30 May 2012. At the closing session of the Schizophrenia International Research Society meeting on Wednesday afternoon in Florence, Italy, the society's president, Robin Murray of King’s College London, United Kingdom, pronounced, “I just loved this conference.” From the audience, Cynthia Shannon-Weickert of the University of New South Wales, Sydney, Australia, echoed this, calling it “the best SIRS meeting ever.” Not that the participants didn't have suggestions for improvement: one audience member mentioned a lack of emphasis on postmortem work at the meeting, but others countered that it was represented, just not as cohesively in the same sessions as in the past. About half of the audience consisted of graduate students, and one suggested a graduate student-specific activity at SIRS 2014 (also scheduled for Florence) that might nurture the large-scale collaborations that will surely be a mainstay of their future research.
Here, SRF brings you a sampling of other conference news, including gene-environment interactions contributing to schizophrenia risk, evidence for brain changes in the disorder, and mechanistic models of psychosis.
GxE = ?
In his plenary talk on Monday, Robin Murray ventured into the complicated topic of gene-environment interactions in psychosis. Ranking cannabis use as one of the best environmental leads, Murray reviewed the evidence for an interaction with catechol-O-methyltransferase (COMT), a gene whose protein regulates catecholamine levels, including dopamine. “So far, it’s not very encouraging,” he said, noting that the initial positive study in 2005 has not since been reliably replicated. However, a new possibility has turned up in the form of AKT1, a workhorse protein kinase, and single-nucleotide polymorphisms (SNPs) in the gene encoding it have, when combined with cannabis use, been found to increase risk for psychotic disorders (van Winkel et al., 2011). Murray reported a similar increase in risk for psychosis in a case-control study of 491 first-episode psychosis patients and 280 controls. Among those who used cannabis every day, the odds ratio jumped from two to seven. Murray noted that AKT1 lies downstream from both the cannabinoid receptor and the D2 subtype of the dopamine receptor, presenting a plausible mechanism for its involvement in the overactive dopamine signaling associated with psychosis. He cited the need to look toward gene-environment methods honed to study cancer, which can deal with thousands of genes at a time. “We are just beginning to bite into the edges of this,” Murray said.
How people react to their environment may influence their risk for developing a psychotic disorder. Inez Myin-Germeys followed with a plenary talk about the experience sampling method, in which people document their moods multiple times a day, in an effort to help define “reactive phenotypes.” She finds that how a person reacts to stress is associated with the well-known COMT Val158Met allele that influences dopamine levels in the synapse: negative reactions to stress were associated with the Met/Met genotype, but only among those with a psychotic disorder (Collip et al., 2011). Myin-Germeys also reported new data showing that this association between negative responses to stress and the Met/Met genotype was found in a stress task adapted for brain imaging among 14 subjects at genetic risk for psychosis. Subjects with the Met/Met genotype had the lowest increase in dopamine signaling in the prefrontal cortex, compared to the Val/Met or Val/Val genotypes.
Tracking brain changes
In a packed session about progressive brain change in psychosis on Sunday afternoon, Matthew Kempton of King’s College London, United Kingdom, reported his findings from a meta-analysis of antipsychotic-naïve people with schizophrenia. Brain imaging studies turn up evidence for progressive brain tissue loss in the disorder, but the nature of the link remains debated: does the shrinkage drive the disease, or does it reflect use of antipsychotic medications, or lifestyle (see SRF related news story)? To try to get around these confounds, Kempton rounded up 58 studies of antipsychotic-naïve people with schizophrenia, and focused on regions of interest measured in three or more studies. Compared to controls, the schizophrenia subjects showed significantly smaller total brain volumes, as well as gray matter and white matter volumes. More specifically, the caudate, thalamus, and hippocampus were each significantly smaller than in controls, consistent with the idea that reduced brain volumes were related to disease, rather than medication.
Kempton also reported a similar pattern of reductions (compared to controls) in first-episode patients who had been treated with antipsychotics, and in chronically ill patients. The average reduction did not differ among these three groups for total brain, gray matter, or hippocampus; however, the smaller caudate found in the antipsychotic-naïve group was not apparent in the first-episode or chronic group, which might indicate that the striatum could be particularly sensitive to antipsychotic medication. Kempton said that the length of untreated psychosis in these groups would be important to track, as it might influence the magnitude of the changes seen. Also, whether these reflect progressive changes afoot as the illness unfolds would best be explored in a longitudinal study—something Kempton said was just getting underway in a multicenter MRI study of people at high risk for psychosis.
In the same session, Neeltje van Haren of University Medical Center Utrecht, The Netherlands, reviewed her data for progressive brain changes associated with schizophrenia. Subdividing brain volume into its component area and thickness, van Haren has found that cortical thickness decreases over five years in schizophrenia (van Haren et al., 2011), but her preliminary data do not show a concomitant decrease in cortical surface area—though it was abnormally small at baseline. She suggested that surface area may reflect an aberrant neurodevelopmental program that is present before or at illness onset, whereas cortical thinning may emerge at illness onset and after.
The mechanics of psychosis
On Tuesday afternoon, Peter Woodruff of the University of Sheffield, United Kingdom, described his experiments probing the brain basis for auditory hallucinations by using healthy controls who experience these while falling asleep or waking up. Preliminary fMRI data show that the auditory cortex in this group has regions of enhanced sensitivity to sound, and that this might reflect altered top-down modulation of these areas by prefrontal cortex. Woodruff discussed these findings in light of the complex interplay between brain regions that predict and receive sensory input. When the brain erroneously identifies a predicted sound as an actual signal, auditory hallucinations might result (Nazimek et al., 2012).
A session sponsored by the International Congress on Schizophrenia Research (ICOSR) on Tuesday featured a systems neuroscience view. Anthony Grace of University of Pittsburgh, Pennsylvania, described his results from studies of the methyl-azoxymethanol acetate (MAM) model of schizophrenia, in which pregnant rats treated with MAM, a DNA methylating agent, give birth to offspring which exhibit a number of schizophrenia-esque abnormalities, including deficits in pre-pulse inhibition, reversal learning, social interaction, and increased locomotor response to amphetamine, suggestive of overactive dopamine signaling (see Grace lecture at the 2008 Society for Neuroscience meeting). In the brains of these rats, Grace finds a noisier hippocampus, with more spontaneously active neurons than that found in controls; this is accompanied by decreases in a kind of interneuron there, too, which suggests that their ability to dampen activity in the hippocampus is compromised. Because hippocampal hyperactivity trickles through the neural circuit to increase dopamine neuron activity, restoring the inhibitory action of these interneurons might help. Indeed, he finds that a drug that selectively potentiates a subtype of GABA receptor found in the hippocampus settles activity there to control levels, and also tempered the dopamine-dependent hyperlocomotion induced by amphetamine (Gill et al., 2011).
Grace presented new results showing that these MAM-treated rats were more sensitive to stress: they made more distress calls when given a foot shock, and they show abnormally high levels of cortisol, a stress hormone, compared to controls. Because cortisol can damage the hippocampus, Grace suggested that keeping stress levels in check may protect the hippocampus, and, consistent with this, he reported that MAM-rats given diazepam before and during puberty did not exhibit hyperlocomotion induced by amphetamine. He suggested that anything that helps control stress may help ward off transition to psychosis in humans.
Continuing with the hippocampal theme, Carol Tamminga of University of Texas Southwestern Medical School in Dallas presented a model for psychosis based on regional alterations in glutamate signaling within the hippocampus (Tamminga et al., 2012). In schizophrenia, imaging studies find increased baseline activity in the hippocampus, and postmortem studies find decreased levels of the GluN1 subunit of the NMDA receptor in the dentate gyrus subregion of the hippocampus. Tamminga hypothesized that a decrease in glutamate signaling within the dentate gyrus could lead to increased activity within its target, the CA3 subregion, through homeostatic plasticity. This would increase synaptic strength between the dentate gyrus and CA3, and Tamminga reported preliminary evidence for this in the form of increased BDNF mRNA and more dendritic spines in CA3 neurons in postmortem tissue in schizophrenia. An increase in sensitivity to input could render CA3, a region responsible for completing memories based on isolated features of that memory, prone to making spurious associations and generating the false content of psychosis. To test this model, Tamminga is developing transgenic mice with a “molecular lesion” of decreased GluN1 levels specific to the dentate gyrus.—Michele Solis.