The Real Hurdle in Discovering New Medicines
There are many self-styled experts who think they know how to improve the productivity of biopharmaceutical R&D organizations. They believe that by having smaller organizations, or building external networks, or having great technologies, or having biomarkers of diseases, or outsourcing large parts of your operations, you can improve R&D productivity. All of these things may indeed help gain efficiencies. But there is no guarantee that they will lead to new treatments. These are just ways to help a company to execute better. They do not solve the core problem: running the key clinical trial to prove or disprove your hypothesis to treat a disease with your new drug. No matter how good you think your preclinical science is, the ultimate test of your hypothesis occurs in long-term clinical trials. And these trials ALWAYS yield surprises which many times unfortunately are negative. Here are two such examples.
Nerve growth factor (NGF) is a protein that modulates pain through sensitization of neurons. Multiple studies in animal models of pain show that NGF can both cause and augment pain. Furthermore, blocking NGF alleviates pain. About a decade ago, scientists at Rinat, a biotech company since acquired by Pfizer, developed an antibody to NGF called tanezumab. Tanezumab worked extremely well in animal models of inflammation and so an obvious path for clinical study was to treat painful osteoarthritis of the knee, a poorly served condition.
The initial results with tanezumab were extremely exciting. Patients, for whom pain medications no longer worked and for others who had been recommended for a knee replacement, suddenly felt great. Given that the sole biological role for tanezumab was to bind to NGF and prevent its harmful effects, it was thought that this antibody would have a great safety profile as it was a very targeted drug. However, tanezumab’s downfall proved to be the great activity it led to. Patients were feeling so good that their rejuvenated active lifestyle resulted in worsening arthritis in the knee and, in some cases, total knee replacement. It turns out that complete elimination of pain in these patients is not a good thing as pain serves as a warning sign that damage is occurring.
There are other pain indications where tanezumab may prove useful. One is in cancer pain and here the risk-benefit profile of this antibody may prove to be of clinical significance. Pfizer is now studying tanezumab for this indication. But these studies will take time and cost tens of millions of dollars more. And there is no guarantee that this effort will be successful.
The explosive increase in the incidence of type 2 diabetes has been discussed in this blog and in numerous other articles. The need for new medications is acute, especially in light of the safety difficulties encountered with two commonly used marketed agents, Avandia and Actos. Thus, many people were closely following the progress of dapagliflozin, a totally new approach in treating this disease. This compound, being jointly developed by Bristol-Myers Squibb and AstraZeneca, is the first of a new class of compounds called SGLT2 inhibitors which lower blood sugar by causing it to be excreted in the urine. Furthermore, dapagliflozin also caused a small but significant drop in body weight, another important risk factor in this population.
Last week, an FDA Advisory Committee voted against the approval of dapagliflozin to treat diabetes. In a two year study of this drug in diabetics, roughly 0.4% of women got breast cancer as compared to 0.1% of women in the control group. There was also an increase in bladder cancer in men (0.3% on drug vs. 0.05% in controls). This risk-benefit profile was not deemed acceptable to the committee. Is this problem related to the mechanism of SGLT2 inhibition, or is it limited to an unknown property of dapagliflozin? No one knows and this likely won’t be known unless another biopharmaceutical company studies a different inhibitor and shows that it doesn’t share this problem.
Both the NGF and SGLT2 programs involved cutting edge science. Both programs were carried out by organizations with great experience in their respective fields. Both sought to develop new medicines to meet a major medical need. They didn’t fail because of lack of effort, resources or talent. They failed because the ultimate proof of a medical hypothesis in drug R&D does not occur in the laboratory but rather in long term clinical trials, where the complexities of human biological pathways are still not well understood. These trials are long and expensive – and they often fail.
Pharmaceutical R&D is a high risk, high reward enterprise. There are no easy pathways to getting a major new medicine approved. The FDA has already approved more new drugs in 2011 than it had all of last year. Furthermore, there are great compounds in late stage development across the industry. That’s the good news. The sad news for patients, however, is that unfortunately a good number of these will fail for reasons like those above. Outsourcing R&D, mergers, reorganizations, etc. won’t change that.