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: firstname.lastname@example.org
January 31, 2006
Talking about reaction-quiz job interviews the other day really seems to have brought out a lot of stories (check the comments to that post). I'm glad to hear that this kind of interrogation seems to be disappearing.
I had more than one experience with this sort of thing when I was on the interview trail. Probably the worst was a lunchtime interview back in 1989 with a director-level person who looked over my background and said "Hmm. . .post-doc with Bernd Giese. . .free-radical chemistry, eh? So, tell me, who would your ten favorite free-radical chemists be?"
The first thought that went through my head was that I didn't even know of ten free-radical chemists. Favorites, my foot. I probably would have been better off saying that, although maybe not in as many words. But instead, I thought furiously, stalling for time: "Well, let me see. . .there's X. . .and Y has done some interesting stuff, although I'm not sure that he'd agree to being called a free-radical chemist. . .and. . ." I ran some time off the clock like this, and was beginning to panic a bit at the shortage of further names that were coming to mind, when I was interrupted by my interviewer. I was foolish enough to be grateful.
"Hmm. . ." he began again, "I see that you mentioned Z on your list. I don't think I would have ranked him that highly, personally. What recent work of his impressed you?" Cursing foully but inaudibly, I stammered out the only paper of Z's that I could even think of (which was indeed recent, and was the only reason I remembered him at all). "I'm not familiar with that one," came the response. "Could you draw those reactions out for me?" Over came a paper napkin for my use. Where the hell, I wondered, was my artichoke and fennel pasta, and could I perhaps fake choking on it?
So there I was, trying to remember and explain a paper I'd looked at once, by someone whose work I didn't follow, wondering all the while with the spare 1% of my brain how I ended up fighting on this particular battlefield. The rest of the interview proceeded along similar lines, and I gave what was surely one of my least impressive performances. I mean, I've had interviews where I knew, mid-way through the day, that I was either going to get an offer or something was so seriously wrong with the place that I wouldn't want to work there anyway. But I lurched out of this one thinking that they'd have to have something seriously wrong with them to call me back. Which indeed they didn't.
+ TrackBacks (0) | Category: How To Get a Pharma Job
January 30, 2006
Business Week, whose reporting on the industry I've been uncomplimentary about in the past, is saying that year-end earnings reports show that European drugmakers are moving past the American-based ones in profitability.
And that's probably true, for 2005 anyway. Merck (and to a lesser extent, Pfizer) got hammered last year, and several other companies turned in pretty unremarkable performances. But the comparison that the magazine is making is, for the most part, an artifact. US companies had some big sellers a few years ago, and the European companies looked like they were falling behind. Now the patent expirations are looming (and some bad news has hit), so the situation is reversing. It happens.
It doesn't make much sense to divide drug companies up just by location of their headquarters in the first place. GlaxoSmithKline, for example, is a British company, but their culture seems more or less identical to that of the American companies. And I'd have to think that Novartis would have more in common with another big outfit like Merck than it would with a smaller European company. I'm not saying that the European companies don't tend to have a different way of doing things, because on average I think that they do. It's just that the comparisons aren't as easy to do as you'd think.
The BW piece goes on about various drug recalls on this side of the ocean, but they make things look more dramatic by not including some of the other European companies like Bayer (which went through a painful recall of its statin drug a few years ago). The article also makes much out of Sanofi-Aventis and their prospects for Acomplia (rimonabant) this year - but how would Aventis look if they were still on their own? That merger issue makes the comparison with the US companies rather complicated, too, since Pfizer's biggest drugs were acquired through buying other companies.
The article doesn't try to draw many larger conclusions about how the European companies might be doing things differently or how the American ones got into trouble. And it's a good thing, too, because it's hard to come up with any generalizations that make sense. (You might be able to sell a specialty-markets versus direct-to-consumer difference, but the piece doesn't bother trying). What readers are left with some sort of weather report: "It was cold a few days ago, but now it's not. Today it's raining, but tomorrow, who knows?".
+ TrackBacks (0) | Category: Business and Markets
January 29, 2006
Here's a question for people who have been out interviewing in the last couple of years (with special emphasis on those who've been seeking their first position in the industry): does anyone still do the old "Let's go to the blackboard and see how many reactions you know" interview? Syntex in Palo Alto was well-known for putting people through this twenty years ago or so.
I got a couple of those back when I was first looking around, and I'd be lying if I tried to tell you that I liked them. (I did OK under these conditions, but while you're up there on you're always worrying that the next question could be the one that makes you look like an idiot. The whole thing is too close to a PhD defense to be anything but unpleasant.
And I never understood the point of this technique to start with. Chemistry is not done solo in a spotlight in front of a hostile crowd. It's important to be able to speak coherently in front of people, of course, but that's what a seminar (and its associated questions) are for. It's also important to realize that a person who can whip out answers quickly might be an even better scientist if they trained themselves to slow down and think a bit.
Knowing the mechanism (under time pressure) for the Nosenko-Golitsin Rearrangment doesn't seem to be a good predictor of research success. That was true back in the late 1980s, when I was the guy at the blackboard, and I think it's even more true now. SciFinder and the other computer databases have really gutted the rationale for committing large numbers of transformations to memory. It's important to know what sorts of things can be done, but there's little point in memorizing exactly which reagents do them and what journal they appeared in. You can always look that up, and when you do you'll likely find a dozen alternatives that you didn't even know about.
These days, being able to sort out all those potential reactions into an order of plausibility is a more important skill than just memorizing them. I think I might work up some questions like that for the next time I interview someone. "Here," I'll say, handing over a sheet of paper. "SciFinder says that you can do this reaction any of these six ways. Which one would you recommend trying first, and why?"
+ TrackBacks (0) | Category: How To Get a Pharma Job
January 26, 2006
Since I was mentioning the other day how little I enjoy reading (or writing) patents, I thought I'd pass along this item from Greg Aharonian's PatNews mailing list. There are plenty of other people, it seems, whose attentions wander while doing IP work. Take, for example, the luckless 3 inventors from the New York area who filed application US 20030004652, which published about three weeks ago.
They've got a (putatively new) system to monitor animal behavior during drug testing. The abstract starts off by describing:
A system and method used to assess animal behavior includes a module having sensors that collects a variety of physical and biological data from a test subject. Interpretation of the data is provided to assess the test subject's behavior, neurology, biochemistry and physiology. The module is useful in observing the effects of a drug on the test animal and providing information on the drug's signature.
OK, fine. I'm pretty sure I've seen things like this before, but who knows, maybe they have something inventive in there. So, you're asking, what are the wonderful features of this new invention? As you'll see, the folks who drafted (and edited) the abstract were asking themselves the same question during the document's preparation. Verbatim, it continues with the following, whose unsteady grammar is the least of its problems:
Another advantage is module's portability that allows it to be used in standard laboratory cages. (NOT SURE ABOUT THIS PORTABILITY) This portability allows the animal to be tested in its own habitat, that can reduce any erroneous data due to stressing the animal when removed to a test cage.
Proofreading, guys, proofreading. You're going to have a tough time with the novelty and enablement requirements if you tell the patent office what you really think, you know.
+ TrackBacks (0) | Category: Patents and IP
January 25, 2006
I'm starting to wonder how we're going to get many drugs even as far as the clinic these days. I don't think I'm alone in feeling that it's just one thing after another.
Modern mass spectrometry techniques can find all sorts of interesting metabolites for your compounds after they've been dosed in animals. If you look hard enough, you're likely to find some structures that you don't want (epoxides, quinone precursors, and other potentially toxic beasts). In the old days, we'd never have seen them, but now we get a chance to worry about them.
There's always our good friend the Ames test to make you hold your breath. It's true that not every Ames positive compound is going to be a mutagen in humans, but it's also true that no one is going to develop one because they happen to feel lucky, either.
Anyone who doesn't hold their breath when their drug candidate goes into two-week rat testing hasn't been in the business long enough. Even if the rodents act normally during the whole thing, you can't breath normally again until the tissues are examined by the histopathologists, Perfectly reasonable-looking rats can be hiding amazing things at the microscopic level.
And for the last few years we've had to worry about QT prolongation. That's a cardiac side effect that can lead to very serious trouble - sometimes. And we have some ways to check for it early in development, which work - sometimes. The list of drugs that came to the market before we realized all this is rather impressive.
All this is why I wonder when I read about small startup companies with three or four compounds in clinical trials. Have they really jumped through all the hoops that a big company has the resources for? And if not, how much worse off are they for it?
+ TrackBacks (0) | Category: Drug Development
January 24, 2006
PCT (Patent Cooperation Treaty) applications are published with an "International Search Report" appendix document. This is done by one of the larger patent offices from the countries in the PCT, and is a preliminary assessment of the patentability of the claims in the application. The searchers try to identify other documents in the patent or open literature that might bear on their novelty or scope.
It would be difficult to pay me sufficiently to do this for a living. I've been involved in many in-house patentability and claim-writing discussions, and they can be agonizingly tedious and frustrating. That's why I've never been able to understand how some pharmaceutical applications are written. They start off with chemical claims whose first generic structure seems to take in the vast reaches of interstellar space. Then the next fifty or sixty claims narrow this down in gratingly tiny steps to oh, about the edge of the Kuiper belt. Then comes a recitation of six hundred and seventy-eight specific names of compounds that are specifically claimed, and on and on.
I have trouble imagining how anyone can go to such pains, and I have trouble imagining why they bother. After all, if you don't enable these claims (patent-speak for "show that you actually made something that fits that description"), they're not worth all that much if it comes down to a fight. And I've seen many claims that couldn't be fully enabled short of putting five hundred people to work on them full-time for about ten years.
For those of you with access to patent images, try the generic structure in US6214850 on for size. It's very concise, as these things go, but rapidly expands like some sort of mutant fungus. For a wordier example, try WO2002042272. This jewel has 139 pages of claims in it, and the number of R-groups that show up in its laughable generic structures goes up into the hundreds.
The whole purpose seems to be to confuse and irritate anyone who tries to read the claims - and to be fair, that's very likely just that the purpose is, since that makes it harder for someone else to figure out what the real subject of these hay-bale-sized applications actually is. But the poor souls at the search agencies are paid to do that, and every so often they lose it. I just came across a search report which starts out with this cri de coeur:
"In these claims, the numerous variables and their voluminous, complex meanings and their seemingly endless permutations, makes it virtually impossible to determine the full scope and complete meaning of the claimed subject matter. . .it is impossible to carry out a meaningful search on same."
+ TrackBacks (0) | Category: Patents and IP
January 23, 2006
Tuesday and Wednesday of this week are the days when Merck's attorneys are scheduled to question Dr. Gregory Curfman of the New England Journal of Medicine and another unnamed staffer. The journal's handling of the published Vioxx data (the VIGOR study) looks as if it could be a key part of the next trial, which will be in lawsuit-friendly Rio Grande City, Texas, with jury selection beginning this week. Note that the late plaintiff had taken Vioxx for all of seventeen days.
Naturally, both sides are saying that they expect the facts to help them out. The plaintiffs will argue that the recent editorial in the NEJM confirms that Merck was trying to hide the risks of Vioxx, and Merck will argue. . .well, they might argue several things. For one thing, they'll try to show that the journal's suspicions about altered data are unfounded. But I also have to wonder if my theory from last month is going to get aired out, too. Will Merck's attorneys ask if the journal has been approached by representatives of the plaintiffs?
We'll have to wait until the the next trial to see if any of this makes it into the courtroom. For now, I think it's interesting and unusual enough for Merck to be putting a journal editor under oath. They must feel as if they've got a pretty good reason for doing it. And they'd better. . .
For pre-trail reading, here's a law professor from Fordham who says the whole spectacle isn't making our liability law system look very good.
+ TrackBacks (1) | Category: Cardiovascular Disease
January 22, 2006
Some readers will have heard that Michael Fumento, well-known pundit and controversialist, has been fired from his post at Scripps-Howard. Eamon Javers of Business Week broke the story of how Fumento had received some $60,000 from Monsanto without disclosing the relationship. If you look in the comments after that article, you'll see a lot of arguing about whether this was money for a book, for op-eds, whether it went directly to Fumento or not, and so on.
The problem is, none of that matters much. This column by Cathy Seipp is a good explanation of why that is. If you're going to set yourself up as a journalist, you can't take money from the people that you're going to be writing about. Even if there's room to argue, you have to at least disclose what's going on.
Update: here's Fumento's published response to Seipp and the Business Week article.
Fumento has been a vigorous defender of the pharmaceutical and biotech industries over the years, of course. And according to Seipp, there had been suspicions that he had been well-rewarded for his enthusiasm. "Has anyone read a pro-biotech piece. . .that doesn’t practically smell like solicited spin?" she asks.
Well, yeah, I have, but it's one that I wrote, so maybe that doesn't count. I'm in an odd position here, because the journalism is a nights-and-weekends thing with me, while the day job is in the very industry that I write about. That gives me some unique strengths - I like to imagine, anyway - but it also means that anything I write can automatically be discounted as the product of someone who's already been paid for.
That's why I spend little or no time talking about the products of the Wonder Drug Factory where I work, because
you can count them on one hand that wouldn't be right. I'm definitely not a spokesman for the company where I work. But for the industry in general? Yeah, informally I guess I am, because I find the work interesting and important enough to speak up for it.
That doesn't mean that I get paid extra for that role, and it doesn't mean that I like everything that goes on in the drug business, either. It is, after all, a business, not some sort of higher calling. Scientific research might qualify as a higher calling, under the right circumstances, but that's only one part of the drug industry. I'm just fortunate that doing the science overlaps as well as it does with making a living.
So there's my disclosure statement. I'm never going to do the under-the-table pay-for-punditry thing, because where I stand should already be clear. Where a byline of mine appears will be where the money's coming from. I'm pro-research and pro-pharmaceutical, because that's what I do for a living. But it isn't love, and it isn't blind.
+ TrackBacks (0) | Category: Press Coverage
January 19, 2006
I currently have a piece up on the Medical Progress Today site, about what an 18th-century minister has to offer modern clinical trial design. (Statistics groupies will have already guessed the subject matter from that clue)!
+ TrackBacks (0) | Category: Clinical Trials
So, are the neuraminidase inhibitors (Tamiflu and Relenza, oseltamivir and zanamivir) going to be of any use against bird flu? The press coverage is a mass of confusion. Headlines range from stuff like "Newer Flu Drugs Work Better" to "Don't Use Older Flu Drugs, Experts Warn" to "Tamiflu Not Effective Against Bird Flu". The problem is, we're looking at two different classes of drugs, and two different sorts of flu.
A few days ago the CDC issued a warning that the prevalent H3N2 flu strains this year have mutations that make the older class of flu antivirals (the aminoadamantanes) ineffective. But I'm not sure how many people even get these any more, since they were never very impressive to start with. I don't think anyone has seriously proposed them as a defense against H5N1 avian influenza, should that ever take off.
But Tamiflu and Relenza are a different story; a blizzard of hype has surrounded their possible use. Lost in the bird-flu noise is their use against "regular" influenza, a market where they've never performed up to expectations - no doubt they're selling rather better so far this season - and the CDC recommended their use for this purpose.
Comes now a report in The Lancet from a group in Rome (available to subscribers here), looking over all the published studies on the various drugs. They also recommend that that adamantanes be retired, and they aren't very positive on the use of the NA inhibitors against the standard forms of flu, which I'd say is in line with the clinical experience. And they found no evidence that either Tamiflu or Relenza is effective against bird flu, which leads to all the jumpy headlines.
But this isn't really a surprising finding, since there hasn't been (to my knowledge) any published study on the use of the drugs against avian influenza in humans. (Cell culture, yes, but that's a long way from the real world). There hasn't been enough time (and there haven't been enough patients, fortunately). A better headline would have been "Tamiflu's Effectiveness Against Bird Flu Unknown", but we already knew that. Didn't we?.
+ TrackBacks (0) | Category: Infectious Diseases
January 18, 2006
I've just spent some time reading a very enjoyable and interesting paper (PDF, and thanks to Tyler Cowen at Marginal Revolution) from Shane Frederick at MIT's Sloan management school. He has a simple test that seems rather well correlated with a person's appetite for financial risk-taking and their ability to postpone a smaller immediate reward for a larger one in the future.
Frederick's test consists of what are basically trick questions. They're the sort of things that have an immediately obvious intuitive answer, but one which is (unfortunately wrong). These take a bit of mathematical and logical thinking to work out, but nothing advanced. You do have to be able to not just run with the first answer that comes in your head, though. (Not doing so, of course, requires you to be the sort of person who double-checks things, preferably from a different angle, before committing to them). Cognitive ability and patience have been linked before, both in the popular imagination and in a few studies that Frederick cites.
Update: there's a possibly confounding variable that I forgot to mention: perhaps the better students taking this test had already been exposed to these types of questions before as brain teasers. I know that this was the case with me; I recognized them as classic forms of not-the-obvious-answer questions. This gets back to the question of how much a person's test-taking performance is due to practice in taking tests. . .
He gave this test to a number of different groups, and his table of results is worth the download right there (I'll give you a hint - if you didn't know already, MIT is quite different from the University of Toledo). But as it turns out, the people who score very low and the ones who score very high on this sort of quiz also answer quite a few other questions differently. Frederick was checking their responses to choices such as "Would you rather have $100 now or $140 next year?" and "Would you rather have $100 for sure or a 75% chance at $200?". What he found was the high-scoring group heavily prefers to wait for the larger payout in the first question, and heavily prefers to take the risk in the second one. (The whole paper details a range of these, of varying levels of risk balanced with immediate or long-term attractiveness).
Another effect that's been noted in past studies is that people are much more willing to take a risk to avoid a loss, rather than taking an equivalent risk when there's a prospect of an equivalent gain. Update: Although this was what I had in my head, I originally had this backwards; thanks to Shane Frederick for pointing this out! His low-scoring cohort is hugely biased toward this mode of thought, Frederick finds, but the effect actually disappears in the high-scoring group. (He also confirms the results of several other studies, including the finding that women tend to be much more risk-averse than men in such situations).
I couldn't help thinking that his high-scoring group is also the group that makes the best scientists. Think about it: not going with the first thing that pops into your head, but always stopping to ask yourself if it's true or not. Checking it in different ways to see if you get the same answer. These are the habits of mind that a good researcher has - and I can tell you from personal experience that some of the least competent chemists and biologists I've known come from the opposite category.
You know the ones - the folks who get an n-of-one number in an assay and go running around telling everyone that they've found something wonderful, only to have to eat the whole thing (again) when it doesn't repeat. The set-up-the-reaction-first (and look in the library later) folks who have to pour even more reactions into the waste can than the rest of us do. Professor Frederick should run his next survey in the science hallways - perhaps we could separate the sheep from the goats.
+ TrackBacks (0) | Category: Who Discovers and Why
January 17, 2006
Many of my readers are professional chemists, but many aren't. For those who've only had sophomore organic (or less), it sometimes comes as a surprise to find out that we actually use a lot of those reactions that you have to learn.
That's a strange thing about chemistry as compared to most of the other hard sciences. I don't think that much that you learn in a second-year physics class comes up in the day-to-day life of a physicist, and the same goes (and how) for most biologists. But I'm here to tell you: we professional organic chemists really do Fischer esterification, the Knoevenagel condensation, and a lot of those other ancient reactions. Not just once in a while, either - we do 'em every day of the week. My lab, for example, has been wrestling the last few days with forming and using a particular Grignard reagent, and Grignards have been a standard part of undergraduate labs for generations.
Why do we stick with these moldy reactions? Because they work, for one thing. Reducing an aldehyde with sodium borohydride, for example, is a procedure that's been around for a good forty years. But it's a mighty rare aldehyde that won't reduce cleanly with the stuff. It's fast, it's cheap, and it's generally easy to clean up the reaction, so why not?
Another reason is that these reactions are close to the fundamental principles of organic chemistry. If you're going to make an ether from an alcohol, for example, it's hard to see how that's going to happen without an oxygen attacking a carbon center somewhere along the line. I mean, you're forming an oxygen-carbon bond, so you can't avoid it. And oxygen is so electronegative, you have to figure that it's going to have some sort of negative charge built up on it, so it's going to be attacking something with a partial positive charge. . .and there's the good ol' Williamson ether synthesis, the classic nucleophilic substitution. To have a completely different ether synthesis, you'd almost have to have a completely different form of oxygen.
And that brings me to an observation that I've made before - that you can be a fine medicinal chemist using nothing but reactions from a sophomore organic textbook. It's a bit humbling to realize that, because catching on to that fact tells you where chemistry stands in drug discovery: not as an end in itself, but as a means to an end. And if those means turn out to be reactions that an eighty-year-old grandmother could learn to run, and that are older than she is on top of it, well, fine. We're not here to use the latest hot reaction, unless it can speed up making a drug. Because that's the point.
+ TrackBacks (0) | Category: Life in the Drug Labs
Not being a world-famous scientist, I wasn't asked to contribute to the Edge.org "Dangerous Ideas" festschrift (see the post below). Of course, I also don't have their founder John Brockman as my literary agent, either. . .but if his people (or any other good agents!) call, I'll definitely pick up the phone.
But I'd like to start a related discussion over here. I'm inviting comments on what people think the most dangerous idea in drug research might be. I realize that we may not all assign the same value (or even sign) to "dangerous", but let's see what happens. Update: What I mean by that is, do we call an idea dangerous because it's false, and its adoption would be damaging to drug research? Or do we mean dangerous, in the sense that it's something that's true but too unpleasant or politically difficult to face? I'll take suggestions in either category. . .
I'll start things off with this one: There are diseases that aren't worth the money it would take the treat them.
+ TrackBacks (1) | Category:
January 16, 2006
In all the server-migrating of the last week, it appears that some of the recent comments have been hosed. I just wanted to let everyone know that it's unintentional - hey, I even let the comments that show up on the Intelligent Design posts around here stand. I also think I've fixed the "comment held for moderation" problem that was afflicting all of us, so if it happens again, please drop me an e-mail.
In other blog-news, I'd like to take this opportunity to welcome the advent of the Scienceblogs site, which incorporates several of the blogs over there on the left-hand side of the page. I'll be updating the blogroll shortly.
There's an interview with me up over at the financial site The Stalwart. Unfortunately, I don't think I've revealed enough to make anyone rich, but do write if I turn out to be wrong about that.
Finally, via Chad Orzel's new Uncertain Principles site over there, I found this useful companion to the Edge.org "Dangerous Questions" page. As you'd figure, not all of the submissions there are of equal quality, so these guys have taken the trouble of digging out the best ones.
+ TrackBacks (0) | Category: Blog Housekeeping
January 15, 2006
The FDA has released some new guidelines intended to make it easier to get drugs into the earliest stages of human trials. Considering how often we fail at those stages of development, anything that increases our number of shots on goal is good news.
The first part of the new approach is a set of official standards for what are known as "Exploratory INDs". (An IND (Investigational New Drug) application is the mound of paperwork that you have bring to the FDA in order to dose human subjects). What they're doing here is making it easier to test limited numbers of people at smaller doses, just to see if you can narrow down your drug candidates with some realistic data. This "microdosing" approach has been tried in Europe for the last few years, and it's definitely time the the agency laid out the rules for its application over here.
I think that for the most part the key data obtained in these trials will be pharmacokinetics - you could take three or four roughly equivalent drug candidates in and see which one had the best blood levels and distribution before committing yourself. (Fancy mass spectrometry techniques can allow you to see very small amounts of compound or metabolites in samples these days). This approach will also work out well if you have some sort of clinical marker that you trust. Sub-effective doses of the drug candidates could still point you to which ones are having the effect you want, if the signal-to-noise ratio of your surrogate marker is good enough. Imaging studies would be another good use for these trials.
If I'm reading their guidelines right, the FDA can let you go into human microdosing with only rat data. That sounds either exciting or alarming, depending on your point of view, but given the low doses involved, I think it's a good thing. For dosing up to pharmacologically active levels, though, they want the traditional rat/dog safety package, which is certainly reasonable.
The second part of the new standards are for manufacturing drug substances for these early trials. What they're doing, it seems, is loosening up the "Current Good Manufacturing Practice" rules a bit for small exploratory studies. This is mostly going to help out the smaller companies and (especially) academia. CGMP is no fun to follow if you're not set up for it, and it is rather odd to make the five-gram-batch folks jump through almost the same hoops as the five-thousand-kilo people. For the larger companies with lots of CGMP capacity, though, I don't see this change making much of a difference.
It'll be interesting to see what sort of coverage this gets in the press - assuming it gets much at all, that is. There's room for a sufficiently motivated complainer to go on about the FDA loosening up on its regulation of unknown toxic drug substances, poisoning America, slippery slope, etc. If any readers spot someone taking this tack, forward the reference to me, and I'll reward you with a laudatory mention here. It'll look great on your CV. Trust me.
+ TrackBacks (0) | Category: Clinical Trials
January 13, 2006
We've had a few incidents recently at the Wonder Drug Factory where people have been using some common solvents like ether or tetrahydrofuran (THF) and ended up with odd results. When they check their reactions, there's something else in there that hasn't turned up before. The same impurity turns up in completely different reactions, too, which narrows the possibilities down a lot. In a couple of these cases, the chemists involved went to the trouble of isolating this pesky impurity and getting NMR spectra of it.
The experienced chemists in my audience are already cringing; I can feel it. No, we didn't blow anything up. But the people involved are now the proud owners of clean NMRs of ether peroxides. These explosive little beasts are an unavoidable byproduct of storing ethereal solvents where ambient oxygen can get to them. Oxygen is just too reactive - which is fine for us, biochemically, since that keeps us alive, but it can be a real nuisance in other situations.
These solvents are usually sold with some inhibitor added, a free-radical sponge like BHT, for example. But over time - or if someone in the supply chain stored things improperly - this will get used up, and then peroxidation moves right along. In extreme cases, such as with the unstoppable di-isopropyl ether, you can even get crystals of the peroxide coming out of solution. I have never seen this in person, and I will be very glad if I never do
Biologists and physicians have, among chemists anyway, a reputation for treating bottles of ether much more cavalierly than we do. A colleague of mine witnessed this at first hand at a former company of hers. A note had gone out to all the departments to check for old ether bottles, went into the possibility of crystal formation, and told everyone to notify the haz-mat team if any bottles were uncovered. In the molecular biology department where my source was working, one of the lab heads promptly marched out and rooted through the cabinets, emerging a few minutes later with a can of ether of uncertain age. This person then held the can up to his ear while shaking it, listening to see if any solid material was sloshing around in there. Which is one way to find out.
+ TrackBacks (1) | Category: How Not to Do It
January 12, 2006
One of the glowing chunks of fallout from the South Korea stem cell scandal is the handling that Science gave to Hwang's two retracted papers. It's no secret that the top journals compete to see who can publish high-profile papers, and for research like this it's pretty much down to Science or Nature. As fate would have it, Nature got the one of Hwang's recent papers that actually doesn't have to be retracted - whether that's a coincidence or not, I don't know. But the scientific publishing world is no doubt speculating about that, and you can already see some uncharitable comments surfacing. For example:
"It sounds as though their processes were rather sloppy," said Dr. Benjamin Lewin, the founder and former editor of Cell, a biology journal known for its rigor. "At a minimum, Science should have been more careful and should never have reached the stage of publishing a paper with identical photos," he said, referring to the fact that some photos of cell colonies in Dr. Hwang's 2005 article were duplicates of one another.
Dr. Lewin said that a journal editor needed to develop an intimate knowledge of his reviewers' strengths and weaknesses, and that "Nature and Science don't have the reputation for rigorous review."
Meow! I have a lot of respect for Lewin, but this sort of thing isn't going to burnish his reputation. (Update: His deep admiration for the folks at Science turns out to go back a long way). The response of my fellow scientists has been to suggest that he try publishing in one of those non-rigorous journals under a pseudonym and see how far he gets. I guess it depends on how you define "rigorous", though, and it's true that there are other definitions besides rejecting 90% of all the papers submitted. The journal may well have rushed these hot papers through the process, since they were clearly the sort of thing that would be worth publishing. (See some thoughts on this here). The reviewers for both papers have no doubt been involved in some difficult e-mail exchanges in the last few weeks. . .
The danger with comments like this, though, is that every journal that publishes papers worth reading has published papers worth retracting. And that includes Call, naturally, although their most recent one was (weirdly) done without the lead author's consent. (Thanks to Ivan Oransky's blog at The Scientist for this). And in a fraud case that I missed last fall, Luk Van Parijs of MIT was fired after faking loads of data in several research papers. New Scientist found, while looking through his publications on their own, that one of his papers (in the journal Immunity) almost certainly contains faked flow cytometry data. The graphs shown are just too similar, and that's not a technique that churns out exact duplicates of anything, ever.
This is quite similar to the problem with Hwang's stem-cell illustrations, which Lewin is saying should have been caught. But Immunity, like several other single-noun-title journals is published by. . .Cell Press. That's the problem here. No one comes out of this business looking good, even if they try.
+ TrackBacks (0) | Category: The Scientific Literature
January 9, 2006
Update: Since the site was down most of Tuesday, I'm leaving this post up another day. Things have only worsened since I put it up, though. . .
I've been withholding my comments on the South Korean stem cell controversy, waiting to see how the story finally settled out. Well, it's good and settled now: the entire enterprise was a fraud. Here's a timeline of the whole sorry business, for people who need a recap. Start at the bottom of that page to experience it in the most painfully realistic way.
My first impulse, in the manner of anyone belonging to a group (biomedical researchers) whose reputations have been dented by such a case, is to point out that, yes, "the system worked". The fraudulent research was discovered and rooted out, papers were retracted, funding lost, brows slapped, all of it. And it hasn't taken that long, either. It's useful to point these things out to people who would like to throw mud on the whole enterprise of science.
See, for example, this blog review of a recent book on scientific fraud. Contrary to its repeated assertions, scientists do indeed realize that fraud happens, because every working scientist has seen it. For starters, most large academic departments have tales of grad students or post-docs whose work could never be trusted. And all of us in research have run into papers in the literature whose techniques won't reproduce, no matter what, and the suspicions naturally grow that they were the product of exaggeration or wishful thinking. The number of possible publications sins alone is huge: yields of chemical reactions get padded, claims of novelty and generality get inflated, invalidating research from other labs doesn't get cited.
It's painful for me to admit it, but this kind of thing goes on all the time. And as long as the labs are staffed with humans, we're not going to be able to get rid of it. The best we can do is discourage it and correct it when we can.
But takes me to the second standard impulse that strikes in these situations, which is to ask what in the world these people were thinking. That's what's always puzzled me about major scientific fraud. The more interesting your work is, the more fame you stand to gain from your results, the more certain you are to be found out if you fake it. There are obscure areas that you could forge and fake around in for years, and journals in which you could publish your phony results without anyone ever being the wiser. Of course, by definition those won't do you much good - heck, you might as well do real work by that point.
But faking the big ones, the worldwide-headline national-hero stuff - you can't get away with that for long, and Professor Hwang didn't. The closest parallels I can think of are the recent Jan Hendrik Schoen case and the thirty-year-old Summerlin mouse scandal. (These and several other infamous cases are summarized here and here. I honestly find it hard to believe that there are others of that magnitude that anyone got away with.
I've never been able to imagine the state of mind of someone involved in this kind of thing. There you are, famous for something you've completely made up. In front of you are the cameras and reporters, while behind you, off in the distance, are hundreds of other scientists around the world busily trying to reproduce your amazing results. Every minute, they get closer to finding you out. How can anyone smile for the television crews under such conditions?
It's tempting to speculate about the state of the Korean scientific establishment and the role of Korea culture itself in this latest blowup. But such things have happened everywhere. The Korean factor certainly led to Hwang being an instant national figure with his face on every magazine and a dozen microphones trained on him wherever he turned. But it's not a Korean failing that did him in, it's a human one.
+ TrackBacks (1) | Category: Biological News
January 8, 2006
Some time ago I wrote about some atoms that I wish I could use. There are still other molecular fragments that Nature has neglected to provide, though, and I'm adding to the wish list:
First off, I'd like some groups that are plain linear spacers of different lengths. Ideally, we could stick these onto carbons and heteroatoms alike. As it stands, the only commonly available group that does this is an alkyne (a carbon-carbon triple bond), and those come with their own baggage. While there are drugs that have these groups in them, they're typically treated pretty roughly by the liver enzymes. A metabolically stable alkynish thing, of variable length, would be a wonderful thing. People have made various weirdo spacers out of small bicyclic systems, true, but the synthetic routes to them are beastly, with no improvement in sight. Until we can source these things by the kilo, drug companies aren't going to be interested.
Next, I'd like some six-membered heteroaromatic rings with something other than nitrogen in them. As it is, pyridines, pyrimidines, and pyrazines are given a real workout in drug discovery programs, and we'd love to have some more options. Unfortunately, the fabric of the universe hasn't accomodated us. If you put an oxygen in there instead, it has to take on a positive charge, and that gives you a highly reactive beast called a pyrylium. It's rare that you can get one of those into a bottle, and even if you did, sending one in for biological testing would be grounds for a referral to the HR department. I see that some zanies have tried, though.
Sulfur can do the same lively thing, although I've never actually seen any of those, and there's probably a good reason for that. Varioius sorts of aromatic rings with a phosphorus atom in them are known, but they're cranky and exotic, like a lot of phosphorus chemistry. Actually, a lot of phosphorus chemists are kind of that way, too, in my experience. Like a lot of phosphorus compounds, those aryls probably reek to the skies, too. As for phosphorus chemists reeking, I'd say my personal data set runs about fifty-fifty.
And finally, I'd also like some big, lumpy polar atoms. As it is, if you want to put a single lunker of an atom onto a molecule, you're looking at a bromine substitution. Iodine's possible and even larger, but those compounds are usually too unstable to sunlight to make good drugs, unless you're doing thyroid receptor ligands, where you might have to have them whether you feel like it or not. It's true that a trifluoromethyl group is kind of like a big halogen atom, too. But all these halogens make your molecule rather greasy, which is something we'd rather avoid. Something the size of a bromine that could hydrogen-bond and help its molecule go into aqueous solution would be a big hit. Quantum mechanics, being perverse, has not obliged.
+ TrackBacks (0) | Category: Odd Elements in Drugs
January 5, 2006
There are plenty of links around the blog world to the Edge.org "Dangerous Ideas" symposium, and much of it makes for good reading. But there are some clinkers. I was going to take some time to disassemble this one from biologist Paul Ewald, but that effort has already been made for me.
Put briefly, Ewald is a believer in the more extravagant reaches of the "New Germ Theory", the idea that many more diseases than we now think are actually caused by infectious agents. But he also seems to believe that even if vaccines were possible for things like cardiovacular disease that no drug company would develop them. Y'know, so we could keep selling our regular drugs instead of curing diseases. But it never seems to occur to people who advance ideas like this that such a vaccine would make an overwhelming amount of money, and that we'd be very interested in it indeed. Think about how much people are willing to pay for, say, medication to lower their cholesterol or keep their blood pressure down. Think how much people pay yearly for insulin or for asthma medication - whole companies are founded on these kinds of franchises, even with fierce competition between drugs.
Now imagine how much people would pay to never have to do that again. Quite a bit, I'd guess, since you no longer have the disease and no longer have to take any drugs for it. And here's the real kicker: the company that comes up with this wonder cure would scoop up the revenue from the entire therapeutic area, because there would be nothing that could compete. No, I think that we'd be quite intrigued.
Now, I have to say that such treatments probably aren't going to turn out to be possible in most cases. I like a lot of the New Germ Theory stuff, but I'm not sure if it can be pushed this far - although I'd love to be proven wrong about that. But the best way to make sure that they don't happen is to remove that gigantic incentive. Any company that makes a serious try at something like this will be taking a very large, very expensive risk, and if they know that there's a patent seizure waiting at the end of it they may decide that the money is better spent elsewhere. It would get done, eventually, but the fastest way - if there is a way at all - is to let profit-minded companies scramble for it.
+ TrackBacks (1) | Category: Drug Prices
January 4, 2006
Via Tyler Cowen at Marginal Revolution I came across this post earlier in the year from a blog called EffectMeasure on the use of rodent models to predict human cancer risks. It's a broadside against the American Council on Science and Health and a petition they filed against the use of high-dose rodent carcinogenicity tests.
Quote the anonymous "Revere":
The main rhetorical lever ACSH employs is the use of high doses in the animal studies, doses that are much higher than usually faced by humans. But as ACSH knows well (but didn't divulge) there is a technical requirement for using these doses. If one were to use doses in animals predicted to cause cancer at a rate we would consider a public health hazard, we would need tens of thousands of animals to test a single dose, mode of exposure and rodent species or strain. This makes using those doses infeasible. Thus a Maximum Tolerated Dose is used, one that causes no other pathology except possibly cancer and doesn't result in more than a 10% weight loss. The assumption here is that something that causes cancer at high doses in these animals will also do so at low doses. This is biologically reasonable. It is a (surprising) fact, that most chemicals, given in no matter how high a dose, won't cause the very unusual and specific biological effect of turning an animal cell cancerous. Cancer cells are not "damaged" cells in the individual sense but "super cells," capable of out competing normal cells. It is only in the context of the whole organism that there is a problem. It is not surprising, then, that very few chemicals would have be ability to turn a normal cell into a biological super cell of this type. Estimates are that is far less than 10%, perhaps only 1% of all chemicals that have this ability. Thus western industrial civilization doesn't have to come to a screeching halt if we eliminate industrial chemical carcinogens from our environment.
We know of no false negatives with this process. Every chemical we know that causes cancer in humans also does so in rodents (with the possible exception of inorganic trivalent arsenic, which is equivocal). The reverse question, whether everything that causes cancer in animals also is a human carcinogen, is not testable without doing the actual natural experiment: waiting to see if people get cancer on exposure, an experiment ACSH is only too happy to conduct on the American people to make their corporate sponsors happy."
I've left out (as did the MR post) the part where he called the ACSH "right wing whores", which is the kind of thing that doesn't enhance the statistical arguments very much. Dropping the invective, I want to take up Tyler Cowen's question: is there anything to this critique? My answer: there might be. But there might not be. It's certainly not as clear-cut as the author would like to make it, cancer epidemiologist though he is, which would seem to be one of the criticisms he's making against the ACSH petition.
Here are some complicating details:
1. The effects of high doses of compounds can be due to their effects on cell division. At such levels, test substances cause irritation and inflammation that promotes cell proliferation. The more cells are forced to divide, the more opportunities there are for the defects that lead to cancer. These effects do not scale well to lower doses. It's the opinion of Bruce Ames (inventor of the Ames test genotoxicity screen) that this problem has completely confounded the interpretation of high-dose animal data. (His article in Angew. Chem. Int. Ed. 29, 1197, 1990 is a good statement of this argument).
2. The statement that "most chemicals, given in no matter how high a dose, won't cause the very unusual and specific biological effect of turning an animal cell cancerous" is not accurate. As Revere surely knows, there are many mutations and pathways that can turn a cell cancerous (which is why I keep harping on the idea that cancer isn't a single disease). Somewhere between one-third and one-half of all synthetic chemicals tested in cell assays or in high-dose animal assays show up as possible carcinogens, depending on your definitions. Interestingly, basically the same proportion of natural products (isolated from untreated foods and other sources) show up as positives, too.
Now, if you want to talk confirmed human carcinogens, then Revere may have a point. There are only some three or four dozen specific chemicals that are confirmed as causes of human cancer. Here's the list. If you read through it, you'll note that many of the 95 agents on it are radioactives or broad categories such as "alcoholic beverages." (Mention should be made of things like nickel, all compounds of which are under suspicion. Check your pockets, though, for your most likely exposure). Specific compounds known as human carcinogens are quite rare. But doesn't that fact support the ACSH's point more than Revere's?
3. Revere's statement that "Cancer cells are not "damaged" cells in the individual sense but "super cells," capable of out competing normal cells. It is only in the context of the whole organism that there is a problem" is also inaccurate. Cancer cells are indeed damaged, right in their growth-regulation and/or apoptosis pathways. A car whose throttle is damaged will run at a higher RPM than a normal model, but I wouldn't call it a "super car". And cancerous cells are often quite recognizably problematic, whole animal or not. They divide like crazy in petri dishes, the same as they do in an animal.
4. The majority of the cancers seen in rat and mouse models are in the liver (which supports the idea that these tumors occur through general strain on their metabolic systems). Human liver cancer is much more rare. The most common human cancer in many countries is lung, caused to a great degree by smoking (which is also likely to have constant-irritant cell-proliferation component). Of the agents on that ICAR list in point #2, only three or four are chemicals (or mixtures) known to induce human liver cancer specifically. This is a significant mismatch.
5. Revere states that "We know of no false negatives with this process. Every chemical we know that causes cancer in humans also does so in rodents. . ." But how about false positives? There are hundreds of compounds that seem to cause cancer in rodents that (as far as we can tell) do not pose a risk to humans. I say "seem to", because these are almost always high-dose studies. But I can even think of some compounds (the PPAR-alpha ligands) that cause all sorts of trouble (including tumors) in rodent livers at reasonable doses, but don't do so in humans. Rodent tox is necessary, but it sure isn't perfect.
There, that should be enough to complicate things. It doesn't make for as dramatic a story as the evil henchmen poisoning America on behalf of their corporate masters, I have to admit. But we'll have to try to get along without the excitement.
+ TrackBacks (0) | Category: Cancer | Toxicology
January 3, 2006
The FDA has recently pulled pemoline (sold in its non-generic days as Cylert) from the market, citing risks of liver toxicity. It's used for narcolepsy, ADHD, fatigue in multiple sclerosis patients, and other indications. In theory, there are several other drugs that are useful for all of these.
In practice, though, we're talking about some poorly understood CNS indications here, and the patient response to drugs of this sort is so heterogeneous that no one understands what's going on. There are people who respond to pemoline that don't respond nearly as well to anything else. And they're an unhappy bunch, because they seem to be willing to take on the liver risks in order to have a drug that works.
One such customer is Teresa Nielsen-Hayden, well-known for the Making Light blog, and this issue has been the hot topic over there, as you'd imagine. You'll notice that the linked blog post takes a rather hostile attitude toward Ralph Nader and his Public Citizen group, because they're taking credit for petitioning the FDA for the drug's removal. The Nielsen-Haydens (no fans of the Bush administration) have been furious at Nader and his people since at least the 2000 election, so this new accomplishment has understandably pushed about all their accessible buttons.
I think they've got a point. I think that if the risks of a given drug are known, that informed patients should have a right to choose that drug's benefits with its risks in mind. Pemoline has had a black-box warning on it for years now, so it's not like its risks have been hidden. Now, it's true that if such a drug remains on the market, some people are going to take it who shouldn't, black box or no black box. But I have to wonder if such people are going to find some other way to get into trouble, no matter how much concerned bureaucracies try to save them.
The situation with Accutane (isotretinoin) and pregnancy is a similar one: there's no way you can miss the fact that it should by a pregnant woman. But every year, some do, despite strenuous efforts to prevent such cases. Mind you, Nader's people have been agitating for years to get that one pulled from the market as well.
Wouldn't a similar registration and/or liver-monitoring regime work for the people who have to have pemoline? Perhaps coupled with some sort of indemnification for its manufacturers? How many more potentially useful medications could we have available under such conditions?
If any of my industry readers have suggestions on where to obtain pemoline, I'd be glad to hear them, although Teresa Nielsen-Hayden and many others would be even happier.
+ TrackBacks (0) | Category: Toxicology
Over the last month or so, the commenting system here has been rather erratic. Despite having Movable Type set to allow anyone to comment, people keep getting "held for moderation" messages - it happens to me, too. At the same time, an increasing amount of junk is making its way through. I'm going in manually every so often to clean out the stables, but that's not the best solution. My intention is to let anyone comment at any time, except the people who are robotically offering us all opportunities to play poker or to all-naturally extend various organs of our bodies. Anyone with experience in dealing with MT's current treatment of junk comments is welcome to offer suggestions on how to fix things.
I should also mention that Corante is going to be trying some server migrations, probably starting this coming weekend. The hope is that posting on this and the other blogs won't be disrupted, but if there's an outage, that'll be the reason. . .
+ TrackBacks (0) | Category: Blog Housekeeping
January 2, 2006
1. Get more done in the lab. That's a pretty generic one, but it gets harder to do the further on you go in your career. At the point I've reached, I could spend a good amount of my time hiding out in my office, banging away on the computer, and no one would be the wiser. No one including me, unfortunately, which is why I need to resist doing that when there's something more useful available.
2. Clean up my hood. I have piles of junk in there now, and while I can work around it, there's no reason to. I'd be more comfortable - and who knows, maybe even more productive - with some of it cleared out. The lab bench needs some pruning, too, since there's stuff there from three projects ago with months of dust on it. Out it goes.
3. Get more done in the office. What with number 1, that doesn't leave me much room to maneuver, does it? What I mean is to do the office work I need to do, without using the lab as a place to hang out and procrastinate. Perhaps these two resolutions could be combined into a broader one, a researcher's version of Kant's Categorical Imperative: to use both sides of my job as ends in themselves, not as means to avoid each other.
4. Go out on more limbs. This is another thing that I can afford to do at this point, and I should do it more often. I have some opportunities to try low-percentage high-reward ideas. Not everyone does - if I've got a special function, that's probably it. ("Bothering people" doesn't count, I'm pretty sure).
I can think of an example that touches on all of these simultaneously. Long-time readers will recall my occasional series on a research project I've been working on for the last two or three years. It's the very definition of a high-risk high-reward idea, and one of my better moments in 2005 was finally getting it to work a little bit. It's been somewhat stalled the last two or three months, though, partly by factors out of my control.
But partly not. I haven't been pushing the stuff as hard as I should have been, either. And I don't hesitate to diagnose fear as the cause. You've probably heard that line about how to travel hopefully is better than to arrive - well, it's not true if you've chosen your destination wisely. But it does show how it's easier to stay in the almost-done state rather than reach a resolution. After all, what if I'm wrong? What if the effect I've seen can't be generalized to anything useful, and I've worked myself up over a triviality? Isn't it better to stay where I am, where I can still think that I'm on to something?
Well, no, it isn't. But I need to keep reminding myself of that, and not look for excuses when it comes time for another key experiment. And that's my advice to the part of my readership that does research for a living: take some risks in 2006. I'm going to. Let's discover something for a change.
+ TrackBacks (0) | Category: Birth of an Idea