Pharmaceutical companies on average spent $1.1 billion to develop and launch a new drug during the late 1990s, according to a 2009 report by Bain Capital. By 2009 the required investment had more than doubled to $2.2 billion, and the return on invested capital for new-drug development had dropped from 9% in 1995–2000 to 4% in 2009.
Prompted by the need to refurbish their pipelines at lower costs and to drive and sustain future profits, drug companies have turned to drug repositioning, or repurposing, as a means of drug rediscovery. The process of drug repositioning identifies novel indications for clinical candidates that have been discontinued for their primary indications for reasons other than safety.
Pfizer’s Viagra to treat erectile dysfunction exemplifies the successful resurrection of a drug as it was originally developed to treat high blood pressure and cardiac problems. Other successfully repositioned drugs include raloxifene (Evista from Eli Lilly) and thalidomide (Thalomid from Celgene).
FDA approval of Eli Lilly’s Cymbalta (duloxetine), a serotonin-norepinephrine reuptake inhibitor for fibromyalgia, offers another example of successful drug repositioning program. Initially developed as an antidepressant, the FDA approved its use for fibromyalgia in 2008.
While drug repositioning offers considerable advantages over conventional drug discovery, it has its own technical, intellectual property, and regulatory challenges.
At CHI’s “Drug Repositioning Summit” held late last year, several companies and research enterprises presented technologies and strategies for facilitating drug repositioning.
Ken Phelps, president and CEO of Camargo Pharmaceutical Services, discussed a regulatory path that was originally designed to facilitate approval of improvements in existing drugs, the 505(b)(2) New Drug Application.
The 505(b)(2) NDA is allowed to reference information that, in a normal 505(b)(1), would be conducted by the sponsor. “Thus, let’s say,” Phelps explained, “that to be approved, your repositioned drug requires a two-year carcinogenic study in an animal. You find in the approval of the original formulation a two-year carcinogenic study. You can reference that study to meet the approval requirements for your drug product.”
In distinguishing between the two approval paths, Phelps noted that both 505(b)(1) and 505(b)(2) have to meet the same standards for approval. “It just depends on where the data comes from.”
With regard to the relative advantages of going the 505(b)(2) route, Phelps said that it involved fewer studies than the traditional 505(b)(1), and, he added, that, “most people see it as entailing lower costs and taking less time. And because the drug is already known, it entails less risk that studies will fail.”
Verva Pharmaceuticals develops therapies for diabetes and obesity. Its lead product, VVP808, is a nonthiazolidinedione insulin sensitizer that is entering a proof-of-concept clinical study for type 2 diabetes. VVP808, a repositioned drug, was previously approved in North America as a glaucoma therapeutic.
Verva was able to identify the compound’s unexpected antidiabetic activity using its Gene Expression Signature (GES) drug discovery platform. According to Vince Wacher, Ph.D., CEO, VVP808’s potential as a diabetes drug was discovered by comparing gene-expression signatures of normal fat cells, insulin-resistant fat cells (made by exposing normal fat cells to TNF-a), and insulin-resistant fat cells treated with a cocktail of diabetes-reversing agents.
In comparing the data, Dr. Wacher said that the company found a 7–15 gene set that serves as a signature for reversal of insulin resistance. “These genes are from different families and aren’t necessarily the target for VVP808. Rather, changes in their collective expression serve as a gauge of insulin resistance.
“VVP808 had not previously been identified to have an antidiabetic effect. Interestingly, while VVP808 inhibits a known enzyme activity in its glaucoma indication, inhibition of this enzyme does not appear to be responsible for its insulin-sensitizing effect in our models.
“In this regard GES provided a ‘two-for one’ with VVP808—identifying a novel insulin sensitizer that should also lead us to a new diabetes target. Our studies so far show that VVP808 doesn’t modify PPAR, DPP4, or other established diabetes targets.”
Pfizer’s Indications Discovery Unit was established in 2007 to reposition failed compounds and identify additional indications for existing clinical compounds. The goal was to develop a coordinated, systematic approach to identify new therapeutic opportunities for these assets.
Michael Barratt, Ph.D., technology leader of the unit, described Pfizer’s approach and presented case studies on discoveries to date. Dr. Barratt explained that his group has “assembled a database of the entire Pfizer clinical compound collection, past and present. We are looking at the highest value compounds where there is a significant amount of data built up from the primary indication that can be leveraged for our efforts.
“Traditionally, large pharmaceutical companies often progress drugs for diseases in a sequential fashion, considering repositioning only when studies in the primary indication have played out.”
In addition to bioinformatics, analysis of human genetics, and clinical data mining, the Pfizer group has developed a panel of over 50 in vitro and in vivo phenotypic models that it calls DRIVAs (disease relevant in vitro (and in vivo) assays), through which the clinical compound collection has been screened. The selection of assays for inclusion in this panel was primarily based on prior knowledge of the predictive nature of the assay and relevance to human disease. Endpoints are thus generally functional in nature, i.e., neurite outgrowth, angiogenesis, cell motility, and glucose uptake.
“Some of the biggest initial challenges to implementing this highly systematic approach were cultural, and these are likely universal across the industry,” Dr. Barratt explained. “Development project teams typically own their compound and are concerned that any additional evaluation of a clinical compound would lead to unfavorable even spurious results or delays to their programs.
“We addressed these legitimate concerns in two ways. First, we partnered with colleagues in our safety sciences group and clinical development teams to develop processes that carefully control the types of studies, as well as the level and duration of drug exposure in preclinical models.
“Second, we gained endorsement and strong support from senior level executives for our strategy and discussed proposed studies with such compounds in detail with the leaders of the development teams for their primary indications. Investing significant time in building understanding and trust across multiple groups and project teams has been a key factor in overcoming early barriers.”
Another challenge, “has just been to select the best opportunities to progress from Pfizer’s vast portfolio of compounds,” Dr. Barratt added. But this scale has also presented its own hurdles. “Physically accessing compound data that was often spread across an array of internal databases and legacy systems (particularly for older shelved compounds that have been subject to various mergers and acquisitions) made the compilation of a cohesive summary of all of the key compound properties difficult.
“We have invested heavily in approaches and technologies to automatically link relevant data from multiple primary databases on a compound-by-compound basis. An added advantage is that these tools are now being adopted by other groups in Pfizer as their value has become apparent more widely in the organization.
“With the group’s objective being to deliver positive proof-of-concept clinical data in areas of unmet medical need, we have already been successful in identifying a number of new opportunities now in Phase I and Phase II trials. We are optimistic that the next two to three years will prove the value of our approach to Pfizer and the industry as a whole.”