Chemical Process Research - ACS Publications - American Chemical

---

Chapter 1

Downloaded by 188.72.126.33 on October 15, 2016 | http://pubs.acs.org Publication Date: November 25, 2003 | doi: 10.1021/bk-2004-0870.ch001

Reflections on Process Research Edward J. J. Grabowski Merck Research Laboratories, Merck & Company, Inc., Rahway, NJ 07065

#Dedicated to my early Merck Mentors - Ed Tristram, Roger Tull, Pete Polack, Erwin Schoenewaldt, Mike Sltezinger, Earl Chamberlin, John Chemerda, Vic Grenda, Seemon Pines and Len Weinstock - a truly formidable bunch of teachers and scientists.

The achievements of an organization are the results of the combined efforts of each individual Vince Lombardi

The kind invitation of the Editors of this Symposium Series to provide an introductory chapter affords me an excellent opportunity to reflect on what has been a thirty eight year career in Process Research - all at the Merck Research Laboratories. There have been monumental changes in organic chemistry during this time, and today the field of Process Research is being pulled in more directions than ever before. Regardless, my belief is that our principal objectives remain the same: the design of practical chemical processes to support drug development, and the discovery and definition of new synthetic methods to support process activities.

© 2004 American Chemical Society Abdel-Magid and Ragan; Chemical Process Research ACS Symposium Series; American Chemical Society: Washington, DC, 2003.

1

Downloaded by 188.72.126.33 on October 15, 2016 | http://pubs.acs.org Publication Date: November 25, 2003 | doi: 10.1021/bk-2004-0870.ch001

2

When I was interviewing for a job in 1965 as the completion of my doctoral research approached, my thought was to be a medicinal chemist and engage in that alluring area called 'basic research'. I suffered all of the preconceived notions about 'process research', and thought that I would have none of it. A number of the process groups that I visited clearly fit the stereotype, until, of course, my interview with the Process Group at Merck. Max Tishler organized the group along non-typical lines, and there was a very strong emphasis on developing new science, studying reaction mechanisms, and participating in the greater chemical community. These activities were a natural part of the group's principal responsibility - the design and development of superb chemical syntheses and manufacturing processes to support Merck's product pipeline. At that time the methyldopa manufacturing process, which even today stands as an outstanding example of what a good process should be, was just coming into its own.(i) The excellence and enthusiasm of the Merck Process Group was clearly evident during my interview, and I decided to make a career change before my career had begun. Thirty eight years later, I still believe that it was the right choice. During the course of my career there have been major changes in what we do and in our responsibilities, and I am not sure that I welcome all of this, nor our response to it, with enthusiasm. In addition to our major responsibilities of designing practical syntheses and manufacturing processes, and preparing bulk drug to initiate early development, we are beset with a myriad of ever-growing regulatory requirements, endless discussions about low-level impurities, when or when not to make process changes, what are process changes, compression of program timelines, increases in drug requirements to begin development, etc., etc. Over the past ten years Process Research has emerged as a recognized field in organic chemistry, with its own journal and scientific meetings. I have always maintained that process research in the pharmaceutical industry must keep a considerable overlap with the academic chemical community. We are one of the few industrial organic chemistry enterprises that has the opportunity to do basic chemical research as a normal part of our job, and we must never let go of that opportunity. This overlap with academic research keeps us in close contact with academicians, who are keenly aware that careers in process research present wonderful opportunities for their students. I would be remiss not to comment on our closest allies in R&D in a vehicle such as this, and I'd like to begin this section with some thoughts on medicinal chemists and medicinal chemistry routes to drug candidates. Clearly, the medicinal chemists are part of the most creative group in the pharmaceutical industry - the discovery effort, which I consider the most challenging and difficult endeavor in our industry. I stand in awe of their contributions, and I continually marvel at the complexity of structure and synthesis at which

Abdel-Magid and Ragan; Chemical Process Research ACS Symposium Series; American Chemical Society: Washington, DC, 2003.

Downloaded by 188.72.126.33 on October 15, 2016 | http://pubs.acs.org Publication Date: November 25, 2003 | doi: 10.1021/bk-2004-0870.ch001

3 structure-activity studies are done. The era of methyl-ethyl-propyl is long since gone, and some of our carbapenem candidates reflect the heights to which discovery efforts aspire.(2) Such complexity in drug candidates reflects the current state of synthetic organic chemistry and organic chemists, and the armamentarium of reactions currently available. I think that process chemists often forget that the goal of medicinal chemistry is to design drug candidates, and not to design practical syntheses. Medicinal syntheses are designed for versatility, convenience and expedience. The drug candidate that results is not known when the program begins, and its initial synthesis is far from ideal by the very nature of the science. Yet process chemists persist in taking delight at setting up the medicinal synthesis of a drug candidate as a straw man to be readily pushed over with great glee. I suggest that process chemists would better serve their cause by not gloating over the inefficiencies of a first synthesis, but rather take the time to extract the best and most useful information from it and build upon it. Second amongst our co-conspirators in R&D are the chemical engineers. We all tread with fear at the first introduction of our chemistry into the pilot plant and casting it into the hands of the chemical engineers. I had once commented to Ed Paul, the head of our chemical engineering group, that the chemists felt that the first pilot plant run for a chemical process served to confirm O'Reilly's corollary to Murphy's Law ("Whenever something can go wrong, it will go wrong."). O'Reilly noted that Murphy was an optimist. With a gleam in his eye, Ed responded that the engineers considered the first pilot plant run as the chemist's final exam! In fact, there is some truth to both views, and often problems on scale-up result from an inadequate understanding of the parameters that control our chemical processes. This is more evident than ever in today's rushed climate, where getting final drug substance as quickly as possible is often considered the major objective. Analytical chemists and pharmaceutical scientists are also our key partners. Analytical chemistry has undergone a revolution in instrumentation, but sometimes I think this creates as many problems as it solves. The pharmaceutical scientists possibly have the most complex challenges. Despite the wonderful advances in science and technology over the past decade, much of what they do still contains a strong component of art. Regardless of the plethora of data we might acquire on a final bulk form, it is still a matter of trial and error to determine a workable drug formulation. And the only test of the response of that formulation to scale-up, is the actual scale-up. One of my favorite topics in process research is chemical lore, and we have all fallen victim to it. Years ago we were designing and developing the first practicable synthesis of imipenem, the first carbapenem antibiotic to reach the market(J) As part of this synthesis lactone 1 (Scheme 1) was solvolyzed in benzyl alcohol, which afforded a 3:1 mixture of the desired benzyl ester and the

Abdel-Magid and Ragan; Chemical Process Research ACS Symposium Series; American Chemical Society: Washington, DC, 2003.

Downloaded by 188.72.126.33 on October 15, 2016 | http://pubs.acs.org Publication Date: November 25, 2003 | doi: 10.1021/bk-2004-0870.ch001

4

unopened lactone. This was the equilibrium ratio. The ester was cyclized to the P-lactam, which was then hydrogenolyzed to the free acid in preparation for a two-carbon chain extension. As we attempted to scale-up the benzyl alcohol solvolysis reaction, we began to recognize that such was impossible. We could not readily remove the benzyl alcohol, and as we handled the system the benzyl ester would revert to lactone. Faced with this dilemma, we wondered about solvolyzing the lactone in methanol. We were happy to note that the equilibrium ratio of the ester to lactone was now 97:3, but we were faced with the problem of hydrolyzing a methyl ester in the presence of the P-lactam. There were two parts to the established lore: everyone agreed that one could NOT successfully effect the desired hydrolysis as the P-lactam was far more reactive than the methyl ester to aqueous base; and the medicinal chemists had already demonstrated that the reaction did not work. Facing likely failure in that we could never develop the benzyl alcohol reaction into a practical process, we initially ran the hydrolysis of the methyl ester lactam with IN NaOH in water. We were pleased that the lore was indeed wrong, and the selectivity for methyl ester hydrolysis over p-lactam opening was >100:1. Ultimately, we developed a process where the hydrolysis was done at a 30-50% concentration of the ester in water with 5N NaOH at 0°C and pH 100:1 for the Na Salt at pH 9 at 50% cone, and 25° is 22 years!

C0 H

1

C0 H O

3 Parts |

I

1 Part

A

OH

O

OH

O

O

OH

H

2

2

^C0 CH

H

H

2

2

C0 Na

3

C0 Bn

5NNaOH/H 0 pH