Hopes Ride on Drug Candidates

Every spring marks the advent not only of new flowers and new leaves on trees, but also of newly revealed drug candidates. An annual symposium at which drug company researchers disclose potential new medicines for the first time has become a tradition at spring American Chemical Society national meetings, and this year was no exception.

The 2005 symposium on first-time disclosure of clinical candidates was co-organized by Balu N. Balasubramanian, director of drug discovery at Bristol-Myers Squibb's (BMS's) Pharmaceutical Research Institute, Wallingford, Conn., and Joel C. Barrish, vice president of immunology chemistry and chemical synthesis at BMS in Princeton, N.J. The symposium was sponsored by Biotage, Uppsala, Sweden.

"Each symposium paints a clear picture of the standards set by big pharma and the biotech industry to advance their clinical candidates," Balasubramanian said after the session.

"As ever, this was one of the best attended sessions" among symposia organized by the Division of Medicinal Chemistry, Barrish added.

Delivering the address for the Alfred Burger Award in Medicinal Chemistry, William Greenlee, vice president of cardiovascular and central nervous system chemical research at Schering-Plough Research Institute, Kenilworth, N.J., got the pill rolling at the symposium by reporting on the discovery of the orally available antithrombotic agent SCH-530348.

SCH-530348 is an antagonist (blocking agent) at the cellular receptor for the clotting enzyme thrombin. It's a potential treatment for acute coronary syndrome--chest pain caused by coronary artery disease, Greenlee said.

Researchers have been working for years to develop an antithrombotic drug that improves the risk-benefit profile of medications currently on the market, such as the anticoagulant warfarin. "The efficacy of currently available antithrombotic agents is balanced by dose-limiting side effects, mainly associated with bleeding," Greenlee said. A thrombin receptor chemistry team led by Schering-Plough Distinguished Fellow Samuel Chackalamannil identified SCH-530348 as a potential improvement over existing medications, many of which also lack oral activity.

Thrombin plays a key role in blood coagulation by catalyzing the formation of the insoluble protein fibrin from a precursor and by binding to the thrombin receptor on platelets, which activates platelets for aggregation. Although this clotting mechanism is essential to prevent excessive blood loss after injury, it can also cause clots to form in arteries, leading to heart attack or stroke. SCH-530348 acts as an antagonist at the thrombin receptor, reducing thrombin-induced platelet activation, but it doesn't interfere with thrombin's essential fibrin-generating mechanism, Greenlee said.

ACCORDING TO Greenlee, Schering-Plough's work on the thrombin receptor antagonist goes back to an earlier study on the natural product himbacine, an early lead in an Alzheimer's disease program. Researchers at the company achieved a total synthesis of himbacine in the course of that study, and one of many analogs prepared at the time became the lead compound for the thrombin receptor antagonist program that led to SCH-530348.

The path from lead compound to SCH-530348 took eight years to traverse and involved the synthesis of more than 2,000 additional analogs, including two early-development candidates that were discontinued, Greenlee said. Nevertheless, "SCH-530348 was worth the wait."

SCH-530348 has a complex structure with seven stereogenic centers. Fortunately, Chackalamannil and coworkers developed a highly efficient synthesis, the key step of which is a highly diastereoselective intramolecular Diels-Alder reaction that generates the drug's tricyclic skeleton.

"SCH-530348 is uniquely selective toward the thrombin receptor and has excellent oral bioavailability," Greenlee said. It's currently in early clinical development.

At Predix Pharmaceuticals, Woburn, Mass., a group led by former senior vice president of drug discovery Dale S. Dhanoa (now president and CEO of Invent Pharmaceuticals, San Diego) discovered a potential new treatment for anxiety, depression, and attention deficit hyperactivity disorder (ADHD), conditions for which better medications are needed. Dhanoa noted that current drugs in this class include benzodiazepines, which can cause dependency and cognitive impairment; selective serotonin reuptake inhibitors, which are associated with sexual dysfunction, appetite loss, and sleep disorders; and buspirone, drawbacks of which include slow onset, hypertension, and dizziness.

According to Dhanoa, the Predix group's goal was to find a potent, selective, and orally active agonist of 5-HT1A (5-hydroxytryptamine receptor 1A), a receptor that binds the neurotransmitter serotonin (5-hydroxytryptamine) and is known to be associated with anxiety and depression. An agonist is a drug that increases the physiological activity of a receptor. The researchers set out to discover a novel drug with good oral bioavailability and receptor specificity but with minimal or no tendency to inhibit hERG, a potassium ion channel expressed in the heart and brain, because potent blockage of hERG activity can cause ventricular fibrillation and sudden death.

Dhanoa and coworkers identified a starting compound by computational screening and optimized it through medicinal and computational chemistry, yielding a lead compound. Modifications to the lead compound further improved its 5-HT1A activity and selectivity but also increased its undesirable hERG inhibitory activity. The latter was eventually reduced by replacing a p-toluenesulfonamide group on the lead compound with cyclohexylmethyl sulfonamide, eliminating the ability of the compound to interact with the hERG binding pocket. This modification yielded a clinical candidate, PRX-00023.

DHANOA ALSO replaced the terminal cyclohexylmethyl group in PRX-00023 with an isobutyl group to yield a backup candidate, PRX-00030. The backup drug has shown remarkably increased oral bioavailabilty in animal studies. PRX-00023 is currently in Phase II clinical trials for the treatment of anxiety, depression, and ADHD, and PRX-00030 may be tested later for those and perhaps other therapeutic indications.

Also reporting on a potential anxiety medication, Leslie J. Street of Merck Sharp & Dohme's Neuroscience Research Centre, in Terlings Park, England, disclosed his group's discovery of imidazotriazines for the treatment of anxiety. The objective was to develop an anxiolytic (antianxiety) drug similar to the benzodiazepine diazepam (Valium) but with fewer side effects, such as drowsiness.

Diazepam is an agonist at the GABA_A receptor, which is a ligand-gated ion channel on neuronal cell membranes. Diazepam makes the ion channel open more frequently, increasing chloride ion flow into neurons. This, in turn, decreases neuronal activity, and less neuronal activity translates into reduced anxiety.

According to Street, the problem with diazepam and other benzodiazepines is that they act nonselectively at four of the main GABA_A receptor subtypes-- ₁, ₂, ₃, and ₅. The GABA_A receptor ₁ subtype has been shown to be associated mainly with sedation. Street and coworkers wanted to develop an anxiety medication that wouldn't cause patients to fall asleep when they took it. They hypothesized that this goal could be achieved by designing compounds that acted at the GABA_A receptor∝ ₂ and ₃ subtypes but not at the ₁ subtype.

"We initially did in vitro screening to identify compounds with functional selectivity for GABA_A receptor ₂ and ₃ subtypes," Street said. A series of imidazopyridines and imidazopyrimidines so identified were screened in animal models. The imidazopyridines showed good receptor-subtype selectivity but poor pharmacokinetics (absorption, distribution, tissue localization, duration of action, and excretion properties). The imidazopyrimidines, on the other hand, showed good selectivity and good oral bioavailability but short longevity of action (half-life).

TO FIND something better, Street and coworkers turned to yet another class of compounds they had identified by in vitro screening: the imidazotriazines. One imidazotriazine they identified became their lead compound. It had good oral bioavailability, good receptor-subtype selectivity, and a long half-life. It acted primarily at the GABA_A receptor ₂ and ₃ subtypes, showed little activity at the ₁ subtype, and had greatly reduced (but nonzero) activity at the ₅ subtype.

Behavioral data from animal studies suggested that the imidotriazine might be a nonsedating anxiolytic agent, and it was advanced into Phase I human clinical trials. "There we found that the relatively long half-life in preclinical species does translate into humans," Street said. "The human data indicated that the half-life is sufficient for once-a-day dosing," a goal of the project, and the data showed that the drug is well-tolerated. Street was not at liberty to say whether Phase II trials of the agent were in progress or planned.

For diabetes therapy, Executive Vice President of R&D Mark D. Erion of Metabasis Therapeutics described a decade-long effort by him and his team that led to MB06322, a potent and selective fructose-1,6-bisphosphatase (FBPase) inhibitor that is designed to control gluconeogenesis in type 2 diabetes. Gluconeogenesis is the biosynthesis of glucose in the body from noncarbohydrate three-carbon starting materials, such as pyruvate and lactate.

"In diabetics, the liver produces too much glucose, leading to high blood glucose levels that in the long term can cause blindness, kidney disease, and cardiovascular disease," Erion said. "This excessive glucose production is a consequence of both the relatively low levels of the hormone insulin in diabetics and the poor response to insulin by their muscle, fat, and liver."

FBPase is a key enzyme in the gluconeogenesis pathway. Inhibiting it would slow down glucose production and thereby lower blood glucose levels in type 2 diabetics.

"FBPase is an enzyme well-known to the pharmaceutical industry," Erion said, but "previous efforts have failed to find potent, specific, and in vivo-active inhibitors." Metabasis focused on mimics of adenosine 5´-monophosphate (AMP), which is an allosteric inhibitor of FBPase. Using structure-based design, Metabasis found AMP mimics capable of binding with high affinity and specificity to FBPase's AMP-binding site.

"Historically, AMP-binding sites have been viewed as difficult drug targets, in large part because of their hydrophilicity and their widespread use by nature for allosteric regulation of enzymes and metabolic pathways," Erion said. His group's design strategy was based on X-ray structures of FBPase-inhibitor complexes obtained by emeritus professor of chemistry and chemical biology William N. Lipscomb and coworkers at Harvard University and on binding-site energetics data obtained with computational techniques developed at Metabasis.

The structural information and energetics data led the group to investigate phosphonic acid-containing compounds as potential FBPase inhibitors. Metabasis chemistry project leader Qun (Max) Dang designed and synthesized MB05032, a promising lead compound that turned out to be a potent and highly specific inhibitor of human FBPase. Its oral bioavailability, however, was less than 2%.

So the Metabasis group focused on finding phosphonic acid prodrugs that they hoped would provide better oral bioavailability and then convert to MB05032 in the body. The researchers identified a series of orally bioavailable bisamidate prodrugs that are stable in aqueous solution and rapidly convert in vivo to the active agent. This led them to the development candidate, MB06322.

Mechanism-of-action studies in rats showed that, after oral administration, MB06322 inhibited FBPase, gluconeogenesis, and glucose production by the liver. "Glucose levels decreased in rats irrespective of their age and initial insulin levels, suggesting that the approach may ultimately result in glucose lowering in diabetic patients with either early or advanced diabetes," Erion said.

Several years ago, Metabasis and a corporate partner, Sankyo Co., Tokyo, advanced MB06322 (renamed CS-917) into clinical trials. In Phase II trials, they found preliminary evidence of efficacy in patients with type 2 diabetes. A few days after Erion's lecture in San Diego, however, two serious adverse events occurred in a Phase I interaction study of CS-917 with the type 2 diabetes medication metformin, requiring reassessment of the development strategy. "The Phase II program is on hold while the companies determine the implications of these findings," Erion told C&EN.

ANOTHER DRUG disclosed at the session was BMS-599626, a cancer drug that acts by dual inhibition of HER1 and HER2 receptor tyrosine kinases, which have been implicated in a variety of cancers. The drug was developed by Ashvin V. Gavai and coworkers at BMS Pharmaceutical Research Institute, Princeton.

Most drugs used in the treatment of cancer are cytotoxic agents that kill actively dividing cancer and noncancer cells indiscriminately, Gavai said. An emerging class of signal transduction inhibitors, however, attacks cancer by a more targeted mechanism--by modulating the activities of cancer-associated receptors like HER1 and HER2. "This is an attractive approach since the signaling inhibitors are likely to provide a significantly improved side-effect profile," he said.

Receptor tyrosine kinases are receptor proteins that include three components: an extracellular domain that binds specific ligands, such as epidermal growth factor; an intracellular enzymatic (tyrosine kinase) domain that catalyzes phosphate transfer from bound adenosine triphosphate (ATP) to a protein; and a transmembrane domain that links the other two domains. The recent approval of two monoclonal antibody drugs--ImClone Systems and BMS's HER1-selective agent cetuximab (Erbitux) and Genentech's HER2-specific drug trastuzumab (Herceptin)--has helped validate the concept that drugs targeting receptor tyrosine kinases can be effective anticancer agents.

Gavai and coworkers wanted to find a compound that would target HER1 and HER2 simultaneously. "Extensive homology of the catalytic sequences of the two receptors supported the feasibility of designing inhibitors that occupied ATP-binding pockets on both," Gavai said.

A BMS medicinal chemistry group identified a pyrrolotriazine nucleus as a novel template for drugs that might compete effectively with ATP for access to its binding site on the kinases. Modeling and structure-activity relationship studies suggested that the pyrrolotriazine ring bound in the adenine pocket of the ATP-binding site on kinases and interacted with the enzymes in several other important ways. A range of substituents was tried on the pyrrolotriazine to improve inhibitory potency and impart desirable pharmaceutical properties.

The result was BMS-599626. "It was selected for clinical evaluation based on its potent, selective kinase inhibition profile and its excellent antitumor efficacy," Gavai said.

When administered orally in HER1-overexpressing human tumor transplant models, BMS-599626 showed superior efficacy to gefitinib, another receptor tyrosine kinase inhibitor. BMS-599626 showed comparable activity to trastuzumab in mice with HER2-overexpressing tumors. It was also efficacious in inhibiting HER2-overexpressing tumors that do not respond to trastuzumab.

Preclinical data suggest that BMS-599626 is suitable for once- or twice-daily oral dosing. "Phase I trials have been initiated in cancer patients," Gavai said.

In next year's first-time disclosures session, Balasubramanian said, "it would be interesting to do a retrospective analysis on how many of the candidates presented at these symposia over the years are still in the development stage or approved and perhaps a discussion of the lessons learned."