Mechanisms of Ototoxicity & Otoprotection

Significance

The Ototoxicity Working Group of Pharmaceutical Interventions for Hearing Loss defined ototoxicity as damage to the inner ear, targeting cochlear and vestibular structures as well as sensory function, due to exposure to certain pharmaceuticals, chemicals, and/or ionizing radiation. Ototoxicity typically focuses on the inner ear; however, ototoxins can also affect central auditory pathways and are, thus, considered neurotoxic^1,2. Ototoxins can also cause kidney damage and associated renal dysfunction^3–5. This review will focus primarily on ototoxic hospitalbased medications received by several hundred thousand infants and children in Western Europe and the Americas each year – aminoglycoside antibiotics and the anti-cancer drug, cisplatin. Other clinically-relevant drugs are listed in Table 1.

Table 1. Major classes of ototoxic drugs (with specific examples).

Adapted from Steyger PS. Mechanisms of Aminoglycoside- and Cisplatin-induced Ototoxicity. American Journal of Audiology. 2021;in press, with permission from Cold Spring Harbor Press.

Class	Examples
Platinum-based therapeutics	Cisplatin, carboplatin, oxaliplatin
Aminoglycosides	Amikacin, gentamicin, tobramycin
Peptide antibiotics	Capreomycin, viomycin, chloramphenicol
Polypeptide antibiotics	Vancomycin (and synergistic with aminoglycosides)
Macrolides	Erythromycin
Cyclodextrins (vehicle)	Derivatives of cyclodextrins, e.g., for Neiman-Pick Syndrome Type 1C
Anti-malarials	Chloroquine, hydrochloroquine
Anti-rheumatics	Chloroquine, hydrochloroquine
Loop diuretics	Furosemide, bumetanide (and synergistic with aminoglycosides or cisplatin)
Non-steroidal anti-inflammatory drugs (NSAIDs)	Acetaminophen (a.k.a. paracetamol)
Anti-neoplastics	Vincristine

Aminoglycoside antibiotics, like gentamicin, are clinically-relevant, broad-spectrum medications to treat life-threatening bacterial infections, often of unknown etiology, including neonatal sepsis, mycobacterial infections, and meningitis. Exacerbated respiratory infections in children with cystic fibrosis are typically treated with tobramycin or amikacin, and individuals with tuberculosis with kanamycin, and these are also aminoglycoside antibiotics. Aminoglycoside-induced hearing loss is typically cumulatively dose-dependent^6,7, and the incidence in humans is as high as 20–63% of those receiving multi-day dosing^7–11. The prevalence of aminoglycoside-induced vestibular deficits is under-reported due to the widespread lack of clinical instrumentation for assessing vestibular function¹².

Liver tumors and glioblastomas in the brains of children are effectively diminished by cisplatin or other platinum-based derivatives like carboplatin and oxaliplatin, and the degree of drug-induced hearing loss is typically dose-dependent^13–15. There are fewer reports of cisplatin-induced vestibular deficits in children due to the lack of routine clinical assessments^12,16, although there are pre-clinical reports of dose-dependent cisplatin-induced vestibular deficits¹⁷. Children with Niemann Pick Type 1C syndrome can be treated with derivatives of cyclodextrins, a primary component of drug formulations, that at higher doses also sequester membrane cholesterols¹⁸. Treatment with these pharmacological agents in clinical settings can cause ototoxicity. Environmental ototoxins include solvents^19–21, and metals (e.g., lead, copper)^22–25.

Uncorrected congenital or acquired hearing loss in infants and children can lead to delayed acquisition of listening and spoken language skills compared to age-matched peers with typical hearing, with concomitant delays in achieving academic, linguistic, and psychosocial milestones^26,27. Loss of vestibular function is also debilitating^28,29. There is an estimated socioeconomic cost >$1.5 million over the lifetime of each pre-lingually deafened child (in 2019 dollars)³⁰.

Crossing the blood-labyrinth barrier

The inner ear is protected by a blood-labyrinth barrier (BLB; akin to the blood-brain barrier) that compartmentalizes inner ear cells and fluids from the bloodstream and the rest of the body. Endothelial cells lining cochlear blood vessels are coupled together by tight junctions to prevent extravasation (i.e., the paracellular trafficking of blood cells, macromolecules, and serum from the capillary into cochlear tissues)³¹. Within the cochlea, there are two major fluid spaces, perilymph and endolymph. Blood-borne ototoxins are readily trafficked into the perilymph surrounding the basolateral membranes of hair cells^32,33, yet this does not appear to be the primary entry route into hair cells³⁴ (see Figure 1). Instead, aminoglycosides and cisplatin cross the BLB of the stria vascularis and, from there, appear to be rapidly cleared into endolymph and enter hair cells across their apical membranes more readily than when directly infused into perilymph^34–37. The cellular and molecular mechanisms by which aminoglycosides and cisplatin cross the BLB and transverse the tight-junction-coupled marginal cells of the stria vascularis into endolymph remain to be determined, and likely include one or more of the following molecular mechanisms: ion channels, transporters, exchangers, or transcytosis^12,38, each permutation specific for each individual class of ototoxins.

Figure 1: Primary trafficking routes of aminoglycosides and cisplatin from the vasculature to cochlear hair cells.

Circulating ototoxins typically enter the cochlea via capillaries in the stria vascularis and are cleared into endolymph via as-yet-unidentified ion channels or transporters, although several candidates exist, e.g., TRPV1 and TRPV4 non-selective cation channels for aminoglycosides, and potentially via OCT2 and CTR1 transporters for cisplatin. Once in endolymph, ototoxins can enter hair cells via one or more of several mechanisms.

From Kros CJ, Steyger PS. Aminoglycoside- and Cisplatin-Induced Ototoxicity: Mechanisms and Otoprotective Strategies. Cold Spring Harb Perspect Med. 2019;9(11), with permission from Cold Spring Harbor Press.

Entry of ototoxins into sensory hair cells

Typically, cochleotoxicity requires ototoxins to enter hair cells³⁹. At the distal tips of most stereocilia in the mechanically-sensitive hair bundle are transduction channels (Figure 2A) that allow the polycationic aminoglycosides to permeate into the cell⁴⁰. These transduction channels are non-selective cation channels such as TMC1⁴¹. Aminoglycoside permeation of TMC1 channels can be blocked by polyvalent cations like magnesium or calcium (Figure 2B), as well as by organic compounds like curare and quinine^42,43. Hair cells also express other aminoglycoside-permeant channels, including several transient receptor potential (TRP) channels, each activated by different physical or chemical stimuli⁴⁴. These include TRPV1 and TRPV4 at the apical membrane of hair cells (Figure 2C)^45–48. The aminoglycoside-permeant TRPA1 channel is likely expressed on the basolateral membranes of outer hair cells (Figure 2D)⁴⁹. Aminoglycosides can also enter hair cells via non-receptor-mediated endocytosis and are trafficked to lysosomes (Figure 2E). Blocking endocytosis (Figure 2F) or impeding intracellular trafficking of aminoglycoside-laden endosomes does not prevent hair cell death^42,50,51. Cisplatin can enter cells via passive diffusion across the cell membrane (Figure 3A)⁵². Hair cells with active (open) transduction channels (Figure 3B) are more susceptible to cisplatin-induced cytotoxicity⁵³.

Figure 2: Aminoglycoside entry into hair cells.

Aminoglycosides preferentially enter mammalian hair cells via the TMC1 channel that consists of two TMC subunits (purple), each with a permeation groove (A). Entry of aminoglycosides can be blocked by curare, quinine and high levels of polyvalent cations (B). Other aminoglycoside-permeant channels on the apical membrane of hair cells include TRPV1 and TPRV4 (C), and TRPA1 on the basolateral membrane of outer hair cells (D). (E) Non-specific endocytosis enables aminoglycoside-laden endosomes to traffic to hair cell lysosomes. (F) Blocking endocytosis does not prevent hair cell death when aminoglycosides can enter hair cells via the TMC1 channel.

From Steyger PS. Mechanisms of Aminoglycoside- and Cisplatin-induced Ototoxicity. American Journal of Audiology. 2021;in press, with permission from the American Journal of Audiology.

Figure 3: Cisplatin entry into hair cells.

Cisplatin has multiple potential entry routes. (A) Neutral cisplatin can diffuse across the plasma membrane, and is readily aquated in the cytoplasm to the more toxic form of cisplatin that can form functionally disruptive adducts with proteins and DNA. (B) Uptake of aquated cisplatin is dependent on functional TMC channel complexes. (C) Cellular uptake of the aquated form of cisplatin can also occur via CTR1 and OCT2 transport proteins when expressed by the cell.

From Steyger PS. Mechanisms of Aminoglycoside- and Cisplatin-induced Ototoxicity. American Journal of Audiology. 2021;in press, with permission from the American Journal of Audiology.

Once in cochlear hair cells, cochleotoxic dosing with aminoglycosides and cisplatin first leads to outer hair cell death in the basal high frequency region of the cochlea, particularly in outer hair cells. Increasing cumulative dosing leads to further outer hair cell death in more apical regions of the cochlea (i.e., at lower frequencies) and inner hair cell death, with an increasing risk of permanent hearing loss^7,15,54. There is one major difference between aminoglycosides and cisplatin, and that is the preferential cochleotoxicity of cisplatin, compared to the equal reporting of cochleotoxicity and vestibulotoxicity induced by aminoglycosides, although the underlying etiology for this difference remains hypothetical⁵⁵.

There are multiple (yet incompletely-defined) molecular mechanisms of hair cell death, and the best characterizations for aminoglycosides and cisplatin are reviewed elsewhere^4,5,12. It is important to stress that ototoxicity is not dependent on hair cell dysfunction or cell death, and can dysregulate non-sensory cells within the inner ear essential for sensitive auditory perception (e.g., the stria vascularis)¹². Below, we focus on clinical settings that can exacerbate the degree of hearing loss and vestibular disorders induced by ototoxins, particularly by aminoglycosides and cisplatin.

Clinical settings that can exacerbate the risk of drug-induced hearing loss

Experimentally-induced inflammation (to mimic host-mounted responses to bacterial infection) exacerbates the degree of aminoglycoside- and cisplatin-induced hearing loss in mice^38,48,56, and potentially for vestibulotoxicity⁵⁷. Pilot studies of neonates treated with aminoglycosides likely confirm these preclinical findings^58,59. Aminoglycoside-induced lysis of bacteria elevates the inflammatory response (i.e., the Jarisch-Herxheimer reaction)^60,61. The mechanisms underlying the potentiation of hearing loss by inflammation remain to be established^62,63. Thus, the very patients with bacterial infections (and therefore inflammation) treated with aminoglycosides are likely at higher risk of ototoxin-induced hearing loss³⁸.

Other clinical settings can also exacerbate the risk of drug-induced hearing loss, and include: renal insufficiency decreasing the clearance of ototoxins from blood^64,65, increasing age with concommitant decreased glomerular filtration rate^66–69; depletion of endogenous antioxidants needed to ameliorate ototoxin-induced generation of toxic levels of reactive oxygen species^70–72; fever (hyperthermia or higher-than normal body temperature)^59,73, and transient ischemia/hypoxia⁷⁴. Aminoglycoside- and cisplatin-induced cochleotoxicity is also greater when these ototoxins are administered during active hours (i.e., nighttime) for rodents^75–78.

Several single nucleotide polymorphisms in mitochondrial ribosomal RNA (e.g., A1555G and C1494T) result in a higher binding affinity to aminoglycosides, leading to mistranslation of mRNA and inaccurate protein synthesis and a greater susceptibility to aminoglycoside-induced hearing loss^79–82. At this time, no genomic polymorphisms have been reported to modulate susceptibility to aminoglycoside-induced hearing loss. Several genomic polymorphisms have been reported to increase the risk of cisplatin-induced hearing loss, including gene variants for antioxidant enzymes, DNA adduct repair enzymes; antioxidant enzymes, drug efflux or membrane pumps^83–86. However, the predictive value of these genomic variants in predisposing individuals for drug-induced ototoxicity is currently poor^87,88.

Hospitalized children frequently receive multiple medications simultaneously, and several clinically-relevant non-ototoxic drugs can potentiate drug-induced hearing loss. For example, the neuromuscular blocking agent pancuronium bromide is frequently prescribed for neonates requiring respiratory assistance (via intubation and ventilation); however, pancuronium bromide can potentiate the cochleotoxicity of loop diuretics prescribed for hypervolemic patients^89,90. Simultaneous dosing with two or more ototoxic therapeutics can synergistically exacerbate the degree of ototoxicity to greater than the sum of the two ototoxins alone. These co-therapeutics include aminoglycosides and loop diuretics in hypertensive individuals^91–93, or aminoglycosides with vancomycin, a glycopeptide antibiotic commonly prescribed in the neonatal intensive care unit^94,95. Enhanced ototoxicity can also occur when cisplatin is co-administered with aminoglycosides, loop diuretics, and cranial radiation^96–99. Thus, when dosing with multiple drugs simultaneously, all therapeutics should be examined for toxic interactions, including potentiated or synergistic ototoxic effects, and alternative dosing regimens identified where possible.

Loud sound exposures exacerbate the degree of aminoglycoside-induced ototoxicity^100–102. Noise exposure also increases cisplatin-induced hearing loss^103–106. These data are relevant when children require pharmacotherapy following recent exposure to intense loud sounds.

Otoprotection

Strategies that protect the inner ear from the toxic effects of ototoxins in children are currently considered to be the same as being developed for adults. Multiple strategies have been investigated, although few have completed Phase III clinical trials and none are yet approved by the Federal Drug Administration (FDA) in the United States of America. The most-intensely investigated strategy to prevent ototoxicity has been efforts to inhibit the toxic generation of reactive oxygen/nitrogen species or the production of proapoptotic factors^15,107,108. Potentially more specific, otoprotective strategies now include blocking systemically-administered ototoxins from cross the BLB and entering the inner ear fluids to prevent toxicity¹², and physicochemical structural modification of ototoxins to prevent entry into hair cells and thus ameliorate ototoxicity while retaining their desired efficacy have also been investigated^109,110, although for derivatives of cisplatin, this has led to reduced clinical efficacy (i.e., toxicity to the tumor) despite decreased ototoxicity¹⁵.

Alternative delivery strategies have focused on local delivery to the middle or inner ear to reduce the systemic toxicity associated with oral or parenteral administration of candidate otoprotective compounds ; however, local delivery has challenges too, principally rapid clearance from the middle or inner ear¹¹¹. One strategy to reduce clearance has been to use hydrogels and nanoparticles to retain the drug in the vicinity of the round window membrane. One novel strategy is to inject drug-loaded magnetic nanoparticles onto the round window and use a contralateral magnet to pull the magnetic nanoparticles into the inner ear. This strategy abrogated cisplatin-induced hearing loss in the high frequency (basal) region of the cochlea¹¹².

Unbiased screening

Many of the above strategies to prevent ototoxicity used a rational, deductive approach to selecting and designing otoprotective dosing regimens, and have been clearly successful preclinically to varying degrees. An alternative strategy has emerged in recent years to use unbiased screening of large compound libraries in silico, in vitro and/or using zebrafish larvae to identify promising ototoxic “hits” or candidate otoprotective compounds.

Computational, or in silico, models use molecular structures and physiochemical properties to predict drug toxicity or efficacy in an unbiased manner, based on reported structure-activity relationships. This approach has been used to predict cardio- or hepatotoxicity^113,114. Two studies have reported in silico screening strategies for ototoxicity, while others use pathway or linkage models (https://lincsproject.org/LINCS) to identify potential otoprotectants^62,115. Promising hits or candidate otoprotective compounds identified in silico must be verified using in vitro or in vivo assays. In vitro assays typically use cell lines, including those derived from the cochlea^116–118. Although cell lines do not possess stereocilia with mechanically-gated TMC1 channels, cell lines are extremely useful for deciphering the molecular and signaling pathways underlying ototoxic or otoprotective mechanisms^118–121.

Like most fish, zebrafish possess on the surface of their body a mechanosensitive lateral line consisting of neuromasts within which are sensory hair cells with stereocilia and mechanically-gated transduction channels, analogous to cochlear and vestibular hair cells. Female zebrafish release hundreds of larvae in each clutch, and within 4–5 days after fertilization have functional neuromasts¹²². These numbers, and ease of access (without a BLB) enable medium-throughput screening of compound libraries to identify hits that induce hair cell toxicity⁶², or interact with hair cell transduction channels to protect neuromast hair cells from aminoglycoside or cisplatin cytotoxicity^123–125. Subsequently, these hits or candidate otoprotective compounds need to be validated preclinically using inner ear explants and in vivo ototoxicity preclinical models, primarily in rodents and higher- order mammals, before advancing to human clinical trials^117,118,124.

Challenges to preventing ototoxicity

We do not yet appreciate the numerous ways each ototoxin can induce cytotoxicity. This is best exemplified by the hundreds of observed aminoglycoside- or cisplatin-binding proteins, most of which remain unidentified and therefore uncharacterized, of which a fraction will have cytotoxic interactions and the remainder incidental, benign, or even cytoprotective interactions^120,121,126. Hair cells and strial cells (like renal proximal tubule cells) are often unable to clear ototoxins from their cytosol^35,127. Most other cells within the body are typically able to clear ototoxins from their cytosol, often by as-yet-unknown mechanisms; the reasons for these differences in toxicologic sensitivities remain highly speculative.

If candidate otoprotective compounds are to be delivered orally or systemically, and the primary site of otoprotection is within the inner ear itself, these compounds must first cross the BLB at sufficient quantities to meet therapeutic levels without being systemically toxic. If this is not possible, local delivery via intratympanic administration might be viable if the candidate otoprotective compounds are efficacious in perilymph. Other crucial research-driven questions have been posed elsewhere¹⁰².

Although knowledge of the mechanism by which therapeutic compounds achieve their efficacy is not needed for FDA approval, the mechanisms by which candidate compounds provide their otoprotective effect are increasingly being requested. This can add years to the discovery pipeline before a drug can be advanced to clinical trials. This phenomenon can be abrogated by repurposing already-FDA-approved drugs for other indications to off-label indications if otoprotection can be shown preclinically (e.g. statins)^128,129. Candidate otoprotective compounds must be effective in the medical settings in which they will be used (e.g., in individuals with host-mounted inflammatory responses to infection or cancer treatment (e.g. radiation). This practice has been robustly implemented for pharmaceutical interventions for loud sound exposures, but not for prevention of ototoxicity in sick individuals.

Several non-pharmacological, yet otoprotective or rehabilitative interventions, are also likely in the near future. The identification of risk factors, including aging, renal insufficiency, inflammation, genetic polymorphisms, or selected co-therapeutics, that predispose individuals to a greater risk of ototoxicity can be extracted from the medical record or tested for. For example, clinical guidelines exist for identifying which currently known mitochondrial polymorphisms predispose individuals to ototoxicity⁸². In such cases, non-ototoxic pharmacotherapies can be prescribed. Even if these alternate routes can be more expensive in the short-term, they would still be much cheaper than the socioeconomic costs over the individual’s lifetime. Additional non-pharmacological otoprotective strategies include developing point-of-care (bedside) diagnostic procedures for sepsis through species-specific detection of microbial DNA sequences¹³⁰ or genomic screening for genomic polymorphisms that enhance susceptibility to drug-induced ototoxicity^86,131.

Nonetheless, at present, the prevalence and incidence of most predisposing factors remain uncertain due to the variability of each subject’s medical and ototoxin- dosing history within a population. Given the need to prescribe ototoxic medications to individuals in life-threatening situations, it is important to implement ototoxicity monitoring protocols as recommended for those with cystic fibrosis¹³². Widespread implementation of ototoxicity monitoring protocols will provide new insights into the prevalence and incidence of cochleotoxicity and vestibulotoxicity that will accelerate the relative significance and impact of ototoxicity on quality of life scores and optimal rehabilitation of those affected by drug-induced hearing loss.

Key Points:

1. Numerous hospital-prescribed medications and environmental factors cause ototoxicity.

2. Ototoxicity encompasses hearing loss (cochleotoxicity) and/or balance deficits (vestibulotoxicity).

3. Ototoxicity has permanent, life-long debilitating consequences, and if uncorrected can lead to unfulfilled scholastic and career trajectories in children, as well as accelerated cognitive decline in aging individuals.

4. Otoprotection to preserve or restore auditory and/or vestibular function includes ototoxicity monitoring, prosthetic and social rehabilitation, and soon, pharmaceutical ototherapeutics to preserve, repair, or restore inner ear functions.

5. Ototoxicity is, ideally, a preventable form of acquired hearing loss or vestibular deficit.

Synopsis.

Ototoxicity refers to damage to the inner ear that leads to functional hearing loss or vestibular disorders by selected pharmacotherapeutics as well as a variety of environmental exposures (e.g., lead, cadmium, solvents). This chapter reviews fundamental mechanisms underlying ototoxicity by clinically-relevant, hospital-prescribed medications (i.e., aminoglycoside antibiotics or cisplatin, as illustrative examples). Also reviewed are current strategies to prevent prescribed medication-induced ototoxicity, with several clinical or candidate interventional strategies being discussed.

Clinics Care Points.

1. The degree of expected ototoxicity from multi- day dosing with known ototoxins can be amplified by risk factors such as renal insufficiency or systemic inflammation.

2. Dosing with multiple pharmacotherapeutics simultaneously (even if separated temporally) can exacerbate the degree of ototoxicity of a known ototoxin.

3. Ototoxicity monitoring is vital when dosing with life-saving ototoxins, especially as any identified hearing loss can then be more rapidly rehabilitated, ameliorating the negative impact of drug-induced hearing loss.