As antidepressants come under scrutiny for their suspected link to suicide, skeptics question how safe and effective the 'miracle drugs' really are. But as Apoorva Mandavilli discovers, the prospects for a better alternative are downright depressing.
In 1988, Julio Licinio was finishing his residency in psychiatry at the New York Hospital-Cornell Medical Center. In those days, the mainstay of treatment for patients with depression was a class of drugs called tricyclics. The drugs caused serious side effects and the risk of a fatal overdose, so patients were closely monitored.
That year, the US Food and Drug Administration (FDA) approved fluoxetine, a selective serotonin reuptake inhibitor, or SSRI, for depression. Marketed by Eli Lilly as Prozac, fluoxetine rapidly gained a reputation for being as effective as tricyclics, but safer and with fewer side effects. Within a year, sales of Prozac rocketed to $350 million. "It was a big turning point," says Licinio, now a professor of psychiatry at the University of California in Los Angeles.
North Wales department of psychological medicine
But that might change: in recent months, the UK has banned the use of most SSRIs in those under 18 and the FDA is meeting on September 13 and 14 to review the link between the drugs and suicide in adolescents. Although most of the controversy surrounds the drugs' use in children, there is renewed attention on whether they are safe even in adults.
If new analysis finds that SSRIs and other drugs of its class are unsafe, what are the alternatives? Pharmaceutical pipelines for antidepressants are discouragingly dry. What's more, any new drug for depression might carry the same risk of suicide-unless pharmaceutical companies can design one that would rapidly lift melancholic moods, says Licinio. "But such a drug doesn't exist."
Severe depression is often characterized by suicidal impulses but most patients are so lacking in energy and paralyzed by their illness-a condition dubbed 'psychomotor retardation'-that they don't act on those impulses. But when patients are treated with antidepressants, different symptoms of the illness lift at different rates. "The psychomotor retardation is the first thing to go, existential sadness is the last thing to go," says Licinio.
Psychiatrists recognize that, by giving people the energy to act out their destructive thoughts, the drugs effectively make people acutely vulnerable to suicide. But family physicians and internists, who have increasingly taken on the treatment of depressed patients, might not know that. "Treatment of depression has this built-in period when people become more suicidal before they get better-we were all taught this," says Licinio. "That never went away, people just forgot about it."
Mysteries of the mind
But for decades, every new contender for that market has been just a slight variation on its predecessors, in large part because no one really knows what depression is or what causes it.
We know, for instance, that depression has a familial or genetic component, that it is attributable in part to life events, particularly those that occur early in life, and that it is accompanied by high levels of stress hormones and anhedonia-a lack of enjoyment in food, sex and other pleasurable things. Beyond that, the details are sketchy.
With only a tenuous grasp on the workings of the depressed mind, scientists have been hard-pressed to find drugs to treat it. "All the molecular targets we point to seem to be linked somehow to the symptoms," says David Schulz, executive director for CNS Discovery Biology and General Pharmacology at Pfizer. "We have no convincing evidence at all that [the existing drugs] contribute to the underlying pathology."
In fact, the current repertoire of drugs is largely a result of luck. The first antidepressants were discovered in the 1950s when people being treated for tuberculosis reported feeling happier and more energetic. Scientists later found that the drugs increase the levels of a group of chemicals called monoamines that transmit messages across nerve cells. Existing antidepressants work either by preventing the breakdown or by inhibiting the reuptake of these monoamines-serotonin, norepinephrine and dopamine-from the synapse.
Tricyclics nonselectively raise levels of serotonin and norepinephrine. Most new drugs are selective for serotonin, and a few-such as reboxetine, marketed as Edronax in Europe-target norepinephrine. Some of the drugs increase appetite, others have a sedative component, yet others dramatically boost energy levels. Because different people respond to different drugs, experts say it is important to have so many variations on the same theme. But overall, differences in the drugs are marginal.
At the level of neurons, the drugs' action is seen within hours, but full clinical effect can take up to eight weeks. "What that tells us is that all the drugs we have are acting so remote from the mechanism that produces the antidepressant effect," says Florian Holsboer, director of the Max Planck Institute of Psychiatry in Munich. "So there must be something else in between."
To tap more closely into the basic underpinnings of depression, companies are exploring targets in stress-hormone regulation, brain-specific proteins and circadian rhythms. For instance, about 75% of Wyeth Pharmaceuticals' programs in depression and anxiety focus on new mechanisms. But drugs that rely on new mechanisms are in the very initial stages of being tested. For the time being, monoamines continue to be the primary targets.
Is more better?
Approved this summer, Eli Lilly's Cymbalta exemplifies a new trend in antidepressant development. Like tricyclics, Cymbalta raises levels of both serotonin and norepinephrine. GlaxoSmithKline's (GSK) Wellbutrin XL, approved in 2003, affects norepinephrine and dopamine. In early August, DOV Pharmaceuticals licensed two triple reuptake inhibitors-which target all three monoamines-to Merck. Another triple reuptake compound at DOV is set to enter trials later this year.
The rationale is that by aiming at multiple receptors, the drugs will either be faster or capture a bigger percent of responders, says Phil Skolnick, chief scientific officer of DOV Pharmaceuticals.
But critics say that the combination drugs are a marketing ploy. "[The theories] are so speculative and they're presented with so much authority," says Jon Jureidini, head of psychological medicine at Women's and Children's Hospital in Adelaide, Australia. When SSRIs first emerged, the companies defined depression as imbalances in serotonin, Jureidini notes. "Now that they're promoting the serotonin [plus] norepinephrine drugs, it's not just all about serotonin anymore. It's about other neurotransmitters," he says. "The science follows the marketing to a certain extent."
Jureidini and others say there is no evidence that the double reuptake drugs are any better. In fact, the FDA's analysis cites the highest number of problems with venlafaxine (Effexor), which raises levels of both serotonin and norepinephrine.
Experts outside the industry are more optimistic about antagonists of the corticotropin releasing factor (CRF), a stress hormone. The link between stress and depression is well-documented and there is evidence that levels of cortisol are elevated in the spinal fluid of depressed patients. "That's probably the best biology that's been worked out for depression," says Ranga Krishnan, chair of psychiatry at Duke University.
The only human trial of a CRF antagonist was a small but promising proof-of-principle study by Holsboer's group in Munich. But at high doses, the compound produced a troubling increase in liver enzymes, and was shelved when Belgium-based Janssen, which held the license, was acquired by Johnson and Johnson (J & J).
In 1996, Pfizer also abandoned a CRF antagonist when researchers there uncovered toxicity in preclinical studies. Still, both Holsboer and Pfizer maintain that the problems were specific to those compounds and have no bearing on the validity of the hypothesis. Nearly every major company is now pursuing CRF antagonists, which are widely acknowledged as the lead alternative to available antidepressants.
Most of the CRF compounds are in phase 1 trials or earlier. Because cortisol regulates immune response to infections, a successful CRF antagonist would have to act on the brain receptors but not in the periphery. "That's not an easy task," notes Krishnan. As a result, he says, "everyone is going very, very slowly and making sure they understand the pharmacology."
Lower on the list of candidate drugs are those that target substance P, a member of the neurokinin family of brain-specific peptides. Merck's report in 1998 that its substance P antagonist was a powerful antidepressant marked the first promising results from a non-monoamine drug1. But in subsequent phase 3 trials, the drug proved ineffective. It is now being marketed for nausea associated with chemotherapy.
Merck, Pfizer and others have since shelved their substance P programs, but a few companies, including GSK, continue to pursue them. Other drugs in early development target receptors for the neurotransmitter glutamate, or attempt to reset the disrupted sleep patterns in depressed patients.
Trials and error
The biggest hurdle in developing drugs that veer from the monoamine hypothesis is that there are few animal models to validate them. The scarcity of models is a catch-22 of sorts: depression is considered a state unique to human beings, meaning animals do not get depressed, so animal models can only reflect pieces of the symptoms or behavior.
Most existing models recreate monoamine defects and may not be predictive for the new classes of compounds. But that's not likely to stop the companies. "If we failed to see activity, we wouldn't be discouraged by it," says Pfizer's Schulz. "If you believe strongly in your hypothesis and there's a lot of science behind it, you should go ahead anyway."
Once in human trials, the compounds are up against a formidable 'placebo effect,' which dominates discussions about antidepressant development. Some studies say that in as many as half of all depression trials, the drugs do not prove significantly better than a placebo. Critics argue that this is because the drugs are no better than placebos. But many experts-including those without financial ties to the industry-say the placebo effect is a real challenge.
Depression is often defined by arbitrary and subjective symptoms rather than concrete physiological data, so the kind of people recruited into trials can vary wildly. Companies also shy away from the severely depressed and increasingly enroll those with mild or transient depression, the very group most likely to respond to placebos.
Future trials are also likely to be scrutinized carefully for any link between the drug and suicide. One analysis by David Healy, director of the North Wales department of psychological medicine, suggests that in mildly depressed or anxious people and even among healthy individuals, the drugs can trigger suicidal thoughts. "The risk of suicide seems if anything to increase with antidepressants," says Healy, who has consulted for many manufacturers.
But Healy's analysis is hotly debated. Some, like Licinio-who has no ties to manufacturers of antidepressants-say the risk is inherent to treating depression. Others say most suicides are by people who are not on antidepressants. "If anything, SSRIs lower the risk of suicide statistically," says DOV's Skolnick.
But most agree that much more needs to be done to understand exactly how the drugs affect depressed brains. "I don't dispute the numbers. What I don't accept from [Healy's] analysis is that there is a direct causal link," says Spilios Argyropoulos, consultant psychiatrist at South London and Maudsley NHS Trust. "It's not a straightforward question and it's not as straightforward an answer as it's made to be."
Gray skies ahead:
There is almost universal agreement among experts that antidepressants are overused and oversold. But even the most skeptical of critics agree that the drugs are sorely needed. "I'm in favor of using these drugs but I think in the case of the people less severely ill, you've got to take much, much more care," says Healy.
Pushing the drugs too hard, too fast, without enough information about their safety, can eventually bring on too many restrictions-and keep them from the people who really need them, Healy says. Lawsuits from various governments are pressuring companies to be more forthcoming about negative results from trials so that people can be aware of the risks long before there is a crisis. GSK has released results from its trials on its website. Merck and J & J have both supported the idea of a clinical trial registry and Eli Lilly has said it would list all trials for its approved drugs by the end of the year.
But it is not clear how much of a difference that will make. Lilly's pledge only applies to its approved drugs: any unpublished information about earlier-stage drugs has to be cleared through a lengthy approval process. Merck will not disclose information on any drugs before they enter phase 3 trials.
Still, the controversy is likely to change the way the industry operates. "The way it's affecting us is that we're a little more conservative in launching trials in the first place," says Schulz. "The days when we throw a lot of money at depression and hope for the best are probably over."
Realistically, companies will also have to abandon their hope of developing a drug that will work in all people, says Holsboer, who predicts that future antidepressants will be targeted to specific subgroups of patients. Holsboer, Licinio and others are trying to find genetic variations that might predict whether someone will respond to a given antidepressant. Others are hunting for genes that might predispose people to depression.
In the meantime, people might begin to think twice before popping the pills, and consider opting instead for psychotherapy. Physicians might also exercise much more caution before prescribing the drugs, says Licinio. "At least this will alert people that this is not cosmetic psychopharmacology," he says. "You have to follow the patients."
- Kramer M., et al. Science, 281. 1640 - 1645 (1998).