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
October 31, 2007
The post here the other day on resistant bacterial infections prompted some readers to wonder why the drug industry isn’t doing more to come up with compounds in this field. It’s not like there’s no money to be made, and it’s not like there’s no history of antibiotic research, after all. But since my industry doesn’t have a history of knowingly leaving money on the table (what industry does?), you’d figure that there’s more to the story.
Money aside, there’s a real problem with finding good targets. For as long as I can remember in the industry, the infectious disease field has suffered from a relatively small target landscape. Almost all the known drugs in the area work through just a handful of basic mechanisms, and adding new ones to the list has been very difficult for at least the last twenty or thirty years.
That was supposed to change, in theory, starting about ten years ago. I interviewed around then at a company that was working in the field, and everyone was quite excited about the bacterial genome sequences that were starting to appear. Surely this would open the sluice gates and let that long-delayed swell of new targets come washing down the flumes. Hasn’t happened. Not yet, anyway.
I have the impression that the same problems that have affected the translation of human genomic data to new drugs have been the problem here as well. In some cases, not as many genes came out as some people were hoping for. And of these, the function of many of them was (to put it mildly) obscure. Of the ones whose use was at least partially known, many of them have proved not to be useful targets for killing the bacteria or limiting their growth. And of the ones that made that cut – and we’re down to an all-too-manageable set by now – screening hasn’t turned up much chemical matter for people like me to work on.
In fact, there’s a persistent feeling among many people in the field that bacterial and fungal proteins have a lower hit rate than you’d assume they would. Even enzymes that are fairly homologous to those in higher organisms, so the story goes, don’t turn up as many hits in the screens as expected. I’m not sure if this is true or not, but as folklore it’s pretty well known. The combination of all these factors with the perceived lack of opportunities for profits (even if you do find something) has made for slow going.
In recent years it’s become clear that the medical need has grown to the point that antibiotic research can indeed be financially worthwhile – but there are any number of financially worthwhile drug outcomes that we haven’t been able to realize. (See obesity, Alzheimer’s, and many other therapeutic areas for examples of multibillion-dollar opportunities waiting for a good idea to come along. Resistant bacteria have their name on one more sword stuck in yet another stone.
Update: there's clearly another reason why developing good antibacterials is hard, and it's the same reason we need more of them. Bacteria are well-stocked with efflux pumps to get rid of molecules they don't like (and with other weapons as well), and they evolve so fast that you can watch them do it. I wrote about efflux on the site a while back - another post is well worth doing soon.
+ TrackBacks (0) | Category: Drug Development | Infectious Diseases
I’ve written before about all the fun you can have in a lab with compressed gas cylinders. We use the things all the time in chemistry, but as pieces of apparatus, they can only be pushed so far. The problem is that they demonstrate their unhappiness by venting great quantities of stuff that you’d rather not breathe (if you’re lucky), or by taking off like an unguided missile and punching holes in the walls and ceilings (if you’re not). That latter behavior is flat-out guaranteed to show up if you abuse them – for one of the more spectacular examples, see here.
I’ve never had one take off on me, fortunately, but I haven’t always stuck to the straight and narrow with the things, either. My worst behavior has usually been with lecture bottles, the dilettante-sized gas cylinders that bench chemists often use. Most chemistry departments have a few of these sitting around, generally charged with foul reagents that are needed every three or four years or so. Sulfur dioxide, boron trifluoride, phosphine – that’s the sort of thing. They’re low-use almost by definition. If there’s a regular need for a gaseous reagent, you buy larger cylinders of it, because lecture bottles are by far the most expensive way to go.
In graduate school, I was setting up some Prins reactions, which take some sort of acid component to make them run. If you use an aqueous one, you generally get an alcohol out of them (from water picking up the final cation), but if you go anhydrous you can get all sorts of other compounds. I needed a bromide, so anhydrous hydrogen bromide it was.
We had a fairly crusty lecture bottle of it around, and I eventually located some dubious-looking small regulator valves. I picked the least-corroded looking one and screwed it on. Lecture bottles have a main metal-faucet style valve up at the top, like all gas cylinders, and once you open that it’s up to the regulator valve to stop things down to manageable flows. I had my reaction set up, so I worked some tubing onto the thing and had a go at opening it up.
No dice. Boy, was that thing tight. I reached into the hood and wrestled around with it, to no avail. I took it out and got a better grip in another part of my hood, away from my reaction setup. Tighter than two quarts of fresh frogs in a half-pint pickle pot, as Walt Kelly once put it – the valve wouldn’t budge.
I’ll skip over a couple of intermediate stages and cut right to the final scene, which is me clutching the darn lecture bottle to my chest, hopping around the lab grunting and cursing as I put all my strength into trying to force the stupid valve open. You won’t see a pictograph of that method in the instruction booklet, I’m pretty sure. A recording of what happened next would have gone something like this: “Urk! Unk! Ark! WHOA!”
The valve opened, finally responding to my Conan-the-Barbarian technique, and the cheap regulator then hissed out an orange cloud of gaseous HBr right into my shirt pocket. Not a good storage compartment, actually. I whooped, shut the valve, laid the gas cylinder down fast and stripped off my shirt where I stood. You haven’t really lived in a lab until you’ve taken off your clothes in it, I always say. I staked my claim to this one by standing in it shirtless, splashing saturated sodium bicarbonate on my chest, and glaring at the remains of what used to be one of my favorite shirts.
+ TrackBacks (0) | Category: How Not to Do It
October 29, 2007
There was an intriguing paper published earlier this month from Manfred Reetz and co-workers at the Max Planck Institute. It's not only an interesting finding, but a good example of making lemonade from lemons.
They were looking at an enzyme called tHisF, a thermostable beast from a marine microorganism that's normally involved in histamine synthesis. It has an acid/base catalytic site, so Reetz's group, which has long been involved in pushing enzymes to do more than they usually do, was interested in seeing if this one would act as an esterase/hydrolase.
And so it did - not as efficiently as a real esterase, but not too shabby when given some generic nitrophenyl esters to chew on. There was some structure-activity trend at work: the larger the alkyl portion of the ester, the less the enzyme liked it. Given a racemic starting material, it did a good job of resolution, spitting out the R alcohol well over the S isomer. All just the sort of thing you'd expect from a normal enzyme.
Next, they used the crystal structure of the protein and previous work on the active site to see which amino acids were important for the esterase activity. And here's where the wheels came off. They did a series of amputations to all the active side chains, hacking aspartic acids and cysteines down to plain old alanine. And none of it did a thing. To what was no doubt a room full of shocked expressions, the enzyme kept rolling along exactly as before, even with what were supposed to be its key parts missing.
Further experiments confirmed that the active site actually seems to have nothing at all to do with the hydrolase activity. So what's doing it? They're not sure, but there must be some other non-obvious site that's capable of acting like a completely different enzyme. I'm sure that they're actively searching for it now, probably by doing a list of likely point mutations until they finally hit something that stops the thing.
So how often does this sort of thing happen? Are there other enzymes with "active sites" that no one's ever recognized? If so, do these have any physiological relevance? No one knows yet, but a whole new area of enzymology may have been opened up. I look forward to seeing more publications on this, and I'll enjoy them all the more knowing that they came from a series of frustrating, head-scratching "failed" experiments. Instead of pouring things into the waste can, Reetz and his co-workers stayed the course, and my hat's off to them.
+ TrackBacks (0) | Category: Biological News
There’s been a steady stream of reports in the news about methacillin-resistant Staph. aureus. It’s not a new problem, but (like other nasty infections) it does get a lot of press when the media start paying attention. Works in reverse, too – on the viral front, have you noticed the much reduced number of bird-flu-will-kill-us-all stories this year as we head toward winter? This despite the likelihood of bird flu killing us all being as high (or low) as ever, as far as I can tell.
But the resistant bacteria problem is certainly no joke, and there doesn’t seem to be any reason why it won’t gradually get worse over time. It struck me the other day that antiinfectives, as a drug research field, might be moving toward a similar spot to oncology. In both cases, you have a problem with rapidly multiplying cells, giving you a serious medical outcome - often in cancer, and increasingly with infections. The average tumor is a lot more worrisome than the average infection, of course, but that’s something we can only say with confidence in the industrialized world, and we've only been able to say it for the last sixty or seventy years. As cancer gradually becomes more manageable and infections gradually become less so, the two might eventually meet – or even switch places, which would be bad news indeed. (In some genetically bottlenecked species, in fact, the two problems can overlap, which is fortunately extremely unlikely in humans).
There are, of course, a lot of differences between the two fields, not least of which is that you’re fighting human cells in one case and prokaryotes (or worse, viruses) in the other. But many of those differences actually come out making infectious diseases look worse. The transmissibility of bacteria and viruses make them serious contenders for causing havoc, as they have innumerable times in human history, and they can grow more quickly in vivo than any cancer. It’s only the fact that public health measures allow then to be contained, and the fact that we’ve had useful therapies for many of them, that makes people downrate the infectious agents. If either (or both) of those change, we’re going to be rethinking our priorities pretty quickly.
What this means for drug development is that some researchers will have to rethink their attitudes towards antiinfective drugs. For serious infections, we're going to have to think about these projects the way we've traditionally thought of oncology agents - last-ditch therapies for deadly conditions. Anticancer therapies have long had more latitude in their side effects, therapeutic ratios, and dosing regimes, and antibiotics for resistant infections are in the same position. For some years now, there's been a problem that new drugs in this field would perforce have small markets, since they'd be used only when existing agents fail. That market may not be as small as it used to be. . .
+ TrackBacks (0) | Category: Cancer | Drug Development | Infectious Diseases
October 25, 2007
A colleague reminded me the other day of a project that he and I had worked on back at the Wonder Drug Factory seven years ago. "Seven years ago", I thought. . .I was the project leader on that one, trying to keep things alive as weird toxicology kept torpedoing everything. In the end, we held it together long enough to get four compounds into two-week tox testing, whereupon every one of them wiped out for yet another set of ugly reasons. Ah, yes. No one's going to have to work on that stuff again, that's for sure.
Hmm, I thought. What was I doing seven years before that? Well, I was back at my first drug industry job in New Jersey. The company had just moved into a new building the year before, and the old site was on its way to becoming a Home Depot. I was spending my days cranking out molecules hand over fist. Boy, did I run a lot of reductive aminations. It's safe to say that during those years I ran the majority of all the reductive aminations that I'll ever run in my life, unless something rather unforeseen crops up. We made thousands of compounds on that project, and I remember pointing out in a talk that nobody makes that many compounds if they really understand what they're doing. This was not a popular line of reasoning, but it's hard to refute, unless saying how much you don't like something counts as a refutation.
And seven years before that? Still in the lab. I was midway through grad school, wrestling with the middle of what turned into twenty-seven linear steps by the time I pulled the plug. (At this point, I began to reflect that I've been doing chemistry for quite some time now). In 1986 I didn't know that I wasn't going to end up finishing the molecule, and I was still hauling buckets of intermediates up the mountainside, only to find them alwyas mysteriously lighter and smaller by the time I got to the top. My response, naturally enough, was to start with larger buckets - what else was there to do?
And seven years before that? That finally takes me over the chemistry horizon, back to my senior year of high school in Arkansas, and to what might as well be a different planet entirely. Although I was interested in chemistry - as I was in most all the sciences, something I've never lost - I'd never heard of a Grignard reagent, and I didn't know what a nucleophile was. Counting up, I see that some time next year will mark the point at which I will have spent a slim majority of my lifetime doing organic chemistry, which is an odd thought. And it makes me wonder what I'll be up to seven years from now. . .
+ TrackBacks (0) | Category: Graduate School | Life in the Drug Labs
October 24, 2007
GSK opened up their books today, and the magnitude of their Avandia problem has become clear. This was a big part of the company’s sales, and the recent cardiovascular worries have really knocked down the numbers. The response, as has been the trend this year, is for the company to announce layoffs.
And man, have there ever been layoffs in the industry this year. There’s a list of the larger ones over at FierceBiotech, and it does not make for cheerful reading. In January, Pfizer announces 10,000 job cuts and closes their Ann Arbor site. That same month, Bayer-Schering closes the doors on research buildings in Connecticut and California (layoffs which were announced in 2006 and are thus not on the Fierce list). Bayer-Schering, who really should have run that B-S initial thing past a couple of native English speakers, announces 6,100 more job cuts in March. In July, AstraZeneca doubles down on its earlier layoff announcement and says that 7,600 jobs will disappear, and J&J announces a 4% reduction in its workforce (5,000 jobs). Then in August, Amgen cut over 2,000 jobs of its own.
In September, most everyone held on to the jobs for the moment But Novartis said this month that they’re going to trim over 1200 positions in the US, mostly through attrition. And now we have GSK with disappointing earnings and an announcement of unspecified layoffs, and bear in mind, this is just the news from the big outfits. The usual turmoil has been going on among the smaller companies (Idenix, Palatin, Sonus, and others), whose fortunes depend more on single drugs.
What a year – and hey, there’s still time to announce more layoffs before the holidays, so we may not be through yet. It’s tempting for some people to look at a list like this and say “Outsourcing! China! India!”. And I can’t deny that some of these jobs have headed there, just as some possible hiring expansions have been muted for the same reasons at other companies.
But outsourcing isn’t the whole story. Many of these job cuts have been in the sales forces, and they’re definitely not outsourcing the sales reps to Shanghai. Ditto for the people in Regulatory Affairs and Legal. Outsourcing is changing the size and shape of layoffs, but it’s not providing the motive force for them. That force, simply enough, is just that we’re not selling enough drugs, mostly because we don’t have enough good drugs to sell. Some areas have had too few projects even to start with (anti-infectives?), and everyone has had too few make it all the way through the clinic and the FDA.
And some of those failures have been extraordinarily large and expensive. Unfortunately, this has been the case for a while now. Over the last few years, we’ve had drugs that have failed terribly late in the clinic (torcetrapib, among others), drugs that have made it through trials but failed at the FDA (rimonabant, among others), and (most expensively of all) drugs that have made it to market and been pulled back early in their product lifetimes, after the big promotion money’s been spent and before any of it gets made back (Bayer’s Baycol and Pfizer’s Exubera – among others).
Add in the ones that never lived up to their planned potential (Iressa, Macugen, yes, yes, among others) and you have a gigantic revenue shortfall. Now, it’s true that not all of these would have made it under any conditions. Drugs fail. But do they fail like this, so relentlessly and so expensively? And it’s not that we aren’t killing all sorts of stuff off earlier in the development pipeline – no, these things are what’s left after the dogs are gone.
What to do? If I knew how to answer questions like that, I'd be dictating this from the deck of my yacht. The glass-half-full perspective is that there sure are a lot of opportunities for anything that can open up some new therapeutic areas or help with drug failure rates in the existing ones. It won't take much, considering where we're starting from. Yesterday I was encouraging people to try out some high-risk ideas, and here, in case anyone was wondering, is an excellent place for them.
+ TrackBacks (0) | Category: Business and Markets | Clinical Trials | Drug Development
October 23, 2007
My apologies for no post today - home events kept me away from the computer for a while, but everything's settled back down now.
I've had several e-mails the last couple of months asking about "Vial Thirty-Three", the saga of which can be found (in reverse chronological order) here. (More specifically, the first time that particular experiment worked was the May 18, 2006 entry, and you can scroll up from there if you wish). When last heard from, I was cranking away on a batch of experiments to finish before the Wonder Drug Factory closed its doors at the end of January.
The last ones got run just before they pulled the electrical plugs out of the walls, and a lot of interesting things came out of them. They were interesting enough, in fact, that they suggested a whole new series of ideas to me during the months I was between jobs. Of course, that did little good, since this isn't the kind of stuff that you can easily pull off in your basement.
But I'm very glad to report that my current employer is interested in this sort of thing, and in plenty more weird stuff besides. That's the good news, and good news it surely is. I have an explicit mandate to look at ideas and technologies beyond what the company is currently doing, and a group to tackle these things full-time. This is just the sort of thing I like to do, and having it as my main job responsibility is so enjoyable that I may never get used to it.
The bad news is that I won't be able to talk about what I'm up to. At the Wonder Drug Factory, my odd experiments were a sideline and were a long shot to work at any rate. I felt safe talking obliquely about them. But now I spend my whole day on this kind of thing - the mutant progeny of Vial Thirty-Three and several other similarly odd ideas. It's a wonderful feeling to see this sort of thing get resourced and watch it move forward, but it's all completely proprietary.
But even if I can't say much, I just wanted to let people know that things are continuing. I'm doing full-time what I used to have to squeeze in as a sideline. Working on this kind of idea has been, in retrospect, one of the best decisions I ever made as a scientist. If any of you have some wild thoughts about experiments that sound a bit weird, but just might work - well, my advice is to somehow make time for them. Sometimes they work. . .
+ TrackBacks (0) | Category: Birth of an Idea
October 22, 2007
Pfizer has pulled the inhaled insulin Exubera from the market, and not because of the FDA, and not because of the lawyers. They’re giving up on it because they can’t take the pain any more. The company sold 12 million dollars worth of the stuff so far this year, a horrifyingly tiny amount. That represents about 0.3% of the insulin market, which we can round off to "zero". The ticket out is a mere 2.8 billion dollar charge against earnings. It's the first time I can remember a company pulling a drug just because it was losing so much money - of course, Pfizer is not a normal company, and these are not normal times, especially for them.
There are plenty of post-mortems around, from the front page of the Wall Street Journal onward. (See the Journal’s Health Blog, Matthew Herper’s blog at Forbes, Pharmalot and the folks at Invivoblog for more). I have my own, naturally, since a disaster of this size admits of many interpretations. Here’s what it says to me:
1. Marketing isn’t everything. The next time someone tells you about how drug companies can sell junk that people don’t need through their powerful, money-laden sales force, spare a thought for Pfizer. The biggest drug company in the world, with the biggest sales force and the biggest cash reserves, couldn’t move this turkey. People didn’t want it, and they didn’t buy it.
The flip side of this is that even the drugs that folks love to hate, the ones that no one can figure out why they do as well as they do, must be doing something for some people. Perhaps other, cheaper drugs would do something quite similar, and we can discuss cost/benefit ratios, but you couldn’t sell them if people didn’t feel that benefit in the denominator. Not many people felt it from Exubera.
2. Internal sales estimates can be a joke. People inside the drug companies have known this for a long time, although they’d often rather not think about it. Analysts have known it, too, but they're forced to pay attention to those numbers anyway. But man, look at the magnitude of this one. Just as Warner-Lambert tried to kill Lipitor before they brought it to market (who needs another statin?), Pfizer was telling analysts a few years ago that their projections for Exubera sales (a billion dollars a year) were just too darn low. Two billion a year by 2010, thank you and please correct the error. Only off by a factor of one hundred, and what’s two log units between friends?
Sales forecasts are not science, and they only bear a superficial resemblance to math (where the phrase "imaginary number" is rather more strictly defined). They are guesses, and some of them are good guesses and some of them are awful, and unfortunately when you first look them over, they all smell about the same.
3. Groups aren’t necessarily smarter. This is the flip side of all the “Wisdom of Crowds” stuff, which only works when a lot of people (who think of a lot of different things) all get a crack at a subject. Inside a company, though, diversity of opinion sometimes doesn’t get much respect, and the problem gets worse in areas like marketing (and worse as you go into the higher ranks). Think of what would have happened to a Pfizer exec who forecast a 0.3% market share and a 2.8 billion dollar charge for Exubera when everyone else was revising their figures up a billion. It would have taken a fantastic amount of nerve to make a call that contrarian, and the rewards for being right (if any) would definitely not have been worth it. Even if someone had a terrible suspicion, it was surely much safer to keep quiet.
Groups of people can, in fact, be quite stupid. People will deliberately not bring their minds to bear on a problem, in order to get along with their co-workers, to not stick their heads up, or just to make the damned meetings end more quickly.
4. Pfizer is in vast amounts of trouble. While not an original thought, it's an unavoidable one. We all know the problems they have, and believe it, they do too. But what to do? I remarked a few weeks ago that Pfizer's situation reminded me of a slow-motion film of a train running toward a cliff, and a colleague of mine said "Yeah, me too, but in this case they're still boarding passengers and loading their luggage".
+ TrackBacks (0) | Category: Business and Markets | Diabetes and Obesity | Drug Industry History
October 18, 2007
My graduate school lab, like most of them, had an assortment of people from different countries. That kept things at all sorts of hours, since we’d get the occasional Japanese post-doc who never really seemed to get off JST and worked the zombie schedule. It also made for some adventures in communication. English was the lingua franca of the lab, naturally, but there were a lot of varieties spoken (and attempted).
And although it’s risky to generalize, I think that the ones with the biggest language gap were the aforementioned Japanese. Friends of mine from the country have blamed the problem on the traditional state of English teaching there, and the way that too many students are taught the language as if its phonics really did conform to what’s available in Katakana.
That’s the writing system used in Japan to render words phonetically. Reading fancy Japanese (Kanji) takes some real practice, but any hack (like me) can plow through Katakana with a chart and a little practice. I’ve been asked many times, in wondering tones, if I read Japanese after I pulled out a useful reagent name from a Japanese patent, but I wasn’t reading Japanese – I was reading English. Sort of.
Sounding out the words makes you sound like the most unfortunate expatriate Japanese post-doc you’ve ever heard. “Cyclohexyl”, for example, comes out as “Sa-ki-ru-he-ki-sa-ru”, and “chloro” is “ko-ru-ru”. I’d probably sound even worse than that if I had to speak Japanese, but it does give you some insight as to where the stronger features of the accent come from.
One way or another, we all did communicate in the end. I remember talking with one of our post-docs, trying to learn some Japanese profanity (a well-known gateway into a foreign language, of course). But I couldn’t get the concept across. “Bad word?” he asked, puzzled. “Curse word?”
An idea hit me. “OK”, I said, pointing at his rota-vap where a 1 mL flask was spinning. “How long did it take you to make that stuff?” That stuff, he informed me, was step 17 of his synthesis, and had taken weeks and weeks of work. “Fine,” I said, “what happens if it falls into the water bath?” “Ah! Terrible!” he said, looking fearful at the very thought. “Right”, I told him: “If that happens, what do you say?”
Enlightenment! “Oh! Yes! Those words! Bad word, yes, now I understand!” A great moment in international understanding. We went on to explore the sorts of phrases that are absolutely guaranteed to come in handy in any research lab, no matter where.
+ TrackBacks (0) | Category: Graduate School
October 17, 2007
Well, Carl Icahn has been buying large boatloads of Biogen stock over the last few months, and as those of you who follow the biotech industry are aware, he’s got the company putting itself on the block. That’s one of the normal sequels to an Icahn buying spree, and it looks like it’s going to work out well for him again (see Medimmune for another recent example).
Will it work out for Biogen, though? Looking over the potential buyers, it should also come as no surprise that the serious prospects make a pretty short list. Let’s see. . .huge, but with little presence in biologicals. Loaded with cash, but with serious pipeline problems and some nasty patent expirations coming on. Not averse to doing great big acquisitions to pump up their numbers. Gosh, who does that leave? Yep, Carl Icahn is playing marriage broker between Biogen and Pfizer.
There are also rumors of Pfizer buying Genzyme, for the exact same reasons. (That deal would at least have the desirable side effect of not further enriching Carl Ichan). If Amgen were doing better, we’d be hearing about Pfizer buying them, too, no doubt, but for now takeover rumors are one of the few things they probably don’t have to worry about. (Although you can make