Researchers Find No Clear Paths on Road to Unraveling Schizophrenia

Once considered largely genetically based, this complex disease may have many points of origin


Twenty years ago, many investigators believed that genetics held the key to understanding schizophrenia, an etiologically heterogeneous disease.1 It seemed only a matter of time before the power of genetic analysis could be brought to bear on this malady, resulting in better drug leads and better ways toward prevention.2

So far, genetic advances have been few. "There was really a misjudgment on the part of some in the field," says Kenneth Kendler, professor of psychiatry and human genetics at Virginia Commonwealth University. Today, the disease's other potential causes, working in tandem with schizophrenia's genetic component, are receiving considerable attention as well.

Courtesy of Robert H. Yolken

The Evidence Is In: Cumulative percentage of Herpes Simplex Virus -Type 2 antibody levels, for cases and controls, showing 75th and 90th deciles for control sample. This 40-year perspective study presented evidence of a correlation between schizophrenia pathogenesis and herpes simplex type II virus. (Archives of Genetic Psychiatry, 5[11]:1032-7, 2001)

The Power of Genetics

In the late 1980s, the National Institute of Mental Health (NIMH) started an initiative aimed at tackling the genetic components of several complex diseases. By pooling the data-collection efforts of many institutions and investigators, researchers envisioned creating a national resource of data and materials for genetic studies that could be shared with the broader scientific community. The initiative's initial focuses: schizophrenia, bipolar disorder, and late-onset Alzheimer disease (AD).

Many who participated had preconceived notions about the sample sizes required to mount the necessary statistical gene-finding power. Steven Moldin, chief of the NIMH's Genetics Research Branch, says that researchers expected 100 families would do the trick. As it turned out, that number was off by an order of magnitude.

Some investigators presumed they would find a straightforward relationship between gene and disease, but they learned that with schizophrenia, as with other complex diseases, simple Mendelian genetics did not apply. Moreover, researchers underestimated the number of loci across the genome that could be involved and the degree of incomplete penetrance—when those carrying susceptibility genes fail to manifest the actual illness.

A new version of the NIMH initiative, expanded to include autism, depression, and obsessive-compulsive disorder, continues today. Within a few years, Moldin hopes to have sample sizes of up to 1,000 families. Six years ago, the institute began providing scientists with materials related to schizophrenia, bipolar disorder, and AD. Scientists, who get the data through a grant or by paying an access fee, are expected to add their results based on that data to the repository. The NIMH has distributed 20,000 DNA samples and more than 200 cell lines to 76 groups worldwide, hailing from academia, biotechnology companies, and Big Pharma.

Outside of some success with AD, namely the identification of a risk factor called ApoE, researchers have not confirmed any susceptibility loci for schizophrenia or bipolar disorder, though studies have suggested several potential sites. "It's been a long road and there have been false starts along the way," says Moldin. Complicating matters, though perhaps for the better, are the several regions of susceptibility overlap between bipolar disorder and schizophrenia, including locations on chromosomes 10, 13, 18, and 22. "It's a very interesting phenomenon because it might provide cross-validation of genetic investigation disorders," explains Miron Baron, a professor of psychiatry at Columbia University who studies both diseases. "Some commonality in a particular subset of these individuals may shed light on particular regions that point to susceptibility for a more general spectrum of psychosis." It remains unclear as to whether schizophrenia and bipolar disorder are homogeneous genotypes with pleiotropic effects, or a single phenotype resulting from the combined effects of multiple genes, or something in between.

Beyond Genetics

Even as geneticists continue to work on improved methodologies,3 they are confronting only half the problem. Monozygotic twin studies suggest that schizophrenia's pathogenic roots are about 50% genetic, leaving a sizable and equally mysterious environmental component. "The environmental perturbation that's necessary in combination with a particular set of gene mutations to cause the disease may be as varied as the genes that provide the susceptibility," comments neurobiologist Pat Levitt, a schizophrenia researcher at the University of Pittsburgh.

Scientists and clinicians have known for years about the possible connection between schizophrenia and, for example, seasonality of birth (more patients with schizophrenia are born in winter months) and perinatal complications. Demonstrating causality, however, has been difficult. Moreover, some or all of these factors may be little more than statistical artifacts.

Research exploring a possible viral contributor, probably occurring during fetal development, has been particularly controversial. One recently concluded 40-year prospective study presented evidence of a correlation between schizophrenia pathogenesis and herpes simplex type II virus.4 The results were based on serological antibody evidence in blood samples taken from pregnant women in the late 1950s and early 1960s. Now decades later, after tracking down those women's children, investigators found that those subjects whose mothers had serological evidence of herpes simplex type II infection had a higher incidence of schizophrenia.

According to study investigator Robert H. Yolken, a neurovirologist at Johns Hopkins Children's Center, such studies, though difficult and time-consuming, provide the best way to examine the pathogenic consequences of prenatal infection. A host of other viruses have been implicated as well, including congenital rubella, borna virus, influenza and other respiratory viruses, with varying degrees of support.

Courtesy of P. Levitt, M. Sabatini, and S. Levey

cDNA Microarray hybridized with two different brain samples.

"Trying to prove cause and effect is extremely difficult," says Yolken, who, along with colleague E. Fuller Torrey of the National Alliance for the Mentally Ill, contends that schizophrenia's genetic component is actually closer to 30%.5 The best hope, he suggests, is to modulate the disease by targeting the viral infection in question. As part of a planned study, Yolken intends to prevent and treat infections in pregnant women—but the study's results wouldn't be known for another 40 years.

Yolken and his colleagues continue seeking the mechanisms of viral infection that might provide a causal link to infection. One hypothesis: certain viruses activate genetic elements called endogenous retroviruses, viral sequences that were integrated into the human genomic germline 35 or 40 million years ago. In findings reported last year, they showed that about one-third of patients with schizophrenia (in the study) have endogenous retroviral sequences upregulated in their spinal fluid during the disease's acute phase.6 After treatment, these levels appear to decrease. As with herpes simplex, Yolken wants to treat viral agents, namely exogenous retroviruses, which are known to upregulate these retroviruses.

Skeptics Abound

"There is limited evidence in favor of a viral etiology," comments Moldin. "The results do not replicate across all samples." But even if studies like Yolken's do not reveal causal, treatment-ready determinants of the disease, they may help researchers understand how the disease manifests itself. Epidemiologist Ming Tsuang, Harvard School of Public Health, is looking for clues to disease onset at a much later stage of pathogenesis. Ideally, Tsuang wants to find neurobiological indicators analogous to high blood-glucose levels in patients with diabetes, or high cholesterol levels in patients with hypertension. Such an indicator, sorely missing in the treatment of mental illness, may lead to ways that will help prevent or slow disease progression.

Regardless of the outcome of such studies, understanding the genetic underpinnings will be essential. "Once we have a firm grip on the genetics, then it will be easier to study the environmental influences," comments Baron. By following up multiple groups of patients with known genetic markers, scientists could incorporate environmental data and see which combinations lead to schizophrenia.

Levitt and his Pittsburgh colleagues plan to take this sort of approach in mice using the newest methodological tool for gene analysis, microarrays. They hope to make murine models that possess some of the suspected schizophrenia susceptibility genes, then use microarrays to analyze gene expression, and examine how particular environmental perturbations or infections lead to severely altered function in the context of the given mutation.

In October 2000, Levitt's group used gene microarrays to examine alterations in transcript expression associated with a neuropsychiatric disorder.7 In analyzing postmortem brain tissue of patients with schizophrenia, they found consistent changes in genes that encode proteins involved in neurotransmitter release. They also found that the same defective synaptic function in different patients had different gene expression patterns. "So the argument goes that schizophrenia ... may be heterogeneous because the combination of mutations and environmental effects may generate an overlapping but distinct combination of altered expression patterns," explains Levitt. Follow-up work implicated one particular candidate gene; called RGS4 (regulator of G-protein signaling), it was the most altered gene among the samples studied.

But the brain confounds even the most advanced analytical tool. Drawbacks to microarray studies include an inability to examine the dynamics of gene expression over time. "You're inferring a lifetime of dysfunction based on a static pattern of altered gene expression," comments Levitt. Sensitivity is also a problem; since the brain is such a complex organ composed of nonuniform cell types, getting a strong gene expression signal from a given cell population can be difficult. Microarray analysis of tumor tissue, for example, is much more straightforward.

So far, nothing has proven straightforward about unraveling schizophrenia—Johns Hopkins researchers recently announced that they found no gene linked to schizophrenia on chromosome 1q, dashing previous hopes about the possible locus.8 Says Kendler, "It's a tough problem. Anybody in this field expecting quick reaction, quick answers—[it is] not to be."

Eugene Russo ( is a contributing editor.

1. S. Bunk, "Research: Researchers pry into schizophrenia's stubborn secrets," The Scientist, 13[18]:14, Sept. 13, 1999.

2. N.S. Halim, "Gene hunters' next challenge," The Scientist, 14[15]:28, July 24, 2000.

3. L. Pray, "The promise that haplotypes hold," The Scientist, 15[23]:21, Nov. 26, 2001.

4. S.L. Buka et al., "Maternal infections and subsequent psychosis among offspring," Archives of General Psychiatry, 58:1032-7, November 2001.

5. E.F. Torrey, "Are we overestimating the genetic contribution to schizophrenia?" Schizophrenia Bulletin, 18:159-70, 1992.

6. H. Karlsson et al., "Retroviral RNA identified in the cerebrospinal fluids and brains of individuals with schizophrenia," Proceedings of the National Academy of Sciences, 98:4634-9, 2001.

7. K. Mirnics et al., "Molecular characterization of schizophrenia viewed by microarray analysis of gene expression in prefrontal cortex," Neuron, 28:53-67, 2000.

8. D.F. Levinson et al., "No major schizophrenia locus detected on chromosome 1q in a large multicenter sample," Science, 296:739-41, April 26, 2002.