Stanford Medicine scientists are generating a periodic table of sorts for psychiatric disorders, providing better understanding of these conditions and paving the way toward targeted treatment.

By combining two massive, publicly available databases – one flagging genes associated with psychiatric disorders, the other showing which cells in which parts of the human brain are making the most use of which of our genes – they’ve implicated certain cell types, located in particular brain regions, in schizophrenia.

Like the periodic table of elements, which has enabled generations of scientists to predict the existence of yet-undiscovered elements and the behavior of those already known, the brain-cell classification system is the product of two sets of observations. In the former, the breakthrough stemmed from organizing the elements in a two-dimensional grid, presenting them not just in order of how many resident protons their atoms housed but also according to their chemical properties. The latter likewise combines two separate series of observations, yielding both confirmations of imaging- and autopsy-derived findings and unearthing previously unsuspected types of cells, in specific brain regions, that may be participants in the pathology of psychiatric disorders.

A study detailing these findings – the first of its kind to rely on fully human data throughout – was published Jan. 20 in Nature Neuroscience.

The still-experimental combinatorial method reveals a new way to learn about psychiatric disorders in general, said the study’s senior author, Laramie Duncan, PhD, assistant professor of psychiatry and behavioral sciences.

A periodic table for cells

The new study confirms many previous findings, from imaging and post-mortem tissue analyses, regarding places in the brain housing structures that are suspiciously small or where nerve-cell signaling appears disrupted in schizophrenia. It also implicates new brain-cell types schizophrenia researchers have not previously focused on. And it has unearthed brain-cell types, in key brain structures, that are common to psychiatric disorders beyond schizophrenia.

“Psychiatric disorders are mysterious, even though they impact at least one-fifth of the population at any given time,” Duncan said. Yet the pace of developing treatments for psychiatric disorders has been extremely slow.

“That’s partly because these disorders are so complex,” she said. “But it’s also because we don’t have a good neurobiological understanding of what’s causing them. A key step forward in understanding why people develop these disorders is to identify some of the precise cell types in the brain that contribute to them.”

Knowing that a particular kind of a cell is involved in a psychiatric disorder, plus knowing how that cell normally works and where it resides, provides a clue as to how that cell type’s dysfunction may be contributing to the disorder. Because the receptors on or in many cell types are known, it also zeroes in on what kind of drugs might work for schizophrenia.

That’s easy to say. But the brain is hard to study. Sampling cells deep in that organ typically requires an autopsy, not a biopsy. In contrast, the method the scientists used is noninvasive: It requires computation, not a surgical operation or even imaging.

The researchers focused primarily on schizophrenia because it’s a serious disorder found worldwide (about 0.5% of all individuals); genes account for roughly 70% to 80% of the variability in people’s likelihood of developing schizophrenia, the biggest genetic contribution of any psychiatric disorder; and it’s more reliably diagnosed than other disorders.

“Schizophrenia is the quintessential psychiatric disorder,” Duncan said. It’s marked by hallucinations (people see or hear things others don’t), delusions (believe themselves to be someone else, often someone famous) and profound difficulties in accomplishing daily activities.

It’s devastating. “For many, the symptoms are so severe that people with schizophrenia end up sleeping on the streets,” Duncan said.

One plus one equals three

One of the two databases the researchers used was drawn from a large genome-wide association study, or GWAS. Genes often come in versions – they’re mostly identical from one person to the next, but relatively tiny differences in and near genes can make big differences in how, and how much, they influence the behavior of the cells in which these genes are active. In a GWAS, investigators scour the genomes of large numbers of people with and without a trait of interest – say, schizophrenia – and compare them to see if those with the trait are disproportionately likely, or unlikely, to carry one or another version of any or many genes.

A recent GWAS analyzed a pool of 320,404 people and found 287 genes in or near which one given version of a genetically variable portion was statistically overrepresented among people with schizophrenia. The team fixed their sights on these “schizophrenia-associated genes” for their study.

The second database was a brainwide catalog showing which genes each different type of cell in a given part of the human brain actively uses, and how much of it. All cells carry effectively the same bunch of genes, but different groups of those genes are used in, say, a nerve cell versus a liver cell – or even a different variety of nerve cell. An inspection of 3,369,219 cells extracted from 105 regions of autopsied human brains defined 461 cell types by their differing patterns of gene usage and where in the brain the cells came from. The researchers combed this database for brain cells, making heavy use of genes identified in the GWAS as schizophrenia-associated. Those cells were considered likely candidates for involvement in the pathology of schizophrenia.

The scientists found 109 cell types that fit this description and bore down on 10 representative cell types that statistically were most strongly associated with the illness.

The two most significantly schizophrenia-associated cell types had similar functions – selectively inhibiting and therefore shaping excitatory activity in the cerebral cortex, the human brain’s outermost and most recently evolved structure. These two cell types separately localized to two different layers of that six-layer brain coating. Both layers have appeared shrunken in postmortem studies of schizophrenia patients’ brains.

The researchers also identified a brain cell type that hasn’t been associated with schizophrenia. It’s located in the retrosplenial cortex, which has been found to play a role in people’s sense of self (the sense of being within your body, not dissociated from it).

“The retrosplenial cortex hasn’t received much attention, but it might be a core component of some circuit important for psychiatric disorders,” Duncan said. “Disruption of one’s sense of self is an experience common to several psychiatric disorders. We’ve found these same cells to be involved in every psychiatric disorder we’ve looked at.”

Two other distinct schizophrenia-associated cell types dwell in a brain structure called the amygdala, considered a hub of nerve-cell activity related to threat assessment and fear. (Fear is one of the defining emotional features for many who suffer from schizophrenia.) Two additional strongly schizophrenia-associated cell types were found in the hippocampus, and one turned up in the thalamus.

These three evolutionarily ancient subcortical structures are precisely the same subcortical structures that, in imaging studies, most reliably show shrinkage in schizophrenic brains compared with healthy people’s brains.

The road to personalized medicine

“Now we have a roadmap showing specific directions to go in understanding this disorder,” Duncan said. “We know exactly which cell types to study further in the lab, we have new targets for drugs, and we are using genetic information from individual patients to predict what medicine a person should take.”

Duncan said she thinks it will take six or seven years before scientists have useful clinical applications such as matching patients to therapies.

“Our study didn’t look at cells in which schizophrenia-associated genes were remarkably underactive,” she said. “We want to refine our model to include underrepresented genes as well as overrepresented ones, so we get an even better understanding of the cell types involved. Plus, we’re expanding the model to more psychiatric disorders. We hope to identify groups of people whose cell-type profiles are characteristic of specific disorders, or subsets of those disorders,” she said. “That may help us predict or discover which old or new drugs, or combinations of them, will work best for a given patient.”

Researchers from the University of Alabama, Birmingham, and the University of San Francisco also contributed to the work.