Choosing the right path in clinical development – as in life – is a huge challenge. For small companies it can be the difference between success and failure, survival and implosion. But divergent clinical data published in the last month on two apparently similar members of the Factor Xa inhibitor class demonstrate that the largest pharma companies can get it wrong too – and the consequences for them may be no less catastrophic.
In helicopter view, drug discovery and development is very simple: choose the right molecule, and choose the right people to treat with it, and success is assured. For a decade from the mid-1990s, much of the industry simplified it even further: by focusing on ‘me-too’ or ‘fast-follower’ compounds (where commercial success depends on being best in class or, failing that, best at marketing) the second requirement, choosing the right people to treat, was taken out the equation. For a while, this worked great because it leveraged the two things that large pharmaceutical companies are good at: choosing good molecules within a class and marketing.
“Faced with a failure, ask yourself whether it is the asset itself or the team the presided over it who should be tarnished”
But with so much resource poured into optimizing existing drug classes, the opportunities for further improvements began to dry up, and by the mid 2000’s the focus had to shift back to real innovation.
And developing first-in-class therapies thrusts the skill of matching the therapy to the right indication back into the limelight.
When it comes to matching therapy to indication, there are so many ways to get it wrong that it should come as no surprise that teams get it wrong, one way or another, almost every time. Persistence is an important characteristic, as stories such as Pfizer’s Viagra™ sildenafil and Amgen’s Enbrel™ etanercept clearly demonstrate. The PDE5 inhibitors first came to market in the niche pulmonary artery hypertension market, and only became blockbusters when their legendary effects on erectile dysfunction were recognized. Similarly, anti-TNFs had failed in multiple inflammatory indications before their Lazarus-like effect in rheumatoid arthritis was discovered.
These drugs are the pharmaceutical equivalents of supernovae – for years they bubble along below the radar, hardly noticed by the world, until one day they burst forth in a new indication, and a new blockbuster is born.
The anti-CD52 antibody alemtuzumab looks like it is ready to explode: as Campath-1H, this was Greg Winter’s very first humanized monoclonal antibody to find therapeutic use, and for more than a decade it has provided useful service as a second line therapy for chronic lymphocytic leukemia. But data from a couple of small clinical studies hinted at better things, suggesting that contrary to the established doctrines, this anti-thymocyte antibody could be hugely beneficial in multiple sclerosis. This was confirmed by the CARE-MS I Trial, reported by Sanofi and Genzyme in October 2011, which demonstrated clear superiority to the current standard of care (Merck Serono’s Rebif™ interferon 1-beta), with almost 80% of patients relapse-free at 2 years. Thirty years from discovery to launch in the optimal indication might not be a record, but its certainly a long while.
As these examples clearly demonstrate, getting it right eventually is better than giving up (at least for the patient – the original inventors and investors in these products were likely out of the picture long before the real potential of what they had invented and developed was ever stumbled upon). But getting it wrong, even once, can have disastrous consequences.
Some first-in-class product candidates have one shot, and are likely to be usable for one thing only, simply because of the mechanism of action. The exciting Phase II clinical data reported in November 2011 demonstrating the cholesterol-lowering potential of REGN727, an antibody against PCSK9 from partners Sanofi and Regeneron, is just such an example. PCSK9 has only one known molecular partner and activity: marking the LDL-cholesterol receptor for destruction. Interrupting this activity should (and the Phase II data suggests that it does) lower LDL-cholesterol. But had it not done, there was little prospect of a silver lining to save the programme. In effect, developing the first-in-class PCSK9 inhibitors was a binary bet.
But in most cases, things are not that simple. Holding a first-in-class anti-inflammatory therapy with an entirely novel mechanism of action, such as the somatotaxins from Funxional Therapeutics, you are faced with at least 150 possible indications and few reliable signposts as to which rainbow harbours the pot of gold at its end.
Its far from straightforward to decide which scenario is preferable. In the first case (where there is only one plausible indication) the decision is easy, but the risk is high – since if you fail with your first shot on goal there is no possibility of residual value in the programme. Every penny invested in turning the Phase II clinical card is lost in an instant. By contrast, in the second case, the same investment to reach the start of Phase II should have enabled multiple shots on goal, and as a result an increased chance of finding the buried treasure.
But the context matters: small companies are unlikely to have the resource to shoot on goal more than once. For small companies, then, there is a real advantage in selecting assets of the first kind. By contrast, large companies can, if they really believe in the potential, benefit from the multiplicity of options in a range of indications.
But being large, doesn’t protect you from following the wrong path with these “multi-indication” product candidates.
BMS and Pfizer’s focus on the lack of power of the ADOPT trial, whose results were reported in November 2011, cannot hide another complete failure for the Factor Xa inhibitor Eliquis™ apixaban, this time in thromboembolism among acutely ill patients. As DrugBaron noted in September, apixaban is headed for an expensive commercial failure (it is the perfect example of a ‘buster’, as opposed to a blockbuster – that is, a drug that is good enough to consume your development cash right through to approval, but not good enough to attract enough sales to repay that investment). Yet it need not have been that way: Xarelto™ rivaroxiban from Bayer is headed on a completely different path, buoyed by positive data in acute coronary syndrome, it is a drug headed for a blockbuster future.
Why is one a ‘buster’ and the other a potential blockbuster? Its almost certainly nothing to do with the molecules themselves, but the development paths selected by their owners. Bayer did trials in indications where Factor Xa inhibitors clearly are superior to heparin or warfarin, while BMS and Pfizer chose indications where the superiority turns out to be minimal (and Boeringer Ingelheim, with the third member of the class, Pradaxa™ dabigatran, is likely somewhere in the middle, with a degree of benefit in the indication they have pursued). As much as $100 billion in lifetime a sales may hang on those decisions.
Developing first-in-class therapies thrusts the process of selecting the right indication right back to centre stage.
Of course, failure may be built in from the start: Effient™ prasugrel from Eli Lilly, another ‘buster’ is mechanistically inferior to AstraZeneca’s Brillanta™ ticagrelor, which more than likely has a blockbuster future. But more often that not, like the Factor Xa inhibitors, failure is engineered by the stewards of the asset. Next time you are faced with a failure, ask yourself whether it’s the asset itself, or the team that presided over it, that should have its reputation tarnished.