Defining the disturbance in cortical glutamate and GABA function in psychosis and its origins and consequences

Introduction

Interest in the role of cortical glutamate and GABA neurochemistry in the pathogenesis of schizophrenia dates back almost half a century [1]. This has been greatly reinforced in the last decade by genome wide association studies (GWAS) which have increasingly revealed that many risk genes cluster around the pre and post-synaptic elements of glutamate and GABA synapses [2–4]. Imbalance between excitatory glutamate and inhibitory GABA function has been implicated in schizophrenia and neurodevelopmental disorders and in processes such as neurotoxicity and neuro-inflammation that influence onset of symptoms, treatment response and long-term outcome [5–8]. The onset of psychosis has long been attributed to an exaggeration or neurotoxic pathology of neurodevelopmental cortical synaptic pruning which is most active in late adolescence [9–11]. Indeed it is increasingly clear from longitudinal imaging studies that loss of frontal cortical grey matter occurs close to the time of transition from the prodrome to psychosis, possibly within a few weeks[12].This may involve loss of synapses rather than loss of neurons as has been reported in histological and gene-expression studies in post mortem brain [13, 14]. Here we investigate the role of abnormal cortical glutamate and GABA synaptic function in the clinical pathogenesis of early and established stages of psychosis and whether inflammation and oxidative stress are mediating mechanisms.

Olney and Farber proposed that impaired NMDA-glutamate receptor signalling is a core deficit in schizophrenia that causes dysfunctional glutamate release[15]. The psychotomimetic effects of drugs that impair NMDA neurotransmission (phencyclidine or PCP, and ketamine) appear to be mediated by disinhibition of glutamate release through failure of restraint by inhibitory GABA interneurons [16, 17]. GABA feedback inhibition is driven by NMDA glutamate receptors on GABA interneurons. In rodent models of psychosis, PCP-like drugs ultimately induce regionally specific excitotoxic damage to glutamate neurones, loss of parvalbumin containing GABA interneurons and cognitive impairment [18, 19]. In humans also, post-mortem brain studies consistently report impaired function in parvalbumin and somatostatin containing subtypes of GABA interneurons [20, 21]. Equally, GABA changes could be the primary driver of glutamate changes and are compatible with the long-standing GABA deficiency theory of schizophrenia[22]. The role of GABA and whether disinhibition of glutamate release occurs early in psychosis and eventual synaptic loss underpins chronicity and negative symptoms are key issues for the present study.

There have been many magnetic resonance spectroscopy (MRS) studies of glutamate and GABA content in brain regions in people with schizophrenia. In keeping with the NMDA model, a landmark meta-analysis by Marsman et al reported overall increased MRS glutamine in medial frontal cortex (MFC)/anterior cingulate cortex (ACC) but decreased glutamate in cohorts comparing patients versus controls and that both metabolites decreased with age of the cohorts at faster rate than controls [23]. The extent to which these static measures reflect glutamate release rather than neuronal integrity, synaptic content and turnover is debatable. However, since synaptic glutamate is converted to glutamine by astrocytes, increases in the latter may have been a better index of increased excitatory neurotransmission in early psychosis, albeit less accurately measured at 3T, than glutamate itself.

There are obvious confounds in combining different cohorts and centres, studied at different MR field strengths and there have been no explicit comparisons of MRS glutamate in early vs established psychosis in the same study. Furthermore, effects of age are confounded with cumulative drug exposure. Indeed, a recent mega-analysis reported decrements in MFC/ACC glutamate in proportion to age and to daily antipsychotic drug exposure[24]. We sought to detect evidence of glutamate/glutamine disinhibition in early phase psychosis and loss in the established phase and to identify potential origins in drug treatment, oxidative stress or neuroinflammation, and the consequences for clinical pathogenesis.

The discovery that rare autoimmune disorders with antibodies against NMDA receptors can present as schizophrenia was a remarkable corroboration of the NMDA deficiency theory while also reinforcing interest in the neuroinflammatory hypothesis of schizophrenia [25]. Circulating cytokine concentrations such as Interleukin-6 (IL-6)-6 and C-reactive protein (CRF) are consistently increased across many cytokine screening studies in schizophrenia [26]. Whether peripheral cytokines penetrate the brain to induce a damaging microglial neuroinflammatory state and grey matter loss [27] or exert direct effects on synaptic development, function and survival is debated [28–30]. We tested these ideas by measuring peripheral cytokines and 2 putative indices of central inflammation, MRS glutathione content in ACC and in one centre, PET radioligand binding to the translocator protein (TSPO), a putative marker inflamed microglia [31]. We examined whether the central or peripheral markers mediated or influenced the group differences in MRS glutamate and GABA measures or their association with clinical symptoms.

We aimed to test the over-arching hypothesis that symptoms of schizophrenia arise from phase-dependent dysfunction in cortical glutamate/GABA synapses; an early neurotoxic phase, possibly related to oxidative stress or inflammation, and a late phase of residual damage, the phases relating respectively to acute psychosis and persistence of negative symptoms. We recruited matched samples of Recent-onset patients with non-affective psychosis with less than 12 weeks of treatment and a group with more than 10 years of treatment together with age-appropriate controls across 3 research centres. The variables of primary interest were the 4 MRS metabolites, glutamate, glutamine, glutathione, and GABA in ACC. The same measures in occipital cortex (OC) were part of a separate connectivity workstream which also enabled us to replicate a previous report of low OC GABA and test for a transcortical deficit [32]. We tested pre-specified predictions using analysis of variance to contrast early and late phase differences between controls and patients. We used correlation and mediation analysis to test for the influence of medication-related confounds, mechanistic inflammatory markers and ROS on the 4 MRS variables and their association with positive and negative symptoms and cognitive performance.

Methods

The study received ethical approval from the National Research Ethics Service Committee Northwest – Lancaster UK, reference 14/NW/0298 on 18/06/2014.

Results

Glutamate and GABA: effects of diagnosis and phase

Greater reductions in ACC glutamate were found in Established patients than in the Recent group relative to their age-matched controls, as predicted, and evidenced by a statistically significant Diagnosis x Phase interaction, Q=5.74, N=183, p=.022 (Figure 1A). There were no significant interactions with Site. Follow-up tests found no evidence of elevated glutamate concentration in the Recent onset group (ES = 0.13 (95% CI: -0.32, 0.59), p=.574 but there was large effect size for the reduction in ACC glutamate in the Established group relative to controls, ES = -0.59 (95% CI: -1.14, 0.09), p=.012. ACC glutamine showed no evidence of a Diagnosis x Phase interaction (Figure 1B), Q=0.60, N=180, p=.440.

• Download figure
• Open in new tab

Figure 1. Effects of Diagnosis and Phase on MRS variables and cytokines

Effects of Diagnosis for MRS variables (A to D) and inflammatory cytokines (E and F). Panel F shows robust effect sizes (AKP, a robust version of Cohen’s d) for effects of diagnosis, with 95% CIs. Supplementary Table 5 shows robust effect sizes and CIs for effect of diagnosis on all variables in each group.

In keeping with a generalised GABA deficit in psychosis, average GABA concentrations (ACC & OC) were lower in patients (Figure 1C), Q=8.08, N=155, q=.036, p=.006), with a Medium ES = -0.48 (95% CI: -0.83, -0.16) for the effect of Diagnosis (Figure 1G). There was no evidence for interactions between Diagnosis and either Phase (p=.102), or Voxel (p=.999, as tested by a 3-way ANOVA with ACC-OC difference as the dependent variable).

Correlation analysis of mechanistic markers and confounders

To test the specified mechanistic correlates of glutamate and GABA dysfunction, robust correlations were computed between the 4 ACC MRS variables and potential influencing variables within groups and, using group mean-centred data, across both patient groups. They are summarised in Figure 2 and shown in detailed scatter plots and matrices in Supplementary Figure 1 and Table 8).

• Download figure
• Open in new tab

Figure 2 Correlation matrix for MRS variables, potential mediators, and clinical correlates

Correlation matrix relating 4 MRS dependent variables (top left) for all patients (n=>100) to potential mediators and to clinical correlates. Size and colour intensity indicate magnitude of correlation, red for positive and blue for negative correlations scaled from -0.6 to + 0.6 using group mean centred data. Dots indicates all p values <.05. Correlations are across both groups except where the grey marker ◥ indicates that the correlations differed between groups at p<.01 and was p<.05 only in the Recent onset group. There were no instances where the correlation was greater and p<.05 only in the Established group. Where groups correlations differed but both at p<.05 in the same direction, the overall average is shown.

Glutamate, glutamine and glutathione content were positively intercorrelated (Figure 2 top left) but none correlated with GABA. There was no evidence of the predicted inverse relationship between decreased GABA content and increased glutamate or glutamine content. Glutamate did not show any significant correlations with clinical measures predicted or otherwise. Glutamine inversely correlated with drug exposure and BMI. In contrast, decreased GABA content correlated only with clinical measures: lower values correlated with greater PANSS positive and negative symptoms, across both groups (PANSS positive: r=-.24, p=.014; PANSS negative: r=-.20, p=.040), with lower IQ, in the Recent onset group, r=0.36, p=.025), and poorer social functioning (PSP) across both groups, r=0.19, p=0.051).

In terms of early inflammatory mechanisms, although glutathione was not depleted in the Recent group, the predicted inverse association with increased levels of IL-6 was observed within this group (r=-.30, p=.037). Only glutathione correlated with age, inversely and within and across each phase group. CRP and IL-6 levels were inversely correlated with ACC glutamate across both groups, significantly so for CRP, suggesting a possible mediating mechanism for low ACC glutamate in Established patients. Neither daily nor lifetime drug exposure was associated with lower ACC glutamate or raised CRP and IL6 levels (middle of fig2). However, greater lifetime exposure to antipsychotic drugs was associated with higher BMI which in turn showed its well-known association with increased cytokines (CRP and IL-6). Nevertheless, lifetime drug exposure was associated with lower IL-6 concentrations (r=-.20, p.023), a possible anti-inflammatory effect which might mask illness-related increases in IL6 that could mediate the changes in ACC glutamate.

Mediation analysis of reduced ACC glutamate in Established patients

The correlation analyses (Figure 2) identified several possible influences that might partially or wholly account for the greater ACC glutamate deficit in Established patients. These included effects related to AP medication exposure, but also potential mediation by co-occurring phase differences in circulating inflammatory cytokines (CRPι1 and IL-6ι1). We evaluated these possibilities in a sequence of mediation models in which we expressed the values of MRS and inflammatory variables as deviations from their age-matched control means (adjusted variables denoted by ι1).

There was no evidence that the negative Phase effect on glutamateΔ (deficit greater in the Established group) was mediated by daily and/or lifetime antipsychotic dose (Figure 3A); none of the indirect effect pathways reached p<.1 and the direct (unmediated) effect of Phase on ACC glutamateΔ remained evident (green line, Figure 3A), β =-0.57, p<.023. However, as the correlation analyses indicated a negative correlation between AP-life and ACC glutamine, we also investigated mediation via the effects of AP-life on ACC glutamineΔ (Supplementary Figure 3B). We found weak evidence (p<.1) for partial mediation of the negative Phase effect on ACC glutamateΔ by AP-life via a negative effect on ACC glutamineΔ (β =-0.18, p=.066). However, this effect was offset by an indirect positive effect of Phase on ACC glutamineΔ itself (β=0.19, p=.059), and accounting for these indirect effects strengthened the evidence for the direct effect of Phase on ACC glutamateΔ (β=-0.66, p=.002)

• Download figure
• Open in new tab

Figure 3. Mediation analysis of phase-dependent loss of ACC glutamate

Models A to C: medication-related mediators of Phase effects on ACC glutamate, CRP and IL-6. Model C revealed excess IL-6 in the Established group (p-003 for the Direct effect of Phase) after accounting for medication-related effects. Model D: Phase-related increases in IL-6 did not mediate the phase-related reductions in ACC glutamateΔ (p=.159 for the Indirect Effect). Coefficients are standardised β values. Total Effect = sum of direct and indirect effect of Phase. Indirect pathways are shown in red if p<.05. Direct non-mediated effect of Phase shown by horizontal connector, green if p<.05. Note that in Model D, the Total Effect excludes indirect effects mediated by the covariates.

The total Phase effect on CRPΔ (greater in the Established group; p<.001) was completely mediated by the indirect pathway from Phase via AP-life and BMI (Figure 3B and Supplementary Figure 3C) leaving no direct i.e. unmediated effect (p>.1). However, for IL-6Δ, while AP-life acted via BMI to increase IL-6Δ, this was offset by the unexpected negative effect of AP-life on IL-6Δ itself, as noted in the correlation analysis, which was independent of BMI (Figure 3C red diagonal). Accounting for these opposed indirect effects revealed evidence for an underlying direct effect of Phase on IL-6Δ (β =0.71, p=.003), namely a residual excess of IL-6Δ in the established group over levels in the Recent onset group, possibly indicative of a Phase-dependent effects of illness.

We investigated whether the revealed effect of Phase on IL-6Δ levels might mediate the ACC glutamateΔ Phase-effect. The Phase effect on IL6 was isolated (Figure 3D, β =.63, p=.009) by covarying the opposing indirect effects of AP-life and BMI shown in Figure 3C. By including glutamineΔ in the covariates, we controlled for its previously mentioned indirect influences on the effect of Phase on glutamateΔ. This model provided no evidence to support an indirect (mediating) effect of excess IL-6Δ on Phase-associated ACC glutamateΔ (Figure 3D, p=.159) and the direct effect of Phase remained evident (β= -0.60, p=.006). See Supplementary Section 9 for a full account of the mediation model series).

Discussion

This study tested the hypothesis that excessive glutamate neurotransmission in early psychosis underpins the emergence of positive symptoms and results in loss of synaptic glutamate and the development of persistent negative symptoms in established illness. We confirmed with well-matched samples recruited explicitly from patients with early and established illness, that the Established group had greater ACC glutamate deficits but found no evidence of excess in the Recent group. The greater deficit in the Established group was not attributable to greater drug exposure in the Established group and was not associated with greater negative symptom scores. There was a generalised decrease of GABA content across ACC and OC which did not correlate with glutamate changes but was associated with greater positive and negative symptoms across both groups. We found no evidence of ACC glutathione abnormalities in patients, although some evidence that might indicate a greater depletion with age. In inverse contrast to ACC glutamate, circulating cytokine concentrations of CRP and IL-6 were greater (significantly for CRP) in established illness than controls raising the possibility of a causal inflammatory process.

In the Recent onset patients, there was no increase in glutamate or glutamine in ACC and neither correlated with positive symptoms as we had predicted. Thus, we found no evidence that dysregulation of glutamate occurs in early psychosis or underlies symptoms. Indeed, early reports of increased glutamine in early psychosis [23] have not been replicated in more sensitive studies at 7T [49–52] with a single report of a trend-level association with symptoms. It is possible that the MRS measures are too static or diffuse in origin to detect dysfunctional glutamate release in early schizophrenia. It is also possible but perhaps unlikely that a neurotoxic process underlying the onset of symptoms had resolved so rapidly as to be undetectable by group differences in MRS measures or a greater spread than controls given that half the Recent group were within 7 months of onset. Nevertheless, the existence of a small neurotoxic subset could be hidden in the variability as suggested by some within-patient studies [53, 54].

There was no evidence of ongoing oxidative stress in Recent onset patients; there was no depletion of glutathione nor were levels of circulating cytokines CRP markedly increased as might have been expected had there been inflammatory neurotoxicity undetected by MRS ACC glutamate or glutamine. However, il-6 was elevated relative to controls across groups and ACC glutathione was negatively correlated with il-6 (Figure 2) in the Recent group, a possible hint of early inflammatory central oxidative stress. As previously reported in the Manchester subset, there was no evidence of central microglial inflammation in Recent onset patients as ACC TSPO radioligand binding was not increased in Recent or Established patients [31]. In summary, there was no MRS evidence of abnormal ACC glutamate, glutamine or glutathione content in Recent onset minimally treated patients with schizophrenia and minimal evidence of evidence of systemic or central immune activation.

The decreases in glutamate in the Established group are broadly in keeping with published studies and meta-analyses. In a meta-analysis of 65 studies in medial prefrontal cortex including ACC, Merritt et al reported reductions in patients with schizophrenia and no changes in glutamine[55]. Some meta-analyses report more marked decrements in ACC glutamate in older cohorts as first reported by Marsman et al [23]. In a mega-analysis combining individual data, both daily antipsychotic drug exposure and age were associated with greater reduction in ACC glutamate [24]. However, our Defined Daily Dose measure showed only a minimal inverse association with ACC glutamate only in the Established group (r=-0.24, p=0.06; Supplementary Table 8), and this did not mediate the phase-related reduction in ACC glutamate (Supplementary Figure 9B). Additionally, we observed for the first time that long-term cumulative exposure to antipsychotic drugs had no effect on ACC glutamate (Figure 3A). However, such exposure was associated with reduced ACC glutamine across both phase groups and so an influence of antipsychotic drug exposure on synaptic glutamate function cannot be ruled out.

Reduced ACC glutamate in the Established but not the Recent group is compatible with the proposed accumulation of neurotoxic damage to glutamate synapses in established schizophrenia [15, 56]. This is corroborated by recent PET studies in which the synaptic marker SVA2 was reduced in patients with established illness but not first episode patients [57, 58]. Furthermore, SVA2 correlated with MRS ACC glutamate content in healthy controls and in patients but to a lesser extent because of their constrained range of reduced levels of SV2A binding. The duration of illness in our Established group spans 25 years but duration was not associated with greater loss of ACC glutamate suggesting that the loss is static after 10 years of treatment or stably maintained by an active process. The contribution to pathogenesis is unclear since reduced glutamate did not show the predicted correlation with negative symptoms (table 2) nor was this observed in a metanalysis of 15 studies with age-related decrease in MRS creatine-scaled glutamate [24]. Indeed, the meta-analysis found that against the background of reduced levels, glutamate reliably but weakly correlated with positive symptoms as we also observed (Established r=.21; p=.111).

We evaluated the possibility that previous or current inflammation contributed to the static reduction in ACC glutamate. Circulating CRP and IL-6 concentrations were increased in the Established group and were weakly predictive of reduced ACC glutamate (Figure 2). However, mediation analysis indicated that whereas that late phase increases in CRP were accounted for by associated treatment and BMI differences, the effect of Phase on IL-6 was enhanced after adjustment for treatment confounds. Nevertheless, there was no evidence that IL-6 increases directly mediated the late phase decrease in ACC glutamate (Figure 3D). Indeed, low ACC glutamate was recently associated with polygenic risk for reduced NMDA receptor function shared with treatment resistance albeit in a younger cohort than the Established group [59].

Reduced ACC glutamate correlated at p<.05 with lower glutathione (Figure 2) and although glutathione levels were not reduced in patients, it alone among MRS and cytokine variables, correlated inversely with age within each phase group. This might suggest that excessive levels of ageing-associated reactive oxygen species indexed in the brain by the age-MRS glutathione decline could induce detectable damage to glutamate terminals in Established which are not detectable in Recent onset schizophrenia in keeping with a recent meta-analysis [60]. In the periphery, reactive oxygen species could underly the chronic mild inflammation indexed by increased IL-6 concentrations associated with schizophrenia [61, 62].

There is consistent evidence for impaired GABA synthesis in cortical interneurons in samples of post-mortem brain in schizophrenia, but it is less clear whether GABA neurones are lost [20]. The interpretation of In-vivo MRS GABA studies is complicated by variable definitions of ROIs and by clinical and treatment status. In 17 MRS studies with ROIs ranging from subgenual, pregenual, anterior, and dorsal cingulate, Egerton et al (2017) found no overall decrease in GABA [63]. More recently, Simmonite et al reported an overall decrease in GABA in 10 studies in ROIs corresponding to our ACC ROI [64]. The authors suggest GABA deficits may be modulated by illness stage (greater in acute) or ameliorated by prolonged medication. However, their subgroup difference was small and we found no evidence for an effect of phase or of medication exposure on GABA deficits in patients. Iwata et al (2018) found 3 studies in OC all 3 reporting reductions in GABA in keeping with our study [65]. Our MRS findings are strongly corroborated by a report of lower GABA concentrations measured directly in cerebro-spinal fluid in 40 prospectively recruited patients with FEP and 21 controls [66]. Furthermore, reduced CSF GABA correlated with PANSS positive scores and functional outcomes as in our study in which positive and negative symptoms correlated with low GABA across Recent and Established groups. The findings suggest that low GABA may have a key role in the pathogenesis of symptoms in schizophrenia but not apparently acting via glutamate disinhibition.

Conclusions

In testing the neurotoxic synaptic loss account of schizophrenia pathogenesis, we confirmed the prediction that MRS glutamate in anterior cingulate cortex is reduced in people with long-term treated symptoms of schizophrenia. The deficit appears to be the stable result of a process which is not apparent in early-stage illness, is not related to exposure to chronic antipsychotic medication and is not mediated by central or peripheral measures of inflammation. There was no evidence of the predicted association with negative symptoms or with other symptoms. GABA content was reduced in anterior cingulate and occipital cortex in the full sample of patients with recent or established schizophrenia. Reductions in anterior cingulate GABA content correlated with positive and negative symptoms, but did not correlate with markers of inflammation, degree of drug treatment or with glutamate. These in-vivo findings strengthen the evidence that impaired GABA function is involved the pathogenesis of the symptoms of schizophrenia. Further analyses to be reported elsewhere are relating these findings to cortical connectivity using magnetoencephalography and assessing dynamic glutamine-glutamate turnover using ^13C-MRS.

Conflict of interest

LP reports personal fees for serving as chief editor from the Canadian Medical Association Journals, speaker/consultant fee from Janssen Canada and Otsuka Canada, SPMM Course Limited, UK, Canadian Psychiatric Association; book royalties from Oxford University Press; investigator-initiated educational grants from Janssen Canada, Sunovion and Otsuka Canada outside the submitted work.