced Feeding Dose-Response Curve. END produced a dose-dependent and time-dependent increase in food intake with only the highest (10-µg) dose significantly increasing intake over the entire 4-h time course (Fig. 1). Since the highest dose produced the greatest magnitude and shortest latency of feeding responses, this i.c.v. dose was used in all subsequent studies.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Alterations (mean ± S.E.M.) in food intake (g) following i.c.v. administration of the endogenous opioid peptide END. 0 µg, ; 0.25 µg, ; 1 µg, ; 10 µg, . Significant differences in food intake were observed following END relative to vehicle treatment after 1 [F(4,52) = 7.48, p < 0.0001], 2 (F = 8.98, p < 0.0001) and 4 (F = 18.06, p < 0.0001) h. The asterisks (*) in this and subsequent figures indicate significant increases in END-induced feeding relative to corresponding vehicle control values (Dunnett's comparisons, p < 0.05). Please note that control intake was minimal during this testing period (2-8 h into the light cycle) in rats adapted to the testing paradigm such that the symbols often encompass the small S.E.M.s in this and subsequent protocols.

Ntx and END-Induced Feeding. END-induced feeding was significantly reduced by the two highest (20- and 40-nmol) Ntx doses, and was transiently potentiated by the lowest (5-nmol) Ntx dose after 1 h (Fig. 2, top), suggesting opioid mediation of END-induced feeding.

View larger version (28K): [in this window] [in a new window]   Fig. 2.   A, alterations (mean ± S.E.M.) in food intake (g) following i.c.v. administration of either vehicle alone, END alone, or END following ventricular pretreatment (1 h) with the general opioid receptor antagonist Ntx at doses of either 5, 20, or 40 nmol. Significant differences in food intake were observed following all END treatment conditions relative to vehicle after 1 [F(4,79) = 16.45, p < 0.0001], 2 (F = 19.47, p < 0.0001), and 4 (F = 29.53, p < 0.0001) h. Difference-score analyses also displayed significant effects after 1 [F(3,36) = 21.03, p < 0.0001], 2 (F = 10.95, p < 0.0001), and 4 (F = 7.95, p < 0.0003) h. B, alterations (g ± S.E.M.) in food intake following i.c.v. administration of either vehicle alone, END alone, or END following ventricular pretreatment (24 h) with the µ-opioid receptor antagonist FNA at doses of either 0.5, 2.5, 5, 20, or 40 nmol. Significant differences in food intake were observed following END treatment conditions relative to vehicle after 1 [F(6,98) = 5.81, p < 0.0001], 2 (F = 9.49, p < 0.0001), and 4 (F = 14.15, p < 0.0001) h. Difference-score analyses also displayed significant effects after 1 [F(5,45) = 2.32, p < 0.05], 2 (F = 2.51, p < 0.044), and 4 (F = 5.05, p < 0.0009) h. The crosses (+) in this and subsequent figures indicate significant decreases in END-induced feeding following either antagonist or AS ODN pretreatment relative to corresponding END-induced feeding values alone (Dunnett's comparisons, p < 0.05).

FNA and END-Induced Feeding. Significant increases in feeding relative to vehicle occurred across the time course following END, and emerged after 4 h following FNA doses of 0.5 and 20 nmol paired with END (Fig. 2, bottom). END-induced feeding was significantly reduced by each of the FNA pretreatment doses: 0.5 (4 h), 2.5 (1-4 h), 5 (1, 4 h), 20 (4 h), and 40 (2-4 h) nmol (Fig. 2, bottom), suggesting µ-opioid mediation of END-induced feeding.

Nti and END-Induced Feeding. Significant increases in feeding relative to vehicle occurred across the time course following END and the lowest (5-nmol) Nti dose paired with END (Fig. 3, top). END-induced feeding was significantly reduced by the two highest (20- and 40-nmol) Nti doses after 4 h, and was transiently (1 h) potentiated by the lowest 5-nmol Nti dose (Fig. 3, top), suggesting -opioid mediation of END-induced feeding.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Alterations (mean ± S.E.M.) in food intake (g) following i.c.v. administration of either vehicle alone, END alone, or END following ventricular pretreatment (1 h) with either the -opioid receptor antagonist Nti (A) or the 1-opioid receptor antagonist NBNI (B) at doses of either 5, 20, or 40 nmol. Significant differences in food intake were observed following END and Nti treatment conditions relative to vehicle after 1 [F(4,79) = 10.83, p < 0.0001], 2 (F = 14.17, p < 0.0001), and 4 (F = 24.85, p < 0.0001) h. Difference-score analyses also displayed significant effects after 1 [F(3,30) = 5.66, p < 0.003], 2 (F = 4.69, p < 0.008), and 4 (F = 7.62, p < 0.0006) h. Significant differences in food intake were observed following END and NBNI treatment conditions relative to vehicle after 1 [F(4,81) = 8.09, p < 0.0001], 2 (F = 13.22, p < 0.0001), and 4 (F = 22.85, p < 0.0001) h. Difference-score analyses also displayed significant effects after 1 [F(3,30) = 3.85, p < 0.019], 2 (F = 5.76, p < 0.003), and 4 (F = 4.48, p < 0.016) h.

NBNI and END-Induced Feeding. Significant increases in feeding relative to vehicle occurred across the time course following END and the lowest (5-nmol) NBNI dose paired with END, and after 4 h following the highest (40-nmol) NBNI dose paired with END (Fig. 3, bottom). END-induced feeding was significantly reduced by the two highest (20- and 40-nmol) NBNI doses after 2 and 4 h (Fig. 3, bottom), suggesting ₁-opioid mediation of END-induced feeding.

MOR-1 AS ODN Probes and END-Induced Feeding. Significant increases in feeding relative to vehicle occurred across the time course following END, following pretreatment with AS ODN probes directed against either exons 1 or 2 of the MOR-1 clone before END treatment, and following pretreatment with an MS ODN probe before END treatment, as well as 4 h following pretreatment with an AS ODN probe directed against exon 3 of the MOR-1 clone before END (Fig. 4, top). END-induced feeding was significantly reduced by AS ODN probes directed against exons 1 (4 h), 3 (2, 4 h), and 4 (2, 4 h) of the MOR-1 clone (Fig. 4, top). The AS ODN probe directed against exon 2 failed to alter END-induced feeding. The specificity of AS ODN probe effects was confirmed by the failure of END-induced feeding to be affected by the MS ODN probe (which differed from the MOR-1 exon 1 AS ODN probe by the sequence reversal of pairs of bases). These data indicate that END-induced feeding is dependent upon the integrity of exons 1, 3, and 4 of the MOR-1 clone for the full expression of this ingestive response.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Alterations (mean ± S.E.M.) in food intake (g) following i.c.v. administration of either vehicle alone, END alone, or ventricular pretreatment over 3 days (days 1, 3, and 5) with AS or MS ODN probes directed against specific exons of either the MOR-1 (top) or DOR-1 (bottom) opioid receptor clones with END administration occurring 24 h later (day 6). Significant differences in food intake were observed following END and MOR-1 AS ODN treatment conditions relative to vehicle after 1 [F(6,81) = 6.27, p < 0.0001), 2 (F = 8.21, p < 0.0001), and 4 (F = 13.92, p < 0.0001) h. Difference-score analyses also displayed significant effects after 1 [F(5,120) = 8.52, p < 0.0001], 2 (F = 7.34, p < 0.0001), and 4 (F = 20.08, p < 0.0001) h. Significant differences in food intake were observed following END and DOR-1 AS ODN treatment conditions relative to vehicle after 1 [F(4,59) = 11.88, p < 0.0001], 2 (F = 14.81, p < 0.0001), and 4 (F = 22.35, p < 0.0001) h. Difference-score analyses also displayed significant effects after 1 [F(3,57) = 3.57, p < 0.019], 2 (F = 2.59, N.S.), and 4 (F = 5.37, p < 0.0025) h.

DOR-1 AS ODN Probes and END-Induced Feeding. Significant increases in feeding relative to vehicle occurred across the time course following END and following AS ODN probes directed against exons 1, 2, and 3 of the DOR-1 clone paired with END (Fig. 4, bottom). END-induced feeding was significantly although transiently (1 h) reduced by the AS ODN probe directed against exon 1 of the DOR-1 clone; no other significant effects were observed (Fig. 4, top). These data indicate the relatively limited actions of AS ODN probes targeting different exons of the DOR-1 clone in the mediation of this response.

KOR-1 AS ODN Probes and END-Induced Feeding. Significant increases in feeding relative to vehicle occurred across the time course following END and following AS ODN probes directed against exons 1, 2, and 3 of the KOR-1 clone paired with END (Fig. 5, top). END-induced feeding failed to be significantly affected by any of the AS ODN probes directed against exons 1, 2, or 3 of the KOR-1 clone (Fig. 5, top). These data indicate the relatively limited actions of AS ODN probes targeting different exons of the KOR-1 clone in the mediation of this response.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Alterations (mean ± S.E.M.) in food intake (g) following i.c.v. administration of either vehicle alone, END alone, or ventricular pretreatment over 3 days (days 1, 3, and 5) with AS ODN probes directed against specific exons of either the KOR-1 (top) or KOR-3/ORL-1 (bottom) opioid receptor clones with END administration occurring 24 h later (day 6). Significant differences in food intake were observed following END and KOR-1 AS ODN treatment conditions relative to vehicle after 1 [F(4,60) = 13.85, p < 0.0001], 2 (F = 16.54, p < 0.0001), and 4 (F = 23.99, p < 0.0001) h. Difference-score analyses also displayed significant effects after 1 [F(3,60) = 5.97, p < 0.0012], 2 (F = 7.84, p < 0.0002), and 4 (F = 9.74, p < 0.0001) h. Significant differences in food intake were observed following END and KOR-3/ORL-1 AS ODN treatment conditions relative to vehicle after 1 [F(4,56) = 8.22, p < 0.0001], 2 (F = 10.89, p < 0.0001), and 4 (F = 16.36, p < 0.0001) h. Difference-score analyses also displayed significant effects after 1 [F(3,60) = 7.28, p < 0.0003], 2 (F = 8.06, p < 0.0001), and 4 (F = 9.62, p < 0.0001) h.

KOR-3/ORL-1 AS ODN Probes and END-Induced Feeding. Significant increases in feeding relative to vehicle occurred across the time course following END and following AS ODN probes directed against exons 1 and 2 of the KOR-3/ORL-1 clone paired with END, and after 2 and 4 h following the AS ODN probe directed against exon 3 of the KOR-3/ORL-1 clone paired with END (Fig. 5, bottom). END-induced feeding failed to be significantly affected by any of the AS ODN probes directed against exons 1, 2, or 3 of the KOR-3/ORL-1 clone (Fig. 5, bottom).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Increased food intake following END was significantly and dose-dependently attenuated by pretreatment with either general (Ntx), µ- (FNA), - (Nti) or ₁ (NBNI)-antagonists. In addition, END-induced feeding was significantly reduced following pretreatment with AS ODN probes directed against either coding exons 1, 3, or 4 of the MOR-1 gene and coding exon 1 of the DOR-1 gene. In contrast, AS ODN probes directed against any exons of either the KOR-1 or KOR-3/ORL-1 clones were ineffective. Furthermore, a control MS probe, that differed from an effective MOR-1 exon 1 AS ODN probe by the sequence reversal of only two pairs of bases, was also ineffective.

The potency of general and selective opioid receptor antagonists in reducing END-induced feeding was not uniform. The use of equimolar antagonist doses revealed that µ (FNA)-opioid antagonism significantly reduced END-induced feeding at far lower doses than general (Ntx), ₁- (NBNI), or  (Nti)-opioid antagonists. It should be noted that the K_I values to displace [³H]naloxone were very similar for the three selective antagonists: FNA (3.27 nM), NBNI (3.55 nM), and Nti (3.22 nM) (Newman et al., 2000), suggesting that minimal affinity differences among these antagonists could not account for the observed potency differences. Furthermore, the antagonist dose range used has previously demonstrated nonselectivity in blocking feeding elicited by selective opioid agonists. Thus, comparable doses of FNA and NBNI, respectively, reduced feeding elicited by both µ (DAMGO)- and  (U50488H)-opioid agonists to the same degree for each antagonist (Levine et al., 1990, 1991). Indeed, whereas a dose of 5 nmol of FNA significantly reduced END-induced feeding across the time course, an equimolar dose of the other antagonists was ineffective. The present study confirmed the attenuation of END-induced feeding by general opioid antagonists (Grandison and Guidotti, 1977; Leibowitz and Hor, 1982; Majeed et al., 1986; DePedro et al., 1995). Selective opioid receptor subtype antagonist participation in END-induced feeding was limited to the goldfish, and indicated activity by µ-selective antagonists (FNA, naloxonazine; DePedro et al., 1996), consistent with the present results. In contrast, END-induced feeding in goldfish was unaffected by either ₁- (NBNI, 5 µg), ₁- (7-benzidilidendenaltrexone, 5 µg), or ₂ (naltriben, 5 µg)-antagonists (DePedro et al., 1996). We also failed to observe NBNI effects at a comparable 5-nmol (3.65-µg) dose in rats. When comparing the respective inability of selective ₁- and ₂-receptor antagonists in goldfish with the ability of the general -antagonist, Nti in rats to reduce END-induced feeding, one should consider the lack of consistent receptor-selective effects observed for selective ₁- and ₂-agonists and antagonists in feeding studies (Yu et al., 1997). Taken together, these antagonist data suggest that µ-opioid receptors appear to play a sizable role in the mediation of END-induced feeding, yet other (- and ₁-) opioid receptors might participate to a lesser degree.

These data underscore inherent limitations in the exclusive use and subsequent interpretation of selective antagonist data. In addition to the use of equimolar doses, it is essential that these antagonists exert functionally specific and selective effects at their respective receptors. Although NBNI has been characterized as a selective and reversible ₁-opioid receptor antagonist (Portoghese et al., 1987), it also displays long durations of action (Horan et al., 1992) and blocks analgesia elicited by µ-, -, and -opioid agonists (Spanagel et al., 1994). Although Nti works with greater potency at -receptors, it can also act as an antagonist at µ-receptors (Portoghese et al., 1988). Since these antagonists only worked at high equimolar doses relative to FNA, it is conceivable that they could be exerting their effects through multiple rather than specific opioid receptors. Since none of the antagonist doses used in the present study completely eliminated END-induced feeding, this suggests that blockade of multiple opioid receptors might be necessary to produce this effect. This is in contrast to the ability of comparable doses of FNA to eliminate feeding elicited by the µ-selective agonists morphine, M6G, and DAMGO (Levine et al., 1991; Leventhal et al., 1997, 1998b), and comparable doses of NBNI to eliminate feeding elicited by the ₁-selective agonist U50488H (Levine et al., 1990). This concept is also consistent with END binding at µ, , , and the proposed  receptor (Chang et al., 1979; Schulz et al., 1979; Akil et al., 1980).

Thus, further converging evidence is required to determine the distinct pharmacological properties of END-induced feeding, and the use of AS ODN probes provided support for primary µ-receptor mediation of this response. AS ODN probes directed against either exons 1, 3, or 4 of the MOR-1 opioid receptor clone significantly reduced END-induced feeding. The unique sensitivity and specificity of END-induced feeding to transcriptional-translational disruption (for review, see Myers and Dean, 2000) by an AS ODN probe directed against exon 1 of the MOR-1 clone was further demonstrated by the failure of a control MS ODN probe, which differed from the MOR-1 exon 1 AS ODN by sequence reversal of two pairs of bases, to alter END-induced feeding. µ-Opioid-selective agonist-induced feeding, reversed by FNA pretreatment, was further distinguished using AS ODN probes directed against specific exons of the MOR-1 clone (Leventhal et al., 1997, 1998b). Thus, feeding elicited by either morphine or DAMGO was blocked by AS ODN probes directed against either exons 1 or 4, but not exons 2 or 3 of the MOR-1 clone (Leventhal et al., 1997, 1998b). The activity of AS ODN probes directed against either exons 1 or 4 of the MOR-1 clone suggests that END shares highly similar molecular binding profiles to that of morphine and DAMGO in eliciting feeding. However, since an AS ODN probe directed against exon 3 of the MOR-1 clone significantly reduced END-induced feeding, this shows that this response shares similarities with feeding elicited by the active morphine metabolite M6G, which is significantly reduced by pretreatment with AS ODN probes directed against either exons 2 or 3 of the MOR-1 clone (Leventhal et al., 1998b). The differential MOR-1 AS ODN sensitivity profiles displayed by morphine, DAMGO, and M6G suggested that they were potentially acting upon different splice variants or different isoforms of the MOR-1 clone (for review, see Pasternak and Standifer, 1995; Rossi and Pasternak, 1997), and the present data indicate that END-induced feeding appears to be mediated by multiple coding exon regions of the MOR-1 gene.

AS ODN probes directed against specific exons of the DOR-1, KOR-1, or KOR-3/ORL-1 opioid receptor clones provided converging data concerning specificity and selectivity of - and ₁-opioid antagonist data. The ability of AS ODN probes directed against coding exon 1, but not coding exons 2 or 3 of the DOR-1 clone to reduce END-induced feeding differs from the selective ability of an AS ODN probe targeting exon 3 of the DOR-1 clone to eliminate feeding induced by the ₂-opioid agonist Deltorphin II (Leventhal et al., 1998b). Since it has been suggested that the DOR-1 gene encodes the pharmacologically characterized ₂-opioid receptor subtype (Rossi et al., 1997), this differential profile of feeding responses induced by END and Deltorphin II suggests that Nti's inhibition of END-induced feeding is not acting through the ₂-subtype. The ability of NBNI at moderate and high doses to reduce END-induced feeding suggests a ₁-mechanism of action. NBNI completely blocks feeding induced by the ₁-opioid agonist U50488H (Levine et al., 1990), which in turn is eliminated by an AS ODN probe directed against exon 3 of the KOR-1 opioid receptor clone (Leventhal et al., 1998a). Since none of the AS ODN probes directed against the KOR-1 clone significantly reduced END-induced feeding, this suggests that NBNI's effect upon this feeding response was not mediated through the ₁-opioid receptor. A last opioid receptor clone, termed KOR-3/ORL-1, mediates the functional effects of orphanin FQ/nociceptin (Meunier et al., 1995), including its central stimulation of feeding (Pomonis et al., 1996). AS ODN probes directed against either exons 1, 2, or 3 of the KOR-3/ORL-1 clone each significantly reduce feeding elicited by orphanin FQ/nociceptin (Leventhal et al., 1998a), yet none of the probes were effective in altering feeding elicited by END, providing further evidence for the independent actions of these opioid receptor peptides in mediating their respective functional effects.

Thus, it appears that the opioid receptor mediating END-induced feeding is the pharmacologically described µ-opioid receptor. The gene(s) responsible for this action appears to be encoded by the MOR-1 clone since END-induced feeding shares similar, although not identical profiles in AS ODN studies using morphine, DAMGO, and M6G. The use of additional modern molecular tools to define precise receptor mechanisms mediating END-induced feeding will allow an understanding into the roles of how release of endogenous opioid peptides under normal and challenged ingestive states control this important homeostatic behavior.