Severe depression is associated with increased microglial quinolinic acid in subregions of the anterior cingulate gyrus: Evidence for an immune-modulated glutamatergic neurotransmission?

Recent studies have focused on the role of immune dysfunction in depression, and analogies to "cytokine-induced sickness behavior" have been established [1]. Sickness behavior is a coordinated set of adaptive behavioral changes that develop in affected individuals during the course of an infection. Disease symptoms include lethargy, depression, failure to concentrate, anorexia, sleep disturbances, reduction in personal hygiene or social withdrawal, and are mediated by proinflammatory cytokines, such as interleukin-1 (IL-1), interleukin-6 (IL-6) and tumor necrosis factor α (TNFα) [1].

Previous research has suggested that these specific monocyte-derived cytokines are increased in the peripheral blood of acutely depressed patients [2‐7] along with elevated monocyte counts [8, 9]. Furthermore, lymphocyte and natural killer cell abnormalities have been described [10‐12]. It is not yet clear, whether these changes in the peripheral blood are associated with corresponding neuroinflammatory responses and alterations in neurotransmission. Peripheral immune processes may be mirrored in the brains of patients with acute depression by microglial cells which represent the brain's mononuclear phagocyte system (MPS) [2, 13]. Indeed, an increased density of microglia expressing human leukocyte antigen (HLA)-DR has recently been observed in the anterior midcingulate cortex (aMCC), the dorsolateral prefrontal cortex and the mediodorsal thalamus of suicidal patients with affective disorders [14]. However, this study of the surface marker HLA-DR did not suggest a mechanism of how modulation of neurotransmission is accomplished.

Quinolinic acid (QUIN), an endogenous modulator with agonistic properties on N-methyl-D-aspartate (NMDA), which is produced by microglial cells, may serve as a potential candidate for such a link between immune and neurotransmitter changes in depression [13]. This hypothesis is based on the observation that the above mentioned proinflammatory cytokines induce a shift from serotonin synthesis to tryptophan metabolism via the kynurenine pathway in glial cells [1, 15‐17], which may ultimately lead to serotonin depletion and particularly an increased production of the metabolite QUIN (Figure 1). MPS cells, such as microglia, macrophages and monocytes, mainly produce the NMDA receptor agonist QUIN, while astrocytes synthesize the NMDA receptor antagonist kynurenic acid (KYNA) because they lack the enzyme kynurenine monoxygenase (KMO) [18‐20]. Analyses of blood and cerebrospinal fluid revealed elevated QUIN levels in cytokine-induced depression and major depressive disorder (MDD) [1, 21, 22], while an increase in KYNA production was related to schizophrenia [23‐25].

These findings may connect immune pathologies to MPS activation in MDD. In addition to serotonin depletion, a direct glutamatergic mechanism has been suggested, which has recently been identified as an important target of antidepressant treatment [26]. In this context, the anterior cingulate cortex (ACC), with its region-specific NMDA and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) glutamate receptor profiles that cover functionally segregated areas, represents an important target region in the central nervous system, although investigations must account for the histo-architectural diversity of this region [27]. The importance of the pregenual anterior cingulate cortex (pACC) in MDD is supported by the pronounced effects of the glutamate modulating NMDA antagonist ketamine on the improvement of clinical symptoms in treatment-resistant MDD patients [28], in which ketamine leads to an increase in glutamate concentration precisely in this region [29].

To our knowledge, this is the first report of microglial QUIN expression in human brain during acute depressive episodes. An increase in QUIN-immunopositive microglia was specific to cingulate subregions with high NMDA receptor densities, like the sACC and the aMCC, but not the pACC, which shows a lower NMDA receptor expression. This increase in QUIN-immunoreactive microglial cell densities was found particularly in unipolar patients. With regard to BD less clear statements can be given. We observed a significant difference between MDD and BD, yet the BD group is also higher than the controls, though this is apparently not significant (Figure 3b). This could be due to the small number of specimens studied. The numeric increase in QUIN-immunopositive cell counts was paralleled by the presence of microglial forms that displayed numerous granular structure processes in the proximity of neurons in the depressed group, supporting an interaction of inflammatory mechanisms and neurotransmission at the time of acute depressive episodes. These findings thus corroborate evidence for acute inflammatory microglial activation in depression, leading to increased levels of the NMDA receptor agonist QUIN in regions with corresponding receptor profiles that have been previously revealed as key structures in non-invasive imaging studies.

Increased levels of QUIN, which is also produced by macrophages and monocytes, have already been found in the blood and cerebrospinal fluid of subjects with cytokine-induced depression or MDD [1, 21, 22]. Thus, our result of increased microglial QUIN expression in suicidal MDD patients is in line with the hypothesis of a systemic MPS activation during acute disease phases of depression [2‐9, 14]. Due to the excitotoxic properties of QUIN, our findings are also supporting the neurodegeneration hypothesis of depression [15]. Therefore, our study provides insight into why immune- and glutamate-modulating therapies may be helpful for acutely ill suicidal patients suffering from depression. Potential candidate drugs include the tetracycline antibiotic minocycline, which inhibits microglial activation by blocking NF-kappa B nuclear translocation [39‐42] or anti-inflammatory inhibitors of cyclooxygenase-2 [43, 44]. Furthermore, severely depressed suicidal patients may benefit from the administration of glutamate-modulating drugs, such as the NMDA receptor antagonist ketamine [28, 45, 46].

It should be mentioned that Laugeray and colleagues observed reduced levels of the QUIN precursor 3-OH-kynurenine (3HK) in the cingulate cortex and increased levels of 3HK in the striatum and the amygdala of mice using an unpredictable chronic mild-stress model for the induction of depressive-like symptoms [47]. The observation of reduced 3HK could be due to either reduced formation of 3HK or increased degradation of 3HK to QUIN, which would result in reduced 3HK level. Since QUIN was not directly measured in this study, a translational validation of these converging results remains subject to future studies. A general drawback of animal studies is that it is unclear if animal models adequately reflect the pathophysiology of human MDD or BD. Moreover, an analysis of ACC subregions was not undertaken in this study, and direct correspondence of subregions in primates and humans differ considerably to those found in rodents. Therefore, the implications on regional glutamatergic throughput in depression, as a function of local NMDA and AMPA receptor profiles, remain difficult to interpret in animal studies.

We have shown that abnormal NMDA receptor function related to microglial activation is highly dependent on the location in the ACC in humans. Non-invasive studies have led to similar distinctions of abnormal cingulate cortex activation in MDD. While sACC hyperactivity has been postulated in a number of studies, the pACC has been less consistently characterized. Grimm et al. [48] found a reduced deactivation during a task study, reflected in smaller negative BOLD responses in a sample of severely depressed patients; this functional deficit was accompanied by decreased pACC glutamate and glutamine levels, which are correlated with the severity of clinical depressive symptoms [49‐51]. Moreover, these glutamatergic deficits have been related to anhedonia and abnormal functional activations in the pACC in humans [52]. Our finding of relatively increased QUIN immunoreactivity, which is potentially associated with serotonin depletion due to changes in the kynurenine pathway, would thus be consistent with the relative hyperactivation in the sACC. The sACC is also a putative target of deep brain stimulation. Importantly, the metabolic activity after deep brain stimulation in the sACC, as measured by positron emission tomography, shows a reduction in hyperactivity similar to a region bordering the aMCC and the pACC [53].

Specifically increased concentrations of the NMDA receptor agonist QUIN in the aMCC and the sACC may also directly contribute to the disturbed balance in glutamatergic throughput, which could explain the rapid onset of antidepressant effects after ketamine [28, 46]. According to Salvadore et al. [54], activity bordering the pACC does indeed predict the responsiveness towards ketamine treatment; therefore, our finding may represent a histopathological surrogate. As shown by Vollenweider and Kometer [55], similar metabolic changes can be found in the sACC and aMCC upon acute ketamine administration. Therefore, the anatomical patterns of such pharmacological challenges fit the observed pattern of microglial histopathology.

The present study has certain limitations that need to be considered: (1) our findings are based on a relatively small number of MDD and BD cases and must be confirmed in a larger sample size; (2) it was not possible to track data on drug exposure or the history of inflammation and infection across the patients' entire life spans, as we could only collect data on psychotropic medication in the three months prior to death; (3) the present study enables us to draw conclusions about the cellular QUIN content, but not released or secreted QUIN in the extracellular space, which potentially interferes with glutamatergic neurotransmission; (4) it remains unclear if increased QUIN immunoreactivity in microglial cells is caused by increased synthesis or reduced degradation of QUIN. Future studies in frozen tissue may address this question by measuring different kynurenine pathway metabolites using high-performance liquid chromatography (HPLC) or mass spectrometry (MS). (5) It is currently uncertain if drugs like glibenclamide, nifedipine, metoprolol, or theophylline which have been applied in five of the control subjects may influence microglial QUIN expression.

Here we present the first study providing evidence that supports a disease-related upregulation of microglial QUIN in depressive disorders, particularly in brain regions known to be responsive to infusion of NMDA antagonists such as ketamine [55]. These results add a novel link to the immune [1, 26] and neurodegeneration [15] hypotheses of depression. Further work in this area could lead to a greater understanding of the pathophysiology of depressive disorders and pave the way for identification of novel biomarkers and therapeutic strategies targeting specific disease subtypes.