Novartis Out, Merck In, a Win for Vertex's Aurora Kinase Program

Earlier this year, Vertex and Novartis reworked their 2000 deal covering R&D of drugs against kinases. The shift has already proved consequential. Vertex had yet to present a compound to Novartis: the original arrangement called for Vertex to conduct all R&D through early clinical proof of concept and the most advanced molecule, VX-680, was just at IND filing. With Novartis unwilling to decide on VX-680 before seeing clinical data, as part of the renegotiation Vertex was able to regain control of the compound then license it to Merck. For that compound, at least, Vertex got from Merck what it couldn't get from Novartis: significant value and an earlier handoff of an asset.

Most biotechs strive to get a lead molecule into the clinic at all costs, building a clinical development infrastructure to establish product value, make the firm more attractive as an in-licensor, and in many cases help create a niche franchise. Discovery-oriented Vertex Pharmaceuticals Inc. , however, has long adopted a contrarian risk-reducing alliance model. It has sought to maximize the number of programs it can put into drug development, largely through partners--often to the dismay of Wall Street, which sees a strategy rooted in outlicensing of promising preclinical compounds as giving up value too early. To make up for the value it sacrifices by foregoing full commercialization, as well as the risks that its partners might slow development of Vertex programs, it generally insists on full control of the discovery end of R&D. (See "Vertex Sells Productivity," IN VIVO, June 2000 (Also see "Vertex Sells Productivity" - In Vivo, 1 Jun, 2000.).)

The biotech's 2000 deal with Novartis AG covering the R&D of drugs against kinase targets extended that control into clinical proof of concept. [See Deal] It called for Vertex to decide which compounds to move forward into the clinic for validation, after which Novartis would choose whether to pick them up for late-stage development and commercialization. But earlier this year, Vertex and Novartis reworked their arrangement. The handoff from Vertex to Novartis is now at preclinical, not subsequent to human clinical proof of concept studies. In exchange, Novartis can conduct R&D in the kinase field, which it had largely ceded to Vertex, including the areas where the collaboration has advanced the science.

The shift has already proved consequential. Vertex had yet to present a compound to Novartis: the most advanced, VX-680, an inhibitor of Aurora kinase, was just at IND filing. With Novartis unwilling to make a decision on VX-680 before seeing clinical data, as part of the renegotiation, Vertex was able to regain control of the drug candidate and four months' later, license it to Merck & Co. Inc. on unexpectedly favorable terms. [See Deal] For that compound, at least, Vertex got from Merck what it couldn't get from Novartis: significant value and an earlier handoff of an asset.

The Vertex/Novartis deal renegotiation came about because in the last four years both firms' R&D strategies had shifted such that the division of rights and responsibilities in the original agreement, and the attendant resource commitments, made less sense than at the outset. For Novartis, the original kinase collaboration was an intentional mega-outsourcing of R&D in a promising area of research. The goal was to deliver up to eight clinically qualified compounds into Novartis' development pipeline in six years. The research program covered all kinase targets except a narrow set that was the subject of prior R&D interest to the companies. (See "Novartis Plugs Gaps in Pharma Pipeline," IN VIVO, May 2000 (Also see "Novartis Plugs Gaps in Pharma Pipeline" - In Vivo, 1 May, 2000.).)

Novartis wasn't alone among Big Pharmas in externalizing R&D. Mired in a productivity quagmire and with deal values for promising late-stage drug candidates in the stratosphere, some large drug companies were shifting towards earlier-stage product deals. Even Merck has in the past year very publicly acknowledged that the kind of innovation that has historically driven it—innovative compounds aimed at targets discovered by Merck internal research—simply will not be enough, particularly given the increasing failure rates of clinical-stage products. (See "Jumpstarting Research: Big Pharma's Out-Licensing Dilemma," IN VIVO, March 2004 (Also see "Jumpstarting Research: Big Pharma's Out-Licensing Dilemma" - In Vivo, 1 Mar, 2004.).) In effect, discovery research continues to be important, but it is now being applied both more broadly: to re-indicate existing compounds, for example, or to figure out their mechanisms and, with that biological understanding, improve the chances of approval.

Those industry trends notwithstanding, Novartis has seen its commitment to internal R&D pay off since the 2000 deal with Vertex. Kinase inhibitors, specifically, have advanced to become a significant part of the pharma's oncology pipeline. (See "The Fits and Starts of Targeted Cancer Drug Development," START-UP, December 2003 (Also see "The Fits and Starts of Targeted Cancer Drug Development" - Scrip, 1 Dec, 2003.).)

"The partnership with Novartis was forged before Gleevec," reminds Vicki Sato, PhD, president of Vertex. Sato was referring to Novartis' targeted therapy for chronic myelogenous leukemia, a disease defined by a molecular translocation called the Philadelphia chromosome, which produces the tyrosine kinase, Bcr-abl, that Gleevec (imatinib) blocks. With the tremendous positive impact of that drug, which was approved in 2001, Novartis' desire to commit increasing internal resources to kinases—and its concurrent need for freedom to operate—has elevated in importance.

Vertex, meanwhile, has had to fight through several prominent clinical setbacks. In 2003, it restructured operations. Its core discovery strategy remained intact, but in opposition to earlier days, Vertex began to focus more on a few drugs for its own commercialization--molecules for which the firm wanted to keep at least Western hemisphere development rights and which figured to be the significant drivers of its valuation. Because it was now more heavily focused on internal programs, Vertex no longer wanted to deliver to Novartis nearly what it had envisioned four years ago.

"We wanted our development infrastructure to be largely committed to the successful development of Vertex-controlled molecules, with energy and focus," explains Sato. The issue was even more significant for the company, she says, as it began to appreciate the potential for an increasing number of compounds to come out of the collaboration with Novartis. "We became concerned that our ability to execute would require considerably expanded development infrastructure, or some dilution of attention," she adds. For Vertex to have met anything like the collaboration's original goal of selecting up to eight compounds during the life of the agreement would have meant initiating at least a half-dozen onerous clinical programs over the next two years. It just wasn't going to happen.

The definition of the hand-off was also becoming a battleground. "Vertex wants to invest so much to get to this point--we think it's a good endpoint--and Novartis of course wants us to take as much risk out as possible," Sato explains. The original definition involved a biochemical endpoint in the clinic, she says, but Novartis was pushing increasingly for a therapeutic proof of principle requiring longer studies and more time.

As part of that handoff discussion, Novartis had to decide what to do with VX-680, Vertex's most advanced compound within the kinase collaboration. The compound had not reached the selection point in the original Novartis agreement. "The original contract would have had VX-680 still under our development control for at least another year," noted Vertex CEO Josh Boger, PhD on the conference call discussing the Merck deal for VX-680. "It was only in the renegotiated contract that VX-680 was extracted from that program."

Vertex wanted to get VX-680 through development as expeditiously as possible, and not surprisingly, they wanted to move it faster than Novartis, Sato acknowledges.

At the time of the renegotiation, Novartis had seen all of the preclinical data Vertex had amassed on VX-680--including the data that supported a February 2004 paper in Nature Medicine showing VX-680 significantly reduced tumors in animal models—a publication that apparently helped stimulate Merck's and others' interest in Aurora kinase as a target. Ostensibly, VX-680 was ready for presentation under the renegotiated Novartis agreement: By not accepting the compound at its current stage of development, Novartis had effectively shown its disinterest in VX-680 and the Aurora kinase target.

"We thought the preclinical package was pretty strong," says Tony Coles, MD, Vertex's SVP, commercial operations, who doubles as head of business development for the company. "We said, ‘It'd be great for us to have an opportunity to hold onto 680 as an option and also consider it under the old agreement if we so chose.'" Four months after the renegotiation, Vertex turned around and outlicensed the molecule.

Merck now gets rights to develop VX-680 in exchange for $20 million up front, $14 million in research funding over two years, milestones, and royalties on product sales. Merck will be responsible for all clinical development and product commercialization and marketing, plus development of follow-on compounds; for what it's worth, the companies have also stated that Vertex will have the opportunity to negotiate a co-promotion right down the line. The partners expect to commence clinical trials in 2004.

Researchers have implicated Aurora kinases, a family of serine-threonine kinases, in the onset and progression of many tumor types. They downregulate the tumor suppressor gene, p53, and are potentially important control points for slowing the growth and spread of cancer cells, including solid tumors. Aurora kinases are overexpressed in colon cancer, breast cancer, and leukemia, and amplification of Aurora genes is associated with progression of colorectal cancer and poor prognosis in some breast cancers.

The companies will conduct a joint research program to characterize VX-680's activity across a broad range of cancer types, and will also develop follow-on compounds under the agreement using molecular profiling approaches and microarray technologies resident within Merck.

Winners and Losers

Vertex got from Merck what it couldn't get from Novartis: an early handoff of VX-680. Moreover, viewed with the hindsight of the Merck deal, the earlier handoff did not seem to diminish the value of the Aurora kinase set of assets. Under the Novartis deal, pre-approval payments for an accepted compound are $35 million. Vertex got $34 million in hard cash and committed funding, and knows that Merck is also likely to develop follow-ons to VX-680 as a matter of course. And with Merck, not only is Vertex getting the earlier handoff of VX-680, it is getting a partner perhaps better aligned to evaluate in-licensed preclinical oncology drug candidates. "Novartis was a good partner [for VX-680]—the challenge for them is they have a lot of compounds coming," says Sato. "What I know with Merck is they will use profiling for patient selection and early assessment of activity."

Both Novartis and Vertex state firmly that VX-680 was not a significant driver for the renegotiation. "There were lots of moving pieces during the negotiations," says Coles. "They served a strategic purpose for both Vertex and Novartis." But giving Vertex the option to develop VX-680 under the old agreement (under which it could still borrow the money for development from Novartis, and have the loan forgiven should Novartis accept the compound for development) was to some extent a quid Novartis was willing to offer. Novartis may expect Vertex to present compounds with more appealing preclinical profiles than VX-680 between now and May 2006, when the agreement ends—both companies say their goal remains presentation of the same number of compounds as expected under the original deal, with the significant difference, of course, that they will not have passed early clinical muster. Novartis may have tacitly given up on Vertex, but the renegotiation did not remove the committed research funding. (Alternatively, the research funding could be viewed simply as the price of obtaining two years' freedom to operate in the kinase field.)

Given how well the renegotiation appears to meet Vertex's needs, securing freedom to operate in kinase R&D must have meant a lot to Novartis. Vertex continues to receive research funding through April2006 consistent with the original agreement--the R&D portion of the Novartis deal accounted for 64% of Vertex's total revenues for 2003, or approximately $44 million--and up to $35million in pre-commercial payments for each preclinical drug candidate that Novartis accepts for development. "The economics of the deal didn't shift," notes Coles.

But for Vertex, the renegotiation is about more than continuing the research to the end of its term. "The revised deal is fine so long as Novartis picks compounds," he acknowledges—just finishing up the research collaboration, in areas where the partners could end up competitors, won't cut it.

Has the invigoration of R&D within Novartis dulled its appetite for in-licensing? The pharma rejects the notion. Citing a October 2003 collaboration with Myogen Inc. on the discovery and development of cardiovascular drugs [See Deal] and its May 2004 antibody discovery with MorphoSys AG [See Deal], Victor Hartmann, MD, Novartis' head of global business development and licensing, declares: "Early-stage deals are still of interest to us. In fact, Novartis Institutes for Biomedical Research has done 33 new deals and renegotiated eight others in 2004," he points out. Nor did Novartis pass on VX-680 to favor an internal program over a collaborative one. "It's not applicable to say we didn't take 680 so we could put the money elsewhere," Hartmann says.