For a guy who once claimed his greatest missed opportunity was not joining a rock band, George Scangos has a lot of loud fans. Scangos has steered Exelixis Inc. of South San Francisco, Calif., from a tiny genetics company to a respected producer of promising cancer compounds to a firm finally within sight of bringing one of its drugs to market, a significant achievement for any biotech.

"George Scangos is a rock star and not just because of the fancy glasses," says Colin Hill, who has tried unsuccessfully to persuade Scangos to hire Hill's bioinformatics firm, Gene Network Sciences Inc., to do some of the data crunching Exelixis needs to "finish the game," as Hill puts it.

That game, or as Scangos put it when he sat down with The Deal in October, a "whole new ballgame," is the expensive, competitive world of big clinical trials. The stakes get higher, the patient populations larger and the price of failure more damaging once a drug moves into the final phases of testing in humans.

-- See other stories in this Healthcare Special Report --
The big trial
Doctor in the house
Unclear diagnosis

"They need to be very targeted and very efficient," says Morningstar Inc. biotech analyst Karen Andersen.

Exelixis just launched its first Phase 3 trial, with a lot of help from the much larger Bristol-Myers Squibb Co. If it succeeds in introducing one of its own drugs, Exelixis will have reached a sort of biotech nirvana.

It's been a steady progression since Scangos took over in 1997. Exelixis has graduated, stepladder-like, from a tiny company doing fruit-fly research to creating a state-of-the-art laboratory for drug discovery to early clinical work, where its drugs are tested in handfuls of volunteers.

Peers give Exelixis high marks for turning its "platform" -- its underlying technological expertise -- into programs with commercial appeal. "They've been responsive to where they needed to transition their platform" and have made sure the drugs they produce attract Big Pharma customers, says Steven Tregay, a former Novartis AG executive who now runs Forma Therapeutics Inc. in Cambridge, Mass., an early-stage biotech hoping to replicate Exelixis' success.

The fruits of Exelixis' platform have accumulated quickly. It now has seven drugs in its pipeline, with seven more in or near the clinic handed off to partners. With its most advanced candidate, XL184, Exelixis recently started its first Phase 3 trial, in 315 patients with a rare, hard-to-treat form of thyroid cancer. Two-thirds of the funding for the trial comes from Bristol, a longtime collaborator that paid in December a near-record $195 million up front to share the rights to XL184. It's the latest in a string of creative arrangements Exelixis has crafted in recent years.

"George is a brilliant strategist," says Ian Malcolm, who served as vice president of strategic marketing for three years at Exelixis before leaving in early 2008. "His creativity is a wonder to behold."

Because its early-stage compounds, some not yet even approved for testing in human volunteers, have become reliable revenue sources, Exelixis is relatively stable. In contrast, most of its biotech peers hunt for cash or desperately cut staff and programs. At the end of February, 52% of the 356 public biotechs had market capitalizations of less than $100 million, and at year's end 120 were trading with less than six months' cash on hand, according to Burrill & Co. LLC in San Francisco.

In December, Exelixis cut 78 staffers, or 10% of its work force, and Scangos and his lieutenants vow they won't need to tap the equity markets through 2011. "Going to the capital markets is out of the question," he said in October. "You have to survive on cash from deals." He says now that still holds true and promises another licensing deal is in the works.

Licensing, initial public offering, secondary offering, private equity development partnership, convertible debt, divestitures, acquisition: The only type of deal he hasn't done is selling the company.

The deepening recession and a stock price of less than $5 a share could change that. Of course, executives prefer to talk about Exelixis' future as an independent drugmaker. "Investors want to see products come to market with us having certain ownership," says CFO Frank Karbe. "That's where the main value sits and what investors appreciate."

But with a product at least two or three years away, with a history of spending a lot of money, and with big, cash-rich drug firms ever hungry for promising cancer drugs, Exelixis might find an unrefusable offer on its doorstep. One billion dollars would be more than double its current market capitalization and a reasonable outlay for firms such as Bristol-Myers Squibb, Sanofi-Aventis SA and others that have stated their intent to shop while prices are depressed.

Morningstar recently ranked biotech acquisition candidates, and Exelixis tied for 15th most desirable on the list. It scored fairly high on pipeline strength and collaborative fit but was tempered by its high burn rate and the patience an acquirer would need to realize a big revenue boost. Operating expenses topped $300 million in 2008 and are expected in the same range for 2009. Scangos takes umbrage when, as often happens, analysts and reporters highlight the firm's cash burn. "Our burn rate is not high; our expenses are high," he says with some exasperation, claiming that money from partnerships and other sources reduces the annual burn rate into the tens of millions. The company ended 2008 with $284 million in cash.

However you call it, the firm has never been shy about spending money. Before its expansion into clinical development, Exelixis built an ambitious state-of-the-art facility for the foundational work of finding targets among a vast group of molecular enzymes called kinases and designing the drugs to hit them. "It was a very grand scale for everything," recalls Ruihong Chen, a former senior scientist who worked on the protein chemistry of drug targets. One example of Exelixis' largess, she says, was the hiring of X-ray crystallographers, experts who tease out the enormously complex three-­dimensional structures of proteins to better understand how a drug might interact with them. "A lot of companies don't have crystallography-aided drug discovery," says Chen. "Exelixis put a lot of money into that."

There was urgency, too. Other firms had a head start in taking small-molecule kinase inhibitors -- drugs that block the activity of kinases, often implicated in tumor growth -- into the clinic, and Exelixis wanted to catch up. The company's goal was to create drugs that attack several pathways, the cascades of signals that trigger tumor activity. Block the signal along the pathway, and it's possible to interrupt the tumor growth. "Certain pathways are known to be abnormally regulated," says Scangos. "We want to have leading inhibitors of each pathway."

With several of its kinase inhibitors in clinical trials and a goal of three more compounds ready for the clinic every year, Exelixis must be more rigorous. Programs will be killed as they develop safety concerns or don't work well enough.

If past deal activity is any indication, they should also find homes for some of them, bringing in up-front cash and off­loading development costs. Kinase inhibitors are a crowded field, and advances such as ones that target multiple kinases at once are expanding, but they've been around long enough to show eye-opening results. A study in the January issue of Nature showed kinase inhibitors with a much higher clinical success rate than cancer drugs generally. Of nearly 1,000 oncology drugs studied between 1995 and 2007, 18% made it through Phase 3 trials. In the kinase inhibitor subset of 137 drugs, the success rate was 47%.

"Factors underlying this difference are likely to include the targeted nature of kinase inhibitors and the improved design of clinical trials; for example, biomarker-driven patient stratification," the authors wrote.

Biomarkers have become a buzzword, but the idea is fairly simple. Every person has different underlying biology that creates subtle variations in similar diseases and their responses to drugs. As we learn more about the biology of patients and their tumors, researchers can include or exclude patients in test groups to narrow the focus of their drugs.

Perhaps the most famous is Genentech Inc.'s Herceptin, a treatment for breast tumors that because of a genetic mutation produce too much of a protein called HER2. It occurs in 20% to 30% of all breast cancer patients, according to the Mayo Clinic. Once the link was established, Genentech scientists began designing a drug to shut off HER2's chemical signals that help tumor growth. It was approved for use in 1998.

Exelixis can exploit this approach, either by tapping into the rapidly expanding pool of public knowledge about biomarkers or through its own internal work. The firm collects tumor samples from some Phase 1 trials for its translational medicine team to analyze. Such analysis can help design "more thoughtful Phase 2 studies," says Scangos. "It can shorten [trial times] for some tumor types, not for others. You can't get serial samples of lung or brain tumors."

Colin Hill of Gene Network Sciences says a company like his can help with the sophisticated data analysis, which most companies won't have the resources to do in-house. "You can use Phase 1 data for cancer because the patients are pretty sick, uncover what the key genetic markers of response are and include them for design of the next phase," Hill says.

For example, if a researcher finds a marker for a small percentage of patients who aren't tolerating high doses of a drug, she can exclude them and crank up the dose.

"Relative to a lot of cancer-focused firms, they're in good position to keep late-stage clinical trial costs down," says Morningstar's Andersen. "They're benefiting from the industry's overall understanding that cancers are specific mutations, and also from the fact that their own drug candidates are very specific."

One example is XL228, now in Phase 1 to treat chronic myelogenous leukemia that has built up resistance to Novartis' Gleevec, one of the first kinase inhibitors on the market. Several Gleevec-resistant mutations exist, and Novartis and Bristol-Myers have drugs to treat some of them. Exelixis is going after a mutation that hasn't been addressed.

Smaller patient groups in a trial can save a lot of money. One rule of thumb is each patient costs $60,000 to $90,000, says Russ Belden, a former Genentech marketing executive turned independent clinical trial consultant. Cutting a trial size in half from, say, 1,000 to 500 patients, is trivial cost savings for a big firm but "a huge improvement for a small company doing its first drug to approval," says Belden. Exelixis would not discuss its clinical patient costs but says Belden's estimates didn't seem out of line.

Starting small can also help a firm gain a toehold with a drug, prove to regulators and doctors it works, then expand it into larger, more lucrative indications.

There are drawbacks, however, to whittling trials into smaller subgroups. First, the market -- if the drug ever gets there -- might be so tiny as not to be practical. Second, smaller sizes can be double-edged swords for the clinics. If a doctor knows a subgroup of his patients will likely respond well to a treatment because of genetic screening, it's strong motivation to sign on. But if the subgroup is too small, a doctor might find it too exclusionary.

"Is it going to be applicable to a large number of our patients or a tiny fraction?" asks Howard "Jack" West, a lung cancer doctor at Swedish Cancer Institute in Seattle who is testing XL184 for Exelixis in an early-stage trial. "What if you had a trial that's potentially useful to your new lung cancer patients, and it took the same amount of effort and money as one that only appealed to 10% of your population? We're all being selective with our budgets, and it does cost a fair amount to maintain the infrastructure for clinical trials. You have to prioritize."

That kind of caution means more competition. Exelixis has many powerful partners -- Bristol-Myers, Genentech, Wyeth (soon to be part of Pfizer Inc.) and GlaxoSmithKline plc -- but it's also competing with them and many others to persuade a limited pool of clinicians to test their drugs. "I'd get on my soapbox and talk about clinical trial accrual," says former Exelixis vice president Malcolm. "The company is beginning to recognize that unless you bang on your chest a little bit, it's tough."

Scangos acknowledges that oncology is a crowded field with a limited number of clinical sites. "But we've not had a problem getting investigators to take our compounds."

He's also had no problem getting companies and, in one case, private equity investors, to buy Exelixis' compounds. Exelixis was one of the first biotechs to partner with Symphony Capital LLC of New York, which invests in drug development. It creates separate companies, seven so far, to which the participating biotechs transfer the rights to a basket of drugs. In Exelixis' deal, signed in 2005, it got a commitment of $80 million and gave up rights to three early-stage compounds. Symphony and Exelixis kept developing them, and Exelixis had the exclusive option of buying them back at an escalating premium. The trials haven't gone well, however, and Exelixis has said it won't exercise its option, which expires in June. That means Symphony paid $80 million and keeps the rights to the three drugs. Symphony is looking for partners and funding for the compounds, "which we believe have significant value," Symphony chairman Mark Kessel wrote in an e-mail.

Exelixis also signed last June an unusual debt deal as a safety net. Deerfield Management LP, already one of Exelixis' top-10 shareholders, agreed to open a $150 million line of credit that Exelixis can draw down through the end of the year in exchange for warrants tied to the stock price at the time of the draw. So far Exelixis hasn't needed the cash, and it might not need it anytime soon if the firm can notch a couple more large licensing deals.

Exelixis' licensing deals have some unusual aspects, too. It persuaded Genentech to pay $40 million for rights to a drug that hadn't yet been tested in humans (though Exelixis is paying for part of the current Phase 1 trial). It has signed deals with Bristol-Myers, Wyeth, Daiichi Sankyo Co. Ltd. and GlaxoSmithKline, the latter a broad-ranging pact in 2002 before Exelixis had much clinical experience. That deal expired in October, with GSK taking control of just one compound, though it had an option on many more. "We got money five different ways from GSK" -- up fronts, milestones, $90 million for R&D, a small equity purchase and an $85 million loan that starts to come due with interest in October -- "and that allowed us to build the company," CFO Karbe says.

When the pact ended, investors saw Glaxo's refusal to choose a second compound -- out of a pool of five -- as a bad omen for Exelixis' pipeline. Shares fell 22%. But Exelixis spun it quickly into gold. XL148, one of the drugs Glaxo passed over, became the cornerstone of the Bristol-Myers deal in December. The details: Bristol paid $195 million immediately and promises $45 million more by the end of 2009. The drug giant pays 65% of development costs, and the firms split U.S. profits 50-50. Bristol then pays royalties on sales outside the U.S. and also got rights to XL281, paying all development costs, up to $465 million in milestones and double-digit royalties. The $240 million committed cash is one of the top five largest biopharma license payments ever.

Before the Bristol deal, Scangos said a big license deal was coming and he wanted to keep at least 50% ownership of XL184. He delivered even as the economy spiraled down through the fall. Now he's saying that talks on another are well under way.

The more license deals, the more entanglements Exelixis creates. The GSK deal walled off compounds for which Exelixis might have found other partners. But entanglements don't prevent acquisition. After all, ImClone Systems Inc. was tied to the hip of Bristol, which owned 20% of ImClone as well as partial rights to its only marketed product, Erbitux. That didn't stop Eli Lilly and Co. from outbidding Bristol when ImClone went on the block.

But if any biotech is poised to keep the wolf from the door during these dark days, it's Exelixis. The signs are everywhere that more focused cancer therapies, designed and tested against an expanding field of genetic data, are the future. Massachusetts General Hospital in Boston said last week it would soon start taking genetic profiles of the tumors of every new cancer patient, the better to target their treatments. The federal government wants industry to steer in that direction, too. "The [Food and Drug Administration] is expecting to evaluate drugs based on biomarkers," says Sonia Pearson-White, a geneticist who oversees several projects at the Biomarkers Consortium, a research group funded jointly by the federal government and the drug industry. "Moving toward targeted therapies is practically a necessity these days."

If Exelixis can start to transform that scientific urgency into lower costs now, it has a good shot at emerging from the recession not merely intact but prepared to stick around for a long time.