May 15, 2005 (Vol. 25, No. 10)
Susan Aldridge, Ph.D.
Early Decisions Can Strongly Influence Eventual Success or Failure
Innovative approaches to medicinal chemistry in a number of important therapeutic areas were under discussion at Scientific Update’s conference “MedChem Europe: Molecules that Matter” held in Berlin recently. Smarter medicinal chemistry is key to getting new drugs on the market, and there was no shortage of inspiring case studies and examples of putting new chemistries into practice.
It is important for a medicinal chemist to know whether a compound is worth synthesizing at allthat is, whether it has the properties of being “druglike” or “leadlike.”
“The factors that lead to failure in drug discovery are equally as important as those leading to success,” said Gilbert Rishton, Ph.D., of California State University, adding that it is the decisions made in early discovery on target class and lead molecule selection that are the most influential in determining eventual success or failure.
In his opinion, it is better to focus upon a target class, such as GPCRs or proteases, rather than a therapeutic area like cancer or CNS. When it comes to therapeutic modalities, antibodies and small molecules are more likely to be successful than relatively untried approaches such as gene therapy, stem cell therapy, and antisense therapy. Protein therapies can be valuable, but they are not really drugs and most have already been exploited.
When it comes to a small molecule, it must be selected for its lead- and drug-likeness. Unfortunately, screening is full of “false positive” hits that do not fulfill these criteria. Such compounds form a covalent bond with the target in the assay, which looks like a hit, but is not.
These compounds, which contain specific functional groups such as acyl halides, alkyl halides, and others, should be removed from collections, according to Dr. Rishton, because they have already led too many discovery programs awry.
What medicinal chemists should look for in terms of lead- and drug-likeness is a noncovalent, high-affinity, chemically-stable and reversibly binding ligand which can be taken forward into optimization and development.
Certain physico-chemical properties, such as molecular weight less than 500, are linked with drug-likeness, and there are also specific properties linked with lead-likenessfor instance, lack of chemically reactive functional groups.
Some of the smaller companies, such as Sepracor, NPS Pharmaceuticals, Neurogen, Neurocrine, Arena and Athena/Elan (in small molecules), and Abgenix and Medarex (in therapeutic antibodies) have already taken this message on board. “The big pharma companies must look at how these organizations have been successful,” said Dr. Rishton.
The Field Approach
Meanwhile, there is a new and better way of defining chemical diversity: in terms of electron fields, according to Tim Cheeseright, D.Phil., director of projects, Cresset BioMolecular Discovery (Letchworth, U.K.). After all, protein targets do not interact with chemical structures; they interact with electrons.
The field approach involves describing the target-drug binding environment more in terms of what the target actually “sees.” If two diverse structures generate the same field pattern, they should show the same biological activity.
The field method has been applied to the study of thrombin inhibitors and shows that chemically diverse actives look similar in terms of their electron fields. The aim now is to apply the method in virtual screening.
In collaboration with the James Black Foundation, a validation was carried out with the CCK-2 receptor (a GPCR involved in digestion and other physiological functions) as a target using two known active molecules to create the field search patterns.
The search across 600,000 commercially available compounds produced 27 candidates with high potency, 24 of which had no structural similarity with known actives against CCK-2.
The Silicon Switch
Meanwhile, Paradigm Therapeutics (Cambridge, U.K.) is exploiting the use of silicon as an isostere for carbon in drug design in cases where the presence of silicon could improve the molecule’s properties. Paradigm is the only company to commercialize this approach, known as the Silicon Switch, and has several programs moving toward lead optimization.
The Silicon Switch involves replacing one or two carbons with silicon in a known drug which, typically, would be either a failed development candidate or otherwise have suboptimal properties.
Organosilicon chemistry is a versatile tool for the medicinal chemist. Paradigm has already identified a number of Silicon Switch opportunities in the literature, many involving important targets such as GPCRs, ion channels, and kinases. Substitution of silicon can have an impact on the potency, selectivity, metabolism, and half-life of a compound.
In proof of concept studies, done in collaboration with Reinhold Tacke, Ph.D., a pioneer in organosilicon chemistry at the University of Wrzburg, the Silicon Switch versions of haloperidol, an anti-psychotic, niguldipine, a calcium-channel blocker, and the retinoid bexarotene, used in treating T-cell lymphoma, were all found to have improved pharmacological properties.
Paradigm is using silicon medicinal chemistry in cases where the carbon counterpart is hard to synthesize or does not exist and is focusing upon privileged structureschemical scaffolds common to proven drugs in a number of categories. “By using silicon you can get druglike qualities very early on which allows for a rapid entry into lead optimization,” said Graham Showell, director of chemistry at Paradigm.
Silicon chemistry is robust and reactions are often easier to achieve than the carbon-based versions. This has led Paradigm to the synthesis of novel protease inhibitors, and they are also working on other programs involving validated GPCR, nuclear hormone, and kinase targets.
Faster Med Chem
Speeding up the medicinal chemistry stage can have a significant impact on overall discovery and development times. In recent years, there has been a dramatic increase in the number of reactions and publications involving microwave synthesis, thanks to its many advantagesnot least of which is speed.
Microwave reaction times are shorter, and some difficult reactions are made possible by the microwave approach. Typically, the microwave version of a reaction uses less solvent, reagents, and catalyst than the conventional version, which is compatible with the trend toward “green” chemistry.
“Microwave synthesis speeds up your existing medicinal chemistry processes,” said Farah Mavandadi, Ph.D., product manager with the discovery chemistry group at Biotage (Uppsala, Sweden). “It allows you to make decisions more rapidly. If a compound is going to fail, it will fail faster.”
Comparison of synthesis times for some well-known drugs demonstrate the clear advantage of microwave. For instance, the conventional synthesis time of acyclovir (Zovirax) is 40 hours, compared to just under half an hour with microwave. Corresponding times for trimethoprim (Bactrim) are 22 hours and 15 minutes.
Recently Biotage has been developing equipment to allow scale-up of microwave reactions to the 10 g1 kg range. The company is also focused upon easing the purification bottleneck, with a combination of solid-bonded reagents and flash chromatography purification. The total compound production time of many important medicinal chemistry reactions has been shown to be significantly shortened by this approach.
Another significant new approach in medicinal chemistry involves the synthesis and development of dendrimers, which are highly defined nanoscale structures. In other words, dendrimers are on the size scale of biological molecules rather than small molecules.
“The use of dendrimers is the solution for making synthetic molecules in nanotech space,” said Tom McCarthy, Ph.D., vp, drug development of Starpharma (Melbourne, Australia).
Dendrimers are constructed generation by generation by a series of steps which increase the number of links and branching molecules outside a central core.
The number of generations determines the eventual dimension of the dendrimers and the nature of the molecules added to its surface determines its pharmacological properties. The central core often includes a lysine residue, as this has two functional amino groups which can be used to extend the branches.
A small molecule drug makes only one contact with its receptor. For a dendrimer drug, many such contacts are possible, because there are many functional groups on its surface. In other words, dendrimers are a platform for polyvalent interactions which may potentially enhance activity.
Starpharma has been developing dendrimer polyanions as HIV antivirals. These are being applied as prevention with the aim of reducing vaginal HIV transmission. The development candidate SPL7013 has been formulated as a gel and tested for efficacy in non-human primate studies. Phase I trials showed the drug to be safe and well-tolerated, and expanded clinical trials are now underway at various international sites.
“Efficacy will look a lot like a vaccine trial,” said Dr. McCarthy. “We feel with this sort of product we can give women control of their exposure to HIV.”
Of course, there is still a strong need for HIV/AIDS treatments as well as preventive approaches. Bart De Corte, principal scientist, drug discovery at Johnson & Johnson Pharmaceutical R&D (Spring House, PA), described 15 years of medicinal chemistry research into non-nucleoside reverse transcriptase inhibitors (NNRTIs). This has led the company to TMC125 (Etravine), currently in Phase II in HIV/AIDS patients.
The first anti-HIV NNRTIs were in the TIBO chemical structure class, followed by a number of other small molecule structural types, all of which could bind to the RT enzyme in HIV.
“We began with wild type HIV, but then single and double mutants came along and these had to be tackled,” explained Dr. De Corte. The DATAs chemical family were potent inhibitors of wild type HIV-1 and single mutants, and the DAPYs family worked on both single and double mutants.
These research efforts culminated in the compound TMC 125, which compares very favorably with other currently marketed NNRTIs in its activity against HIV-1 and mutant strains. Trials in both treatment-nave and NNRTI treatment-resistant patients have shown a dramatic decrease in viral load.
Meanwhile, J&J has also been looking at TMC278 (Rilpivirine) which is proving to be efficacious as monotherapy given over seven days. “If anything, TMC278 is better than TMC 125,” said Dr. De Corte. “We are really thrilled about this compound.”
Hartmut Rehwinkel, Ph.D., principal scientist at Schering (Berlin), described the development of selective glucocorticoid receptor agonists (SEGRAs). There is a clear need for alternatives to conventional glucocorticoids, also known as steroids, which act by two different mechanismstransactivation and transrepression. The latter is thought to favor anti-inflammatory effects over side effects.
“We have identified compounds that show sufficient dissociation between anti-inflammatory and side effect activity,” said Dr. Rehwinkel. For instance, compound A does not increase blood glucose levels in subcutaneous application in mice compared to prednisolone. It also induces less skin atrophy than standard glucocorticoids and has fewer systemic side effects.
Jrg Holenz, director, medicinal chemistry, Esteve Laboratories (Barcelona, Spain), spoke about new and selective ligands that bind to 5-HT6, a recently cloned member of the serotonin receptor family which is a GPCR expressed exclusively in the brain which may have a role in several CNS conditions.
Esteve, a recent entrant into this field, has been looking at a pharmacophore framework model to discover leads. This involves aligning molecules from the literature to discover a common core which can be built upon. “This gave us an idea of a molecule which had all the necessary features,” said Dr Holenz.
The model was then applied to the discovery of various series of novel potential ligands. Application of structure affinity relationships then led to a number of compounds with a high degree of selectivity for the 5-HT6 receptor. One of these has already shown good results in cognitive enhancement and combatting obesity in animal models.
A new approach to treatment of type 2 diabetes was described by Max Dang, Ph.D., associate director of medicinal chemistry, Metabasis Therapeutics (San Diego), involving the development of novel phosphonates that can inhibit the enzyme Fructose 1,6-bisphosphatase (FBPase).
“More than 50 percent of diabetics do not meet the American Diabetic Association’s therapeutic targets, even given existing drugs,” commented Dr. Dang.
Hepatic glucose production (HGP) involves two pathwaysglycogenolysis and gluconeogenesis. FBPase is involved in the latter and is known to be upregulated in diabetes. The enzyme therefore represents a potential new drug target, and one that should be safe, given that people with a genetic deficiency of FBPase are normal.
The normal substrate of the enzyme is AMP so the company wanted to design novel AMP mimetics, starting off with novel benzimidazole phosphonates, then moving to thiazole phosphonates, which proved orally bioavailable as prodrugs.
The company is now working with novel bisamidate prodrugs of organophosphonates, which have a good to high oral bioavailability and are proving efficacious in animal studies.