How To Reach C&ENACS Membership Number


 

May 27, 2002
Volume 80, Number 22
CENEAR 80 22 pp. 53-66
ISSN 0009-2347


DRUG DEVELOPMENT
Process chemistry results in pharmaceutical manufacturing routes that are safe, efficient, and scalable

CELIA M. HENRY, C&EN WASHINGTON

The pharmaceutical industry is being squeezed. With patents expiring on a number of profitable drugs and not many potential big sellers nearing the end of the development pipeline, the pressure is on to shorten the time required to get drugs on the market.

REACTION JOCKS Process chemists must quickly develop routes for manufacturing drugs.
PFIZER PHOTO
Scientists in process chemistry are feeling the economic pressures. They must devise practical syntheses for increasingly complex drugs in a constantly decreasing amount of time. Along the way, they must deal with issues such as cost, safety, efficiency, and environmental impact. Much of the success of process chemistry depends on a working relationship with organic chemists in academia, who provide new reactions--and new process chemists.

In the past decade, process chemistry has started to come into its own as a discipline. Michael J. Martinelli, a research fellow in chemical process R&D at Eli Lilly in Indianapolis, defines process chemistry as the interface of organic chemistry with business. "The chemistry itself is an applied form of organic chemistry targeted at specific molecules. It's more than just the ability to make the molecule. It's the ability to make it with a high degree of specificity and quality, cost effectively, with low impact to the environment."

People point to the American Chemical Society journal Organic Process Research & Development, which started in 1997, as an indication of the growth of the field as a separate discipline. In addition, a number of meetings focus on the topic, including a Gordon Conference on organic reactions and processes. A new conference in the field, sponsored by the ACS ProSpectives conference series (C&EN, Jan. 28, page 74), took place in Barcelona in February. And another ProSpectives conference on process chemistry is planned for February 2003 in San Juan, P.R.

Edward J. J. Grabowski, executive director of process research at Merck in Rahway, N.J., worries a bit about process chemistry developing into a field that is separate from organic chemistry. "I've always felt that the roots of process research are in fundamental science, in basic organic chemistry," he says. "We have to maintain a considerable overlap with basic organic research as one of our key process research activities." Grabowski strongly believes that process chemists have a responsibility to publish and to participate in meetings so that they can "contribute back to organic chemistry."

8021GrabowBOS
Grabowski
8021johnscott1
Scott
PROCESS CHEMISTRY is usually equated with scale-up, but characterizing process chemistry simply as the scale-up of a synthetic route does a grave disservice to the organic chemists who have chosen to focus their creative efforts in this field.

"Process research and development is not about making large amounts of compound," says John W. Scott, executive director of process R&D at Bristol-Myers Squibb Pharmaceutical Research Institute in New Brunswick, N.J. The synthesis of large quantities of a compound is really just a "by-product" of the real job, which is the acquisition of "process knowledge," Scott says.

Process chemistry has evolved from what Scott calls a "cottage industry"--in which each project was approached individually as a "once off"--to a "business process." In this business process approach, the goal is to identify aspects of process development that are common to all molecules. Scott doesn't have an exact number for those attributes, but he guesses that it's somewhere around 70% of the whole. That part of the development can then be attacked in the same manner for each molecule, allowing the chemists to focus their creativity where it's needed: on the 30% that is unique to the compound being developed.

For example, much process research is now done with automated parallel experimentation, analogous to combinatorial chemistry in drug discovery. A type of mathematical modeling called design of experiments is used to set up the experiments. Multiple variables are changed simultaneously to find the optimum process, as opposed to the more classical approach of changing one variable at a time (OVAT) and finding the best value for one parameter before moving on to the next.

Two things are missed by optimizing variables sequentially rather than simultaneously, Scott says. First, you are never totally certain that you have found the global, as opposed to a local, maximum. The reason for this is that the maximum is determined usually not only by the sum of the individual factors, but also by second- and higher order interactions of variables, which cannot be determined in the OVAT approach. Second, you don't know what happens if you subsequently vary a factor, information critical in determining how tightly that factor must be controlled in the final process.

"In one of the simpler experimental designs, you pick a low, medium, and high value for each variable," Scott says. "Then through a randomization process, you set up multiple parallel experiments in which combinations of variables are assessed for their impact on the desired output, typically reaction yield. You then get the computer to do a regression analysis of the impact of the variables on the experimental outcomes. You get a lot more information for the same number of experiments. You not only are sure of having the absolute maximum, but you also know how critical the variables are. Are you sitting on top of the Eiffel Tower, where variable control in the manufacturing setting will need to be very tight, or on the Sydney Opera House, where there are multiple optima and close control of some variables is more important than others?"

Some process chemists recommend that development issues should be considered early in the drug timeline, before a candidate drug is selected. Such "early integration" can help determine whether a compound is "druggable," Martinelli says.

FACTORS THAT MUST be considered in making such a decision include the cost to make the material, the projected dose, and the anticipated market value. Sometimes, Martinelli says, a particular molecule is just not practical when those aspects are considered. "Early integration of drug development into discovery will help determine if the molecule is developable in the context of the market and in the context of the biology."

Hans-Jürgen Federsel, head of project management at AstraZeneca's site in Södertälje, Sweden, believes that process chemists should be involved as early as the lead-optimization phase. Typically, Federsel says, process chemists would have about six months before they needed to provide material to support the preclinical development. "That worked when the molecules were fairly simple, but now we're seeing very complex molecules with several-stage syntheses," he says. "We don't have the capacity in our process R&D to really be able to get the chemistry right in this six months, to be able to quickly support the preclinical work."

Interacting with discovery chemists during the early phases of a project gives process chemists a chance to figure out the consequences of moving certain compounds into development, according to Federsel. "We could also be proactive in the sense that we try to influence our colleagues in medicinal chemistry, trying to go for simplicity instead of overemphasizing the complexity," he says. Federsel disapproves of the trend toward making drug molecules increasingly complex.

However, many process chemists advise against thinking that the synthesis developed during drug discovery must be practical. "The most important thing about the discovery route is that it is able to generate a lot of structural diversity," says Tamim F. Braish, director of pharmaceutical process development at Pfizer in Groton, Conn. "The discovery chemist is looking to make as many different types of analogs as possible. A diverse synthesis where they're making a lot of intermediates allows them the luxury of making many compounds that can be tested. Discovering new drugs is the goal, and coming up with the perfect process from day one is an unnecessary burden for the medicinal chemist."

Grabowski agrees. "It is the medicinal chemist's job to design new drug candidates," he says. He calls finding new drug candidates the "most formidable job" in the pharmaceutical industry. "I don't think the medicinal chemists should worry about process research considerations at all. If you ask the medicinal chemists to even begin thinking about process considerations, you're going to put a pair of handcuffs on them. Medicinal chemists do great chemistry, but their objectives are different from our objectives."

The cost of a synthesis, known as "cost of goods" to those in the business, is an important consideration for process chemistry. Braish uses a "back of the envelope" formula for determining a ballpark figure for a given synthesis. The sum of the
number of rings, functional groups, heteroatoms, stereocenters, and regiocenters multiplied by 100 gives a rough estimate of the target cost per kilogram when a process is fully scaled up and commercialized. Cost of goods includes cost of raw materials, manufacturing costs to make the drug, waste disposal, and licensing fees.

Another factor that process chemists must deal with is so-called green chemistry, in which they aim for energy-efficient processes that minimize or even eliminate the production of waste (C&EN, April 22, page 30). Roger A. Sheldon, a chemistry professor at Delft University of Technology, in the Netherlands, uses an "E factor" to measure the efficiency of various chemical industries, in terms of kilograms of waste per kilogram of desired products. According to Sheldon, bulk chemicals have an E factor of less than 1 to 5, compared with 5 to greater than 50 for fine chemicals, and 25 to more than 100 for pharmaceuticals.

IN FAIRNESS to the pharmaceutical industry, Sheldon acknowledges that pharmaceutical syntheses require more steps, which push up the E value. Most of that waste (80 to 90%) is inorganic salts, which means that "organic chemists make more inorganic salts than organic compounds," he says. Ways to reduce the E factor include using catalytic processes and reactions that are atom efficient. Plus, Sheldon says, process chemists should minimize the number of different solvents they use. "Of course, the best solvent is no solvent," he notes.

In addition to doing their own process research, pharmaceutical companies are also farming out some of the process development to outsourcing companies. "The reliance upon third-party manufacturers and the whole outsourcing world is becoming increasingly important," Martinelli says. "It gives us some extra capacity."

One such company is ISP Fine Chemicals in Columbus, Ohio. According to Satish C. Nigam, manager of process R&D, the company manufactures advanced intermediates for pharmaceutical clients. The information his firm receives from pharmaceutical companies ranges from just the identity of the target molecule--in which case ISP needs to devise the process from the ground up--to an advanced-stage process where the company has little freedom to change the chemistry. The biggest challenge, Nigam says, is providing clients with the synthesis they want at the price they want to pay. He calls outsourcing process development a "win-win situation" that "helps the customer reduce the time required for development."

Scott Denmark, professor of chemistry at the University of Illinois, Urbana-Champaign, notes that these custom chemicals companies are intermediaries between academia and the pharmaceutical industry.

Academia doesn't teach modern process chemistry, Bristol-Myers Squibb's Scott says. However, it fills two critical roles. The first is to provide new synthetic methods--the tools--to enhance the process chemist's toolbox. The second is to develop good scientists, whom Scott defines as people "who understand the scientific method, recognize a problem when they see it, know how to develop solutions to that problem, and go out and demonstrate solutions to that problem."

Kumar Gadamasetti, chief operating officer of X-Mine in Brisbane, Calif., believes that process chemists should regularly interact with academic scientists so that the academics can serve as "ambassadors" to help process chemistry "attract the best of the best." Fostering this interaction was one of the factors driving the selection of speakers at the Barcelona meeting, which Gadamasetti helped organize. The program featured a mix of academic and industrial scientists.

8021braisch1
Braish
PFIZER PHOTO
8021cov
PHOTO BY V. JAIN
Gadamasetti
ACADEMIA IS also a continuing source of new reactions for process chemistry. There are sometimes questions of how applicable an academic synthesis might actually be to process chemistry, whether it's scalable, for example. However, Scott says that he "would be the last person to dictate to the academic community what it ought to be doing." New reagents and new applications for existing reagents are "exactly what we need in industry. Our job is to figure out how to apply them most effectively in the context of the compounds we are developing," he says.

T. V. (Babu) RajanBabu, a professor of chemistry at Ohio State University, lets the needs of process chemistry guide most of the work he does. Before moving to Ohio State, he worked in industry for 14 years at DuPont. "I was in an industrial environment even though the research we were doing was more or less like academic research," RajanBabu says. "I probably have a better idea of what is feasible in terms of safety, what solvents are allowed, what chemicals are allowed, and, more importantly, what chemicals are not allowed than most of my academic colleagues."

Although RajanBabu works on basic problems, he wants to be able to see practical applications. "In choosing a project, we always use the following criterion: What is in it for a process chemist if all our wishes come true at the end of our research? If, at the end of our research, we are just producing pure science and nothing else, I'm likely to walk away from it."

Therefore, RajanBabu's group focuses its efforts on taking readily available, inexpensive materials and converting them into useful organic products. "We've got things like ethylene and carbon monoxide and hydrogen cyanide that are abundantly available, but we don't use them as much as we should in our fine chemical syntheses because of the lack of available science on how you do this," RajanBabu says. "We're interested in whether we can take these compounds, which are very stable, and use them for making fine chemicals. Done correctly, this effort will also contribute to environmentally benign manufacturing."

Denmark also takes some of the concerns of process chemists into consideration when he's developing reactions, particularly focusing on the generality and robustness of reactions. "We try to be very thorough in providing users with a manual, so they can anticipate whether our reaction would be applicable to their case," he says. Much of Denmark's work focuses on discovering new ways to make carbon-carbon bonds and on asymmetric chiral catalysis.

The trend of academic scientists patenting their work and seeking licensing fees prevents some new reactions from being used in industry as much as they might otherwise, according to Scott. Licensing fees are part of the cost of goods, a part that process chemists will avoid if possible. "Every process chemist I know views licenses in the same way. You pay them if you have to, but you work hard to determine if there is an equally efficient, license-free approach to the desired synthetic transformation," Scott says.

RajanBabu strongly believes that federally funded research should be in the public domain. Denmark, on the other hand, doesn't believe that academic scientists should refrain from patenting their work. However, he says, universities need to be realistic in the licensing fees they seek. "Many institutions view each discovery as the next billion-dollar income for the university. That causes problems in getting chemistry implemented," he says. Often the reactions have not yet proven their usefulness for scale-up. "We have to educate technology-transfer offices at universities that reasonable licensing fees are really all we can expect," he says.

CONFERENCES GIVE process chemists a chance to share examples of their work with researchers from both academia and industry. At the ProSpectives meeting in February, Braish presented the development of a synthetic route for the antibiotic trovafloxacin--trade named Trovan--as a case study in process chemistry. Trovan made it to the market but then was pulled because it caused liver damage in some patients, including 10 deaths. It is now available only as an intravenous form in a hospital setting when other antibiotic options have failed.

Trovafloxacin consists of a naphthyridone with a side chain of fused five- and three-membered rings. The synthetic route that came from drug discovery had a number of "issues," Braish said. It included a reaction with diphenylphosphoryl azide that couldn't be scaled to more than 100 g at a time. Other steps involved ethyl diazoacetate and Jones oxidation, which uses chromic anhydride.

"These are all great reagents for small-scale chemistry," Braish said. "But for a commercial process, you have to worry about the safety, economics, robustness, efficiency, and most importantly, the environmental impact. You can imagine running a large-scale Jones oxidation and having to deal with chromium disposal on a large scale. That is virtually impossible to do nowadays."

AFTER TRYING several routes to the side chain that didn't give them the product they needed, the Pfizer team settled on a convergent five-step route that started with the reaction of bromonitromethane and N-benzylmaleimide. This reaction formed the cyclopropane, provided the nitrogen functionality, and gave the desired stereochemistry in a single step.

However, the bromonitromethane reaction was not without its own problem. Tars were also formed in the reaction. By adding molecular sieves, the chemists were able to remove the tars and increase the reaction yield. "The sieves provided a surface for the tars to be deposited on," Braish said. "It allowed us to filter the sieves and the tars together. It was an easier manipulation of the reaction mixture." Eventually, the reaction was fixed to eliminate the tar formation completely, and the Pfizer team had to balance the needs for synthetic improvements while providing the large amounts of bulk needed to support multiple simultaneous clinical trials.

?Once the side chain was made, it was a fairly straightforward matter to protect it, couple it to the naphthyridone, and deprotect it. Altogether, the process required seven steps: five to make the side chain, one coupling step, and one deprotection.

The Pfizer group members were under severe time pressure as they devised the synthesis, because the preclinical and clinical development proceeded rapidly. The period between the two filings with the Food & Drug Administration--the investigational new drug (IND) to go into clinical tests and the new drug application (NDA) to market the product--was only four and a half years. "Time-line compressions are real," Braish said.

Another case study was presented at the Barcelona meeting by Margaret M. Faul of Lilly. She described the process development for a new class of retinoid-X receptor (RXR) agonists for treatment of type 2 diabetes. This work was performed as part of a collaboration with Ligand Pharmaceuticals, San Diego.

Faul described the synthesis of a compound identified as 411629, which is an analog of Targretin (bexarotene), a Ligand product already on the market for the treatment of T-cell lymphoma. 411629 consists of a tetrahydronaphthalene subunit connected by a cyclopropane bridge to a para-substituted pyridylcarboxylic acid.

The first-generation synthesis had nine steps, with a 12% overall yield. The intermediates required chromatographic purification. And incorporation of the cyclopropane ring represented half of the total synthesis cost.

8021Cole
Cole
ISIS PHARMACEUTICALS PHOTO
8021Margaret_Faul
Faul
LILLY PHOTO
In the synthesis of the tetrahydronaphthalene subunit, Lilly scientists were able to increase the two-step yield from 48% to 98% by making the key Friedel-Crafts alkylation catalytic in AlCl3, Faul said. They improved the selectivity in the synthesis of the pyridylcarboxy ester subunit. The two subunits were then coupled by a nitromethane-mediated Friedel-Crafts acylation reaction. Taking advantage of both the pyridine nitrogen and para-substituted carboxy ester functionality, the scientists incorporated the cyclopropane ring by chelate-controlled addition of CH3MgCl to the ketone, followed by treatment of the olefin with trimethylsulfoxonium ylide [
J. Org. Chem., 66, 5772 (2001)].

This second-generation synthesis took eight weeks to devise. It consisted of seven steps with a 24% overall yield and required no chromatographic purification. It has been used to produce the compound in kilogram quantities for clinical development. In addition, the chemistry developed for this process was valuable in identification of a second generation of RXR agonists for preclinical evaluation.

DNA and RNA oligonucleotides are not the kinds of compounds that are usually associated with process chemists, but that's what the process chemists at Isis Pharmaceuticals in Carlsbad, Calif., are tackling. At the ProSpectives meeting, Douglas L. Cole, vice president of development chemistry and pharmaceutics at Isis, described the work the process chemists have done to significantly reduce the cost of producing the company's antisense drugs.

One of the challenges in producing oligonucleotide drugs is the "tyranny of yield," Cole said. Yield is important in every process, but it becomes especially important when the drug consists of 20 nucleotides that must be coupled in a specific order. "There are going to be n–1 coupling reactions, however long this oligonucleotide drug is," Cole said. "They have to go well over 98.5% yield for the overall synthesis to be efficient, but the beautiful part is, get it right and you can then make drug after drug efficiently."

A MAJOR PART of the cost of oligonucleotide drugs is the cost of the nucleosides used as starting materials. Originally, the commercially available nucleosides were isolated from salmon milt (sperm-containing fluid). Some of the nucleosides that Isis uses are still obtained that way, but the company has also been able to add synthetic nucleosides to its supply chain as well. One reason for switching to synthetic nucleosides is that it's easier to certify that they don't contain bovine products, Cole said, something that's especially important in Britain and elsewhere in Europe.

The process chemistry work at Isis has focused on taking the purchased nucleosides and converting them into the nucleosides used in the drugs, some of which are alkylated or otherwise modified, and on improving the coupling chemistry.

When Isis started its process work in 1991, the reactions were all at the micromolar scale, producing only a few milligrams per synthesis. "It could consume months to make 100 g, and it did. To qualify a drug through the IND process and do a Phase I trial in humans, you're going to use 200 g if you're very careful," Cole said. "That was quite a chore to take chemistries and methods that were designed to make milligrams and turn them to the task of making hundreds of grams. The costs were quite high at that point."

One of the ways that Isis researchers reduced costs was by deciding early on that they would not use typical stirred vessels for oligonucleotide synthesis, but rather packed-bed reactors. When the work began, synthesis in a packed bed took 22 hours. "We reduced that cycle time by optimizing those reactions--the conditions, the solvents, and the reagents we used. We have the full synthesis cycle down to five hours. We can use our equipment more than once a day on a synthesis," Cole said.

The amount of solvent was reduced by using reactors in which the solid support is packed tightly. The smallest of these reactors are basically 35-cm-diameter tubes. The support is trapped between porous end plates that allow the reactants to be flowed through the bed. The molar excess of reactants was also reduced from 10- to 14-fold in the molecular biology versions of the reactions to only 1.7-fold in the current process. The total process includes a single preparative chromatography purification, which Cole said is "as scalable as the chemistry."

Altogether, the process improvements have resulted in a 50% overall yield for 20 cycles, the most conservative estimate based on initial bead loading. The costs, compared with the original 1991 synthesis, have fallen by 99%. "This 99% reduction in costs has occurred across the board," Cole said. "Consider everywhere that a process chemist would look and all the stones that you would turn over, and we've done it."

Process chemists do indeed look under every stone. They play a vital role in the pharmaceutical industry, developing large-scale syntheses that are safe, efficient, and economical. As the pharmaceutical industry comes under increasing economic pressure, the importance of process chemistry will only continue to grow.



Top


Chemical & Engineering News
Copyright © 2002 American Chemical Society



 
Visit Eastman

Advertise Here
Related Stories
Pharmaceutical Business
[C&EN, Jan. 28, 2002]

Green Chemistry For Pharma
[C&EN, Apr. 22, 2002]

Cutting-Edge Conferences
[C&EN, Jan. 28, 2002]

E-mail this article to a friend
Print this article
E-mail the editor
   
 

Home | Table of Contents | Today's Headlines | Business | Government & Policy | Science & Technology | C&EN Classifieds
About C&EN | How To Reach Us | How to Advertise | Editorial Calendar | Email Webmaster

Chemical & Engineering News
Copyright © 2002 American Chemical Society. All rights reserved.
• (202) 872-4600 • (800) 227-5558

CASChemPortChemCenterPubs Page