ICOSR 2011—The Promise of Genetic Neuropathology for Schizophrenia
As part of our ongoing coverage of the 2011 International Congress on Schizophrenia Research (ICOSR), held 2-6 April in Colorado Springs, Colorado, we bring you session summaries from some of the Young Investigator travel award winners. For this report, we thank Bart P.F. Rutten, of the Maastricht University Medical Centre in The Netherlands.
On Sunday afternoon, 4 April 2011, at the International Congress on Schizophrenia Research, Paul Harrison, Oxford University, Oxford, U.K., chaired the session, "Neuropathological contributions to the etiological understanding of schizophrenia: recent and future studies." He started by pointing out that human postmortem studies are crucial for elucidating the pathophysiology of schizophrenia since some key processes may not occur in other tissues, or species. The studies presented in this session provided excellent examples of novel genetic and epigenetic designs for postmortem brain studies that could elucidate biologic mechanisms involved in the etiology and pathophysiology of schizophrenia.
In the session’s first presentation, entitled “Molecular and cellular abnormalities in cortical circuits in schizophrenia: what do we know and where should we go?”, David Lewis, University of Pittsburgh, Pennsylvania, provided an overview on his group's findings on altered expression of GABAergic genes in frontal cortex, and discussed possible relations with γ oscillations. His postmortem work showed that GAD67 mRNA levels were lower in dorsolateral prefrontal cortex (DLPFC) of schizophrenia patients compared to controls. He extended the robustness of these findings by showing that these differences were not accounted for by variations in RNA measurement themselves, by psychotropic medications or substance use, or by predictors or measures of disease severity. Thus, these results suggest that decreased GAD67 mRNA levels in DLPFC may be a conserved feature of the disease process of schizophrenia.
As a logical next step in high-quality research, he addressed whether the protein levels of GAD67 in DLPFC of schizophrenia patients are decreased as well. When comparing schizophrenia patients to controls, his group found that the terminals of parvalbumin-immunoreactive cells (putative basket cells) displayed lower levels of GAD67 protein in layer 3 of the DLPFC, and that this reduction was much greater than in total cortical tissue, consistent with a cell type-specific deficit in GAD67 protein. Analyses of mRNA of two subunits of the GABAA receptor (i.e., the α1 subunit and β2 subunit) that co-assemble to form functional receptors were also lower in layer 3. In particular, GABAA α1 subunit mRNA was markedly lower in layer 3 pyramidal cells, but was unchanged in GABA neurons. These latter findings nicely illustrate that measures of gene products (here, GABA-related genes) in specific cell types may reveal substantially larger and more consistent differences in schizophrenia than previously found. The findings of GABAergic alterations in DLPFC may be functionally related to the observed impairments in prefrontal γ oscillations during cognitive control tasks in subjects with chronic schizophrenia, as well as in antipsychotic naive schizophrenia patients.
Thomas Hyde, National Institute of Mental Health, Bethesda, Maryland, presented his group's work on “Abnormalities in GABA signaling in schizophrenia.” He started his presentation showing that a SNP (rs 3749034) in the promoter region of GAD1 was associated with lower expression in DLPFC of schizophrenia patients, but not in that of controls. At the RNA level, GAD1 expression can form the truncated transcript GAD25, and the full-length transcript GAD67. He presented further evidence of developmental alterations in GAD1 expression with predominance of GAD25 expression in fetal brain and low fetal levels of GAD67, whereas GAD67 expression was high in the second decade of life. Such trajectories of GAD67 over the lifespan show interesting variations between brain regions. A rather stable level of expression was found for the hippocampus, while the DLPFC displayed quite dynamic changes in expression, with a low level of expression in the fetal period that increased postnatally. On the other hand, lifespan trajectories of GAD25 expression were quite similar in hippocampus and DLPFC, with relatively high levels of expression during the fetal period and lower levels postnatally.
These developmental trajectories of GAD25 and GAD67 levels may have functional impact, especially given that GABA exerts an excitatory function early in development and becomes inhibitory later in development. It has been proposed that molecules such as NKCC1 and KCC2 mediate this switch. Interestingly, patients with schizophrenia have been documented to show lower KCC2 expression in the hippocampus, particularly those individuals that carry a specific SNP in the GAD1 gene. Furthermore, KCC2 is known to have multiple transcripts in the brain (11 in human brain). An increased level of expression has been observed in schizophrenia cases for one of these transcripts—the transcript related to exon 6—whereas decreased expression of this transcript was observed in patients with bipolar disorder and unipolar depression (as compared to controls). Thus, these results from state-of-the-art expression analyses in postmortem brain suggest that, in schizophrenia cases, allelic variation in the GAD1 promoter SNP is associated with an “immature” expression pattern of hippocampal GABA signaling elements.
In his presentation, Paul Harrison (also on behalf of Sharon Eastwood, Oxford University, Oxford, U.K.) presented general aspects on the field of genetic neuropathology and discussed recent findings on his DISC1 studies in his presentation entitled “Contributions of postmortem studies to the understanding of DISC1 and its role in psychosis; an illustration of genetic neuropathology.” After illustrating that the brain’s anatomy and function can be investigated by comparing brains on the basis of genotype, Harrison pointed out that many genes have human- and brain-specific regulation, and that the study of the human brain is thus essential for understanding the etiology of human psychiatric illnesses. He focused the rest of the presentation on his group's recent research on the gene disrupted in schizophrenia-1( DISC1) that has been a focus for many neuroscientists and cell biologists over the last few years. Two common SNPs in DISC—Leu607Phe and Ser704Cys—have frequently been associated with psychiatric disorders. Other studies have also shown that DISC1 splice variants are upregulated in schizophrenia, that these splice variants are associated with the risk polymorphisms, and that these risk polymorphisms can also affect expression of binding partners to DISC1.
DISC1 protein is located in centrosomes (among other cellular locations), which function as the microtubule organizing centers via an interaction with PCM1 (pericentriolar material 1); PCM1 has itself also been associated with schizophrenia. Earlier work from the group showed that undifferentiated neuronal (SH-SY5Y) cells in vitro expressing the Phe607 variant of DISC1 had less centrosomal PCM1 than cells expressing the Leu607 variant. The group then examined whether the DISC1 SNPs also affected centrosomal PCM1 in human brains. They found that PCM1 is only seen at the centrosome in glia and not neurons (in human cerebral cortex), but that within glia, DISC1 genotype affected centrosomal PCM1 just as it did in vitro, with both SNPs having an effect. To complete the picture, Harrison and colleagues then confirmed that PCM1 is also centrosomal in glial cells in vitro, and that centrosomal PCM1 is lost when SH-SY5Y cells are differentiated into neurons. The studies provide an example of translational or genetic neuropathology, using postmortem human brains and experimental models iteratively, whilst the results reveal novel aspects about the biology of, and interactions between, DISC1 and PCM1.
In the last presentation of the session, entitled “Epigenomic approaches to the etiology of schizophrenia,” Ruth Pidsley (also on behalf of Jonathan Mill, Institute of Psychiatry, London, U.K.) discussed epigenetics and presented recent data on DNA methylation. Pidsley illustrated the concept of epigenetics, and possible relationships between epigenetics and 1) developmental periods of environmental sensitivity, 2) paternal age, 3) parent-of-origin effects, and 4) monozygotic discordance for psychiatric disorders. She also pointed out that choice of tissue type is important for epigenetic studies, but that epigenetic changes in peripheral tissue types may be detected, and may correlate with intracerebral changes. The field of epigenetics is still young, and one important development is that reference epigenomes are currenly being established, for example, through large initiatives such as the NIH Roadmap Epigenomics Mapping Consortium.
One evidently crucial question is whether (and if so, how) epigenetic profiles in peripheral tissue types correlate with other tissue types. Pidsley presented recent data that between-individual differences in epigenetic profiles were smaller than between-tissue type (blood and several brain regions) differences per individual. Pidsley reviewed earlier work by Jonathan Mill and colleagues showing small, but statistically significant, differences in methylation status in postmortem samples of frontal cortex of patients with major psychotic disorders (schizophrenia and bipolar disorder). They found that gender had a substantial influence on the observed epigenetic differences, and also found a correlation between methylation status and antipsychotic medication.
Recent unpublished work of Pidsley, Dempster, and colleagues on DNA from blood lymphocytes of 22 monozygotic twin pairs discordant for major psychotic disorder further identified several loci with differential DNA methylation. Another line of their research observed that methylation in the last exon of IGF2, an important growth factor, in DNA extracted from brain tissue correlated positively with brain weight in males (higher brain weight correlated with higher methylation percentage). However, IGF2 methylation in the imprinting control region between the IGF2 and H19 genes showed an inverse correlation with cerebellum weight. It was also found that certain SNPs in IGF2 showed a parent-of-origin effect on cerebellum weight. Thus, these findings suggest a role for epigenetic regulation of IGF2 in brain and cerebellum weight. In conclusion, epigenomic approaches in postmortem human brain tissue seem viable and may produce relevant findings.—Bart P.F. Rutten.