I wrote here about a very unusual dinitro compound that's in the clinic in oncology. Now there's a synthetic chemistry follow-up, in the form of a paper in Organic Process R&D.
It's safe to say that most process and scale-up chemists are never going to have to worry about making a gem-dinitroazetidine - or, for that matter, a gem-dinitroanything. But the issues involved are the same ones that come up over and over again. See if this rings any bells:
Gram quantities of (3) for initial anticancer screening were originally prepared by an unoptimized approach that was not suitable for scale-up and failed to address specific hazards of the reaction intermediates and coproducts. The success of (3) in preclinical studies prompted the need for a safe, reliable, and scalable synthesis to provide larger supplies of the active pharmaceutical ingredient (API) for further investigation and eventual clinical trials.
Yep, it's when you need large, reliable batches of something that the inadequacies of your chemistry really stand out. The kinds of chemistry that people like me do, back in the discovery labs, often has to be junked. It's fine for making 100mg of something to put in the archives - and tell me, when was the last time you put as much as 100 milligrams of a new compound into the archives? But there are usually plenty of weak points as you try to go to gram, then hundreds of grams, then kilos and up. Among them are:
(1) Exothermic chemistry. Excess heat is easy to shed from a 25-mL round-bottom flask. Heat is not so easily lost from larger vessels, though, and the number of chemists who have had to discover this the hard way is beyond counting. The world is very different when everything in the flask is no longer just 1 cm away from a cold glass wall.
(2) Stirring. This can be a pain even on the small scale, so imagine what a headache it is by the kilo. Gooey precipitates, thick milkshake-like reactions, lumps of crud - what's inconvenient when small can turn into a disaster later on, because poor stirring leads to localized heating (see above), incomplete reactions, side products, and more.
(3) Purification. Just run it down a column? Not so fast, chief. Where, exactly, do you find the columns to run kilos of material across? And the pumps to force the stuff through? And the wherewithal to dispose of all that solid-phase stuff once you've turned it all those colors and it can't be used again? And the time and money to evaporate all that solvent that you're using? No, the scale-up people will go a long way to avoid chromatography. Precipitations and crystallizations are the way to go, if at all possible.
Reproducibility. All of these factors influence this part. One of the most important things about a good chemical process is that it works the same flippin' way every single time. As has been said before around here, a route that generates 97% yield most of the time, but with an occasional mysterious 20% flop, is useless. Worse than useless. Squeezing the mystery out of the synthesis is the whole point of process chemistry: you want to know what the side products are, why they form, and how to control every variable.
Oh yeah. Cost.Cost-of-goods is rarely a deal-breaker in drug research, but that's partly because people are paying attention to it. In the med-chem labs, we think nothing of using exotic reagents that the single commercial supplier marks up to the sky. That will not fly on scale. Cutting out three steps with a reagent that isn't obtainable in quantity doesn't help the scale-up people one bit. (The good news is that some of these things turn out to be available when someone really wants them - the free market in action).
There are other factors, but those are some of the main ones. It's a different world, and it involves thinking about things that a discovery chemist just never thinks about. (Does your product tend to create a fine dust on handling? The sort that might fill a room and explode with static electricity sparks? Can your reaction mixture be pumped through a pipe as a slurry, or not? And so on.) It looks as if the dinitro compound has made it through this gauntlet successfully, but every day, there's someone at some drug company worrying about the next candidate.
1. Anonymous on February 22, 2012 10:35 AM writes...
Don't worry, with the direction that pharma is currently headed, there won't be ANY chemistry pretty soon, let alone discovery and process groups.
Permalink to Comment2. Anonymous on February 22, 2012 10:37 AM writes...
Outsource it, like everything else.
Permalink to Comment3. Hap on February 22, 2012 10:59 AM writes...
But that only works when the outsourcing group is 1) cheap (relatively) and 2) provides consistent material that does what it's supposed to (so the FDA doesn't come knocking, and so that you can can continue to sell or test the drug while the patent and exclusivity clocks continue to run down). That means that someone has to care about the problems in the post, even if the discovery team (wherever it is) doesn't.
Permalink to Comment4. quintus on February 22, 2012 11:24 AM writes...
Send the process to China
Permalink to Comment5. simpl on February 22, 2012 11:45 AM writes...
Outsourcing of chemical intermediates, whether to BASF or China, works much better than a lot of business outsourcing (like IT support, document management, analytical methods development). The internal chemists keep a good hold of the keys like partner choice, specifications or synthetic path, and the partners are sufficiently competent.
Permalink to Comment6. processchemist on February 22, 2012 12:09 PM writes...
Welcome to the (once) wonderful world of scaled up chemistry.... that received no better treatment than the one reserved to medchem... I wonder if in AZ there's still a chemist remembering the great tradition of ICI and AZ about scaled up chemical reactions... If some of you think that industrial medchem culture is fading, process chemistry began to fade years before. And currently you can find project managers with very little knowledge about the issues/requirements of scaled up chemistry in a pharma context. Two quick examples:
1) A 20 l rotavapor working in GMP anyone? (Wow, you begin your clinical trials with an investigational API without any known drug form - and we're talking about tablets...)
2) In this process is crucial to use MIBK distilled over phosphoric anhydride (can we talk about water by KF? Then let us take care of the details, PLEASE).
Permalink to Comment7. Curt F. on February 22, 2012 12:11 PM writes...
I love this post. It's essentially an homage to chemical engineers and chemical engineering. I guess a lot of the folks who actually do this work view themselves as chemists. But (a) they should learn to embrace their engineering identity!, and (b) it doesn't matter what you call these folks (and whether they're in an outsourcing firm in China or in the US), they do good important work that does not often have high visibility to external observers of an organization.
Permalink to Comment8. andrewD on February 22, 2012 12:21 PM writes...
Curt FThey do when we screw up!
Permalink to Comment9. Chemjobber on February 22, 2012 12:28 PM writes...
"they should learn to embrace their engineering identity!"
I'm pretty comfortable with my identity as a process chemist, thanks. An engineer? That's a lot of math that can't be absorbed easily.
Permalink to Comment10. CTChemist on February 22, 2012 1:00 PM writes...
Anyone know the projected daily dose for this thing?
Permalink to Comment11. processchemist on February 22, 2012 1:07 PM writes...
@ Curt F.
I worked with engeneers and, believe me, we're two different kind of animal species (and the ones I know would say the same thing).
Permalink to Comment12. Biotechtranslated on February 22, 2012 1:30 PM writes...
"Where, exactly, do you find the columns to run kilos of material across?"
At the old Upjohn pilot plant!
I have no idea if it's still there, but a few years back there was a 2 story tall column, about 60 cm (2 ft) in diameter.
The technicians would load it with approximately 10 barrels of silica gel (50 kg each I think) and then gravity feed the column from 55 gal drums of solvent.
The guys with experience could pull the fractions off, test via TLC and strip the solvent as fast as it came down the column.
Of course this wasn't standard procedure for a scale-up, but if it had to get done and you had no other way, it was possible.
Mike
Permalink to Comment13. CMCguy on February 22, 2012 6:51 PM writes...
"There are other factors, but those are some of the main ones. It's a different world, and it involves thinking about things that a discovery chemist just never thinks about." The missions and operations of med chem vs process are indeed different therefore its hard to be overly critical of routes and procedures applied to bench scale however having suffered through enough translation scale up projects I often wish more med chemists (and more so academics) would have better awareness of certain issues they can create by choices. Exotic and expensive reagents or conditions may be fun to try out in the lab yet typically will ultimately require modification of such techniques to progress in development.
IMO a fundamental aspect that can often be seemingly ignored is the impact of reaction concentrations. Direct scale-up conversions of dilute solutions common on the bench would directly calculate to increasingly larger vessel requirements in the plant whereas learning to consider and attempt adjustments to doing chemistry more concentrated could go a long way to a more scalable path forward.
Permalink to Comment14. anonymous on February 22, 2012 7:22 PM writes...
Looks like a damn good alkyating agent....just sayin'. OH, that's right, it's for cancer!!! Never mind.
Permalink to Comment15. partial agonist on February 23, 2012 11:42 AM writes...
After you take a pill of that stuff, you had better not jump up and down too much
Permalink to Comment16. Anonymous on February 23, 2012 10:58 PM writes...
Can you say "LACRAMATION" !!!! DOH
Permalink to Comment17. GladToMoveToProcess on February 24, 2012 1:53 PM writes...
@11: Right on! Most engineers I've worked with didn't know that much chemistry ("Couldn't we use ethanol for this hydride reduction?"), but they knew pipes, pumps, heat exchangers and the like really well. It was a joy to work with the good ones.
Permalink to Comment18. processchemist on February 25, 2012 5:34 AM writes...
@17
And some of the ones I worked with are much better than an industrial or process chemist to dimensionate reactions where phase transfer and fluid dynamics are the controlling factor (but don't ask them to investigate about the more performing catalyst.... )
Permalink to Comment19. Tyrosine on March 1, 2012 6:57 AM writes...
Regarding problems 1 and 2, exotherms and inadequate mixing, we've had some really nice success with flow chemistry - scaling up to tons.
Permalink to Comment