Identification of MK-944a: a second clinical candidate from the hydroxylaminepentanamide isostere series of HIV protease inhibitors.

Recent results from human clinical trials have established the critical role of HIV protease inhibitors in the treatment of acquired immune-deficiency syndrome (AIDS). However, the emergence of viral resistance, demanding treatment protocols, and adverse side effects have exposed the urgent need for a second generation of HIV protease inhibitors. The continued exploration of our hydroxylaminepentanamide (HAPA) transition-state isostere series of HIV protease inhibitors, which initially resulted in the identification of Crixivan (indinavir sulfate, MK-639, L-735,524), has now yielded MK-944a (L-756,423). This compound is potent, is selective, and competitively inhibits HIV-1 PR with a K(i) value of 0.049 nM. It stops the spread of the HIV(IIIb)-infected MT4 lymphoid cells at 25.0-50.0 nM, even in the presence of alpha(1) acid glycoprotein, human serum albumin, normal human serum, or fetal bovine serum. MK-944a has a longer half-life in several animal models (rats, dogs, and monkeys) than indinavir sulfate and is currently in advanced human clinical trials.

Transcription control and neuronal differentiation by agents that activate the LXR nuclear receptor family.

LXR and PPAR receptors belong to the nuclear receptor superfamily of transcriptional activating factors. Using ligand-dependent transcription assays, we found that 5-tetradecyloxy-2-furancarboxylic acid (TOFA) transactivates chimeric receptors composed of the glucocorticoid receptor DNA binding domain and the ligand binding regions of PPARalpha, PPARbeta (NUC-1) and LXRbeta (NER) receptors. In the same assays, ligands for PPARs (oleic acid, WY-14643 and L-631,033) and LXRs (hydroxycholesterols) maintain their respective receptor selectivity. TOFA and hydroxycholesterols also stimulate transcription from a minimal fibrinogen promoter that is under the control of AP-1 or NF-kappaB transcription factor binding sites. In addition to their effects on transcription, these LXRbeta activators induce neuronal differentiation in rat pheochromocytoma cells. TOFA and the natural LXR agonist, 22 (R)-hydroxycholesterol, stimulate neurite outgrowth in 55 and 28% of cells, respectively. No neurite outgrowth was induced by the related 22(S)-hydroxycholesterol, which does not activate the LXR family. These results suggest that the hydroxycholesterol signaling pathway has a complex effect on transcription that mediates the activity of TOFA and hydroxycholesterol on neuronal differentiation in pheochromocytoma cells.

Structural requirements and cell-type specificity for ligand activation of peroxisome proliferator-activated receptors.

The mammalian peroxisome proliferator-activated receptor (PPAR) family consists of three different subtypes, PPARalpha, hNUC1/PPARdelta and PPARgamma. Selective agonists have been identified for PPARalpha and PPARgamma but not for hNUC1, and consequently little is known about the genes that are controlled by this receptor. Using ligand-dependent transcription assays in COS-7 cells, we screened a variety of PPAR activating agents to identify a selective activator of hNUC1. We found that the potent peroxisome proliferator, Wy-14643, and the PPARgamma-selective thiazolidinedione, BRL 49653, were poor activators of hNUC1 (EC50s of > 100 microM). Short chain fatty acids (FAs) appeared more selective for PPARalpha than for hNUC1, whereas the very long chain FA, erucic acid (C22:1) was more selective for hNUC1. Using erucic acid as a probe, we conducted a topological similarity search of the Merck Chemical Collection and identified a fatty acid-like compound, L-631,033 4-(2-acetyl-6-hydroxyundecyl) cinnamic acid, that was a selective activator of hNUC1 (EC50 of 2 microM), but was much less selective for PPARalpha or PPARgamma (EC50s of > 100 microM). Structure-function analysis of PPAR activation by L-631,033 structural analogues showed that receptor selectivity depends on the position of the carboxyl group relative to the phenyl ring on the molecule. Transfection experiments in several cell types: an osteoblastic cell line (MB 1.8), a mouse liver cell line (ML-457), rat aortic smooth muscle cells (RSMCs) and COS-7 cells revealed differences in the activation profile of specific ligands. The most notable differences were observed in RSMCs, where transactivation by L-631,033 and Wy-14643, but not by BRL 49653, was markedly reduced, and in MB 1.8 cells, where oleic acid failed to activate PPARs. These findings identify certain structural features in PPAR-activating agents that modulate PPAR activation, and suggest that as with other nuclear receptors, activation is cell-type specific.

Design and synthesis of nonpeptide peptidomimetic inhibitors of renin.

The desire to replace the amide backbone of renin inhibitors with a new scaffold led us to explore vinylogous amides (enaminones). An initial attempt proved unsuccessful, a result explained after the fact via docking experiments. Based on this lesson, we designed a different vinylogous amide scaffold which incorporated one or more pyrrolinone rings into the backbone. Three of the four compounds gave IC50S in the 0.6 to 18 microM range. These compounds did not inhibit HIV-1 protease. Taken together, the results reported herein provide insights into the role of hydrogen bonding and steric interactions for binding to renin.

A priori prediction of activity for HIV-1 protease inhibitors employing energy minimization in the active site.

We have observed a high correlation between the intermolecular interaction energy (Einter) calculated for HIV-1 protease inhibitor complexes and the observed in vitro enzyme inhibition. A training set of 33 inhibitors containing modifications in the P1' and P2' positions was used to develop a regression equation which relates Einter and pIC50. This correlation was subsequently employed to successfully predict the activity of proposed HIV-1 protease inhibitors in advance of synthesis in a structure-based design program. This included a precursor, 47, to the current phase II clinical candidate, L-735,524 (51). The development of the correlation, its applications, and its limitations are discussed, and the force field (MM2X) and host molecular mechanics program (OPTIMOL) used in this work are described.

Synthesis, antiviral activity, and bioavailability studies of gamma-lactam derived HIV protease inhibitors.

Incorporation of a gamma-lactam in hydroxyethylene isosteres results in modest inhibitors of HIV-1 protease. Additional structural activity studies have produced significantly more potent inhibitors with the introduction of the trisubstituted cyclopentane (see compound 20) as the optimum substituent for the C-terminus. This new amino acid amide surrogate can be readily prepared in large scale from (R)-pulegone. Optimized compounds (36) and (60) are potent antiviral agents and are well absorbed (15-20%) in a dog model after oral administration.

The development of cyclic sulfolanes as novel and high-affinity P2 ligands for HIV-1 protease inhibitors.

Design and synthesis of a novel series of protease inhibitors incorporating conformationally constrained cyclic ligands for the S2-substrate binding site of HIV-1 protease is described. We recently reported urethanes of 3-tetrahydrofuranyl as P2 ligands for HIV-1 protease inhibitors. Subsequently, we have found that the urethane of 3(S)-hydroxysulfolane further increased the in vitro potency of these inhibitors. Furthermore, introduction of a small 2-alkyl group cis to the 3-hydroxyl group of either heterocyclic system further enhanced enzyme affinity. The cis-2-isopropyl group thus far offered optimum enhancement of the inhibitory properties. This led to the discovery of inhibitor 43 (IC50 3.5 nM, CIC95 50 +/- 14 nM) of comparable in vitro antiviral potency to the current clinical candidate 1 (Ro 31-8959) but of reduced molecular weight due to the exclusion of the P3 quinoline ligand. Also, it has been demonstrated that the octahydropyrindene derivative 34 is an effective replacement of the P1' decahydroisoquinoline derivative.

Potent HIV protease inhibitors: the development of tetrahydrofuranylglycines as novel P2-ligands and pyrazine amides as P3-ligands.

A series of protease inhibitors bearing constrained unnatural amino acids at the P2-position and novel heterocycles at the P3-position of compound 1 (Ro 31-8959) were synthesized, and their in vitro enzyme inhibitory and antiviral activities were evaluated. Replacement of P2-asparagine of compound 1 with (2S,3'R)-tetrahydrofuranylglycine resulted in improvement in enzyme inhibitory as well as antiviral potencies (compound 23). Interestingly, incorporation of (2S,3'S)-tetrahydrofuranylglycine at the P2-position proved to be less effective. The resulting compound 24 was 100-fold less potent than the 2S,3R-isomer (compound 23). This stereochemical preference indicated a hydrogen-bonding interaction between the tetrahydrofuranyl oxygen and the residues of the S2-region of the enzyme active site. Furthermore, replacement of P3-quinolinoyl ligand of 1 with various novel heterocycles resulted in potent inhibitors of HIV proteases. Of particular interest, compound 2 with (2S,3'R)-tetrahydrofuranylglycine at P2 and pyrazine derivative at P3 is one of the most potent inhibitors of HIV-1 (IC50 value 0.07 nM) and HIV-2 (IC50 value 0.18 nM) proteases. Another important result in this series is the identification of compound 27 in which the P2-P3-amide carbonyl has been removed. The resulting compound 27 has exhibited improvement in antiviral potency while retaining the enzyme inhibitory potency similar to compound 1.

Synthesis and antiviral activity of a series of HIV-1 protease inhibitors with functionality tethered to the P1 or P1' phenyl substituents: X-ray crystal structure assisted design.

By tethering of a polar hydrophilic group to the P1 or P1' substituent of a Phe-based hydroxyethylene isostere, the antiviral potency of a series of HIV protease inhibitors was improved. The optimum enhancement of anti-HIV activity was observed with the 4-morpholinylethoxy substituent. The substituent effect is consistent with a model derived from inhibitor docked in the crystal structure of the native enzyme. An X-ray crystal structure of the inhibited enzyme determined to 2.25 A verifies the modeling predictions.

Renin inhibitors containing conformationally restricted P1-P1' dipeptide mimetics.

A series of renin inhibitors containing lactam-bridged P1-P1' dipeptide mimetics based on the ACHPA (4(S)-amino-5-cyclohexyl-3(S)-hydroxypentanoic acid) design was studied. The inhibitors were obtained by aldol addition of various lactams with N alpha-Boc-L-cyclohexylalaninal, followed by Boc group removal and acylation with Boc-Phe-His. The aldol diastereomer having the S configuration at the two newly generated stereogenic centers gave optimal enzyme inhibition. Potency was further enhanced in the gamma-lactam ring series by substitution with small hydrophobic groups to mimic the P1' side chain of the renin substrate. Thus, 2(S)-[(Boc-L-phenylalanyl-L-histidyl)amino]-3-cyclohexyl-1(S)-hydroxyl-1 - (1,5,5-trimethyl-2-oxopyrrolidin-3(S)-yl)propane (34) has an IC50 of 1.3 nM in the human plasma renin assay. A variety of substituents on the lactam nitrogen are tolerated and can be used to vary the physical properties of the inhibitor. By using a model of the human renin active site, the conformation of 34 in the enzyme-inhibitor complex is proposed. This modeled conformation is very similar to the solid-state conformation of 2(S)-[(Boc-L-phenylalanyl-L-histidyl)amino]-3-cyclohexyl-1(S)-hydroxyl- 1-(1-methyl-2-oxopyrrolidin-3(S)-yl)propane (36), the structure of which was determined by single-crystal X-ray diffraction analysis. The most potent ACH-PA-lactam renin inhibitors show good selectivity when assayed against other types of aspartic proteinases. By varying the lactam ring substituents, potent and selective inhibitors of cathepsin D and cathepsin E can be obtained.

Characterization of an active single polypeptide form of the human immunodeficiency virus type 1 protease.

The pepsin-like aspartyl proteases consist of a single polypeptide chain with topologically similar amino- and carboxyl-terminal domains, each of which contributes 1 aspartic acid residue to the active site. This structure has been proposed to have evolved by gene duplication and fusion from a dimeric enzyme composed of two identical polypeptide chains, such as the aspartyl protease (PRT) of human immunodeficiency virus type 1 (HIV-1). To determine if a single polypeptide form of the HIV-1 protease would be enzymatically active, two protease coding regions were linked to form a dimeric gene (pFGGP). Expression of this gene in Escherichia coli yielded a protein with the expected molecular mass of 22 kDa. The in vitro kinetic parameters of PRT and FGGP (where FGGP is the single polypeptide form of the HIV-1 protease with 2 glycine residues connecting the two subunits) for three peptide substrates are similar. Construction and analysis of a CheY-GAG-FGGP fusion protein demonstrated that FGGP is capable of precursor processing in vivo. Mutation of one or both of the active site aspartates to either asparagine or glutamate rendered the enzyme inactive, demonstrating that both active site aspartate residues are required for enzymatic activity.

Novel glutamic acid derived cholecystokinin receptor ligands.

Novel aryl amide analogues of glutamic acid dialkylamide have been synthesized to test for a possible structural analogy between glutamic acid and benzodiazepine CCK antagonists such as compounds 2 and 24 (lorglumide and MK-329, respectively). In support of the structural model, certain of these hybrid compounds are more potent in pancreas CCK radioligand binding assays than corresponding lorglumide-type reference compounds. Modifications previously found in the benzodiazepine antagonists to result in brain CCK/gastrin receptor selectivity were also incorporated to produce an aryl urea series of glutamic acid analogues. None of these compounds were brain CCK/gastrin selective; however, one was potent and selective in the pancreas binding assay. The model appears to be most useful in the design of selective ligands for the pancreas type CCK receptor.

Cholecystokinin antagonists. Synthesis and biological evaluation of 3-substituted benzolactams.

A series of 1,3-substituted benzolactams are reported that are potent nonpeptidal antagonists of the peptide hormone cholecystokinin (CCK). Design considerations were based upon the natural product CCK antagonist asperlicin and the potent benzodiazepine antagonist series exemplified by L-364,718 (1). Compound 19, the most potent compound in the benzolactam series, had an IC50 = 3 nM for inhibition of binding of 125I-CCK-8 to CCK receptors in rat pancreatic tissue, and its racemic analogue 8 was found to be orally active in inhibiting CCK-induced gastric emptying in mice, with an ED50 = 2.6 mg/kg po. The effects of ring size, substitution at positions 1 and 3, and stereochemistry at position 3 are discussed. Conformational studies of compound 19 and L-364,718 have delineated similarities that these molecules share in their core conformations and substituent orientations.