Industry Voices: The Fibrosis Stampede
By Michael Gilman
Fibrosis is wound repair gone bad. It's the process that erodes our organs, slowly and inexorably, in response to a lifetime of injury and insult. Fibrosis burns out the kidneys of diabetics, condemning patients to a punishing regime of dialysis. It turns the livers of hepatitis C patients crusty and cirrhotic, making liver transplant the only way out. In a particularly insidious form known as idiopathic pulmonary fibrosis or IPF, it scars the lungs of unlucky people whose only real transgression has been to breathe, slowly suffocating them over the course of a few years. More than 40% of deaths in the developing world are attributable to fibrotic diseases.
Yet there are no approved anti-fibrotic drugs in the U.S. and, until recently, very few even in advanced development. How could that be? And why, now, a renaissance that sees one drug, pirfenidone, approaching approval in the U.S. for IPF and a burgeoning pipeline of followers with potentially profound disease-modifying activity?
The answer, surprisingly, is not advances in biology. We've known for years what causes fibrosis. There's one major bad actor in fibrotic organs, a rowdy variant of the otherwise placid fibroblast, called a myofibroblast. The myofibroblast quarterbacks wound repair; in response to injury, it cranks up its metabolic activity, proliferates rapidly, spews out piles of collagen and other matrix proteins, and develops muscle-like contractile properties to physically close wounds. In an acute injury, like a cut or a burn, its work done, the myofibroblast goes to ground and reassumes its quiet life as a resident tissue fibroblast. But in injury that is chronic and lifelong, myofibroblasts persist and eventually take over, filling the organ with collagen, laying down scar tissue, crowding out and eventually destroying the functional tissue in the organ. With time, the organ can no longer do its job. It fails.
So, if you're after an anti-fibrotic drug, you want to stop the myofibroblast. And we know how to do that, too. The myofibroblast is largely owned and operated by a cytokine called TGF-beta. Two decades of data make clear that inhibiting TGF-beta activity prevents and perhaps even reverses fibrosis in animals; we've cured mice by the thousands. Sure, there's more to the story than just TGF-beta--all amply on display at this week's Keystone Symposium in Big Sky, MT--but for fibrosis TGF-beta is ground zero. The key technical challenge is how to block the anti-fibrotic activity of TGF-beta without messing with its other mission-critical activities in the body. And there are a number of good solutions to that problem.
Then what's the holdup? Fibrosis is a huge and completely unmet medical need, and the biology seems very ready for prime time. Why don't we have drugs? And why is the development pipeline finally filling now?
One reason the industry has been slow to tackle fibrosis is, in my view, a function of the organizational paradigm that has, until recently, ruled pharma companies. Fibrosis is not a disease; it's a pathology. And, as such, it runs orthogonally to the vertical therapeutic area structure that has dominated pharma for many years. While scientists in the renal group, for example, may recognize the utility of an anti-fibrotic compound in their portfolio, they are less likely to appreciate (and poorly incentivized to explore) the utility of the agent in liver and lung. That's made it challenging for companies to appreciate the strategic importance of fibrosis and build critical mass in the area. But in recent years, that's changed, spurring pharma activity and interest in this area. The evidence is right here in Big Sky; the Keystone meeting is vastly over-subscribed and a full third of the attendees are from industry.
A more significant barrier, however, has been clinical development. These are tough diseases for clinical trials. For one thing, they are slow. It can take a diabetic 30 years to develop significant kidney fibrosis; 20 years of hepatitis C infection before liver fibrosis becomes apparent. Thus, demonstrating an effect on clinical progression takes a long time. Secondly, the likely registration endpoints--pulmonary function tests in IPF, for example--are noisy and insensitive, not to mention imperfectly correlated with organ survival, so that large numbers of patients are required to obtain a statistically robust response.
Consequently, it is exceedingly difficult--I'm going to say nearly impossible--to run a typical Phase II study to evaluate the activity of a compound against clinical parameters in these diseases. The trials are, by their nature, simply too short and too small. That creates a gut-wrenching business problem: You may need to go all the way to Phase III, and spend hundreds of millions of dollars, before you know if you have an active drug and should have invested even a nickel in its development. Developers of Alzheimer's drugs are up against much the same problem. And it's a risk that most organizations simply aren't willing to stomach.
So what's changed? First of all--and this is certainly one of the developments propelling pharma interest in IPF--we finally have some clarity on the path to regulatory approval. For this, we largely have our colleagues at InterMune ($ITMN) to thank. Despite being dealt a challenging hand of clinical cards for their drug pirfenidone, they've succeeded in securing approval in the EU and have worked hard with the FDA to obtain agreement on what they need to show for approval in the U.S. Importantly, they've taught us that approval is possible based on showing robust effects on pulmonary function; you don't need to prove an effect on patient survival.
Second, although IPF development is littered with high-profile and expensive clinical failures, smoking wrecks at the side of the road, these trials have nevertheless shown us the way. Specifically, they've provided excellent data on how untreated patients fare and illuminated the natural history of the disease. Therefore, we now know how rapidly IPF patients lose lung function and that translates into solid information on how big and how long these trials need to be.
Third, the industry is finally coming to accept if you're interested in fibrosis, you need to be looking for something other than functional endpoints in your Phase II study and that, given our sophisticated grasp of the underlying biology of fibrosis, biomarker endpoints can be uniquely informative and valuable. Indeed, biomarkers can and should serve as critical go/no-go data for development of an anti-fibrotic and enable you to meaningfully evaluate your drug--is it pushing all the right buttons in target tissue?--on a much more modest investment and subdue those intestinal butterflies when you're writing the nine-figure check for the pivotal trials.
Isaac Newton said (more or less), "We stand on the shoulders of giants." Nowhere is that more true than in science--and especially in clinical development. We have a few decades of biology under our belts that provide us with a firm scientific foundation for tackling fibrosis. But what's really turned the tide in recent years are the hundreds of millions of dollars spent on failed trials, the rivers of tears shed, the armies of jobs lost. These are the experiences that have slowly but inexorably blazed the trail, bucked us up, and forced us to rethink the problem. That's why there are so many exciting drugs now motoring down that path. And none too soon. Patients are waiting.
Michael Gilman is senior vice president of early stage programs at Biogen Idec. This is his second tour of duty at the company, having returned last month when Biogen Idec acquired Stromedix, the company he founded in 2007.