On the Medicinal Chemistry Pedagogy

29 November 2023

This is an editorial of sorts adjacent to topics covered in this outstanding editorial recently published by Dean Brown in J. Med. Chem. It’s required reading on its own, but a pre-read is not necessary to understand what follows. For background, I’ve worked in big pharma twice and also at a biotech twice, so I feel like I have a good grounding in the similarities and differences. I’ll also caveat though that my experiences are necessarily anecdotal, and your mileage may vary. That’s why I’m clearly marking this as an editorial and not citing a bunch of references.

On the training of medicinal chemists, the pedagogy for the entire 20+ years of my career has been on-the-job training. It’s still the case today as when I started my career in 2002 that the great majority of industrial medicinal chemists come out of academia with a background in some flavor of synthetic organic chemistry — often knowing little-to-nothing about the practice of medicinal chemistry. The training is an apprenticed model, where more senior medicinal chemists in an organization impart the ways of the Force to the new hires.

What has changed, as the J. Med. Chem. editorial points out, is that the medicinal chemistry structure of large pharmas — where most of this learning happened in the past — has been forever rearranged. Instability and change are the new normal. The times of spending one’s career at a single company are increasingly in the rear-view, driven by massive rounds of downsizing that have shifted more and more of the medicinal chemistry workforce out of big pharma and into biotech. In parallel, synthetic organic chemistry, which used to be the exclusive domain of in-house medicinal chemists, has been increasingly outsourced to countries with cheaper labor rates than the Western bases of most pharmas, notably China and India.

A reasonable concern coming along with these industry-wide shifts is that the wheels will come off the longstanding pedagogy, which depends on a sufficient talent density to teach the next generation of medicinal chemists. Bruce Booth wrote about this talent problem eloquently and at length back in 2014. Here I will argue that this concern has not been realized over the last decade, and that the pedagogy remains functional — perhaps with some changes that are for the better.

In my sense of things, the forced exile of talent from big pharma to biotech has simply shifted the learning environment to a different set of smaller venues that could have significant upside for learning and career development. Many of those big pharma medicinal chemists didn’t hang it up when they were laid off — they just went somewhere else. And with them, they brought their skills to teach to the new hires.

As an example, at my current job, when I first joined, it was a startup environment with <40 people. I had 15 years of experience as a medicinal chemist… and I was the most junior senior medicinal chemist (!) we had on staff. Our density of medicinal chemistry talent per capita was (and is) far in excess of any big pharma I’d ever worked for. Despite having many years under my belt by then, the subsequent 7 years have witnessed some of my greatest professional growth as a chemist because I was surrounded by talented folks who challenged me every day.

Now that we’re a little more mature of an organization and we’re hiring in more early career folks, I pitch to potential hires what a fantastic environment we have to learn medicinal chemistry in, because we have great chemists to learn from — and all of those senior people are invested in spending time teaching our new folks. Increasingly, I think this kind of environment is where our talent is going to be trained in the future, and is a scenario Bruce Booth envisioned in the piece referenced a few paragraphs up. The concern about the scalability of the approach for training given the size of biotechs vs. big pharma has been mitigated in part by the massive influx of capital into the biotech space and high rate of startup creation in the last few years of the 2010s and into the early 2020s. Hiring was crazy brisk for a while and many folks were getting snapped up right out of school again in a way I hadn’t seen since my first job interviews in the early 2000s. (It’s since cooled off, of course, thanks to high inflation / interest rates, the Inflation Reduction Act, etc. — but we had quite a ride up there.)

Moreover, the variety of backgrounds the senior folks came from has allowed us to weigh institutional ideas and practices from many other legacy organizations and select the best of them to encode into our own medicinal chemistry credo. Kinda like viruses doing recombinant gene exchange, but more deliberate. Large organizations, by contrast, may provide a big training ground, but they’re also enormously prone to groupthink concepts that are simply wrong, and thereby propagate these wrong ideas to the next generation.

As an example of this latter point, few things have been as abused over the years by big pharma as the free drug hypothesis. Exhibit A: witness this paper, which makes a big point of trying to “improve” plasma protein binding in a series of compounds to improve free drug exposure… which, if you understand the free drug hypothesis, is nonsense. You may recognize one of the co-authors. (The rest of the paper is fairly solid, and I mean no disparagement to my old friends, but the point re: free fraction stands, and we approached this incorrectly here.) I wouldn’t publish this paper with this title today, because I know better now — but I learned better by being part of other organizations with different ideas that challenged my thinking about the free drug hypothesis. Would I have learned that staying in one place for 20 years? Eventually, probably, but more slowly.

Although the uptake of the free drug hypothesis among medicinal chemists has improved in the last 20 years, it was not historically encoded into the DNA of many large pharma cultures. Stirring that concept around in the great biotech melting pot where contrasting opinions had to duke it out has allowed this idea to emerge stronger in our collective thinking as a discipline. Indeed, the chemist changing jobs can be a powerful change agent by bringing new ideas with them to their new job. The probability of such impact is magnified in biotech, where there are less fish in the individual ponds and far less organizational inertia to overcome.

Now, it’s not necessarily true that you’ll join a small biotech as an early career chemist and be fortunate enough to be surrounded by a group of experienced medicinal chemists to learn from. This is where I would strongly advise folks looking for their first job: make sure you have enough chemists (~5 minimum with at least a few 10+ year veterans) around you to learn from at your potential employer. If you’re going to be the only chemist on staff, you’re being set up to fail — so don’t do that to yourself. The smallest startups where chemistry is a one-person band are going to need an experienced chemist in that role, and I’d beware anyone who tells a newly-minted chemistry PhD that they will “just pick it up” in that kind of environment. They won’t. It’s mostly true that small startups won’t even want to hire a fresh PhD because they can’t afford the time and capital to train someone, but slightly larger biotechs have more flexibility in that regard as they start planning for the long haul.

It is true, though, that folks are being thrust into roles of greater responsibility in biotech much earlier in their careers than would have ever traditionally been the case in big pharma. Biotech has had no choice but to adopt this model, due to heavy (and likely irreversible) outsourcing of synthetic chemistry and, until recently, an acute labor shortage. The historical model was to spend 4-5 years mostly at the bench, focused on doing wet chemistry while learning med chem from the experienced folks, and gradually transitioning into a project leadership role. The vast increase in outsourcing of wet chemistry over the last 20 years means that more entry level folks are being hired into project leadership or design roles right out of the gate — because those are the roles that are left to be had.

I’ll offer here my unpopular opinion that the “lead from the bench” approach is largely a failed paradigm. It’s good in principle to remain directly connected to the project synthetic chemistry, and the virtues of this approach (especially facile cross-fertilization of synthesis to design, and vice-versa) are spelled out well in the J. Med. Chem. editorial. In practice, though, it’s more important to be in contact with people doing synthesis (or design) rather than try to do both yourself. This division of your time is dilutive to your efforts. At my last big pharma (and last biotech) I was put into a project leadership / design role while also being expected to spend the majority of my time at the bench doing wet chemistry. These are two jobs that really require 100% focus to do well, and the result is people burning it at both ends for 200% and a high rate of burnout.

Case-in-point: as a PhD chemist, you’re expected to be able to tackle big, complex, thorny synthetic chemistry problems. Anyone who’s done a synthetic organic chemistry PhD can tell you that you need regular, uninterrupted stretches of time in the lab to make headway. But the “lead from the bench” model pulls you in the opposite direction: instead you’re spending half your time going to team meetings, updating management, etc. — and often at times that are controlled by other people. Your lab time starts getting chopped up into an hour here, two hours there, a day on, a day off. Pretty soon you find yourself tackling mundane problems in the lab instead of difficult ones because those are the ones you have time to make headway on with such a disjointed schedule. Then you have to ask what value that adds vs. outsourcing those mundane tasks to CRO chemists.

A dozen years ago now, Pfizer was roundly criticized in many quarters for dividing their synthesis and design teams into separate roles. This move presaged what has become standard practice in small biotech these days. Synthetic chemistry is often heavily outsourced and the internal synthetic footprint is small — to the extent that only design roles may remain. Having lived a long time with a foot in both worlds, I think it’s beneficial on balance to separate these functions — as long as those teams are in communication, we acknowledge the real world where these divisions aren’t absolute, and we center the notion that good ideas can come from anywhere.

The preceding discussion raises a question: is the historical model of taking 4-5 years to ease into a project leadership role with greater design responsibility the correct model? It’s been my experience that very quickly, a talented chemist knows enough to take something on with support. Getting really good at the nuts-and-bolts of drug discovery is at least a 10 year endeavor, but you gotta start with something sometime. So why not sooner rather than later? Maybe not a full-blown lead optimization effort, but certainly something earlier in the portfolio where the stakes are a little lower and there’s room for some mistakes to be made while ascending the learning curve. At my first big pharma job, I witnessed a significant degree of discontent, and ultimately turnover, because the supply of talented chemists in the <5 year experience bucket who were ready to take on leadership roles far exceeded the number of projects available to lead.

Can this idea be taken to the extreme of throwing a new hire right into the deep end and asking them to just start leading something, while learning medicinal chemistry on the fly? In my experience, this model can work, as long as: 1) you’ve hired a person who shows great adaptability and problem solving skills (so interview them well, friends); and 2) the support network of experienced folks is there to provide the high level of guidance these new hires will need for a while. I’ve seen enough early career people take flight in this way now (and what a joy to behold when it happens!) that I question the dues-paying model that I grew up with. This is also a heads-up to new job seekers: more than your forebears, the med chem jobs you’ll find in biotech are increasingly distant from the organic chemistry skill set you probably trained with in grad school, so emphasizing transferable skills is more important than ever.

In summary, I think the model for training medicinal chemists in industry is alive and well. The central venue has shifted from big pharma to biotech, and with that has come more cross-pollination that allows good practices to propagate and stagnant ideas to be discarded more quickly. Perhaps we’ve learned that the duration of the apprenticeship period can be shortened significantly with proper support in place, and maybe focusing on design & project leadership over simultaneously being at the bench isn’t so bad. Despite the changes wrought on the industry over the last 20 years, the role of experienced medicinal chemistry mentors remains as centered and important as ever. I’m cautiously optimistic that we’re passing down the knowledge needed to train the next generation of medicinal chemists.