PsychEducation.org (home)                                                                                                                      How Mood Changes Affect the Brain (start)

 

Chapter 4:  Connecting Anxiety and Depression via the Serotonin Transporter Gene

or

Why Are There Shorts and Longs?
(revised/updated 7/2005)

Warning:  This is one of the more difficult chapters. Don't skip it just yet. But if you start to feel lost, move on to Chapter 5. You won't have missed anything essential for understanding the big picture of genes and molecules and depression.  This chapter tries to answer the question: Why are there short and long versions of the serotonin transporter gene?   

Summary:  As you have seen in Chapters 1-3, the serotonin transporter gene is somehow involved not just in depression, but also in anxiety and alcohol use. Recently a team of researchers at the National Institute of Mental Health (NIMH) showed how these different conditions might be related, and the serotonin transporter gene was right in the middle of their explanation.  Another gene involved in memory may also be part of this story. Finally, did you know that some monkeys leave home a year earlier than others?  Right -- the ones with short versions of the serotonin transporter gene.  This gene somehow affects social development. 

Acknowledgement:  Dr. Daniel Weinberger is the head of the NIMH team. He presented some of the following story at the American Psychiatric Association meeting in Atlanta, May 2005.  If you'd like to see his team's work without my "translation", see their June 2005 press release and one of the published reports, co-led by Lukas Pezawas and Andreas Meyer-Lindenberg.

Link to Chapter 5

 

  Chapter 4 Outline

 


Quick Review

May I remind you what the Serotonin Transporter is, and what the length difference is -- and then we'll look more closely at why some of the research that might explain why length appears to have such an effect on people's experience of the world.

First, recall that the Transporter is the "recycling device" which pulls serotonin back in from the space between neurons, the "synaptic cleft", for re-use.  Here's that same schematic drawing we saw earlier, describing  where the Transporter is and what it's doing: 

Secondly, recall that the gene for the Transporter (the serotonin transporter, or SERT gene) comes in two lengths, a long and short. Each version is called an allele -- nice simple language, huh?.   Remember that the difference between the two lies not in the information for making the Transporter itself (the "translated region") but in the portion of the gene that controls how often the translated region is read, namely the "promoter region": 

  Shutting Off Your Amygdala

What is the effect of a short SERT gene, compared to a long one?  Dr. Weinberger's team at the NIMH answers this way:  part of your brain that is supposed to decrease fear responses doesn't do this job very well.  And we already knew that prolonged negative emotions like fear can lead to depression. So it could be that this short SERT alleles lead to an increased risk of depression through less-well-restrained anxiety, fearfulness, and the resulting negative outlook on life (if life events, as shown in Chapter 1, crash down hard and repeatedly). Let's look at the results the NIMH has produced along these lines. 

First, remember that the amygdala is very directly involved in experiencing fear (see Chapter 2).  Dr. Weinberger's extended their earlier work on the amygdala and the SERT gene by examining the size of the amygdala in people with two longs, versus people with a short or two shorts.  Interestingly, they studied people without anxiety or depression (a.k.a. "normals").  They found a very strong result: people with one or two shorts had amygdala sizes 25% smaller than those with two longs. A picture of this finding is coming up, but requires a bit of explaining first.  

The following view of the brain may look a bit strange to you ("where are we?").  There is a Brain Tour which will make this view more understandable, written in plain English and making no assumptions about how much brain anatomy you already know. That Tour introduces the "cingulate cortex", the front curve of which is lit up like a rainbow in the picture below.  You'll want to know about the role of the cingulate in mood, as explained in that Tour, before you go on, so as to recognize the top/front/back as shown here.  The other structure you'll want to know about, also lit up in blue below, is the amygdala. If you are not already quite familiar with the amygdala from Chapter 2, take the Brain Tour of that structure also.   

                                                                    (top)

(front) (back)

                                                             (neck, so to speak) 

So, people with at least one short SERT gene end up with smaller amygdalae, and even more dramatically, smaller cingulate cortices (plural for cortex) as well.  So what?  This size difference led the NIMH team to examine the relationship between these two brain structures, using a new system for analysis of fMRI data that sheds light on how structures connect and influence one another (this technique, which you're likely to hear much more about in the next few years, refers to the study of what they call "functional connectivity"). 

By that analysis, there is a control loop for fear, outlined in this figure (same basic view as in the figure above, but more of a close-up): 

Or at least, that's what the loop looks like in "long/long" SERT gene people. The amygdala, shown as a yellow oval, sends signals to the bottom part of the cingulate, shown in red. In the current working theory developed from this and other research, a signal then travels within the cingulate, reaching the part just above, shown here in blue. That part of the cingulate is thought to decrease or somehow slow down the activity of the amygdala.  The result is that when the amygdala becomes active, this control loop damps it back down:  something like "calm down, calm down, it's just a garden spider, we've seen this before, this one is not a dangerous one." 

By comparison, the NIMH team found that in normal individuals who have at least one short SERT allele, the loop looks more like this: 

As you can see, the signal from the amygdala into the cingulate is not as strong.  The activity of the loop is not as strong either.  The result is that the amygdala is not being damped down as strongly as it is  in a person with two long alleles.  The result might be something like "calm down, it doesn't look like a dangerous spider, although I suppose...."  In the words of one of the investigators:  "This study suggests that the cingulate’s ability to put the brakes on a runaway-amygdala fear response depends upon the degree of connectivity in this circuit, which is influenced by the serotonin transporter gene.”  Got that?  If not, walk through that sentence again:  this is the "bottom line" conclusion from their research, their current working guess as to how the different gene lengths affect anxiety. 

Interesting theory. Does it actually work? Does it correctly predict what one might find in the real world? The NIMH team tested their theory by measuring "harm avoidance" in their study participants.  This is a very standard measure of people's inclination to stay out of danger. People who are "fearless" have low harm avoidance; and conversely, people who are "fraidy-cats", have high harm avoidance scores on this standard test.  Think about it: what should the NIMH team have found, when they looked at the harm avoidance scores of their "normal" participants? 

I know, you're thinking: "let's see, the short gene carriers (one or two S alleles) are less able to damp down their amygdalae, so they should have higher harm avoidance scores." That's what I would have thought. But they found something more subtle, yet very consistent with their working model.  They found a strong correlation between their "functional connectivity" measure and harm avoidance.  Huh?  Relax, it's just a garden research result...  Let's try that again.  

Remember, they already knew (as presented in Chapter 2) that the amygdalae of people with two shorts are more active when exposed to a fearful face than the amygdalae of people with two longs.  But now they are trying to correlate the gene-based difference in fear with a "real life" variable, namely the extent to which "normal" people prefer to avoid danger.  Here, the amygala activation in response to an angry face was not very strongly correlated to these people's scores on harm avoidance scales.  The researchers needed to look for something more directly related to this  "real life" fearfulness.  Lo and behold, the "functional connectivity" measure they had obtained was much more strongly related.  In other words, the difference between blue people and yellow people (between those with two shorts versus two longs), which one can see not just in depression vulnerability but also in fearfulness,  is not just in their amygdalae.  At the moment, it appears to be in their amygdala/cingulate interaction somehow.   

One more time, phrased a little differently, as I think this is a pretty tricky thing to understand... The strength of the very relationship they are examining in all this, between the amygdala and the cingulate -- that is what predicts harm avoidance scores.  This relationship is more strongly affected by the SERT gene length than is amygdala activity alone. In order to sort out the harm avoidant "normals" from the more courageous (or foolish, depending on how you look at it) "normals", the team needed a very strong signal coming from the SERT gene effect, if there was such a signal to be found. When they looked at how the two regions, the amygdala and cingulate, were interacting, that's when they found such a strong signal. This strengthened their belief in the model shown above. 


Are you really curious about these length differences in the SERT gene? Do you want see additional suggestive research? Here is a bit more, much of it reflecting the work of Dr. Klaus-Peter Lesch, now at the Psychobiology program at the University of Weurzburg in Germany; and some educated-guess thinking from Dr. Weinberger, head of the NIMH team working on all this. 

What does the Serotonin Transporter do? Why does it matter how long the promoter is? 

If you remember much of the story you learned once about how DNA encodes information, you'll remember that DNA is made of building blocks called nucleotides, and that there are 4 such blocks ("bases", typically labeled A, C, T, G).  Those four bases encode our genetic information.  But, they also determine how that genetic information is read, in the promoter region.  The length difference in the promoter region is due to a pair of bases:  GCGCGCGCGC..., called a "tandem repeat". The "long version" of the serotonin transporter gene is longer because of these.  Somehow this tandem repeat also makes the gene produce more protein product, which goes to work on grabbing serotonin from the synapse.  The result: cells with "two longs" take up twice as much serotonin from they synapse as cells with a short or two shorts.Lesch  

(If you're thinking ahead, you may be wondering if this isn't somehow backwards from what serotonergic antidepressants do, and that therefore this whole story is not holding together properly.  If you need to get that straightened out before you go on, skip down to the rather complicated explanation about this that appears at the bottom of this page).  

The GCGCGCGCGC... repeat is apparently a region where regulatory molecules can bind to the gene and influence how much of the serotonin transporter is made.  What regulatory molecules?  One answer is "glucocorticoids", such as cortisol.Glatz  You'll see more about cortisol as a stress hormone in Chapter 8.  We've known for a long time that cortisol was somehow involved in stress reactions, and mood changes.  We didn't know how.  This might be one of the connections.  

Dr. Glatz, and the research group from Germany under the direction of Klaus-Peter Lesch, showed in a series of experiments that the region that differs in people who have two shorts, versus the people with two longs, is precisely the region that is involved in regulation by cortisol.  Most of this work has been done in cells living in a Petri dish, so there is a ways to go yet to make sure this is part of the story in intact humans.  But a possible connection has been made: there is an explanation, or at least the beginnings of one, for how gene length differences might explain why different people have such different responses to the world, as shown in Chapter 1.  

Wondering About Why 

In the big picture of things, what might explain why there is a difference in gene length?  Why might evolution have lead to two different lengths? (You don't believe in evolution? Here's some brief help with that) What keeps both lengths around?  Here are two very small bits of information that may have nothing at all to do with the answer, but are intriguing in this respect.  First, getting two longs appears to make people more susceptible to a lung condition called pulmonary hypertension. Eddahibi   Certainly there must be something bad about getting two longs, since otherwise you'd think the two-long people would out-reproduce the two-short people (who would be removing themselves from the gene pool by committing suicide, as was shown by the research in Chapter 1).  

By the same logic, you'd think there must be something good about getting a short gene, or it would have been selected out of the gene pool a long time ago.  There is no good answer to this latter question.  The research results below just begins to hint at some of the potential function of the gene and these length differences.  

First, in monkeys who are running around in the wild, there's an advantage to getting out of the house early and moving on, for adolescents.  They get a chance at meeting new monkeys and getting on with the business of reproduction.  But they also may pay a big price, if they head out too early: they may end up at the bottom of the heap in their new group, with little access to females for reproduction, and little food to boot.  So, there's a risk in leaving early.  (I hope I haven't offended any monkey researchers with this gross summary; please correct me).  At any rate, here's the tantalizing result (two shorts, s/s; a short and a long, s/l; two longs, l/l):Trefilov

The "two shorts" monkeys leave "home" more than a year earlier than the "two longs".  There is some function to this difference in gene length, and it has something to do with social relationships.  This guess was suggested by some of the researchers who've been working on this, especially Dr. Lesch in Germany, who looked at other primates and found that only the most evolutionarily recent primates have the long allele (prosimians don't; catarrhini, including humans, do).  They speculate that the difference in allele function is somehow related to "the complexity of socialization" in recent primates.Lesch

Dr. Lesch and his colleagues, including Dr. Christine Barr, whose work we also saw in Chapter 3, have produced multiple articles on this topic.  They are looking the serotonin system and social success.  Low serotonin levels are associated with ending up lower on the ladder in monkey society.  They are also associated with violence.  Although monkeys with low serotonin levels are not more frequently aggressive, when they do become violent, it more frequently escalates to dangerous levels of fighting.Lesch  Isn't that interesting.  Mind you, these are brain serotonin levels we're talking about here; you can't measure these with a simple blood test (Dr. Lesch and colleagues surely wish you could, as it would make their research much easier!).  

Finally, in one of the most recent articles from Dr. Lesch' group, they show that people with "two shorts" have a bigger electrical event go off in their brain, in an emotion-related processing center of the frontal cortex, when they make an error.Fallgatter   They found that people with "two shorts" had a more powerful electrical signal of this type after they made an error in a computer game.  These errors were very minor, but had consequences:  participants were in a position to get more money for the experiment if they did not make many errors, and the test was designed to adjust itself so as to be difficult for everyone who played it (isn't that nasty? -- but they needed the same relative degree of difficulty to be able to evaluate the results, so it wasn't just that they wanted to keep their pay-out down!).  

The implication seemed to be that people with "two shorts" might be more negative in their evaluations of events, or perhaps more importantly, their own personal role in events (that's more my interpretation than the researchers';  they are more cautious and stop short of trying to dig that much meaning out of their result).  For now, we can say that the Serotonin Transporter gene, and its length differences, appear to be involved in some sort of social event analysis or evaluation.  Beyond that, we can await with great interest further results in this very active area of research.   Dr. Lesch does go so far as to wonder out loud about how these transporter differences, and the way their influence on adults seems to be so strongly affected by childhood, might affect the way we humans deal with one another on the world stage.Lesch (see the Conclusion, page 17 of 24 on your scroll bar).   How appropriate for a European community member to lead the way with such wondering. 

(Update 2011: one other detail: having a short gene is also associated with seasonal shifts in mood and behavior.Sher

Interactions With Other Genes

Now you're really entering some complex territory.  If you're lost now, bail out to Chapter 5.  But if you're still wanting to know everything you can get your hands on about genes and mood and anxiety, read on. 

Dr. Weinberger proudly presented the work of his NIMH team at the American Psychiatric Association meeting in May 2000.  He tried to tie all their work together into a way of thinking that might help guide us as research proceeds on the genetics behind mood and anxiety disorders.  

To explain their thinking, he explained that there is another gene being studied in relationship to mood and anxiety, the gene for BDNF -- Brain-Derived Neurotrophic Factor. You'll see much more about BDNF in Chapter 9.  It is directly involved in the molecular chain leading to depression, and back out again via effective treatments, all of which appear to increase BDNF -- including antidepressants, exercise, even electroconvulsive therapy.

There are two versions of the BDNF gene.  One allele has the amino acid valine in a particular part of the BDNF protein chain.  The other allele has a different amino acid in that spot, methionine.  If both of your parents gave you a valine version, you'd be a "val/val" person.  If one gave you the valine-version, but your other parent gave you a methionine-version of BDNF, you'd be "val/met."  You can tell what your parents were doing if you're met/met.  Okay, here's the point of all this: val/val people have better memories.  Quite a bit better, in fact.  

But, there's price for that.  Val/val people are also more "neurotic."  That term is rather ancient in psychiatry, and rather vague. In this case, it means that people with val/val alleles score higher on a very standard test of "neuroticism" that's been around a long time (as reflected in the terms used). People with high "neuroticism" scores on this test have more "negative emotionality" than those with lower scores, emotions such as anxiety, low mood, and hostility (I'm quoting from the test description there).  This val/val effect was shown in a study by a team from the University of Michigan (Sen, Nesse et al), as shown in the graph below. 

           

Looks pretty impressive, the differences -- but overall this is a pretty small effect. You don't need to go get yourself a NEO-PI test right away, or get this particular gene tested (as we'll consider in the next chapter).  However, there does appear to be something important going on here. 

According to Dr. Sen and colleagues, the Val allele is the common one and the Met one doesn't appear in evolutionary history until recently; might this imply that the met version serves some function necessary in human brains, but not earlier in evolutionary history (need the evolution essay yet?) That's not the only potential explanation, but tempting. It appears to have tempted Weinberger and colleagues, as follows. 

Weinberger thinks that perhaps met/met BDNF "doesn't listen well" to serotonin-based information about threat.  He thinks it might be a sort of "deaf ear" to the higher anxiety signal that result from short SERT alleles (as described above).  By contrast, val/val appears to possibly "exaggerate" the short- SERT gene anxiety effect. 

In this way of thinking, if you get s/s SERT genes, but also get met/met BDNF genes, you're less vulnerable to the "increased anxiety" effect of s/s SERT; whereas if you get val/val BDNF (superhearing) with short/short SERT genes, then vulnerability to anxiety and depression could be much higher.  

Here is Dr. Weinberger's analogy to help imagine how these systems might interact.  Imagine that val/val BDNF alleles, with their memory-improving capacity, make your brain function like a 200 mile-per-hour race car.  If you've got a hot rig like that, you'd better a good driver who's capable of handling a fast, but temperamental car.  In this analogy, that's the long/long allele pair for the SERT gene: the driver won't get over-anxious and allow the car to get out of control.  That's important, because if you smash your car into the wall very often, your car won't run very well.  In real life, if you take too many stress-hits, you end up depressed.  

By comparison, if you inherit the short/short pair for SERT, and thus are less able to handle anxiety-producing situations such as conflict, trauma, and loss -- you are a more cautious and potentially distractible, frighten-able driver.  In this case, you might be better off with a slower but more crash-resistant car, one that you can smash up against the wall quite a few times without changing how it performs very much. In this analogy, that's the met/met allele pair for the BDNF gene.  

By this analogy, perhaps the "slower" met allele is was selected for (evolution-speak) in humans to help people with "two shorts" get through life better.  If two shorts makes you more cautious, and two met's makes you less likely to worry about things, for some people that could make a durable, reliable combination that in dangerous times might be better than the high-performance but "higher-maintenance" val/val and long/long combination. Of course at this point that's almost entirely a guess, but it gives us a beginning of a model which might help understand these genetic variations in humans. 

Here's a link to Chapter 5:  "Hey, where do I get my genes tested?

If you're still curious and would like to see everything I've run into so far on the short/short, long/long story of the SERT gene -- everything that might have some usefulness to me as a practicing psychiatrist anyway -- read on for a few more interesting details. 


Antidepressant responses 

You might think that "two shorts" people, because they obviously have some sort of serotonin problem, would be the ones who would most benefit from serotonergic antidepressants.  But, it's definitely not going to be that simple.  In fact, the results are rather the opposite: 

Percentage of patients with side effects
from Serotonergic antidepressants  (SRI
's

 

s/s

s/l or l/l

 Insomnia

78

22

 Agitation

67

7

As you can see, people with two short alleles (s/s people) have many more of the over-energized side effects from SRI's (sorry, forgot to link this study when I wrote this, and have since lost track of the reference; but see below, the story thickens).  These are the very side effects the U.S. Food and Drug Administration (FDA) recently connected to an increased risk of suicide in patients taking SRI's.FDA 

 


Serotonin Transporter Gene Length and Bipolar Disorder

Update 10/2006: 
Is the short/short pair associated with antidepressants causing mania, in people with bipolar disorder? Two studies suggest this is true, but there are two which do not show this relationship (the middle two in the table below). 

Lead Author

Year

Results

Mundo 

2001

*(see slide below)
 

2003

Short/short associated with rapid cycling but not mania
Serretti 
 

2004

Large sample, no differences seen
Masoliver 
 

2006

*(see slide below)

Remember that it's hard to do science. One can miss a "real" finding just on chance, sometimes. So here we have three studies (Mundo, Masoliver, and we'll count Rousseva as antidepressant-induced worsening) which suggest that having the short/short pair is not good if you have bipolar disorder.  Looks like that pair makes a person more susceptible to having a bad reaction to an antidepressant. Knowing your gene type might tip the odds quite a bit, as in the bar graph above:  if you have two shorts, your chance of having mania if given an antidepressant is 60%. That's much higher than the general estimates of this risk, which are usually no higher than 30-40%, and are often much lower (4%, as in one analysis, is almost certainly far too low; references in Antidepressant Controversies). But, knowing you have two shorts still leaves a 30% chance of not having such a reaction. And knowing you have two longs leaves a 40% chance of having the reaction.  So knowing what your gene pair is won't help you decide whether to take an antidepressant or not, not at this point. 

A few more tidbits on the relationship of allele length and bipolar disorder:  

Other serotonin transporter gene research

The short Transporter gene has been associated with failing to respond to an antidepressantArias, Kim,  and others , but there are similar studies which did not find this difference.  Several studies have found that inheriting the long/long genes is associated with a more rapid response to serotonergic antidepressants.Pollock, Durham.  The interaction of these genes and mood is complex.  One study even found that the long/long people responded better to placebo!Rausch  

Similarly contradictory results have been found for "tryptophan depletion".  Drinking an amino-acid mixture that has been specially prepared to lack tryptophan, the amino acid from which serotonin is made, causes people's serotonin levels to drop very quickly and dramatically.  In several previous studies, this nasty drink has been used to show that people with a history of depression are more susceptible to becoming depressed again when they drink this stuff.  If this serotonin transporter story was simple, you might think that inheriting two short alleles would make people more likely to have depressive symptoms when they drink this mixture.  Ineed, in one study, people with "two shorts" were more susceptible than people with two longs.Neumeister  But in another study, it was the long/longs who were more susceptible, if I am reading their results correctly.Moreno 

Among suicide victims -- not those who tried, but those who tried and actually died -- the long/long allele pair was twice as common as the short/short pattern.Du  (Perhaps short alleles make a person more susceptible to depression, but if she/he has the protection of two longs and still gets depressed enough to consider suicide, then life may be so bad that she/he is more likely to judge it worth leaving?  A reader who thinks she or he is s/s and so knows a thing or two about this, suggests that perhaps s/s people are used to being depressed and/or handling things poorly, so they can better cope with suicidal thoughts; whereas, the l/l person who finally gets depressed is probably so unfamiliar with feeling depressed and suicidal that he or she may be more affected by the depression, and might even attempt suicide as a result. These are just working guesses on how these results might be interpreted.)  

Another group found no association between suicide and the short/short allele inheritance, but instead found an association with a different gene, for tryptophan hydroxylase (the enzyme which can govern just how much serotonin one makes from tryptophan, the amino acid from which serotonin derives)Abbar; but another group did not find a tryptophan hydroxylase effect on suicide.Kunugi   

On the whole, the data seems to be coming together more for the serotonin transporter story than for other gene research, though this is an outsider/beginner's assessment.  Thanks for a broad literature search, from which most of these citations derive, by Shawn at http://www.neurotransmitter.net/ .

For another source on this topic, there is a literature review at the National Library of Medicine complete with summaries: http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=182138

Link to Chapter 5


Review by Dr. Lesch.

For the most complete story connecting the Serotonin Transporter to human behavior, there's a technical summary by the master himself, Dr. Klaus-Peter Lesch, from 2000.  It is an old document now, but you can see how far he was thinking even then.  It would be wonderful to see a more recent such summary, but as far as I know, none is available, certainly not online.  Much of the material cited here appears in his chapter in Molecular Genetics and the Human PersonalityHere are the three sections referenced above in the Review by Dr. Lesch

  1. Serotonin and violence:  page 596 to the Conclusion (page 16 on your scroll bar, starting with the very last sentence); 
  2. Speculation about human behavior relative to the gene length differences starts on page 590 in the document itself ( page 10 of 24 on your scroll bar);
  3. The apparent contradiction between higher Transporter activity in "two longs" people, versus the fact that serotonergic antidepressants (SRI's) lower activity of the transporter (page 13 on your scroll bar, page 593 in the document itself).  Here Dr. Lesch emphasizes the role of development with these gene differences in place:  in other words, it appears not to be the absolute amount of Transporter activity that matters, but perhaps what the Transporters were doing as the monkey or human was growing up, in terms of the effect of these gene length differences on adult behavior. 

Link to Chapter 5