(FORTUNE Magazine) – Do you enjoy delayed gratification? Consider a career in pharmaceutical research. The wait between a drug's discovery and its approval is usually well over a decade. Often it takes forever: Most experimental drugs fail, and world-class researchers can spend their entire careers working on flop after flop. But once in a while this landscape of frustration is lit by an intense glow. Such was the case one day in 1995 when Jeanette Wood discovered that a compound stored 40 years in a bottle on her company's shelves had the makings of a cancer breakthrough.

A pharmacologist at Ciba-Geigy, now part of Novartis, Wood had set out several months earlier to find compounds that could block angiogenesis, the growth of new blood vessels. Tumors induce the formation of vessels to feed their runaway growth; she was seeking a drug to deny that blood supply, arresting cancer without the harsh side effects of chemotherapy. Week after week, she and her research team did test-tube studies screening "candidate" molecules--over 100,000 in all. Finally they hit pay dirt. The 40-year-old compound from Ciba's chemical archives blocked the activity of a key enzyme that helps trigger angiogenesis. And it did so at concentrations measured in parts per billion, a sign of extraordinarily high potency. But Wood kept her excitement on a tight leash. It doesn't pay to put much store in an untested drug.

In fact, just one in 5,000 to 10,000 early-stage drug candidates make it all the way to market, according to Datamonitor, a London research firm. From that dismal statistic flow most of the industry's woes, including the $897 million it costs to ready a medicine for market (according to the Tufts Center for the Study of Drug Development) and the resulting high drug prices that infuriate almost everyone. Still, there are signs that the industry's productivity is on the rise, helped in part by the influx of novel medicines from biotech.

One sign of the uptick in pharma's productivity is the burst of creativity surrounding enzymes called kinases. Acting like relay switches within cells, kinases are involved in turning on basic operations such as cell division. When they go wrong, horrible things can happen, most notably cancer. A new generation of designer drugs tailored to block errant kinases is now coming down industry pipelines--including PTK, the experimental cancer drug discovered in Wood's lab. Novartis and Germany's Schering are co-developing PTK (for the record, its full designation is PTK787/ZK 222584) to treat cancers of the colon, brain, kidney, and other organs. To understand the traps and hurdles that these next-generation medicines must survive on their way to market--and to gauge how close we really are to a designer-drug revolution--FORTUNE delved into the story of PTK. It's the tale of an arduous, often frustrating, decade-long journey that could lead to the first anti-angiogenesis pill--or no product at all.

The huge potential of kinase-based medicines was first brought home by Novartis's Gleevec. Launched in 2001 to treat a cancer of blood cells called chronic myeloid leukemia, or CML, it's the closest thing yet to a magic bullet, a drug that wipes out diseased cells without hurting normal ones. Every patient in an early trial of Gleevec showed a major improvement; some advanced cases even went into remission. Genentech's vaunted cancer drug Avastin, which debuted last year and is the first purely anti-angiogenesis medicine on the market (others are known to have anti-angiogenic effects), also works by blocking a molecule that activates a kinase. It is administered in clinics via intravenous infusion, so an anti-angiogenesis pill like PTK would mean a major advance in ease of use. Another example is Tarceva, a new lung-cancer drug developed by U.S. biotechs OSI Pharmaceuticals and Genentech along with Switzerland's Roche Group. Scores of additional kinase-based drugs are on the way to treat cancers, diabetes, and other diseases.

Like Gleevec, PTK sprang from a line of research once seen in pharmaceutical circles as almost quixotic. Scientists have known for more than two decades that abnormal kinases are deeply implicated in cancer. A deviant kinase known as Bcr-Abl, for instance, was found in the 1980s to cause the runaway dividing of white blood cells that underlies CML, the leukemia Gleevec is used to treat. But selectively inhibiting kinases like Bcr-Abl seemed next to impossible. The body's cells contain hundreds of different kinases with similar structures. A drug molecule whose shape allows it to nestle tightly with one of the enzymes (and thus jam its operation) tends to nestle with many others as well. That, in turn, poses a high risk of severe side effects.

Around 1990, however, researchers discovered subtle differences in kinases' structures, allowing at least somewhat selective hits on them. That set off an intense race to develop kinase-inhibiting medicines. A team at Ciba-Geigy zeroed in on Bcr-Abl and in 1993 discovered a compound that, in test-tube studies, shut it down without messing up many other kinases. But the company's marketers were skeptical--they figured the potential drug would never repay its development costs, for the leukemia it would treat is one of the rarer cancers.

If business as usual had prevailed, that calculus probably would have scuttled the drug. But in 1996 Ciba merged with Sandoz to become Novartis, putting decisions about the project in the hands of the new company's CEO, Daniel Vasella. A medical doctor who had worked his way up Sandoz's corporate ladder before the merger, Vasella, now 51, felt that Gleevec's potential as a major cancer advance trumped the downside of its small market. Personal history helped motivate his support: As a boy, he had watched his older sister fight an agonizing battle with cancer; she died at 18.

Vasella's hunch was reinforced by Jörg Reinhardt, the new company's head of pharmaceutical development. A trim, precise German native who plays the cerebral Mr. Spock to Vasella's more emotive Captain Kirk, Reinhardt is a gatekeeper whose views can make or break projects. He thought pursuing Gleevec made sense from a risk-management perspective: With its deepened pipeline and pockets, he says, the post-merger company was poised to take more risks than either of its smaller predecessors could.

It turned out to be one of the best bets in pharmaceutical history. The first patients on the drug did so well that even Reinhardt got excited--when that happens, it's "a definite sign that something is up," Vasella wryly noted in a 2003 book he co-wrote about Gleevec called Magic Cancer Bullet. The FDA approved the drug less than three months after Novartis applied to put it on the market. Then--Secretary of Health and Human Services Tommy Thompson hailed it as initiating the "wave of the future" in medicine. Last year Gleevec's sales topped $1.6 billion; its success reinforced the Vasella philosophy that had kept the drug alive. Says Mark Fishman, head of drug discovery: "There are many diseases pharma doesn't touch due to projections of inadequate market size. But the projections are often wrong. There are so few really good new drugs that if you make one, people will use it."

Even before the Gleevec project took off, Jeanette Wood's drug-hunting team joined with counterparts at Schering in 1995 to target a "pathway" of interacting chemicals known to play a key role in angiogenesis. Its foremost component is VEGF (pronounced vej-eff), a molecule secreted by cancer cells to trigger angiogenesis. Tumors can't get much larger than BBs without calling VEGF into play to form new blood vessels. VEGF's message is conveyed through "receptors," molecules that jut through the outer surfaces of blood-vessel-forming cells. The receptors' tails, which reside inside the cells, include kinase regions. VEGF molecules secreted into the blood dock on the receptors' outer parts. That activates the inner kinase sections, which in turn switch on other chemical signals in the cells to start the angiogenesis process.

Wood, 53, led Ciba's quest for molecules that could gum up the VEGF kinase. An unassuming New Zealand native with a keen mind, motherly mien (she has two sons, ages 19 and 21), and invaluable pluck, she had joined the company in 1980 after graduate school. "I was a little post-doc coming to the land of the Big Pharma companies," she says. (Ciba, Sandoz, and Roche all had headquarters in Basel, Switzerland's industrial center on the Rhine.) Her gender made the move particularly daunting. "I came from the first country in the world to give women the vote to one that was almost the last to do so," she says. Females were generally seen as homemakers in Switzerland, she recalls, and "only underprivileged women were supposed to work" outside the home. "I always took delight when nonworking mums commented on how nice and well adjusted my boys are."

Wood's grit and rigor impressed colleagues, and she proved her mettle by leading the discovery team for aliskiren, a potential blockbuster for treating high blood pressure that's now in late clinical tests. As usual, good came with the bad: In the mid-1990s, Ciba shut down the cardiovascular program she had specialized in, and about the same time, her brother and one of her lab's technicians both died from cancer. Soon after this "first shocking personal exposure to cancer," she says, she threw herself into the anticancer research that led to PTK.

The first step was fashioning the molecular equivalent of Cinderella's slipper. Working with Schering and the Tumor Biology Center in Freiburg, Germany, Wood's team developed a way to isolate the VEGF kinase in the test tube and detect whether subsequently added compounds had the right molecular shape to mesh with it in a way that blocked its signal. With this "screening assay" in hand, they began combing Novartis's chemical library for princess compounds.

Like most Big Pharma players, Novartis has a vast collection of compounds assembled over the years as potential drugs. They're shelved in various labs--hundreds of thousands of painstakingly cataloged bottles and vials waiting to be mined for medical miracles. The laborious screening by Wood's team went on for months before they tried the compound that became PTK--about four pounds of the stuff was stored in a big glass bottle, recalls retired chemist Peter Traxler, who worked on the project. It was first synthesized in the 1950s by a University of Bern professor with a genius for building molecules. He contributed many novel samples to the company labs but died in the 1980s, never realizing that one of his creations had the makings of a powerful weapon against cancer.

Wood recalls that her technicians showed some low-key excitement on the day of the momentous discovery in 1995. But there was no celebration in the lab. After taking note of the screening "hit," Wood matter-of-factly brewed her usual mug of strong black tea with milk and plunged right into the next phase: experiments to see whether PTK promiscuously cozied up to a broad array of kinases, a fault that would have returned it to the shelf. Her team found that it does hit on multiple kinases, but their luck held--the extra targets are also involved in angiogenesis and the spread of tumor cells, hence likely to enhance the drug's desired effect.

The team also confirmed that cells can absorb enough PTK to have an effect. Then they demonstrated that when given orally to mice, PTK is rapidly taken up in the blood and distributed around the body, indicating that it could be made into a pill.

In other studies, the researchers implanted tiny, porous chambers containing VEGF in the flanks of mice to induce new blood vessels to grow, then gave them doses of PTK to prove that it blocked the growth around the implants. Next came experiments with "nude mice," weird-looking, hairless rodents whose immune systems are so weak that human cells can be planted inside them without rejection. PTK slowed the growth of various human tumors placed under their skin. Even more exciting, the drug blocked the metastatic spread of tumors in rodents. Other studies showed that despite holding back tumors' deranged angiogenesis, PTK didn't impair the healing of sutured incisions in rats.

After months of lab work, the compound seemed too good to be true. Even the company's exacting chemists liked it. Their job is like juggling snakes--they must ensure that a drug can be readily formulated, remains stable when stored, resists destruction by stomach acids, gets into patients' bloodstreams, and doesn't break down into toxic compounds when metabolized. It's almost unheard-of for a randomly selected compound to meet all these requirements without major tinkering that takes months to years. But PTK had only one significant defect: Its crystals tended to absorb water and degrade when stored in humid conditions.

Solving that problem fell to Guido Bold, 50, a veteran chemist whose self-effacing manner belied his excitement about PTK. He and his colleagues rummaged through their bag of tricks and came up with a way to formulate the drug so that it was both water-resistant on the shelf and readily made into pills.

But then PTK's amazing run of luck came to an end. The improved version of the drug had shown no significant toxicity in mice and dogs after a month of dosing. But a bizarre side effect turned up in long-term studies of rats. After the rodents took the drug for three months, their duodenums--the part of the digestive tract that connects the stomach to the intestines--enlarged. The entire project screeched to a halt.

PTK's chance to reach the market suddenly seemed slim. Anti-angiogenesis represented an unproven way to fight cancer at the time, so Novartis's upper management saw the project as a high-risk venture--the uncertainty of a payoff gave little reason to keep going if safety was in doubt.

Wood was determined to keep PTK alive. "I considered it my baby," she says. She and other proponents pushed for further tests to analyze the rat effect, arguing that the chance to get the first anti-angiogenesis pill on the market warranted the effort. Besides, the side effect had never been seen before, and seemed to occur only in rats; maybe it had more to do with their metabolic idiosyncrasies than the drug. As PTK's fate hung in the balance, Wood worked off stress by gardening. "I have a very beautiful garden that was created during that time," she says.

The fact that Schering was collaborating on PTK helped her cause--the German company's scientists had also been testing the compound and felt it was too promising to kill. After days of uncertainty, Novartis gave the PTK team a chance to restore their sick baby to health. Wood and colleagues spent the next year painstakingly establishing that the rat effect is not a form of cancer; that it does not occur in mice, dogs, or monkeys; and that the rats' swollen duodenums return to normal when the animals are taken off the drug. Precisely how PTK causes the problem still isn't clear. But the rigorous safety studies put the drug back on track.

Another near-death experience loomed for PTK after its clinical trials started in 2001: A similar anti-angiogenesis drug developed by biotech Sugen (now part of Pfizer) failed to show efficacy in a clinical trial. The announcement, in early 2002, "made our management terribly skeptical" about PTK, says Wood. "They talked about giving it a lower priority"--which would probably have been another death sentence. This time her chemist colleagues came to the rescue: They synthesized the failed drug, enabling studies at Novartis that showed it wasn't as effective or as well-tolerated in animals as PTK.

Early clinical results with PTK gave it a much-needed boost. In one trial, half of 28 patients with advanced colon cancer who got PTK in combination with standard chemotherapy experienced at least some tumor shrinkage, and tumors disappeared completely in one case. The drug's side effects were mostly minor--there was no sign of the duodenal swelling that had affected the rats, but the highest doses tested were associated with dizziness, fatigue, and other adverse effects.

PTK now was ready for the acid test of drug development--the protracted, costly, nail-biting Phases 2 and 3 of clinical trials. Normally Phase 2 tests give preliminary readings on whether a drug really works, while Phase 3 trials demonstrate efficacy convincingly enough to win approval. But PTK's encouraging results suggested a shortcut: Skip the Phase 2 tests and immediately push into large, pivotal Phase 3 trials. The early data gave some support for that. So did the fact that Genentech had grabbed the lead in the race to bring out the first anti-angiogenesis drug--its blood-vessel inhibitor, Avastin, was launched last year to treat colon cancer--and Novartis wanted to catch up.

But leapfrogging would add considerable risk, for Phase 2 would dispel lingering doubt that PTK really worked and establish the optimal dose. Skipping Phase 2 increased the chance that hundreds of patients would wind up getting zero benefit in worthless studies that could cost $100 million or more. On the other hand, jumping to Phase 3 could significantly shorten PTK's time to market, potentially saving many lives and racking up hundreds of millions of dollars of sales that otherwise wouldn't be.

Pushing for the leap, PTK's champions argued that a novel technique had made lengthy Phase 2 trials unnecessary. In an early trial, researcher Bruno Morgan and colleagues at Britain's Leicester Royal Infirmary had used MRI scans of patients' tumors to show that PTK, as expected, interfered with the blood flow in the tumors.

"This is where [the PTK project] got very different from research on traditional cancer drugs," says Gregory Burke, Novartis's head of oncology development. When conventional chemotherapy drugs succeed, they cause tumors to shrink, affording a fast way to assess efficacy and get the dose right--the researchers in early clinical trials just ramp up the dosage in patients until they see signs that tumors are melting away. But anti-angiogenesis medicines work not by shrinking tumors but by stopping their growth. That has forced researchers to cast about for new "biomarkers" to measure whether the drugs work. The Leicester team's blood-flow method provided one and revealed the dose needed to get PTK's desired effects. But was it good enough to take a huge gamble on?

That question was addressed in late 2002 at an intense meeting of Novartis's steering committee on new drugs, a group that includes Reinhardt and Vasella. "Some people said, 'It's too high a risk to go ahead,' " recalls Reinhardt. "They wanted to see more patients responding in Phase 2 first. But others said, 'If you do that, you lose two years, and you still don't know much more than you do now.' "

With perhaps $100 million worth of trials on the table, recalls Reinhardt, "we debated back and forth for two hours. Finally we decided to go ahead. At the end of the day, the most interesting thing was that the [MRI] data were so plausible. PTK was doing what it was supposed to do." In April 2003, Novartis and Schering launched two major Phase 3 studies of PTK in colon-cancer patients.

Wood had relinquished oversight of PTK as it advanced in the clinic, but she and her team ardently followed its progress and continue to study its mode of action. One day this spring she walked into work to find her team in a state of high anxiety--they had just heard the news that PTK had failed an interim analysis of data from one of the ongoing Phase 3 trials, and that the failure had caused Novartis's stock to drop 1.9%. "We were shattered," says Wood. "We all thought, Oh, gosh, our research in this area will be stopped. Management will never believe us again."

The gloom lifted, however, when they got the full story later in the day. The interim analysis had failed to show clear-cut efficacy. But the news wasn't all bad. For the analysis, two groups of physicians had assessed scanner images of patients' tumors to decide whether their cancers were improved, stable, or progressing. Some of the patients were being treated with PTK plus standard chemotherapy, while others got chemo without PTK. One set of doctors judged that PTK extended the time it took for the cancers to worsen. The other group said the drug failed to show a statistically significant extension of progression-free survival. In other words, "the jury is still very much out on PTK," says William Li, president of the Angiogenesis Foundation, a nonprofit in Cambridge, Mass., that supports research on the topic.

Disappointed? Imagine how the Novartis scientists felt. To resolve the ambiguity, they must now wait until mid-2006 to see whether the patients who got PTK actually live longer than those who didn't get it. It's still quite possible that PTK will prove to extend lives, says Li. "The criteria used to assess progression-free survival are very stringent," he explains. "For instance, in a patient with five tumors, four might shrink [after doses of PTK] while the other one grows. That would count as a case of progression," even though the overall tumor reduction might well enable the patient to live longer. It is also likely that some kinds of patients respond better to the drug than others--Novartis recently identified a blood enzyme that may show whom the drug will benefit most. So PTK may well turn out to be a winner. Of course, it may not--but that's the hundred-million-dollar-plus risk a pharmaceutical house takes when it commits to an experimental drug.

Regardless of PTK's fate, Novartis's gutsy leap to the Phase 3 study based on the MRI scans of tumor blood flow represents a major advance in cancer research, says King Li (no relation to William Li), an expert on medical imaging at the National Cancer Institute in Bethesda, Md. The scanning method has opened the door for rapid vetting of novel combinations of drugs to starve tumors of blood. Devising such cancer "cocktails" in the past has taken years of trial and error involving hundreds of desperate patients. The new method promises to identify powerful drug combinations over mere months of testing--Novartis's move has helped push the war on cancer into the age of the magic fusillade.

Wood remains cautiously hopeful about PTK. "A quick win in Phase 3 would have been nice," she says. "But there is something happening in patients." The animal data on PTK, she adds, suggest that it may not show its full cancer-fighting power until it is tried in patients whose cancers aren't as advanced as those in the trials. In particular, PTK may potently inhibit the metastatic spread of tumors--which, in most cases, is what ultimately kills.

Meanwhile, the designer-drug trend continues to bloom. On the same day that the PTK setback was announced, Germany's Bayer, with partner Onyx Pharmaceuticals of Richmond, Calif., announced that a kinase inhibitor they're developing significantly extended kidney-cancer patients' progression-free survival in a Phase 3 trial. And Novartis reported in December that Gleevec continues to rack up heartening statistics. In a study of 1,106 patients, 98% of early-stage CML patients treated with the drug had achieved normal white-blood-cell counts, a sign their cancers weren't progressing. In follow-up tests 3½ years later, 89% continued to show normal blood counts. The fact that some leukemias came back shows that resistance to Gleevec remains a killer issue. Still, no doctor or cancer victim could look at the drug's record of success without feeling that it has a certain magic they'd love to see with PTK.

In Basel, Jeanette Wood isn't waiting for miracles--she's hard at work studying how PTK works, looking for better ways to apply it to the cancer fight. "Only now are we starting to really understand what it's doing" in the body, she says, adding with a flash of irony, "The fact that PTK wasn't 100% successful in the first Phase 3 trial maybe justifies my lab's existence" as the struggle continues to make it a winner. This summer the garden she tends when she's under stress could be very productive indeed. ■

FEEDBACK dstipp@fortunemail.com