5 Apr 2017
by Lesley McCollum
The GABA neurotransmitter continues to pose a challenge for schizophrenia researchers—robust findings in postmortem tissue have fingered GABA abnormalities in schizophrenia; these data also have compelling support from animal models. But imaging studies in patients don’t seem to line up. The lack of cohesion between findings in postmortem tissue and imaging was evident in a Sunday morning symposium at the International Congress on Schizophrenia Research annual meeting in San Diego, where researchers brought published and new data to the table to continue building the case for GABA’s role in the disorder.
Brad Ruzicka of McLean Hospital at Harvard Medical School highlighted published data on alterations in GAD1—an enzyme involved in the synthesis of GABA—from gene expression and copy number variation studies, as well as alterations in DNA methylation of GAD1-related genes in people with schizophrenia. These studies, conducted using impressive precision with laser capture microdissection to examine specific cells, parsed alterations present in subregions of the hippocampus, specifically CA3/2, revealing circuit-specific alterations.
New imaging studies appear to be continuing the trend of discouraging GABA results in patients, however. Gemma Modinos of King’s College London reported unpublished results from a new study using magnetic resonance spectroscopy (MRS) to measure GABA levels in the dorsomedial frontal cortex of 20 clinical high-risk (CHR) patients compared with 21 healthy controls, which turned up negative. She also presented the results of a new meta-analysis by Alice Egerton and colleagues of schizophrenia patients, including three CHR studies, that were inconclusive.
But as Modinos noted, the absence of evidence is not evidence of absence. She highlighted the striking variation among studies in voxel placement—the three-dimensional region where MRS was measured—even for studies targeting the medial prefrontal cortex (mPFC), indicating a potential source of conflicting human findings. In a comment after the session, Lawrence Kegeles of Columbia University and New York State Psychiatric Institute emphasized the point that voxel placement may be critical to the results of a study. He suggested that sampling more areas using multivoxel placement may be a way forward to better define GABA abnormalities.
Standing in for Hilleke Hulshoff Pol of the University Medical Center Utrecht, the Netherlands, Modinos also presented the findings of a published imaging study that did find reduced GABA levels in the mPFC of schizophrenia patients (see SRF news story). This study was conducted using ultra-high strength 7T MRS with high resolution (most studies use 3T). Philip McGuire of King’s College London, discussant for the session, noted the larger differences identified in the 7T study compared with past 3T studies, and suggested this may be what’s needed for reliably measuring GABA abnormalities in schizophrenia. Patients with lower IQs had higher GABA levels in the published study, but new data in a follow-up study didn’t find a connection between IQ and GABA levels in healthy people.
The difficulty with measuring GABA in humans was a common thread through the clinical talks of the symposium. McGuire commented that the extreme differences between the level of specificity of Ruzicka’s findings are a world away from the relatively huge voxels used in imaging, which he added was “sobering to hear.”
But not all GABA imaging data were so dismal. Modinos also presented new research linking hippocampal perfusion with dorsomedial prefrontal cortical levels of GABA in CHR patients. The levels of GABA were higher in CHR patients who later developed psychosis. Hyperperfusion is a well-replicated imaging finding in schizophrenia, but no clear basis for it has yet been discovered. McGuire noted that this “striking” relationship between GABA and perfusion in CHR patients provides some of the first human evidence linking hyperperfusion with the neurotransmitter, a potentially important finding for understanding the role that both play in the pathophysiology of schizophrenia.
Encouraging advances were also presented by Anthony Grace of the University of Pittsburgh, who summarized compelling evidence on the effects of stress in the MAM (methylazoxymethanol acetate) animal model of schizophrenia and normal rats, elegantly linking genetic predisposition with the impact of environmental exposures. Findings published since the last meeting (see SRF news) reveal that normal rats with a lesion of the mPFC have a similar response to stress as MAM rats, suggesting that stress combined with a predisposition to the effects of stress—caused by dysfunction of the mPFC and subsequent GABAergic signaling—may underlie susceptibility to schizophrenia.