by Barry Werth
The larger problem was to understand—and then show—how a drug or combination of drugs could achieve sustained viral response (SVR): in other words, a cure. The key with killing off viruses is not how to exterminate them (you can use bleach) but how much of a substance to dose people with that will be both effective and safe. To rate the effectiveness of a molecule, researchers traditionally test how much of the substance it takes to knock down a targeted biological activity by half (IC50). Since trillions of HCV particles are made daily in infected people, numerical drops are measured in “logs”—each log being a power of 10. A 2-log drop is a 100-fold decrease; a 4-log drop, a 10,000-fold decline.
Inside Vertex, there were well-founded qualms about VX-950’s potency, and Kwong made it her mission to dispel them by developing a better test of effectiveness. “The reason that we were in trouble is that BILN-2061 was 350-fold more potent than we were in a certain assay,” she recalls. That assay compared effectiveness at a 2-log drop. But when Kwong ran a multi-log drop essay comparing 2061 side by side with VX-950, “we weren’t in spitting distance of each other,” she recalls.
Kwong and Vertex now realized that the longer you incubate with the drug, and the more drug you put in, the more effectively you could clear the virus. While the hepatitis C team pressed ahead with animal studies to establish an optimal range of dose regimens for human testing, Ian Smith and Ken Boger launched timely financial and legal maneuvers aimed at reducing the company’s exposure while squeezing more value from its assets. Smith went back to some of Vertex’s significant bondholders and proposed a private restructuring of their notes, offering, essentially, a 6-to-1 split. “Typically, if you’re a bondholder, you hold the company hostage,” he says. “You don’t go, ‘Yeah, I’ll take a risk on your equity.’ They say, ‘Give me the damn cash. I’ll take sixty cents on the dollar instead, as opposed to taking your stock, which might be worthless.’ ”
Since the crash of pralnacasan, Vertex had seen perhaps its most precious asset among investors—the benefit of the doubt—dry up. “We were being treated like everyone else,” Boger recalls. Of sudden and equal concern was brewing internal doubt. John Alam, in particular, was facing the realization that Vertex’s target-based discovery engine might be steaming down the wrong track: that outside of anti-infectives, a few cancers, and some rare genetic diseases, the idea that you could inhibit a single target with a small molecule to stop a disease was “a false belief.” Still, Smith found that there remained sufficient hope in Vertex’s business. He was able to round up enough investors willing to convert their shares, removing $320 million in debt from the balance sheet.
Ken pressed to retrofit the terms of the Novartis partnership by arranging for an earlier, more rapid stage transfer of Vertex’s drug candidates to Novartis for clinical development. So far Novartis hadn’t accepted any of Vertex’s molecules, claiming that Vertex hadn’t shown sufficient clinical relevance. Ken, after persuading Sato and Boger that the present arrangement was unsustainable, convinced Novartis to come to terms, splitting the research efforts to give Novartis first-option rights on Vertex compounds, while allowing it to push ahead in kinases on its own. He recalls:
I proposed a way to redo the deal to knock out this proof-of-concept stuff, and that said to Novartis, “You have to pick the compound when we propose it for development, and if you don’t pick the compound that we propose, then we not only get the compound, we get the whole target.” I was pretty sure that I could get Novartis to go along with it, because I knew they had a problem that was ongoing.
They were violating the agreement, big-time, and I had made a decision from a business development and legal standpoint not to call them on this. I was in a “Don’t ask, don’t tell” mode. I knew they were breaching the agreement by doing work on the side with kinases that were supposed to be exclusive to us, and the people who were doing the work were really influential inside Novartis. They were the people who were supposed to be killed off by the deal with us, but they didn’t get killed off, they grew. I figured, let’s just let this sit out there and use this someday.
Novartis agreed to the amended deal in early February. Vertex now was free to shop those kinase projects that Novartis had rejected, and Boger approached an old friend at Merck who’d just been put in charge of a new cancer effort, proposing a collaboration in Aurora kinases, a group that Vertex had first shown to be a potentially important class of targets. The company had a potent inhibitor that profoundly reduced tumor growth in cancer models. At the same time, Sato, Kwong, and Tony Coles’s business development team took Kwong’s data on the road, meeting with dozens of potential partners in HCV. As Porges had warned, the company had until June to line up another collaboration, or the view on Wall Street, still scraping near its nadir, would again turn negative.
Boger had told the Harvard Business School team, “A lot of biotechs are founded on the German academic model, with a couple of principal investigators and their closest troops. As the firm expands, later employees are considered less important. In contrast, we believe that the last person in the door is just as important as the first. We consciously reject the German model for the Silicon Valley model.” But what if the last person in the door is an actual German academic? And what if he’s handed the keys? The gradual immersion of chief scientist Peter Mueller into Vertex’s social experiment was bound to roil Boger’s hypothesis.
At an off-site, all-day Saturday meeting—on Valentine’s Day—Mueller proposed a new strategic vision for the company, attempting to galvanize Boger’s and Sato’s fervor for hepatitis C into action. Sizing up Vertex’s situation from the perspective of getting a drug out the door and becoming a sustainable pharmaceutical company, he argued that it first had to free itself from its dependence on Big Pharma. “With those types of arrangements, there are some obligations linked to it, and you have not the freedom to operate that you would otherwise have,” he recalls. “You’re more or less a slave of those partners. They provide development support, but at the end of the day it doesn’t drive a small company to the next stage of evolution.
“The other thing was we had in the pipeline molecules that I would say were so-so. We had to maintain those activities, but to move forward, you have to demonstrate that you can do something on your own. Given the small amount of dollars that we had, those activities were very, very limited.”
The issue was how to advance the company. The group was restive and large—maybe fifty in all—more or less the same body that had conducted the portfolio review. The turbulence, shocks, reversals, and doubts of the previous nine months lingered, and feelings were raw, conflicted. More than a few regarded Boger as not quite the same Pied Piper he had been, a casualty of his own inflated expectations. Merimepodib was advancing but was “crippled,” Alam says—the best of an uninspired lot.
Mueller made his case. “I strongly came forward with the idea that we have to develop the anti-infective drug VX-950, because it was probably the highest chance of success from a clinical development point of view,” he says.
There was a huge debate. Can we afford that? It’s very difficult. It’s long trials because standard of care is forty-eight weeks. Cost of goods of the molecule were outrageous, when you have to produce a couple of kilograms, and a kilogram costs $2.5 million, and it’s cumbersome to make. People get nervous because it takes away money and capacity from other projects.
It also meant that we had to make a commitment to build a development organization. That was actually the more significant piece, because then you have to go and do financing. We didn’t have the development capabilities that you need to make a drug all the way through the end. You had to basically build this more or less from scratch, with all the functions that are needed underneath. There is chemical development, analytical chemistry, formulation development, a quality environment. We also had to think about a manufacturing buildout, because we couldn’t afford to put a plant in the ground for a couple of hundred million bucks. The c
linical end was the same thing. You need a set-up in clinical that has all the components that you need. We’re going to need a strong regulatory environment, strong clinical development environment, strong clinical operations that can handle bigger and complex trials across the globe.
Boger agreed with Mueller that VX-950 was Vertex’s best hope for a medicine that transformed the lives of patients. This was a decision he had positioned himself to face ever since he’d committed Vertex to getting as many projects as possible out of the lab and into patients. He was not without other options, which complicated his choices. He told the group that Vertex would mount a full-scale development effort with VX-950, effectively freezing some IMPDH projects that were further advanced. “That was a tough call,” Boger says:
It wasn’t a company killer. It wasn’t we have to keep this going because this is all we have. But that made it even tougher. This was a program that was being run at the thirty-five-million-a-year level for which suddenly the whole bill was ours—unexpectedly. It doesn’t matter if Lilly says publicly that they’ve canceled all their infectious disease programs. Everybody thinks that they wouldn’t cancel anything they thought was valuable. So you have to convince Wall Street, convince the board, convince everybody, that it didn’t matter what Lilly said, this was worth finding thirty-five million a year to keep going, knowing that it was going to go to a hundred million real soon. And why this? If we had that kind of financial resources, why didn’t we put it behind something that wasn’t tainted that way. The typical biotech thing to do was to say, “Well, we’re putting this program on hold until we find another partner”—which would have happened. Instead, we just went full speed ahead.
Mueller prides himself on his Bavarian understatement. He knew better than anyone else, including Boger, the full implications ahead for the organization, and he admired the courage it took to take on the risks. Boger often told people that he wanted Vertex to be the most feared pharmaceutical company and the most fun pharmaceutical company: feared because of its fearlessness, fun because it was more exciting to dare, excel, and win against all odds than simply outperform your rivals. Sato thought the combination derived from the fear of falling short of one’s sacred ambitions, of not being good enough, and the deep personal pleasure and raw thrill one takes in an almost psychedelic level of hard work and engagement in several major challenges all at once—at life, exultant, breathing hard but evenly, on the cutting edge. Even at the bottom of Vertex’s fortunes on Wall Street, she recalls, “The Kool-Aid was strong.”
“So we made this commitment that we could solve all those problems and do it,” Mueller says. “That was actually the fundamental moment when the decision was made: ‘Okay, we do this drug, and along the line we build an organization to make it all happen.’ This is one of the boldest moments that I have seen from a corporate decision-making point of view. I must say I give kudos to Josh and his guts. He stood up and said, ‘Yeah, we do this.’ ”
As the second quarter advanced, Vertex management regrouped while making the sort of partnership deals that the company, now more than ever, needed to do to survive. Although it had more than $460 million in cash, Smith expected a net loss for the year of $140 million to $150 million, meaning its balance would soon drop near $300 million. Despite better-than-expected royalties from Lexiva, at its current burn rate the company had two years or less before it ran out of funds. Losses as far as the eye could see, the founding promise of all biotechs, had started to add up to real money, ratcheting up pressure on the executive team.
In May the company announced that the Cystic Fibrosis Foundation would pay $21 million in direct research funding for the next two years. Bob Beall, pleased that Vertex had proved it could find small-molecule compounds that corrected the defective ion transport in the lab, was eager to subsidize its late-stage drug discovery effort. As Aldrich had predicted, Vertex’s full commercial rights in CF—a so-called orphan category since it has so few patients, and which affords, among other inducements, accelerated time frames to approval and superpremium pricing—looked more valuable now that development deals with big companies had lost much of their allure for both sides.
Sato scrambled hard to find a limited partner to subsidize the work in HCV, one that would deliver near-term revenue and supply badly needed development support while the company geared up VX-950—and itself—to conduct large-scale human trials. She was under the gun to bring in something fast: the first human study was scheduled to start in June. With Boger uncharacteristically staying out of the discussions, she approved a $33 million collaboration with the Japanese drugmaker Mitsubishi for development and commercialization rights to the drug in the Far East. It wouldn’t cover the bills but it was something. Vertex retained exclusive rights to the rest of the world.
In mid-June, the day the agreement was signed, VX-950 was given to the first of thirty-five healthy volunteers in Europe to test its safety, tolerability, and pharmacokinetics in an early-stage dosing trial. That morning in Cambridge, chemical engineer Trish Hurter, forty, joined Vertex, taking over its formulation group. Hurter, compact and hugely energetic, a spirited horsewoman, was from South Africa, having gotten her PhD from MIT before migrating into and out of the paper industry, having two children, and managing, most recently, formulation development at Merck. At Vertex, where most discussion galloped along like the dialogue in The Social Network, Hurter was instantly in a speed class by herself: “Walk fast, talk fast, ride fast, drive fast, type fast,” she explains. She adds, in another context: “Shit happens, move on.”
What the start of human testing signified is that VX-950 could be coaxed, despite being less soluble than marble, into a state where it could be taken orally, get through the digestive track, and be absorbed, three to four hours later, through the lining of the small intestine and into the bloodstream. What the experimental subject got was not a pill but a cloudy suspension made in a beaker from ingredients stable for only a day or so. Concentrations detected later in the blood were a measure of how much reached the plasma—the exposure rate.
VX-950 “loves being crystalline and loves hanging on to itself,” Hurter explains. After two years and many millions of dollars, Vertex chemists had developed a process to keep it from locking into a lattice by melting it at high temperature, then infiltrating it with polymers as it flash-cooled. The polymers, like infinitesimal chains of pop-beads, mingled with the drug, diverting and delaying the atoms from reconnecting long enough for the substance to be absorbed. The process plainly worked, but yields were scant and unreliable and results erratic, swinging from batch to batch.
“They had done this tox study in Arkansas in July, and the exposures they got from dogs were really bad; way worse than they expected,” Hurter recalls. “Then, a little later, in one of the multiple-dose panels in humans in Germany, they got almost no exposure, even though they’d been getting good exposure before. They’d thought they had stability for twenty-four hours, because that was the data that people had generated in February in Boston. There had been a heat wave in Europe, so the general hypothesis was that the drug was crystallizing. We didn’t know what was going on, basically.”
Hurter dug into the chemical process, an amorphous system like glass in which the drug substance is kept in a jumbled-up state that prevents it from finding itself and crystallizing. She hadn’t worked with an amorphous system before. All Vertex could spare for her experiments was two grams of the drug. Her group was supposed to have thirteen people, but the previous director had left during the winter, taking a couple of chemists with him, and within two weeks after her arrival, it was down to five. Unwilling and unable to wait until she could conduct a more thorough search, she hired two temps.
“The chaos was intense,” she recalls, “but fun. Nine-fifty is definitely way up there in terms of challenging molecules. Merck didn’t have anything like that in their repertoire at the time. It’s insoluble in everything; it’s not even soluble in solvents. All the normal things you d
o just aren’t enough for it. It’s big, so it’s hard to get through the GI membrane, but to get it through the membrane, you need high concentrations; and yet it’s hard to get it soluble—you’ve got to work on both aspects. It’s also very difficult to make and very expensive to make. They had twenty-two steps with really low yields. All kinds of hazardous process steps. A real bear.”
On a site visit to Aurora, she found herself in a car from the airport sitting next to physical chemist Pat Connelly. One of Vertex’s early scientists, Connelly had come to the company straight from a Yale postdoc, and within five years he had left to start his own structure-based discovery company specializing in antibacterials. After selling that business and starting another, he’d returned to Vertex earlier in the year. Thomson and Murcko had asked him to help develop its CMC (chemistry, manufacturing, and controls) environment. He and Hurter discussed the problems with VX-950 for five hours during the flight back from San Diego.
Connelly appreciated her urgency, especially regarding the minimal human exposure during the Phase Ia trial in Belgium. It wasn’t hard to imagine the sheer unacceptability of Vertex heading into its first trials with VX-950 in infected patients, scheduled to start on November 1, with a drug that couldn’t be guaranteed to show up in the blood. Connelly had spent much of his early career investigating the effects of heat on the bonding of proteins and small molecules, using a supersensitive calorimeter that can measure temperature differences of a millionth of a degree. The machine was still in the lab.
