Corticosteroids secreted by the adrenals after stress exert a negative feedback action to suppress the enhanced HPA axis activity (Fig. 2.2). This phenomenon was demonstrated by a classical endocrine experiment in 1938 by Dwight Ingle (see Raff 2005). He was the first to show that ACTH was needed for adrenal growth and steroid secretion by administering the peptide to hypophysectomized animals. Next, Ingle showed that corticosterone given to the ACTH-treated animal did not affect adrenal weight while it suppressed adrenal weight in the intact animal. Hence, corticosterone exerted in high doses pituitary feedback on ACTH release.

Subsequent research demonstrated different levels of corticosteroid feedback operation. The first level is on the anterior pituitary level—as noted by Ingle—and mediated by GR expressed in the corticotrophs. This feedback site responds to potent synthetic glucocorticoid such as dexamethasone and very high levels of endogenous cortisol and corticosterone (de Kloet et al. 1974). The onset of suppression occurs with a delay of 30 min, and in case of dexamethasone, the suppression may last several hours, even more than 12 h as used in the dexamethasone suppression test with or without CRH (see Box 2.1). The rise and fall of the dexamethasone suppression test in endocrine psychiatry is wonderfully described in “The riddle of Melancholia” (Shorter and Fink 2010).

The second level is at higher brain regions harboring circuits that process stressful information and that communicate transsynaptically with the GABA-ergic network surrounding the PVN. The steroid feedback is complex in these circuits and operates over different time domains depending on the nature and severity of the psychological stressor. The coordinate action exerted by corticosteroids via membrane and genomic MR and GR adds to this complexity (Dallman and Hellhammer 2011; de Kloet 2014).

Figure 2.3 shows the role of MR and GR in processing of stressful information in the limbic brain with the goal to support coping and adaptation. Corticosteroids affect virtually every step from detection and perception of a salient event triggering emotional arousal and appraisal processes until coping, adaptation, and memory storage of the experience to be prepared if a similar encounter occurs in the future. Thus, first corticosteroids affect the detection threshold and perception of auditory information. Lack of steroids was found to enhance detection at the expense of perceiving the significance of the acoustic signal (Henkin and Daly 1968). Next, arousal is triggered (Pfaff et al. 2007) and is necessary for the limbic structures to function optimally in assessment of the valence of a novel experience and selection of an appropriate coping style.

Using a large variety of behavioral tests, Melly Oitzl et al. (Oitzl et al. 2010; Oitzl and de Kloet 1992) have carefully dissected the role of MR and GR during stress coping and adaptation. Thus, corticosteroid appeared to rapidly promote appraisal processes of newly acquired information, retrieval of contextual information, and selection of an appropriate coping style. Since these MR-mediated actions proceed rapidly, they most likely are exerted by the membrane receptor variant regulating excitatory transmission; GR becomes activated only with high amount of corticosteroids induced by stress. GR activation is important for restoring cellular homeostasis and promoting allostatic processes, behavioral adaptation, and memory storage of the experience and coping style. By doing so the input from higher brain regions subsides resulting in attenuation and at last termination of stress-induced HPA axis activity because of adaptation (Box 2.2 and Fig. 2.3).

Firstly, MR activation directs coping style. Lars Schwabe et al. (2010a) demonstrated that exposure to a stressful situation switches the coping style. Under resting conditions the rodent uses multiple cues in order to memorize the location of a food resource. If exposed to stress, the animal switches rapidly to a simpler stimulus response. In rodents the pathway activated chronically by a stressor switches from hippocampus toward the dorsal striatum supporting habit-like behavior (Dias-Ferreira et al. 2009). The phenomenon is also observed in humans: with fMRI it was shown that during stress the amygdala-hippocampus pathway rapidly switched to the amygdala-striatum connectivity (Schwabe et al. 2013; Vogel et al. 2015, 2016).

The switch from hippocampus to striatum was observed in males. If the same experiments were performed in females, the opposite results were obtained. Females under resting conditions were rather poor in spatial performance as compared to their male counterparts. Under stress the situation was reversed, females performed better, and these differences were eliminated in the MR forebrain knockout mice (ter Horst et al. 2012). Thus, context and sex determine the outcome of the MR-mediated functions in coping with stress. Anti-mineralocorticoids blocked the switch from hippocampus to striatum in rodents and man (Schwabe et al. 2010b, 2013; Vogel et al. 2015, 2016). Moreover, active vs passive coping style in mouse and rat lines correlates with MR expression in the hippocampus (Cabib and Puglisi-Allegra 2012; Veenema et al. 2003).

Secondly, GR activation promotes adaptation and memory storage. It appeared that GR-mediated effects on memory storage required the presence of noradrenaline (Joels et al. 2012; Roozendaal and McGaugh 2011). For this purpose GR mediates a plethora of activating and suppressive actions in discrete brain regions. Thus, in the CRH neurons of the amygdala GR stimulates the synthesis and release of CRH, while the reverse occurs in the PVN (Zalachoras et al. 2016). In the amygdala, GR promotes and extends MR-mediated glutamatergic excitation (Karst et al. 2010). In the hippocampal pyramidal neurons, MR enhances excitability, which is subsequently suppressed by subsequent stimulation of GR by higher concentrations of corticosteroids (Joels and de Kloet 1989, 1990, 1992). In addition multiple neuropeptide systems (oxytocin, vasopressin) are activated by stress which exert in specific behavioral domains their context-dependent effects on processes modulating the stress response. For instance, oxytocin stimulates bonding and social support, which facilitates coping with a stressful situation (Barrett et al. 2015; Young 2015).

Third, the limbic MR is important for the tone of the HPA axis and sympathetic nervous system. For instance, the higher the hippocampal MR expression, the lower the basal pulsatile and stress-induced HPA axis activation, and thus the average amount of corticosteroids secreted over 24 h is decreased. Under MR antagonists applied intracerebroventricularly enhance basal and stress-induced HPA axis activity (Ratka et al. 1989) and act as an anxiolytic (Korte et al. 1995) and anti-aggressive (Kruk et al. 2013) agent. MR antagonists also decrease the blood pressure response to a stressor (van den Berg et al. 1990). This effect mediated by MR appeared to depend on the condition of 30 min warming the animal which is needed to do a proper tail sphygmographic measurement of the blood pressure. Using this warming/stress condition of the indirect tail cuff method, the direct telemetric recording revealed that MR antagonist blocked autonomic outflow and, interestingly, now suppressed the stress-induced HPA axis response (de Kloet et al. 2000; Van den Berg et al. 1994).

Collectively, these observations have led to the formulation of the corticosteroid receptor balance hypothesis:

Upon imbalance in MR: GR-regulated limbic -cortical signaling pathways, the initiation and/or management of the stress response is compromised. At a certain threshold this may lead to a condition of HPA axis dysregulation and impaired behavioral adaptation, which can enhance susceptibility to stress-related neurodegeneration and mental disorders. (de Kloet 2014, 2016; de Kloet et al. 1991, 1998, 2005, 2016; de Kloet and Molendijk 2016; Holsboer 2000)

Genetic variants of MR, GR, and their regulatory proteins such as, e.g., FKBP5, have been identified that appeared associated with HPA axis regulation, emotional expressions, and cognitive performance. Genetic variation may alter control in the promotor and translation region and result in an altered primary structure. The GR variant N363S was found hypersensitive to cortisol and associated with an unhealthy metabolic profile, while E22/23EK is linked to steroid resistance and enhanced risk of depression. The Bcl-1 polymorphism predicts cardiovascular risk and contributes to individual differences in emotional and traumatic memories as well as PTSD symptoms after intensive care treatment (Quax et al. 2013).

In the MR gene, the rs5522 (minor allele frequency 12%) is an A/G SNP located in exon 2, which causes an amino acid change (I180V) in the N-terminal domain of the protein. Roel de Rijk discovered that this G-allele is a loss of function variant as shown by a reduced transactivation capacity in vitro (DeRijk et al. 2006). These G-allele carriers showed increased HPA axis and autonomic reactivity in response to psychological stressors. Moreover, Bogdan reported an association of MR gene variation with depressive symptoms and deficits in reward-motivated learning induced by stress and heightened stress-induced amygdala activity (Bogdan et al. 2010, 2012). Interestingly, the same G-allele is with a high odd ratio considered a risk factor in reverse remodeling in heart failure patients undergoing cardiac resynchronization therapy (De Maria et al. 2012).

Another MR SNP, rs2070951 (C/G), minor allele frequency 49.3%, is located 2 nucleotides before the translation start site. The G-allele produces less MR in vitro and is associated with increased renin and aldosterone and elevated blood pressure (van Leeuwen et al. 2010).

The rs5522 and rs2070951 are in linkage disequilibrium, and if merged, three common haplotypes can be identified. Haplotype (hap) 2 (CA, frequency 35%) is a gain of function variant as was shown from the increased transactivation capacity and increased translation of MR protein in vitro, while hap 4 (GG) is very rare and produces strongly reduced MR activity as compared to hap 1 (GA, frequency 49%) and hap 3 (CG, frequency 12%) (Hamstra et al. 2015; van Leeuwen et al. 2011).

Hap 2 carriers had lower scores on the Trier Inventory for Chronic Stress (TICS) subscales “excessive demands at work” and “social overload.” In females, hap 2 appeared associated with dispositional optimism, optimistic risk decision-making in gambling tests, less rumination, and less feelings of hopelessness. GAIN cohort study (N = 3600) has demonstrated that hap 2 carriers are protected from depression (Hamstra et al. 2015; Joels et al. 2008; Klok et al. 2011). Further, this haplotype moderates the effect of childhood maltreatment and depressive symptoms in a population-based cohort (N = 665) and an independent clinical cohort from the Netherlands Study of Depression and Anxiety (NESDA, N = 1639) (Vinkers et al. 2015).

Selye showed that a relative excess of mineralocorticoids was pro-inflammatory, while excess of glucocorticoids increased the risk for infection and expressed this view in the pendulum hypothesis. The balance hypothesis, however, is based on one single corticosteroid hormone which maintains homeostasis via two distinct and co-localized receptor types that carry the pharmacological activity of Selye’s two hormones: the MR and GR (de Kloet 1991). As was pointed previously, the MR in the brain, heart, and fat cells binds cortisol and corticosterone rather than aldosterone and does so with a tenfold higher affinity than GR.

Over the past 30 years, the MR/GR balance has been challenged using endocrine, pharmacological, and genetic approaches. The outcome of these challenges was measured on the molecular levels using genomic approaches and on the cellular level with neuroanatomical and electrophysiological techniques, and behavioral and physiological responses were recorded in a great variety of paradigms (de Kloet 2014, de Kloet and Joels 2016; de Kloet and Molendijk 2016; de Kloet et al. 2016; Joels et al. 2012).

Below are some general characteristics of MR/GR imbalance:

Genetically selected rat or mouse lines or strains that display overexpression of MR have a reduced HPA axis tone as expressed by lower basal and stress-induced levels of corticosterone (Harris et al. 2013; Veenema et al. 2003). The male animals have an active coping style if dealing with an inescapable stressor, a high sympathetic tone and reduced 5HT function (Veenema et al. 2003). They show less anxiety in the home environment and improved cognitive performance in maze learning and fear-motivated tasks. Their behavior once learned perseverates. This phenotype is mimicked in mice with forebrain MR overexpression, particularly in the face of reduced GR. It seems as if limbic overexpression of the MR facilitates during stress the switch from costly time-consuming declarative hippocampal learning and memory processes to a rapid and effective striatal habit performance as coping style. However, these dominant high MR expressing animals become prone to anxiety in novel situations where they have lost control (de Kloet et al. 2016).

Exposure to chronic stress decreases the expression of hippocampal MR. Likewise rats or mice exposed to adverse early life conditions have at later life reduced MR. Reduced hippocampal MR expression is observed at senescence and is a characteristic of the depressed patient’s hippocampus measured postmortem. Antidepressants increase the synthesis of hippocampal MR. Rats with viral overexpression of MR in the dentate gyrus showed improved short-term memory and were protected against the impairing effect of 3 weeks of corticosterone in a nonspatial object recognition paradigm (Ferguson and Sapolsky 2007). In mutant mice, forebrain MR overexpression restored impaired learning induced by chronic stress but only in a low arousing task. This behavioral change in the MR overexpression mice was paralleled by a normalization of hippocampal dentate gyrus function (Kanatsou et al. 2015).

That chronic stress affects the hippocampus is obvious from the profound neuroanatomical changes: the CA3 pyramidal neurons atrophy and dentate gyrus neurogenesis is reduced (McEwen 2016). Using microarrays it was found that the widely diverse gene patterns were reduced to only a few pathways that regulate chromatin organization, epigenetics, apoptosis, and inflammatory responses in the dentate gyrus. One highly responsive gene network revealed by this procedure is the mammalian target of rapamycin (mTOR) signaling pathway which is critical for different forms of synaptic plasticity and appears associated with depression (Datson et al. 2013; Polman et al. 2012).

Tinnitus is a phantom sound indicating malfunction of the central auditory system. The causes of tinnitus include damage to the inner ear and consequent changes in the auditory system. The damage may, for instance, be due to aging, noise exposure, infections, altered vascular integrity, and inflammatory responses because of hypertension or atherosclerosis and local head or neck injuries (Knipper et al. 2013).

The auditory system comprises the neuronal cochlear circuit connected with the auditory cortex via the olive nucleus and the midbrain geniculate nucleus. This circuit enables arousal via the brainstem-midbrain reticular system and communicates with limbic circuitry (McIntosh and Gonzalez-Lima 1998; Middleton and Tzounopoulos 2012). Each acoustic stimulus received by the ear and passed via the process of auditory transduction into the central auditory pathway undergoes assessment leading to emotional reactions. The majority of acoustic signals are evaluated as neutral but part is appraised with positive or negative emotional weight. This assessment is possible due to the connectivity of auditory brainstem and auditory cortex with limbic circuits in the amygdala, hippocampus, and prefrontal cortex regions (Kraus and Canlon 2012); it is active not only during the awake phase but also during phases of the non-REM sleep (Portas et al. 2000). Thus, the connectivity between the auditory and limbic systems is involved in detecting adversity, danger, and regulation of the HPA axis (Fig. 2.4).

In patients with disturbing tinnitus, the persistent phantom sound is continuously evaluated and classified by the limbic system as adverse and thus negative (Rauschecker et al. 2010). This may in turn lead to a long-term dysregulation of the HPA axis characteristic for a condition of chronic stress (Fig. 2.5). Some of the consequences of the tinnitus-induced chronic stress effects are, for instance, insomnia, as the limbic system signals danger and keeps the victims of tinnitus awake. It would be of interest to accommodate these findings to the current knowledge of the action of corticosteroids, since previously it has been reported that tinnitus patients display hypocortisolism upon exposure to severe psychosocial stressors (Hebert and Lupien 2007). One scenario is therefore that this “hypocortisolism” provides an insufficiently large cortisol signal that is not capable to control the central stress reaction evoked by auditory adversity (see Sect. 3). This would suggest the existence a cortisol sensitive tinnitus connectome or neuronal network underlying an acoustic-induced allostatic load/chronic stress phenotype (McEwen and Wingfield 2010). Recent evidence indeed suggests tinnitus-specific connectivity of a functional limbic neuronal network involved in processing of emotionally loaded and emotionally neutral acoustic information which could be the tinnitus signature of such an altered phenotype (Georgiewa et al. 2016).

Not all of the subjects with tinnitus are disturbed by its sound; however, those who are suffer greatly from tinnitus-related insomnia and concentration problems. In addition, about 50% of patients with tinnitus has additional mental comorbid condition(s) such as depression or anxiety (Pattyn et al. 2016; Zirke et al. 2013), and these are known to be set off or amplified by the emotional stress and MR/GR imbalance (de Kloet et al. 2016).

Tinnitus may cause emotional distress and, finally, stress-related pathology. At the same time, emotional exhaustion or the pathology accompanying posttraumatic stress disorder was suggested to be predisposing for tinnitus (Fagelson 2007; Hebert et al. 2012; Hinton et al. 2006). The argument is presented by the seminal experiments of Sylvie Hébert et al. (Hebert et al. 2012; Hebert and Lupien 2007; Mazurek et al. 2015). These authors reported a blunted cortisol response to the Trier Social Stress Test and enhanced suppression of the morning rise in cortisol by a low dose of exogenous dexamethasone administered at 11 pm on the previous day (Simoens and Hebert 2012). Collectively, these data reveal an HPA axis phenotype of tinnitus resembling that of fibromyalgia, chronic fatigue syndrome, posttraumatic stress syndrome, and atypical depression, which are all characterized by a relative underexposure to cortisol during stressful conditions (Chrousos and Gold 1992). Such a reduced cortisol secretion maybe the consequence of an overactive limbic MR conveying an enhanced inhibitory tone over the HPA axis. The recently uncovered cytokine signature of tinnitus would fit in a phenotype of an altered functional ratio of MR over GR activity (Betancur et al. 1995; de Kloet et al. 1994) causing prevalence of pro-inflammatory cytokine synthesis (Szczepek et al. 2014).

That cortisol is of relevance for auditory processing is known for a long time. Adrenally deficient patients were shown to have lower detection threshold in the frequencies 500 and 1000 Hz than the healthy controls but have a deficit in speech discrimination (Henkin and Daly 1968). This increased detection and decreased perception cannot be ameliorated by deoxycorticosterone, but the detection threshold was normalized upon ACTH, prednisolone, and fludrocortisone treatment, the latter with either dexamethasone, prednisolone, or cortisone (Henkin and Daly 1968). Corroborating this early clinical finding, recent study demonstrated that although the rats with impaired adrenal function have intact function of the outer hair cells in the inner ear, their distortion produces otoacoustic emissions (DPOAE). These adrenally deficient animals also have significantly elevated auditory brainstem responses (ABR) which are consistent with impaired tone and speech perception in people (Dogan et al. 2015) and indicative of neuronal processing rather than sensory malfunctioning. Further, in support of clinical findings, dexamethasone reversed this impairment auditory information processing. Hence, it seems that enhancing GR function contributes to reinstatement of normal auditory function. Another argument for a positive action of corticosteroids on the auditory system is the therapeutic use of synthetic corticosteroids (prednisone, dexamethasone) to treat inner ear illnesses such as sudden sensorineural hearing loss (SSHL)—a condition that is always accompanied by tinnitus (Hobson et al. 2016; Leung et al. 2016) or idiopathic tinnitus (Barreto et al. 2012; Dodson and Sismanis 2004).

Corticosteroid receptors are expressed in the inner ear and auditory networks in the brain. Several areas of the inner ear are richly endowed with MR and GR (Fig. 2.1) (Kil and Kalinec 2013; Terakado et al. 2011). The current notion is that glucocorticoids prevent the hearing loss via GR because of their anti-inflammatory and immunosuppressive action, while the aldosterone-selective MR is involved in maintenance of ion homeostasis required for optimal hearing (Meltser and Canlon 2011). Since 85% of subjects with tinnitus have some degree of hearing loss (Mazurek et al. 2010), it would be very interesting to examine whether prevention of hearing loss is connected with prevention of tinnitus. Also the cochlear neuronal network expresses differentially in discrete nuclei MR and GR. However, so far no systematic studies have been reported on the function of these brain receptors in the onset and modulation of tinnitus.

Synthetic corticosteroids are used since decades as systemic or local therapy for tinnitus. Dexamethasone and methylprednisolone are most commonly used, and the administration routes vary from per os, intravenous injection to intratympanic injections (see Table 2.1). Recent systematic review scrutinized clinical studies that used the latter method and concluded lack of effectiveness (Lavigne et al. 2016). However, the authors also recognized that the extreme heterogeneity of the clinical protocols and the lack of long-term follow-up undermined their disappointing conclusion.