The Case of the Notorious Depression Drug

San Francisco magazine, May 2007

A psychiatrist turned drug developer in Silicon Valley has invested 10 years to answer one question: could the compound in the infamous abortion pill RU-486 incidentally ease one of the most harrowing forms of depression?

In the late 1980s, doctors at a hospital in the Dutch port city of Rotterdam admitted a 43-year-old man who, in the space of a few weeks, seemed to have completely lost his mind. Once stable and articulate, he now barely spoke. And the few words that did escape exposed dark, terrifying visions. He feared that people were trying to stick needles into him as part of some voodoo ritual. He was suicidal.

The problem, it turned out, was in his body, not his head. He had cancer of the adrenal gland, and as a result his body was pumping out massive amounts of the stress hormone cortisol. In high levels, cortisol can cause acute psychosis. The man’s cancer was inoperable, but the doctors had a hunch about how to relieve his mental anguish. They gave him a drug they hoped would tamp down the hormonal deluge. Amazingly, within 24 hours all of his psychotic symptoms were gone.

It was the kind of dramatic turnaround doctors dream of. It’s no wonder, then, that a report a few years later of the Dutch patient’s recovery caught the eye of a young Stanford University doctor by the name of Joseph Belanoff, a newly minted MD who was just starting his psychiatric training. He may not have known much, but he knew enough to recognize that something remarkable was going on.

Belanoff was interested in a disorder related to psychosis known as psychotic depression—a debilitating mental illness thought to affect 2 to 3 million Americans. Besides feeling sad and hopeless, people with psychotic depression also hear voices or suffer delusions and are five times as likely to commit suicide as those suffering from regular depression. It’s probably what plagued Andrea Yates, the Texas woman who in 2001 drowned her five children because she was convinced she’d been such a poor mother that they were “doomed to perish in the fires of hell.” One psychotic depressive Belanoff saw was a woman who had tried to burn down her house with her young son in it. “As the parent of a young child,” he recalls, “I realized just how horrible this illness is.” 

And it’s one with terrible treatment options. The main one is electroconvulsive therapy, or ECT (commonly known as shock therapy), in which a patient is sedated and zapped with a low electrical current until he or she suffers a seizure. ECT works (no one quite knows why), but between its association with One Flew Over the Cuckoo’s Nest and the fact that it often erases swaths of memory, only a fraction of patients who stand to benefit from it get it—about 150,000 a year. Instead, most are treated with a combination of antidepressant and antipsychotic drugs, which, though effective, have nasty side effects, including movement disorders, significant weight gain, and diabetes. They also take weeks to kick in.

The Dutch report, however, suggested a treatment that solved both these problems: it worked virtually overnight and seemed to have no side effects. Belanoff wondered, could it perform the same miracle in his patients? He was convinced it might, because he was familiar with the theory that psychotic depression is caused by abnormally high levels of cortisol, the same problem afflicting the Dutchman.

There was just one hitch: the drug was RU-486, aka the abortion pill—one of the most controversial medications ever developed. Branded as a “death drug” and “human pesticide” by abortion opponents, the pill had sparked major protests in Europe and the United States. Opposition in this country was so fierce that for nearly two decades the pill was banned here.

The Food and Drug Administration ultimately approved RU-486 in 2000, and since then more than 750,000 women have used it to end a pregnancy, though under stringent regulations: only doctors can order it, the pregnancy can be no more than nine weeks along, and patients must return for two follow-up visits. But controversy over the drug has barely ebbed—every year congressional opponents file bills to withdraw federal approval—especially after several young women died after taking it. “People didn’t want to touch it with a 10-foot pole,” says Belanoff.

But that didn’t deter him. For the past 10 years, Belanoff has doggedly worked to develop this most notorious of drugs into a treatment for some of the most desperately ill patients he’s ever seen. The story of his unlikely quest offers a rare window onto the trials and tribulations of modern drug development, a difficult and risky business even under the best circumstances. This unlikely entrepreneur—a Wall Street trader turned psychiatrist turned drug company executive—has been surprisingly effective in his mission. He’s gained the support of some of Silicon Valley’s savviest investors and drug development experts and raised close to $100 million through public and private stock offerings for Corcept Therapeutics, the company he cofounded. He’s also contended with investors’ caution, colleagues’ skepticism, and problematic clinical trials. Critics say he should have thrown in the towel long ago. Yet he’s seen enough encouraging evidence to sustain his hopes that the drug is indeed the answer to one of psychiatry’s most intractable illnesses. If he’s right, it could be an incredible breakthrough in psychiatric treatment.

But then again, Silicon Valley is littered with the ghosts of once-promising breakthroughs.

Belanoff is a slight man with straight brown hair that rebels against his efforts to push it back into the standard businessman’s helmet. He looks a lot like the comedian Martin Short. At 49, he could easily pass for 35. Though his long commitment to developing RU-486 bespeaks a certain passion, Belanoff is resolutely undramatic in the way he talks about it. He’s a low-key person of steady mood and habits, the sort who rises nearly every day at the same time—6:15 a.m.—and is rarely rattled. In explaining a complicated point, he often speaks in outline form and can go on for long stretches without losing his thread. But he’s also a warm and earnest conversationalist who seeks out common points of interest and listens attentively. It’s not hard to imagine people confiding their troubles to him.

He didn’t set out to be a head doctor, though. “A lot of people go through life assuming there’s just one thing for them to do,” Belanoff says. “That’s never been true for me.” Instead, he’s shuttled smoothly between worlds that for most people are mutually exclusive. He studied English at Amherst College, but after graduating in 1979, he took up work in the number-centric world of Wall Street. There he specialized in technical arbitrage, a field devoted to leveraging profits by buying a product in one market and selling it in another. “The goal is to create a risk-free transaction, where you make a small amount of money because of market inefficiency,” he explains.

That was appealing to someone who is drawn to risky endeavors and yet is too rational and levelheaded to enjoy the uncertainty of truly high-stakes gambles. What Belanoff likes is working out ways to ensure the risk will yield a worthwhile reward. It’s a mind-set that goes back to childhood, when he would collect thousands of pennies to find the few that were valuable. As a college student, he once sought to recoup a summer’s worth of expenses by playing the blackjack tables at Reno, but only after teaching himself to count cards.

He excelled as a trader—“He’s one of the smartest men I have ever met in any business,” says Sandy Herzfeld, his former boss at Moseley Securities—and the money was nice. He could buy an apartment, put his college girlfriend (whom he later married) through medical school, and amass a small fortune in savings. But after several years, Belanoff decided Wall Street wasn’t the place to find meaningful work that could keep him engaged for a lifetime. Medicine seemed to offer just that, so while he continued working as a trader by day, he began taking premed courses at night. “I’d show up for the organic chemistry lab in a three-piece suit,” he recalls.

At age 30, Belanoff had one career behind him and another unfolding before him that he knew he was going to love. He particularly enjoyed interacting with patients and ended up winning a prize for his compassionate bedside manner. No one was surprised when he decided to become a psychiatrist, and from the beginning he was drawn to the cutting edges of the field.

That’s one reason he decided, in 1992, to move with his wife and two children to Palo Alto to train under Alan Schatzberg, the newly appointed chairman of Stanford’s psychiatry department, who had himself just arrived from Harvard. Schatzberg is a leader in biological psychiatry, a field that explores the little-understood intersection between mind and brain. One of Schatzberg’s research interests is psychotic depression; indeed, he was one of the first to posit a connection between the disease and high cortisol levels.

Years earlier, Schatzberg and a Harvard colleague, Anthony Rothschild, had noticed that many of their psychotically depressed patients had abnormally high amounts of the hormone circulating in their bloodstream. Some researchers dismiss this as a symptom of the disease, but based on additional tests of the body’s system for governing cortisol production—a complex feedback loop involving the hypothalamus and the pituitary and adrenal glands—Schatzberg concluded it might actually be causing the psychotic symptoms. In these patients, the system seems to be thrown into overdrive, generating so much cortisol that it triggers changes in the brain that can lead to psychosis.

Though the theory is elegant, and biologically plausible, some experts didn’t, and still don’t, buy it—noting, for instance, that not all psychotic depressives have high cortisol. Yet the key, as Rothschild maintained, is that the hypothesis was testable. “If you could do something to lower cortisol levels or block their effect, you could see if this was true,” he says.

Enter RU-486.

The drug was invented in 1980 by researchers at the French pharmaceutical giant Roussel Uclaf who were looking for a new approach to birth control. They wanted to find a compound that would induce abortions very early in pregnancy. To cover up what they knew would be controversial research, the story goes, they said they were focusing on compounds that also had another property: the ability to block cortisol. (Many assumed they were working on a treatment for Cushing’s disease, a rare disorder involving excess cortisol.) The best of the bunch was mifepristone, or as the scientists dubbed it, RU-486.

Turns out, Schatzberg could not have custom-designed a drug better suited to test his theory. Unlike other cortisol-blocking drugs, which blanket the whole hormone-production system, mifepristone hones in on one crucial cog in the brain, a receptor that’s activated only when levels of the hormone are acutely high. Belanoff and Schatzberg thought the drug might deliver two benefits for the price of one: breaking the chain of chemical changes in the brain that lead to psychosis, while “resetting” the overheated cortisol-production system so the body will stop churning out so much of it. Ideally, they thought, one short-term use of the drug (for a period of seven days) would effect long-lasting improvement.

If they were right, they would be on to something extraordinary. Most psychiatric drugs have been discovered “by serendipity,” says Dennis Charney, former head of the federal government’s mood and anxiety disorders research program. Antidepressants, for example, are in part the legacy of research in the 1950s showing that patients became less depressed on tuberculosis drugs, which happened to boost levels of certain brain chemicals. In fact, experts still don’t understand the precise workings of many of the most successful psychiatric drugs. “Prozac has sold $20 billion,” says Belanoff, “and we still have no idea how it truly works.”

In this case, there was a well-developed theory of how a disease comes into being and a drug whose mode of action correlated directly with that theory. If psychotically depressed patients did respond to the drug, it would be the first time a psychiatric treatment was developed in response to a clear understanding of a disease’s mechanism. It could, as Charney and his colleagues later wrote, usher in a “paradigm shift in the treatment of depression.”

Belanoff wanted to put Schatzberg’s and his theories to the test, but he had difficulty getting enough of the drug to even get started. Abortion protesters had threatened to boycott all of Roussel’s products, so the company had decided against seeking the FDA’s approval. Instead, it turned over the U.S. license to the Population Council, a small nonprofit in New York. Belanoff convinced the group to give him enough of the medicine to test on five patients. Each would get either 600 milligrams of the drug or a placebo for four days, and then the pills would be switched for the next four days.

The first patient was admitted to Stanford University Medical Center shortly before Christmas in 1996. He was a 50-year-old high-tech executive who had been completely healthy until he’d had to put his ailing mother into a nursing home a few months before. At that point he became depressed and began to have trouble eating, concentrating at work, and sleeping. He started worrying that his coworkers were talking about him and plotting to get him fired. By the time his distraught wife brought him to the emergency room, he was fully psychotic. While being shown to his room on the psychiatric ward, the man became convinced he was a condemned prisoner being escorted onto death row.

Because the study was blinded, no one knew whether the man was taking mifepristone or a placebo. But a day or two later, when Belanoff visited the ward, one of the nurses came up to him beaming. “Come see your patient,” she said. Belanoff walked into the man’s room to find him calm and completely coherent. He called Schatzberg: “You’ve got to come see this guy.” Schatzberg, too, was stunned by the change. In more than 20 years of treating people with psychotic depression, he’d never seen anyone recover so quickly. The improvement was so dramatic that Schatzberg jokingly told Belanoff he wanted to “break the blind”—figure out which pills the man had taken—then and there. By New Year’s, the man was well enough to go home, and within a few weeks he was back at work. “It was really amazing,” Belanoff recalls in his understated way.

When the two doctors completed the study, they found that four of the five patients had improved after taking the mifepristone, including the executive. None made as striking a recovery as their first patient did, but the drug appeared to alleviate their psychotic symptoms as well as—or better than—standard antipsychotics could. And what would normally take weeks had occurred over the course of four days.

As small as the study was, it was enough to convince Belanoff and Schatzberg they were on to something big, so they began the process of securing patents on the use of mifepristone as a treatment for psychotic depression. They hoped to interest a pharmaceutical company in developing the drug, and since they were both Stanford employees, they enlisted the university’s Office of Technology Licensing to shop the license to 35 companies. Every single one turned it down.

The contentious abortion issue was just one of the reasons big pharma shied away. Pharmaceutical companies are conservative in general, wary of the complications of inventing completely new drugs when simply modifying existing ones is profitable enough. Scientists created the antidepressant Lexapro, for instance, by rejiggering the molecular structure of another antidepressant, Celexa. Pharmacy shelves are packed with this type of “me-too” drug. As a result, says Alan Leong, director of research at Seattle-based Biotech Stock Research, “innovation happens disproportionately in small companies.”

It soon became clear to Belanoff that the only way this innovation would happen was if he made it happen himself. He, unlike Schatzberg, had the skills to start a business—but he wasn’t thrilled by the idea. “It was a real philosophical fork in the road,” says Belanoff. “I had worked very hard to get to medical school. I really liked being a doctor.” Yet it now looked as if the best thing he could do as a doctor would be to devote himself to developing the drug. “If we can get it approved,” he finally told himself, “that’s also helping patients.”

For every biotech start-up that makes good on a brilliant idea, many more die trying. The risks are even greater when it comes to psychiatric drugs, which present, as Leong puts it, “a minefield of unexpected diversions.” For starters, there are no objective markers—such as blood glucose levels or heart rate—that allow researchers to measure changes in patients’ state of despair, for example, or the intensity of their hallucinations. And sometimes people really do get better purely because they think they’re getting better (the placebo effect). Even experienced practitioners aren’t always sure they’re making an accurate assessment. “I’ve been doing this for more than 20 years. I can tell you it’s very hard,” says Rothschild.

But as friends with whom he’s participated in NCAA office pools can attest, Belanoff lays only very careful bets. In this case, his skills told him the risks were manageable and outweighed by the potential rewards. Not only for his patients, but for himself. The approximately 3 million founder shares he would acquire would be worth, at peak value, about $36 million. (In the end, he’s made far less from the investment: the shares he’s sold over the years—to diversify his net worth, he says—have brought in about $1 million.)

How do you start a biotech company? Belanoff had no idea, until a friend introduced him to David Singer, then a 35-year-old entrepreneur who had successfully founded and run one biotech company. (Singer, coincidentally, is married to San Francisco contributing writer Diana Kapp.) Singer was convinced enough by Belanoff’s earnest and well-documented pitch to help launch the company. He also joined Schatzberg and Belanoff in kicking in $10,000 toward the start-up. It was enough to incorporate the new company, which they dubbed Corcept Therapeutics (for cortisol receptor). The company’s product would be a higher-dose version of mifepristone, or, as it would be called, Corlux (they’ve since developed other products).

Now it was time to start raising real money—so Belanoff rummaged through his closet for his old Wall Street wardrobe. “I had five gray suits and five blue suits,” he says. “Luckily, they still fit.”

Singer had tutored Belanoff in the art of asking for money, but it started to look as if those lessons were all for naught: none of the usual Silicon Valley venture capital firms wanted to hear a pitch that involved RU-486, which in 1998 had yet to win federal approval. One of the few who did agree to listen was Len Baker, a partner at Silicon Valley’s oldest venture capital firm, Sutter Hill Ventures in Palo Alto. Though skeptical at first, Baker and his partners were sufficiently intrigued by what Belanoff, Singer, and Schatzberg had to say to invite them back for another meeting. And another. And another.

After nearly a year of intense question-and-answer sessions, Baker knew he had to come to a decision. So he consulted one of the advisers who had been attending the meetings—Jim Wilson, the legendary 63-year-old biotech veteran. Wilson had headed half a dozen companies and was president of Syntex, maker of one of the first birth control pills and the painkiller Naprosyn. Review the research, Baker told him, and then give us a firm yes or no.

Initially he was ambivalent about the investment. On the pro side, Corcept would be targeting a large patient population, with a product based on a very plausible scientific theory. But there were significant cons, not least of which was RU-486’s hot-button status. The FDA had been reviewing it for several years already and seemed as though it might never give the drug the green light. Wilson also worried that the kind of patents Corcept was seeking—which covered only the “use” of the drug, not the drug itself—were hard to protect from infringement.

Yet after several days of reflection and talking with a few experts, Wilson surprised himself by deciding all those obstacles were surmountable. He became so convinced that he has invested a substantial amount of his own money—he won’t say how much—in the venture.

On Wilson’s advice, the Sutter Hill partners also said yes. In exchange for an initial infusion of $650,000, followed over the next few years by about $13 million, the firm wanted Belanoff to commit to being CEO full-time and to bring on Wilson as chairman of the board.

The only condition Belanoff insisted on was that he still be able to see patients one day a week at Stanford in order to keep his hand in his practice. Ironically, though he spends his days at Corcept thinking about the problem of psychotic depression, he hasn’t seen a single patient with the disease since the first five-person study. “That would be a conflict of interest,” he says.
In July 1999, years after he had read the Dutch report, Belanoff took a leave from Stanford and began working from a small office at Sutter Hill. Partners there joked that they were the only venture capitalists in the Valley with an in-house shrink.

New drugs are a dime a dozen.

But only a handful of the thousands created each year end up being put to the test in clinical trials, which are designed to determine whether they are safe, effective, and as good as, if not better than, drugs already on the market. Any drug that makes it through that gauntlet still has to pass muster with the FDA, which approves, on average, only one in five. All told, companies usually spend 10 to 15 years and more than $800 million to move an experimental drug from lab bench to bedside.

Belanoff was still having trouble securing enough mifepristone to even get to square one. This time, the only viable source was China, so he hired someone to travel to Shanghai and buy enough pills for a study of 30 men and women hospitalized with psychotic depression. Belanoff and Wilson thought it best to keep expectations low and agreed in advance that if the study failed, they might just shut down the company and move on. But within a week or less of taking the drug, nearly two-thirds of the patients showed significant reductions in their psychotic symptoms. The only side effects were a minor rash and, for some of the women, occasional cramps. 

As Belanoff and his colleagues knew, however, this was the kind of study that tends to make drugs look good, since all participants took the drug. Nonetheless, Schatzberg was impressed enough to boast, “The treatment may be the equivalent of shock treatments in a pill.”

He wasn’t the only one who was excited. The New York Times and other papers picked up the story, and the FDA put the drug on the fast track for approval, the first time the agency had ever done so for a psychiatric drug. When the study was published in the journal Biological Psychiatry, it was accompanied by a laudatory editorial signed by Charney and two other leading scientists in the field. “If replicated in a placebo-controlled study,” they wrote, “this finding will have important implications for the treatment of mood disorders.”

Investors now eagerly climbed on board. In early 2001, two venture capital firms, Alta Partners and Maverick Capital, joined Sutter Hill to collectively contribute $28 million; and in coming years the trio would invest another $12 million. (To this day, Belanoff proudly points out, none have sold a single share.)

Belanoff and Wilson decided they would pour every penny possible into the clinical trials, which they would now be able to afford. If the drug didn’t pan out, they would at least “fail efficiently,” says Wilson. That meant running a lean operation: hiring as few employees as possible—the number eventually topped out at 11—and farming out as much of the work as possible, from refining their version of mifeprestone to running clinical trials.

Belanoff surrounded himself with experienced advisers pulled largely from Wilson’s and Schatzberg’s extensive networks. “I used to say that we were the only company in Silicon Valley with no employees under 40,” Belanoff says. Robert Roe, who joined as president in 2001, was an MD and, like Wilson, a veteran of Syntex, where he had shepherded 19 new drugs through development and approval. Another Wilson connection, board member Joe Cook, had resuscitated Amylin Pharmaceuticals, the maker of a novel diabetes drug, levitating it to a $5 billion market cap after it was nearly killed by a series of failed clinical trials.

It’s striking how much support rallied around little more than the tantalizing aroma of potential. The initial trials were small (35 people total) and not particularly rigorous. Indeed, critics would later call into question Belanoff and Schatzberg’s early claims of success, accusing them of making more of the data than was warranted. Both were aware there would be no real evidence one way or another until the company completed its two Phase 2 trials—placebo-controlled, randomized studies involving a sizable number of patients.

In December 2003, Belanoff and Roe met at the Stanford Park Hotel in Menlo Park to wade through the stacks of pages from one of those trials: a study of 221 patients who had been given either the drug or a placebo for seven days and then followed for eight weeks. To see whether this drug would help patients stay well for at least as long as shock therapy or antipsychotic medications do, they had assessed the volunteers at five different intervals during the study.

Poring over the charts and tables, Belanoff and Roe found much to be pleased about. By one measurement, 48 percent of the mifepristone group had fewer psychotic symptoms within seven days, compared with 35 percent of the control group. And the effects did appear to last. Among the patients followed for a full eight weeks, 53 percent of those who took the drug were still better at the end of the study, compared with 22 percent in the placebo group. For the first time, Belanoff felt the thrill of certainty. “It wasn’t until that point that I thought the medicine might really work,” he recalls.

Cheryl B.*, a businesswoman from Olympia, Washington, didn’t need any studies to make her a believer.

The 49-year-old first fell into a deep depression in 1989, almost immediately after her son was born. Lying in her hospital bed, she began having “very paranoid thoughts” about protecting her newborn and hurting herself. Neither she nor her doctors recognized that she was developing postpartum psychosis. (She later discovered both her mother and grandmother had suffered from the same problem.) Each day she had thoughts of suicide, and it took every ounce of energy to resist the urge. She heard voices and saw scary visions. She couldn’t sleep.

She was hospitalized, but that didn’t help, and the stresses in her life continued to escalate. Her baby was colicky, and then her husband joined a cult. Then the marriage broke up, leaving her to struggle as a single mother. Deep depression and psychotic episodes dogged her for years. “I was treated with all the available drugs, but none of them did much,” she says. She knew ECT might help but feared it would damage her intellectual capacity. So for more than 10 years, she says, she “just sort of hung on.”

In 2002, she was seeing her family doctor, who knew about Corcept’s clinical trials and suggested that she sign on. After a decade of unhappiness, she was pretty much willing to try anything, so she agreed. Midway through her first week in the trial, she woke from her first full night of sleep in years, and that day of feeling rested was the beginning of a tremendous improvement.

Over the next few days, she could feel the fog lifting from her brain, and her mental skies stayed clear for more than six months. But then the fog began to creep back. Devilish faces appeared in the trees, laughing at her. The insomnia also returned, as well as the urges to kill herself. Driving her car one day, she swerved to avoid a figure in the road and crashed into a post—but when she looked around, nobody was there.

Cheryl contacted Belanoff to ask for another dose of the drug. The trial protocol allowed for follow-up treatment, so again she took the drug for a week. And once again, midway through the week she woke from “an amazingly wonderful night’s sleep.” Her mental clarity returned.

But after about six months, so did her depression and psychotic episodes. So she began looking desperately for other sources of the drug. She found a group in Arlington, Virginia, the Feminist Majority Foundation, that is authorized to dispense a small number of pills through a “compassionate use program.” But, Cheryl says, her application was turned down. “I tried every avenue to get this drug legally in this country, but I couldn’t get it,” she says. So she took the illegal route.

Cheryl now travels once a year to Asia to buy the drug, which she brings surreptitiously back into the country. Until or unless Corcept wins approval for it, she considers the smuggling her only option. “If I needed to go without it, I would be in serious trouble,” she says. “I don’t want to jeopardize my chance at being able to live my life.”

The ultimate test of a drug is in Phase 3 trials, which follow the same rigorous standards of Phase 2 trials (that is, they’re randomized and placebo-controlled), but the field must be large enough to ensure that the results aren’t the product of some statistical quirk.
Failure at this point is common. The annals of psychiatry are filled with unsuccessful trials of drugs that ultimately made it to the market. The FDA requires manufacturers to present at least two successful clinical trials, which can be surprisingly hard to do. Eli Lilly had to test Prozac in five trials in order to get two in which the drug was shown to work. The makers of Zoloft and Paxil had to run even more. There are many reasons a drug might bomb that have nothing to do with whether it really works, says Arif Khan, a psychiatrist at Northwest Clinical Research Center in Bellevue, Washington. One of Amylin’s drug trials failed “because a few volunteers failed to show up for the final assessment,” says Corcept board member Joe Cook.

Such trials are also extraordinarily expensive, costing on average $30,000 per patient. To mount them, Belanoff would need more money—a lot more money. The answer was to go public with Corcept’s stock. Belanoff and others had earlier considered an initial public offering, but then came the attacks of 9/ 11, after which the economic mood of the country wasn’t right for issuing risky biotech stock. In the spring of 2004, though, Corcept was among the first to go ahead.

As IPOs generally are, the process was brutal. For Belanoff it meant three weeks of nonstop pitching all across the United States and in Europe. To drive home the moral of his story, he included in each presentation a video of a patient receiving shock treatment. “That got people’s attention,” says Roe.

Belanoff had a final conference call in the Milwaukee airport with the local investment bankers and then boarded a red-eye for San Jose, uncertain whether the stock offering would really go through. The market was still ticklish enough that even a tiny tremor could derail the whole deal.

In the end, the offering raised $54 million—not quite as much as the Corcept team had hoped for, but enough to finance the next round of clinical trials. The company began organizing three, to run more or less simultaneously. Each would involve a few hundred patients scattered over dozens of sites.

Investors’ hopes were rising, but so were some critical voices. One pair of skeptics was Bernard Carroll and Robert Rubin, who did some of the pioneering research on the role of cortisol in depression and related disorders and helped lay the foundation for Schatzberg’s hypothesis. Formerly chairman of the Duke University School of Medicine’s department of psychiatry, Carroll is now with a private group in Carmel, the Pacific Behavioral Research Foundation. Rubin is chief of psychiatry for the V.A. Greater Los Angeles Healthcare System and professor of psy¬chology at UCLA.

Together, they questioned how far Schatzberg and Belanoff have taken their theories. Their doubts hardened when they read the early 30-patient study. When they took apart the data, they found no advantage for the drug over the placebo. It looked to them as if the results had not been properly analyzed. They also questioned whether all of the patients in the studies were actually psychotically depressed.

They claimed to find the same shortcomings in Belanoff’s other Corcept studies as well as in one that Schatzberg has produced at Stanford. What irked them was not just the alleged poor methodologies but the rising drumbeat for the drug in both the popular press and scientific publications. “This was as bad a case of talking up a drug as you will ever find,” says Carroll. “They were all negative papers when you applied the standard, scientific rigorous analyses to the data,” says Rubin. “There’s not a shred of evidence that this drug makes any difference in psychotic depression.”

Other experts agree the studies done to date have yet to prove mifepristone’s power. Even Rothschild, coauthor of the cortisol hypothesis, questions the claim that mifepristone can have a long-term effect by “resetting” the stress hormone axis. “From what I know of the theory, I’m skeptical the drug would work beyond seven days,” he says. Still, Carroll and Rubin are alone in calling the company’s research methods into question. And the editor of the journal in which the 30-patient study had appeared declined to publish their critique.

Belanoff is plainly exasperated by the continuing attacks, which he says misrepresented the claims he and Schatzberg have made for the drug. The early studies, he says, were intended simply to test the hypothesis, to “show a hint of whether or not it was worth continuing.” Such small pilot studies are rarely subjected to the statistical analyses Rubin and Carroll call for, he says. “Theirs is sort of an absurd argument.”

He also adamantly rejects any suggestion that he has—or could ever have—manipulated the data to support his hopes for the drug. “This is an entirely data-driven business,” he says. “In the end, your data will always win out.”

By last fall, there was finally some real data to debate: the results of the first two of the three Phase 3 trials. Findings from the third trial came in this March. Each was designed a bit differently. The first enrolled about 220 patients at 23 sites in the United States and eastern Europe, and each person was given either 600 milligrams of the drug or a placebo pill for seven days. The second trial, of 247 patients, followed the same protocol, but all the sites were in eastern Europe. And in the third, 443 people were enrolled at 45 sites, some in eastern Europe but mostly in the United States. And this time, volunteers received either a placebo or one of three different dosages: 300 mg, 600 mg, or 1,200 mg.

Unfortunately, the results were disappointing—but all in the same frustratingly weird way. In each one, the researchers saw no significant difference after one week between the patients who got the drug and those who got placebo pills. Yet, after eight weeks, the vast majority of patients in both groups were much better.

Belanoff was initially puzzled. He knew placebo effects are common in psychiatric drug trials, but the skyscraping 80 to 90 percent response rates were “almost inconceivable.” People in the office began joking grimly that all it takes to cure psychotic depression is enrolling in a Corcept trial.

Critics like Carroll and Rubin saw the results as vindication of their belief that the drug has never worked. But Belanoff thought they were so off-the-chart strange that he became convinced something unusual was going on. He and Roe began combing through the data looking for clues.

They soon discovered dramatic differences between the ways in which patients at the various sites had responded. It turned out that the sites with the wildly high placebo rates were mostly in eastern Europe, in Serbia, Croatia, Bulgaria, and Romania. Belanoff isn’t sure why the results were so off-kilter, but says there are lots of ways to muddy the waters. And he’s still trying to sort them out. At some of the sites, for example, the doctors were reportedly recruiting patients to the study with promises of “a miracle drug.” That in itself could raise expectations so high that a patient might recover regardless of which pill he or she got. Whatever the cause, Belanoff believes that if the eastern European sites had not been included in the trials, the results would have been more positive.

That’s why no one in the company thinks the trials were a total bust. In fact, the last one, which involved 443 patients, offered the kind of clear signal of the drug’s efficacy Belanoff had been waiting for. The data showed that, regardless of the dose, once a patient had a high enough level of the drug circulating in his or her bloodstream (patients’ ability to metabolize the drug varied), the psychotic symptoms were significantly reduced. “This is really an encouraging result for us,” says Belanoff. “It gives us a path forward.”

Yet that path depends not only on FDA approval but also on more support from investors. By now, Corcept has burned through most of the $93 million it has raised, and the value of its stock has plunged. But there is a silver lining. “One of the benefits of your stock taking a hit on bad news is that it makes it easier to finance the company,” says Wilson. He is hopeful they can raise the $20 million or more it would take to run a new set of Phase 3 trials. Indeed, in late March, he and other current investors plunked down another $9 million for Corcept stock—a sign that they are still behind Belanoff in his assertion that the drug works, and that ultimately the company can design trials that will show that.

That’s enough to keep Belanoff going, he says. He never imagined that after 10 years, he’d still be stuck so near the starting gate. But in the world of drug development, the second half of the race is what really counts.