Drug Discovery: Some Imcompetent Reflections

The Feds have now made it official: The pharmaceutical industry doesn't know how to discover useful new drugs.

As reported by Gardiner Harris in the New York Times:
http://www.nytimes.com/2011/01/23/health/policy/23drug.html
--the government has run out of patience with the slow pace of drug discovery and plans to invest $1B in a new National Center for Advancing Translational Sciences within the NIH, specifically to do some of the drug development that private industry once did but, apparently, no longer does.

Now, I am not a research scientist, nor do I play one on TV, and I have no idea how to discover new drugs. Despite my near-total incompetence on these matters I will offer some reflections, many of which were previously offered in HOOKED or in previous blog entries.

First, what is wrong with the current state of industry drug discovery? This has a two-part answer--new drugs are quite hard to discover these days, and Pharma has essentially gotten itself out of the drug discovery business.

The big reason drugs are hard to discover is what I've previously quoted economist Uwe Reinhardt about, the paucity of low-hanging fruit. Once we have a basic understanding of what causes various common diseases, we can figure out what sorts of molecules might make useful drugs to treat them, but after a while, we have discovered most of the easy-to-find molecules that produce drugs that are both safe and effective. The as-yet-undiscovered molecules are fewer in number and much harder to find, or else more likely to cause adverse reactions while being no better than existing drugs. Until such time as we advance our basic understanding of disease, we can expect a lull in the pace of discovery. It seems as if we're more or less in one of those lulls just now. (If you think about all the things a molecule has to do inside of and in interaction with the human body to become a safe and effective drug, it's a miracle there are any useful drugs at all.)

Now, even given that good new drugs are hard to find, it remains the fact that Pharma seems ill-equipped just now to find them anyway. There are three main reasons for this so far as I can figure.

• The drug industry has been into me-too drugs in a big way. Rather than go after a whole new class of drugs, it's much simpler and safer to tweak an existing molecule slightly and then try to market the result as the "new purple pill" or whatever. (The wimps at the Patent Office enable this by granting patents in apparent disregard for the criterion of non-obviousness, which is one of the standard legal requirements for patents. I am reasonably sure that if I could afford to hire the right attorneys, I could take out a patent on the period at the end of this sentence.)
• The industry has outsourced all risk to biotech start-ups. These biotech companies are often based on new and even wacky ideas. At least 9 out of 10 will go bust. The big drug companies can wait around and then buy out the occasional, rare biotech company whose idea actually pans out and looks like it could turn into a drug.
• The third reason is actually at the root of the first two as well. The companies have put their bean counters almost totally in charge of the research agenda. If a line of research seems like it's not going to turn into a billion bucks in sales by next week, they axe it--even if, pursued for a couple more years, it could have produced sterling results.

OK, so now maybe we understand why then industry has for a good decade or more suffered from a near-empty R&D pipeline. Then next question we might ask is whether the Feds can do any better. The news report is not that promising. The new National Center, we are told, will do things like use robotic screeners to turn over a large number of candidate molecules. This sounds like the current industry business model. (I always had a lingering suspicion that scientific discovery ought not be delegated to robots.)

Clearly the new Center is supposed to be a part of the love affair that the NIH has declared with the idea of "translational science." Depending on who's talking, you can get very different ideas about what "translational science" is. On the positive side, translational science seems to mean having scientists in different fields form interdisciplinary teams, communicate with each other, and exchange ideas throughout the course of research. If you try to implement this model, you find that it's hard to do because the average scientist was simply not trained in this way. That in turn suggests the need to train future scientists more in the translational setting from the get-go. All that, I suggest, is for the good, though unlikely to produce quick results.

On the negative side, sometimes it sounds as if what "translational science" means (in terms of moving research findings more quickly "from the bench to the bedside") is "let's do what they do in private industry" or "let's do what will make us the most money soonest." If the industry R&D model is a bust, then why the heck model our science after it?

To my poor ignorant mind, it seems that ultimately, we'll discover new, useful drugs to the extent that we invest in more basic science research and learn more about the mechanisms of disease. That's not sexy, and it certainly cannot be preprogrammed to suit Wall Street's quarterly earning reports. But it seems to explain, for instance, why we seem poised for real advances in cancer drugs in the next decade or so, because we now know a lot more about tumor cell genetics and enzymology. Does this mean that we won't occasionally get lucky and discover good drugs purely by serendipity? Of course that can happen; what about aspirin and penicillin? But we probably can't count on that as our best and sole approach.