Casting Doubt on Neurogenesis as an Explanation of Depression

This post over at Neuroskeptic reignites a debate -- if it ever really stopped -- as to the role of impaired adult neurogenesis in causing depression and the function of anti-depressants in stimulating neurogenesis to treat the disease.

This is one of those hot topics in neuroscience. If you look away for just a second the entire field changes, so I thought I would do a little update on where the field stands.

Short summary: Whether neurogenesis is at fault in the etiology of depression is still a very controversial idea among neuroscientists with mixed evidence. Many anti-depressants stimulate neurogenesis, but whether this why they work at treating the disease is not known.

First a little background:

The early theories about depression suggested that anti-depressant medications worked by increasing the concentration of certain neurotransmitters -- like serotonin, norepinephrine, and dopamine -- in synapses in the brain. However, these hypotheses didn't quite square with the observation that it takes about 2-3 weeks for anti-depressants to be clinically effective even though the increases in neurotransmitters happen right away.

So scientists went searching for other effects of anti-depressant medications, effects that more closely paralleled the time course of their clinical effectiveness. They happened upon one in stimulating neurogenesis. We know that animals generate new neurons throughout their lives in many regions, among them the hippocampus. We also know that this neurogenesis in the hippocampus is stimulated by a variety of anti-depressant medications. This observation incited researchers to generate an alternative neurogenesis theory of depression: depression is a disease of impaired adult neurogenesis and anti-depressants work by stimulating this cell growth.

So far, so good?

This theory has some good data to back it up. Most strong is a study where they irradiated the hippocampi -- to stop neurogenesis by killing new cells -- and showed that they could block the anti-depressant effects of some drugs in a mouse model of depression. This study had some detractors though, who said that you could not be certain of the effects of radiation and that the mouse model used wasn't a very good one. (More on mouse models of depression in a second.)

What has happened in recent years since that study is that researchers have gone scrambling looking for evidence for or against the neurogenesis hypothesis? This search for evidence has taken many forms. One of these has been to search for examples of where anti-depressants are effective, but neurogenesis is not present. In this vein, Neuroskeptic talks about Couillard-Despres et al. in Molecular Psychiatry. I won't repeat their excellent discussion, but to summarize:

The researchers, Couillard-Despres et. al. from the University of Regensburg in Germany, found that fluoxetine (Prozac) enhances hippocampal neurogenesis in mice - as expected - but found in addition that this only holds true in young mice. In middle-aged and older mice, there was no such effect.

The idea is that anti-depressant medications may be effective in older mice, but neurogenesis is not present. Hence that would throw a wrench in the neurogenesis and depression hypothesis.

The result above begs two immediate questions. First, are anti-depressant medications effective in elderly humans? The answer to that is, yes, about as effective as for everyone else.

Second, and more importantly, are anti-depressant effective in elderly mice? This gets into the more complicated issue of how one creates a mouse model for depression. I have talked about mouse models for depression at length before, so I will just summarize it here. Basically, we create mouse models of depression by using behavioral tasks that cause the mice to give up -- or to initiate what is called learned helplessness. It is impossible to know whether the mouse is "sad" in the human sense of the term, so rather than attempting to psychologize mouse behavior we look for models that 1) show similar negative symptoms to depression -- like low activity -- and 2) are amenable to correction by anti-depressant therapy.

There are a wide variety of behavioral paradigms that fit these criterion. For example, one of the most common is that forced swim test (FST). In the FST, you put the mouse in a beaker of water. For a while it will swim around and try and get out, but eventually it will stop swimming and float there. We say that we have achieved learned helplessness in that animal. We also know that anti-depressant medications significantly extend the time the animal will keep trying.

(As an aside: many of you may respond to the descriptions of these experiments with disgust. These treatments are indeed barbaric. We as scientists justify their use because depression is a horrible and life-destroying disease. Only through research can we hope to cure it. Further, many interventions that we would like to test could be dangerous in humans, and we need to confirm that they work in mice before we try them. So, yes, I recognize the barbarity of some of these tests, but this barbarity is justified by the legitimate goal of treating disease.)

Recognizing that anything we say about mouse models of depression is conditional -- mice are not humans -- are anti-depressants effective in older mice subjected to our behavioral models? The answer is maybe. I could only find two papers that had actually checked the subject, and both used 40 week old mice. That isn't particularly elderly for mice living in captivity -- the average lifespan is about 2 years -- but it is long enough that Couillard-Despres et al. in Molecular Psychiatry no longer saw significant increases in neurogenesis in mice of this age.

Both these papers used the FST, but they two produce somewhat contradictory results, one found that anti-depressants were effective in mice of this age while another found some types of anti-depressants (tricyclics and buproprion) were effective while others (SSRIs) were not.

Therefore, it is very difficult to interpret whether the observation that neurogenesis at a result of anti-depressant treatment decreases in older mice is a blow to the neurogenesis theory of depression. In order to say so, we would need clear evidence that anti-depressants work but neurogenesis doesn't happen.

Taking a step back, this is not to suggest that there are not serious problems with the neurogenesis theory of depression already. Like I said, this theory is still controversial. (I recommend these three reviews for a taste of the evidence.)

An interesting consensus idea is that stimulating hippocampal neurogenesis might explain how anti-depressants treat some aspects of depression but not others. Anti-depressant medications may initiate multiple processes of relevance to depression of which stimulating neurogenesis is only one. This would explain why some drugs cause neurogenesis and some do not; different drugs have differential effects on the depression phenotype. It would also explain why there are divergent results in animal models; different models focus on different aspects of the disorder and it matters which one you pick. This consensus view is not without evidence. (More here.)

If I had to summarize the field, in general, I would say it is still a mess. There have been so many claims and counter-claims that it is hard to make sense of what is right. This is, unfortunately, often how cutting edge science works. A field moves forward so fast that it is difficult to keep up if you are not a specialist. (I am not anywhere close to a specialist in this field, so if I have made a substantive error, please let me know.)

Just to quickly summarize the field and where it is going:

1) The theory that depression results from an impairment in hippocampal neurogenesis is probably an over-simpification. There are many examples, such as the one in aged mice discussed above, where neurogenesis and anti-depressant effectiveness do not proceed hand-in-hand.

2) Stimulating neurogenesis may account for the effectiveness of anti-depressants in treating some symptoms of depression, and in this regard some medications may be more effective than others.

3) Be skeptical of any over-arching claims made in this field. The facts are far from in.

Hat-tip: David Dobbs