Evidence of a Neuroinflammatory Model of Tinnitus

Abstract

Emerging literature has highlighted the relationship between inflammatory and neuroinflammatory biomarkers and tinnitus. Neuroinflammation may help to explain the mechanisms underpinning hyperactivity in the cochlea, cochlear nucleus, inferior colliculus, medial geniculate body, and the auditory cortex in those with tinnitus. Glial activation and pro-inflammatory cytokines may cause excitatory-inhibitory synaptic imbalance. Advancing our understanding of these mechanisms may help elucidate the pathogenesis of tinnitus and lead to improvement in subtyping subjective tinnitus. The chapter explores our current understanding of the neuroinflammatory model within the context of the classical auditory pathway and what we can infer about the underlying mechanisms based on these studies.

1. Introduction

Tinnitus is an auditory perception with no external stimulus. It often presents as a ringing or buzzing sound in one or both ears or in the head and can be intermittent. Tinnitus can be subjective or objective. Objective tinnitus can be categorized into subtypes based on known sound sources. These categories include neurological disorders (myoclonus), vascular disorders, temporomandibular disorders (TMD), and patulous Eustachian tube. Palatal myoclonus is a series of sporadic muscle contractions by the tensor veli palatini muscle in the soft palate. The condition can be categorized as either essential or symptomatic through symptomatic manifestation or the presence of physical lesions in the cerebellum or brainstem [1]. Stapedial myoclonus describes similar muscle contractions of the tensor tympani and the stapedial muscles leading to the subject hearing a constant clicking sound from the vibration of the tympanic membrane [2]. Some case reports suggest individuals with TMD and tinnitus can modulate their tinnitus through head or jaw movements. The pathogenesis for tinnitus remains unclear, but tinnitus can be alleviated by treating the TMD [3].

Patulous Eustachian tube (PET) often presents with aural fullness, autophony, and tinnitus [4]. PET is caused when the Eustachian tube fails to close. Vascular disorders cause pulsatile tinnitus. This occurs when the sound of blood flow becomes audible due to increased blood pressure disrupting laminar flow [5]. Compression of the ipsilateral internal jugular vein may accentuate or attenuate the sound of pulsatile tinnitus, depending on whether it is arterial or venous in etiology [6]. Pulsatile tinnitus, presenting with or without asymmetrical hearing loss, is regarded as a criterion for neuroimaging leading to a possible acoustic neuroma diagnosis, which may require surgical intervention [7, 8]. Pulsatile tinnitus is also a symptom of semicircular canal dehiscence alongside autophony, similarly to PET [9]. Table 1 shows a list of causes of subjective and objective tinnitus.

Subjective tinnitus	Objective tinnitus
Ototoxicity	Patulous Eustachian tube
Noise exposure	Neurological disorders
Traumatic event	Vascular disorders
Ménière’s disease	Superior semicircular canal dehiscence
Hearing loss

Table 1.

Known causes of subjective and objective tinnitus.

2. Possible subtypes of subjective tinnitus

The term “tinnitus,” on its own, typically refers to subjective tinnitus, which is a perceptual phenomenon with no physical counterpart or means to hear the sound by auscultation [10]. Subjective tinnitus can be idiopathic and can present as any type of sound, it can last from seconds to being constant. Subjective tinnitus is usually idiopathic but can present with a number of inflammatory diseases, one of such being Ménière’s disease. Low-frequency narrow band tinnitus noise can be one of the earliest symptoms of Ménière’s disease, which worsens in later stages, often becoming more intense before an episode of vertigo [10]. Cisplatin-based chemotherapy and many ototoxic medications can result in the development of subjective tinnitus, though the mechanisms underpinning this relationship remain unknown [11].

Subtyping tinnitus may be important for understanding tinnitus etiology and developing targeted treatments for each subtype, however, the rarity of subtypes (based on presentation) makes it difficult to conduct large-scale studies with representative samples. For example, “typewriter tinnitus” is a type of intermittent staccato-like tinnitus sound that has been thought to be related to vascular compression based on case-based neuroimaging and its response to the anticonvulsant carbamazepine [12]. However, there were only 12 cases of “typewriter tinnitus” recorded.

Subjective tinnitus is correlated with excitatory-inhibitory synaptic imbalance leading to hyperactivity in the auditory pathway; however, the cause of this balance has remained the subject of debate [13, 14, 15]. Recent studies indicate this process is underpinned by neuroinflammation through pro-inflammatory cytokines. Markers of this inflammation serve as potential biomarkers for the development of subtypes of subjective tinnitus [16, 17].

3. The current neuroinflammatory model of tinnitus

The synaptic connection between two neurons is strengthened by the firing of the presynaptic neuron shortly before the postsynaptic neuron, and this is known as Hebbian plasticity. Conversely, anti-Hebbian plasticity is when the neurons fire out of sync or at the same time. The former leads to long-term potentiation (hyperactivity) and the latter leads to long-term depression; it is this temporal relationship in neuronal firing, which is involved in spike-timing-dependent plasticity (STDP) [18]. Changes in STDP and an increased spontaneous firing rate (SFR) are referred to as hyperactivity in the auditory pathway.

The neuroinflammatory model of tinnitus is still in its infancy with emerging literature limited to animal studies. The following sections will explore our current understanding of the neuroinflammatory model within the context of the classical auditory pathway and what we can infer about the underlying mechanisms based on these studies. Cytokines, microglia, and activated platelets all of these had documented associations with tinnitus, which could pave the way for reliable inflammatory biomarkers of tinnitus.

Short-term noise exposure can lead to a temporary hearing loss, whereas repeated or long-term noise exposure can lead to a permanent hearing loss, both of which are often accompanied by tinnitus. Similarly, the amount of ototoxic drug exposure can lead to temporary or permanent hearing loss and tinnitus [19]. Noise-induced hearing loss is caused by damage to inner and outer hair cells in the cochlea by acute noise exposure, the hearing loss is typically high-frequency and often presents with a high-pitch tinnitus sound [20]. Emotional and physical trauma can also contribute to tinnitus. There is a higher prevalence of tinnitus in individuals with post-traumatic stress disorder than in those working in noisy environments [20]. Many tinnitus treatments focus on stress management with the aim to reduce a tinnitus sufferer’s levels of anxiety or distress, this helps alleviate tinnitus itself, or its negative effects on one’s mental state [21].

In human populations, several non-auditory conditions, such as depression, stress, and traumatic brain injury, are risk factors for tinnitus [22]. These same conditions promote pro-inflammatory cytokine production in cerebrospinal fluid and contribute to neuroinflammation [17]. Genetic studies investigating single nucleotide polymorphisms have found associations between alleles contributing toward cytokine expression and susceptibility to develop noise-induced tinnitus. The frequency of the genotype for IL1α −889 C > T was found to be significantly associated with tinnitus in the elderly with a history of occupational noise exposure [23]. IL6 −174 G > C allele and TNFα −308 G > A allele frequency have been shown to be significantly associated with tinnitus in the elderly with a history of occupational noise exposure [24, 25].

Microglia may also play a role in neuroplasticity, beyond being simple scavengers that monitor and phagocytose waste products after neurodegeneration [26]. Microglia also produce pro-inflammatory cytokines, reactive oxygen species, and chemokines [27] as well as playing a role in neural maturation, aging and neuroplasticity, their pro-inflammatory cytokines regulate the function of neurons in synaptic plasticity [17]. A rodent study by Wang et al. [16] demonstrated neuroinflammatory events in the auditory cortex, following noise exposure. The study found microglial activation led to tumor necrosis factor-alpha (TNF-α) production and TNF-α further activated microglia, this feedback loop leads to an excitatory-inhibitory imbalance, which could be a cause of noise-induced tinnitus in the primary auditory cortex [13]. Following acoustic trauma, microglia upregulate TNF-α and IL-1β (interleukin-1 beta) cytokines in the cochlear nucleus on a much larger scale than in the auditory cortex [28]. This process of glial activation in neuroinflammation may inadvertently contribute to the pathogenesis of tinnitus, while trying to stabilize neuronal activity in the auditory pathway.

3.1 Tinnitus induction methods

Recent studies investigating tinnitus in animal models have typically used salicylate to induce tinnitus or noise exposure. The studies then measure their performance on the gap-prepulse inhibition of the acoustic startle reflex (GPIAS) to determine whether or not tinnitus is present [29]. The GPIAS paradigm relies on animals with tinnitus failing to elicit an acoustic startle response – a defensive reflex in response to loud noise. A silent gap in a continuous noise can inhibit this reflex in animals but the inability to perceive this gap due to tinnitus leads to the reflex being elicited [30].

4. In the context of the classical auditory pathway

4.1 The cochlea and cochlear nucleus

Roberts [19] proposes maladaptive plasticity within the auditory system as a reason for the development of tinnitus following noise-induced hearing loss. This phenomenon is supposedly due to homeostatic neuroplastic changes in neuronal firing leading to increased spontaneous activity in the auditory pathway to compensate for reduced input from the cochlea due to inner hair cell damage.

Martel et al. [31] suggest sodium salicylate increases SFRs and alters STDP through activating N-methyl-D-aspartate (NMDA) receptors, leading to tinnitus, as shown in Figure 1 [32]. Ralli et al. [33] found that memantine, which acts as a noncompetitive inhibitor of the NMDA receptor in the cochlea, was effective at reducing salicylate-induced tinnitus in rats but concluded that the side effects of consuming the dose required would outweigh the benefits. Salicylate-induced tinnitus may also be generated by an increase in the SFR in fusiform cells in the dorsal cochlear nucleus (DCN) as well as a change in STDP due to reduced auditory input from the cochlea [22].

Figure 1.

The current neuroinflammatory model of tinnitus in the context of the classical auditory pathway. The figure shows theoretical causes and effects of hyperactivity in the cochlea, cochlear nucleus, inferior colliculus, medial geniculate body, and auditory cortex, based on findings in the current literature.

It is this change from regular spiking to bursting activity in DCN fusiform cells that may underpin both noise-induced and salicylate-induced tinnitus in animal models [31, 34]. However, this is only true in spontaneous auditory nerve activity when high doses of sodium salicylate are used; moderate doses are capable of inducing tinnitus with no significant change in spontaneous neuronal firing [35]. Greater doses of salicylate are known to have more severe and irreversible effects on the auditory system [36]. Wu et al. [37] propose that noise-induced tinnitus also leads to increased parallel fiber excitation in DCN fusiform cells leading to an increase in SFRs. A reduction in inhibitory synapses on these DCN fusiform cells leads to burst firing and thus hyper-excitation. Variations in NMDA receptor expression across parallel fiber synapses on DCN and cartwheel cells may account for the differences in their Hebbian and anti-Hebbian plasticity, respectively [34].

Cartwheel cells are interneurons that release gamma-aminobutyric acid (GABA) and glycine – two inhibitory neurotransmitters, following noise exposure they are thought to reduce their activity. This downregulation at GABAergic and glycinergic synapses causes hyper-excitation in the fusiform cells of the DCN [34, 38].

Brozoski and Bauer [39] found that weeks after cochlear nucleus ablation, noise-induced tinnitus remained the same in rats; however, the DCN was considered ablated in the study even in cases where as much as 40% of it remained unaffected, as highlighted by Manzoor et al. [40]. Hyperactivity in the remaining DCN could still have been the cause of the noise-induced tinnitus in this case.

Vogler et al. [41] found that on average SFRs in the ventral cochlear nucleus (VCN) doubled in animals with noise-induced tinnitus, most significantly in primary-like and onset cells. Though many studies have explored DCN ablation when monitoring tinnitus in animal models [39, 40, 42], very few have examined hyperactivity in the primary-like cells and onset cells in the VCN and the inferior colliculus, which may sustain tinnitus post-ablation [41].

4.2 Inferior colliculus

Following acoustic trauma, hyperactivity in the inferior colliculus originates from activity in the afferent neurons in the cochlear contralateral to it, however, this dependency on the cochlea lessens the longer the hyperactivity remains in the cochlea [43]. This was demonstrated by [44] who performed cochlear ablation 2 weeks after acoustic trauma and found reduced hyperactivity, whereas in [45] where the same procedure was performed for 12 weeks after the acoustic trauma, found no significant difference in hyperactivity. Hyperactivity in the inferior colliculus may originate from the DCN or be an independent process altogether. Manzoor et al. [40] found that DCN ablation significantly reduced inferior colliculus hyperactivity in hamsters; however, ablation was performed 3 weeks after the development of noise-induced tinnitus. Though hyperactivity in the inferior colliculus likely originates from hyperactivity in the DCN [34], it is possible that hyperactivity in the inferior colliculus could to become endogenous given enough time after the onset of tinnitus [19].

Salicylate-induced tinnitus is known to lead to hyperactivity in the inferior colliculus and the secondary auditory cortex [36]. Tinnitus severity is correlated with an increase in TNF-α [46]. Measuring mRNA expression in mice with salicylate-induced tinnitus, Hwang et al. [47] found that TNF-α and IL-1β were significantly increased in the cochlea and the inferior colliculus as well as NMDA subtype 2B (NR2B) gene expression suggesting NMDA receptor action, which may be involved in long term potentiation leading to hyperactivity [35].

4.3 Medial geniculate body

The medial geniculate body is a known intermediary between the auditory cortex and the inferior colliculus in the classical auditory pathway; however, its role in tinnitus has not been well-examined [48]. Roberts [19] suggests following hyperactivity in the inferior colliculus, thalamic neurons switch from tonal firing to burst firing in the medial geniculate body. This change is caused by membrane hyperpolarization, which activates calcium channels causing thalamic neurons to carry less well-defined nonlinear inputs [48]. This may lead to low-frequency oscillations propagating to the auditory cortex [19].

Iba-1 (ionized calcium-binding adapter protein-1) expression can be used as a marker for increased microglial activation. Xia et al. [15] found Iba-1 was upregulated in the medial geniculate body and the primary auditory cortex in rats, evidencing microglial and astrocyte activation in salicylate-induced tinnitus [49]. The primary auditory cortex also showed increased IL-1β expression in its mRNA. IL-1β alongside other pro-inflammatory cytokines can regulate the excitatory-inhibitory balance through interactions with receptors on neuroglial cells [15].

4.4 Auditory cortex

Deng et al. [50] found greater microglial activation in the auditory cortex of rodents when subjected to 86 dB SPL noise at 8 kHz, following the intracerebroventricular infusion of TNF-α. This also impaired gap detection and prepulse inhibition that suggest the development of tinnitus. The same microglial activation was not present when subjected to the noise alone or when having received the TNF-α infusion alone. Similarly, Basura et al. [51] found stimulus timing-dependent responses followed anti-Hebbian timing rules in the primary auditory cortex of guinea pigs subjected to noise exposure, however, only those who developed tinnitus showed an increase in SFRs. Blast-induced traumatic brain injury results in microglial activation through an alternative pathway from acoustic trauma, one which is possibly independent of TNF-α interaction [14, 16]. Salicylate ototoxicity also leads to increased SFRs in the secondary auditory cortex [52].

Noise-induced hearing loss decreases activity at GABAergic synapses and increases activity at glutamatergic synapses in the auditory cortex, leading to an excitatory-inhibitory imbalance [53]. This may lead to spontaneous synchronous neuronal firing, which would cause long-term potentiation in the auditory cortex inducing tinnitus following acoustic trauma.

5. Peripheral inflammation

Peripheral inflammation and an increase in pro-inflammatory cytokines, such as TNF-α, can lead to neuroinflammatory processes in the brain [54]. In the past 30 years, it has become apparent that platelets are involved in more than just thrombocytosis in vascular injuries, despite this, little remains understood about their cytokine interactions and their role in neuroinflammation [55]. A meta-analysis by Ahmed et al. [56] synthesized data from six studies including 451 tinnitus sufferers and 367 controls, and found mean platelet volume (MPV) was significantly increased in populations with tinnitus, including in normal hearing populations. This suggests greater platelet activation in the tinnitus population and may serve as an easy-to-obtain and inexpensive biomarker of tinnitus [57, 58]. However, further research in this area must prioritize a well-reported methodology including the type of hematology analyzer used and the precise means through which the samples were collected. Additionally, further research must exclude individuals with existing inflammatory pathology as recommended in [56]. By reproducing the association with better documentation and improved methodology, we can better understand the exact mechanisms that underpin MPV as a candidate biomarker of subjective tinnitus and its relationship with neuroinflammatory processes.

6. Current limitations and further research

The meta-analysis by Ahmed et al. [56] used only a human population to demonstrate an increase in MPV in people with tinnitus compared to controls without tinnitus. The current literature examining cytokine interactions in the auditory pathway has primarily used animal studies. This is because current techniques for measuring neural activity in the human brain lack the cellular resolution to pinpoint the exact structures involved in tinnitus [59]. Guinea pigs and rodents are likely to be used since the human cochlear nucleus is cellularly similar to the rodent cochlear nucleus [60]. The GPIAS is commonly used to determine the presence of tinnitus in animals; however, the GPIAS has not been proven to be a valid measure of tinnitus in humans [29]. It is also not possible to distinguish between subtypes of subjective tinnitus based on the presentation in animals [48].

One limitation of this model that has not been discussed in the literature thus far is the relationship between the neuroinflammatory model and hearing loss or stress. Hyperactivity in the DCN, for example, has been shown to be affected by attentional and emotional responses just as tinnitus is [61]. Studies thus far do not seem to define the direction of this relationship. McKenna et al. [62] suggest that those without tinnitus have habituated to spontaneous neuronal firing but tinnitus occurs due to brief lapses in the ability to filter out these sounds. Those who become overly conscious and aware of these phantom sounds may then focus on this triggering a “fight or flight” response impeding their ability to filter out the sounds. It is this feedback loop that may lead to greater stress, which in turn increases neuroinflammation. A systematic review by Calcia et al. [63] found psychosocial stressors lead to microglial activation in the hippocampus, prefrontal cortex, and possibly other regions of the brain [64]. Deng et al. [50] overcame this limitation by comparing three conditions, one with the infusion of TNF-α, second with noise exposure, and the third with both. It was only in the lattermost condition where microglial activation occurred suggesting a causal relationship between noise-induced tinnitus and neuroinflammation.

The role of the VCN in leading to hyperactivity in the contralateral inferior colliculus and the role of the medial geniculate body require further investigation as the current literature on this subject is sparse. Future studies may wish to reexamine the positive feedback loop between microglia activation and TNF-α expression in the auditory cortex, which was observed by Wang et al. [16]. Further studies examining the association between MPV and tinnitus should work toward developing a more homogenous standardized methodology to recreate current findings [65].

Conflict of interest

The authors declare no conflict of interest.