In this study, we identified abnormalities of GR mRNA and protein expression in bipolar disorder and schizophrenia in the lateral OFC. These abnormalities particularly implicated the GR-1B mRNA transcript variant in bipolar disorder, the GR-1H mRNA transcript variant in schizophrenia, and the GRα-D1 protein isoform in both bipolar disorder and schizophrenia. Interestingly, the transcript-specific patterns of GR mRNA dysregulation within the lateral OFC in bipolar disorder and schizophrenia differed from the more generalised GR mRNA dysregulation identified in the DLPFC in these disorders .
In bipolar disorder, decreased expression of the GR-1B mRNA transcript variant was identified in the lateral OFC, while no changes were observed in levels of pan GR, GR-1C or GR-1H mRNA transcripts. This pattern contrasts with the dysregulation seen in the DLPFC, which was characterised by decreased GR-1C and GR-1H mRNA transcripts in bipolar disorder, a subtle decrease in pan GR mRNA, and no change in GR-1B mRNA . These findings suggest that distinct dysregulation of GR mRNA expression occurs in the lateral OFC and DLPFC in bipolar disorder, and implicate the transcriptional regulatory mechanisms governing GR-1B mRNA expression in the lateral OFC in bipolar disorder. GR-1B mRNA transcript expression is driven by a unique upstream promoter region , and can be influenced by sequence variation in the NR3C1 promoter region in the human DLPFC . Furthermore, GR mRNA transcript variants, including GR-1B, are likely to be also regulated by tissue-specific transcription factors, which mediate tissue-specific GR action . It is possible that these regulatory mechanisms, and/or others yet to be defined, are involved in dysregulation of GR-1B mRNA in bipolar disorder. GR promoter methylation may also play a role, since GR promoter hyper-methylation has been associated with decreased GR-1B mRNA expression in the hippocampus of child abuse sufferers . Greater variation in GR-1B mRNA levels was observed in bipolar disorder cases than schizophrenia cases in the lateral OFC. This was not a gender effect, since males and females displayed equivalent variability in GR-1B mRNA expression in bipolar disorder. However, it may have been due in part to the influence of age of illness onset on GR-1B expression in bipolar disorder, since more cases with later age of onset are present in the bipolar disorder group than the schizophrenia group, and decreased GR-1B mRNA correlated with later age of onset. However, the effect of this illness parameter on GR-1B expression is difficult to interpret, since earlier age of onset has been linked to increased illness severity , but is associated with higher GR-1B mRNA levels among bipolar disorder cases, more reminiscent of normal controls. Although GR-1B expression (which represents approximately 32% of total measured GR mRNA) was decreased in bipolar disorder, there was no difference in pan GR mRNA levels between bipolar disorder cases and controls. This may arise because other GR mRNA transcripts including GR-1C (which represents approximately 66% of total measured GR mRNA) may have diluted this diagnostic effect, despite themselves being unchanged in bipolar disorder.
In schizophrenia, a significant decrease in levels of the GR-1H mRNA transcript variant, along with a trend towards a decrease in GR-1B mRNA expression, were observed in the lateral OFC. As observed for bipolar disorder, these mRNA abnormalities in schizophrenia were more circumscribed in the lateral OFC than in the DLPFC, where decreases in pan GR mRNA and all transcript variants (GR-1B, GR-1C and GR-1H) were observed . The GR-1H transcript includes exon 1H, the GR alternative first exon which is located most proximal to exon 2. The regulation of GR-1H expression by transcription factors in its promoter region, and the function of GR-1H, have not been characterised. Furthermore, GR-1H mRNA represents only a small fraction (approximately 1.2%) of total measured GR mRNA in this study, as in other studies . As a result, the selective GR-1H mRNA deficits which we observe may have limited impact on GR signalling in the lateral OFC in schizophrenia.
The effects of antidepressant use in general, and fluoxetine use in particular, were explored in this study, since previous work has showed selective effects of fluoxetine on total GR and GR-1F mRNA expression in rodent hippocampus . In our study, GR-1C mRNA expression was increased in fluoxetine users relative to non-users. The direction of this change is consistent with the previous study. However, unlike previously reported, we observed no effects of fluoxetine on GR-1F mRNA expression in the OFC, suggesting that antidepressant effects may vary between species and/or brain regions.
The primary GRα protein abnormality that we observed was increased abundance of a truncated GRα isoform, putative GRα-D1, in both bipolar disorder and schizophrenia in the OFC. This same increase in GRα-D1, of a similar magnitude, was seen in bipolar disorder and schizophrenia cases in the DLPFC . In vitro experiments have revealed that the abundance of the GRα-D1 isoform is determined not only by mRNA transcript levels, but also by post-transcriptional mechanisms [46, 54]. Consistent with these findings, over-expression of the GRα-D1 isoform was observed in both the DLPFC and lateral OFC in bipolar disorder and schizophrenia, despite divergent patterns of GR mRNA dysregulation in both regions. The absence of consistent correlations between GR mRNA and GRα protein measures in this study also suggests post-transcriptional regulation of GRα protein abundance. GRα-D1 has been previously shown to function as a transcription factor at glucocorticoid response elements , and to activate and repress the transcription of numerous target genes . As a result, upregulation of GRα-D1 has the potential to influence diverse aspects of cellular function. Since the OFC is an integral component of the brain’s reward circuitry , and is involved in integrating sensory information [4, 6], it is possible that abnormal GR signalling may impact these cognitive functions, particularly during the experience of stress.
In both psychotic illnesses, a lesser involvement of the lateral OFC than the DLPFC in GR mRNA deficits was seen. One possible reason for this observation may relate to the experience of stress. It is plausible that more widespread GR mRNA dysregulation in the DLPFC than the lateral OFC arises due to chronic illness-induced stress in both illnesses. Stress has been shown to down-regulate GR mRNA expression [55–58]. The DLPFC may be more sensitive to this effect, having stronger connections than the lateral OFC (BA11L) with other stress-sensitive regions which are involved in regulating HPA axis activity, such as the hippocampus and hypothalamus [3, 4, 6, 9, 10]. To explore this possibility we examined the relationship between GR mRNA expression and suicide, which may be associated with stressful experiences prior to death. We observed an influence of suicide on expression of multiple GR mRNA transcript variants in the DLPFC , but this influence was limited to the GR-1F and GR-1H mRNA transcripts in the lateral OFC, suggesting that the lateral OFC may be less sensitive than the DLPFC to the (stressful) effects of suicide. In all cases, GR mRNA or GRα protein changes in individuals with suicide were in the opposite direction to diagnostic differences, and therefore were not driving diagnostic group differences. Taking into account the divergent GR mRNA and protein findings in this study, and in previous work in the DLPFC [40, 41], it is likely that multiple processes are involved in GR dysregulation in the frontal cortex in psychotic illness, potentially representing a mix of primary, consequential and/or compensatory changes. Some of these changes may be anatomically specific, whereas others may be more ubiquitous.
Analysis of GR mRNA expression in the lateral OFC, in the context of sequence variation in the NR3C1 gene, did not reveal any relationship of NR3C1 genotype to lateral OFC gene expression. This finding is in contrast to our previous observations in the DLPFC, in which the rs10052957 (Tth111l) and rs6190 (R23K) SNPs impacted expression of the GR-1B and GR-1C mRNA transcript variants respectively . Such polymorphisms may, therefore, impact NR3C1 gene expression in an anatomically specific manner. This effect could be mediated by brain region-specific transcription factors, which may influence regional patterns of GR mRNA transcript variant expression . Such DLPFC and lateral OFC region-specific transcription factors could differentially interact with NR3C1 polymorphisms, manifesting effects of genotype in the DLPFC but not the lateral OFC. Alternatively, the effect of genotype on GR gene expression may represent a gene × environment interaction, which is not evident in the lateral OFC because this region may be less susceptible than the DLPFC to the effects of stress or other environmental influences. The effects of NR3C1 genotype on GRα protein which was previously observed in the DLPFC  were not seen in the lateral OFC in this study. The mechanisms by which genotype may impact GRα protein isoform levels in a brain region-specific manner remain to be elucidated.
In this study, our observations of both similarities (at the GRα protein level) and differences (at the GR mRNA level) between schizophrenia and bipolar disorder are consistent with the similarities and differences between the two illnesses more generally. Bipolar disorder and schizophrenia share similar psychotic symptoms, genetic and environmental risk factors, HPA axis abnormalities and some similar neuropathological changes [19, 23–25, 60, 61]. However, the two illnesses differ from each other in the extent of their affective symptoms, cognitive disturbances and in other neuropathological changes [62–66]. Our findings support the notion that a complex relationship exists between schizophrenia and bipolar disorder at a neurobiological level.
Overall, we identified abnormal GR mRNA and GRα protein expression in the lateral OFC in schizophrenia and bipolar disorder. Depending on the functional properties of the GRα-D1 isoform, these changes have the capacity to impact cellular stress responses of neurons within the lateral OFC. It has been shown that the stress hormone cortisol, acting through GR in other brain regions, can impact glutamatergic and GABAergic neurotransmitter signalling [67–72], and in excess can cause neuronal loss and impair brain function [73–75]. It is plausible, therefore, that high cortisol levels in some individuals with schizophrenia and bipolar disorder may act via glutamatergic and GABAergic mechanisms to contribute to abnormalities (structural and functional) in the lateral OFC network, in a process either mediated by, or resulting in, GR mRNA and GRα protein dysregulation in the lateral OFC. The possible roles of stress and GR dysregulation in the pathophysiology of schizophrenia and bipolar disorder, potentially through interaction with the glutamatergic and GABAergic neurotransmitter systems in the prefrontal cortex, warrant further study.