Discovery of (R)-7-cyano-2,3,4, 5-tetrahydro-1-(1H-imidazol-4-ylmethyl)-3- (phenylmethyl)-4-(2-thienylsulfonyl)-1H-1,4-benzodiazepine (BMS-214662), a farnesyltransferase inhibitor with potent preclinical antitumor activity.

Continuing structure-activity studies were performed on the 2,3,4, 5-tetrahydro-1-(imidazol-4-ylalkyl)-1,4-benzodiazepine farnesyltransferase (FT) inhibitors. These studies demonstrated that a 3(R)-phenylmethyl group, a hydrophilic 7-cyano group, and a 4-sulfonyl group bearing a variety of substituents provide low-nanomolar FT inhibitors with cellular activity at concentrations below 100 nM. Maximal in vivo activity in the mutated K-Ras bearing HCT-116 human colon tumor model was achieved with analogues carrying hydrophobic side chains such as propyl, phenyl, or thienyl attached to the N-4 sulfonyl group. Several such compounds achieved curative efficacy when given orally in this model. On the basis of its excellent preclinical antitumor activity and promising pharmacokinetics, compound 20 (BMS-214662, (R)-7-cyano-2,3,4, 5-tetrahydro-1-(1H-imidazol-4-ylmethyl)-3-(phenylmethyl)-4-(2-thie nyl sulfonyl)-1H-1,4-benzodiazepine) has been advanced into human clinical trials.

Discovery and structure-activity relationships of imidazole-containing tetrahydrobenzodiazepine inhibitors of farnesyltransferase.

2,3,4,5-Tetrahydro-1-(imidazol-4-ylalkyl)-1,4-benzodiazepines were found to be potent inhibitors of farnesyltransferase (FT). A hydrophobic substituent at the 4-position of the benzodiazepine, linked via a hydrogen bond acceptor, was important to enzyme inhibitory activity. An aryl ring at position 7 or a hydrophobic group linked to the 8-position through an amide, carbamate, or urea linkage was also important for potent inhibition. 2,3,4, 5-Tetrahydro-1-(1H-imidazol-4-ylmethyl)-7-(4-pyridinyl)-4-[2-(t rifluo romethoxy)benzoyl]-1H-1,4-benzodiazepine (36), with an FT IC(50) value of 24 nM, produced 85% phenotypic reversion of Ras transformed NIH 3T3 cells at 1.25 microM and had an EC(50) of 160 nM for inhibition of anchorage-independent growth in soft agar of H-Ras transformed Rat-1 cells. Selected analogues demonstrated ip antitumor activity against an ip Rat-1 tumor in mice.

Potent, cell active, non-thiol tetrapeptide inhibitors of farnesyltransferase.

All previously reported CAAX-based farnesyltransferase inhibitors contain a thiol functionality. We report that attachment of the 4-imidazolyl group, via 1-, 2-, or 3-carbon alkyl or alkanoyl spacers, to Val-Tic-Met or tLeu-Tic-Gln provides potent FT inhibitors. (R*)-N-[[1,2,3,4-Tetrahydro-2-[N-[2-(1H-imidazol-4-yl)ethyl] -L-valyl]-3-isoquinolinyl]carbonyl]-L-methionine ([imidazol- 4-yl-ethyl]-Val-Tic-Met), with FT IC50 = 0.79 nM, displayed potent cell activity in the absence of prodrug formation (SAG EC50 = 3.8 muM).

Development of highly potent inhibitors of Ras farnesyltransferase possessing cellular and in vivo activity.

Analogs of CVFM (a known nonsubstrate farnesyltransferase (FT) inhibitor derived from a CA1A2X sequence where C is cysteine, A is an aliphatic residue, and X is any residue) were prepared where phenylalanine was replaced by (Z)-dehydrophenylalanine, 2-aminoindan-2-carboxylate, 1,2,3,4-tetrahydroisoquinoline-3-carboxylate (Tic), and indoline-2-carboxylate. The greatest improvement in FT inhibitory potency was observed for the Tic derivative (IC50 = 1 nM); however, this compound was ineffective in blocking oncogenic Ras-induced transformation of NIH-3T3 fibroblast cells. A compound was prepared in which both the Cys-Val methyleneamine isostere and the Tic replacement were incorporated. This derivative inhibited FT with an IC50 of 0.6 nM and inhibited anchorage-independent growth of stably transformed NIH-3T3 fibroblast cells by 50% at 5 microM. Replacing the A1 side chain of this derivative with a tert-butyl group and replacing the X position with glutamine led to a derivative with an IC50 of 2.8 nM and an EC50 of 0.19 microM, a 26-fold improvement over (S*,R*)-N-[[2-[N-(2-amino-3-mercaptopropyl)-L-valyl]-1,2,3,4- tetrahydro-3-isoquinolinyl]carbonyl]-L-methionine. This derivative, (S*,R*)-N-[[2-[N-(2-amino-3-mercaptopropyl)-L-tert-leucyl]-1,2,3,4 - tetrahydro-3-isoquinolinyl]-carbonyl]-L-glutamine, was evaluated in vivo along with (S*,R*)-N-[[2-[N-(2-amino-3- mercaptopropyl)-L-tert-leucyl]-1,2,3,4-tetrahydro-3- isoquinolinyl]carbonyl]-L-methionine methyl ester for antitumor activity in an athymic mouse model implanted ip with H-ras-transformed rat-1 tumor cells. When administered by injection twice a day at 45 mg/kg for 11 consecutive days, both compounds showed prolonged survival time (T/C = 142-145%), thus demonstrating efficacy against ras oncogene-containing tumors in vivo.

Vav cooperates with Ras to transform rodent fibroblasts but is not a Ras GDP/GTP exchange factor.

Vav is a proto-oncogene specifically expressed in cells of hematopoietic origin. Its gene product contains a series of structural motifs, including SH2 and SH3 domains, suggestive of a role in signal transduction. The Vav protein also possesses a Dbl-homology (DH) domain previously found in regulators of the Ras superfamily of small GTP-binding proteins. Recently, Vav has been reported to be the major Ras GDP/GTP exchange factor (GEF) in hematopoietic cells [Gulbins et al., Science 260, 822 (1993); J. Immunol. 152, 2123 (1994)]. The following observations are inconsistent with such a role: (i) Vav proteins do not exhibit Ras GEF activity in standard GDP/GTP exchange assays; (ii) Cells overexpressing Vav do not have increased levels of GTP-bound Ras proteins; (iii) Overexpression of Vav does not overcome the growth inhibitory activity of RasN17, a mutant that blocks Ras signaling by inhibiting Ras GEFs; (iv) Transformation of NIH3T3 cells by Vav oncoproteins is not inhibited by a farnesyl transferase inhibitor that completely blocks transformation by both Ras and its well characterized GEF, RasCDC25 and (v) The morphology of Vav-transformed NIH3T3 cells is dramatically different from that induced by Ras and RasCDC25. Whereas these observations make it unlikely that Vav functions either as a RasGEF or as an upstream regulatory element of Ras, we have observed that Vav can cooperate with normal Ras proteins to transform NIH3T3 cells. These results suggest that Vav and Ras may mediate signal transduction by distinct, but interactive mitogenic pathways.

Molecular structural requirements for binding and activation of L-alanine taste receptors.

L-Alanine binds to and activates specific taste receptors ofIctalurus punctatus, the channel catfish. In order to determine the structural requirements for receptor binding and activation in this model system, a number of analogues of L-alanine were tested using a neurophysiological assay and a competitive ligand binding assay. These assays measured the ability of analogues to activate taste receptors and to displace L-[(3)H]alanine from L-alanine binding sites. Of those derivatives with modifications of the sidechain, L-serine, glycine,ß-chloro-L-alanine and 1-amino-cyclopropane-1-carboxylic acid were the most potent analogues with IC50s similar to and neural responses slightly decremented from that of L-alanine. Derivatives containing branched sidechains or sidechains of otherwise increased volume were considerably less active. All modifications of theα-carboxylic acid and theα-amine, including amides, esters and various isosteres, led to substantial reduction in the analogues' ability to displace L-[(3)H]alanine and, in most cases, very weak stimulatory capability. However, L-lactic acid was a reasonably strong stimulus, but a poor competitor, suggesting that it acts at a different receptor site. Overall, these results indicate the importance of the charged amine and carboxylic acid groups for binding to and activation of the receptor for L-alanine. Moreover, modifications around the chiral center of L-alanine support the hypothesis that receptor binding and activation are separate processes in this model taste system.

Structure/activity relationships in the L-alanine taste receptor system of the channel catfish, Ictalurus punctatus.

In order to understand the molecular determinants of amino acid taste receptor binding and activation, structure/activity studies were performed using analogs of L-alanine in a competitive ligand binding assay and a taste neurophysiological preparation. The presence of both the amine and carboxylic acid in a charged and unhindered form is a primary requisite for a ligand to both bind and activate L-alanine receptors. Although a number of carboxylic acid derivatives are moderately good stimuli, their neural activity derives from action at receptors different from L-alanine receptors. Of the molecular parameters examined, chirality and molecular volume of the side chain are the most important factors in determining the binding and stimulatory efficacy of L-amino acid taste stimuli. Electronegativity of the side chain did not correlate with receptor site binding. Heterologous ligand-induced enhancement of the binding of L-[3H]alanine by a purified taste membrane preparation is described. Neurophysiological experiments support the hypothesis that this phenomenon may be a basis of peripheral sensory interactions.

Synthesis and beta-adrenergic antagonist activity of stereoisomeric practolol and propranolol derivatives.

A series of stereoisomeric practolol and propranolol derivatives has been synthesized in which the N-isopropyl group of the drug was replaced by an asymmetric heptanoic acid terminated by a substituted p-toluidide or p-(trifluoromethyl)anilide. The asymmetric epoxide, 3-(p-acetamidophenoxy)-1,2-epoxypropane, was allowed to react with a preformed enantiomeric 6-aminoheptanoic acid amide to yield the stereoisomeric practolol congener derivatives. An asymmetric drug precursor epoxide was prepared from p-acetamidophenol and enantiomeric 3-(tosyloxy)-1,2-epoxypropane 6-aminoheptanoic acid amides were allowed to react with one of the enantiomers of 3-(1-naphthyloxy)-1,2-epoxypropane. This drug precursor epoxide was prepared either by combining 1-naphthol with enantiomeric 3-(tosyloxy)-1,2-epoxypropane (the Sharpless epoxide) or by combining 1-naphthol with enantiomeric 3-(tosyloxy)-1,2-propanediol followed by epoxidation. Pharmacological studies carried out for the practolol derivatives demonstrated a significant dependence of enhanced potency and tissue/subreceptor specificity on both the configuration of the drug asymmetric carbon and the configuration of the spacer asymmetric carbon. The compounds containing the S configuration at the drug asymmetric center and the R configuration at the spacer asymmetric carbon exhibited an increase in potency over the other stereoisomeric congener derivatives and the progenitor drug. For the propranolol congener derivatives, a large decrease in potency was observed for all of the stereoisomers over the progenitor drug. The propranolol stereoisomers containing the S configuration at the drug asymmetric center were more active than those containing the R configuration at that center.