Abstract
The recently isolated peptides endomorphin-1 and endomorphin-2 have been suggested to be the endogenous ligands for the mureceptor. In traditional opioid receptor binding assays in mouse brain homogenates, both endomorphin-1 and endomorphin-2 competed bothmu1 and mu2receptor sites quite potently. Neither compound had appreciable affinity for either delta orkappa1 receptors, confirming an earlier report. However, the two endomorphins displayed reasonable affinities for kappa3 binding sites, withKi values between 20 and 30 nM. Both endomorphins competed3H-[d-Ala2,MePhe4,Gly(ol)5] enkephalin binding to MOR-1 receptors expressed in CHO cells with high affinity. In mouse brain homogenates 125I-endomorphin-1 and125I-endomorphin-2 binding was selectively competed by mu ligands. 125I-Endomorphin-1 and125I-endomorphin-2 also labeled MOR-1 receptors expressed in CHO cells with high affinity. Autoradiography of the two125I-labeled endomorphins demonstrated regional patterns in the brain similar to those previously observed for mudrugs. Pharmacologically, the endomorphins were potent analgesics. Although they were equipotent supraspinally, endomorphin-1 was more potent spinally. Endomorphin analgesia was effectively blocked by naloxone, as well as the mu-selective antagonists β-funaltrexamine and naloxonazine. In CXBK mice, which are insensitive to supraspinal morphine, neither endomorphin was active, consistent with a mu mechanism of action. Finally, the endomorphins inhibited gastrointestinal transit. In conclusion, these results support the mu selectivity of these agents.
The opioids were among the earliest neuropeptides identified in the nervous system. The enkephalins were the first, followed soon afterward by the dynorphins and β-endorphin (Evans et al., 1988; Reisine and Pasternak, 1996; Pasternak, 1993). The enkephalins are the endogenous ligands for the delta class of opioid receptors and dynorphin A is the endogenous ligand for thekappa1 receptor. The mu receptor was the first opioid receptor identified in binding assays and its importance is further emphasized by its importance in mediating the analgesic actions of most clinically used analgesics (Reisine and Pasternak, 1996; Pasternak, 1993). Yet, the search for the endogenous ligand for the mu receptor has lagged far behind the other opioid receptor subtypes. A number of opioid peptides have high affinity for mu receptors, including dynorphin A and β-endorphin. However, dynorphin A labelskappa1 receptors far more potently thanmu sites and β-endorphin binds equally well tomu and delta sites. Thus, many investigators felt that these peptides were not the endogenous ligand for mureceptors based on these selectivity profiles.
The recent identification of endomorphin-1 and endomorphin-2 has opened a new area of research in the mu opioid system (Zadinaet al., 1997). Although related to each other, the sequences of endomorphin-1 (Tyr-Pro-Trp-Phe-NH2) and endomorphin-2 (Tyr-Pro-Phe-Phe-NH2) are quite distinct from traditional opioids in which the first four amino acids are Tyr-Gly-Gly-Phe followed by either methionine or leucine. Yet, both peptides label mu receptors with high affinity and selectivity, raising the possibility that they may represent two endogenous mu receptor ligands. To further evaluate these two peptides, we have extended these initial studies on the endomorphins.
Materials and Methods
Male CD-1 mice (24–30 g: Charles River Laboratories, Raleigh, NC) were housed in groups of five with food and water ad libitum. Animals were maintained on a 12-hr light/dark cycle. Fresh calf brains were obtained from a local slaughterhouse (Max Insel Cohen, NJ). Endomorphin-1 and endomorphin-2 were synthesized at our institution’s Microchemistry Facility, purified by high-performance liquid chromatography, their structures verified by mass spectroscopy and the peptide content (58% for endomorphin-1 and 74% for endomorphin-2) determined by Rockefeller University’s Protein Technology Center. Naloxonazine and naloxone benzoylhydrazone were synthesized in our laboratory as previously published (Luke et al., 1988; Hahn et al., 1982) and norBNI and naltrindole were purchased from Research Biochemicals International (Natick, MA). Naloxone, β-FNA and the other opioids and opioid peptides were provided by the Research Technology Branch of the National Institute on Drug Abuse (Rockville, MD). Chemicals were purchased from either from Fisher Scientific (Pittsburgh, PA) or Sigma Chemical Co. (St. Louis, MO). Halothane was obtained from Halocarbon Laboratory (Hackensack, NJ).
Preparation of membranes.
Brains were homogenized in 50 volumes of treated buffer (50 mM Tris, pH 7.4 at 25°C, 1 mM EDTA and 100 mM NaCl) (Clark et al., 1989; Pasternak et al., 1975; Pasternak and Snyder, 1975). The brain homogenate then was incubated with phenylmethanesulfonyl fluoride (0.1 mM) for 15 min at 25°C water bath, centrifuged at 20,000 × g for 30 min and the resulting pellet resuspended in 6 volumes of the original wet weight in sucrose (0.32 M). Membranes from CHO cells stably transfected with MOR-1 were prepared as previously described (Brownet al., 1997b) and stored in sucrose (0.32 M). Homogenates stored at −80°C retained binding for at least 4 wk. Protein content was determined by the Lowry method (Lowry et al., 1951).
Iodination of endomorphin-1 and endomorphin-2.
The peptides were iodinated with Na125I (Du Pont, Wilmington, DE) and chloramine T for 80 sec at room temperature with a peptide/NaI molar ratio of 10:1, after which the reaction was terminated with sodium metabisulfite, as previously described (Mathis et al., 1997). The iodinated peptides were purified by high-performance liquid chromatography over a Rainin Microsorb-MV C18 reverse-phase column (Woburn, MA) using an acetonitrile gradient (25–50%) over a 50-min period. Both 125I endomorphin-1 and 125I-endomorphin-2, eluting at approximately 37% acetonitrile/0.1% TFA, were readily separated from their noniodinated peptides, which eluted at about 27% acetonitrile/0.1% TFA
125I Endomorphin-1 and endomorphin-2 binding.
125I-Endomorphin-1 or125I-endomorphin-2 binding (0.2 nM) was performed in potassium phosphate buffer (50 mM, pH 7.4; 0.5 ml) with MgCl2 (5 mM) at a tissue concentration of 10 mg wet weight/ml for brains or 0.06 mg protein/ml for MOR-1/CHO cells. Specific binding was determined in the presence and absence of either 1 μM of the corresponding unlabeled peptide. The entire mixture was then incubated at 25°C for 1 hr and filtered over no. 32 glass fiber filters (Schleicher & Schuell, Keehne, NH) which had been presoaked for 1 hr in 0.5% polyethylenimine and washed twice with ice cold Tris buffer using a Brandel cell harvester (Cambridge, MA). The filters were then counted on a Packard Cobra gamma counter. The other opioid receptor binding assays were performed as previously described (Clarket al., 1988, 1989).
Autoradiography.
Brains were removed from male CD-1 mice and quickly frozen in isopentane at −50°C. The brains were cryosectioned at 10 μm and thaw-mounted onto gelatin coated slides. Tissue sections were preincubated with buffer (50 mM potassium phosphate buffer, pH 7.4 and 5 mM MgCl2) for 15 min to remove endogenous ligands. For 125I-endomorphin-1 experiments the buffer also contained 0.1% bovine serum albumin to reduce binding to white matter. The sections were then incubated with125I-endomorphin-1 or125I-endomorphin-2 (1 nM) for 2 hr at 25°C. Nonspecific binding was determined in adjacent sections with the corresponding unlabeled peptide (1 μM). After the incubation, sections were rinsed in fresh buffer for 20 min at room temperature (125I-endomorphin-1) or 0°C (125I-endomorphin-2). Sections were then dried under a stream of cool air and exposed to Hyperfilm (Amersham Corp., Arlington Heights, IL) for 47 hr (125I-endomorphin-1) or 17 hr (125I-endomorphin-2).
Analgesic assays.
The endomorphins were administered i.c.v. or intrathecally under light halothane anesthesia as previously reported (Rossi et al., 1995, 1997). Antinociception, termed analgesia for convenience, was assessed in the radiant heat tail-flick assay with baseline latencies between 2 to 3 sec and a maximum cutoff score of 10 sec to minimize tissue damage. Analgesia was determined quantally as a doubling or greater of baseline tailflick scores. Dose-response curves were analyzed to generate ED50 values and 95% confidence limits were generated from dose-response curves using a computer program based on the Litchfield-Wilcoxin approach (Tallarida and Murray, 1987). Peak analgesia was seen at 10 min for endomorphin-1 and 15 min for endomorphin-2. Results reflect peak analgesic values unless stated otherwise.
Gastrointestinal transit.
Groups of mice were treated i.c.v. with endomorphin-1 (12 μg) or endomorphin-2 (3 μg) 15 min before a 0.5-cc charcoal meal (2.5% gum tragacanth,10% activated charcoal in water). The mice were killed 30 min later and the distance the charcoal traveled was measured.
Results
Endomorphin binding to opioid receptors.
The initial report describing the endomorphins indicated that they weremu-selective (Zadina et al., 1997). First, we examined the affinity of the two endomorphins in traditional opioid binding assays in brain homogenates (table1). Both endomorphin-1 and endomorphin-2 competed binding to mu receptors with high affinity. As noted with most opioids, the affinity of these compounds for themu1 receptors was greater than for themu2 receptor. We also found that the endomorphins had little appreciable affinity for eitherdelta or kappa1 binding sites, with Ki values greater than 500 nM. They did, however, show moderate affinity for the kappa3 site, lowering binding withKi values between 20 and 30 nM.
We next determined the affinity of the endomorphins against the recently cloned mu receptor MOR-1 (Wang et al., 1993; Chen et al., 1993) stably expressed in CHO cells (table 1). As expected, the compounds competed3H-DAMGO quite effectively. Prior studies have suggested that the binding to the expressed MOR-1 receptor (Wanget al., 1993; Chen et al., 1993) resemble most closely the profile to mu2 receptors. In our studies, we also observed a closer correspondence between binding results in the expressed MOR-1 clone and the values observed in againstmu2 binding in brain (table 1).
125I-Endomorphin binding in mouse brain and MOR-1/CHO membranes.
To extend previous reports, both endomorphins were iodinated and examined in binding studies. Initial studies established that binding reached steady state levels by 60 min at 25°C in potassium phosphate buffer (pH 7.4). In addition, there was little change in specific binding between pH 7 and 7.5 (data not shown) and the physiological pH of 7.4 was used for all subsequent assays. Lowering the temperature to 0°C drastically reduced specific binding whereas increasing the temperature to 37°C had little advantage. Binding was performed with tissue at 10 mg wet weight/ml, although binding remained linear up to tissue concentrations as high as 15 mg wet weight tissue/ml. As previously observed in traditional opioid receptor assays (Pasternak et al., 1975a), magnesium enhanced binding and was included in all assays.
125I-Endomorphin-1 and125I-endomorphin-2 both displayed saturable binding in mouse brain homogenates (fig.1a) with affinities in the low nanomolar range (table 2). The two radioligands also labeled the mu receptors expressed in the MOR-1/CHO cells. The affinity of the ligands was similar to that seen in mouse brain membranes. The differences in Bmax values between the two radiolabeled peptides in the MOR-1/CHO assays presumably reflects the use of different membrane preparations.
In the MOR-1/CHO membranes, the competition profile was consistent with a mu site (table 3). Both morphine and DAMGO competed binding quite effectively, withKi values similar to those seen with traditional mu radioligands in this assay (Clark et al., 1988, 1989; Brown et al., 1997). Endomorphin-1 and endomorphin-2 both competed binding in these assays quite potently. It is interesting to note that their Ki values against 125I-endomorphins (table 3) were quite a bit lower than against 3H-DAMGO in the same transfected cell line (table 1). Because both assays measured the binding to the same receptor, these results imply that the endomorphins may sit in the binding pocket somewhat differently from DAMGO.
We next characterized 125I-endomorphin binding in mouse brain membranes. Competition studies confirmed the high affinity of unlabeled endomorphins for these sites (table4). Both endomorphin-1 and endomorphin-2 displayed Ki values in the low nanomolar range and showed a potency similar to that seen in the MOR-1/CHO membranes with Hill coefficients close to unity (table 3). Traditionalmu ligands also lowered binding quite potently. Morphine, DAMGO, fentanyl, M6G and naloxone all hadKi values similar to those previously reported in mu receptor binding assays (Clark et al., 1988, 1989). However, a number of these agents revealed shallow competition curves with Hill coefficients far less than unity, particularly against 125I-endomorphin-1 binding.
The delta-selective peptide DPDPE and thekappa1 drug U50,488H were not very effective against radiolabeled endomorphin binding in brain, consistent with the proposed mu selectivity of the radioligands. However, the results with MeONtx a potent and selective M6G antagonist in vivo, were unexpected. In binding studies, MeONtx competes3H-morphine binding in brain (Ki 62 nM) quite poorly and is even less active against MOR-1 receptors expressed in CHO cells (Ki 255 nM) (Brown et al., 1997a). Its relatively poor affinity in traditional mu receptor binding assays contrasts with a high affinity for the3H-M6G site in mouse brain (Ki 12 nM) (Brown et al., 1997a). In our study MeONtx competed125I-endomorphin binding in brain homogenates as effectively as 3H-M6G.
125I-Endomorphin binding also was sensitive to the endogenous opioid peptides (table 4). [Met5]enkephalin was slightly more potent than [Leu5]enkephalin. The others also lowered binding with potencies similar to those seen against traditional mu radioligands. The competition curves for α-neoendorphin and dynorphin B were shallow with Hill slopes lower than unity, particularly against125I-endomorphin-1.
Pharmacology of endomorphin-1 and endomorphin-2.
Both endomorphin-1 and endomorphin-2 were potent analgesics with peak effects seen at 10 and 15 min, respectively. All subsequent studies were performed at peak effect. Both compounds were fully active supraspinally and spinally, with no indication of ceiling effects. Endomorphin-1 was significantly more potent spinally than supraspinally and, at the spinal level, it was significantly more potent than endomorphin-2 (fig. 2; table5). The response of both agents were readily reversed by naloxone (fig. 3). β-FNA, a highly selective mu antagonist, effectively reversed the actions of both endomorphins, as previously noted (Zadina et al., 1997). Naloxonazine is another mu antagonist, but its actions are limited to mu1 and M6G receptors (Paul et al., 1989; Ling et al., 1986; Hahnet al., 1982; Pick et al., 1991; Paul and Pasternak, 1988). Although naloxonazine significantly lowered endomorphin-1 and endomorphin-2 analgesia, its actions were not as complete as β-FNA. Neither the kappa1antagonist norBNI nor the delta antagonist naltrindole were active against either endomorphin (fig.4).
CXBK mice have proven valuable in the assessment of muanalgesics (Moskowitz and Goodman, 1985a; Reith et al., 1981a; Baron et al., 1975a; Pick et al., 1993a). Morphine and DAMGO are not active analgesics after supraspinal administration in CXBK mice, consistent with the reported deficiency in mu1 binding sites in this strain. However, othermu drugs, including M6G, heroin, 6-acetylmorphine and fentanyl retain their analgesic potency in CXBK mice, presumably acting through a different subset of mu receptors (Rossi et al., 1996). Both endomorphin-1 and endomorphin-2 displayed a profile similar to morphine (fig. 5). Neither compound had analgesic activity in CXBK mice at a dose which produced over 70% analgesia in control CD-1 mice.
Morphine and other mu drugs effectively decrease gastrointestinal transit (Reisine and Pasternak, 1996). As with traditional mu drugs both endomorphins significantly inhibited gastrointestinal transit in mice (fig.6).
Autoradiography of 125I-endomorphin binding in mice.
Finally, we examined the regional distribution of125I-endomorphin binding in mouse brain (fig.7).125I-Endomorphin-1 and125I-endomorphin-2 binding patterns were established in sections from the striatum through the brain stem. Overall, the pattern of labeling was quite similar to that previously reported for mu receptors (Goodman and Pasternak, 1985;Atweh and Kuhar, 1977a, b, c; Moskowitz and Goodman, 1985; Kuharet al., 1973). In the striatum (fig. 7a and b), both peptides exhibited higher labeling in the patches as compared to the surrounding matrix. High binding levels were also seen in the nucleus accumbens core and shell, as well as the anterior and medial thalamic and the amygdaloid nuclei (fig. 7c-h). At the level of the brainstem, both peptides labeled the periaqueductal gray, superior colliculus and interpeduncular nucleus (fig. 7I and h). Finally, a layered distribution of binding was observed with both peptides throughout the cortex and hippocampus.
Discussion
The original description of the two endomorphins revealed that both compounds had a profound mu selectivity (Zadina et al., 1997). In this initial study both endomorphins competed mubinding over 1000-fold more potently than either delta orkappa1 receptors (Zadina et al., 1997). In the current studies, the two endomorphins also displayed very poor affinities for delta and kappa1receptors and high affinity for both mu receptor subtypes. Most opioids label mu1 receptors more potently than mu2 sites and the same trend was seen with the endomorphins, which displayed a 5- to 10-fold greater affinity in the mu1 binding assay. The high affinity of the two endomorphins for mu receptors was confirmed in competition studies against the cloned mu receptor MOR-1.
The initial study exploring the selectivity of the compounds did not examine kappa3 binding. We found that the endomorphins also competed kappa3 receptors moderately well, lowering binding with Ki values between 20 and 30 nM. Although the endomorphin-1 and endomorphin-2 remain selective for mu receptors, their selectivity is not as great as initially proposed.
Radiolabeling the endomorphins did not appreciably affect their affinity in assays using either mouse brain homogenates or MOR-1/CHO cell membranes. 125I-Endomorphin-1 displayed an affinity in the MOR-1/CHO cells which was slightly better than endomorphin-1 itself although 125I-endomorphin-2 labeled sites with an affinity similar to that seen with the noniodinated endomorphin-2. The maintenance of affinity after iodination contrasts with the typical 10-fold decrease in affinity seen when traditional opioid peptides are iodinated. The iodine is located on the N-terminal tyrosine in all the compounds, suggesting significant differences in the way the traditional opioid peptides and the endomorphins sit in the binding pocket.
Overall, the competition studies were consistent with a high affinity of the 125I-endomorphins for mureceptors. Typical mu drugs were quite potent in these studies although kappa and delta selective agents were not. In the MOR-1/CHO cell assay that contains only a single site, the Hill coefficients were close to unity, contrasting with the shallow competition studies in brain. For example, morphine has a Hill coefficient close to unity in the MOR-1/CHO membrane binding assay, but only about 0.5 against 125I-endomorphin-1 in brain. The meaning of the low Hill slopes in the brain is not clear, but they might represent evidence that the labeling in brain may not reflect a single site. These issues will require further study.
In vivo, the endomorphins are potent analgesics, both spinally and supraspinally. In the first description of endomorphin-1,Zadina et al. (1997) reported significant analgesic activity after both spinal and supraspinal administration, with a 3-fold greater potency after spinal administration (Zadina et al., 1997). We also observed this difference between the two sites. Indeed, our ED50 values are remarkably similar to the earlier ones.
Previous work from our laboratory has suggested that differentmu receptor subtypes are responsible for spinal and supraspinal morphine analgesia (Pick et al., 1991, 1993;Ling and Pasternak, 1983). Mu1 receptors mediate supraspinal morphine analgesia although mu2receptors are responsible for spinal analgesia. The affinity of endomorphin-1 for the mu receptor subtypes in binding assays does not explain its greater spinal activity because it competedmu1 binding approximately 5-fold more potently than mu2 sites. Thus, the reasons for the enhanced spinal activity remain unclear. However, both the earlier results and the current ones find significant spinal/supraspinal differences in the potency of endomorphin-1.
Endomorphin-2 was active at both levels of the neuroaxis, with no significant difference in the sensitivity of the two sites. However, spinal endomorphin-2 was significantly less active than endomorphin-1. An earlier study exploring the analgesic activity of the endomorphins at the spinal level found that the two compounds were not significantly different in a thermal assay (Stone et al., 1997). Although our results with endomorphin-1 were quite similar to the earlier report, endomorphin-2 was far less active in our studies, perhaps due to different assays. Whereas we used a quantal radiant tailflick assay, the earlier report used a warm water immersion tailflick assay with graded responses. The question of peptide stability also must be considered.
The mu characteristics of the endomorphins were further supported by a variety of studies. Pharmacologically, endomorphin analgesia was reversed by the mu-selective antagonist β-FNA, implying that they were acting through mureceptors. Although the actions of both endomorphins were significantly reduced by the mu1 antagonist naloxonazine, the blockade was not as great as with β-FNA, particularly for endomorphin-1. However, the significance of this difference is not clear. Neither the kappa nor the delta-selective antagonists were effective, but the general opioid antagonist naloxone potently blocked the response, confirming its opioid nature. The inactivity of the two endomorphins in CXBK mice, a strain that is insensitive to morphine (Moskowitz and Goodman, 1985a; Reith et al., 1981a; Baron et al., 1975a; Pick et al., 1993a), also supported their mu selectivity.Mu drugs typically inhibit gastrointestinal transit and both endomorphins had similar actions. Finally, the regional distribution of125I-labeled endomorphins resembled that seen with traditional mu radioligands (Goodman and Pasternak, 1985; Kuhar et al., 1973).
In conclusion, the endomorphins are highly selective endogenousmu ligands. The binding selectivities and pharmacology of the peptides in the current studies support the possibility that they may represent the endogenous mu receptor ligands, as originally proposed (Zadina et al., 1997). However, several features of these peptides raise the possibility that their actions may involve more than just traditional mu receptors. A number of opioids competed binding with shallow hill slopes and preliminary studies suggest that the endomorphins retain analgesic activity in an MOR-1 knockout mouse model (King M, Schiller A, Pintar J and Pasternak GW, in preparation). The significance of these observations is not yet clear, but they do suggest that additional studies are needed to more fully define these potential endomorphin systems.
Footnotes
-
Send reprint requests to: Dr. Gavril W. Pasternak, Department of Neurology, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021.
-
↵1 This work was supported, in part, by grants DA02615 and DA07242 from the National Institute on Drug Abuse to G.W.P. and a core grant from the National Cancer Institute to Memorial Sloan-Kettering Cancer Center. G.C.R. was supported by Mentored Scientist Award DA00310, G.W.P. was supported by Research Scientist Award DA00220 from the National Institute on Drug Abuse and S.L. was supported by Training Grant CA09461 from the National Cancer Institute.
- Abbreviations:
- β-FNA
- β-funaltrexamine
- i.c.v.
- intracerebroventricularly
- MeONtx
- 3-methoxynaltrexone
- CHO
- Chinese hamster ovary
- TFA
- trifluoroacetic acid
- DAMGO
- [d-Ala2, MePhe4, Gly (01)5]enkephalin
- Received January 21, 1998.
- Accepted April 13, 1998.
- The American Society for Pharmacology and Experimental Therapeutics