The advantages of today's powerful drug-hunting technologies are offset by what many see as a loss of freedom to stretch one's mind around novel ideas. ‘Too much computer and not enough brain,’ says former Merck biochemist Alfred Alberts, who helped invent Mevacor, the first successful statin, as well as its $6-billion-a-year successor Zocor. Strategies to unearth blockbusters today are ‘not working,’ says Alberts, who retired in 1995 after 20 years at Merck. ‘I think that's fairly clear.’ – The Brains Behind Blockbusters

A careful reading of the history of drug discovery shows successful drug development efforts of the 20^th century fall into a small number of basic motifs. The original insight, the germ of the idea, is accessible to even the uninitiated. These were all ‘great ideas’ spawned from a 20^th century explosion in basic biology. I suspect that when future authors look back over the history of drug development in the 21^st century they’ll be able to construct a similar simple list, only based on a 21^st Century explosion in biologic understanding. For example, Janeway Jr. writes ‘as early as the 1960’s textbooks could describe these cells, now the central focus of immunology, as having no known function.’ It’s pretty clear that insights from this emerging field of immunology will loom large in the 21^st century list. We invest in emerging motifs in biology in the hopes of finding that next blockbuster drug.

Are these motifs a priori, or are they merely convenient cubbyholes where we a posteriori place successful drug discoveries? Frequently in the 20^th century it was a matter of biology catching up with drug development. It didn’t matter that at first it was biologically unknown that important anti-metabolite drugs interfered with DNA chain elongation during cell replication. Much more important was the observed animal physiology or human clinical outcomes. However once the biology catches up we take advantage of the deeper insights to further improve the drugs and develop new blockbusters. For example, Prostaglandin was not elucidated until the 1950’s which led to a deeper understanding of how aspirin worked, and subsequently to improved blockbuster pain medications. Motifs are good in that they make for a compelling narrative for funding and can generate excitement in the development teams. But they are a starting point, and are often subject to revision.

The proof is in the pudding. We take the original insight, design a series of tests, and then based on the results decide whether or not to continue with the insight, discard it, or (as is usually the case) adjust it. The adjustment process is called ad hoc hypothesis generation⁺, and can be good or bad. It’s good in that history shows that many successful drugs are based on insights that are often just a starting point, and need adjustment. It’s bad in that it can lead to years of fruitless efforts for an original insight that should never have been continued. It can be safely assumed all unsuccessful drug programs start down the path shown in the figure to the right but just never make it to the confirmation stage. Also, many successful drugs followed an adjustment process that spanned decades; an unsustainable model for industrial R&D.

The key lesson learned in innovative drug development is Anything Goes⁺! Successful drug discovery can come from any of a dozen different starting points, subject only to the caveat that the original starting point is rarely correct. So the starting point can be a new motif, an extension of an existing motif, an observational curiosity in the laboratory, or anything! Hypothesis-free research is, surprisingly, often quite productive. If the starting point can generate lasting enthusiasm and passion in the research team, then it’s miles ahead of any artificial means of selecting and enforcing starting points (e.g., therapeutic areas, commercially-driven planning). The most important question to be answered for successful drug development in the 21^st century is how we will allow individuals or teams to ‘stretch their minds’ around novel ideas without bankrupting the R&D effort, both financially and reputationally. How do we open up options for novel thinking, and then shut them down in a way that is not destructive of the passions and enthusiasms of future novel thinking.

We expect a couple of rounds of disconfirming evidence⁺ (ad hoc hypothesis generation) during this initial period. The commercial proposition first proposed at recruitment is not expected to be the one that is brought up before the graduation committee. Research teams are free to position their technology in its most favorable light during bootcamp. We seek evidence of commercial potential as supported by a few rounds of disconfirming evidence.

---

‘Direct mechanistic evidence for how any vaccine works in humans is sparse.’ Toward an AIDS Vaccine. The long-held theory of Original Antigenic Sin OAS suggested that new virus strains would evade rapid detection when the immune system had been trained for previous viral strains. However patients vaccinated in prior seasons for earlier influenza strains showed no limits to response for the 2008 season influenza strains. Rapid cloning of high-affinity human monoclonal antibodies against influenza virus, Jens Wrammert, et. al., Nature Letters, Published online 30 April 2008.