Directed Screening Approach
With the renewed interest in potential clinical applications of GH, Merck researchers in 1988 became interested in discovering an orally active nonpeptidyl mimic/peptidomimetic of GHRP-6. Extensive structure-activity relationships for GHRP-6 had already been published (5,8,9). Aromatic residues were favored at positions 2, 4, and 5 and a basic amino terminus was important for GH releasing activity. In addition, preliminary evidence at Merck suggested that the GHRP-6 receptor (hereafter referred to as the GHS receptor) may be G-protein linked. Based on this information, compounds from the Merck Sample Collection were selected for screening in a GH releasing rat pituitary cell culture assay (26). Data from this assay are presented in this chapter as EC50's - the dose required for half maximal GH release. From this effort, benzolactam 1 (10,27) was discovered and shown to release GH in a dose-dependent and specific manner with an EC50 = 3 ^M. Notwithstanding its modest potency in this assay (cf. GHRP-6, EC50 = 10 nM), it was a remarkable achievement considering the rarity in 1988 of nonpeptide mimics of peptide agonists. The carboxylic acid moiety in 1 was initially replaced by a tetrazole, a well established carboxylic acid bioisostere in many angiotensin II antagonists, to give the more potent racemic analog 2 (EC50 = 120 nM). Resolution of racemic 2 identified the R-enantiomer 3 (L-692,429) as the biologically active isomer (EC50 = 60 nM) (Fig. 2.)
Benzolactam 3 exhibited little or no activity at 10 ^M in over 50 other receptor binding assays, except for modest activity (IC50 = 6 ^M) as an angiotensin II antagonist. Molecu-
Hexarelin His-D-2-MeTrp-Ala-Trp-Z)-Phe-Lys-NH2 Fig. 1. Selected growth hormone-releasing peptides (GHRPs).
Fig. 2. Benzolactam growth hormone secretagogue lead structures.
lar modeling of 3 and GHRP-6 places the benzolactam ring and its C-3 chiral center onto the D-Trp residue and its a-carbon in GHRP-6, respectively. The basic amine side-chain in 3 occupies the same region as the N-terminal amino group in the hexapeptide (27). DeVita has recently published cyclic analogs of 3 that exhibit potent GH releasing activity and thus, lends support for a bent conformation for GHRP-6 and the close proximity of the side-chain amine and the biphenyltetrazole in the bioactive conformation of 3 (28). Mechanistically, 3 is identical to GHRP-6 in vitro. Rat pituitary cells maximally stimulated by 3 are unaffected when treated with GHRP-6 (and vice versa) but remain responsive to GHRH treatment. The hexapeptide antagonist His-D-Trp-D-Lys-Trp-D-Phe-Lys-NH2 blocks the GH releasing properties of GHRP-6 and 3. Subsequently, 3 was shown to have a Kj = 63 nM compared to 6 nM for GHRP-6 in a rat pituitary membrane receptor binding assay (29).
Clinical Evaluation of L-692,429
L-692,429 (3) was shown to release endogenous GH in rats, pigs, sheep, dogs, and rhesus monkeys when administered intravenously. In dogs the release of GH was shown to be dose-dependent with a minimum effective dose of 0.1 mg/kg (30). L692,429 had little effect on other hormones except for slight elevation in Cortisol. Unfortunately, 3 showed poor oral efficacy in dogs (>30 mg/kg) owing to poor oral bioavailability (2%) (31).
Even though excellent clinical efficacy with GHRP-6 had been demonstrated, L-692,429 was tested intravenously in humans in order to validate our peptidomimetic approach to GH release in humans. In healthy young males L692,429 (t1/2 = 3.8 h) was found to release GH in a dose-dependent fashion with a minimum effective dose of 0.2 mg/kg (32). As observed with all the GHRPs, there were small transient increases in cortisol and prolactin after L-692,429 administration. No significant changes in other pituitary hormones or changes in insulin-like growth factor (IGF)-1, glucose, or insulin levels were observed. L-692,429 was well tolerated with only a transient flushing or warm sensation being reported. In healthy elderly (71± 5 yr) subjects L-692,429 has been reported to release GH, although the response is somewhat less than in healthy young men (33). L-692,429 has also been shown to partially reverse glucocorticoid suppression of GH secretion and may therefore be useful in reversing the catabolic effects of prednisolone (34).
The validation of L-692,429 as a peptidomimetic of GHRP-6 in humans and its excellent safety profile prompted a major program at Merck to discover a more potent analog of L-692,429 with good oral bioavailability for development as an oral GH secretagogue. To address these issues of potency and oral bioavailability, an extensive investigation of the structure-activity relationships for 3 was undertaken. This effort has been reviewed recently (35-37) and only the key results will be discussed herein. The basic amino group in 3, as with the GHRPs, is critical for GH releasing activity. Modifications of this amino group afforded the 2(R)-hydroxypropyl analog 4 (L-692,585) that was 20-fold more potent than 3 in releasing GH in the rat pituitary cell assay (EC50 = 3 nM) (38,39). In the rat GHS receptor binding assay (29) 4 exhibited a Ki = 0.8 nM, which is 60-fold more potent than 3 (Ki = 63 nM) (Fig. 2). This analog represented a benchmark since it was the first peptidomimetic that was more potent than the hexapeptide GHRP-6 (EC50 = 10 nM) in the rat pituitary cell assay. In dogs it was shown to be active at intravenous doses as low as 5 ^g/kg: 20-fold more potent than benzolactam 3 and twofold more potent than GHRP-6 (40). Unfortunately, the bioavailability of 4 in rats and dogs was not improved over 3.
The zwitterionic character of both 3 and 4 most likely contributes to its poor absorption in animals. Since the basic amine is critical for GH releasing activity, much of the early medicinal chemistry on the benzolactam lead focused on removing the negatively charged tetrazole. Because the GHRPs did not require a negatively charged group for potent GH releasing activity, functionalities capable of forming hydrogen bonds with the GHS receptor were investigated as replacements for the tetrazole in the benzolactam lead (Fig. 3). Neutral heterocycles (e.g., triazole analog 5), carboxamides (e.g., 6), and ureas (e.g., 7) were all found to be excellent neutral surrogates for the negatively charged
Fig. 3. Neutral 2'-biphenyl tetrazole replacements.
Fig. 3. Neutral 2'-biphenyl tetrazole replacements.
tetrazole, thus confirming the hydrogen bonding role for the 2'-substituent in this lead (37,41,42). Combined with the 2(R)-hydroxy side-chain, these neutral surrogates afforded very potent analogs (e.g. 8 and 9) as expected (Fig. 4). However, in spite of these profound structural and physico-chemical changes to these molecules, an improvement in oral bioavailability was not forthcoming. This seemingly unsolvable problem with the benzolactam lead prompted Merck researchers to continue screening for new GH secre-tagogue structures.
Was this article helpful?