One of the hardest parts about treating psychiatric illness is deciding which drug to use. Hopefully most doctors are not too swayed by the Paxil pens or Prozac magnets that end up in their offices (though that's a great way to get those names to the top of the recall list), but often, no matter what you start with, psychiatric prescribing tends to go through several iterations. Start with a drug, it doesn't work or has bad side effects. Start with another one. Increase the dose, decrease the dose, and so forth and so on, hoping the whole time for one dose or one drug that will produce the best effect with the least side effects.
A lot of research goes in to this, too. Surveys on what drugs work for which people. Which symptoms get better results from which drugs. While a lot of this is funded by drug companies, much of it isn't. And the final goal is always the same, finding the drug that will do the best (and thus, sell the best) in the most people for a particular symptom.
But how do you determine which patients are being helped by a drug, and in which ways? You can run lots of surveys and ask, but people's reports may vary, and their rating of their own symptoms is highly subjective. Right now, it's all we have to go on. But it'd be AWFULLY nice if we could evaluate, say, a patient's brain scan, a know from the activity in there how they would respond to a particular. And this paper, with that in mind, is looking at taking a step in that direction.
Brühl, Jäncke, and Herwig. "Differential modulation of emotion processing brain regions by noradrenergic and serotonergic antidepressants" Psychopharmacology, 2011.
Full disclosure: the authors were funded by a national grant from Switzerland. Sci is funded by nothing, but motivated by chocolate cupcakes and coffee.
The idea behind this was to look at which brain areas respond to different types of antidepressants. But how do you know which brain areas to look at? And what to look for? Luckily, there is some data on this. People with depression respond differently to things like pictures of negative emotional stimuli (like a sad face or other things that are upsetting, like photos of funerals, etc). Depressive patients, when they see these photos, show increased activity in brain regions like the thalamus, hippocampus, insula, the anterior cingulate cortex, and the ventral and dorsolateral prefrontal coxtex. We do know that while people are being treated effectively with antidepressants, their brain responses in these areas come back down to normal.
But the question now is how do different antidepressant drugs change signaling in the brain. Do they act differently? And could this help us determine which course of treatment is best?
To look at this, the authors took 21 healthy people (though they ended up with 16, some people just CANNOT sit still in a scanner) and put them in an MRI. They gave them either citalopram, a selective serotonin reuptake inhibitor (usually marketed as Celexa), reboxetine, a selective norepinephrine reuptake inhibitor (marketed in Europe as Prolift, among other things), or a placebo. In the case of both of the antidepressants, they work by inhibiting the recycling of chemical messengers in the synapses between neurons. Instead of getting taken back up into the neuron, the neurotransmitters hang around, increasing stimulation. Whether or not this is the main effect that causes the antidepressant effects of these drugs is very much up for debate, but many patients do find relief from their symptoms with these drugs.
In the scanner, they looked at a series of happy, neutral, or sad faces, while the scientists looked at their brain activation patterns. Not only that, they gave the participants signals followed by pleasant or unpleasant pictures, to let them know what kinds of things they were going to see, to allow them to anticipate the emotional impact.
This was only a single dose of drug, not the weeks long treatment that it takes to see clinical effects, so none of the participants really felt anything. But their brains did appear to respond differently to the drug, depending on what it was.
When the authors gave reboxetine, a norepinepherine related drug, they got a specific set of changes:
On the left are the two brain areas where they saw the most activity, the thalamus (a big egg shaped area in the middle of your brain that does a lot of signal modulating), the medial occipitotemporal gyrus, and the cerebellum (which has been becoming a larger player in the mood fields recently). In the center block, you can see measures of anticipation. This is the bit where the patients were shown signals that they were about to see a negative or positive picture. The reboxetine caused the patients brains (in the thalamus, occipitotemporal gyrus, and cortex) to respond more to the symbols preceding the pictures.
But when they used a serotonin reuptake inhibitor instead of a norepinephrine reuptake inhibitor, they got something else entirely.
You can see here that citalopram changed activity in entirely different areas of the brain. Instead of the thalamus or the cerebellum, citalopram changed activity compared to placebo in the inferior frontal gyrus and the ventrolateral prefrontal cortex. Not only that it changed how people responded to the stimulus before a picture...in the opposite direction in those areas, with activity decreasing.
The idea here is that, by looking at the different areas activated by specific kinds of antidepressants, we may be able to determine where changes with chronic antidepressant treatment are likely to occur. Further studies could maybe also identify which patients are likely to respond to which drugs.
So, you might wonder, why do two different antidepressants that are supposed to reduce the same symptoms have very clear differences in how they behave in the brain? Mechanism, dudes, mechanism. Reboxetine acts on norepinephrine transporters, while citalopram acts on serotonin transporters. These transporters tend to be located in areas where there is a lot of particular norepinephrine or serotonin activity. Thus, there are more norepinephrine transporters in areas like the thalamus and cerebellum, and more serotonin transporters in areas like the prefrontal cortex. And these brain areas all do different things, and so what looks like DECREASED activity somewhere may have a behavioral result that is similar to INCREASED activity in other places. The results with citalopram and fluoxetine could thus end up being equivalent behaviorally, even though they are acting in very different places.
Despite the possibilities for how people with depression may react differently to different antidepressants, there are a lot of gaps in this paper (hey, it's just the one experiment, but you know me, I always wish they'd taken it further). Frist, ummmmm...reboxetine doesn't work (though I'm sure no one knew that when they started this study). They did give a nod to the recent meta analysis showing no effects in the discussion of the paper (of course they have to), but they say that maybe they were just giving the reboxetine to the wrong patients (I dunno, dudes, that was a LOT of patients). But there ARE other SNRIs that DO still appear to work, like desipramine. They could have used that one as a control for reboxetine. But I'm not too upset about the reboxetine. We'll probably be seeing studies with reboxetine come out for the next year or so, as people catch wind of the meta study and finish the papers they were working on from back when reboxetine was supposed to be effective. This is clearly one of those papers, and was probably already in review or even accepted before the meta study came out.
No, the major issues to my mind were HOW the patients were dose, and WHO the patients were. First, the patients were given a single dose. we know now that the mood enhancing effects of antidepressants often take several weeks to show effects. So my question is...do the areas activated CHANGE in intensity or area over time? In what way? Are there differences in drug reaction in people on antidepressants vs those who are just beginning? Knowing the acture effects is fine, but seeing how they change over time is better.
Secondly, the patients were not depressed. They specifically recruited patients who didn't have any psychiatric illness. While this is good for seeing the pure effect of the drug, we don't give non-depressed people antidepressants (presumably). Perhaps depressed patients respond differently and in different brain areas. It's possible that you'd see a normalization of different activity almost immediately.
Finally, while they took the brain scans, I'd really like to see what the participants thought. How distressing did they find the pictures? Sure they didn't FEEL anything, but did it change their behavior at all? Other studies have looked at this a little, but I'd like to see the effects here.
Despite all this, I think studies like these may help in the long run. Help to determine where antidepressants begin to act, how they differ from each other, and how different people may respond. And it may help us in searching for the right drugs to give people, rather than going through the constant trial and error associated with current psychiatric treatment.
Brühl AB, Jäncke L, & Herwig U (2011). Differential modulation of emotion processing brain regions by noradrenergic and serotonergic antidepressants. Psychopharmacology PMID: 21359508