Derek Lowe takes a look back over his 20 years as a medicinal chemist

I’ve recently marked my 20th year of drug discovery research, which prompted me to think about what has changed since I started work in the industry. If my 27-year-old self were to skip ahead 20 years, I wondered, what would stand out?

At the lab bench, the biggest change in synthetic chemistry would be the ’rise of palladium’. It seems strange now, but metal-catalysed couplings were still somewhat exotic in 1989, and many working medicinal chemists had never run one. The idea that these reactions would eventually earn the same ’can’t-live-without-them’ status as amide formation would have raised some eyebrows. 

Automated column chromatography (with pre-packaged columns) would have also impressed me greatly, considering the number of columns I was running by hand. Even automated nuclear magnetic resonance sample handling wasn’t in general use back then. And speaking of autosamplers, routine walk-up liquid chromatography-mass spectrometry machines would also have struck me as the height of luxury, if I can mention that word without sounding too much like Monty Python’s ’Four Yorkshiremen’ comedy sketch, but their use would have explained the decline in the use of thin layer chromatography plates.

Quantum leap 

And as with every other part of human life, the extraordinary rise in computing power would have been immediately apparent. Electronic lab notebooks were not even considered in 1989. When I joined my first industrial research group, management was suspicious of the idea that every chemist should have a computer. Shouldn’t they be at their lab benches instead? And should project data really be kept on one of these ’servers’? I would have been simultaneously amused and alarmed though by the ’still-working-on-it’ state of molecular modelling in 2009, which 20 years ago was widely thought to be rapidly taking over the research world. 

What about the general practice of medicinal chemistry? Here I would have noted the expected improvements in high-throughput screening and a general rise in assay capacities. I would have been especially interested in the increased amount of pharmacokinetic and metabolic data that is now generated for any compound of interest. Projects in 1989 had much less data to work with and it was much slower to collect. The amount of genomic information available now would have also amazed me. But now comes the difficult part: hearing about these things would have surely caused me to expect higher success rates. After all, weren’t these advances just what we needed? With all this information and all these wonderful ways to share it and evaluate it, wouldn’t we finally know what we were doing? 

And that’s the realisation that would have crept up on me: the business of really understanding disease states, and of doing something meaningful about them, was even more complicated than I’d ever thought. None of these new technologies, I would have noted with unease, had apparently been enough to remove the huge uncertainties that still hovered over every research project.  

A closer look would have confirmed this - I’d been hired in 1989 to work on schizophrenia and Alzheimer’s drug projects. But 20 years later, I’d note that the best schizophrenia drugs would still be murky multireceptor agents, with no solid drug-friendly mechanism in sight. And finding out that in 2009 people were still arguing (with good reason) about whether β-amyloid was the cause of Alzheimer’s would have come as a terrible shock.  

No progress? 

And that brings us to the present day, where many of our biggest unmet medical needs are still with us 20 years on. We can hope that the tools we have now will look similarly outdated after another 20 years, but what about our progress against disease? We’ll probably have to win each piece of ground the way we have in the past, through sheer effort - in which case, the continued rise of automation and computing power will be critical. But we can hope there are critical masses of knowledge, ones we haven’t reached yet, that will accelerate things. 

Strange as it may sound, it’s our ignorance that gives me room for optimism. There are clearly many important things about human biology that we just don’t understand very well, and there are surely some key features that we haven’t even suspected. So it’s my hope that drug discovery will eventually look almost nothing like it does now. Will that take another 20 years - or longer? 

Derek Lowe is a medicinal chemist working on preclinical drug discovery in the US