The genetics of early-onset psychosis: Could it aid understanding of schizophrenia?

psychosis schizophrenia
(Thomas Zapata/Wikimedia Commons)

At age six, Matthew (not his real name) began hearing voices coming out of the walls and the school intercom, telling him to hurt himself and others. He saw ghosts, aliens in trees and color footprints. Joseph Gonzalez-Heydrich, MD, a psychiatrist at Boston Children’s Hospital, put Matthew, at age 9, on antipsychotic medications, and the hallucinations stopped.

It’s rare for children so young to have psychotic symptoms. Intrigued, Gonzalez-Heydrich referred Matthew for genetic testing.

Sussing out psychosis

Boston Children’s partner diagnostics company, Claritas Genomics, ran a full chromosome microarray analysis. It found that Matthew had lost a chunk of DNA: one copy of the chromosome area 16p13.11 was missing.

early onset psychosis
(Craig Finn/Creative Commons)

Could other children with early psychosis have the same deletion? Several others came to light. One girl had a DNA duplication at 16p13.11 rather than a deletion. By age 4, she was hallucinating monsters, a big black wolf, spiders crawling in her ears and a man with blood on his face. She also had been diagnosed with a Chiari malformation, seizures, autism spectrum disorder and speech delay.

Gonzalez-Heydrich, together with Catherine Brownstein, MPH, PhD, director of Boston Children’s Molecular Genetics Core Facility, and other colleagues, launched a genetic study of very-early-onset psychosis — psychotic symptoms emerging before the age of 13. To date, they have enrolled about 50 families.

The project began with a five-year gift from a family who had lost a son to suicide, which they followed with an additional gift establishing the Tommy Fuss Center for Neuropsychiatric Disease Research at Boston Children’s. With further support from Boston Children’s Manton Center for Orphan Disease Research and the hospital’s Translational Neuroscience Center (TNC), Gonzalez-Heydrich hopes to put a new lens on psychosis and to unlock better treatments.

From genetics, better antipsychotics?

Most antipsychotics today are based on chlorpromazine and clozapine, which work through dopamine and/or serotonin receptors and were developed in the 1950s.

early-onset psychosis

“These treatments suppress symptoms, but the majority of patients still have a lot of functional impairment,” says Gonzalez-Heydrich. “We can suppress delusions and hallucinations, but it’s been harder to treat cognitive and emotional symptoms like withdrawal, insufficient motivation and poor concentration. Those can be very impairing.”

The medications cause weight gain, pose a risk for type 2 diabetes and can themselves reduce motivation, concentration and sense of enjoyment, so patients often want to go off them. Long-term, the drugs can cause tardive dyskinesia, a movement disorder.

genetics psychosis

The genetic mutations — identified by Brownstein and colleagues through chromosome microarray analyses and, in some cases, whole-exome sequencing —are uncovering new biological pathways that may be involved in psychosis. These, in turn, could lead to new treatments with different mechanisms of action.

And they could shed light not only on early-onset psychosis but on schizophrenia in general, which affects about 1 percent of the population.

Delving into DNA

Brownstein and Gonzalez-Heydrich are working closely with the TNC’s Robin Kleiman, PhD, to investigate the genetic variants, beginning with the 16p.13.11 duplication/deletion cases.

16p.13.11 spans at least four genes that could plausibly be related to psychosis: NTAN1, part of a pathway involved in stabilizing certain important proteins in neurons; RRN3, part of the mTOR pathway implicated in various neurological and psychiatric conditions; NDE1, thought to regulate neurodevelopmental processes like neuron migration and differentiation; and PDXDC1, known to be asssociated with the gating of sensory stimulation of the sort known to be abnormal in schizophrenia.

early-onset psychosis

Using some of Matthew’s skin cells, Kleiman and colleagues in the TNC’s Human Neuron Core created induced pluripotent stem cells. They then converted these cells into two different types of neurons that are affected in schizophrenia. Each lab-created neuron carries Matthew’s DNA, allowing the researchers to model the effects of his 16p.13.11 deletion and look for possible signaling disturbances that may explain how his mutation changes neuronal function.

“Each mutation involves a bit of detective work to see what that mutation does and how it fits into what is known about the pathophysiology of psychosis,” says Kleiman. “Eventually, we hope to screen potential drugs to see if there are existing medications that could reverse an observed neuronal phenotype.”

Another mutation, in the gene ATP1A3, was found through whole-exome sequencing in a boy who, at age 6, had begun hearing voices of little boys saying “bad things,” telling him to hurt himself and others. The boy also had selective mutism, self-injurious behavior, aggression and mild motor delays. The voices, self-injurious behavior and mutism went away once he started antipsychotic medication.

The ATP1A3 gene turns out to encode a protein that helps transport ions into and out of electrically active cells. Kleiman plans to collaborate with researchers at Vanderbilt University who have created assays to study another ATP1A3-related condition.

“We think this protein is important in helping neurons recover after they fire an action potential,” she says. “Again, understanding how mutations affect this process may help us establish drug screens, using the patient’s neurons to test potential drugs and drug combinations.”

One gene, many disorders

If We Could Undo Psychosis (Jane Kelly/
If We Could Undo Psychosis (Jane Kelly/

A given neurogenetic mutation can produce different symptoms, Kleiman notes. The effects depend on which circuits are most vulnerable to the mutation, the child’s stage in development and how the mutation interacts with the child’s genetic background.

For example, a significant mutation affecting a protein important during infancy could lead to autism spectrum disorder, whereas a less severe mutation might not have consequences until adolescence — and then cause symptoms of psychosis. A mutation’s effects can be further modified by environmental factors, other genes and the brain’s own attempt to compensate.

That could explain why copy number variants at 16p13.11, for instance, have been linked to non-psychotic disorders including intellectual disability, autism, epilepsy and attention-deficit hyperactivity disorder (ADHD).

“A lot of neuropsychiatric disorders are very heterogenous — that’s what makes them tough to develop treatments for,” says Kleiman. “Everyone has a different collection of genetic and environmental insults triggering the vulnerability of brain circuits. Long-term, what’s going to be required is a precision medicine approach, with biomarkers that stratify patients according to the dysfunction of different pathways.”

A biomarker for psychosis risk?

Biomarkers that appear before psychotic symptoms are of particular interest. Together with genetic tests, they could allow caregivers to intervene before a child or adolescent pivots into full-blown psychosis.

Remain Unnamed (caelanthealien/
Remain Unnamed (caelanthealien/

“The theory is, by the time you get to hearing voices, you’ve had a lot of damage to your brain,” says Gonzalez-Heydrich. “Ultimately, the goal would be to prevent a psychotic break.”

Stressful events can trigger psychosis in people with an underlying vulnerability, as can drugs like methamphetamine, PCP and marijuana. One study subject with a 16p13.11 deletion had a first-time manic-psychotic break in midlife, apparently triggered by Adderall, a common ADHD medication. His son, with the same deletion, developed first-time psychotic symptoms soon after.

Working with Boston Children’s neurologist Frank Duffy, MD, Gonzalez-Heydrich and colleagues are using advanced, quantitative EEG techniques to investigate differences in brain connectivity, sensory responses and plasticity in children at clinical risk for psychosis. Tracked over time, such EEG markers could potentially help identify children likely to “convert” to schizophrenia, Gonzalez-Heydrich says. EEG markers will also be important metrics in clinical trials, allowing researchers to test whether treatment is having an effect.

The rare informing the common

Now age 11, Matthew continues to do well on traditional antipsychotics. But his case has an important place in schizophrenia research. Though genetic variants like his are rare, they have a stronger effect than the more common mutations. That makes them easier to study and model in patient-derived cells and in animals.

“Some 30 percent of risk for schizophrenia is from common genetic variants,” says Gonzalez-Heydrich. “All increase risk individually by a very small amount. With childhood cases, we are seeing more rare variants that probably cause more severe psychosis. Hopefully, these kids will point us to some common pathways that are also involved in adolescent/young adult–onset schizophrenia. That’s what we’re looking for.”