Major depression like diabetes or cardiovascular disease is a complex disease. Like those diseases, depression is driven by both genetic factors and a multitude of established environmental (stress, alcohol and drug use, lack of sleep) risk factors. But finding the genetic risk factors in depression has been more difficult than other complex diseases, requiring much larger sample sizes, as each factor only has minute effects.
In the era of candidate genes, the interaction between serotonin transporter promoter length polymorphism and stress in depression was among the most controversial and cited stories in genetics. While still intensely debated, studies designed to assess the association as individuals encounter a specific stressful situation (military service, post-partum period, natural disaster etc.) were consistently found to identify the gene x stress interaction in systematic reviews and meta-analyses that assessed for study design. So as the field transitioned to the GWAS era, we thought that maybe the broader genomic risk for stress-induced depression could be teased out by using a large group of people who are initially well, and then pushed to their limits, and measuring who will become depressed? But is that ethical? Well, it is done every year to medical interns! At the end of medical school, having happily finished that part of their journey to become a physician, they experience 70+-hour work weeks, grueling schedules that involve getting up very early and staying up all day and sometimes night, and, inevitably, more than a third becomes depressed. Through Polygenic Risk Scores (PRS), summing the effects across all common genetic variants from data available through large consortia and public datasets, we found that high PRS for depression predicted which interns would get depressed under stress. Importantly, low PRS was even better in predicting who would be resilient, i.e. not become depressed despite the stressful environment. What use do PRS have for the typical person? If everyone could get their PRS, what would you tell a person who is in the highest 10% of genetic risk for depression? Is generically avoiding stress a good suggestion? Ask a Buddhist or a self-help consultant, and they will say yes, probably people at high risk for depression – whether from PRS or because of family history – should try to avoid stress, as well as drugs and alcohol which are also known to trigger depression. Ask psychiatrists – and we did! – and you may get all kinds of answers, from “don’t ever tell a person to avoid stress, it’s like telling them to avoid life” to “yes, and stress reduction is good for all kinds of disorders, not just depression”.
Future PRS work could help to develop more specific and actionable guidance. Perhaps additional PRS investigation will reveal different vulnerabilities to different types of stress. For example, some PRS profiles may be more vulnerable to depression in the face of sleep deprivation and maintaining good sleep may be particularly indicated. For other profiles, substance-induced depression may be a particular risk. Another area where PRS may have actionable effects is in prevention. Interventions to prevent depression are reliably more effective than interventions to treat depression after symptoms are already advanced. However, prevention approaches have been limited because there has been little guidance on who and when to target interventions. Combining PRS with other known personal and environmental predisposing factors in a risk model could allow those most likely to develop depression to prioritize preventative interventions for themselves before exposure to a predictable stressful environment such as medical internship, child birth or deployment. For example, we have previously shown that in the population of training doctors, mobile CBT training before internship is effective in reducing progression to depression.
We also need to recognize that PRS, and genomics more broadly, are not deterministic. In our study, some residents who are genetically at high risk still had low depression score. What is it in their early life environment or current lifestyle that makes them so resilient? Already, we know that high work load and lack of sleep are part of the difficulties, and part of the protection comes from interns who are naturally early risers and hence less bothered by the early shifts. If other such factors can be identified, maybe they can be used to make the experience of internship less depressogenic.
Margit Burmeister, Yu Fang and Srijan Sen, Molecular & Behavioral Neuroscience Institute, University of Michigan