In 2000, researchers from the Yale University School of Medicine made a surprising discovery that would start to change the way we think about the causes of depression. Ronald Duman and his colleagues chronically administered different classes of antidepressants to rats, and found that this stimulated the growth of new neurons in the hippocampus. As a result, researchers and clinicians began to think of depression as something like a mild neurodegenerative disorder, rather than as a chemical imbalance in the brain.

Earlier studies had already suggested that depression involves shrinkage of the hippocampus caused by death of cells in that part of the brain, and the findings of Duman’s group further suggested that antidepressants may be effective because they prevent or reverse the loss of cells. Despite a lack of direct evidence for this, the findings have opened up an avenue for the development of new treatments for depression: earlier this year, for example, a company in San Diego began a clinical trial to test the efficacy of a compound which stimulates neurogenesis.

Researchers from the University of Texas Southwestern Medical Center now provide direct evidence that antidepressants work by stimulating neurogenesis. They show that a receptor for a molecule called brain-derived neurotrophic factor (BDNF) governs neurogenesis induced by antidepressants, and is required for sensitivity to these drugs. As well as showing that the effectiveness of antidepressants is indeed dependent on their ability to stimulate neurogenesis, the findings may also begin to shed light on why a large proportion of depressed patients do not respond to medication.

BDNF and other growth factors play essential roles during embryogenesis, but their functions are not restricted to development, and BDNF in particular has previously been implicated in depression. BDNF levels in the rodent hippocampus are elevated in response to chronic exposure to antidepressants; infusion of BDNF into the hippocampus stimulates neurogenesis and produces antidepressant-like effects in behavioural tests; and mutant mice with reduced BDNF levels have blunted responses to antidepressants.

These observations strongly suggest that BDNF is involved in mediating the response to antidepressants of neural progenitor cells (NPCs) in the hippocampus. It is because of the presence of NPCs that new cells can be generated in the hippocampus throughout life. They have stem cell-like properties and retain the ability to divide asymmetrically, thus propagating themselves while at the same time generating new neurons which can be incorporated into existing hippocampal circuits.

With this in mind, Li et al generated several strains of mutant mice lacking the BDNF receptor TrkB. The mutants created conditional knockouts; in other words, they were engineered in such a way that they lacked the TrkB receptor only in specific cell types at specific times. (The researchers used the Cre-LoxP system to generate the mutants.) In one strain of mutants, expression of the receptor was disrupted in NPCs either during embryonic development or in adulthood; in another, TrkB expression was disrupted in differentiated cells throughout the hippocampus.

Those mutants in which TrkB expression was deleted during embryonic development were born with brains of a normal size. Beginning at postnatal day 10, they exhibited a 30% reduction in hippocampal volume, which persisted throughout adulthood. This is interesting, as it suggests that the mechanisms governing cell proliferation during embryonic and postnatal development are different from one another. In the second strain of mutants, in which the receptor was deleted in adulthood, there was a significant decrease in the number of newly-generated hippocampal neurons. In the third strain, deletion of the receptor from differentiated neurons had no apparent effect.

These findings show that the actions of BDNF on progenitor cells, mediated by TrkB, is necessary for the normal proliferation of cells in the hippocampus, both in newborn and adult mice. They were confirmed by a further experiment in which cells were isolated from the dentate gyrus of the embyronic and adult mutants lacking TrkB in the progenitor cells, and then grown on culture dishes. Under these conditions, the cells spontaneously aggregated to form spherical clusters called neurospheres, which increased in size when BDNF was added to the culture medium. Thus, the growth factor acts directly on NPCs to induce proliferation.

The mutants lacking the TrkB receptor in the NPCs were then treated with fluoxetine (Prozac), imipramine (a tricyclic antidepressant) or saline (salt water) for 3 weeks and the amount of time they took to begin feeding was recorded and compared to healthy control mice. This is a behavioural measure of depression-like symptoms – in animal models, “depressed” mice take longer than healthy ones to reach for food. In this test, the healthy mice that were treated with either antidepressant showed a much shorter latency to feeding compared with the controls given saline and the mutants.

The animals were subjected to the tail-suspension test, which is a measure of stress. When suspended by their tails, healthy animals and those treated with antidepressants quickly become immobile; this is a sign of despair in response to a stressful situation from which they cannot escape. This time, the control mice treated with either fluoxetine or imipramine quickly became immobile when suspended, whereas the mutants treated with the same drugs did not.

When the animals’ brains were examined, it was found that the mutants treated with both antidepressants and those who had exercised on the running wheel had significantly fewer newly-generated hippocampal cells than controls. Likewise, voluntary exercise on a running wheel induced hippocampal neurogenesis in control mice but not in the mutants. In all these tests and analyses, mice lacking the TrkB receptor in differentiated hippocampal neurons were comparable to the controls.

Together, these findings show that deleting the BDNF receptor TrkB from hippocampal progenitor cells, but not differentiated neurons, make embryonic and adult mice insensitive to the effects of two different classes of antidepressants. The study provides a direct link between BDNF, neurogenesis and the beneficial effects of antidepressants. It shows that two classes of antidepressants – SSRIs and tricyclics – exert their effects by stimulating neurogenesis, and that this effect is mediated by BDNF. The molecular mechanism by which this occurs is, however, still unknown, and is likely to be the focus of future work.


Li, Y et al (2008). TrkB Regulates Hippocampal Neurogenesis and Governs Sensitivity to Antidepressive Treatment. Neuron 59: 399-412. DOI: 10.1016/j.neuron.2008.06.023


  1. #1 Becca
    August 21, 2008

    So if grad school makes me depressed, does that mean it is actually making me dumber?

  2. #2 Jen
    August 22, 2008

    Sorry if this is a little off-topic, but I’ve never heard of engineering lacking receptors in certain cells at certain times. My brother was a particularly precocious toddler, saying things like “Put the avocado pit in the ashtray” at 12 mos, and was diagnosed with schizophrenia at 27. My father tells me that based on his research this is not uncommon- too precocious too early can mean adult onset mental illness later. Your post here makes me curious about how much we know about what governs adult onset conditions, particularly ones with a precociousness component. In this study, were they trying to mimic adult onset conditions with the conditional knockout mice in which “expression of the receptor was disrupted in NPCs either during embryonic development or in adulthood”? Is the “either” an indication of how difficult it is to directly affect one time period over another?

  3. #3 Abel Pharmboy
    August 22, 2008

    Mo, makes me wonder where the field is with prospectively-designed BDNF-mimetics. I know that BDNF was being investigated by itself as a treatment for parkinsonism but the company had to devise an indwelling transcranial delivery system. Not sure if the TrkB binding site is amenable to small molecules that can cross the blood-brain-barrier. I’d wager this is old news to pharma folks in the know.

  4. #4 Mo
    August 22, 2008

    Jen: The conditional knockout is now a commonly used technique. The technique is very time-consuming: a KO mouse takes at least a year to generate, but it is no harder to disrupt a gene during embrygenesis than it is to knock it out in adulthood. The key is to use different promoters, genetic elements which regulate where and when a particular gene will be activated. The researchers knocked out TrkB in enbryos and adults to show that the effect of BDNF on progenitors occurs throughout life.

    As for the causes of adult onset conditions, they vary depending on the condition. Some are inherited, some sporadic and others have genetic components which make
    one more susceptible. Some conditions are likely to be casued by a combination of genetic and environmental factors. In the case of schizophrenia, the causes are still unknown, but several very recent studies identified rare chromosomal deletions associated with the condition.

    Abel: Here’s a 2003 paper about potent BDNF mimetics, from the Journal of Biological Chemistry, so it would appear that small molecules can bind to TrkB. I haven’t heard anything about this kind of work recently, but I would imagine that it’s ongoing.

  5. #5 Ryan Morehead
    August 22, 2008

    Also, if BDNF is injected into the Nucleus Accumbens (shell?) it has a depressive effect. I think proBDNF infusions to NAc have the opposite effect.

    The interaction between NAc and hippocampus in depression is something in which I’m particularly interested.

  6. #6 Rob
    August 22, 2008

    On a broader note, and I don’t want this to sound like a rant, doesn’t it seem odd that the mechanisms of SSRIs are being studied in the rodent hippocampus given that (a) SSRIs have already passed the FDA gamut (b) a new meta-analysis has revealed that SSRIs are no more effective than placebo and (c) it seems unclear whether neurodegeneration in the hippocampus is a cause or consequence of “depression” in rodents. I guess I don’t understand why time & resources are being invested into the question, “how do antidepressants work?” when their efficacy in human populations has already been shown to be minimal.

  7. #7 Jon
    August 23, 2008


    Pharmaceutical companies want to have the cake and steal ours, too.

    Just take a look at the anti-psychiatry movement.

  8. #8 ML Cohen
    August 23, 2008

    Rob (and #7):

    Meta-analyses of populations are certainly important, but soo too is the emerging field of “personalized medicine”.

    A 30-year non-blinded double-crossover study has demonstrated tremendous efficacy in at least one subject. In other words, you can have my Zoloft when you pry it out of my cold dead hands ;]

    All the best,

  9. #9 Jon
    August 23, 2008

    It does not matter whether it has worked for you or not #8. The fact remains that these drugs are dangerous and yet are marketed as the end-all to neurological/behavioral disabilities/maladaptations when they are merely a method of treatment. No one wants to take away the drugs from those for whom they actually work. However, that they do affect people harmfully (to the point of suicide) — well, a more serious vetting process needs to be put into action. It is so often the case that a doctor will simply give a drug to a patient because that doctor was put into the position to guarantee a certain quantity will be sold, a practice which has little to do with preventative care for the patients themselves. Indeed, I do not need to name a certain company that was recently found to have paid off doctors in order to give its drug’s records a better appeal. In the end, I will never trust pharmaceutical companies to do the right thing when profits are their driving interest.

    The Anti-psychiatry movement has its merits which must be taken seriously, regardless of your cold, dead hands.