Towards an Integrated View of Early Molecular Changes Underlying Vulnerability to Social Stress in Psychosis

Abstract

Psychotic disorders are heterogeneous and complex, involving many putative causal factors interacting along the course of disease development. Many of the factors implicated in the pathogenesis of psychosis also appear to be involved in disease onset and subsequent neuroprogression. Herein, we highlight the pertinent literature implicating inflammation and oxidative stress in the pathogenesis of psychosis, and the potential contribution of N-methyl-d-aspartate receptors (NMDARs). We also emphasize the role of peripubertal social stress in psychosis, and the ways in which hippocampal dysfunction can mediate dysregulation of the hypothalamic-pituitary-adrenal axis and cortisol release. Finally, we propose a model wherein inflammation and oxidative stress act as a first hit, producing altered parvalbumin interneuron development, NMDAR hypofunction, microglial priming, and sensitivity to a second hit of peripubertal social stress. With a greater understanding of how these factors interact, it may be possible to detect, prevent, and treat psychosis more effectively.

Introduction

Schizophrenia is a debilitating mental disorder characterized by positive symptoms (e.g., abnormal perceptions and beliefs), negative symptoms (e.g., anhedonia and social withdrawal), and cognitive deficits. It is a multifactorial illness usually diagnosed in late adolescence or early adulthood that has a complex developmental course beginning long before the first episode of psychosis and progressing long thereafter [1, 2]. Herein, we will focus on the molecular basis of psychosis development and neuroprogression, first describing the N-methyl-d-aspartate receptor (NMDAR) model of psychosis, and then outlining the literature on the important roles of inflammation, oxidative stress, and social stress. We have chosen these factors because they appear to have important interactions in the development, onset, and progression of psychosis (Fig. 1), notwithstanding other neurochemical/functional/genetic alterations also involved [2–4].

Fig. 1.

Factors suggested to contribute to the neurodevelopment and neuroprogression of psychosis. Distal factors, which generally include prenatal insults, play an important role in predisposing the child to later social stress. Social stress further impacts the development of the brain, leading to an altered stress response and dysregulation of neurotransmitter systems such as glutamate and dopamine. Onset may begin with mild symptoms, as in the clinical high-risk population, but progresses to psychosis as episodes compound existing biological and psychosocial impairments. Eventually, neuroprogression may lead to treatment resistance and further deterioration unless psychosocial or pharmacological interventions divert the trajectory away from psychosis towards resilience and health.

NMDAR Hypofunction and Psychosis

There is a great deal of interest in the function of NMDARs expressed on cortical fast-spiking GABAergic interneurons, also known as parvalbumin interneurons (PVIs). PVI NMDAR hypofunction provides a potential common target for the inflammatory, oxidative, and social stress pathways that will be discussed in this chapter. In addition to being affected by metabolic stressors, NMDAR hypofunction may also instigate further damage after the onset of psychosis. These receptors appear to help regulate the same stressors that caused their dysfunction in the first place, suggesting they may, in part, account for neuroprogression at the molecular level [5]. These factors are briefly outlined here to provide context for the remainder of the chapter.

NMDARs are ionotropic channels that are gated by the neurotransmitter glutamate. They play a key role in mediating long-term potentiation and other synaptic modifications that are dependent at the level of neuronal activity [6]. NMDAR alterations were first implicated in psychosis through studies of NMDAR antagonists, such as phencyclidine and ketamine, showing that their effects on humans were similar to symptoms seen in schizophrenia [7]. Following these findings, other lines of evidence converged on NMDAR hypofunction presenting with schizophrenia phenotypes [8, 9]. Disruptions in NMDARs are especially relevant in PVIs because these neurons regulate pyramidal neurons in relevant brain areas. Interestingly, even mild antagonism of NMDARs can lead to lasting alterations in the PVI number [10]. Moreover, PVIs are the last subset of interneurons to develop, and they undergo significant maturation during puberty, ostensibly an important period in the onset of psychosis [11, 12].

Inflammation and Psychosis

Inflammation has been implicated in several psychiatric (e.g., depression) and neurodegenerative disorders, whereas its role in psychosis is still under intense study [13–15]. Multiple lines of evidence from epidemiological, preclinical, and clinical studies have repeatedly shown signs of immune alterations in the pathogenesis of schizophrenia. The effect of inflammation on psychosis may be mediated by its impact on PVI NMDARs over multiple stages of development [16].

Epidemiological Studies

Epidemiological studies have shown an association between maternal infection during pregnancy and the subsequent development of psychosis in the child. The first to describe this connection were studies of the 1957 influenza pandemic which showed increases in psychiatric hospital admissions for psychosis among those who were in their second trimester of fetal development during the outbreak [17, 18]. Later studies found that the relative risk of developing schizophrenia following a first-trimester influenza infection can be as high as 7-fold [19]. These findings have since been replicated in several other infectious diseases, and studies suggest that it is the mother’s immune reaction, rather than the infectious agent itself, that increases the risk of schizophrenia [13, 20]. It is estimated that if maternal influenza infections were eliminated, as many as 14–21% of schizophrenia cases would be prevented, suggesting that alterations in the immune system are likely involved in the pathogenesis of psychosis [21].

Evidence from Preclinical Models

Rodent models have also provided supporting evidence for the role of immune activation in the development of psychosis, including the disruption of PVI development and NMDAR function. There are several maternal immune activation (MIA) models used to study this process, including simulated viral infection with polyriboinosini c:polyribocytidylic acid, poly(I:C), simulated bacterial infection with lipopolysaccharide (LPS), and frank infection with live pathogens [13]. Noninfectious MIA models such as poly(I:C) or LPS have reliably produced a psychosis-like phenotype in rodent models, including characteristic anatomical and behavioral alterations [22, 23]. In addition, multiple changes in key neurotransmitter systems have been noted in the MIA model, including hippocampal NMDAR hypofunction and enhanced sensitivity to acute dopaminergic stimulation, both of which have been noted in schizophrenia [24].

Early immune activation not only appears to affect normal brain maturation in the developing fetus, it also seems to have a role in predisposing the brain to future inflammation. Studies have shown that microglia, the most numerous brain immune cells, may be “primed” if activated during fetal development, causing them to be more easily activated later in life [25, 26]. Prenatal microglial activation can also alter glutamatergic synaptic function later in life, probably through their direct interaction with NMDAR-expressing neurons [27, 28].

Adding another layer of complexity, proinflammatory cytokines can alter the activity of enzymes, such as tryptophan 2,3-dioxygenase and indoleamine 2,3-dioxygenase (IDO), leading to altered tryptophan metabolism to favor kynurenine and its subsequent conversion to quinolinic acid (QUIN), an NMDAR agonist, and kynurenic acid (KYNA), an NMDAR antagonist. In psychosis, multiple studies have demonstrated a shift towards the KYNA side of the catabolic pathway, as evidenced by elevated KYNA levels [29–32] and a decreased QUIN/KYNA ratio in CSF [33]. KYNA elevations in psychosis are contrary to findings in depression, where induction of IDO appears to favor the QUIN catabolic pathway, leading to elevations in the QUIN/KYNA ratio [34]. The apparent disparity between the two illnesses may be explained by alterations in IDO and its downstream enzymes, such as kynurenine 3-monooxygenase (KMO), which shunts kynurenine into the QUIN synthesis pathway. Consistent with this hypothesis, KMO enzyme activity was reduced in the prefrontal cortex in postmortem samples of patients with psychosis, which may explain relative elevations in KYNA compared to depression [35]. However, more research is required to fully understand the differences in tryptophan catabolism between psychiatric illnesses.

In rodent models of psychosis, the effects of KYNA administration appear to mimic those of NMDAR antagonists, such as ketamine and phencyclidine [36]. Furthermore, KYNA may have a permanent impact on the developing brain - rodents exposed to KYNA during early development continue to display higher levels as adults and present psychosis-like symptoms [37]. Taken together, these findings suggest that inflammation may prime the brain for NMDAR hypofunction by increasing the level of KYNA in psychosis.

In vivo Human Studies

In studies of patients with schizophrenia, there is evidence of a chronic inflammatory state with an increase in proinflammatory cytokines and reductions in anti-inflammatory cytokines [38, 39]. Polymorphisms in some proinflammatory cytokine gene complexes, e.g., interleukin (IL)-1, may be linked to increased likelihood of developing schizophrenia [40]. Furthermore, although imaging studies of neuroinflammation using positron emission tomography have yielded inconclusive results in participants with psychosis [41, 42], increases in peripheral inflammatory markers have been observed in first-episode psychosis and are linked to the severity of psychopathology [43, 44]. Finally, treatment with antipsychotic medication may act, at least in part, through an anti-inflammatory mechanism by modifying cytokine levels [39, 45]. These findings suggest that inflammation may be linked to neuroprogression in psychosis.

Oxidative Stress and Psychosis

Oxidative stress from reactive oxygen species (ROS) has also been implicated in the pathogenesis of psychosis and may be an important factor in the onset and later progression of the illness. ROS are an unavoidable by-product of aerobic metabolism, produced primarily in the mitochondria of cells throughout the human body. These chemical species are strongly implicated in the process of aging due to their ability to damage cellular components such as lipids and proteins [46]. Under normal circumstances, endogenous antioxidants, such as glutathione (GSH), neutralize these factors and protect human tissues from excess damage [47]. Without sufficient GSH to balance ROS, neurotoxicity can occur, which may explain some of the structural and functional changes occurring in psychosis [48].

Preclinical Studies

A number of animal studies have begun to tease apart the complex time-dependent relationship between oxidative balance and NMDAR hypofunction in the development of psychosis. Without protective antioxidants in development, damage can ensue; studies have shown that GSH depletion in mice during postnatal days 5–16 leads to permanent psychosis-like deficits [49]. Interestingly, PVIs appear to be particularly vulnerable to oxidative stress. This may be due to the especially high metabolic demands accompanying their fast firing rate and the fact that they rely on ensheathing perineuronal nets for protection during development, which are themselves vulnerable to oxidative stress [5, 50]. Furthermore, GSH deficits can reduce NMDAR function, which may in turn lead to decreases in the cortical PVI number [10, 51]. Conversely, increasing GSH levels has a beneficial effect - treatment with the GSH-replenishing precursor N-acetylcysteine has been shown to prevent PVI abnormalities in various rodent models of psychosis [52, 53]. The level of GSH may normally be regulated by a feedback loop between NMDARs and the glutamate cysteine ligase (GCL) catalytic subunit, a component of the enzyme responsible for GSH synthesis. Deletion of the GCL catalytic subunit triggers oxidative stress leading to impairments in cortical plasticity that can cause long-lasting failure of PVIs [54]. The expression of the GCL catalytic subunit is regulated by NMDAR activity in response to higher synaptic activity, an important protective function whose absence can cause apoptosis in cortical neurons [55, 56]. In turn, the activity of NMDAR is regulated by the oxidative status at its GRIN2A subunit [57]. All of this points to a close relationship between GSH and NMDARs in mitigating oxidative damage in the brain, particularly in vulnerable fast-spiking PVIs.

Clinical Studies

In patients with psychosis, our understanding of the effects of oxidative stress is limited. Some studies have found evidence of excess protein and lipid oxidation in blood, CSF, and postmortem tissue as well as alterations in central and peripheral GSH levels [47, 58–60]. Furthermore, there appears to be a correlation between GSH levels and the severity of psychotic symptoms in schizophrenia. Recent clinical trials have shown that N-acetylcysteine administration in schizophrenia can reduce negative and cognitive symptoms, including mismatch negativity, an indicator of NMDAR function [61–63]. Gene and protein alterations related to oxidative stress have also been noted. For instance, familial variations in the GCL modifier, a subunit of the GSH synthesis enzyme, may increase the risk of schizophrenia up to 4-fold [64]. On the other hand, levels of oxidative markers appear to vary significantly based on the clinical status in crosssectional studies [47]. Furthermore, a recent meta-analysis found no effect of the clinical status on GSH synthesis [65]. Thus, although there seem to be perturbations in oxidative stress in clinical studies, replication of findings and large-scale longitudinal studies are needed in order to understand changes in the oxidative status over the course of psychosis.

Connections between Inflammation and Oxidative Stress

Oxidative stress and inflammation interact at multiple levels, and it is likely that many of the deleterious effects of inflammation are mediated by oxidative stress and vice versa. Oxidative stress induces inflammation via activation of nuclear factor (NF)-κB [66, 67]. The immune system is also a source of oxidative stress. For example, microglia generate ROS to destroy pathogens, which is also toxic for neurons and can sensitize the brain to more damage by activating signaling molecules such as NF-κB, consistent with the neuroprogression of psychosis [68]. Proinflammatory cytokines, such as IL-1β, IL-6, and tumor necrosis factor (TNF)-α, induce production of free radicals through NF-κB [69]. Studies also show that the development of psychosis in immune activation models may, in fact, be mediated by oxidative stress due to low levels of GSH [70, 71]. Specifically, one study using an MIA model showed elevations in multiple markers of oxidative stress in the hippocampus of male offspring, including a decreased ratio of GSH to its reduced form, GSSG. In this same study, replenishing GSH by administering N-acetylcysteine reduced oxidative markers back to baseline and delayed detrimental effects in male offspring [71]. Taken together, these findings suggest that immune activation and oxidative stress may have a close reciprocal relationship in psychosis development and neuroprogression.

Social Stress and Psychosis

Stress, particularly social stress, is another important etiologic factor implicated in multiple developmental stages of psychosis [72–76]. Studies suggest that this relationship is mostly mediated via the dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, which dictates the release of the stress hormone cortisol. In youth at risk for psychosis, high baseline cortisol levels predict clinically significant symptoms after 2 years [77]. Moreover, in those with schizophrenia, the level of cortisol has been positively associated with the severity of psychotic symptoms [78]. It is worth noting, however, that some studies show that cortisol is not consistently elevated in psychosis but rather appears to display disturbed HPA axis regulation; patients with psychosis, for example, display a blunted response to dexamethasone suppression, lower cortisol awakening response, and flattening of diurnal cortisol variations [79–82].

Social Stress, Hypothalamic-Pituitary-Adrenal Axis Dysregulation, Inflammation, and NMDAR Hypofunction

Despite the reported links between cortisol, HPA axis dysregulation, and the increased risk of psychosis, the exact mechanisms are poorly understood. However, one major theory implicates hippocampal dysfunction as the driving factor. This comes from several lines of evidence reviewed extensively elsewhere [83]. Key findings are as follows:

(a) evidence from primates suggests that the hippocampus plays a key role in regulating the HPA axis through negative feedback by decreasing the corticotropin-releasing hormone level; (b) elevated cortisol levels are known to damage the hippocampus; (c) there is an inverse relationship between cortisol secretion and hippocampal volume in humans; (d) in schizophrenia, hippocampal volume is reduced more than in any other brain region; and (e) the magnitude of hippocampal volume reduction is correlated with the severity of cognitive symptoms in schizophrenia [83].

Inflammation may play an important role in mediating these alterations in the hippocampus. It has long been known that social stress is capable of increasing inflammatory markers, both in acute and chronic stress [84]. Acute psychosocial stress challenges increase the levels of both cortisol and ILs in healthy volunteers, participants with negative affect, and individuals exposed to early life stress [85–87]. In terms of chronic stress, those raised in a harsh family have been shown to have higher levels of peripheral markers of inflammation, as do those who self-report as being lonely or having low social status [88–90]. It is possible that the inflammation related to stress can lead to microglial activation, which may further damage PVIs in the hippocampus [91]. Once an inflammatory state is induced, it may give feedback to cause further dysregulation of the stress response by disrupting the normal anti-inflammatory properties of glucocorticoids [92–94]. Interestingly, reducing stress through the pubertal administration of benzodiazepines was able to prevent the onset of psychotic symptoms and elevations in dopamine in the methylazoxymethanol rodent model of psychosis [95].

Towards a Common Molecular Pathway

Studies combining measures of oxidative stress, inflammation, and social stress with PVI dysfunction provide further insight into the complex relationships described so far and provide the fodder for an integrated model. In general, these studies have suggested that neonatal oxidative stress or immune activation make the developing brain more susceptible to the effects of social stress during puberty. For instance, one animal study combined oxidative stress, NMDAR hypofunction, and social stress along development, showing that selective deletion of an NMDAR subunit increased oxidative stress in a way that was exacerbated by later social isolation [96]. Another study testing a similar hypothesis found that pharmacologically inhibiting the formation of ROS prevented PVI deficits induced by social isolation [97]. In terms of neuroinflammation, one study demonstrated that maternal treatment with poly(I:C) followed by a peripubertal stress task led to a synergistic increase in offspring levels of inflammatory cytokines, psychosis-like symptoms, and microglial activation in response to stress [98].

All of these preclinical studies appear to support a disease pathway similar to the “two-hit” model of psychosis (Fig. 2). In summary, a maternal inflammatory state may cause microglial activation and priming, redox imbalance, and vulnerabilities in highly sensitive hippocampal and cortical PVIs. These factors increase vulnerability in the affected individual such that during crucial periods of neuronal growth and development, stressors can cause overactivation of the HPA axis, microglial activation, and further PVI damage. The hippocampus is thus negatively affected by both PVI dysfunction and cortisol levels, and is no longer able to provide negative feedback to the HPA axis [99]. The activation of oxidative and inflammatory mechanisms throughout the onset and course of the disease may contribute to disease progression.

Fig. 2.

Proposed model linking the molecular and psychosocial mechanisms involved in the development and neuroprogression of psychosis. In the “first hit,” prenatal insults due to maternal cytokine activation and oxidative stress may simultaneously alter the development of parvalbumin interneurons (PVIs) and predispose the developing brain to further inflammation by priming microglia for subsequent overactivation. Chronic social stress during puberty provides a “second hit” that triggers hippocampal dysfunction and hypothalamic-pituitary-adrenal (HPA) axis dysregulation, triggering the primed microglia, and leading to inflammation, oxidative stress, and further PVI dysfunction. All this culminates in excitation/inhibition imbalances and altered control of dopamine (DA) pathways, leading to the first psychotic symptoms and later neuroprogression through the accumulation of allostatic load.

Conclusion

Herein, we discussed the molecular mechanisms underlying the pathogenesis and neuroprogression of psychosis. There are still many unanswered questions, especially considering that many studies are in preclinical models of the disease. However, advances in imaging techniques, longitudinal studies, and clinical trials may help to refine our understanding of the illness in living humans. Indeed, there are already a number of clinical trials that take advantage of the mechanisms described here, including NMDAR agonists, antioxidant therapies, anti-inflammatory treatments, and psychosocial interventions to prevent stress/oxidative stress and help prevent the worsening of symptoms over time [2, 5, 13, 100]. In the not-too-distant future, clinicians may be able to reliably identify early psychosis in vulnerable individuals before it has fully emerged, and initiate preventative therapies or provide social support. In doing so, it may be possible to reduce the considerable societal and personal burden of schizophrenia.