In this podcast we talk to Professor Cathryn Lewis, Professor of Genetic Epidemiology & Statistics, Head of Department, Social, Genetic & Developmental Psychiatry Centre, King’s College London.
Cathryn discusses the work of her research group and how determining the polygenic component of mental health disorders can be accurately measured, and how to use genetics to assess people’s risk of mental disorder.
Cathryn also explains how are polygenic risk scores helpful for child and adolescent mental health professionals, and why should they take an interest in this, and how to translate research into clinical use.
Cathryn Lewis is Professor of Genetic Epidemiology & Statistics at King’s College London, where she leads the Statistical Genetics Unit. Her academic training is in mathematics and statistics, and she has been involved in genetic studies since her PhD. She co-chairs the Psychiatric Genomics Consortium Major Depressive Disorder Working group and leads the NIHR Maudsley BRC Biomarkers and Genomics theme. Her multi-disciplinary research group identifies and characterises genetic variants conferring risk of disease, including depression, schizophrenia, and stroke. A major focus is a risk assessment, determining how the polygenic component of mental health disorders can be measured accurately and communicated effectively. Bio and photo via KCL Follow Cathryn on Twitter @cathrynlewis
Interviewer: Hello, welcome to the In Conversation podcast series for the Association for Child & Adolescent Mental Health, or ACAMH, for short. I’m Jo Carlowe, a freelance journalist with a specialism in psychology. Today I’m interviewing Cathryn Lewis, Professor of Genetic Epidemiology & Statistics at King’s College London and Head of Department at the Social, Genetic and Developmental Psychiatry Centre. Cathryn’s research group identifies, and characterises, genetic variants conferring risk of disease, including depression. If you’re a fan of our In Conversation series, please subscribe on iTunes, or your preferred streaming platform. Let us know how we did with a rating, or review, and do share with friends, and colleagues. Cathryn, welcome. Thank you for joining me. Can you start by introducing yourself?
Professor Cathryn Lewis: Hi Jo. It’s a real pleasure to join you and the Association for Child & Adolescent Mental Health, here today. I am Cathryn Lewis, Professor of Genetic Epidemiology & Statistics at King’s College, London. And I’m Head of Department at the Social, Genetic & Developmental Psychiatry Centre. And that’s an interdisciplinary research centre. We investigate the interplay of genetics, and environment, of nature, and nurture, across all aspects of psychiatry, social, genetic, and developmental.
Interviewer: Cathryn, your academic training is in mathematics and statistics. So how did you come to develop an interest in mental health?
Professor Cathryn Lewis: That’s a good question, and it was really pure serendipity, like a lot of things in my working life. I never had a long term game plan. I certainly never planned to be an academic. But, at each stage, I’ve just taken decisions on my career based on what the most interesting science and the people that I enjoyed working with. As you said, my undergraduate degree was in mathematics, but I found pure maths too abstract, too focused, and, to be honest, too difficult.
And so, I looked for something more applied, that might suit me better. I did a Master’s degree in statistics, to explore that area, and, slightly to my surprise, I loved it.
I loved how statistics gives you a set of tools to explore data from real life problems, and gives you real insights into what’s happening, whether that’s exploring trends, across time, or a definitive answer like drug A being better than drug B. And I think in the pandemic, we’ve all seen the value of collecting data, of analysing it properly, and understanding it better, to show what we know, and, crucially, what we don’t know. So I got hooked, and then I studied for a PhD in statistics, and started applying my statistical tools to analyse genetic data.
And genetics has been a really exciting field to be in over the last few decades, because the technology has developed really fast, and we’re constantly expanding our tools to be able to investigate genetics of disease and health.
And, of course, with each new technological tool, with each new laboratory machine, we need new statistical methods to analyse the data that are generated. So life has never stood still, academically. I started working on cancer genetics, and then immune mediated traits, applying my skills to each of those. And contributing, for example, to the identification of BRCA1 gene, for breast cancer. But to come back to mental health, that was pure chance that got me into this field.
I developed a statistical method for pooling genetic linkage studies. And a US psychiatrist, Doug Levinson, tracked me down, at a conference, and asked if I wanted to apply my method to schizophrenia. And, so often happens, in academia, one collaboration, one project, leads to another. And I’m now fully embedded in mental health.
One of the things I like best about psychiatry, and mental health, in general, is that it’s a collaborative and collegiate environment that’s often present.
And my statistical tools provide one aspect of research, but I can’t do anything alone. And all my projects are in collaboration with psychiatrists, psychologists, geneticist [inaudible 00:05:02], computer scientists. And research is so multifaceted that working together gives you an added breadth, and depth, that you can’t get with a small group.
As you work together, you start to understand everyone else, strengths, their priorities, their tools, and, suddenly, when you get a good group of collaborators, the whole becomes greater than the sum of the parts.
And, for me, that’s where the magic of research happens, and, together, you really achieve something you can’t do alone. And, you know, those days where you close down the computer, at the end of the working day, knowing, interesting, or something [s.l. riveting 00:05:47], that you you hadn’t known that morning, that’s the magic of research, for me.
Interviewer: So let’s hear more about your research group? What are the main areas that you work on?
Professor Cathryn Lewis: I lead the Statistical Genetics Unit at King’s College, London. And so, we analyse genetic data to identify the genetic variants, that contribute to psychiatric disorders. We characterise how they work, how they lead to disease, and what their impact is on risk, on prognosis, and on treatment.
Depression is the major focus of my work, and I lead the Depression Working Group in the Psychiatric Genomics Consortium, which is the big international group that leads discovery of the genetic component of depression, both adult, and adolescent, onset.
Interviewer: Cathryn, a major focus of your work is risk assessment, determining how the polygenic component of mental health disorders can be accurately measured. Can you say more about this component of your work?
Professor Cathryn Lewis: So I often get asked why we might want to know the genetics that contribute to mental disorders, particularly when these traits are highly polygenic. And so, there’s no single gene that we can focus on, in assessing risk. And there’s two reasons why I’m really excited about my work, to uncover the genetics that underpin depression, and other disorders.
So firstly, we have no biology for mental health traits. There’s no blood test, or biomarker, that we can use, in diagnosis. We don’t know why, or how, genetics, and the environment, combine, and interact, to put someone at high risk. But genetics can take us closer to that biology. For example, in our recent genetic study of depression, we identified 100 variants that were associated with depression. We also showed that these variants were [inaudible 00:07:54] highly enriched in genes that were expressed in the brain, but not in other tissues.
And that’s reassuring and matched expectation. And then if you drill a bit further down, you find [inaudible 00:08:05] are particularly highlighted in the prefrontal cortex, and the anterior cingulate cortex, which are important for higher level executive function, and emotional regulation, which are often impaired, in depression. So, again, that’s reassuring that our genetic studies are matching what we know from imaging studies, and such like.
And finally, our depression genetic findings connect us to genes that are expressed in neurons, but not oligocytes, or astrocytes.
Again, that takes us closer to that underlying biology, and confirms the validity of the genetic variants that we are finding, for depression.
And the idea is that we could take those, and they could help us identify potential new drug targets, for treating depression. And we know that we desperately need new perspectives on depression, effectively. So whether expanding our options, for new drugs, or highlighting repositioning for current drugs that we have, for different disorders, genetics can give a handle on those. The second reason I’m excited about finding the genetic predisposition to depression is that each genetic variant that we identify gives us new information, that slightly modifies someone’s risk of developing the disorder.
So we’ve identified over 100 variants associated with depression, and we have, in the pipeline, a really exciting study that will substantially expand that. We know that 100 variants is the tip of the iceberg for the genetics of depression. And although each of those variants plays a very small risk in depression, there are many hundreds more, probably thousands, to be identified, before we can come close to the full understanding of the genetic underpinnings of depression. And some of those variants are relevant, not only to depression, but across mental health disorders, particularly for bipolar or schizophrenia.
Interviewer: Given the current understanding that you have, how do you use genetics to assess people’s risk of mental disorder?
Professor Cathryn Lewis: We know very little about why any particular person would develop a mental disorder, or, equally, while others are protected, given they may live in the same high risk environment. But genetics gives us the first insight into this. So for those 100 variants that, we’ve identified, already, individually, each of them only has a tiny impact, and moves the needle of risk minisculely to the left, or the right, towards an increased, or decreased, risk.
But cumulatively, those variants can give us valuable information, and we can combine our knowledge of the genetic predisposition across the genome, across all the chromosomes, to calculate a single risk measure, that combines all the information together.
And that’s called variously a genetic risk score, a genetic profile. I’m going to use the term polygenic risk score, because those three terms capture three important components, that mental disorders are highly polygenic, and that the number that we calculate gives us some information, on an individual’s level of risk. And it’s a score. It’s a single number that captures someone’s genetic liability to disorder. And that’s an advance we’ve had over the last five years really, we’ve never previously been able to do that.
If we calculate these polygenic risk scores, in a group of people, then their scores form a normal distribution, a bell shaped curve. And that means that, for most people, their genetic risk of depression, or any other disorder, is around average. And so, genetics is not going to tell them very much.
But some people will be at the lower end of that distribution, as they carry, by chance, fewer of the risk variants. And others, of course, those that we’re much more likely to be interested in, are at the top end of that distribution. They carry more risk variants and are at the highest risk of depression. So that’s the theory behind it, and it works very well. But, currently, our polygenic risk scores don’t give us much information. They don’t give us enough information to be able to move them into clinical care yet.
To give you an example, in our depression polygenic risk scores, if you look at the 10% of people that have the highest scores, they’re two and a half times more likely to be diagnosed with depression than the people at the bottom end of the distribution, with the lowest 10% of scores. So that level of information isn’t enough to use clinically. And, I suppose, I also want to point out that these scores are exactly that, they’re a score. They’re not a diagnostic tool, but they could be a stratifier of risk, for example, to identify those people who are at the highest risk of disorder, and we might want to to follow up on or use that information in some way.
Interviewer: Well let’s look a bit more at that, then? So we know that polygenic risk scores have been widely applied in research studies. They help further understanding of the longitudinal course of mental health problems. But how does one translate the research, into some, kind of, clinical use, in particular, then, if you find if somebody has a risk score in the top 10%? Is there anything that can be done for them? How do you apply it to patients?
Professor Cathryn Lewis: Really good question. So, as you said, we’ve got a lot of experience of using these in research studies. And we can show, for example, that study participants that are diagnosed with a clinical disorder have, on average, a higher polygenic risk score than controls that haven’t been diagnosed.
But that shift in the mean value, the shift in the bell curve, upwards, to higher risk, for cases of a disorder, is really quite a subtle change. And in a recent paper published in Genome Medicine, Psychiatrist Evangelos Vassos, and I, reviewed the potential for doing what you’ve just talked about, for using polygenic risk scores, in clinical care.
And we highlighted substantial challenges that remain. So one of those is the effect sizes just aren’t really large enough, at the moment. But genetics moves very quickly, and we know that these scores will get more predictive. But another challenge is that currently our scores predict much better in people who have European ancestry, because that’s where much of genetic research has been performed.
But, of course, if we move any score into the clinical setting, it’s really important that that performs equally well across people of different ancestries, from different ethnic groups.
And so genetic studies has some catch up work to do in making sure that we have equally rich genetic information, from around the world, and in different ethnic groups. It’s…I think it’s really…I can’t emphasise enough, it’s crucially important that genetics doesn’t exacerbate health disparities that already exist, for different groups. It has to be part of the solution, not the problem.
But, also, if we’re going to use genetic risk scores in clinical care, they have to be actionable. Just having that number, that score, is not enough. And when we use them in cancer, that action is clear. With a high polygenic risk score for breast cancer, for example, you need more regular screening.
You might need to be offered mammograms at an earlier age than average women in the population.
But it’s not clear to me what the equivalent would be for various psychiatric rates.
What might we do with an adolescent who has high polygenic risk scores with depression? How can we use that information for diagnosis, for treatment, or for follow up? And these are the, sorts of, questions that I’m really interested in working with psychiatrists with, to explore how that might work.
Interviewer: It’s really interesting. Could it be used as a form of, sort of, triage, given the limited resources, in mental health, as to who should receive treatment, and interventions, as a priority?
Professor Cathryn Lewis: Yes, it could possibly be used as triage. It could be used as help in diagnosis.
But, I think, before we use this, we’re going to need studies to figure out where it is helpful. Just because this information is available doesn’t necessarily mean that we should instantly use it. We need to know that that’s really going to have a positive role in clinical care, that clinicians having these scores is going to enable them to treat their patients more effectively.
Interviewer: And Cathryn, how are polygenic risk scores helpful for child, and adolescent, mental health professionals? Why should they take an interest in this?
Professor Cathryn Lewis: I think there are three reasons why child and adolescent mental health professionals should be interested in genetics, in general. So we know that developmental disorders, adolescent onset mental health disorders, are highly heritable. We’ve known that for decades, from twin studies, and from family studies. And now, for the first time, we have the chance to convert that umbrella level heritability figure into individual level risk scores, that’s based on molecular information. So why wouldn’t we? And, secondly, although I’ve focused on the polygenic component, there are some single gene, rare, large effect variants, some structural variation, that are relevant to autism, intellectual disability, and other developmental disorders.
And although they may only account for a small proportion of cases, if they exist, it gives you a real hint towards causation. And so, that’s really important to know about.
And thirdly, genetics is a new source of information that should, perhaps, be combined with all the other information, that child health professionals gather. They know a huge amount about their patients, and their families. And, of course, the notable thing about genetics is that it does not change. So the information we gather in a polygenic risk score at age eight, or 18, or 80, is exactly the same. So genetics can be a solid thread that would continue through fluctuating symptoms, and different diagnoses, across someone’s care.
Interviewer: You’re very widely published. Are there any studies that you can highlight that will be of particular interest to people working in child, and adolescent, mental health?
One paper that I really like to highlight, it’s not mine, but it’s an editorial written by Angelica Ronald that was published in the JCPP in April 2020. And she talks about polygenic scores, and gives a clear SWOT analysis, strengths, weaknesses, opportunities and threats of polygenic scores, and the potential for them being of value in child, and adolescent, psychiatry. And for those listeners that want to follow up in this area, I really highly recommend that paper by Angelica Ronald.
Interviewer: Great, fantastic. Cathryn, what else are you working on currently?
Professor Cathryn Lewis: So, one of the things I haven’t mentioned that we’re working on is pharmacogenetic studies. And these are genetic studies where we identify genetic variants that control, not risk, but response to drugs, both adverse events and efficacy. And I think this is an area where there could be a real impact in clinical care, and a contribution to what people are calling “Precision medicine”, prescribing the right treatment, for the right patient, at the right time.
So, for example, there are over 22 different antidepressants available, but there’s great variability in who responds to which antidepressant, with very few guidelines about which drug is likely to work, for whom.
And we hope that by identifying the genetic component of drug response, that might be a helpful strategy, to add to prescribing guidelines.
So we’ve shown that response to antidepressants is partly genetically determined, and we’ve recently identified a genetic profile to predict who is likely to respond to an antidepressant. Currently, this is really a proof of principle study, the effect is very weak. It has statistical significance, but not yet clinical significance. Like, with so many genetic studies, we need to do a lot more research before that’s relevant, in a clinical setting. But I’m really excited by this area of pharmacogenetics as a potential application for polygenic risk scores in psychiatry.
We know that only about a third of people respond to the first antidepressant they’re prescribed. Many people have to try several antidepressants before they find one that gives them effective treatment of their symptoms.
And if we could use genetics to inform which antidepressants someone is prescribed first, even a modest increase, in response rates, could have a really big impact on that person’s care, by helping them respond earlier, and helping them recover from depression.
Interviewer: Are there any key findings, from other recent work, that you care to share with us?
Professor Cathryn Lewis: One of the things that polygenic risk scores allows us to do is to look at the relationship between different disorders and see how much of that is genetically determined. So we’ve been exploring that to look at the known connection between autoimmune disorders, and depression, and the relationship there is bidirectional. So a diagnosis of depression increases someone’s risk of rheumatoid arthritis, for example, and vice versa. And it has always been an epidemiological puzzle of why that would be, and what the mechanism is? Is it biological, is it natural response to being diagnosed with a chronic disorder? But then why would it work in the other direction, as well?
And we’ve been using genetics to see if there’s a common genetic underpinning to depression, and rheumatoid arthritis, that causes that bidirectional association and risk.
And we’ve just put up a preprint on Med Archive showing that genetics doesn’t seem to be the answer there to that increase in risk, in both directions, between depression, and autoimmune disorders. That’s intriguing, but it’s…I think it’s equally important to dismiss those ideas and so that then leaves us to go ahead and say, “Well, right, what is that causing that relationship between the disorders?”
Interviewer: Cathyrn, is there anything else, in the pipeline, that you’d like to mention?
Professor Cathryn Lewis: Maybe I can just mention one intriguing result that we’re pondering over, at the moment, which comes from a study we’ve defined on treatment resistant depression using UK Biobank data. And UK Biobank is this massive study of half a million people from the UK.
And we’ve shown that depression cases that have treatment resistant depression, so, they’ve not responded to at least two antidepressants. Intriguingly, they have higher polygenic risk scores for ADHD, than depression cases who don’t have TRD. And, again, that’s, you know, that’s really curious to me. What does it tell us about diagnosis, or treatment? And there’s just so much more to find out, to really figure out this genetic landscape, not for a single disorder, but looking across disorders, and how that can inform us, about psychiatry and care.
Interviewer: It’s such a rich area, isn’t it? Finally, what is your takeaway message, for those listening to our conversation?
Professor Cathryn Lewis: My take home message is that genetics is slowly reaching the point where it will be of relevance to clinicians, including child, and adolescent, mental health professionals. And as genetic data becomes more available, using this at all stages of health care, in risk projection, in diagnosis, in treatment, in prognosis, may become feasible. But I do want to highlight that genetics is not a silver bullet. Psychiatry is too complex for any single tool to provide the whole answer.
But those working in child, and adolescent, psychiatry know their patients, and their families, so well, that genetics may play a role, in combination with all the rich sources of information that they gather.
And I do hope that over the next decades that genetics will be relevant. It will be a useful source of information, to help with diagnosis, to help with decisions about treatment, and also prognosis. Because we know that that is so variable, across cases. We’re not at that point yet, but research is progressing fast. And I look forward to the day where we can move that progress into tools, for effective clinical care.
Interviewer: Cathryn, thank you so much. For more details on Professor Cathryn Lewis, please visit the ACAMH website www.acamh.org and Twitter at ACAMH. ACAMH is spealt ACAMH. And don’t forget to follow us on iTunes, or your preferred streaming platform. Let us know if you enjoyed the podcast with a rating, or review, and do share with friends, and colleagues.