Cellular stress mechanisms of prenatal maternal stress: heat shock factors and oxidative stress

Abstract

Development of the brain prenatally is affected by maternal experience and exposure. Prenatal maternal psychological stress changes brain development and results in increased risk for neuropsychiatric disorders. In this review, multiple levels of prenatal stress mechanisms (offspring brain, placenta, and maternal physiology) are discussed and their intersection with cellular stress mechanisms explicated. Heat shock factors and oxidative stress are closely related to each other and converge with the inflammation, hormones, and cellular development that have been more deeply explored as the basis of prenatal stress risk. Increasing evidence implicates cellular stress mechanisms in neuropsychiatric disorders associated with prenatal stress including affective disorders, schizophrenia, and child-onset psychiatric disorders. Heat shock factors and oxidative stress also have links with the mechanisms involved in other kinds of prenatal stress including external exposures such as environmental toxicants and internal disruptions such as preeclampsia. Integrative understanding of developmental neurobiology with these cellular and physiological mechanisms is necessary to reduce risks and promote healthy brain development.

1. Introduction

Prenatal brain development occurs via a complex interplay between maternal, fetal, and environmental signals. Although it has been recognized for centuries that environmental influences on a pregnant mother relate to offspring health and development, intentional manipulation of the maternal environment with the goal of benefitting child development is relatively new [14]. In the past few decades, vitamin supplementation, psychotherapy, and good relationships with obstetric health professionals have been used to increase positive outcomes for mothers and their offspring [88]. Today, we take for granted that attentive prenatal care is essential for optimal embryonic and fetal development. However, current efforts do not address many other poorly understood aspects of the maternal-fetal interaction. Of particular interest are the effects of maternal psychological stress on prenatal neurodevelopment [118].

Critically, the physiologic state created by stress during pregnancy can have adverse effects on offspring brain development. Prenatal perturbations in neurodevelopment due to stress have been demonstrated to increase the risk for psychiatric disorders including attention deficit hyperactivity disorder (ADHD), autism spectrum disorder (ASD), major depressive disorder, bipolar disorder, and schizophrenia [57, 112, 149, 199], with male offspring being at higher risk than females [57, 112, 128]. To this point, pregnant women who experience bereavement [94, 112], war [193], natural disasters [99, 106], or high daily stress [57] have offspring with higher rates of these disorders. Additionally, stress may arise from poorly managed mood disorders during pregnancy, and mood disorder severity is linked to altered neurodevelopmental outcomes in offspring [26, 36, 38, 57, 112, 128].

Though poorly understood, a complex interplay of molecular and cellular mechanisms contributes to the aberrant and pathological outcomes induced by prenatal maternal stress. Stress hormones— glucocorticoids—are an obvious area of inquiry, but it has become clear that other mechanisms are involved including inflammation, hypoxia, and growth-factor signaling [157]. These physiological mechanisms induce oxidative stress and interact at the cellular level with the heat shock response, complicating how prenatal stress contributes risk for psychiatric disorders.

To address these complex risks, there is a critical need to better understand the cellular and molecular mechanisms by which the maternal physiologic stress response alters neurodevelopment. For new interventions to be developed and current guidelines refined, an understanding of neurodevelopmental mechanisms, placental intermediaries, and maternal physiology is needed. Physiological stress responses are linked through multiple mechanisms to cellular stress responses [75]. In this review, we will examine the roles of heat shock systems and oxidative stress in prenatal stress and its effects on neurodevelopment, as well as how these cellular systems may interact with maternal stress physiology. Elucidating the relationship between these processes is essential for understanding and targeting these mechanisms to buffer the effects of prenatal stress on offspring neurodevelopment.

2. Brain development and prenatal stress

2.1. Neurodevelopmental processes affected by prenatal stress

There are numerous neurodevelopmental processes that may be impacted by maternal stress, many of which may interact with the heat shock response (HSR) and oxidative stress. Maternal, placental, and embryonic factors, as well as the interactions between these systems, may impact neurodevelopmental processes during prenatal stress. We will explore these interactions in subsequent sections of this review, but we will first examine mechanisms in the developing brain that are susceptible to prenatal stress. Among these affected processes are changes in offspring GABAergic systems, neuroimmune activation, and hypothalamic-pituitary-adrenal (HPA) axis development [180].

Development of the GABAergic neuron system is particularly vulnerable to prenatal stress. During embryonic development, GABAergic neuronal progenitors are born in the medial and caudal ganglionic eminences, from which they migrate tangentially into the developing neocortex [72, 152, 184]. GABAergic progenitors differentiate into interneurons as they migrate and integrate into the neocortex, ultimately forming the inhibitory neural circuitry required for cognition, learning, and motor tasks [7, 56]. Prenatal stress late in mouse gestation reduces GABAergic progenitor production [190], while prenatal stress early in gestation delays the tangential migration in GABAergic interneurons [180]. Multiple components of the maternal physiologic stress response have effects on GABAergic neuron development. Our group and others have found that these embryonic brain processes may be sensitive to maternal inflammation [71, 142] and oxidative stress [28], for example. In addition, embryonic glucocorticoid exposure impedes pyramidal neuron radial migration in the developing cortex [62]. Thus, maternal physiologic stress response may impede multiple aspects of neuronal migration events in the embryonic forebrain. Impaired heat shock systems may be a means by which cell migration is affected, due to interactions with the proteins necessary for cell movement [52, 74, 129].

Furthermore, the effects of prenatal stress on GABAergic interneuron populations persist well into adulthood. For instance, protein levels of the GABA-synthesizing enzyme GAD67 are decreased in the frontal cortex with prenatal stress, and the maturation of postnatal GABAergic neuron development is markedly delayed [119, 126, 190]. GABAergic neuron development may be linked to both oxidative stress and heat shock signaling. Postanatal parvalbumin neuron development, for instance, is sensitive to redox imbalance [50]. Intracellular heat shock protein (HSP) production promotes GABAergic interneuron genesis in vitro [92].

Recent evidence suggests that the maternal stress response elicits changes to the embryonic immune system, which in turn can impact neurodevelopment. Microglia are resident macrophages of the central nervous system that play important roles in brain development. Derived from yolk-sac monocytes, microglia migrate from the vasculature into the developing brain by embryonic day 9 in mice [186]. Like other phagocytic cells, microglia are innate immune cells that regulate inflammatory responses, though they also have a parallel role in sculpting neurodevelopment. For instance, microglia regulate circuity in the forebrain by modulating axon growth [179]. As in other immune cells, cellular stress mechanisms like oxidative stress and the HSR are intrinsically linked to microglia function [84, 183]. This is critical, as microglia control neural progenitor cell populations during embryonic development by either promoting cell survival or reducing the progenitor pool via phagocytosis [124, 187].

Prenatal stress affects microglia populations in both the embryonic and adult brain [49]. During normal development, microglia transition from an amoeboid state to progressively more ramified morphologies, and maternal stress appears to accelerate this process in offspring brain [177]. In addition, prenatal stress increases multivacuolated microglia density in the cortical plate at embryonic day 14 in mice [71]. Furthermore, these changes in microglial activation appear to persist into adulthood, as adult rodents exposed to prenatal stress exhibit less ramified and more readily-activated microglia populations in the cortex and hippocampus [49, 71]. These observed changes in microglia proliferation and maturation suggest that prenatal stress may induce a lasting deleterious, pro-inflammatory environment in the offspring brain.

The growth and maturation of the hypothalamic-pituitary-adrenal (HPA) axis is another neurodevelopmental process that is affected by prenatal stress. In the hypothalamus, for instance, early increases in apoptotic activity is observed in response to prenatal stress [61]. Prenatal stress may have lasting deleterious effects on postnatal HPA functioning. For example, animal studies suggest a hyperactive (higher baseline activity) or hypersensitive (greater response to stressful stimuli) HPA phenotype in prenatally stressed offspring. Adult rat offspring from prenatally-stressed litters exhibit greater HPA sensitivity to stressful stimuli, and prenatally-stressed rhesus monkeys have been found to exhibit higher baseline cortisol and ACTH plasma levels during adulthood [43, 78, 198]. Female offspring tend to show greater susceptibility to the effects of prenatal stress on HPA axis development, but there is some discrepancy between studies regarding sex-specific effects [93]. Changes in offspring HPA functioning following prenatal stress also involve the hippocampus, which regulates the HPA axis via glucocorticoid receptor (GR)-induced feedback inhibition [60]. Decreased hippocampal volume and GR levels, for example, are observed in non-human primates [66]. The effects of prenatal stress on the developing HPA axis are evidently sufficient to produce permanent maladaptive changes to this neuroendocrine system.

More recent evidence suggests an epigenetic mechanism underlying changes in HPA functioning associated with prenatal stress. Prenatal stress increases promoter methylation of the GR gene NR3C1 in infants, and NR3C1 methylation levels in turn predicts infant cortisol reactivity [66, 178]. Due to the heritability of DNA methylation patterns, prenatal stress may have cascading negative impacts on multiple generations [127]. Furthermore, higher cortisol reactivity may predispose female offspring to experience heightened stress during their own pregnancies, facilitating a cycle of stress and adverse neurologic and psychiatric outcomes. GABAergic neurons in the developing hypothalamus and pituitary gland are particularly susceptible to heat stress during pregnancy [204], however little is known about the direct role of heat shock signaling in mediating altered HPA development in prenatally stressed offspring.

The effects of prenatal stress on the formation of overall brain structure and function has also been studied. In humans, structural and functional changes in multiple fore- and midbrain regions have been identified using imaging techniques (see Van den Bergh et al. 2018 for a more extensive review) [192]. The cellular and molecular processes underlying these large-scale changes are not fully understood, but likely involve alterations to neuron migration, microglial activation, and HPA axis development among others. Interestingly, the cellular changes associated with inflammatory and glucocorticoid-mediators of stress have significant links to heat shock signaling [108, 134]. As such, more research is needed to address whether heat shock signaling functions as a protective mechanism against these processes, or instead acts as a critical intermediary of prenatal stress effects on neurodevelopment.

2.2. Heat Shock mechanisms, prenatal stress, and neurodevelopmental processes

Heat shock factors are basic components of cellular function. The proteins induced in the heat shock system, HSPs, are protein chaperones that re-fold denatured proteins in response to high heat, pro-oxidants, and other forms of proteotoxic stress [3]. In the absence of significant stressors, HSPs play a role in normal protein synthesis and maintenance to allow for normal processes of cell growth and survival [111]. Developmentally, they are particularly important during periods of high cell growth due to the increased demand for proteins in synthesizing cellular structures [41]. HSPs also help mitigate harmful intracellular protein aggregation [44].

Importantly, HSPs play significant roles in neurodevelopmental processes. For instance, they aid in supporting somal growth, cell migration, and outgrowth of axons and dendrites [129]. Apoptosis, which is an imperative, controlled process during embryonic brain development, is dependent on HSPs through the ubiquitin proteasome system [98]. The relatively low oxygen levels of the embryonic environment may also induce the HSR, which in turn may promote angiogenesis necessary for oxygenation of the brain [89]. For a review of the neurodevelopmental roles of heat shock systems, see [130].

Given the important role of heat shock factors in the cellular stress response and their additional roles in facilitating neurodevelopmental processes, HSPs are likely involved in the mechanisms by which prenatal stress alters embryonic neurodevelopment. Alterations in protein folding and stability can dramatically change many of the cellular processes that are critical for proper embryonic brain development, many of which are affected by prenatal stress. Possible effects of prenatal stress involving protein stability include permanent changes in neuronal and glial number, location and growth of neurons and glia, extension of neural processes that underlie functional circuits, mitochondrial functioning, and the epigenome [1]. Maternal prenatal stress could theoretically alter embryonic brain proteins in this way, and the protective effects of the HSR could play a corrective role, although there are currently no studies of prenatal stress that demonstrate this. For example, induction of the HSR in the embryonic brain in response to prenatal stress could aid in restoring homeostasis, as seen for the protection of differentiating neurons after arsenite exposure by conditioning heat shock [201] allowing for normal development to progress. The protection of embryonic brain by HSR is also demonstrated by greater vulnerability of cortical neuron production to maternal mercury, alcohol, or seizure exposure in mice lacking heat shock factor 1 (HSF1) [74]. However, given the critical roles of HSPs in facilitating neurodevelopmental processes, HSR induction by prenatal stress may alternatively induce significant changes the developmental trajectory of the brain. For example, HSF1 and HSF2 impair normal neuronal radial migration gene transcription and post-translational modification in the setting of prenatal alcohol exposure [52]; inhibitory neuron tangential migration, impaired after prenatal stress [180], involves the same genes (e.g. Cdk5) which may indicate a role for HSF1 and HSF2 in prenatal stress. In either case, the HSR may, in part, provide an explanation for why some individuals are resilient against the effects of prenatal stress, while others are more susceptible.

In fact, very little is understood about how maternal stress physiology affects heat shock mechanisms of the developing brain. The classic stimulus for HSR, thermal changes, have yet to be studied in the embryonic brain as a consequence of maternal psychosocial stress. If such changes were to occur, the induction of HSPs may underlie interesting susceptibilities to psychiatric illness. Indeed, almost 200 genes, including those for HSPs and those associated with ASD and schizophrenia, are altered when neuronal aggregates reflecting embryonic stages of development from human inducible pluripotent stem cells are exposed to high temperature [115]. Regardless of whether thermal changes occur, proteostasis and proteotoxicity have been neglected outcomes in the study of any offspring effects after exposure to prenatal stress. While no changes in HSP25, 32, or 70 are seen in mouse offspring brain after maternal exposure to either restraint stress or radiofrequency fields [58], many other questions remain about other aspects of the HSR and prenatal stress, including whether specific developing cells may be more susceptible based on their stage (e.g. proliferative, differentiating) or subtype (e.g. forebrain or hindbrain, glial or neuron).

Despite limited evidence of heat shock systems’ involvement with prenatal stress effects on embryonic brain, the mechanisms broadly implicated in prenatal stress shed some light on how heat shock systems may be involved. For example, HSP90 may be essential for response of the embryonic brain to GR activation, which we describe in detail later in this review. Embryonic brain GR activation may occur as a consequence of prenatal stress [158] and is critical to HPA axis maturation which is influenced by prenatal stress. While this may implicate a negative impact of HSPs in stress effects on HPA axis maturation and other neurodevelopmental effects, HSP90 and other HSPs may also mitigate physiological stress on cells, which could alternatively be a critical protection for developing neurons. As we will describe later in this review, HSPs may protect cells in pro-inflammatory conditions, a likely result of maternal stress. In fact, Torii and colleagues demonstrated that a pro-inflammatory maternal immune activation protocol in mice activated a heat shock reporter in 25–30% of offspring brains [188].

Because of the nutritional requirements for protein homeostasis, impacts of maternal stress on caloric intake may also impact heat shock systems. For instance, intrauterine growth restriction in rats due to caloric restriction is associated with increased HSP90 [9] and maternal hyperglycemia in mice activates the HSR [188], both in offspring brain. Hypoglycemic conditions may alter expression of factors involved in heat shock in the embryonic brain, however, there is some discrepancy whether and how nutrient levels regulate HSP expression. One study found that low glucose levels induce Hsp7 gene expression in mouse neuroblastoma cell lines by upregulating HSF-1 activity [175]. On the other hand, other studies have shown that high glucose and amino acid levels upregulate HSF-1 through mTOR-mediated phosphorylation, suggesting that nutrient deprivation to the fetal brain may actually lead to negative HSF1 regulation [42]. In support of the latter, amino acid deprivation had been demonstrated to inhibit the DNA binding activity of HSF-1 and downregulate heat shock element (HSE)-controlled genes [79].

In addition to maternal psychological stress, prenatal exposure to environmental chemicals may alter neurodevelopment in children through similar mechanisms involving heat shock signaling. For example, prenatal exposure to sodium arsenite and cadmium have been demonstrated to induce HSP accumulation in the neuroepithelial tissue of mouse embryos [81]. In this regard, it is unclear whether HSR induction occurs as a protective mechanism or if heat shock signaling may play a significant role in mediating the adverse effects of environmental chemical exposures.

In support the theory that the HSR offers protection against the toxicity of environmental chemicals, ethanol and methylmercury at doses below the threshold for acute toxicity activate HSF1 in the murine embryonic cerebral cortex [74]. Interestingly, knockout of HSF1 in addition to ethanol and methylmercury exposure significantly increases cortical dysplasia. Furthermore, reintroduction of exogenous HSP70 protein to HSF1 knockout mice rescued the phenotype observed in response to ethanol and methylmercury treatment, suggesting that HSR activation in embryonic brain protects against prenatal toxicant exposure [74]. Interestingly, the same study demonstrated that neural progenitor cells generated from patients with schizophrenia exhibited significant increases in cell-to-cell variation in heat shock response to ethanol and methylmercury [74]. As such, genetic mutations leading to altered heat shock response may increase the vulnerability of neural progenitors to environmental insults, thus contributing to the development of schizophrenia.

In addition to their roles in protecting the embryonic brain from stress, HSPs also play crucial roles in the regulation of neurodevelopment as discussed previously, rendering them a possible target for toxicant-induced neuronal dysfunction. For example, prenatal exposure to polychlorinated biphenyls (PCBs) in zebrafish induce significant reductions in neuronal levels of HSC70, a member of the heat-shock protein 70 family, alongside significant reductions in neuronal serotonin [102]. HSC70 is constitutively expressed in neuronal synapses and may play a role in synaptic plasticity and vesicle recycling [139, 191]. As such, HSP70’s role in synaptic vesicle cycling may underlie the depletion of neuronal serotonin in the embryonic brain with PCB exposure [102]. Ultimately, empirical evidence linking prenatal stress and the HSR is clearly lacking. However, what is known about embryonic brain outcomes and physiological mechanisms of prenatal stress implicates heat shock systems. In particular, the HSR to oxidative stress is an important consideration, given the significant role oxidative stress may play in mediating the effects of prenatal stress on neurodevelopment [116].

2.3. Oxidative stress mechanisms, prenatal stress, and neurodevelopmental processes

The impact of maternal stress on developmental processes in the embryonic brain may result from molecular mechanisms in response to cellular stress. Oxidative stress results from a range of adverse exposures and physiological changes and is of particular interest due to its close relationship with heat shock signaling. The chief mediators of oxidative stress are reactive oxygen species (ROS), generated largely through the processes of oxidative phosphorylation and aerobic metabolism. ROS are also generated in significant amounts through reactions catalyzed by oxidoreductase enzymes, non-enzymatic reactions catalyzed by endogenous metal ions, and by neutrophils and macrophages via NADPH oxidase. Members of the ROS family include the superoxide anion (O₂⁻), hydroxyl radicals (OH^•), hydrogen peroxide (H₂O₂), and lipid peroxides (LOOH). Excess generation of ROS can result in cellular damage in the forms of lipid peroxidation, protein misfolding, and DNA damage, all of which may significantly interfere with cellular functions. Furthermore, high levels of ROS induce the opening of the mitochondrial permeability transition pore (MPTP), leading to cell death via apoptosis [162, 167].

Importantly, the embryonic brain is particularly susceptible to the effects of oxidative stress, as expression of antioxidant enzymes are relatively low in embryonic brain and the cellular environment has relatively high concentrations of catalytic metals such as Fe²⁺ and Cu⁺ that promote the production of hydroxyl radicals from hydrogen peroxide. Furthermore, neural tissue has a relatively high abundance of membrane polyunsaturated fatty acids, which can be readily oxidized by ROS to produce highly reactive byproducts such as 4-hydroxynonenal (4-HNE) and malondialdehyde (MDA), which form stable adducts to protein and DNA nucleophiles [85, 172].

Maintaining cellular redox balance is critical for proper brain development, as both the over- and underproduction of ROS can negatively impact cellular functioning and survival [39]. In developing neurons, ROS play a regulatory role in multiple important cellular pathways. For instance, ROS are implicated in the activation of nuclear factor erythroid-derived 2-like 2 (NRF2) and Redox factor-1 (REF-1) signaling [159]. Both of these pathways are important for astrocyte survival and function, with NRF2 inducing cytoprotective repair mechanisms and REF-1 playing a critical role in controlling inflammation [12, 18]. Additionally, ROS regulate the transcription of genes involved in inflammation, cell proliferation, cell adhesion, and angiogenesis through activation of redox-sensitive transcription factors such as the nuclear factor NF-κB and AP-1 [78].

Neural stem cell proliferation and differentiation also both rely on ROS-mediated processes. For example, NAPDH oxidase 2 (NOX2), an enzyme that generates ROS, regulates neuronal stem cell development by mediating key transcription factor activity [138]. Furthermore, ROS mediate the mitogen-activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K) pathways, both which play a role in neural stem cell proliferation [11]. In addition, activation of epidermal growth factor (EGF) receptors by H₂O₂ plays a critical role in promoting neural progenitor proliferation [86].

ROS levels are also specifically maintained for cell fate determination. Higher levels of intracellular ROS are present in cells that differentiate into neurons, while lower levels are present in those with a glial lineage [189]. As such, the significant roles that ROS play in key neurodevelopmental processes demonstrates how disruptions to the oxidative state of the embryonic brain may disrupt its formation and long-term functioning.

Mediators of the maternal stress response, such as increased inflammation or cortisol production, impact the redox balance in the embryonic environment. In this respect, GSH levels in the embryonic brain decrease in response to both maternal inflammation and restraint stress [105, 120]. Prenatal stress also increases lipid peroxidation in offspring postnatal brain, suggesting that maternal stress during pregnancy may have long-lasting effects on brain redox balance [25].

Recent evidence suggests that ROS play significant roles in mediating the effects of maternal stress on specific neurodevelopmental processes. For example, increased ROS [28] and blockade of mitochondrial oxidative phosphorylation disrupt GABAergic interneuron tangential migration [114] in a similar manner to that seen with maternal restraint stress [180]. The impact of prenatal stress on offspring HPA axis development may result from the susceptibility of embryonic hypothalamic neuronal and microglia progenitors together to redox shifts, as has been shown in response to fetal alcohol exposure [33]. Additionally, application of the antioxidants N-acetylcysteine (NAC) and astaxanthin both block changes in neuronal migration and microglial activation seen with prenatal stress, indicating that oxidative stress is a requirement for the observed phenotypes [28, 29]. Administration of the antioxidant NAC has also been shown to protect against learning and memory deficits in prenatally stressed mice [24].

Taken together, the aforementioned studies support ROS-mediated mechanisms in the effects of prenatal stress on brain growth and development. Furthermore, the efficacy of antioxidants such as NAC in negating the effects of prenatal stress indicate that buffering against oxidative stress may be a potential intervention for preventing adverse neurodevelopmental outcomes in children.

3. Relationships between heat shock and oxidative stress mechanisms

Given the role that HSPs play in protecting cells under stressed conditions, it is not surprising that a significant relationship exists between oxidative stress and heat shock systems. Similar to conditions of heat stress, increased levels of ROS induce protein unfolding through a variety of posttranslational protein modifications. Under conditions of oxidative stress, the most vulnerable protein targets are the sulfur-containing side chains of cysteine and methionine; however other side chains such as histidine, tryptophan, lysine, tyrosine, and arginine can be readily oxidized [23]. These modifications can lead to the loss of a protein’s secondary and tertiary structure, which in turn reduces its solubility, stability, and activity [23]. To protect cells from damage caused by misfolded proteins, heat shock factors demonstrate redox sensitive activation through a variety of mechanisms, including direct induction of heat shock factors by reactive electrophiles, transcriptional regulation via the antioxidant response elements, and production of HSPs as a consequence of ROS-mediated protein unfolding [2].

Redox-mediated activation of heat shock factors

Heat shock factor 1 (HSF1) functions as a sensor for the unfolded protein response and is responsible for the transcription of various HSPs. Under non-stressed conditions, HSF1 monomers are sequestered in the cytoplasm by the HSP90 multichaperone complex consisting of HSP90, HSP40, and HSP70. During cellular stress, increased levels of unfolded proteins compete for binding with these chaperones, leading to the release of HSF1. Upon its release, HSF1 translocates to the nucleus where it trimerizes and binds heat shock elements HSE to induce HSP transcription [132]. As such, redox-mediated activation of HSF1 can be achieved indirectly through an increase in misfolded proteins due to oxidative damage. HSF1 may also be transcriptionally upregulated by oxidative stress via NRF2 interaction with antioxidant response elements (AREs) located upstream of the Hsf1 gene transcription start site [151].

Direct mechanisms of HSF1 activation by oxidative stress have been postulated to be similar to those of the redox sensitive transcription factor NRF2, in which HSF1 is activated via thiol oxidation or sglutathiolation. For example, H₂O₂ directly activates HSF1 through oxidation of cysteine residues 35 and 105, which upon disulfide bridge formation promote HSF1 trimerization and HSE binding [2, 117]. Endogenous electrophiles produced as a consequence of oxidative stress, such as alpha-beta-unsaturated aldehydes like 4-HNE, also activate HSF1 through similar mechanisms involving the oxidation of certain amino acid residues. In addition, HSF1 activation may be induced indirectly by nitric oxide, which inactivates HSP90 through s-nitrosylation of cysteine 597, in turn promoting the dissociation of HSF1 from the HSP90 chaperone complex [123].

Protective effects of heat shock proteins against oxidative stress

In regard to the antioxidant effects of heat shock proteins, several unique modes of action have been proposed. For example, HSP27 contributes to intracellular glutathione reduction by increasing the activity of glutathione reductase (GSR). This increase in activity may result from the stabilizing effect that HSP27 chaperones have on the enzyme G6PDH, which leads to increased levels of the GSR cofactor NADPH. Additionally, although the underlying mechanism remains unclear, HSP27 reduces intracellular iron levels, further protecting cellular structures from oxidative stress associated with Fe²⁺-mediated generation of hydroxyl radicals [10].

In addition to the direct antioxidant effects of HSPs, protection from oxidative stress is also accomplished through the stabilization or degradation of misfolded proteins. Similar to other heat shock proteins, HSP70 identifies misfolded proteins by recognizing exposed hydrophobic residues and utilizing ATP-dependent mechanisms to refold them. In addition, co-chaperones such as Bag-1 and CHIP can modify the effects of HSPs on misfolded proteins, allowing for protein degradation in cases of significant damage. For example, CHIP inhibits the ATPase activity of HSP70 and utilizes its E3 ubiquitin ligase activity to ubiquitinate oxidized proteins and direct them to the proteasome for degradation [45].

HSPs also play a protective role in oxidative stress-induced apoptosis through interactions with apoptotic proteins. For example, HSP70 and HSP90 bind Apaf-1 and inhibit its downstream apoptotic effects [164]. In addition, HSP27 inhibits cytochrome C-mediated apoptosis, either through direct binding to cytochrome C or by inhibiting its release from mitochondria [69, 147]. Heat shock proteins also counteract activation of inflammatory pathways following oxidative stress. For instance, HSP70 suppresses both the upregulation of Cox-2 and the production of nitric oxide in response to oxidative stress by inhibiting NF-κB gene activation [76, 196].

In sum, HSPs are activated by redox status both directly and indirectly and may protect from oxidative stress via antioxidant properties, correction of misfolded proteins, and their anti-apoptotic effects in the context of cellular stress. With growing appreciation for the role of oxidative stress in prenatal stress effects on offspring brain [28], HSP activation in developing neural cells may be similarly implicated.

4. Mechanisms of prenatal stress outside of the developing brain

4.1. Maternal hormonal and inflammatory factors in prenatal stress

While embryonic brain is a potential site of heat shock mechanisms in prenatal stress effects, HSPs and oxidative stress also have a role in the complex physiology of maternal stress factors that constitute a major waypoint in prenatal stress effects. Glucocorticoids in maternal circulation are critically elevated with prenatal stress [166, 170, 200], and there is considerable interest in the complex crosstalk between glucocorticoid and heat shock signaling pathways. GR-mediated gene expression may occur in multiple maternal tissues and is significantly enhanced by HSF-1, suggesting a potential for synergy between these two pathways in response to stress [91]. Consistent with this notion, HSP70 and HSP90 facilitate the assembly of GR heterocomplexes and ensure proper conformation for steroid binding [48, 133]. Additionally, HSF1 upregulates FK506-binding protein-52, which helps translocate GR to the nucleus following ligand binding [91]. On the other hand, GR activation inhibits the transcriptional activity of HSF1 in response to both heat and chemical shock by preventing the binding of HSF1 to Hsp promotors [109, 110]. Evidently, these two pathways exhibit reciprocal regulation, wherein heat shock proteins act to potentiate the effects of the GR under low-glucocorticoid concentration conditions while high concentrations of glucocorticoids act to suppress the heat shock response [108]. Furthermore, maternal cortisol release depends on maternal adrenal redox signaling [6] and may be influenced by the response of HSPs to pro-oxidants. Lastly, GR activation suppresses NRF2 transcriptional activity [4], one of the major mediators of endogenous antioxidant upregulation. Therefore, influences of maternal glucocorticoids during prenatal stress that ultimately affect the embryonic brain may be intimately affected by heat shock responses and may influence oxidative stress response. These may involve the interaction of GR activation, HSR, and antioxidants in other maternal cells that have a significant role to play in the cascade of prenatal stress effects (e.g. immune cells, maternal brain, vasculature). Consequences may also result from direct passage of maternal glucocorticoids to the embryonic brain or through effects of glucocorticoids on the placenta, as discussed in detail in the next section.

Psychological stress is also associated with immune activation and increased levels of pro-inflammatory cytokines such as TNF-α and IL-6 [47]. Furthermore, both stress and inflammation can result in offspring with alterations in brain function [136]. The immune component of maternal physiology during stress may depend on HSR systems. Maternal immune cells may be activated by extracellular HSPs and use heat shock proteins and redox signaling intracellularly in the process of cytokine release [206].

Preeclampsia, a pregnancy disorder characterized by hypertension and proteinuria, is associated with adverse neurodevelopmental outcomes in offspring [141] like maternal psychological stress. In women with preeclampsia, heat shock systems have been highly implicated [83, 144, 163]. Screenings aimed at identifying markers associated with preeclampsia have identified differential expression of HSPs in maternal blood as a common characteristic among patients [82]. In mothers with pregnancies complicated by preeclampsia, circulating levels of HSP70 are significantly correlated with increased pro-inflammatory cytokines, further confirming the link between heat shock systems and inflammation [131].

Overall, maternal factors involved in prenatal stress effects on the developing brain may be both a product of and result in heat shock and oxidative stress. The role for these cellular stress systems in the maternal response to psychological stress suggests more potential for translatable interventions, given the more immediate access of maternal physiology to intervention. However, maternal physiology cannot be considered in isolation; its close partner in supporting embryonic development, the placenta, must be considered in translational efforts as well.

4.2. Placenta as a mediator of prenatal stress

The placenta’s role in regulating maternal-fetal interactions makes it a prime target for understanding the mechanisms by which maternal stress alters neurodevelopment, including the role of HSPs and oxidative stress. The placenta has numerous functions, including the transfer of nutrients and oxygen to the fetus, protection of the fetus from maternal factors, and the secretion of growth factors and hormones that regulate fetal growth [63]. Consistent with these roles, maternal stress is associated with reduced birth weight and altered nutrient transfer, suggesting that altered placental functioning may underlie the mechanisms responsible for altered neurodevelopment [34, 121, 140]. Furthermore, given that the placenta consists of cells of both embryonic and maternal origin, sex differences in placental function are present throughout gestation, mediating sex-specific responses to maternal stress [13, 161].

Maternal stress has been associated with increased placental production of proinflammatory cytokines, which may upregulate heat shock protein expression in the embryonic brain. For example, rodent models of prenatal stress have demonstrated significant increases in the placental production of IL-6, IL-1β, and TNF-α [35, 135]. Placental production of cytokines in response to stress has been directly linked to altered neurodevelopment, as demonstrated by the protective effects of anti-inflammatory treatments on offspring neurodevelopment [35]. Consistent with this observation, proinflammatory cytokines have been demonstrated to upregulate the expression of placental HSPs [17, 155]. Placental regulation of embryonic inflammation, therefore, may depend on HSPs. Oskvig and colleagues demonstrated that increased maternal cytokine levels after lipopolysaccharide (LPS) administration, a model of bacterial infection, did not change expression of inflammatory mediators such as IL-6, IL-1B, and TNF-a in the embryonic brain. Furthermore, cytokine level changes in this model diminished in magnitude from maternal serum to amniotic fluid to embryonic brain, indicating that maternal cytokine fluctuations are buffered, perhaps due to placental actions [142] that may be influenced by the HSR.

Active placental mechanisms maintain amniotic and fetal cortisol levels lower than maternal cortisol [21, 65]. 11β-Hydroxysteroid dehydrogenases (11β-HSD) regulate glucocorticoids by converting inactive cortisone into cortisol and vice versa. The principal isozyme expressed in the placenta is 11β-HSD2, which serves to protect the embryo from maternal fluctuations in cortisol [5, 22, 37, 182, 202]. Placental 11β-HSD2 expression and activity, which is an oxidative process, increases with gestational age [168, 171]. Importantly, IL-1β and TNF-α, which increase with maternal stress, appear to contribute to the regulation of 11β-HSD proteins [46, 53, 122]. While these pro-inflammatory cytokines augment the expression and reductase activity of 11β-HSD1 [46, 53], they negatively regulate the expression and activity of 11β-HSD2 in a dose-dependent manner [46]. Maternal anxiety, depression, and psychological stress in humans, as well as chronic maternal stress in rodents, have been demonstrated to reduce placental expression and activity of 11β-HSD2, in turn leading to fetal glucocorticoid overexposure [121, 148, 153]. On the other hand, maternal exposure to cold temperature in rats increases placental 11β-HSD2 expression [113], which is associated with placental HSP70 production, suggesting that placental HSR may regulate fetal glucocorticoid exposure.

Maternal stress is also associated with restricted umbilical artery blood flow, which can impair the transfer of flow-limited substrates such as oxygen and carbon dioxide between maternal and fetal circulation [137, 176]. Stress-induced elevations in maternal serotonin, a vasoconstrictor, may contribute to reduced blood flow [31, 32]. Placental hypoxia promotes oxidative stress and is linked to adverse neurodevelopmental effects seen in animal models of prenatal stress, such as abnormal neuronal migration and reduced myelination [16, 104]. Hypoxia in late pregnancy can affect signaling molecules in the fetal hippocampus, resulting in impaired neuronal migration, similar to the migration delays seen with prenatal stress or exposure to pro-inflammatory cytokines and glucocorticoids [68, 181]. Placental hypoxia during stress stimulates the release of prostanoids and cytokines, including IL-1β and TNF-α [122, 160], but depending on developmental timing, may alternatively cause no change to placental cytokine production [68]. Consistent with the role of HSPs in mitigating oxidative stress, Hsp70 is upregulated alongside biomarkers of oxidative stress in placental tissues of patients with preeclampsia [15, 80, 143]. In addition, upregulation of Hsp27 in the placentas of preeclamptic patients has also been noted in several studies [64, 174]. The links between hypoxia, inflammation, the HSR, and oxidative stress as outlined in previous sections suggest a significant role for targeting these cellular mechanisms in the placenta to protect against prenatal stress which may be similar to these demonstrated effects of hypoxia and preeclampsia.

In addition to reduced overall placental blood flow, restrictions to glucose, lipid, and amino acid transfer have been observed in animal models of prenatal stress [121]. For example, altered activity of the placental glucose transporter GLUT-1 is found in both animal models of prenatal stress and humans [135]. Glutamate transporters in other tissues are induced through the HSR [70]. Thus, placental nutrient transporters may also be linked to the actions of HSPs. The impacts of prenatal stress also may precipitate through changes in placental oxygenation and oxidative state, both of which have demonstrated impacts on placental nutrient transport [8].

The placenta also has critical endocrine functions that participate in fetal neurodevelopment. In this regard, maternal stress alters placental production of neurotransmitters, growth hormones, neurotrophins, and cytokines, all of which have critical neurodevelopmental functions. The production of these diffusible factors—the placental secretome—depends on the viability and functioning of trophoblasts and other cells, which can be disrupted through oxidative stress and HSR mechanisms. The placental secretome also has direct impacts on embryonic brain where the HSR and ROS mediate responses. For example, maternal IL-1β decreases placental production of BDNF specifically in female offspring of rodent models of prenatal stress [73]. Given that previous studies have demonstrated a suppressive role of BNDF in the stress-induced upregulation of embryonic brain HSPs, decreased placental BDNF may enhance the heat shock response to hypoxia in fetal brain [103]. Animal models of prenatal stress have also demonstrated alterations in placental production of IGF-1 and IGF-2, which are crucial regulators of fetal growth and neurodevelopment [195]. In addition to its role in regulating fetal growth, IGF signaling has also been implicated in regulating expression of heat shock factors through an mTOR-mediated mechanism [42].

Maternal stress is also associated with increased serotonin, norepinephrine, and dopamine in maternal circulation [90]. These increases in circulating neurotransmitters are coupled with reduced MAO-A and increased serotonin transporters in the placentas of stressed mothers [30, 153]. Placental serotonin processing may be a critical mediator of maternal physiological impacts on embryonic brain [67, 194] and may be influenced by cellular stress mechanisms such as HSR and ROS. Serotonin can trigger the HSR in C. elegans via a temperature independent mechanism in some cell types [185] which may also occur in placenta and/or embryonic brain.

5. Heat shock systems, downstream mechanisms, and neuropsychiatric disorders

While it is unclear whether there is a direct connection between prenatal stress, the HSR, and eventual psychopathology in offspring, HSPs have been implicated in the etiology of several psychiatric disorders [19, 20, 40, 95, 97, 101, 107, 146, 169]. As previously discussed, HSPs are anti-inflammatory, anti-apoptotic, and attenuate the effects of oxidative stress [55, 203]. The HSR, perhaps because of elevated HSP70, prevents activation of NF-κB and downstream inflammatory response [77]. Indeed, hyperthermia-induced HSR protects against inflammation caused by combined injection of LPS, IL-1β, and IFNγ in mice [77]. NF-κB is also a mediator of oxidative stress, and induced HSP70 attenuates the effects of this system by reducing NF-κB translocation to the nucleus [55]. These links suggest that HSPs may have a role as biomarkers for these other contributing mechanisms.

Important to psychiatric disorders, the HSR causes nuclear translocation of GRs [165], and HSPs play a critical role in GR folding, transport, and function, as previously described. For this reason, HSP dysfunction is implicated in the etiology of many stress-related psychopathologies, particularly affective disorders.

5.1. Affective disorders

A higher risk for bipolar disorder is associated with prenatal stress [100]. Peripheral blood lymphocytes in individuals with bipolar disorder have lower HSP70 and high GR-HSP70 heterocomplexes; these same cells had low cytosolic BAX, suggesting a susceptiblility to apoptosis [19, 20]. In addition, Cheng and colleagues found elevated serum HSP60 and HSP70 in a cohort of hospitalized patients with bipolar disorder [40]. Gene expression for HSP40, which interacts with HSP70, is also elevated in brain and peripheral blood cells of individuals with bipolar disorder [87]. An integrated transcriptome and methylome analysis of peripheral blood cells also implicated HSPA1L (HS70 protein-1 like) in youth with or at risk for bipolar disorder [59].

Major depressive disorder is associated with prenatal stress [156] and HSP70 abnormalities [173]. Pae and colleagues found that HSP70 single nucleotide polymorphisms are associated with worse symptomology and poorer response to antidepressant treatment [146], suggesting lower HSP70 expression and less induction of HSP70 upon treatment. However, higher baseline serum HSP70 in women predicted the development of depression over a three year period [150]. In mice, knockout of HSF1, which may affect HSP70 induction, results in depressive-like phenotypes [190]. There is significant interest in further understanding the role of HSPs in the stress susceptibility that contributes to bipolar and major depressive illness.

A particularly well-studied HSR protein that may play a role in stress susceptibility and neuropsychiatric disorders is FK506 binding protein 51 (FKPB5). FKBP5 is a co-chaperone of HSP90, which regulates GR sensitivity. Others have reviewed in detail how dysregulation of FKBP5 may disrupt normal cortisol signaling, contributing to the development of affective disorders [27, 125]. In fact, a meta-analysis found that three specific polymorphisms in the FKBP5 gene confer susceptibility to early life stress, causing an increased risk for developing PTSD or depression in response to stressful events [197]. This suggests that early life stressors (including prenatal stress) and genetic predisposition may work synergistically to promote mood and anxiety disorder development [125].

5.2. Schizophrenia

Prenatal stress contributes to risk for schizophrenia [56], and the pathophysiology of schizophrenia may result from reduced protective capacity of HSPs during inflammatory and oxidative challenges to neurodevelopment. Consistent with this, antibodies to HSP70 and HSP90, which may disrupt their function, are elevated in individuals with schizophrenia [96, 169]. Polymorphisms in the HSP70 gene are also associated with schizophrenia in multiple populations of different ethnic backgrounds [97, 101, 145]. HSPA12A, a member of the HSP70 family expressed highly in neurons, particularly in the frontal cortex, is significantly reduced in the dorsolateral prefrontal cortex in individuals with schizophrenia [154]. Interestingly, induced pluripotent stem cells derived from patients with schizophrenia show greater variability in Hsp70 transcription when exposed to methylmercury or ethanol [74] which may suggest dysfunctional upstream regulation in response to stress, perhaps by HSF1. As previously described, HSF1 is an upstream regulator of the HSR and HSP70 in particular [205]. Environmental insults including ethanol, methylmercury, and induced seizure exposure increase HSF1 binding to the Hsp70 promoter in rodent embryonic cortex and human fetal cortical cells [74]. HSF1 knockout disrupts normal cortical development and increases vulnerability to seizures from these exposures, most likely due to cell cycle arrest or premature exit, as well as increased apoptotic activity [74]. These mechanisms may contribute to the risk developing schizophrenia from prenatal adversity.

5.3. Childhood disorders

ASD and ADHD are child onset neuropsychiatric disorders with links to prenatal stress [1]. However, few studies have examined the heat shock system in peripheral tissues, genetics, or postmortem brain of individuals with ADHD or ASD. One finding that does address similar links to those found in schizophrenia is that antibodies to HSP90 are observed in some children with ASD [54]. Additionally, HSP70 is also elevated in some children with ASD, similar to individuals with mood disorders [51]. Deeper investigation of the links between heat shock systems and these early onset disorders may yield not only important biomarkers but also potential pathophysiological mechanisms that could be targeted in treatment development.

6. Conclusion

Intrinsic biological responses that protect against cellular stress are prime candidates for elucidating both how adversity impacts prenatal brain development and how resultant disruptions could be prevented. These include cellular responses to oxidative stress, notably heat shock systems. As a result, HSPs and antioxidants are targets for drug development to treat diseases of the brain and many other systems. Interestingly, dysregulation of heat shock and redox regulation systems may contribute to the incidence of neuropsychiatric disease. On the other hand, heat shock and oxidative stress may also be indirectly induced during pathophysiological processes because of their important connections with multiple cellular functions. This latter possibility highlights how these cellular mechanisms have the capacity to protect the brain. To consider how targeting HSPs and antioxidants could prevent neurodevelopmental disruptions of prenatal stress, substantial progress is necessary in understanding the role of heat shock systems in specific cell subtypes and processes in embryonic brain. In addition, the complexities of the interacting systems in prenatal brain development – from maternal physiology to placenta function to embryonic growth – also require coordinated discovery of the distinct roles of the heat shock and oxidative stress responses in each system.

Highlights.

• Prenatal maternal stress increases risk for neuropsychiatric disorders in offspring

• Heat shock factor and oxidative stress mechanisms are implicated in prenatal stress

• These cellular stress mechanisms may be targets for neuropsychiatric interventions