"The worst is over now," says Thomas E. D'Ambra, chief executive officer of Albany Molecular Research (AMR), which provides custom chemistry services for drug discovery and development. AMR did well in 2002, but it wasn't easy. Chemistry custom services are estimated to have brought in sales of about $73 million in 2002, up 28% from 2001. That growth was achieved despite AMR's loss of a long-term contract early last year when DuPont Pharmaceuticals was acquired by Bristol-Myers Squibb, D'Ambra says. "We had to hustle just to stay even and then grow on top of that."

D'Ambra believes that the belt-tightening of the pharmaceutical industry--due to expiring patents on blockbusters, disappointing sales of new products, unmet Wall Street expectations, and aborted launches--has peaked and drug companies will not cut R&D any further. He offers Eli Lilly as an example of a company poised to have a great 2003 and beyond. The company lost a major revenue stream when generic versions of its blockbuster drug Prozac became available. But now, several of its New Drug Applications have been approved or are pending approval.

Companies are gearing up for the recovery of the pharmaceutical industry. Reflecting the high demand for novel compounds for drug discovery, several companies based in England are expanding. At Tripos Receptor Research (TRR), construction is under way to more than double laboratory space to 65,000 sq ft by 2004. Key Organics is adding lab space, expanding its compound-handling facility, and planning to hire up to 15 Ph.D. chemists in 2003. Peakdale Molecular will add two labs by April and will be hiring as many as 12 chemists by spring. And newcomer Onyx Scientific hopes to complete a new chemistry unit and recruit up to 12 additional bench chemists by year's end.

Estimates of the total market for this custom synthesis segment vary, but $400 million, or 1% of pharmaceutical R&D spending, "is probably a good ballpark figure," says Gareth Jenkins, business development manager at Peakdale Molecular. In 2001, member companies of the Pharmaceutical Research & Manufacturers of America spent about $30.3 billion in R&D.

GENESIS Molecular modeling aids the search for druglike compounds.
COURTESY OF TRIPOS RECEPTOR RESEARCH

THE MARKET'S RAPID growth means double-digit rises in revenues not only among major players, but also among smaller companies. Jenkins says Peakdale Molecular's sales are growing at least 50% per year. Likewise, Onyx Scientific has grown by leaps and bounds since its inception in June 2001--with sales of about $400,000 in 2001 and about $3.5 million in 2002. The company expects sales to hit $6 million in 2003, CEO Tony Flinn says.

Drug discovery begins with a validated target, usually a protein that regulates the progression of a disease. Genomics and other technologies are yielding a plethora of targets, but most of these are not associated with a known ligand. The goal of discovery research is a compound that will bind to the target and influence disease progression.

?To identify a chemical starting point, drug companies screen the target against vast numbers of structurally diverse compounds. If the structure of the active site is known, the initial screen can be done computationally, or "in silico." The goal of screening is to find "hits." These are compounds that are active in a screen; that is, they bind to the target in a manner that meets specified criteria.

"A hit only gives you an idea of the drug molecule," explains Padmakumar R. Kaimal, director of discovery chemistry at Synthon Chiragenics. A hit rarely has the right combination of activity, selectivity, and bioavailability required of a drug. The structure has to be tweaked for maximum bioactivity and a desirable ADMET (adsorption, distribution, metabolism, excretion, and toxicity) profile. This tweaking produces "leads," which are compounds with properties that suggest their value as potential drug candidates.

A lead is tested in an animal model of the disease to prove that it works and does not have adverse activity such as carcinogenicity, teratogenicity, or mutagenicity, Kaimal explains. The pharmacological properties of the compound are also established during animal testing. Data are obtained regarding the compound's pharmacokinetics--the fate of the compound when ingested, how long it is retained in the body, and what metabolites are formed--and pharmacodynamics--what happens to the body when the compound is ingested, for example, whether tumors are formed.

CHEMICAL VARIATIONS of a lead are usually designed to generate compounds with optimal pharmacological properties. This process is called lead optimization. A compound that is both effective and safe in animal models becomes an investigational new drug (IND), or a drug candidate. In the U.S., an IND application must be filed with the Food & Drug Administration to move a drug candidate into human clinical trials.

Numerous compounds of diverse structures are designed, prepared, and tested to arrive at one candidate. Pharmaceutical companies cannot prepare all the compounds for all their discovery programs. Outsourcing fills the gap. Merck, for example, contracts out medicinal chemistry and synthesis of libraries, reagents, scaffolds, and lead analogs, according to Steven M. Hutchins, director of outsourcing in the research operations department at Merck.

Companies supplying compounds for drug discovery usually are generalists in their chemistry, synthesizing compounds that above all are unique and medicinally relevant. Differentiation comes in various ways.

AMR has expertise in biocatalysis, natural products, and asymmetric synthesis, developed from its own R&D. Developing technology that can be licensed to customers is a plank of the business. For 2002, AMR will receive about $50 million in royalties from Aventis for licensing AMR patents related to the allergy drug Allegra. "That's a great example of the power you can bring to the bottom line from owning technology," D'Ambra says.

Ricerca Biosciences is tapping into high-quality but low-cost synthesis through a partnership with gvk bioSciences, a drug discovery and development company based in Hyderabad, India. "We have access to up to 100 medicinal chemists in India," Ricerca CEO Prabhavathi Fernandes says. "We do the design here and then send the compound over. The arrangement gives us a price advantage."

One of Ricerca's strengths, Fernandes says, is awareness on behalf of the client. For example, Ricerca keeps an eye on a compound's polymorph form, a solid-state property that affects bioavailability. If there is more than one form, Ricerca alerts the client and designs a chemical process that generates the preferred polymorph.

AWARENESS IS also expressed in valuable advice. For example, Fernandes discourages use of ultrapure compounds in very early toxicological studies, because the level of purity may not be attainable in scale-up and the toxicological studies would have to be redone. On the other hand, she advises clients to start process design as soon as a compound starts to look good in toxicology, because the impurities in one route may be different from those in another route. "In most cases, it's not expensive to improve the process. If anything, it is probably cheaper in the long run, as process changes can result in changes in polymorph form as well as in impurities," she says.

TRR says it weds chemical informatics to bench chemistry to deliver compounds of high biological relevance in a short time. The goal is achieved through constant dialogue between molecule designers and bench chemists, who are supported by powerful software.

Compounds of interest are identified by using ChemSpace, developed by TRR's parent, Tripos, a chemical informatics company. Using ChemSpace, designers create a virtual library of druglike compounds that are similar in shape to a lead structure. Chemists prioritize the library members on the basis of medicinal relevance, feasibility of synthesis by high-throughput robotic methods, amenability to standard analytical techniques, availability of reagents, and whether the compound type is already in the market or in TRR's screening collection. The designer takes the information and refines the compound set. The dialogue goes back and forth until a working library is designed and eventually prepared.

"I don't believe molecular modeling can be separated from the wet lab," TRR Director Peter Hecht says. "Informatics-driven chemistry needs to take into account what is possible not only in silico, but also in vitro. Otherwise, you will have this wonderful solution that looks great on your screen but is completely unsynthesizable."

Another key tool at TRR is ChemCore, a chemical-synthesis tracking database. Not only does ChemCore track the availability of reagents, an important factor in early delivery, it also tracks compounds from inception to delivery. By capturing all the information associated with a compound in a database that can be searched by various criteria--structure, substructure, structure similarity, name, generic name, acronym, reactions, products, and so on--ChemCore enhances productivity and gets compounds to customers faster.

Meanwhile, at Key Organics, chemistry is carried out in gram scale in glass vessels one reaction at a time, instead of simultaneously in 96-well plates in milligram scale. Gram-scale synthesis ensures that the compound is in stock whenever the customer needs it, says the company's director, Colin Deane. It also means that every compound comes with a synthesis that is amenable to scale-up, adds Roger Moses, the company's principal chemist.

Key Organics is known for heterocyclic chemistry. That more than 80% of its 34,000-compound screening collection are heterocyclics reflects the great interest in this chemistry, Deane says.

ONE WAY the company comes up with unique and novel structures is to dig through the literature for interesting but little-used chemistry. Moses gives as an example N-hydroxy-2-oxopropionimidoyl chloride, for which a practical synthesis was reported in 1968. He says that at the time he was exploring how to use this reagent, he found only five examples in the literature.

Moses says he used the reagent in [3 + 2] cycloadditions for a range of alkenes synthesized by Key Organics chemists. And the products were further functionalized. "All in all, I made some 34 compounds via this strategy," he tells C&EN.

Peakdale Molecular also relies on "one-at-a-time custom synthesis, almost handcrafting the molecule," Jenkins says. The company is strong in heterocyclics, especially nitrogen heterocyclics, as well as carbonylations at unusual positions in the molecule. "That helps us access novel intermediates that drug companies are interested in."

Meanwhile, Onyx Scientific is focused on traditional medicinal chemistry, making analogs and helping develop a lead. On the basis of this strength, the company is looking at risk-sharing arrangements, whereby it will offer reduced rates for its chemistry services in exchange for royalties if a project is successful. Although only a very small part of the business could be done on that basis, Flinn says the company is considering the arrangement as another way to increase the value of the business.

Thinking back on 2002, a bleak year indeed, D'Ambra says: "If you go back to 1992, that was the year of the health care reform debate. For the first time, the drug industry laid off workers and cut back research. We're reaping what happened at the bench then. But the good news is that by the mid-1990s, R&D spending had picked up phenomenally. A lot of new drugs are in the clinic right now and will be hitting the market in 2004, 2005, and beyond. The boom is going to come back."

Chemical & Engineering News
Copyright © 2003 American Chemical Society