4 May 2012. At the 2012 Schizophrenia International Research Society meeting in Florence, Italy, the Tuesday, 17 April, plenary talks focused on roles for the immune system in schizophrenia. The packed audience, combined with several related symposia, suggested that the topic has moved beyond fringe status in research attention. The idea that the immune system somehow contributes to the development of schizophrenia stems largely from epidemiology, which has consistently revealed associations between the disorder and risk factors that can be seen through an immunological lens, such as winter birth, urban birth, and maternal infection during pregnancy. Several talks highlighted the importance of verifying findings with immunological markers such as antibodies or inflammation-related molecules like cytokines in blood samples, which might also help shed light on the immune system’s actions in the brain.
Though maternal infection during pregnancy is a prominent example of an immune-related risk factor for schizophrenia, Preben Mortensen of Aarhus University, Denmark, cited recent evidence from his group that this risk may be part of a more general familial susceptibility to infection itself, as infections in fathers and hospitalizations due to infection also contributed to schizophrenia risk in a recent study from his group (Nielsen et al., 2011). This doesn’t rule out a specific effect of maternal infection, however, and in an attempt to understand how that might happen, he described a candidate gene approach in which infection of a pregnant mother with herpes simplex virus 2 (HSV-2)—verified by blood samples—was considered in combination with single nucleotide polymorphisms (SNPs) in genes for N-methyl-D-aspartate (NMDA) receptors, which may be underactive in schizophrenia. This revealed that HSV-2 infection during pregnancy modulated the risk conferred by specific SNPs (Demontis et al., 2011), and genomewide analysis of SNPs and how their risk is modulated by maternal infection with cytomegalovirus is now underway.
Whether the infectious agents themselves somehow increase risk, or the body’s reaction to them does, is a major question for research into the immune system’s involvement in schizophrenia. Noting that physical illness can trigger or worsen psychiatric symptoms, Barbara Sperner-Unterweger of the Medical University Innsbruck, Austria, focused on the body’s own immune response. She reviewed evidence for increased cytokines in schizophrenia and depression, and noted that specific molecules may mark acute psychosis, whereas others reflect an underlying trait (Miller et al., 2011). Sperner-Unterweger described how cytokines have consequences in the brain by influencing the kynurenine pathway that modulates neurotransmitter signaling, but making definitive links is complicated by the fact that cytokines also vary according to a person’s health.
The long arm of maternal infection
Alan Brown of Columbia University, New York, detailed his work exploring a connection between maternal infection and schizophrenia. Verifying infection through archived samples of maternal blood available in the Childhood Health Development Study (CHDS) cohort, he has found that having the flu during the first half of pregnancy or infection with Toxoplasma gondii (the parasite residing in the cat’s litter box), increases risk for schizophrenia in offspring (Brown et al., 2004 and Brown et al., 2005). These exposures are also associated with poorer executive function in adults with schizophrenia (Brown et al., 2009), and Brown described new work in which worse performance on several cognitive measures, including working memory, was associated with low birth weight among schizophrenia cases. Though birth weight might seem a coarse measure, Brown said it could serve as a proxy for disruptions to brain development caused by prenatal infection, which could then be reflected by neurocognitive measures in adulthood.
The risk associated with maternal infection may not be specific for schizophrenia. In a symposium on Sunday, Brown presented unpublished work, also based on the CHDS birth cohort, in which maternal exposure to influenza—based on a physician’s diagnosis rather than on blood samples—produced a fourfold increase in risk for bipolar disorder. He suggested that timing might be important, where infection later in pregnancy tilts risk toward bipolar disorder, and earlier infection tilts toward schizophrenia.
In a symposium later that afternoon, Åsa Blomström of the Karolinska Institute, Stockholm, Sweden, reported the first replication of Brown’s maternal blood toxoplasmosis finding, and extended the risk-associated factors to cytomegalovirus for the first time. Using archived neonatal blood spots from a Danish biobank containing 3.5 million samples, Blomström assessed prenatal exposure to T. gondii, cytomegalovirus (CMV), and herpes simplex viruses (HSV-1 or HSV-2) by measuring maternal antibodies to these pathogens. Higher-than-usual levels of antibodies to T. gondii and CMV were associated with schizophrenia compared to controls, with odds ratios of 2.1 and 2.2, respectively.
Later in life
In a Sunday symposium, Faith Dickerson of Sheppard and Enoch Pratt Hospital, Baltimore, Maryland, examined a role for retroviruses in schizophrenia. Retroviruses are RNA viruses that can infect the nervous system, and our genome contains endogenous retroviruses that embedded themselves into the DNA of our ancient ancestors. There they lie dormant, until an acute infection or immune response reactivates them. Retroviruses have come under suspicion in schizophrenia because they reside in chromosomal regions that overlap with areas implicated in the disorder, and previous research in small schizophrenia samples has found aberrant retrovirus expression in the blood, cerebral spinal fluid, and postmortem brains.
Dickerson presented results from a large study in which she examined the blood in cases of schizophrenia (n = 163 recent onset, n = 268 multi-episode) and controls (n = 235). She reported an increased level of antibodies to two of these, murine leukemia virus (muLV) and Mason-Pfizer monkey virus, in recent-onset cases of schizophrenia, but not in the chronic cases or in the controls. The blood samples of these participants were negative for retrovirus RNA, ruling out an acutely occurring retroviral infection. Dickerson also described a new longitudinal study of mania, with preliminary results finding an increase in antibodies to feline immunodeficiency virus in the blood of people when they were admitted to the hospital for mania symptoms. The increases were still present a few days later, but not six months later; antibodies to muLV were increased at all three time points.
In the same symposium, Mikhail Pletnikov of Johns Hopkins University in Baltimore, Maryland, described research focused on understanding the effect of infectious pathogens on the brain in mouse models, and found that timing may matter. He reported that T. gondii infection of mice in early adolescence (4.5 weeks), but not in adulthood (nine weeks), led to hyperactivity and sensitivity to NMDA receptor blockade, features used for modeling schizophrenia in animals. Pletnikov pointed out that T. gondii has a tyrosine hydroxylase homolog which could influence catecholamine synthesis in the brain. But it remains to be seen whether these altered mouse behaviors reflect direct action of the parasite in the brain, or stem from immune system activation against the parasite.
In the same symposium, Dorothy Schafer of Harvard University, Boston, Massachusetts, outlined some possibilities of how immune system activation may sculpt the brain’s networks. She focused on the immune system's complement cascade molecules C1q and C3, which have an unexpected role in synaptic pruning (Stevens et al., 2007). Schafer presented mouse data suggesting that these molecules tag weak synapses between retinal ganglion cells and lateral geniculate neurons for removal, and that microglia then recognize these tags and come to engulf the synapses, disposing of them as though they were pathogens. Though there is some evidence for abnormal microglia activation in schizophrenia (Garey, 2010), the specific involvement of the complement cascade in the disorder is so far unclear.
The return of MHC
During the Tuesday plenary session, Peter McGuffin of Kings College, London, U.K., gave a primer on a molecule that has helped revive interest in the immune system in schizophrenia—the major histocompatibility complex (MHC). Over three decades ago, McGuffin’s work into the genetic basis of schizophrenia led him to the region of chromosome 6 containing the MHC—a 3.6 Mb region containing 140 genes, half with immune system function like complement system molecules and cytokines. This lead eventually fizzled out, but MHC is now making a comeback, given the latest GWAS, which finds signals in the region (see SRF related news story). The region is a Pandora’s box of variability, making it difficult to pinpoint the exact location of the association signal, and McGuffin wagered that there is a 50-50 chance that it is immune related.—Michele Solis.