Genes, depression and schizophrenia

Kat -   We'll be returning to the hunt for depression-related genes later but now we're switching to a different mental health condition - schizophrenia. Affecting around one in a hundred people in the UK, schizophrenia can be extremely debilitating, as well as distressing for the family and friends of sufferers and a wider social issue too.  Professor Mike Owen heads up the MRC Centre for Psychiatric Genetics and Genomics in Cardiff, and I asked him what we know so far about the genetic links to the disease.

Mike -   So, we know that for example that if you have a first degree relative, a close relative with the disorder, say, a parent or a sibling, then your risk of getting the disorder is about 10 times increased.  So, it's about one per cent generally in the population - a risk of getting schizophrenia.  If you have a close relative, it's about 10 per cent.  And we know that that could be shared genes, but it could be shared environments as well.  It could be people living in the same sort of settings have similar experiences.  So, you can do twin studies, adoption studies that show that that clustering in families, if you like, is largely due to shared genes.  But there's a lot of evidence from those sorts of studies.  More recently, now, there's many of us have been doing molecular genetics - looking at the DNA of people to try and identify the specific genes involved.

Kat -   From your talk, one of the key problems was that schizophrenia isn't just one condition.  It seems to be a wide range of conditions.  How are some of the genetic studies starting to shed light on this complexity?

Mike -   Well, the problem I guess is that what we call schizophrenia is a syndrome.  It's a set of symptoms that occur together and it's a bit like trying to do genetics of shortness of breath, I guess, or of chest pain.  Chest pain could be due to a heart attack or it could be due to stomach problems or muscular problems, and so forth.  So yes, probably, it is heterogeneous.  It contains various conditions.  The other problem we have is that it's not nicely circumscribed.  There are related disorders - bipolar disorder for example, which we hear a lot about in the press - kind of merges into it symptomatically.  There's no kind of clear blue water between the two. 

So, what the genetics is telling us is that genes for psychiatric disorders are not specific.  The ones we've discovered so far are not generally speaking specific to one disorder.  They're predisposed to multiple disorders, suggesting that these diagnostic categories that we use don't carve nature at the joints, if you like.  There are reasonable sort of surface description of more complicated things going on lower down.

Kat -   So, in terms of the genes that have come out on some of these studies, perhaps your top suspects, what do they tell us about some of the underlying biology of what's going on?

Mike -   Well, the striking thing has been the degree of convergence onto sets of genes. Genes code for proteins, they contain the instructions for proteins.  And so, what we can do is we can look at the proteins - the genes we found are coding for and see whether they cluster in particular functional groups.  One of our most striking finding today is that there is a clustering of proteins that are involved in synapses. 

Synapses are the sort of structures that allow nerve cells to communicate with each other.  They're also the structures that basically confer the brain's ability to code information.  So, the strength of synapses changes with experience and that codes information, if you like, and that plasticity is the basis for learning and memory, and helps the brain develop.  So, we've identified genes that focus down on a small set of proteins that are involved in this synaptic plasticity.

Kat -   And of course, what people with mental health problems, with schizophrenia, and their families will want to know is about, how will this knowledge help to treat the condition better.  Where are we with coming up with more effective treatments.  They don't seem to have been any breakthroughs for decades really.

Mike -   No, there haven't really been any breakthroughs for 40 odd years.  This work is very much aimed at trying to understand what is going wrong in the brains of people with severe psychiatric disorders and in the hope that we can identify new treatment targets.  Now I have to say, this work is going to take a long time.  The next step is to start to understand the function of some of these genes in the brain and we're starting to do that - well, in three ways.

Looking in people, we can look at people who carry specific genetic factors and we can study their brains with brain imaging methods.  We can use animal models.  We can manipulate the genes and model the risk genes in animals and study their brains and behaviour.  A really exciting new technology is to use stem cells.  We can take cells from people, skin cells and convert them into nerve cells and grow them in a dish and study their developments and the way they relate to other neurons, etc. 

When you identify risk genes for disorders like this, the next phase is intensive kind of neuroscience and that's what's going to happen over the next five, 10, 15 years.  I mean, we have to go this way.  It's the only way I think to get effective treatments.  It's very hard to predict how long anything in science will take.  It could be a breakthrough next week, but I don't want to give false hope either, but there's nothing around the corner, but you know, this is solid progress.  There's solid progress and I think we're entering a new phase of the game now.

Kat - That was Professor Mike Owen from the MRC Centre for Psychiatric Genetics and Genomics in Cardiff. And now it's time for a roundup of this month's genetics news.

Writing in the New England Journal of Medicine, scientists from the University of California say they have new evidence that autism begins in the womb. The conclusions come from studying brain tissue from 22 children with and without autism that had died in accidents or from unrelated diseases, aged between 2 and 15 years. The scientists discovered key differences in the outer layer of the brain, called the cortex, in ten out of 11 children with autism, compared to one in 11 children without the condition.

The changes were dotted about in brain regions involved in social and emotional communication, and language, and are likely to have arisen during the second or third trimester of pregnancy. Importantly, the researchers think that their findings might explain why some toddlers with autism show signs of improvement if treated early enough, as younger brains are more likely to rewire themselves. So the race is on to find out whether these changes hold up in further studies, and how the information might be used to inform earlier identification of autistic children so appropriate interventions can be made as early as possible.

Kat - Continuing our reports from the Genetics Society Spring meeting, focusing on the genetics underlying psychiatric disorders, it's time to hear from the winner of the annual Genetics Society's Medal, Professor Jonathan Flint from the University of Oxford. At the meeting he presented fascinating early results from his team's hunt for genes involved in major depression, and told me more about the study.

Jonathan -   So we thought, along with other people, that the most critical thing to do is to try and reduce heterogeneity because there are many different ways that you could become depressed - many different causes, both environmental and genetic.  Unless you try and reduce heterogeneity then your analysis is either going have to be so large as probably to be unwieldy and we felt also that it's very difficult to understand what the interplay and interacting factors might be.  So, the study we designed was to hone in on a group where we thought that the causes would be as homogeneous as possible. 

So for example, we know that the genetic effects are different in men and women and we know that women have higher rates of depression than men.  So, it seems sensible just to pick one sex and we chose women.  And we know that recurrent depression is more likely to be heritable than non-recurrent depression, so we choose heritable depression.  And we know the causes are slightly different if your onset is very, very early or later in life.  So, we restrict the age categories at which the depression occurs.

Kat -   And as well with your study, you didn't carry this out in the UK.  Tell me about where you did it.

Jonathan -   No, we didn't carry it out in UK.  So, because we have to go through all of these restrictions, although the disorder is common - so if you just took one episode of depression as a measure of the occurrence in a lifetime -  the prevalence rate is something between 10 to 15 per cent.  So, it's pretty common.  But because we've restricted it to recurrent and only one sex, and a particular age group, and they have to have certain other features, that made it rarer. And we also require interviewers to spend a long time with their patients.  So, the number of people who were prepared to give up that time is of course, even less. 

We wanted it to be genetically homogeneous because we're very interested in the genetic causes.  All of those features rather restricted where we could do the study.  In the end, we decided to try China because they're obviously an enormous population and it has a very good healthcare system, and the doctors are very well-trained.  We can relatively restrict genetic homogeneity by only going for Han Chinese, which is what we did.

Kat -   Tell me about the size of the study.  How many women were you looking at?

Jonathan -   We decided to choose 6,000 cases and 6,000 controls as our target.  We made that decision about 2006, 2007, so about seven years ago now.  On the small side in fact for what we now know. And ideally, we'd like probably 10 times that number.  You can argue about this but we thought by restricting it, we'd have a chance of finding something.  By May of 2012, we had indeed reached our sample of 6,000 cases and 6,000 controls.

Kat -   So, as well as interviewing these women and getting the doctors to talk to them and find out, you were also looking at their genes as well.

Jonathan -   We sequenced everybody.  We have their genome sequences in a large database now, yes.

Kat -   And so, can you tell me about some of the things that you've started to find?  What seems to be coming out from this data?

Jonathan -   Well, so far, but this is a little preliminary, we've discovered - we think - one locus in the genome that's more likely to make people depressed.

Kat -   So, that's one kind of genetic area, one gene.

Jonathan -   Yes and it's clearly not the only one.  This is a very small effect and everything tells us, including our own data, that it's highly polygenic.  But there's only one which reaches accepted statistical significance and we've yet to replicate that.  But it's an intriguing observation because it lies over a gene called Sirtuin-1.  That's intriguing because that gene has been implicated in the a whole series of phenotypes, many of which are indeed related to the phenotypes of depression.  So, for example, depressed people typically lose their appetite.  They don't sleep very well.  They become a little sluggish.  And knockouts with that gene also manifest many of those phenotypes.  And perhaps most importantly, the Sirtuin-1 locus has been regarded as at least a moderator of metabolic phenotypes.  One way of conceptualising depression is that there's a certain form of it called melancholia, in which by these so-called vegetative symptoms predominate.

Kat -   So, these sound like quite intriguing links.  What next for this study?  How do you go about trying to work out if this is actually true?

Jonathan -   So, the thing we'd like to do most of all is replicate and we've been scouring the world's genetics groups to see if they've got a similar sample to ours.  The problem has been, because we've been so restricted in what we collect, no one has got a similar sample.  So, we have a little problem there.  The alternative is for us to collect another sample, so we're trying to do that as well.  But the other thing that we can do is we can use our sequence data in rather interesting ways. 

So, when people normally carry out these studies, they interrogate the genome at maybe one million positions and these are typically those which vary with frequencies of greater than five per cent.  But we've got sequence data and we can almost interrogate everything.  Not quite because of the way we've done it.  So, that allows us to look at the genes we think might be involved from the genome-wide association study and look for other signals.  So, one of the signals that we look for is whether there might be small deletions, mutations in the genes that we think are involved and now, and are they enriched in cases or controls.  There is a little hint that indeed might be so.  So, that provides an independent signal, because it's not incorporated in the genome-wide association, that what we found is real.

Kat -   It seems to be that with some diseases such schizophrenia, there's a lot of genetic data coming out with autism as well.  Depression seems a bit of an impenetrable box.  Where do you think this is going to be heading and do you hope to find a hatful of genes, or how do you see the field progressing?

Jonathan -   We could, as I think most people would agree, go ahead by collecting a larger sample.  There's little doubt that if you collect a big enough sample, you'll find more things.  There's a quite interesting linear correlation with the number of loci you find in the sample size.  So in a simple way, if you double your sample size, you'll double the number of loci that you find.  So, if we find with 6,000, one locus, we go up to 60,000, we'll find 10.  So, that's one simple approach one could take.  One could refine the phenotype further and the information we're collecting already tells us that's an important way to go.

Kat -   So, just picking people with very, very specific symptoms.

Jonathan -   Yes.  So maybe, for example, what the locus we found is specific for highly recurrent depression and for a subtype of depression called melancholia which is, if you like, some sort of measure the severity.  And it might be possible to subdivide it further.  Once we know what type of depression shows these features, shows the genetic linkage, the genetic association, then we could go ahead and specifically look for those.  That's one thing that we could do.

Kat -   And more broadly, with the kind of studies that we've been hearing about in the meeting, looking more in-depth at the genetics, the underlying causes of some of these mental disorders, where do you think perhaps in five or 10 years we might be with understanding some of these really complex and very challenging, and in many cases, distressing illnesses?

Jonathan -   Well, I think we have to take the genetic findings for what they are - namely, that it's unbiased approach to the underlying biology.  And while that's good in a sense that, for most psychiatric conditions, probably for all, we don't really understand what the causes are.  So, any clue is good.  On the other hand, it means that there's an awful lot of work in turning a genetic signal into what we all want which is mechanism and possibly later some way of treating that condition.

Kat -   That was Professor Jonathan Flint from the University of Oxford.