Last week, Gene Network Sciences disclosed the first pharmaceutical customer for its computer modeling technology — a drug-development deal with Johnson & Johnson Pharmaceutical Research and Development.
Under the agreement, J&JPRD will use GNS’s pathway inference technology along with its mechanistic modeling approach to study the mechanism of action of an undisclosed pre-clinical oncology compound.
“We have information about what types of targets are inhibited by a compound,” Hans Winkler, senior director of functional genomics at J&JPRD, told BioInform. “So we’re working on trying to figure out whether it is really those targets that contribute the most to the mechanism of action of the compound … which is relatively complex because it inhibits several different targets and we don’t know what the relative contribution is.”
The agreement follows a string of deals indicating that drug-discovery firms are beginning to take computational systems biology a bit more seriously. Last month, Entelos announced that it had gained commercialization rights to any therapeutic compounds that may result from an expanded agreement with Organon [BioInform 02-21-05]. In the fall, Bayer teamed up with Physiomics to co-develop technology for clinical response prediction [BioInform 09-20-04].
In addition, pathway informatics firms like Ingenuity, GeneGo, and Jubilant Biosys have all recently reported new or expanded deals with pharmaceutical firms [BioInform 02-07-05].
Even academic groups are reporting mounting interest from industry. Tim Gardner, a researcher at Boston University’s Center for Biodynamics and co-author on a paper in the March issue of Nature Biotechnology that describes the use of network modeling to investigate the mechanism of action of a potential cancer compound [Nature Biotechnology 23, 377-383 (2005)], told BioInform that his research team has “spoken with Pfizer” about a possible collaboration, and that several other commercial groups “have expressed interest in pilot studies.”
GNS CEO and co-founder Colin Hill said that the J&J contract is “validation for both the technology platform and the commercial model, the actual business model, of doing a drug-development alliance deal. Our model is to do these deals around specific compounds, or at least specific compound classes, and we believe that it’s a rather repeatable type of business, and scalable.”
GNS was launched five years ago with the goal of applying computational systems biology approaches to cancer research. Since then, GNS has worked with other pharma companies in oncology, but “this is the first one that we’ve made public,” Hill said.
Financial terms of the agreement were not disclosed, but Hill said it is “a pretty significant deal” for the young company.
Hill said that GNS is seeing more interest from drug-discovery firms looking to use in silico approaches to gain a better understanding of how compounds affect molecular pathways. The approach is starting to “percolate through to more of the mainstream folks,” Hill said, though he was quick to acknowledge that it’s still not a “bread-and-butter” approach within pharma.
“I think there’s starting to be a greater appreciation and acceptance [among pharma] that there is this more sophisticated way to go about understanding what these compounds do, and that this understanding of compound mechanisms is really connected to improving drug-development success rates,” he said.
Winkler said that “there is a certain degree of validation of the technology already out there,” and that J&J is “testing more specifically the things that we have in mind.” GNS convinced J&J to try the approach based on its published research papers, as well as anonymized results from prior pharma projects, Winkler said.
The collaboration is still in its early days, but Winkler said it is going well so far. “We see very clearly now … that the major mechanisms that we thought were going to be active were active — they have definitely confirmed that. And we’re beginning to see some other mechanisms that we didn’t expect, but we have to validate them.”
Evaluating the effectiveness of the technology may prove difficult, however. Winkler said that because the J&J team is applying a new technology to an experimental compound, “it’s a bit of a two-edged sword.”
On the one hand, he said, “we want to learn a little bit about the compound, on the other hand we want to learn something about the technology, so if it is all successful, and we are able to identify an alternative mechanism, that would be a good success criteria. On the other hand, if we fail to do so, it’s not going to be entirely clear whether it’s due to failure of the technology or whether our compound just doesn’t inhibit another mechanism.”
Nevertheless, he said, the approach has already proven useful, particularly in the area of microarray analysis, because “it really gives you some mechanistic information about what the up- and down-regulation of these sets of genes actually mean — what are the underlying pathways that are inhibited or activated.”
Winkler added that the approach could potentially “be very powerful in assessing safety.” But even though a number of academic and commercial researchers have speculated that recent toxicity concerns around Vioxx and other drugs may drive adoption of in silico approaches, Hill was cautious in his assessment of the use of computer modeling in this area.
“I don’t want to make too bold a claim about [safety concerns] now driving what’s happening in systems biology, but I think it’s all part of the mix,” he said. “I think there is more acknowledgment … that we really need to reexamine how things are being approached. And if there is a way to see some of these pitfalls ahead of time, I think [pharmaceutical companies are] more open to adopting them.”
Some observers have noted that the NIH’s recently launched Roadmap for Medical Research, which emphasizes pathway-based studies, as well as the FDA’s Critical Path initiative to speed the adoption of new drugs via emerging technologies, have also helped stimulate interest in computational systems biology.
“I think it’s the sum total of all of these things,” Hill said. “Plus the fact that the data is better, and I think there’s also increased competition among various drug development efforts, that everybody’s looking for an edge.”