Nine years ago, Rusty Gage shattered a neuroscience dogma when he showed human brains give birth to new neurons. Today, a company is eager to take those findings to the clinic.
By Kerry Grens

One day in mid-January 2006 Todd Carter, the director of biology at BrainCells, Inc., was sitting in front of a computer screen counting blue neurons with green snaking processes. An automated neuron counter had given him some promising results, and he was going back to check the raw data. These neurons started off as a sheet of human neural stem cells, and when Carter applied a compound to them they proliferated and developed into neurons. When he did the experiment again he received the same results: neurogenesis. As he counted the number of neurons, glia, and stem cells, the data matched - this compound was making new neurons at rates as good as antidepressants. It was just what he was looking for.

An enlarged view of a rat dentate gyrus granule cell layer highlighting newly generated neurons

In searching for a potential new drug that would stimulate neurogenesis similar to the way Prozac and other drugs act on animal brains, BrainCells had screened 539 compounds on countless neural stem cells before coming across this one. The compound, now called BCI540, seemed to promote neurons with reasonable potency and was not toxic to cells. When Carter blocked the compound from acting, neurogenesis vanished. "Based on the antidepressants that have been run, we knew what we were looking for as to what would be a good antidepressant," he says.

Carter decided to probe the drug's mechanism, and on January 27, 2006, the data came back; the drug was unique, not acting on any of the same neurotransmitters that antidepressant drugs do. BrainCells' scientists gathered to review this latest bit of BCI540's data. Carter was thrilled by the compound's unique action. "It was really exciting to know this was real, and unique. [I said,] 'Let's get it in vivo'." He turned to his colleague Andrew Morse, the company's director of pharmacology, and said, "I think I know what you'll be working on for the next month."

For three weeks Morse gave several groups of animals different doses of BCI540 and then measured their responses in two behavioral assays. In the first, called novelty suppressed feeding, Morse placed the animals in a big open chamber and measured how long it took them to eat a pellet of food placed in the middle. Chronic antidepressant use will expedite an animal's approach to the food, and that's what Morse observed. "...we got a very clear result, it looked like a classic antidepressant." In the second test Morse used, novel object recognition, antidepressants tend to reduce an animal's avoidance of a new object and increase exploration. "The novel object [assay] was not as straightforward. But we don't expect all the assays to give us the same signal." But when Morse examined the brains afterward the effect of the drug was as good as an antidepressant. "Before the data were all in the numbers were jumping off the screen. We got a very clear neurogenic effect," Morse says.

BrainCells is hoping that BCI540 is the punchline of a scientific story that began nine years ago when Rusty Gage, at the Salk Institute in La Jolla, Calif., published the first evidence of neurogenesis in human brains. Since then, BrainCells' founders have shown that all the antidepressants now in use, from agomelatine to Prozac, induce neurogenesis in animals. That made them think that other compounds that induce neurogenesis might be good antidepressants, and they hope BCI540 will be the next.

"The failures in central nervous system drugs have tended to be late and for failure of efficacy, in part because there hasn't been a good biological basis for [the action of drugs on] many of these diseases, particularly in psychiatry," says BrainCells CEO Jim Schoeneck. His company's approach is to try to weed out those failures ahead of time through their neurogenesis screens. But whether neurogenesis will be enough to overcome a complex, widely variable, and chronic disease is far from clear.

Atop Potrero Hill in San Francisco, wall-to-wall windows in Luca Santarelli's apartment boast an impressive view of the city's skyline, hemmed in by the Bay and Golden Gate Bridges whose peaks slowly disappear into the maws of fog. Seated at his dining room table, Santarelli, now head of Roche Palo Alto's CNS research group, recalls leaner times when he was a postdoc in Ren￩ Hen's laboratory at Columbia University. At the time, Santarelli was subletting an apartment from a woman who lived on Manhattan's East 49th Street. Dutifully he collected her mail, never giving much attention to her Scientific American subscription until one day in 1999. On the cover of the magazine, adjacent to a frothing tsunami, were the words, "New brain cells: growth hints at neural repair."

"I was totally captured by the story," Santarelli says. The article, written by Gerd Kempermann and Fred (Rusty) Gage, described how neurogenesis was discovered, and laid out hypotheses on the regulation and function of the phenomenon. Santarelli had heard of neurogenesis; just one year earlier Gage, at the Salk Institute in La Jolla, Calif., published his seminal paper demonstrating evidence of neurogenesis in the dentate gyrus of human brains,¹ and Liz Gould and Bruce McEwen at Rockefeller had published a suite of studies on the effects of stress on neurogenesis in rodents.² But the field was sparkling with newness, and huge questions remained. "At the time there was speculation about the modulatory factors of neurogenesis. Kempermann already showed exercise and an enriched environment increased it. But what are the molecular components?"

There were hints that serotonin might be involved. A year earlier, in 1998, Gould, now at Princeton, teamed up with her next-door laboratory neighbor, Barry Jacobs - known to some as Mr. Serotonin - to see if the neurotransmitter might have an effect on neurogenesis. They applied 8-OH DPAT, an agonist to the 5-HT1A serotonin receptor, or fenfluramine, which stimulates massive serotonin release. "And lo and behold we had loads of proliferation,"³ Jacobs says. Shortly following, two French investigators showed that depleting serotonin decreases neurogenesis in the hippocampus.⁴

At the same time Santarelli had begun work on a knockout mouse whose 5-HT1A serotonin receptor was lost. Though he had not published the work yet, Santarelli found that these mice did not respond to a number of antidepressants. He began to put pieces of the puzzle together: "I started wondering whether or not antidepressants could have some of their efficacy by changing the process of neurogenesis," Santarelli says. "And I thought, maybe my mouse is going to tell me something. Not only do I have a mouse where there's a disruption in the serotonin system, but also a disruption in the response to drugs." When Santarelli brought his ideas back to Ren￩ Hen, "he thought I was crazy."

Ren￩ Hen sits in his corner office on the seventh floor of the New York State Psychiatric Institute in upper Manhattan. Leaning forward with his elbows on his knees, Hen remembers that he didn't want to invest too much time in Santarelli's neurogenesis project, and says that his hesitation was warranted. "When you think of classic hippocampal functions, you think [of] learning and memory," says Hen. "You wouldn't have thought changes in the hippocampus could change mood."

Nevertheless, Hen agreed. The plan was to ablate neurogenesis and test whether antidepressants still worked. Santarelli and a graduate student, Michael Saxe, went for a cheap trick: They decided to use low doses of radiation to kill dividing cells in the hippocampus.

Searching around Columbia's medical campus, they found a 35-year old Siemens Stabilopan X-ray System that had been retired from use in breast cancer treatments. What they needed next was a lead apron that would fit over the mouse and allow only the hippocampus to be irradiated. Next door to the radiation room was a carpentry shop run by an expert in probe design for radiation research. In exchange for mouse-sized aprons with tiny slits above the hippocampus, Santarelli would buy him Jack Daniel's bourbon or 12-packs of beer. "He didn't drink on the job!" Santarelli jokes.

Meanwhile, the stirrings of a new theory in depression were beginning to emerge. In 2000 Gage and Jacobs, who took a sabbatical from Princeton to join Gage's laboratory for a year, proposed a theory of depression in which decreases in neurogenesis in the dentate gyrus precipitate depression, and pharmaceutical interventions that increase serotonin improve neurogenesis and relieve depression.⁵ In the meantime, Ron Duman's group at Yale, who assisted Santarelli on his project, found that antidepressant treatment increases neurogenesis.⁶

The findings confirmed some of Santarelli's suspicions, but Hen reminded him it was merely an association. The ablation experiments would show whether neurogenesis was necessary for antidepressants to be effective. After some time spent getting the radiation protocol optimized, the experiments worked. In August 2003 Santarelli and his collaborators published an article in Science showing that antidepressants do indeed require neurogenesis to change behavior.⁷ "I was thrilled," says Gage, who was anxious to learn more about the functional role of neurogenesis. "I thought it was great."

"From the moment we found neurogenesis was necessary for antidepressants, one of the natural consequences of that finding was, if we identify compounds that stimulate neurogenesis, we could maybe get new antidepressants," says Hen. There is certainly a market out there for new therapeutics, he adds. Though dozens of antidepressant drugs are on the market, there is still room for more, agrees John Rush, a professor of psychiatry at the University of Texas Southwestern Medical Center. "Our current therapeutic armamentarium certainly leaves a lot to be desired," Rush says. "A lot of the drugs out there are similar." He says about one-third of patients with depression are resistant to antidepressant treatments.⁸

Moreover, Rush says, about 10%-15% of patients switch drugs because of side effects such as insomnia, anxiety, and sexual dysfunction. Hen says that stimulating neurogenesis might bypass the serotonergic system (thought to be responsible for the side effects) and perhaps improve the tolerability of the drug. Still, Rush says, "most of the drugs are pretty well tolerated ... even a new drug, even if it doesn't affect those neurotransmitters, could affect other brain functions and have [its own] side effects."

As Hen and Santarelli recruited cofounders (including Nobel laureate Eric Kandel) and began discussing their business plans, another group was also hatching plans to capitalize on neurogenesis. Oxford Biosciences, a venture capital group based in Boston, knew there was opportunity in neurogenesis, but they weren't sure what it was. In 2003 they hosted a neurogenesis think tank in Westport, Conn., at which Kandel, Gage, Hen, Santarelli, and a few others attended and presented ideas. Gage had developed protocols for positive neurogenesis controls, optimal neural stem cell cultures, and for labeling and measuring proliferation, differentiation, and survival of new neurons; Santarelli had already received his results from the neurogenesis ablation experiments. Essentially two business ideas shook out of the meeting: Hen and Santarelli wanted to stimulate neurogenesis to treat depression, and Gage had developed in vitro assays that could be used to screen compounds for their effects on the different stages of neurogenesis.

Gage's assays could be used in a general way to measure a drug's neurogenic effects, but Hen's proposal would give those screens focus. "What every good venture capitalist tries to do is to put the best of the best under one tent," says Ellen Baron, a partner at Oxford Bioscience. She and another partner, Jonathan Fleming, offered to develop one "supercompany" that would target Gage's screening tools toward depression. Oxford Biosciences attracted other investors, including Bay City Capital, Technology Partners, and Pappas Ventures, and raised $17.7 million for their initial investment.

Baron and Fleming also thought that BrainCells could bring drugs back to life - drugs that other companies had developed and abandoned. "One way one uses these [screening tools] are as new eyes for existing compounds, because then you can leapfrog the potential of that compound into the clinic," Baron says. By choosing drugs that have already passed through safety trials, BrainCells could potentially shave years off drug development and save millions of dollars as well.

Leasing drugs from companies hasn't been an easy process, however. Before BCI540 surfaced in BrainCells' screens, several other drugs showed promising results. BrainCells approached a handful of companies to begin negotiations on in-licensing their compounds, but when BrainCells presented the neurogenesis data from the potential drug, "they no longer wanted to give us the compound," says Carrolee Barlow, BrainCells' chief scientific officer. The problem, she says, was the size of the companies they were talking to. "A company that is very small and only has two drug assets, if one is on the shelf, it's more valuable on the shelf than if it's gone." In-licensing is like a permanent lease, and Barlow says that once companies discover that their compound might have useful properties, they are not willing to give it away. What those companies would eventually do with those compounds was a mystery. "We stopped looking to repurpose drugs from very small biotech companies," Barlow says.

BCI540 resulted from that shift in strategy. The owner is $2 billion-a-year Mitsubishi Pharma, which tried the drug for a nonpsychiatric neurologic disorder on more than 300 patients in the United States. Citing competitors who may want to take advantage of their technology, BrainCells won't disclose the identity of the compound or its mechanism of action. Barlow will say only that it does not act upon serotonin, norepinephrine, or dopamine pathways as traditional antidepressants do. She says that patients took the drug for six months with side effects no different from placebo. None of the patients were assessed for changes in mood. Ultimately the drug failed in Phase II trials for lack of efficacy for its original indication.

Once Morse's in vivo results showed that BCI540 was neurogenic and acted like an antidepressant in the novelty-suppressed feeding assay, BrainCells decided to go after the compound. "It was really exciting because we knew it was already safe in humans," says BrainCells' Carter. BrainCells approached Mitsubishi Pharma with a proposal that each company would disclose information about the compound to one another and decide if in-licensing was desirable. Five months of negotiations followed to establish the terms of an in-licensing agreement, which BrainCells' CEO Schoeneck says is a fast turn around. Oxford Biosciences' Baron says the money that was raised originally for the company in the first round was sufficient to cover the costs of in-licensing. Of the $25 million raised for the company so far, they've spent $18 million. Now the company is working on raising another $37 million to bring the drug to humans. Barlow says the board agreed to pursue BCI540 because of the excitement of finding a drug that acts as well as antidepressants in their assays, but doesn't work on serotonin.

BrainCells plans to enter Phase IIA clinical trails once the funds are in hand. A little less than one hundred patients with major depressive disorder will be given a dose of BCI540 either once or three times per day for six weeks. The pills have already been manufactured and if all goes as planned the results should be available sometime in 2008. Barlow says that if BCI540 can meet efficacy goals, the idea is to collaborate with a large pharmaceutical company in the development of a new antidepressant.

BrainCells has a few back-up plans in case BCI540 fails. Schoeneck says BrainCells' screening platform can be used to partner with other companies that want to uncover neurogenic properties of their compounds. The company has already partnered with Organon, a division of Akzo Nobel, to screen Organon's compounds that stalled along the clinical pipeline for neurogenic properties. Barlow says BrainCells' technology might have other applications, such as a treatment for macular degeneration or hearing loss. BrainCells also continues to screen compounds for potential antidepressants. Still, success is no guarantee. As Hen says, "We proved [neurogenesis] is necessary. Now what they are trying to do is prove it's sufficient."

Fritz Henn, at Brookhaven National Laboratory, is skeptical. At the same time Santarelli and his colleagues were demonstrating the role of neurogenesis in antidepressants, Henn was looking at whether a decrease in neurogenesis could induce depression. Henn hypothesized that every animal with impaired neurogenesis should show a change in behavior. But when he induced a decrease in neurogenesis by exposing animals to the same stress, only a fraction of the animals exhibited depressive behaviors.⁹ "Given our findings ... I've argued that looking for drugs that specifically increase neurogenesis doesn't seem to be the right approach," says Henn.

Henn's hunch is that the key to relieving depression lies in synaptogenesis. "What really matters in depression is not how much you knock down neurogenesis, but how many cells really integrate into the system." Where and how is unknown. "There is more to depression than just neurogenesis," says Bruce McEwen at Rockefeller. A suite of changes occurs in the brains of people with depression: hippocampal volume reduction, decreased density in glial cells and neuronal size in the prefrontal cortex, and changes in blood flow and glucose metabolism in the hippocampus and amygdala. Stress also causes extensive dendritic remodeling in the hippocampus. "So the question is," McEwen says, "whether these small molecules that work on neurogenesis in the hippocampus also work on these other parts of the brain involved in depression. I don't think neurogenesis is the be-all and end-all of depression, but it's certainly very important."

As BrainCells brings its compound to humans, the experiment could help answer a looming question that Princeton's Jacobs has posed: "Do any of these animal models have anything to do with human clinical depression?" Investigators at BrainCells admit behavioral assays in rat are imperfect models for human mood disorders. Irwin Lucki at the University of Pennsylvania points out that one of BrainCells primary models, novelty-suppressed feeding, is more a proxy for anxiety than depression. "It remains to be demonstrated that models of depressive behavior and neurogenesis are related," Lucki says.

Still, Lucki says screening compounds for their neurogenic properties is appropriate for determining their effects as a chronic antidepressant (he's currently working on a similar project with Wyeth). Santarelli, who is no longer involved in BrainCells because of a potential conflict with Roche, where he now works, says going after neurogenesis is worth a shot, simply to try something new. "There isn't much novelty in depression in the pipeline," Santarelli says. "The only way to break out the mold is to do things like this."