6 Jul 2014
This is Part 1 of a two-part series. See Part 2.
July 7, 2014. Genomics, connectomics, proteomics, methylomics, transcriptomics—"There are many -omics that are now in our scientific dictionary," said moderator Lynn DeLisi, Harvard University, Boston, Massachusetts, as she opened the SIRS 2014 morning plenary session on Sunday, April 6. Over the next three hours, researchers and clinicians on two separate panels focused first on the genomics, then on the connectomics, of schizophrenia.
The common and the rare
Patrick Sullivan, University of North Carolina at Chapel Hill, kicked off the session with a talk on the search for common variants, each of small effect, in schizophrenia using genomewide association studies (GWAS) (see SRF related news report).
Sullivan first touched briefly on exome sequencing, which deals exclusively with the coding portion of the genome. No single genes have emerged, but several pathways have been hits: neuronal calcium signaling, postsynaptic density/cytoskeletal (ARC) protein complex, NMDA receptors, and binding partners of Fragile X mental retardation protein (FMRP), he said. Data from another 10,000 cases are currently undergoing analysis.
Turning to GWAS, Sullivan noted that the cost has dropped dramatically in the past several years, enabling large studies. The largest published analysis to date, with 13,000 cases, identified 22 loci associated with schizophrenia (see SRF related news report). A Psychiatrics Genome Consortium (PGC) mega-analysis (of pooled raw data) of 36,000 cases finding 128 significant loci in 108 different places has been submitted for publication. This type of polygenic signal is consistent with other complex traits such as height, weight, and inflammatory bowel disease, he added. Many of the identified gene sets have convergent evidence from exome sequencing, copy number variant (CNV) and/or common variant analyses.
Sullivan then discussed the risk profile score, a computation of the number of schizophrenia risk alleles carried by a single person. There is a lot of overlap between cases and controls, but people with a greater number of risk alleles are more likely to be cases than controls, he said. Of special interest, Sullivan said, are the outliers: the minority of low-risk allele carriers that have schizophrenia, and the minority of high-risk allele carriers that do not.
Another member of the PGC, Jonathan Sebat, University of California, San Diego, provided an update on the search for rare variants that confer a larger risk for schizophrenia and tend to fall within de novo mutation hotspots in the genome. Using the same strategy employed for GWAS, the PGC has also looked at CNVs in large sample sizes. Sebat presented preliminary data from a mega-analysis that used the same analysis pipeline for a variety of samples. With this approach, much of the variability between individual studies fell away, Sebat said. On average, there were 1.2 times more DNA deleted in cases than in controls.
None of the CNV loci that met the criteria for genomewide significance for schizophrenia (such as 2p13, 3q29, and 22q11) are new; every single one has already been reported in pediatric clinical populations who, unlike schizophrenia patients, routinely undergo clinical genetic testing as part of the diagnosis process, said Sebat. This indicates that this same strategy could be used in schizophrenia, he added (see SRF related news report).
At the end of his talk, Sebat touched briefly on gene-based association studies as well as pathway-based gene set analyses that are in progress. The latter point to anomalies in several pathways, including postsynaptic density proteins and those associated with FMRP.
Bring in the environment
Tiina Paunio, National Public Health Institute, Helsinki, Finland, discussed the interplay between genes and environment in schizophrenia. The most convincing evidence for a combination of both factors in the disease comes from twin studies, she said. The less than 100 percent concordance rate for monozygotic twins suggests that environment may play a major role.
Environmental risk factors in the illness can be divided into two categories, Paunio said. The first are factors that can be assumed to exert their effects during early development, such as advanced paternal age, season of birth, and birth complications. The other category concerns factors of later development, such as migration, urban upbringing, and early childhood trauma.
Only a few studies have examined gene-environment (G x E) interactions in schizophrenia, with inconsistent findings, Paunio said. The challenges of these studies include defining environmental factors, measurement errors in such factors, insufficient sample sizes, and inconsistent use of statistical modeling. A review of major G x E studies in psychiatry concluded that a substantial amount of publication bias among replication attempts exists (Duncan and Keller, 2011), necessitating caution going forward, she added, noting that the identification of true G x E interactions is difficult.
Paunio then turned to the mechanism by which G x E interactions occur. She presented two examples of altered epigenetic regulation in schizophrenia: a dysregulation of the GABAergic system as evidenced by altered histone modifications of the GAD67 gene (see SRF related news report) and abnormal regulation of microRNAs such as miR-137 (see SRF related news report).
In consumer hands
Francis McMahon, National Institute of Mental Health, Bethesda, Maryland, rounded out the genomics half of the Sunday plenary talks by discussing the ethical implications and commercial uses of genetic testing in psychiatry. Traditionally, genetic testing has been used to test for causal genetic and chromosomal lesions, such as the gene mutation that causes Huntington's disease. There are also more recently developed tests for genetic risk factors that are not directly causal, such as ApoE for late-onset Alzheimer's disease and CNVs in schizophrenia, McMahon said.
A number of these tests, such as single nucleotide arrays and sequencing ("personal genomes"), are now available direct to the consumer (DTC) through the Internet. Many of the companies selling these tests have come under fire for their over-reaching claims of the health information such tests can provide.
"Consumer genetic testing is, to a certain extent, here," said McMahon. "We in psychiatry now need to keep it in line with a lot of the uncertainties that come out of the genetic studies that we've pursued," he added.
There are some advantages to DTC testing, he said, such as respecting patients' desires to be informed about their genetic makeup, equalizing data between clinicians and patients, and increasing awareness of genetics among the public. The disadvantages include the potential for exploitation due to a company's profit motive conflicting with a consumer's best interest; the loss of context when health risks are judged apart from factors such as family history and lifestyle; and misinformation about the true specificity and predictive value of the genetic information. Proper counseling before testing and during reporting of results is important, emphasized McMahon, as is follow-up afterwards.—Allison A. Curley.
This is Part 1 of a two-part series. See Part 2.