About this Author
College chemistry, 1983
The 2002 Model
After 10 years of blogging. . .
Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases.
To contact Derek email him directly: email@example.com
January 12, 2003
I've been reading George Dyson's interesting history of Project Orion, the late-1950s attempt to design a spacecraft powered by sequential nuclear explosions. (A borderline crazy idea, it very likely would have worked. The big question became whether it should be allowed to work at all.)
He quotes his father, Freeman Dyson, about the early days of the project:
"Everybody did a little of everything. There was no division of the staff into phgysicists and engineers. The ethos of engineering is very different from that of physics. A good physicist is a man with original ideas. A good engineer is a man who makes a design that works with as few original ideas as possible."
There's a lot of truth to that. So in which category is work in medicinal chemistry? The answer isn't immediately obvious, especially for people just starting out in the business. In graduate school, the emphasis is (rightly) on the pure science: as many original ideas as possible (as long as you can get them to work, one way or another.) So when freshly coined PhDs or post-docs join a drug company, they're sometimes under the illusion that unusual new chemistry is what's called for at every opportunity.
And nothing could be further from the truth. From an organic chemistry standpoint, medicinal chemistry can be downright boring. The sooner that new researchers figure that out, the better off they are. You can do perfectly respectable medicinal chemistry using nothing but reactions and ideas from an undergraduate textbook. (As I've pointed out, those reactions got to be classics because they tend to work, which is just what you need.)
The point of medicinal chemistry isn't chemistry; that's just the means to the end. We do just as much cutting-edge chemistry as we have to, and no more. That stuff takes a lot of time to figure out - and we have plenty of other problems that are waiting to take plenty of our time. The chemistry had better just quietly work for the most part, if you're going to have a chance.
The original ideas come when it's time to decide what molecules to make, and when it's time to figure out why you're getting the biological effects from them that you are. In those areas, we'll take all the original thinking that anyone can provide. Any weird brainstorms about how to make a compound more potent or more selective are welcome. And if making those new molecules calls for nothing more than ancient reactions, yawners that bore the pants off everyone who does them, then so much the better: that means that the molecules will be made quickly and in a good quantity. (One of the worst binds you can be caught in is to have a wonderful lead structure that you can't find a way to make enough of.)
So, when it comes to chemistry, we're engineers. When it comes to medicines, though, we'd better be the next best thing to poets.
+ TrackBacks (0) | Category: Who Discovers and Why
January 9, 2003
I don't want to give the impression that there are hundreds of gems buried among the papers that no one references. Sometimes no one references them because they're not worth very much, or because no one can get ahold of the actual article. I had a old reference turn up the other day from the local "Proceedings" journal of an obscure Egyptian university - I should have threatened our library staff with a photocopy request. You probably couldn't find it short of London; I seriously doubt that any reference library on this continent has a copy. Heck, you'd probably have trouble tracking it down in Egypt. No one will ever know if it's any good.
But most journal articles in chemistry just disappear from view, because they say what they have to say and get off the stage: "We made compound Z for the first time," or "This palladium catalyst is great when you have exactly the sort of starting material that we have," or "Sometimes this reaction works well, and sometimes, darn it all, it doesn't."
These aren't groundbreaking classics, but they're still valid work. And thanks to modern literature-searching tools, they'll be found whenever someone might really need them (if ever.) Some paper that sits composting quietly for years can suddenly turn out to be vital for another researcher who wasn't even born when it first appeared (I've been that researcher a couple of times myself.) At least twice in my career I've gone to copy a paper out of a bound volume of an old journal and realized that a few years before I'd copied the paper right next to it, for a completely different research project. Last time that happened, I looked at the next paper after that one, wondering if I'd need to come copy it a few years from now. (On closer inspection, I hoped not.)
I've always enjoyed being back in the wilderness of the bound journals in a large library. Of course, as time goes on, I can't help but notice that some of these journals that I can remember seeing seeing as new issues are now in the back storage room. Hmmm. You mean to say they've filled out this entire shelving unit with Journal of Organic Chemistry since I was reading it my senior year of college? Let's see, at this rate, it'll be out to. . .here by the time I retire. Hey, J. Alfred Prufrock measured out his life with coffee spoons; things could be worse.
+ TrackBacks (0) | Category: The Scientific Literature
January 8, 2003
I've had some interesting e-mail on the subject, which I thought I'd address here for the curious. One person mentioned the possibility of ricin dissolved in DMSO. I have to say that that's a nasty thought, because DMSO certainly does increase skin permeability. But I don't know how soluble a large peptide like this would be - even in DMSO, which is generally a solvent of last resort in chemistry. And even if you could get some of the protein in there, odds are excellent that it would denature, change its conformation as it went into solution. Most enzymes shift around so much going into solvents like DMSO that they lose their activity completely. Not all of them, though - but I would put ricin in the category of unlikely to survive the transition. It has an important disulfide bond that would probably be labile to oxidation on storage in DMSO as well.
Others have mentioned food adulteration. If my guesstimate of a gram or two for lethality is right, a big problem would be that the stuff would probably alter the taste of whatever you added it to. I certainly have no idea of what ricin tastes like - and I'm not about to find out, because sublethal doses are still pretty unpleasant. But it's unlikely to be unnoticable. The heat of cooking is an even better denaturant than any organic solvent, usually, but ricin is said to be unusually heat-stable. That's not saying much in protein chemistry, though - boiling water is considered insane heat in the protein world. It's not likely to be a useful agent in someone's french fries; you'll just have to count on the acrylamide.
+ TrackBacks (0) | Category: Chem/Bio Warfare
Spent some quality time in the library at work today, digging into another aspect of a project that I'm working on. As you get deeper into the literature on a given scientific subject, some things happen over and over. There will be articles that everyone refers to, the standards that are like showing a form of ID: "Yes, you can take me seriously, because I'm referring to the big papers that everyone in this field should know about."
And there will be papers that somehow got lost in the shuffle, things that are a lot more important than they look, that no one paid enough attention to. You really have to know the field well to recognize these when you see them. And there's always a nagging doubt: "Why doesn't anyone talk about this? What's wrong with it, anyway? If it were good, people would have referenced it. . .right?" The literature-searching tools we have now are gradually giving these papers a better chance to be noticed, but if they're published in an out-of-the-way journal, they still won't get read the way they should be.
Sometimes these papers are lost more in time than in space. More than once I've found that a topic I thought was the latest rage had been anticipated years before. Sometimes the nomenclature has changed enough so that people don't realize that the earlier work is relevant, and sometimes people just don't bother looking at the old literature. A recent project of mine turns out to have relevant papers from thirty years ago, which is remarkable since the same underlying idea is still of interest. Very few modern papers reference these at all; you have to look closely.
What strikes me every time I learn more about a field, though, is how the details start resolving as I get closer. From a distance, when you don't know much about an area, the main points of it look large and chunky: Enzyme X is involved in doing Reaction Y, and it's found in tissue Z. Then as you start to get into the primary literature, all these start breaking into pieces. . .turns out there are several subtypes of Enzyme X - at least, some people say there are, but this other group says that they can't verify that. . .and it does Reaction Y, all right, but it can also do three others, one of them both in forward and reverse - seems to vary depending on the species you look at. . .and here's a paper saying that it's in tissue Z, sure, but it's a lot more important in this other organ where you can barely find it. . .and so on. All these solid rocks of knowledge start turning into mica, exfoliating into piles of complicated details.
The same thing happens when science as a whole approaches a new area. Broad ideas are all we can see at first, but further inspection is rewarded by puzzling anomalies. We may come to a point where we understand things, but the complexity just keeps on increasing the closer we look. It's like a fractal image - just when you think you can see the outline of the whole thing, you realize that those curves have tiny curves on them, which really look as if they themselves have. . .and so it goes. It's just how the world is put together. And I have to say, it would be a lot less interesting if it were easier to understand.
+ TrackBacks (0) | Category: The Scientific Literature
January 7, 2003
There's a report today that British authorities have rounded up several terrorist suspects in London - and that they had small quantities of ricin. So, what is the stuff, how bad is it, where did they get it, and what did they plan to do with it?
Ricin's a protein from castor beans - yep, the same ones used to prepare castor oil. The parent plant is sometimes used as a warm-weather ornamental, and used to be an industrial crop. The leaves aren't a problem, but the beans contain up to 5% ricin, which is a rather high yield for a natural product. It's quite toxic, although there are certainly worse things out there. Botulinum toxin, for example, is a thousand times more potent, but you can't grow anerobic bacteria very well in your back yard.
The purification methods for ricin are in the open literature, and aren't particularly challenging. It's probably one of the easiest toxins to isolate. For that matter, you can order various forms of it from biochemical supply houses. I looked at a few catalogs today, and it's quite cheap, by the standards of peptidic natural products (which are usually priced rather steeply.)
And what does the stuff do? Briefly, it's a very potent inhibitor of protein synthesis, which it accomplishes by attacking one subunit of the ribosome (the central RNA-to-protein machinery of the cell.) Rather than just binding to ribosomes and gumming them up, ricin is actually an enzyme all by itself. It tears up a specific adenine base in the ribosomal RNA, which disables the whole thing, and then it moves on to the next ribosome. One ricin molecule can turn over thousands of times, and needless to say, a cell can't lose thousands of ribosomes and expect to survive.
Ricin's a reasonably large protein, and it suffers from the defects of large proteins. The least dangerous way to be exposed to it is by eating it, since most of it gets digested, and much of the rest has trouble crossing from the gut into the bloodstream. In rodents, oral dosing is about 4000 times less potent than inhalation, which is the worst way to be exposed. The assumption is that if ricin were weaponized, it would be treated like anthrax spores and dispersed for maximum effect. The US and Britain carried out research that led to a prototype of a ricin bomb during World War II, just another one of many nasty weapons that actually didn't get used in that conflict.
Needless to say, there's not a whole lot of public data on just how toxic ricin might be in that form, and it would certainly depend on particle size, static charge, and all the other variables we learned about during the anthrax scare. We have a single public data point about injected ricin, though: Georgi Markov, a Bulgarian exile who worked for Radio Free Europe. One day in 1978, he felt a sharp pain as a stranger poked him with the tip of an umbrella. He began to feel ill within a few hours, and three days later, he was dead. A small pellet containing ricin had been injected into the muscle of his leg, as it turns out, in one of the more exotic assassinations known to have been carried out by the KGB. The best guess is that at most half a milligram proved lethal.
Which sounds pretty bad - but consider that terrorists are unlikely to be able to give masses of people intramuscular injections. And if they want to use inhalation, which is certainly the way to cause real damage with the stuff, they're faced with manufacturing problems similar to the use of anthrax spores.
It's not particularly water-soluble, so dumping it into a reservoir would be a waste of time. And adulterating food would be almost useless, although I've seen mentions of this possibility since the news story broke today. It takes a good handful of the beans themselves to kill an adult (and they have to be crunched up, too, because whole beans tend to pass unchanged through the digestive tract.) A back-of-the-envelope calculation for the pure toxin suggests that it would take a gram or two to reliably kill someone by ingestion. That adds up to a few hundred casualties per pound of ricin, but only if you can get all your victims to eat enough of it.
How worrisome is the news from London? It depends on how much ricin these people had, and what form it was in. I'm betting that it was straight precipitate from the beans, and not something ready to disperse for inhalation. In which case, the suspects were set up to commit retail murder. And not wholesale, fortunately.
(For as much detail as anyone could want, see this PDF, a book chapter written by two colonels from the Army's Medical Research Institute at Fort Detrick.)
+ TrackBacks (0) | Category: Chem/Bio Warfare
January 6, 2003
I've been staying away from all the Clonaid / Raelian hoo-hah. As soon as I realized who was behind this, I rolled my eyes and braced for the worst. I first read about the Raelians in Donna Kossy's extraordinary book Kooks (which I see is now in a second edition, which I must purchase very soon indeed.) With that as background, it's hard to take anything these people say seriously.
My opinion of the human clone claims can be easily expressed: bullshit. Look, you fools: extraordinary claims require extraordinary evidence. Come up with multiple blood samples now for DNA microsatellite analysis, in full view of multiple witnesses, or shut up. This is an important issue, and watching all of you hit each other with pies and try to cram yourselves back into the midget car isn't very instructive.
There. I feel better now.
+ TrackBacks (0) | Category: Current Events
Dwight Meredith over at PLA pointed out to me that the UC-Davis study on the prevalence of autism in California is online. It hasn't been published in a journal yet, and the JAMA paper I mentioned last week doesn't reference it. But the editorial comment in the same issue does.
As it should, since there's certainly an issue to be resolved. The Davis authors feel that their evidence makes it more likely that autism is actually increasing, even after correcting for wider diagnostic criteria, and so on. They still couldn't correct for all the potential differences in case finding, though, and it's unknown how much this has affected the final numbers. The JAMA editorial points out a recent paper analyzing the same California data which concluded that "diagnostic substitution" had occurred - a decrease in the "mental retardation" category had been taken up by an increase in the autism category.
Dwight's view, I believe, is that there has indeed been a real increase in autism - although short of the epidemic that some in the press have spoken of. I look forward to seeing how more data prove or disprove this - if there really is an increase, it's a tragedy, of course, but it could also provide a rare chance to uncover some important facts about the etiology of the condition. You don't get many good shots at the causes of a complex syndrome like this.
I think that's one reason the thimerosal provision that worked its way into the Homeland Security bill upsets me. Unlike many, I don't see it as evidence of a conspiracy to cover up wrongdoing (although one of the worst parts is that it provides spectacular ammunition to those who do.) I think that the less political maneuvering and grandstanding there is on this topic, the better. Most things would be improved that way, come to think of it. I did a quick Google search while writing this post, and since it had "autism" as a search term, up popped a sponsored link on the right-hand side of the page: "Child vaccines are linked to autism. Free case review by our lawyers." I'm glad these guys are so certain.
It's going to be hard enough to figure all this out without all the bricks flying through the air. As I've said, I think that thimerosal is a red herring. But if autism really is on the increase - and I'm still on the fence about that - then finding the real cause would be the most important research priority in the whole field.
+ TrackBacks (0) | Category: Autism
January 5, 2003
We spend a lot of time in drug discovery thinking about ratios. As we accumulate data about our compounds, we start ranking them by how selective they are - "This one's 10x versus the other receptor subtype and that one's 50x," you'll hear someone say, or "We've got to get compounds at least 100-fold over that other enzyme or side effects are going to kill us." Generally you have several secondary assays that the compounds have to jump through along the way, and the ratios are what everyone looks at.
And when compounds start to get dosed in animals, you try to look for cutoffs that can tell you which compounds are worth trying in longer assays. Maybe they only work, for example, when the ratio of peak blood concentration, Cmax (or time-averaged total exposure, AUC) to the binding potency is 100x or more. (Other things being equal, that means you could get the desired effect with a really potent compound that doesn't get into the system all that well, or a weaker compound that hangs around a long time.)
So, how good are these numbers? There's the problem - not as good as we tend to think. Even experienced medicinal chemists can get caught over-interpreting data when it's expressed in ratio form. The problem is, on a graph we all expect to see error bars (and we get pretty antsy if they aren't there - it means someone didn't run the experiment enough times, or they're sloppy about making their graphs, or they're trying to pull a fast one.) And for single data points from an assay, we try to remember the variability - looking at the various runs that went into the number you see is always recommended.
But when things get expressed as ratios, all that disappears. We throw around "40-fold" as if it's different from "20-fold," and it takes a conscious effort to remember that it almost certainly isn't. The variability of biological assays would completely curl the hair of a physicist or physical chemist - at times it curls ours in med-chem, and we're supposed to be used to it. Plus or minus 100% is considered a nice, tight assay for many systems - really, it is. They get worse as the system gets more realistic, too - cloned proteins are usually tighter than isolated ones, which are invariably tighter than cell assays, which are certainly tighter than tissue preps, and anything's less variable than some of the animal assays. If you have one of the jumpier ones in the denominator of your ratio, well, prepare to get all sorts of crazy results.
This is why no one, and I mean no one of any competence at all, really trusts "N of 1" data, especially if it's saying something interesting or unusual. If you haven't run the assay again, you're often better off not telling anyone about your numbers until you have. I have seen many people fall flat on their faces because they couldn't resist trumpeting some startling result that later turned out to be junk. It's tough, because we live for startling results. But we die by error bars, and they rule the drug discovery world in the end.
+ TrackBacks (0) | Category: Drug Assays
January 2, 2003
There's a new study out in JAMA (free full text here) on the incidence of autism in the US population. Before getting to what the article actually says, it's worth seeing what the media are saying it says. The New York Times headlines it "Study Shows Increase in Autism", and Yahoo runs it as "Study Confirms Marked Rise in Autism."
The AP ("CDC Study Finds Autism To Be Less Rare") and Reuters ("Atlanta Study Finds Rise in Autism Diagnoses" do better. That's because it's very hard to tell if there's a real rise taking place or not. The numbers are going up, but the interpretation isn't as easy as it sounds. To quote the authors:
Debate continues about whether the overall prevalence of autism has increased or whether past rates underestimated true prevalence. This debate is difficult to resolve retrospectively.
It's difficult for these reasons:
In the United States, the increase in the number of individuals receiving services for autism may be attributed to several factors. Changes in diagnostic criteria have expanded the concept of autism to a spectrum of disorders. Heightened public awareness of autism also has had an effect, due in large part to efforts of parent and advocacy groups, availability of more medical and educational resources, increased media coverage of affected children and families, and more training and information for physicians, psychologists, and other service providers. Also, in 1991, the US Department of Education added autism as a category for special education services, possibly leading to increases in the number of children classified with autism because of the availability of these services. The mandate for early intervention services for children with DDs, including autism, also has contributed to greater attention being placed on autism. At the same time, studies are suggesting that some children with autism respond well to early, intense educational intervention. The combined influence of these factors has probably contributed to the identification of more individuals with autism. However, it remains unclear whether specific environmental, immunologic, genetic, or unidentified factors also have contributed to these higher reported prevalence rates.
I think that's a very fair statement. Some environmental factor might be at work, if the increase is a real one. If so, tracking it down is going to be a major undertaking, because instead of one smoking gun, there might be several - insufficient by themselves, but adding up to something. These things are very hard to unravel, because it's almost impossible to get statistically meaningful samples that represent the range of variables that you're trying to check.
As for clues to any environmental causes, one footnote from the article that I have access to is from 2000, in Environmental Health Perspectives (Medline abstract here.)
Epidemiologic studies indicate that the number of cases of autism is increasing dramatically each year. It is not clear whether this is due to a real increase in the disease or whether this is an artifact of ascertainment. A new theory regarding the etiology of autism suggests that it may be a disease of very early fetal development (approximately day 20-24 of gestation). This theory has initiated new lines of investigation into developmental genes. Environmental exposures during pregnancy could cause or contribute to autism based on the neurobiology of these genes.
I find this idea somewhat more plausible than the thimerosal hypothesis, or any other environmental factor acting in the first years after birth. For what it's worth, it makes more sense to me that any causative agent would be something that's present in very small amounts, which would overall have greater leverage to cause broad-based harm earlier in brain development. That's an Occam's razor approach, which doesn't always cut the right way, but that's how I'd call it now - if there's an environmental cause at all, and if there is indeed a rise in autism. And we're still not sure about either one.
+ TrackBacks (0) | Category: Autism