Analysis of a Million-Plus Genomes Points to Blurring Lines Among Brain Disorders

Is lower academic achievement in early life tied to the same gene changes as an increased risk for Alzheimer’s in older age? That is one of dozens of possible deductions to be drawn from the largest genomic study of brain conditions ever conducted, research that obscures what often have been considered clear diagnostic borders.

According to the findings, published June 22 in Science, conditions such as schizophrenia, major depressive disorder (MDD) and bipolar disorder share a suite of overlapping genetic variants rather than having distinct genetic signatures.

In addition to the genetic links between educational attainment and Alzheimer’s risk, the results link neuroticism to anorexia nervosa, anxiety disorders, MDD and obsessive-compulsive disorder (OCD). Neurological disorders like Parkinson’s and multiple sclerosis, however, have few variants in common with each other or with psychiatric conditions.

This mother lode of findings comes after a six-year delving into genomes representing more than a million people, a quest for unusual genetic signals that track with one or more of 42 disorders and traits.

Researchers from 600-plus institutions worldwide, grouped into 25 consortia, pooled their genomic data for the effort, dubbed the “Brainstorm Consortium.” Their goal was to probe the immense data set for links among gene variants, brain disorders and physical and cognitive traits.

And they found many, many links.

“One of the big messages is that psychiatric disorders turned out to be very connected on the genetic level,” says Verneri Anttila, the first author on the paper and a postdoctoral research fellow at the Broad Institute. The implication is that “current diagnostics don’t accurately separate the mechanisms” for the conditions, he says, which might be a factor in explaining the struggle to find new treatments.

But because the study was a “hypothesis-free approach,” as Anttila describes it, showing only statistical associations among genes, not proof of a common genetic basis, the findings are only a starting point for digging deeper “to better understand how these disorders arise,” he says.  

To deliver this treasure trove of starting points, the Brainstorm Consortium applied a statistical method that teases out truly distinct signals from noise in the genome and tracked how those signals—small changes in DNA sequences that represent gene variants—associated with psychiatric and neurological diagnoses of 265,218 patients compared to 784,643 unaffected people.

Then, they looked at how variants that tracked with brain disorders related to educational achievement and physical and cognitive measures. Ultimately, the analysis involved 25 conditions, including MDD, autism, epilepsy, schizophrenia, post-traumatic stress disorder (PTSD) and migraine.

To spice it up a bit, the investigators also looked at 13 behavioral traits, including cognitive and personality measures, and four physical factors—Crohn’s disease, vascular disease, height, and body mass index.

From this brew of inputs, the researchers extracted statistical links between genetic variants and different disorders and identified overlaps across almost all psychiatric conditions they examined. Schizophrenia, anxiety disorder, MDD, bipolar disorder and attention deficit hyperactivity disorder (ADHD) all shared variants. Tourette syndrome, OCD and MDD clustered together, as did anorexia nervosa, schizophrenia and OCD.

Standing out from this crowd at the end of the shared-variant spectrum was schizophrenia, which overlapped with all of the psychiatric disorders except anxiety. PTSD, meanwhile, showed no significant association with any of them.

Although neurological disorders tended to remain distinct, an exception was migraine, which had many variants in common with ADHD, Tourette syndrome and MDD. Migraine also was associated with the personality trait “neuroticism,” which in turn overlapped with several psychiatric conditions, including MDD, OCD and schizophrenia.

Crohn’s disease, intended to represent immune disorders, had no variants in common with any other condition assessed, but vascular disease, stroke and MDD had commonalities.

Some of these associations are surprising, and some are not, says John Hardy, chair of molecular biology of neurological disease at the University College London Institute of Neurology, who was not involved with the study. Hardy calls the study a “reliable and well-organized piece of work.” The overlap among psychiatric diseases isn’t a big shock, he says, because “they share symptoms and can sometimes be confused with each other.”

What stands out for him is a lack of overlap between Alzheimer’s and Parkinson’s, both neurodegenerative conditions. Based on these findings, he says, “they really do seem distinct, which probably means that the mechanisms of cell death are different.”

Anttila mentions a couple of other curveballs. “I was personally surprised by the lack of such correlations between neurological disorders and psychiatric ones,” he says, noting that he would have expected depression, for example, to show overlap with some neurological diseases.

The other unexpected result for Anttila and his colleagues was the link between cognitive factors and the psychiatric and neurological disorders. “Having the genetic variants that predispose someone to Alzheimer’s disease at age 70 to be significantly correlated with those predisposing them towards worse results in school at age 12 was definitely a surprise to us,” he says.

Some of the non-overlapping results were unanticipated, too. Even though about a third of people with autism also have seizures, this probe of genomic information turned up little in the way of common variants between autism and epilepsy. Anttila says that this analysis was “one of the smaller ones” in the work, which might explain the lack of a link.

In fact, having fewer than 10,000 cases in any one of the many analyses in the study could have yielded misleading results because of the small sample size, not because there is no association, says Cathryn Lewis, professor of genetic epidemiology and statistics at King’s College London. Lewis is a member of the U.K.’s Psychiatric Genomics Consortium and contributed samples to the study but was not involved in the analysis.

Another possible explanation for the lack of shared variants between autism and epilepsy is that the study focused on commonly occurring, but not on rare, candidates. “The methods are definitely sound, well-respected, and widely used,” she says, but because they involve looking only at common variants, they won’t capture rare ones.

With dozens of new investigative starting points involving these common variants in-hand, what’s next? “The immediate takeaway from this study is that now that we have identified these connections, we can better understand how these disorders arise,” says Anttila. “There may be some deeper genetic mechanisms at play here [that] predispose individuals to multiple disorders, rather than just a single one.”