The Antidote: Inside the World of New Pharma (11 page)

“Josh wanted to reorganize around science and wanted to make Vicki president of the company, and for me, it just seemed a good time to go. I did think it wasn’t the best thing for Vertex. I have a lot of respect for Vicki, but that was not an arrangement that there was any chance was going to work for me, frankly. And I don’t think that anyone else thought it was going to work, other than Josh.”

JANUARY 22, 2001

Two days after George W. Bush was inaugurated in a bone-chilling drizzle, the FDA approved Schering-Plough’s Pegintron for the treatment of patients with chronic hepatitis C. The drug was the first once-weekly injection of genetically engineered alpha interferon, a biological molecule released to help uninfected host cells resist new infection by a virus. In a study of more than a thousand patients comparing Pegintron to the company’s decade-old Intron A, which was shorter acting and taken three times a week, it doubled the cure rate to 24 percent when taken for forty-eight weeks. More than half the patients in the study complained of flu-like reactions: fevers, chills, muscle aches, sweating, exhaustion. A third of them reported being depressed. Roche, too, had filed for approval for a longer-acting interferon, Pegasys, setting up a marketing war as both companies began further studies combining the new medicines with a second broad-spectrum antiviral, ribavirin.

In the collaboration with Lilly, Vertex rediscovered that HCV wouldn’t yield to the usual stratagems. The front end of any conventional drug discovery effort consists of a screening assay, wherein a multitude of compounds are tested for biological activity. Most screens are deliberately set up so that about 1 percent of the molecules are hits, but Vertex and Lilly devised one so sensitive that any detectable hit would cause a signal. “Lilly screened their entire sample collection against HCV protease,” Boger recalls, “and they had none. Zero verified hits. It was useful to have done that experiment, because even for a company like Lilly,
it did sort of cement internally that there was no other way to do this except design.”

For the chemists and modelers, the goal wasn’t just to invent a molecule that blocked the protease but to fulfill the towering requirements that make all anti-infectives, especially direct-acting antiviral drugs, so hard to develop. The key problem is resistance: the virus, replicating rapidly, evolves variants that allow it to evade a new threat to its survival. Medicinal chemists are skilled at molecular subterfuge, substituting groups of atoms that cause corresponding groups on a target to bind to them instead of their usual partners. But a variety of other properties ultimately rule: the mix of features that make a compound “druggable”—that is, safe, soluble, and stable enough to become an approved medicine. In the end it must be formulated to deliver a pure, precise dose in milligrams to patients while being manufactured—cheaply, reliably, safely, and competitively—in multi-ton lots.

“There are lots of ways to design a potent compound against an active site, and one of them is to put down a lot of grease that touches a lot of grease,” Boger says. “You can lob down a lot of stuff that touches a lot of stuff. But the problem is, if you rely on that kind of strategy, you’re just asking something like a viral enzyme to mutate and knock your compound out. So we did a lot of dynamics with the enzyme structure to see what kind of flexibility it had. And we did a lot of work trying to ask the question, not what’s the best way to make a potent compound, but what’s the best way to make a potent compound that the enzyme will have the hardest time kicking out—and, oh, by the way, because of the nature of the active site, that doesn’t turn out to be brick dust [about as soluble as sand]. It was a really hard problem. It took a long time.”

Vertex and Lilly clashed at every turn. Key to designing a drug is determining where, and in what concentrations, it should collect in the body. The rule of thumb for most drugs—and most drugmakers—is that a molecule should be small enough and soluble enough to circulate in the blood and be excreted in the urine. Large lipid-loving compounds attracted to fats and waxes—grease—are removed from the plasma and gather in the liver, which makes fat and absorbs toxins. No one knew for certain where HCV hid. But a year into the project Kwong had a key
insight while sitting in an educational session at a meeting on liver diseases. A transplant surgeon showed a slide comparing RNA levels of the virus in patients just before and after they had their livers replaced. There was a precipitous drop. “
Boom
,” she recalls. “Back then it was controversial where the virus replicated. Well, my God, I don’t know where else it replicated, but it definitely replicates in the liver. I came back, and I said to the team, ‘We need to target the liver.’ ”

Lilly resisted the idea, passively at first. Its corporate culture was conventional, rigorous, Midwestern, and orthodox, and also, tinged with fresh embarrassment and regret over the dissolution of another recent partnership in antivirals. Lilly had sponsored Agouron’s HIV program, then decided it didn’t want to develop the molecule and opted out of the collaboration; as a result, Agouron, in partnership with Japan Tobacco, took back worldwide commercial rights to what would become a billion-dollar drug. In discussions with Lilly’s scientists, Tung and others from Vertex encountered mounting disagreement over their fundamental goals, even the drug-like profile they thought they had committed to pursuing together. “One of the questions was, is the virus replicating solely in the liver, or are there other reservoirs outside the liver where it’s replicating elsewhere and becoming resistant?” Tung recalls. “We had a tremendous disagreement with Lilly over this issue. Their received wisdom is you’ve got to have significant blood levels of the drug.”

Lilly’s concept of a druggable molecule was limited by what researchers already knew about all existing drugs and their chemical interactions with the five hundred or so known protein targets in the body—a statistical approach preferred throughout Big Pharma for narrowing leads that’s heavily biased toward compounds that are small, soluble, and well behaved. After much intellectual soul-searching, Tung, Thomson, Murcko, and Pravin Chaturvedi, Vertex’s head of pharmacokinetics, developed a drug profile that fell outside all its major parameters. “Lilly was arguing for a very mundane, well-known computational algorithm that basically works only within its memory banks,” Thomson says. “It can interpolate and piece together successful combinations of the pieces of drugs that we’ve already tried. But it can’t invent new ones. In essence, they were trying to tell us that their new ultraslick horse and buggy was
far more sophisticated than our Ferrari because they didn’t get what the technology was that made the Ferrari.”

Sato, becoming directly involved, insisted that the collaboration focus on molecules that went to the liver. Hepatitis C, after all, was a disease of that organ. More crucially, an amended agreement was needed before Vertex could start to develop relevant animal models to test if its compounds stopped the virus where it needed to be stopped. “Vicki certainly, as she will always do, let them know who’s boss, and they responded accordingly,” Tung recalls. The Lilly team, already souring on the partnership, continued to make new variants on Vertex’s design even as more resistance to the program erupted among the process chemists at Lilly, who were struggling to scale up production of the final class of compounds so they could be tested in animals and humans. Vertex chemist Dave Deininger recalls, “Their preclinical development people said, ‘We can’t do anything with this. It’s got none of the characteristics that say, “Oh yeah, that’s gonna be a great drug.” It’s too big, it’s too greasy, it’s peptidal. It’s a rule breaker of all of the rules.’ ”

Highly crystalline, the eventual drug candidate—VX-950, made by a Lilly chemist and based on a Vertex design—was “pretty much brick dust,” as Boger put it. It was less soluble than marble. Vertex had no formulation group of its own, yet as its faith in Lilly’s group faded, Boger could see that HCV might reprise HIV at Vertex: a research triumph but a commercial disappointment, based largely on his partner’s difficulty in getting its molecules into pills and down people’s throats. It was a problem he recognized but had no way of controlling.

A breakneck expansion strategy, especially in a fast-growing high-tech sector, guarantees a business leader a rare degree of power and influence within the company. As chairman and CEO, Boger had more or less blank-check support from the board of directors, most of whom he had recruited after the original members moved on to newer start-ups, and from its first chairman, former head of the war on cancer and pioneering venture capitalist Benno Schmidt Sr., retired.

His executive team now reported to Sato, who elevated the triumvirate that had made its mark with HIV and ever since then had best
anticipated and met the broad challenges Boger laid out—the hits, you might say, in his social experiment. Thomson became vice president of research; Murcko, chief technology officer and chairman of the scientific advisory board; and Tung, vice president of chemistry and head of compound selection. Under the redesign, Tung reported to Thomson, but Murcko, with a relatively small group, retained an independent standing as the company’s in-house big thinker, practical visionary, and prodder.

Boger aggressively promoted chemogenomics as the future of the industry, if not a panacea for everything that ailed Big Pharma, something very close to it. With Agenerase on the market, eight drugs in human testing, and hopes of putting five to seven early-stage drug candidates in the clinic during the next twelve months, he told investors that by 2005, Vertex would start submitting two to three drugs a year for FDA approval. The figure, as he noted, was twice that of Novartis and Vertex’s newly reconstituted partner in ICE, French drugmaker Aventis, formerly HMR. It was an exorbitant claim, one depending on strong, unambiguous findings from its clinical trials and flawless execution of many functions that the company had yet to incorporate.

Even as the genomics bubble burst amid spectacular flameouts of once-promising single-product biotech companies, and as the slump only deepened in pharmaceutical R&D, Boger’s brash evangelism kept the company’s share price afloat above $40, twice what it was a year earlier. “My goal is to grow up and overtake Pfizer, not to be Genentech,” he told
BusinessWeek
. Pfizer was the world’s largest and richest drugmaker, purveyor of Viagra, the fastest-selling pharmaceutical product in history. It recently announced it would spend $5 billion on research in 2001. Former highflier Genentech, the world’s prototype biopharmaceutical firm, still had no blockbuster after twenty-five years in business, though recently it had begun selling two breakthrough cancer drugs. Powerful scientifically, lauded for its academic-like corporate culture, a pioneer in off-label sales with its strenuous promotion of recombinant human growth hormone for children whose only disability was that they might grow up to be short, the company had become synonymous with the industry’s overly aggressive marketing and oversold hopes. Its stock, which crested at $85 in the bubble, now traded at about $20.

To optimize the research engine it was assembling, Vertex went looking for a company to buy that could quickly do for it what it couldn’t do for itself: namely, provide a way to screen large decks of compounds against its fast-growing array of newly discovered targets, then test the molecules in cells to see what effect they had. High-throughput screening was the very opposite of structure-based design: an attempt to focus and accelerate discovery not with precise molecular knowledge but with advanced robotics, miniaturization, and proprietary methods for culling new hits. But in the new era of digital information sharing and genomics—call it
infonomics
—Boger believed the two approaches could complement each other.

One business soon stood out: Aurora Biosciences Corporation of San Diego. Cofounded by future Nobel Prize–winning chemist Roger Tsien, the company led the industry in developing assays, screening, and cell biology. It had drug discovery programs in numerous areas, but its $60 million in revenues came chiefly from providing screening services to more than fifteen major life-sciences companies and research organizations, and its management felt that a merger with Vertex could help it evolve from a screening site into a fully integrated drug discovery company. Aurora’s three hundred employees worked in a gleaming white two-story industrial building in a grassy R&D park on a bluff overlooking, to the east, the freeways and metastasizing sprawl north of downtown. A good golfer with a Santa Ana wind at his or her back could stand in the parking lot, facing an uphill lie, and hit the famed Torrey Pines championship course with a strong drive and a long three-wood.

After three months of negotiations and strategizing about how to put the two companies together, Vertex announced on May 1 that it would acquire Aurora for $592 million in stock. “We are going to seize as much of this ground as we can,” Boger told the
New York Times
. “This is not something we can come back to in twenty years. The fun will all be over.” From Aurora’s perspective, the acquisition promised considerable, but not total, independence. After the merger, Aurora would operate as a wholly owned subsidiary of Vertex, but all Aurora partnerships would need the approval of Vertex management. Aurora’s chairman, chief executive, and president, Stuart Collinson, would sit on Vertex’s board.