Genome Studies Open Window on Schizophrenia Etiology

This is the second in a two-part series on advances in research on the contributions of the genome to the development of mental illness. The first installment appeared in the November 15 issue.

Genomewide association studies (GWAS) have in recent years identified genetic variants linked to development of schizophrenia, part of a series of rapid advances that have energized the field of psychiatric genetics.

Based on multiple analyses of data from the international collaboration known as the Psychiatric Genome Consortium (PGC)—some of which have been published, and more are soon to be in press—the genetic architecture of schizophrenia is fast emerging, Patrick Sullivan, M.D., a professor of genetics and psychiatry and director of the Psychiatric Genomics Department at the University of North Carolina School of Medicine, told Psychiatric News.

Genetic architecture describes the correspondence between phenotypes (e.g., symptoms) and underlying genotypes (e.g., disease-causing genetic variants), particularly how many different loci are involved and the effects of these loci.

The earlier search for genetic variations with a strong, large effect has turned up only a few significant findings. Although more work is being done, scientists have “looked pretty hard” for variants with large effects and should have found them, according to Sullivan. Rather, it appears that schizophrenia is partly caused by “hundreds of variations in hundreds of different genes, where the effects of each variation are very subtle,” he noted.

Jonathan Sebat, Ph.D., an associate professor of psychiatry and the chief of the Center for Molecular Genomics of Neuropsychiatric Diseases at the University of California, San Diego (UCSD), explained the genetic causes of schizophrenia in a somewhat different way. (Sebat coordinates the copy-number variation [CNV] group in the PGC, which is finishing a separate study on schizophrenia and anticipates its publication within a few months, he told Psychiatric News.)

Previous research has shown that CNVs, which includes deletions or duplications of large chunks of DNA, contribute to schizophrenia as well as to autism and bipolar disorder, and specific CNVs have been located. However, the extent of CNVs’ contribution to psychiatric disorders remains a subject of debate. Sebat estimated that de novo CNV mutations, which are usually seen in only one person in a sample population, contribute to from 4 percent to 5 percent of all schizophrenia cases. “So far all the CNVs associated with schizophrenia [that we know of] are rare variations, none is common,” he explained.

Sebat said that in one patient with schizophrenia, a rare CNV or a few variants could account for the total cause of his or her disease; another patient, also diagnosed with schizophrenia, may have different genetic variants and different biological mechanisms. Do they have the same disease or different diseases that look similar? Often the underlying pathology may fundamentally differ. Therefore, the complexity of psychiatric genetics also lies in the heterogeneity in a symptom-based classification system.

“Symptoms alone have turned out to be not very reliable for predicting the causes of the disorders,” Sebat said. “A constellation of similar symptoms can come from lots of different genetic and biological errors. Until we understand the underlying causes of mental disorders, we cannot correctly reclassify them.”

Progress Made With Other Disorders

Since 2010, PGC groups have published meta-analyses on attention-deficit/hyperactivity disorder (ADHD), bipolar disorder, major depressive disorder (MDD), and cross-disorder overlaps. The progress is dependent on the contribution of more genotyping data from researchers.

The MDD genetic architecture is more challenging, Sullivan acknowledged, as the disorder is very heterogeneous, more common, and less heritable than, say, schizophrenia. In an analysis of 9,000 cases and as many controls, the group found hints of risk genes but none of statistical significance. “We need larger samples,” Sullivan said, which are currently accumulating. “By the end of the year we may have 30,000 to 40,000 cases in the database.”

He believes the progress in schizophrenia can be a model for other psychiatric disorders. “Using relatively inexpensive and standard tools, we will be able to learn a lot about the [underlying genetics] of any psychiatric disorders,” he said.

In studies published this year, the cross-disorder group demonstrated risk loci associated with all five disorders (schizophrenia, bipolar disorder, MDD, autism, ADHD), with strong evidence implicating the calcium channel signaling system, which is important for neuronal communications, as a pathway conferring broad genetic risks to psychiatric disorders, according to Jordan Smoller, M.D., Sc.D., a professor of psychiatry at Harvard Medical School and director of psychiatric and neurodevelopmental genetics at Massachusetts General Hospital.

Smoller coordinates the PGC’s cross-disorder analyses. He said that the group is extending the analyses to include new datasets on anorexia nervosa, OCD, and Tourette syndrome.

“Our findings show that genetic risk variants do not necessarily respect the boundaries of our diagnostic criteria,” Smoller said. “The genome hasn’t read the DSM.” He hopes that soon the causes of psychiatric disorders gleaned from genomic studies will be incorporated into the classification and definitions of psychiatric disorders rather than the classification relying solely on clinical symptoms.

“The PGC represents a sea change in the way research is done in psychiatry,” said Smoller. “In the last five years, in part as a result of the PGC, we have gone from having almost no confirmed genetic risk loci [in psychiatry] to having 150 to 200 of them. It’s dramatic progress.”

He believes that the success has convinced more investigators to join the consortium, which led to accelerated success.

Researchers in the consortium also benefit from state-of-the-art tools and technology. For example, several PGC coordinators emphasized that the consortium has engaged top-notch genetic statisticians to develop sophisticated tools to ensure that analyses are performed correctly and consistently.

Genotyping data from each individual, recorded with microarray “chips,” routinely contain 500,000 to 1 million single-nucleotide polymorphisms (SNPs). In GWAS, patients and controls are compared on these SNPs to identify significant differences that may be associated with a disease. Because of the large number of comparisons, the statistics are more complex than clinical or epidemiological research, and the standard for drawing conclusions is therefore much higher.

The PGC receives funding from many public and private sources on both sides of the Atlantic. The National Institute of Mental Health (NIMH) is one of the founding members and has continued to be a funding source.

Thomas Lehner, Ph.D., M.P.H., chief of the Genomics Research Branch at NIMH, told Psychiatric News that the institute is “thrilled to support the effort and watch the field advancing so rapidly.”

“What we find so attractive about the PGC is that it organizes the field around common goals with a team-science approach and vastly increases the scale of experiment,” said Lehner. “It puts together very smart people to think about the problems and come up with effective solutions, including experts inside and outside the field of psychiatry. Without the PGC, progress would have been much delayed.”

Moving Discoveries From Lab to Clinic

One of the promises of genetic research is to reveal the pathophysiology of the illnesses. In psychiatric as well as nonpsychiatric illness, genomic research has turned up multiple risk loci previously unknown or unsuspected. These discoveries uncover underlying etiology of various brain dysfunctions and provide directions for research into gene-environment interactions and brain biochemistry. They also suggest potential targets for developing new drug treatments.

Genetic findings also guide personalized diagnosis and treatment. For example, known CNVs in autism can help stratify the disorder into different subtypes, and at UCSD, genetic testing for autistic patients is routinely used in clinical diagnosis, according to Sebat. He believes the same can and should be done for schizophrenia.

“For the first time, we actually have traction in understanding the genetic architecture of mental disorders, and the PGC is one of the driving forces behind the progress, said Lehner. ■