Around a third of people complain of some sleeplessness, and one in 10 meets diagnostic criteria for clinical insomnia. The costs, in terms of well-being, physical health and productivity, are enormous. From twin studies, researchers know the inability to fall or stay asleep has a genetic component, but the identities of the culprits were mostly unknown.
Now, two studies published Monday in Nature Genetics provide first peeks at the biological basis of insomnia, implicating specific brain regions and biological processes, and revealing links with heart disease and psychiatric disorders like depression. Both are genome-wide association studies (GWASs), which examine DNA from many thousands of individuals to determine where genetic markers related to health, disease or a particular trait reside.
The first study, from a team led by geneticist Danielle Posthuma of Vrije University Amsterdam, analyzed the genomes of over 1.3 million people, making it the largest GWAS of any complex trait to date. They used data from the UK Biobank, a large, long-term genetics project, and from the direct-to-consumer genetics company 23andMe to identify 202 areas of the genome linked to insomnia, implicating 956 genes, a big advance from the seven found previously. “I’m pretty confident the vast majority of these are real,” says geneticist Stephan Ripke, a GWAS expert at the Berlin Institute of Health who was not involved in either study. “But we need to confirm this in more, separate cohorts from different countries and researchers.”
The researchers then investigated which brain regions and cells these genes frequently turn up in. This analysis implicated the axons (output connections) of neurons as well as parts of the cortex and deeper “subcortical” brain regions like the striatum, involved in movement. It also tagged “medium spiny neurons,” which occupy most of the striatum as well as neurons in other regions, including the hypothalamus. These findings tally with brain-imaging studies suggesting dysfunction of some regions in insomnia, and with animal studies implicating these cells in sleep regulation. “Before our study we knew little about which genes, pathways and cells were involved,” Posthuma says. “We now have concrete hypotheses that can be tested.”
The second study, from a team led by geneticist Richa Saxena of Massachusetts General Hospital, interrogated over 450,000 genomes, again from the UK Biobank. They identified 57 regions, implicating 236 genes, and confirmed these results in analyses of two separate data sets. One of these used clinically diagnosed patients, a contrast to the other data compendium that was based on less reliable, self-reported symptoms. The team went further by analyzing data from nearly 84,000 UK Biobank participants who had worn motion detectors for a week to observe tossing and turning or sleepwalking, enabling them to link genetic findings with actual measures of sleep. “This shows the findings are valid for different definitions of insomnia-related symptoms, including some that are measured objectively,” according to Virginia Commonwealth University statistical geneticist Mackenzie Lind.
The two studies found significant overlap between genes implicated in insomnia and those related to psychiatric and metabolic traits. Genes for traits, including depression, anxiety, schizophrenia, coronary artery disease and type 2 diabetes were sometimes the same. The findings suggest insomnia is more strongly related to neuropsychiatric disorders than to other sleep-related traits such as whether someone is a morning person. “That was a big surprise,” Saxena says. “Implying that at the genetic level it’s a disorder that’s likely linked to psychiatric disease and mood regulation, and it’s not necessarily just about sleep regulation.”
Both teams also used a technique (Mendelian randomization) that allowed them to infer what might be causing what by comparing their findings with GWAS results for other conditions. The two studies found insomnia may cause depression and coronary artery disease, and the larger study also found causal risk effects for BMI (body mass index) and type 2 diabetes. “One of the motivations for using genetics to study sleep was to tease apart where it’s causal where it’s not,” Saxena says. “So eventually interventions can be targeted to areas where things are causal.” Not all researchers are confident in these tests, however. “The genetic overlap is sound,” Ripke says. “But there’s debate about these Mendelian randomization tests; I wouldn’t take this for granted.”
Both studies implicated a gene involved in restless leg syndrome, which “makes sense, given it’s also a sleep disturbance,” Saxena says, although her team also found this may have been partly due to undiagnosed cases of RLS in their data sets. In fact, it is probable insomnia is not a singular condition but a cluster of symptoms grouped together, which can have a range of underlying causes. It could be a consequence of childhood trauma in one patient, due to disrupted circadian processes in another or just resulting from restless leg syndrome in another. “If that’s the case, then we’ll really be able to dissect that with the genetics,” she says. “Understanding if there are different types of insomnia, and how can we study them and maybe treat them separately, that’s the hope for the whole field.”
Going forward, these findings provide entry points for researchers to dive into the biology of insomnia. “We’re following two strategies now,” Posthuma says. Implementation is proceeding by “increasing sample size even further,” she notes, “and setting up lab experiments to prove causation and show how implicated cell types influence insomnia.” Studies like this may illuminate new therapeutic targets. Although treatments exist, access to therapies like cognitive behavioral therapy cannot meet existing demand.
The behavioral therapy demonstrates, however, why this line of research is worth pursuing. The genetic overlap between insomnia and mood disorders may point toward why cognitive behavioral therapy may be effective for both sleep and anxiety. Current drugs, for their part, have limited efficacy, can be addictive and have side effects. “Identifying new variants that contribute to risk helps pinpoint new biological targets,” Lind says. This search, she adds, is “a step toward the eventual goal of using genetic information to predict risk and treatment outcomes, although we’re not at this point yet.”