Developmental Potential for Endomorphin Opioidmimetic Drugs

Morphine, which is agonist for μ-opioid receptors, has been used as an anti-pain drug for millennia. The opiate antagonists, naloxone and naltrexone, derived from morphine, were employed for drug addiction and alcohol abuse. However, these exogenous agonists and antagonists exhibit numerous and unacceptable side effects. Of the endogenous opioid peptides, endomorphin(EM)-1 and endomorphin(EM)-2 with their high μ-receptor affinity and exceptionally high selectivity relative to δ- and κ-receptors in vitro and in vivo provided a sufficiently sequence-flexible entity in order to prepare opioid-based drugs. We took advantage of this unique feature of the endomorphins by exchanging the N-terminal residue Tyr1 with 2′,6′-dimethyl-l-tyrosine (Dmt) to increase their stability and the spectrum of bioactivity. We systematically altered specific residues of [Dmt1]EM-1 and [Dmt1]EM-2 to produce various analogues. Of these analogues, [N-allyl-Dmt1]EM-1 (47) and [N-allyl-Dmt1]EM-2 (48) exhibited potent and selective antagonism to μ-receptors: they completely inhibited naloxone- and naltrexone-induced withdrawal from following acute morphine dependency in mice and reversed the alcohol-induced changes observed in sIPSC in hippocampal slices. Overall, we developed novel and efficacious opioid drugs without deleterious side effects that were able to resist enzymatic degradation and were readily transported intact through epithelial membranes in the gastrointestinal tract and the blood-brain-barrier.


Introduction
Morphine, which represents the quintessential agonist for μ-opioid receptor, has been used as a pain-killing drug for millennia. Since natural occurring opioid antagonists are nonexistent, naloxone and naltrexone were derived from morphine and currently find use in drug addiction and alcohol cessation programs; however, these alkaloid-derived antagonists exhibit numerous deleterious side effects. In 1975, the endogenous opioid peptides enkephalins (H-Tyr-Gly-Gly-Phe-Met-OH/Leu-OH) were discovered [1], followed sequentially by the endorphins [2], dynorphins [3], and the endomorphins [4], all of which are involved in the modulation and attenuation of pain and regulation of homeostatic mechanisms.
Of the endogenous opioid peptides, endomorphin-1 (EM-1: H-Tyr-Pro-Trp-Phe-NH 2 ) and endomorphin-2 (EM-2: H-Tyr-Pro-Phe-Phe-NH 2 ) exhibited high μ-opioid receptor affinity (K i = 0.36 and 0.69 nM, resp.) with high selectivity: 4,000-and 13,000-fold preference over the δopioid receptor and a similar 15,000-and 7,500-fold preference for μ-receptor relative to κ-opioid receptors [4]. These data underline the potential importance of these opioid ligands in all phases of human homeostatic mechanisms. Considering this premise, our research was directed toward the eventual development of endomorphin opioidmimetics, which would exhibit agonist and antagonist properties with potentially minimal side effects. We review the approach in this field, focusing basic research on key factors in the rational development of novel and highly efficacious opioid drugs able to resist enzymatic degradation and readily transported intact through epithelial membranes in the gastrointestinal tract and the blood-brain barrier.

Properties of Endomorphin Analogues
Opioid peptides and their G-protein-coupled receptors (δ, κ, and μ), which are distributed in the central nervous system 2 International Journal of Medicinal Chemistry and peripheral tissues, were initially classified on the basis of their functional pharmacological activity. However, despite a common mode of biological action as agonists, the structural differences among opioids permitted a division into two separate classes based on their N-terminal message domain: namely, H-Tyr-Gly-Gly-Phe-, a sequence that comprises the enkephalins, endorphins, and dynorphins, while H-Tyr-Pro-Trp/Phe-defines endomorphins-1 and -2. It is the unique sequence of the latter opioids that gave rise to their flexibility in the production of bioactive analogues.

Synthesis of Stereoisomeric Analogues of Endomorphin-2 and Their
Activities. Initially, in order to gain insight on the interaction between opioid ligands with their receptors, we substituted d-amino acids into endomorphin-2 [5]. The rationale for the use of d-amino acids is their ability to generally affect biological activity due to a subtle change induced in peptide conformation that, if bioactive, can lead to enhanced stability against enzymatic degradation [6].
Endomorphin-2 and d-amino acid containing stereoisomers were prepared by Fmoc solid-phase method using Fmoc (9-fluorenylmethyloxycarbonyl) amide resin as follows: solid support, Fmoc-d-or l-Tyr(Bu t )-OH, Fmoc-d-or l-Pro-OH, Fmoc-d-or l-Phe-OH, and HBTU/HOBt/DMF, DIEPA/NMP were used. After each coupling reaction, the Fmoc group was removed by piperidine/NMP. For the final deblocking, dried protected peptide resin was suspended in TFA/H 2 O, and the reaction mixture was stirred at room temperature for 2 h. The material was filtered and ether added to filtrate to precipitate the peptides, which were collected by filtration and lyophilized from 1 M HCl to >98% purity. Receptor binding data are detailed in Table 1 (2-17) [5]. All d-amino acids containing analogues exhibited less binding affinities to the μ-opioid receptor (K i = 24.3-2,755 nM), resulting in the loss of high selectivity over δopioid receptor (K i δ/K i μ = 2.6-177). Interestingly, although [d-Pro 2 ]EM-2 (12) exhibited only low affinity towards the μ-receptor (K i = 512.4 nM), it substantially exhibited more potent and longer activity in an in vivo tail flick test in mice compared to EM-2 [7]. These data clearly indicate an enhanced bioactivity most likely due to its resistance to proteolytic degradation, presumably by dipeptidyl peptidase IV [8]. [2 ,6 -Dimethyl-l-tyrosine 1 (Dmt 1 )]EM-2 Analogues: Structure-Activity Relationships. In order to develop potentially more potent analgesics, 2 ,6 -dimethyll-tyrosine (Dmt) was substituted for Tyr as the N-terminal residue, since Dmt markedly increases the affinity and bioactivity of numerous opioid peptide agonists and antagonists [9,[14][15][16]. Optically pure 2 ,6 -dimethyl-l-tyrosine was prepared as previously described [17].

Synthesis of
As summarized in Tables 1 and 3, substitution of Dmt 1 in EM-1 and EM-2 and in C-terminal deletion analogues profoundly affected all the measured parameters. In each case, the affinity of [Dmt 1 ]EM-1 (19) and [Dmt 1 ]EM-2 (20) towards the μ-opioid receptor increased 6.6 and 4.6 times compared to the parent molecules (1, 2), respectively, and increased δ-opioid receptor affinity by 270-and 327fold. The functional bioactivity of [Dmt 1 ]EM-1 (Table 3, 19) increased μ-bioactivity by 15-fold over EM-1. Interestingly, [Dmt 1 ]EM-1 (19) was transformed to potent mixed μagonist/δ-antagonist, while the bioactivity of [Dmt 1 ]EM-2 (20) greatly increased both μand δ-agonist bioactivities by 98-and 184-fold greater than EM-2, respectively. Similarly, the deletion of C-terminal carboxyl group of [Dmt 1 ]EM-2 to yield H-Dmt-Pro-Phe-NH-C 2 H 4 -Ph (22) also exhibited mixed μ-agonist/δ-antagonist properties, but with over an order of magnitude less activity than those observed for 19. The marked change in the Dmt-containing analogues relative to both receptor interaction and bioactivity could be a result of an alteration in the topography of the peptide. In fact, the 1 H NMR spectra of EM-2 analogues revealed that the rotamers around the Dmt-Pro amide bond existed predominantly in the cis configuration [9].
These data substantiate that N-terminal Dmt-containing ligands permit development of novel bioactive opioidmimetics for potential therapeutic and clinical applications. The methyl groups on the tyramine ring of Dmt undoubtedly play a dominant role in the interaction within the opioid ligand-binding domains either by direct interaction with hydrophobic side chains of receptor residues or more interestingly by stabilization of favored cis conformer in solution prior to and during binding, or a combination of both mechanisms.

Synthesis of μ-Opioid Receptor Ligands Incorporating
Unique Tyrosine Analogues. The enhancement of opioid
As summarized in Table 1, the alkylated Phe 3 analogues essentially enhanced the affinities for both μand δ-opioid receptors in these [Dmt 1 ,Xaa 3 ]EM-2 ligands (40-46). Of these analogues, the highest μ-opioid selectivity occurred with [Dmt 1,3 ]EM-2 (43) (K i δ/K i μ = 878). One analogue of considerable interest is [Dmt 1 ,Tmp 3 ]EM-2 (44) with a 44-fold enhancement toward δ-opioid receptors relative to [Dmt 1,3 ]EM-2 (43). This suggested that the hydrogen donor capacity of the hydroxyl group of Dmt was apparently less effective in affecting receptor interaction when substituted within the sequence of the peptide than the hydrophobicity of a 4 methyl group; that is, the hydroxyl group may contribute a negative influence when it occurred as an internal residue. κ-Opioid receptor affinities for Dmt derivatives (40-46) were quite weak relative to the interaction of these peptides to both μand δ-opioid receptors [13].
These data permitted us to conclude the following: (i) the bulky side chain of Trp in combination with Dmt 1 caused either a steric hindrance in the conformation of the peptide or a shift in hydrophobicity to potentiate the induction of δ-opioid antagonism; (ii) [Dmt 1 ,Emp 3 ]EM-2 (45) and [Dmt 1 ,Imp 3 ]EM-2 (46) exhibited dual μ/δ-agonism similar to that seen for [Dmt 1 ]EM-2 (20), while compounds 40-44 had δ-opioid antagonism ranging from a weak pA 2 = 6.59 to a potent pA 2 = 9.05. Thus, these bifunctional molecules are targets in the design of new antinociceptive opioids that could potentially alleviate acute or chronic pain with a low degree of physical dependence and tolerance [25].

Transformation of [Dmt 1 ]EM-1 and [Dmt 1 ]EM-2 into Potent and μ-Selective
Antagonists. The development of potent and selective opioid antagonists, especially μ-opioid receptor antagonists, is very important in order to delineate critical biochemical, pharmacological, and physiological roles played by these receptors and for their possible application as clinically and therapeutically relevant agents. Table 1 revealed that [N-allyl-Dmt 1 ]EM-1 (47) exhibited better affinity compared to [N-allyl-Dmt 1 ]EM-2 (48); however, in terms of their in vitro functional bioactivity (Table 3), [Nallyl-Dmt 1 ]EM-2 (48) exhibited somewhat better μ-opioid antagonism with pA 2 = 8.59 versus pA 2 = 8.18 for [N-allyl-Dmt 1 ]EM-1 (47) [10]. Furthermore, both antagonists are defined as neutral μ-antagonists due to their lack of inverse agonist properties determined by functional guanosine 5 -O-(3-[ 35 S]thiotriphosphate) assays in vitro from membranes of cells grown in the presence of morphine or alcohol [26]. They also completely inhibited naloxone-and naltrexone-elicited withdrawal symptoms following acute morphine dependency in mice [26]. [N-Allyl-Dmt 1 ]EM-2 (48) induced a dose-dependent suppression of an ethanol-induced increase of sIPSC frequency with full reversal at 300 nM that was several orders of magnitude more potent than naltrexone [27]. These results suggest a potential therapeutic application in the treatment of drug addiction and alcohol abuse without the adverse effects observed with inverse agonist alkaloidderived compounds, such as naltrexone and naloxone that produce severe withdrawal symptoms.

Agonists.
The presence of Dmt in lieu of Tyr 1 in opioid peptides enhanced affinities, bioactivity, and analgesia. In order to assess the possible effect of Dmt per se on opioid activities, H-Dmt-NH-CH 3 was prepared and examined [28]. This compound had K i μ = 7.45 nM and K i δ = 460 nM values that were nearly equivalent to those of morphine. However, the in vitro bioactivity in a GPI assay was three orders of magnitude lower than that of EM-2 and [Dmt 1 ]EM-2 and essentially inactive in the MVD assay. Its analgesic response relative to morphine was insignificant (0.64% in hot-plate test and 1.3% in tail-flick test). According to the message-address concept of opioid functionality [29], Dmt would be considered an important pharmacophore interacting within opioid receptors as an integral component of the message domain even though it had no intrinsic activity of its own. Thus, to test this hypothesis, we set out to construct ligands containing two message and address domains.
Compound 57 produced naloxone reversible analgesia by i.c.v., s.c. and oral (po) administration. While i.c.v. analgesia was 50-and 20-fold more potent than morphine in the tail-flick and hot-plate tests, respectively, both s.c. and p.o. were somewhat less active than morphine. These results demonstrated that compound 57 crossed epithelial membrane barriers in both the intestine and microcapillaries in mouse brain to interact with brain μ-opioid receptors. Similar conclusions were obtained by Igarashi et al. [31] and Koda et al. [32]. These results indicated that pyrazinone derivatives could be potential candidates for clinical and therapeutic applications in the treatment of pain arising from postoperative procedure or cancer, associated with birth, or act as possible veterinary drugs.

Development of μand δ-Opioid Receptor Antagonists by
Dimerization of Dmt-Tic with Diaminoalkanes or Diaminoalkylpyrazinones. We expanded our studies with Dmt through the synthesis and analysis of the biological properties of unique series of dimeric H-Dmt-Tic (2 ,6 -dimethyl-ltyrosyl-1,2,3,4-tetrahydroisoquin-o-line-3-carboxylic acid) analogues linked either through diaminoalkanes of variable length (66-68) or by symmetric or asymmetric 3,6diaminoalkyl-5-methyl-2(1H)-pyrazinone derivatives (59-65). Salvadori et al. [14] first reported that H-Dmt-Tic-OH had not only δ high affinity (K i δ = 0.022 nM) but also extraordinary selectivity for the δ-opioid receptor   [20] a IC 50 value is the concentration required to 50% inhibition of the electrically induced contraction in a muscle. b pA 2 is the negative log of the molar concentration required to double the agonist IC 50 value in order to achieve the original response. c Not tested. d Not determined.
In the series of dimeric H-Dmt-Tic-OH, ligands (59-70) listed in Table 4 were devoid of δ-opioid receptor mediated agonism; all the compounds were exceptionally potent δantagonists with pA 2 values ranging from 10.42 to 11.28, which represent orders of magnitude greater than that of both naltrindole (pA 2 = 9.20) and H-Dmt-Tic-OH (pA 2 = 8.48). In contrast to their μ-opioid receptor affinities (Table 2), the compounds exhibited very weak to nonexistent μ-agonism, especially, 69 and 70, which exhibited pure and potent δand μ-antagonism in the same molecule. In fact, the μ-opioid receptor antagonism of 69 and 70 exceeds that of other known peptidic [33] and nonpeptidic [34] antagonists.
The extraordinary dual δ/μ-antagonism of 69 and 70 qualifies these compounds as potential pharmacological tools for application in the clinical and therapeutic treatment of drug addiction and alcohol dependency. Considering that the bis-Dmt analogues containing alkylpyrazinone are orally active and pass through the blood-brain barrier [19,31,32], we would anticipate that 69 and 70 might show similar properties or may be even more potent due to their increased hydrophobicity [35].

Conclusion
Based on the structure of endomorphins (H-Tyr-Pro-Trp/Phe-Phe-NH 2 ), which exhibited very high selectivity International Journal of Medicinal Chemistry 9 toward μ-opioid receptors, we developed various analogues and examined their activities by alterations of a specific residue. From the studies on the stereoisomers of EM-2, [D-Pro 2 ]EM-2 (12) exhibited more potent and prolonged analgesia [7] although it exhibited low μ-affinity [5], indicating an enhanced bioactivity due to a presumed resistance to enzymatic degradation by dipeptidyl peptidase IV [8]. Substitution of Tyr 1 by Dmt yielded [Dmt 1 ]EM-1 (19) and [Dmt 1 ]EM-2 (20): the former, containing Trp 3 , had mixed μ-agonism/δ-antagonism properties, and the latter, with Phe 3 , exhibited dual μ/δ-agonism. The differences between bulkiness of Trp and Phe defined their biofunctional properties, suggesting the existence of fine differences in the stereo geometry of the ligand-binding site between μand δopioid receptors. These data provided us with methodology to design ligands with agonism or antagonism towards their respective receptors. Thus, we could develop various compounds with dual μ-/δ-agonism or mixed μ-agonism/δantagonism in the same molecule.