Role of glucocorticoid- and monoamine-metabolizing enzymes in stress-related psychopathological processes

Abstract

Glucocorticoid signaling is fundamental in healthy stress coping and in the pathophysiology of stress-related diseases, such as post-traumatic stress disorder (PTSD). Glucocorticoids are metabolized by cytochrome P450 (CYP) as well as 11-β-hydroxysteroid dehydrogenase type 1 (11βHSD1) and 2 (11βHSD2). Acute stress-induced increase in glucocorticoid concentrations stimulates the expression of several CYP sub-types. CYP is primarily responsible for glucocorticoid metabolism and its increased activity can result in decreased circulating glucocorticoids in response to repeated stress stimuli. In addition, repeated stress-induced glucocorticoid release can promote 11βHSD1 activation and 11βHSD2 inhibition, and the 11βHSD2 suppression can lead to apparent mineralocorticoid excess. The activation of CYP and 11βHSD1 and the suppression of 11βHSD2 may at least partly contribute to development of the blunted glucocorticoid response to stressors characteristic in high trait anxiety, PTSD, and other stress-related disorders. Glucocorticoids and glucocorticoid-metabolizing enzymes interact closely with other biomolecules such as inflammatory cytokines, monoamines, and some monoamine-metabolizing enzymes, namely the monoamine oxidase type A (MAO-A) and B (MAO-B). Glucocorticoids boost MAO activity and this decreases monoamine levels and induces oxidative tissue damage which then activates inflammatory cytokines. The inflammatory cytokines suppress CYP expression and activity. This dynamic cross-talk between glucocorticoids, monoamines, and their metabolizing enzymes could be a critical factor in the pathophysiology of stress-related disorders.

Lay summary

Glucocorticoids, which are produced and released under the control by brain regulatory centers, are fundamental in the stress response. This review emphasizes the importance of glucocorticoid metabolism and particularly the interaction between the brain and the liver as the major metabolic organ in the body. The activity of enzymes involved in glucocorticoid metabolism is proposed to play not only an important role in positive, healthy glucocorticoid effects, but also to contribute to the development and course of stress-related diseases.

Keywords:

• Post-traumatic stress disorder (PTSD)
• cytochrome P450 (CYP)
• 11β-hydroxysteroid dehydrogenase (11βHSD)
• monoamine oxidase (MAO)
• inflammatory cytokines

Introduction

Stress has a key role in the etiology of numerous central nervous system (CNS) disorders including depression (Checkley, 1992; Tafet & Nemeroff, 2016), chronic fatigue syndrome (Fenouillet et al., 2016), anxiety disorders (Jezova & Hlavacova, 2008), and post-traumatic stress disorder (PTSD) (Brooks, Amlot, Rubin, & Greenberg, 2018). Similarly, stressful life events influence the development and course of many somatic diseases (Jezova & Herman, 2016; Roy, Riley, & Sinha, 2018; Simas, Nunes, Crestani, & Speretta, 2018).

Stress was originally defined as a nonspecific neuro-endocrine response of the body to the stimuli disturbing its homeostasis (Selye, 1973). Later on, the term allostasis has been introduced and defined as the ability to sustain homeostasis in the presence of stressors (Sterling & Eyer, 1988). Correspondingly, allostatic load is a burden on the body’s intrinsic systems required to sustain homeostasis under stress (McEwen, 2000). When allostatic load overcomes the body’s capabilities, subsequent exposure to stressors leads to the development of stress-related disorders. The vulnerability to stress-related disorders is determined by the nature, intensity, and duration of the stress situation and also by the organism’s intrinsic properties, such as interactions between biological molecules responsible for the allostasis (Jezova, Jurankova, Mosnarova, Kriska, & Skultetyova, 1996).

McEwen (McEwen, 2000) described four major types of allostatic load characterized by different glucocorticoid response patterns to stressors. The first type (“major hits”) results from repeated exposure to unknown stressful stimuli and it is characterized by recurrent increase in glucocorticoid concentrations. The second type (“lack of adaptation”) is due to an abnormally high glucocorticoid response after repeated exposure to known stressors. The third type (“prolonged response”) demonstrates abnormally long duration of glucocorticoid response to stressors. Finally, the fourth type of allostatic load (“inadequate response”) reflects an abnormally low glucocorticoid response to stressors, and this insufficient response to stressors has been observed in the following patients; those with PTSD (Smith et al., 1989; Strohle, Scheel, Modell, & Holsboer, 2008; Zaba et al., 2015), depression (Lopes et al., 2012), allergy (Buske-Kirschbaum, Ebrecht, & Hellhammer, 2010; Hlavacova, Solarikova, Marko, Brezina, & Jezova, 2017), high trait anxiety (Duncko, Makatsori, Fickova, Selko, & Jezova, 2006; Jezova, Makatsori, Duncko, Moncek, & Jakubek, 2004), and panic disorder (Jezova, Vigas, Hlavacova, & Kukumberg, 2010; Petrowski, Herold, Joraschky, Wittchen, & Kirschbaum, 2010).

The development of a blunted glucocorticoid response to stressors might be therefore an important element in the pathophysiology of PTSD and other stress-related disorders. This review is summarizing the regulation of glucocorticoid signaling in relation to psychopathological processes, with a focus on glucocorticoid-metabolizing enzymes.

Factors regulating glucocorticoid signaling

Circulating glucocorticoid concentrations are determined by their biosynthesis in the adrenal cortex and by their metabolism in the local tissues and the liver. Increased glucocorticoid synthesis and/or decreased metabolism leads to hypercortisolism and decreased synthesis and/or increased metabolism produce hypocortisolism (Figure 1). In addition, glucocorticoid docking by corticosteroid-binding globulin is also an important mechanism regulating circulating glucocorticoid levels (Qian et al., 2011).

Figure 1. Role of production and metabolism of glucocorticoids in the regulation of their concentrations in the blood. The circulating levels of cortisol (CORT) are balanced by the production of this enzyme in the zona fasciculate of the adrenal cortex, its metabolism to dihydrorocortisol and tetrahydrocortisol with subsequent glucuronidation to tetrahydrocortisone glucuronide (THCG) in the liver and clearance of THCG via the kidneys. Decrease in cortisol biosynthesis and/or increase in its metabolism lead to hypocortisolism. Increase in cortisol biosynthesis and/or decrease in its metabolism lead to hypercortisolism.

Glucocorticoids penetrate biological cells and their nuclei and directly affect the expression of specific genes via the nuclear glucocorticoid receptors (GRs). Target organ sensitivity to glucocorticoids is therefore regulated by the density and activity of GRs. GR expression is auto-regulated by glucocorticoids so that hypercortisolism usually results in decreased and hypocortisolism in increased GR expression (Leistner & Menke, 2018; van Zuiden, Kavelaars, Geuze, Olff, & Heijnen, 2013). The first three types of allostatic load have increased glucocorticoid levels associated with decreased sensitivity to these hormones (McEwen, 2000). These glucocorticoid signaling changes are similar to those observed in Cushing’s (Granner, Wang, & Yamamoto, 2015) and glucocorticoid resistance syndromes (Nicolaides & Chrousos, 2018), but to a lesser extent. Finally, the fourth type of allostatic load has decreased glucocorticoid bioavailability and increased target organ sensitivity, and similar alterations have been observed in some stress-related disorders (McEwen, 2000).

GR affinity to glucocorticoids is regulated by co-chaperone proteins. These include FK506-binding protein 5 (FKBP5) and its overexpression leads to decreased GR sensitivity and impaired negative feedback regulation of the hypothalamic-pituitary-adrenocortical (HPA) axis (Matosin, Halldorsdottir, & Binder, 2018; Tatro, Everall, Kaul, & Achim, 2009). It is possible that GR modulation by FKBP5 has a role in allostatic load and in stress-related pathophysiological processes because altered FKBP5 gene expression has been recorded in PTSD patients (Kang, Kim, Choi, So, & Kim, 2019) and also in patients with depression (Rao et al., 2016) and alcohol-use disorder (Qiu et al., 2016).

Although GRs are the primary target of glucocorticoids, these hormones also bind to mineralocorticoid receptors (MRs). Actually, MRs have higher affinity to glucocorticoids than GRs (de Kloet, Reul, & Sutanto, 1990). The MR-mediated glucocorticoid signaling boosts attention and memory retrieval; these factors are important for the initiation of the behavioral response to a stressful stimulus (Oitzl & de Kloet, 1992). In contrast, GR-mediated signaling regulates memory storage and consolidation (Nikzad, Vafaei, Rashidy-Pour, & Haghighi, 2011) and it is critical for termination of the stress response (de Kloet, Meijer, de Nicola, de Rijk, & Joels, 2018). The balance between GR- and MR-mediated signaling is therefore important in the stress response, and its disturbance can contribute to stress-related psychopathological processes.

Regulation of glucocorticoid signaling by 11-β-hydroxysteroid dehydrogenase

The balance between GR- and MR-mediated glucocorticoid signaling is controlled by multiple factors. Of particular importance are the enzymes belonging to the of 11-β-hydroxysteroid dehydrogenase (11βHSD) family which act mainly at the local, tissue level (Chapman, Holmes, & Seckl, 2013). The 11βHSD has two isoforms: 11βHSD1 and 11βHSD2.

The 11βHSD1 isoform is localized on the luminal surface of the endoplasmic reticulum and it is nicotinamide adenine dinucleotide phosphate (NADPH)-dependent. It is highly expressed in the brain, adrenal glands, liver, and adipocytes and it converts inactive cortisone to cortisol. Its main NADPH source comes from the glucose-6-phosphate-dehydrogenase (H6PDH) reaction, and therefore glucose-6-phosphate deficiency is often accompanied by decreased 11βHSD1 catalytic activity (Tu et al., 2008). 11βHSD1 over-expression in white adipose tissue induces hyperphagia (Michailidou et al., 2012) and it is also associated with enhanced liver gluconeogenesis (Draper & Stewart, 2005). Moreover, the 11βHSD1 isoform contributes up to 30% of extra-adrenal cortisol synthesis in healthy humans (Basu et al., 2004). In the brain, 11βHSD1 is expressed in brain neurons and glial cells, and relatively high densities of this enzyme have been recorded in the pituitary, hypothalamus, and hippocampus (Varga et al., 2016; Wyrwoll, Holmes, & Seckl, 2011).

The 11βHSD2 isoform interacts with MRs and is particularly active as a selective MR protective factor in aldosterone target tissues (Edwards et al., 1988; Funder, Pearce, Smith, & Smith, 1988) and it provides local conversion of bio-active glucocorticoids into the inactive ketone form. It is localized in the cell endoplasmic reticulum, but also profoundly expressed in the placenta and fetal tissues where it creates a barrier preventing maternal glucocorticoid transfer to the fetus (Wyrwoll et al., 2011; Yehuda & Seckl, 2011). Varga et al. (2013) also recorded higher 11βHSD2 expression in the rat’s early postnatal period.

The 11βHSD2 isoform was long considered absent in brain tissue until highly sensitive mRNA probes revealed its expression in the nucleus of the solitary tract (NTS) (Evans et al., 2016; Gasparini et al., 2018; 2019; Geerling & Loewy, 2007). NTS is involved in the regulation of electrolyte balance, salt appetite, and blood pressure. Since glucocorticoid concentrations in the brain are 100–1000 times higher than those of mineralocorticoids and the affinity of MRs to glucocorticoids and mineralocorticoids is comparable, the MRs located in brain areas not expressing 11βHSD2 are glucocorticoid-preferable (Geerling & Loewy, 2008). Co-expression of 11βHSD2 and MRs makes NTS neurons sensitive to mineralocorticoids, via a mechanism involving intracellular inactivation of glucocorticoids (Geerling & Loewy, 2007). The 11βHSD2-mediated mineralocorticoid-sensitivity of MRs was considered unique for the NTS. A more recent study however reported, that 11βHSD2, as well as 11βHSD1, are also expressed in the hypothalamus and hippocampus (Varga et al., 2016). In the same study, the highest concentration of 11βHSD2 mRNA was detected in the hippocampus of neonatal vasopressin-deficient Brattleboro rats, suggesting that hippocampal MRs might be mineralocorticoid-sensitive as well, particularly in the early postnatal period and/or under specific physiological or pathological conditions, such as low vasopressin concentrations.

The affinity of 11βHSD2 to the substrate is one hundred times higher than that of 11βHSD1 (Chapman et al., 2013). Therefore, in the structures in which these enzymes are co-expressed, such as hypothalamus and hippocampus, 11βHSD2 might play a dominant role in glucocorticoid balance. Indeed, the administration of glycyrrhetinic acid, a nonspecific inhibitor of 11βHSD1 and 11βHSD2, decreased CRH concentration in the hypophysial portal blood (Seckl, Dow, Low, Edwards, & Fink, 1993), via a mechanism presumably involving hypothalamic 11βHSD2 rather than 11βHSD1.

The inhibition of 11βHSD2 led to the state of apparent mineralocorticoid excess (Ferrari, 2010). Moreover, Evans et al. (2016) demonstrated, using NTS-specific knockout mice, that reduced 11βHSD2 activity in the brain did not intrinsically cause hypertension, but stimulated salt intake and the transition from salt resistance to salt sensitivity. The authors then suggested that 11βHSD2-expressing neurons integrate salt appetite and the blood pressure response to dietary sodium through an MR-dependent pathway.

While the suppression of 11βHSD2 activity leads to increased mineralocorticoid signaling on a local level, 11βHSD1 inactivation in glucocorticoid-feedback sites results in stimulation of the entire HPA axis. The 11βHSD1 gene knockout in mice promoted adrenal hypertrophy and augmented ACTH-stimulated circulating corticosterone concentrations (Abrahams et al., 2012). It is therefore possible that 11βHSD1 expressed in the hypothalamic paraventricular nucleus (PVN), and perhaps in some other brain areas, affects both local and circulating glucocorticoid levels. Interestingly, patients with type 2 diabetes treated with selective 11βHSD1 inhibitors (Rosenstock et al., 2010), and also 11βHSD1 knockout mice, exhibited increased ACTH levels, and adrenal hypertrophy (Harris, Kotelevtsev, Mullins, Seckl, & Holmes, 2001). This was summarized in Yehuda & Seckl’s (2011) minireview supporting a metabolic hypothesis for low cortisol activation in stress-related psychiatric disorders.

Table 1 highlights that 11βHSD1 and 11βHSD2 modulate glucocorticoid signaling on both local tissue and systemic circulation levels, where increased 11βHSD1 tissue activity increases corticosteroid bioavailability, and augmented 11βHSD1 activity suppresses HPA axis and possibly promotes adrenal hypotrophy. In contrast, increased 11βHSD2 activity leads to decreased tissue glucocorticoid availability in aldosterone target cells.

Table 1. Regulation of glucocorticoid signaling by 11-β-hydroxysteroid dehydrogenase types 1 and 2 and by cytochrome P450.

Enzyme	Influenced by	Effects on glucocorticoid signaling
11βHSD1	Chronic stress (↑) Glucocorticoids (↑) Acetylcholine (↑)* Serotonin (5-HT; ↑)* Growth hormone (↑)* Norepinephrine (↓)* Dopamine (↓)*	Local effect: Tissue glucocorticoids (↑)
		Central effects: HPA axis hypo-reactivity (?) Adrenal hypotrophy (?)
		The central effects are attributed to the 11βHSD1 expressed in PVN neurons.
11βHSD2	Chronic stress (↓) Glucocorticoids (↓)	Local effects: Tissue glucocorticoids (↓) MR-mediated signaling (↓)
		Central effect: Apparent mineralocorticoid excess (↓)
		The central effects are attributed to the 11βHSD2 expressed in NTS neurons
CYP3A	Glucocorticoids (↑) Central monoamines (↓)* Some antidepressants (↓)* Oxidative tissue damage (↓) Inflammatory cytokines (↓)	Central effect: Circulating glucocorticoids (↓)
		This effect is attributed to the CYP expressed in the liver

Repeated stress-induced glucocorticoid release may result in activation of 11-β-hydroxysteroid dehydrogenase type 1 (11βHSD1) and cytochrome P450 type 3A (CYP3A; CYP3A1 and CYP3A2 in rodents and CYP3A4 and CYP3A5 in humans) and also in inhibition of 11-β-hydroxysteroid dehydrogenase type 2 (11βHSD2). Hepatic CYP activation decreases the circulating levels of glucocorticoids, as observed in some animal models of post-traumatic stress disorder (PTSD). Since 11βHSD1 knockout or inhibition increases adrenocorticotropic hormone (ACTH) levels and induces adrenal hypertrophy, the activation of this enzyme by chronic stressors may have the following opposite effects: adrenal hypotrophy and hypothalamic-pituitary-adrenal (HPA) axis hypo-reactivity. Although the adrenal hypotrophy was indeed observed in an animal model of PTSD, the role of 11βHSD1 in chronic stress-induced HPA axis hypo-reactivity has not yet been directly examined. NTS: nucleus of the solitary tract; ↑: stimulatory effect; ↓: inhibitory effect; *this effect was observed for the enzyme expressed in the liver.

Obut et al. (2009) suggested that repeated, but not acute, stressors activated 11βHSD1 and inhibited 11βHSD2, and other authors consider that the glucocorticoid stimulatory effect on 11βHSD1 is mediated, at least in part, by the GRs (Pal’chikova, Kuznetsova, Selyatitskaya, Cherkasova, & Kuz’mina, 2016). Since 11βHSD1 inhibition results in HPA axis activation and adrenal hypertrophy, its activation by repeated stress may have opposite effects. It is therefore possible that low glucocorticoid concentrations observed in some animal models of PTSD (Boero et al., 2018; Kondashevskaya et al., 2017; Lazuko et al., 2018), and the adrenal hypotrophy reported in one of these models (Manukhina et al., 2018), were at least partly caused by increased 11βHSD1 activity. While it is not clear if 11βHSD2 inhibition results in altered circulating glucocorticoid levels, it may attenuate GR-mediated glucocorticoid signaling at the tissue level. These hypotheses, however, require clarification.

Cytochrome P450 and control of glucocorticoid concentration

Cytochrome P450 (CYP) is a microsomal and mitochondrial enzyme superfamily which catalyzes oxidation of various endogenous and exogenous biological molecules. These include steroid hormones, catecholamines, and medications including antidepressant drugs (Munro, McLean, Grant, & Makris, 2018; Rendic, 2002). The CYP3A subfamily comprises several microsomal enzymes involved in glucocorticoid metabolism and it irreversibly metabolizes corticosterone to 6β-corticosterone in rodents (isoforms CYP3A1 and CYP3A2) and cortisol to 6β-cortisol in humans (isoforms CYP3A4 and CYP3A5) (Peng et al., 2011).

CYP3A activity is stimulated by glucocorticoids (Dvorak & Pavek, 2010), and dexamethasone induces CYP3A mRNA and CYP3A immunoreactive protein expression in cultivated rodent (Kocarek, Schuetz, Strom, Fisher, & Guzelian, 1995) and human (Cooper, Cho, Thompson, & Wallace, 2008) hepatocytes, and in the rat liver (Kishida et al., 2008; Telhada, Pereira, & Lechner, 1992) and in the mouse liver (Yanagimoto, Itoh, Muller-Enoch, & Kamataki, 1992) and brain (Stamou, Wu, Kania-Korwel, Lehmler, & Lein, 2013).

In addition to its regulation by glucocorticoids, CYP3A activity is also modulated by other hormones and neurotransmitters, including monoamines. While α- and β-adrenoceptor agonists stimulated CYP3A1 and CYP3A2 expression in cultivated rat hepatocytes, their in vivo administration showed that the α-adrenergic agonist decreased CYP3A1 and CYP3A2 expression in the liver and in vivo β-adrenergic receptor antagonists increased CYP3A1 and CYP3A2 mRNA levels in that organ. These findings indicate that the complex network of central and peripheral pathways may override the hepatic adrenergic receptor effects (Daskalopoulos, Lang, Marselos, Malliou, & Konstandi, 2012). Accordingly, specific brain norepinephrine neuron lesion by 6-hydroxydopamine led to the activation of several hepatic CYPs, including CYP3A, while norepinephrine administration to the cerebral ventricle had the opposite effect.

It has also been suggested that central norepinephrine negatively regulates hepatic microsomal oxidation of glucocorticoids and that this effect is mediated at least in part by growth hormone (Kot et al., 2015). In addition to catecholamines, hepatic microsomal oxidation of glucocorticoids is also influenced by central 5-HT pathways, because selective lesion of 5-HT neurons and inhibition of 5-HT synthesis both led to robust activation of multiple CYPs, including CYP1A and CYP3A (Kot & Daniel, 2011). Dopamine appears to directly modulate hepatic CYP3A, CYP2C, and CYP2D activities by a D₂-receptor-mediated mechanism (Daskalopoulos et al., 2012); and more recently, the D₂ receptor antagonist sulpiride has been shown to down-regulate CYP1A1, CYP1A12, and CYP1B1 expression in the rat liver (Harkitis et al., 2015).

Peripheral catecholamines, including adrenal epinephrine, suppress microsomal oxidation of glucocorticoids, and Daskalopoulos et al. (2012) suggested that this effect may contribute to the increased levels of glucocorticoids in response to acute stressors. However, this is quite unlikely because peripheral catecholamines were not found to participate in HPA axis activation in acute stress (Jezova et al., 1987; Makara et al., 1986).

Next, microsomal oxidation of glucocorticoids by CYPs in both the brain and liver is also modulated by inflammatory cytokines, and it has long been accepted that the bacterial toxin lipopolysaccharide (LPS) can suppress CYP1A activity in the brain via mechanisms mediated by tumor necrosis factor alpha (TNF-α), interleukin-1α (IL-1α), interleukin-1β (IL-1β) and interleukin-6 (IL-6) (Nicholson & Renton, 1999). Finally, the LPS-induced downregulation of CYP3A in the liver also appears to be mediated by TNF-α (Li et al., 2008).

Since CYP is fundamental in the modulation of the circulating levels of glucocorticoids, it is possible that abnormal expression or activity of this enzyme is responsible, at least in part, for the blunted glucocorticoid signaling in certain stress-related disorders. In patients with depression showing a blunted response in the dexamethasone suppression test (DST) plasma half-life of dexamethasone is shorter compared to that in depressive patients exhibiting a normal DST response (Holsboer, Wiedemann, & Boll, 1986; Stokes, Sikes, Lasley, & Stoll, 2002).

Regulation of monoamine oxidize activity by glucocorticoids

As described above, the glucocorticoid metabolism effected by 11βHSD and CYP is controlled by monoamines which are metabolized by monoamine oxidase (MAO). The following three putative mechanisms enable glucocorticoids to modulate MAO activity.

First, MAO is a mitochondrial enzyme, and glucocorticoids control general mitochondrial functions. An example here is glucocorticoid dose-dependent regulation of three key characteristics of mitochondrial function: oxidation level, membrane potential, and calcium holding capacity (Du et al., 2009). Second, MAO activity is dependent on the lipid microenvironment, and the ability of glucocorticoids to regulate mitochondrial lipid peroxidation enables them to modulate MAO activity (Ekstedt & Oreland, 1976; Singer, 1995). Finally, glucocorticoids regulate MAO expression, and it has been established that dexamethasone increased krueppel-like factor 11 (KLF11) mRNA and protein levels in cultured neuronal cells. Over-expression of KLF11 resulted in increases in both the MAO-A mRNA concentration and its enzymatic activity (Grunewald et al., 2012; Harris et al., 2015). MAO activity correlated with the microsomal oxidation phenotype (Tseilikman et al., 2016) and the microsomal oxidation inhibitor proadifen reduced MAO activity in both the brain and liver (Kozochkin et al., 2017).

Potential changes in MAO activity in stress-related psychopathologies are not well understood. While one study recorded that increased MAO activity was detected in the platelets from patients with PTSD compared to healthy subjects (Davidson, Lipper, Kilts, Mahorney, & Hammett, 1985) a more recent study gave the opposite result (Pivac et al., 2007) and GR blockade in an animal model of PTSD diminished the stress-induced increase in MAO activity (Tseilikman et al., 2015).

Therefore, the elevation in glucocorticoid concentrations following acute stress exposure could contribute to the increase in MAO activity. This increase in MAO activity may hypothetically result in decreased monoamine transmission which is one of the pathophysiological mechanisms in several mood disorders.

The role of glucocorticoids, CYP, and MAO in oxidative tissue damage

It is well established that stress conditions increase the concentration of free radicals, and that enzymes of the CYP and MAO families generate reactive oxygen species (Benedetti Strolin & Tipton, 1999). It is therefore possible that increases in free radicals during stress are mediated by mechanisms involving changes in CYP and MAO. Treatment with both the GR blocker mifepristone (RU486) and the MAO type B (MAO-B) inhibitor selegilin diminished stress-induced production of reactive oxygen species, as measured by increased lipid peroxidation in the rat brain cortex (Tseilikman et al., 2015). Earlier studies showed that both MAO-B (Tazik et al., 2009) and MAO-A inhibitors reversed dexamethasone-induced apoptosis in cultivated human neuroblastoma cells (Johnson et al., 2010).

It is therefore most likely that stress-induced generation of reactive oxygen species by MAO-A and MAO-B is involved in the pathophysiology of certain neurodegenerative disorders, such as Parkinson disease (Goldstein, Jinsmaa, Sullivan, Holmes, Kopin, & Sharabi, 2016a, 2016b; Goldstein & Lieberman, 1992 ). In addition, apoptotic and neurodegenerative processes induced by oxidative tissue damage have also been reported in depression (Leonard & Maes, 2012; Leonard, 2017) and PTSD (Miller & Sadeh, 2014; Schumacher et al., 2015). As stated earlier, glucocorticoid-induced stimulation of CYP and MAO expression and activity can lead to increased concentration of reactive oxygen species, and these may contribute to the neurodegenerative and apoptotic processes observed in several stress-related disorders. CYP- and MAO-induced oxidative tissue damage is also likely to activate inflammatory cytokines including TNF-α, IL-1β, and IL-6 (Al Ghamdi et al., 2015; Megha et al., 2015; Tuorkey, Abdul-Aziz, & Zidan, 2013). It is well known that increased inflammatory cytokine levels activate the HPA axis (Serrats, Grigoleit, Alvarez-Salas, & Sawchenko, 2017) and down-regulate CYP expression (Morgan, 2001). This cytokine-induced inhibition of CYP suppression has a function in stress-related liver injury (Tseilikman et al., 2012), and more recently, the cytokine-induced activation of the HPA axis is reported to be blunted in both depression (Niemegeers et al., 2016) and panic disorder (Quagliato & Nardi, 2018).

Brain-liver axis: mechanisms involved

There are three potential mechanisms that enable monoamines to regulate hepatic glucocorticoid metabolism: (1) monoaminergic innervation of the cell bodies of sympathetic and parasympathetic nerves innervating the liver; (Jensen, Llewellyn-Smith, Pilowsky, Minson, & Chalmers, 1995), (2) direct monoaminergic innervation of the liver (Ruddell, Mann, & Ramm, 2008), and (3) neuroendocrine pathways connecting the brain and liver (Bromek, Wojcikowski, & Daniel, 2013; Rysz, Bromek, & Daniel, 2016).

Brain 5-HT neurotransmission and hepatic CYP activity have mutual influence. The selective lesion of 5-HT neurons or inhibition of 5-HT synthesis leads to robust activation of hepatic CYP (Kot & Daniel, 2011), and administration of the 5-hydroxytryptophan 5-HT precursor to the lateral cerebral ventricle has the opposite effect (Rysz et al., 2016). It can therefore be suggested that the functional deficit in 5-HT neurotransmission characteristic in depression and other stress-related disorders is associated with altered expression of CYP isoforms. This likelihood is supported by brain 5-HT neuron lesions leading to CYP3A over-expression (Kot & Daniel, 2011). Accordingly, the Flinder sensitive line (FSL) rats which serve as an animal model of depression are characterized by decreased extracellular monoamine levels (Dremencov, Weizmann, Kinor, Gispan-Herman, & Yadid, 2006; Dremencov et al., 2004; 2005) and also by increased hepatic CYP3A1 activity (Kotsovolou et al., 2010). Moreover, high hepatic CYP activity was found to be associated with high brain MAO-A activity and low plasma 5-HT levels (Tseilikman et al., 2016). Herein, Figure 2 highlights the neural and endocrine pathways modulating the interaction between brain monoamines and hepatic CYP, 11βHSD1, and MAO.

Figure 2. Regulation of hepatic glucocorticoid-metabolism enzymes by central monoamines. Dopamine (DA) neurons of the ventral tegmental area and serotonin (5-HT) neurons of the raphe nuclei innervate the dorsal nucleus of the vagus nerve. DA inhibits, and 5-HT-stimulates cholinergic (ACH) vagus efferents projecting to the liver, via dopamine-2 (D₂) and serotonin-3 (5-HT₃) receptors, respectively. Vagal ACH in turn stimulates hepatic 11-β-hydroxysteroid dehydrogenase type 1 (11βHSD1), most likely, via the nicotine acetylcholine receptors (NAChR). Serotonin also positively regulates somatostatin (ST) and growth hormone-releasing hormone (GHRH) secretion from the paraventricular (PVN) and arcuate (ARN) nuclei of hypothalamus, respectively. Norepinephrine (NE), originating from the locus coeruleus, induces opposite effects on the ST and GHRH. ST and GHRH are released into the hypophysial portal blood and these balance growth hormone (GH) secretion from the pituitary. GH is released into the general circulation and arrives in liver where it positively modulates 11βHSD1 and cytochrome P450 (CYP). Serotonin and NE projections incorporated in the parasympathetic and sympathetic efferents may also directly modulate hepatic CYP and monoamine oxidase (MAO). The serotonin-2A (5-HT_2A) receptors may be involved in direct 5-HT regulation of CYP and MAO activity.

Liver is excessively innervated, and CNS-liver projections may therefore be involved in the central control of hepatic glucocorticoid metabolism. The liver is innervated by post-ganglionic sympathetic nerves originating in the celiac and superior mesenteric ganglia, and the cell bodies of the pre-ganglionic nerves projecting to these ganglia are in the intermediolateral column of segments T7-T12 in the spinal cord. The cell bodies in the spinal cord receive serotonergic innervation, and 5-HT input may stimulate norepinephrine release from the post-ganglionic sympathetic neurons (Jensen et al., 1995). The main type of adrenergic receptor expressed in the liver is the β-adrenergic receptor; particularly the β₃-adrenergic receptor subtype (Douris et al., 2017; Negres et al., 2016). The α-adrenergic receptors, mainly the α₁-adrenergic subtype, are also present in the liver (Huan et al., 2017). In addition, Daskalopoulos, Malliou, et al. (2012) found that activation of the hepatic adrenergic receptors suppressed CYP expression, and it is therefore possible that 5-HT indirectly regulates hepatic CYP activity by activating the visceral adrenergic pathways that suppress CYP expression.

Hepatic parasympathetic fibers originate in the vagal dorsal motor nucleus (DMV) in the medulla oblongata and from the enteric ganglia located at the hepatic hilus and hilar spaces (Jensen, Alpini, & Glaser, 2013; Lautt, 2009 ). The DMV receives serotonergic and dopaminergic innervation (Babic & Browning, 2014), and Browning & Travagli (2019) provide supporting evidence of their functional role. One example is the authors’ 1999 in vitro result that 5-HT stimulated the activity of DMV neurons projecting to gastrointestinal organs (Browning & Travagli, 1999). This stimulatory effect on the gastrointestinal-projecting vagus neurons is most likely mediated by 5-HT₃ receptors (Mussa, Sartor, & Verberne, 2008; Mussa & Verberne, 2008). In contrast, dopamine may inhibit gastrointestinal-projecting vagal neurons; most likely by a D₁-mediated mechanism (Zheng & Travagli, 2007).

The post-synaptic parasympathetic neurons release acetylcholine, and both muscarinic and nicotinic cholinergic receptors are expressed in the liver (Koo & Liang, 1979; Soeda et al., 2012); an example here is that exposure to nicotine during gestation resulted in enhanced 11βHSD1 mRNA expression in the fetal liver (Xu et al., 2012). It is, therefore, possible that central 5-HT and dopamine regulate hepatic function, including 11βHSD1 activity, by 5-HT₃, D₁ and cholinergic nicotine receptors.

In addition to its indirect modulation of liver-projecting sympathetic and parasympathetic fibers, 5-HT has a direct regulatory influence on hepatic functions. The 5-HT fibers are incorporated in both sympathetic and parasympathetic hepatic nerves. Some hepatocytes express 5-HT_2A receptors, 5-HT transporters (SERT), and MAO; and the 5-HT_2A receptors there could be involved in mitochondrial toxicity and oxidative tissue damage (Ruddell et al., 2008).

Brain and liver also communicate through neuroendocrine mechanisms, GH has a major function in brain monoamine and hepatic CYP interaction. GH positively regulates CYP2C11 expression in the liver (Bromek et al., 2013; Rysz, Bromek, Haduch, et al., 2016), and its secretion from the pituitary is controlled by the hypothalamic inhibitory somatostatin peptide and the stimulatory growth hormone-releasing hormone (GHRH). While norepinephrine can activate the release of hypothalamic regulatory peptides with subsequent modulation of GH and GH-dependent CYP2C11 expression in the liver (Bromek et al., 2013), 5-HT in contrast, can either activate or inhibit this cascade (Rysz et al., 2016).

GH influence on liver enzymes can be important under stress conditions and its activated release in response to different stress stimuli is well known. There is an adaptation mechanism in the GH response to some types of stressors which involves the central dopamine system (Jezova, Radikova, & Vigas, 2007), and although this response to certain stressors was blunted in patients with panic disorder compared to healthy subjects (Jezova et al., 2010), direct evidence for GH’s role in the stress-induced down-regulation of CYP expression has only been established in fish species (Saera-Vila, Calduch-Giner, Prunet, & Perez-Sanchez, 2009). Further studies are therefore required to confirm its function in stress-induced regulation of CYP in mammals.

Interaction between brain monoamine transmission and liver CYP activity is reciprocal. Complementary to central monoamine regulation of hepatic CYP, the hepatic CYP activity can influence brain monoamine neuron excitability. We have recently found that CYP inhibition by proadifen hydrochloride (SKF525) resulted in inhibition of spontaneous firing of 5-HT neurons in the rat dorsal raphe nucleus (Grinchii, Paliokha, Tseilikman, & Dremencov, 2018).

Central mechanisms involved in the modulation of CYP activity by antidepressants

Although antidepressant drugs were primarily designated for treatment of depression, they have also demonstrated efficacy in the treatment of other stress-related disorders, such as generalized anxiety disorder (GAD), social anxiety disorder (SAD), and PTSD (Williams, Stein, & Ipser, 2018). The stress-induced increase in glucocorticoid levels activates MAO (Du et al., 2009) and this leads to decreased 5-HT levels which may subsequently stimulate CYP (Rysz et al., 2016). However, antidepressant drugs which stimulate central monoamine transmission may have opposite brain-mediated effect on hepatic CYP.

Liver CYP is the major metabolizer of antidepressants. As CYP substrates, antidepressants affect the activity of this enzyme, and the direct effects of different antidepressant drugs on the activity of hepatic CYP sub-types have been described in numerous experimental studies summarized in recent reviews by researchers including Ornoy & Koren (2018) and Nassan et al. (2016). However, at least some antidepressant drugs can affect hepatic CYP expression and activity by CNS-mediated mechanisms. Since CYP inhibition attenuates 5-HT neurotransmission (Grinchii et al., 2018), the inhibition of this enzyme by antidepressants could interfere with the primary therapeutic effect of these drugs on 5-HT transmission.

To the best of our knowledge, the effect of intracerebral administration of antidepressant drugs on hepatic CYP expression and activity has not yet been directly investigated. However, it has been reported that 5-hydroxytryptophan injection into the lateral cerebral ventricle increased brain 5-HT concentrations and diminished the activity of CYP1A, CYP2A, CYP2B, CYP2C11, and CYP3A and reduced the mRNA levels of CYP1A2, CYP2A2, CYP2C11, CYP3A1, and CYP3A2 in the liver (Rysz et al., 2016). It is, therefore, possible that intra-cerebroventricular administration of selective serotonin reuptake inhibitors (SSRIs) has a similar effect on liver CYP. Future direct experimental studies should test this hypothesis.

Other types of antidepressant drugs which do not directly enhance brain monoamine levels may act directly on the hepatic CYP without central mechanism involvement. One example here is Haduch et al.’s. (2006) finding that central effects are not involved in mirtazapine-induced increase in hepatic CYP2D activity.

Conclusions

Glucocorticoid-metabolizing enzymes may be at least partly responsible for development of the blunted glucocorticoid response to stress stimuli in patients with some stress-related psychopathologies. Two lines of evidence support this hypothesis. First, repeated stress-induced elevation in glucocorticoid levels results in activation of 11βHSD1 and inhibition of 11βHSD2, and this latter 11βHSD2 suppression can lead to apparent mineralocorticoid excess. Second, elevated glucocorticoid concentrations induced by stressors stimulates the synthesis and expression of CYP and MAO. Table 1 herein highlights that increased CYP expression and activity can lead to subsequent decrease in glucocorticoid concentrations due to their increased metabolism.

Acknowledgments

We thank Mr Raymond Marshall for the English language editing and Ms Katsiaryna Shabatsenka for the graphical design of the illustrations.

Disclosure statement

Authors declare no conflict of interest and no financial interest in the publication of this manuscript.

Additional information

Funding

The work of the authors was supported by the Russian Science Foundation [grant # 17-15-013418 to VT, ED, and OT], Scientific Grant Agency of Ministry of Education of Slovak Republic and Slovak Academy of Sciences [grant VEGA-2/0046/18 to and ED and LL] and Slovak Research and Development Agency [grant APVV-15-0388 to ED, LL, and DJ].

Notes on contributors

Vadim Tseilikman

Vadim Tseilikman, MSc (1982, Chelyabinsk State Pedagogical University, Chelyabinsk Russia), PhD (1992, Siberian State Medical University, Tomsk, Russia), DSc (1999, Institute of General Pathology and Pathological Physiology, Moscow, Russia), was a Head of Biochemical Department of the South Ural State Medical University (2005-2016, Chelyabinsk) and he is currently a Director of School of Medical Biology of the South Ural State University (2017-). His research is focusing in brain-liver interactions in stress;

Eliyahu Dremencov

Eliyahu Dremencov, MMedSc (Hebrew University, Jerusalem, Israel, 2000), PhD (Bar-Ilan University, Israel, 2004) is an Independent Research Fellow in the Institute of Molecular Physiology and Genetics (IMPG), Centre for Biosciences, and Institute of Experimental Endocrinology, Biomedical Research Center, Slovak Academy of Science (since 2013), Head of the IMPG Neuropharmacological Laboratory (2016-), Head of the IMPG (2018-), and a Head of the International Neurohepathological Laboratory in the School of Medical Biology, South Ural State University, Chelyabinsk, Russia (2017-);

Olga Tseilikman

Olga Tseilikman, PhD (2001, South Ural State Medical University, Chelyabinsk, Russia), DSc (2005, Omsk State Medical University, Omsk, Russia), is a Professor in the School of Medical Biology, South Ural State University, Chelyabinsk, Russia. Her research is focusing in brain-liver interactions in stress;

Michaela Pavlovicova

Michaela Pavlovicova, PhD (Comenius University in Bratislava, Slovakia, 20013) was a Research Fellow (2003-2015) and an Independent Research Fellow (2015-2016) in the Institute of Molecular Physiology and Genetics (IMPG), Centre for Biosciences, Slovak Academy of Sciences, Bratislava, Slovakia;

Lubica Lacinova

Lubica Lacinova, PhD (1989), DSc (2004), Professor (2018), is a Leading Research Fellow in the Institute of Molecular Physiology and Genetics, Centre for Biosciences, Slovak Academy of Sciences, Bratislava, Slovakia, and Professor of Biophysics at University of Saints Cyril and Methodius, Trnava, Slovakia;

Daniela Jezova

Prof. Daniela Jezova, PharmD is a full professor of Pharmacology at the Faculty of Medicine of Comenius University and the Head of the Laboratory of Pharmacological Neuroendocrinology being a part of the Department of Endocrine Regulations and Psychopharmacology of the Institute of Experimental Endocrinology, Biomedical Research Center, Bratislava, Slovakia. She has been a vice-president of the Slovak Academy of Sciences and vice-president of All European Academies. She is serving as an international expert in European research networks including the panel of European Research Council (ERC) on Endocrinology, Pathophysiology and Physiology. Her research team focused on stress research and psychopharmacology in both humans and animals belongs to the evaluated top research teams in Slovakia.