Mechanisms Involved in Ototoxicity

Abstract

The modern era of evidence-based ototoxicity emerged in the 1940s following the discovery of aminoglycosides and their ototoxic side effects. New classes of ototoxins have been identified in subsequent decades, notably loop diuretics, antineoplastic drugs, and metal chelators. Ototoxic drugs are frequently nephrotoxic, as both organs regulate fluid and ion composition. The mechanisms of ototoxicity are as diverse as the pharmacological properties of each ototoxin, though the generation of toxic levels of reactive oxygen species appears to be a common denominator. As mechanisms of cytotoxicity for each ototoxin continue to be elucidated, a new frontier in ototoxicity is emerging: How do ototoxins cross the blood-labyrinth barrier that tightly regulates the composition of the inner ear fluids? Increased knowledge of the mechanisms by which systemic ototoxins are trafficked across the blood-labyrinth barrier into the inner ear is critical to developing new pharmacotherapeutic agents that target the blood-labyrinth barrier to prevent trafficking of ototoxic drugs and their cytotoxic sequelae.

Ototoxicity was first recognized when mercury vapors were used for treating head lice in the 11th century.^1,2 Cinnabar, a naturally occurring mercuric sulfide that is part of Chinese herbal medicine, is often used today as a sedative for infants in Asian countries.³ During the 1840s, antimalarial drugs (e.g., quinine^4–7), and in the 1850s, nonsteroidal anti-inflammatory drugs (e.g., salicylate^6,8,9) were first reported to induce auditory dysfunction, albeit transiently. The modern era of ototoxicity emerged in the 1940s following the discovery of streptomycin and its efficacy for treating tuberculosis.^10,11

Following the 1940s, subsequent decades have been associated with new clinical reports of chemically-induced auditory dysfunction and deafness, particularly loop diuretics in the 1960s, antineoplastic drugs (cisplatin and derivatives) and solvents during the 1970s, and metal-chelating agents in the 1980s.¹ Auditory, vestibular, and toxicology research laboratories around the world are investigating the mechanisms by which these classes of ototoxins induce their cytotoxic effects, as each of these compounds have little or no pharmaceutical or chemical overlap. These are briefly reviewed below.

The sensory structures of the auditory and vestibular systems lie behind a blood-labyrinth barrier that is similar to the blood-brain barrier. Theoretically, only ions, amino acids, sugars, and other compounds essential for cellular function within the inner ear should be transported through the cells that form the blood-labyrinth barrier. Any breakdown in the cellular integrity and/or increase in paracellular permeability of the blood-labyrinth barrier immediately induces loss of the endolymphatic potential (EP) with consequent elevation of sensory thresholds. Ototoxins, however, are able to traverse the intact blood-labyrinth barrier, yet the trafficking mechanisms by which they do so are largely unknown. This review will also discuss potential trafficking mechanisms by which ototoxins cross the blood-labyrinth barrier to enter the cochlear tissues and fluids to induce their cytotoxic effect.

AMINOGLYCOSIDE ANTIBIOTICS

In 1944, streptomycin was the first aminoglycoside antibiotic to be isolated and was found to be bactericidal against both gram-negative and gram-positive bacteria, including tuberculosis pathogens.¹⁰ The following year saw the first of many reports of aminoglycoside ototoxicity in tuberculosis patients,¹² initiating a search for other bactericidal aminoglycosides that were less ototoxic.¹ Today, the most frequently used aminoglycosides include amikacin, gentamicin, and tobramycin. Despite their ototoxic potential, aminoglycosides remain an essential part of the clinicians’ prescriptive arsenal against bacterial infections, particularly in premature babies as prophylaxis against bacterial infection in the United States.¹³ However, in Asian Pacific Rim countries, aminoglycosides are cheap, efficacious, and available as over-the-counter medications, making aminoglycosides one of the world’s most popular medications.

Aside from their bactericidal effects, aminoglycosides also induce acute nephrotoxicity and chronic hearing loss and/or vestibular deficits, depending on their dosage and route of administration. Typically, aminoglycosides are administered parenterally in doses based on body weight. The toxicity of aminoglycosides is thought to depend on total dose. A variety of dosing regimens have been described for aminoglycosides: once daily, twice daily, or short-term infusion; however, ~4 to 14% of subjects will experience ototoxicity.¹⁴ When more sensitive diagnostic tools for auditory sensitivity are used, such as high-frequency audiometry, the rate may increase to as high as 70%.^15,16

Aminoglycoside-induced ototoxicity is synergistically enhanced by prior or simultaneous noise/sound exposure capable of inducing temporary or permanent threshold shifts.^17–19 Poor renal function, fever (temperature and/or bacterial-induced inflammatory responses), poor nutritional or low antioxidant status, and administration of loop diuretics all enhance the risk of aminoglycoside-induced hearing and vestibular deficits.^20–24

Aminoglycosides induce both acute physiological and permanent functional effects. The acute physiological effects include the blockade of ion channels in a variety of tissues, most pervasively cation wasting of calcium, sodium, potassium, magnesium, and glucose in urine excreted from the kidneys.^25,26 Other acute physiological effects include extracellular blockade of calcium influx and intracellular calcium signaling by activating the calcium-sensing receptor.^27–34 For inner ear sensory hair cells, aminoglycosides block the depolarizing transduction current of the mechanoelectrical transduction (MET) channel on hair cell stereocilia.³⁵ They also slowly permeate through the hair cell MET channel and other nonselective cation channels, preventing functional ion channel kinetics.^36–38 Aminoglycosides are also taken up by endocytosis into kidney proximal tubule cells and inner ear sensory hair cells.

Following entry into inner ear sensory hair cells, aminoglycosides induce a wide variety of acute effects depending on uptake mechanism, cell type, and intracellular conditions. These include the elevation of intracellular calcium levels³⁹ and the generation of toxic levels of reactive oxygen species (ROS).^21,40,41 These physiological changes can initiate several different mechanisms of hair cell death (often simultaneously), including apoptosis (programmed cell death initiated by the dying cell) and necrosis (exogenous factors, e.g., drugs, that induce cell death independently of cell regulation). Which mechanism of hair cell death is initiated depends on the type of aminoglycoside exposure, for example, acute or chronic exposure with low or high dosing regimens. The experimental design itself is also critical because apoptotic death pathways are frequently observed during in vitro experiments, but in vivo experiments will display necrotic and caspaseindependent death pathways more frequently.^42–44

In the kidney, aminoglycoside-induced toxicity and cell death in the proximal tubule is typically acute as tubular epithelial cells can be regenerated.⁴⁵ In contrast, within the inner ear, systemically-administered aminoglycosides preferentially enter basal cochlear outer hair cells, which are consequentially more susceptible to aminoglycoside-induced cytotoxicity.^46–48 Continued treatment with aminoglycosides extends the region of toxicity to inner hair cells, and to outer hair cells in more apical regions of the cochlea. Aminoglycoside-induced hair cell death mechanisms may continue for up to 4 weeks after cessation of drug administration. Most vertebrates have the ability to spontaneously replace the loss of inner ear hair cells and regain sensory function through regenerative proliferation. However, in mammals including humans, the ability to regenerate inner ear sensory hair cells has been lost,⁴⁹ resulting in a permanent decrease in auditory and vestibular sensitivity.

The Composition of Cochlear Fluids: A Primer for Drug Trafficking

The mammalian cochlea is longitudinally divided into three fluid compartments, the scala tympani and scala vestibuli, filled with perilymph. Sandwiched between these two scalae is the scala media, filled with endolymph (Fig. 1). Endolymph is a unique extracellular fluid composed of high potassium (K⁺, 157 mmol/L), very low calcium (Ca²⁺, 0.023 mmol/L), and very low sodium (Na⁺, 1.3 mmol/L) ions, and is similar to cytoplasmic molar concentrations.⁵⁰ Endolymph is distinctly different from perilymph found in the other two major fluid spaces within the cochlea, the scalae vestibuli and tympani. Perilymph is typically composed of low K⁺ (5 mmol/L), low Ca²⁺ (1.3 mmol/L), and high Na⁺ (145 mmol/L) and is similar to cerebrospinal fluid and other extracellular fluids in the body.⁵⁰ The cochlear fluids do not intermix due to the presence of impermeable tight junctions between adjacent cells lining the scala media and prevent the paracellular flow of ions and water. The specific ionic composition of endolymph is generated by a high density of ion transporters, ion exchangers, and ion ATPases located in the stria vascularis, located on the lateral wall.⁵⁰ The active trafficking of these ions also generates the positive EP that is +80 mV compared with perilymph and is highly dependent on strial trafficking of K⁺ transport into endolymph. Thus, the stria vascularis has a major ion and fluid regulatory role, similar to the proximal and distal tubules of the kidney. Ototoxic drugs that pharmacologically disrupt ion regulation are frequently also nephrotoxic.

Figure 1.

Potential aminoglycoside trafficking routes within the cochlea. (1) Aminoglycosides enter hair cells by apical endocytosis or permeating MET channels on stereocilia in vitro, and presumptively from endolymph in vivo. (2) Systemic aminoglycosides could cross the strial blood-labyrinth barrier (BLB) in the stria vascularis (SV) by trans-strial trafficking from strial capillaries to marginal cells, and subsequent clearance into endolymph, or (3) by traversing the spiral ligament or spiral limbus BLB into perilymph, before being (4) transported across the cellular components of Reissner’s membrane or the basilar membrane into endolymph, or (5) trafficking of aminoglycosides from perilymph via gap junctions in fibrocytes (F) to the stria vascularis, and trans-strial trafficking into endolymph. Alternatively, (6) aminoglycosides within the scala tympani may enter hair cell directly across their basolateral membranes. Thick line represents the impermeable tight junction coupling between adjacent cells. Diagram is not to scale.

Intracochlear Trafficking of Aminoglycosides

Aminoglycosides enter hair cells via apical endocytosis or by permeating through stereociliary MET channels in vitro.^36,51 In vivo, the apical membrane and stereocilia of hair cells are immersed in endolymph that fills the scala media (Fig. 1). If aminoglycosides enter hair cells from endolymph, the positive (+850 mV), EP will electrophoretically drive the cationic aminoglycosides into the negatively polarized hair cells via aminoglycoside-permeant ion channels, like the MET channel. This driving force clearing aminoglycosides from endolymph could account for the low concentration of aminoglycosides in endolymph (compared with perilymph) in pharmacokinetic studies.^52,53

If aminoglycosides enter hair cells from endolymph, how do systemically administered aminoglycosides cross the blood-labyrinth barrier to enter the endolymphatic scala media? Recent studies revealed that one aminoglycoside gentamicin preferentially loads the stria vascularis compared with the spiral ligament and other cochlear regions. Within the stria vascularis, trans-strial trafficking of fluorescently tagged gentamicin from the strial vasculature into marginal cells is inhibited by the untagged drug following systemic injection.^54–56 Trans-strial trafficking of gentamicin (or other drugs) from the strial vasculature into marginal cells will require active transport against the electrical barrier imposed by the generation of the EP at the interface between intermediate cells and marginal cells (i.e., a gentamicin or aminoglycoside transporter; Fig. 2). Once in marginal cells, aminoglycosides would passively diffuse out into endolymph down the electrical gradient via nonselective cation channels that are aminoglycoside-permeant (e.g., transient receptor potential, vanilloid class member 4 [TRPV4]).^37,57

Figure 2.

Hypothesized trans-strial trafficking pathway for systemic aminoglycosides. Aminoglycosides within the capillaries (1) enter endothelial cells (E) and could permeate into intermediate cells (I) via gap junctions (2), before clearance from endothelial cells and/or intermediate cells into the intrastrial space against the electrical gradient (3). (4) Aminoglycosides are then translocated by the presumptive aminoglycoside transporter across the basolateral membrane of marginal cells (M), and then (5) cleared across the luminal/apical membrane of marginal cells into endolymph. B, basal cell; F, fibrocytes.

Systemically-administered aminoglycosides also traffic into perilymph, to ~10% of serum levels. In addition, perilymphatic uptake and clearance of aminoglycosides follows serum pharmacokinetics closely.⁵² Perilymph perfusion with aminoglycoside-laden artificial perilymph did not affect hair cell transduction for 24 hours,⁵⁸ suggesting that if aminoglycosides enter hair cells across the basolateral membranes, they require time to exert their cytotoxic effect on hair cell function. Alternatively, time may be required to transport aminoglycosides into the endolymph (via the stria vascularis?) where they can then block or permeate MET channels.^36,58,59 Higher (nonphysiological) doses of aminoglycosides in perilymph perfusate were found to induce hair cell dysfunction and cause cytotoxicity more rapidly.⁶⁰ This suggests that these drugs can damage hair cells directly from the basolateral perilymphaticdomain.

CISPLATIN

Cisplatin is clinically essential to treat a wide variety of epithelial and metastatic cancers, especially in pediatric cases. More than one million patients receive cisplatin or its derivates in North America and western Europe.⁶¹ Sixty percent of patients receiving multiple doses of cisplatin experience some degree of hearing loss including profound deafness, neurotoxicity, and acute nephrotoxicity.^62–66 Antineoplastic drugs like cisplatin target the genome of proliferating cells to form cisplatin DNA adducts, disrupting genetic regulation within tumor cells, and inducing cytotoxic processes that can cause tumors to stop growing or go into remission.⁶⁷

However, cochlear and proximal tubule epithelial cells do not proliferate rapidly (and mammalian hair cells not at all). In these cells, cisplatin is thought to permeabilize mitochondria to release proapoptotic factors or generate toxic levels of ROS, each of which can initiate cell death mechanisms.^68–71 Cisplatin-induced hair cell toxicity occurs preferentially in the basal region of the cochlea, which is sensitive to high frequencies, subsequently spreading to more apical regions, which are sensitive to lower frequencies.⁷² Cisplatin also induces degeneration of the stria vascularis, decreasing the number of marginal and intermediate cells, as well as spiral ganglion cells,^73,74 along with deterioration in auditory performance.

Trafficking of Cisplatin

The cellular uptake of cisplatin is poorly understood, although recently, several active transport mechanisms in and out of the cell have been proposed. Several copper transporters are thought to transport cisplatin, including copper transporter-like 1 (CTR1),^75,76 ATP7A and ATP7B (copper-transporting adenosine triphosphatases 7A and 7B, respectively), and the organic cation transporter 2 (OCT2). A thorough description of the distribution of these transporters has yet to be described in the cochlea, although recently CTR-1 was described to be present in the stria vascularis among other cochlear locations.⁷⁷

The evidence for the CTR1 copper transporter being involved in cisplatin comes from multiple experimental approaches. Transfection of human and yeast cells with the CTR1 gene to induce protein expression of the CTR1 transporter increased cisplatin uptake by 50%.^75,78,79 Cultured cell lines that have been shown to be more resistant to cisplatin treatment have significantly lower expression of CTR1 and reduced intranuclear DNA adduct formation following cisplatin treatment.^80,81 Treatment with high levels of copper causes the internalization of the CTR1 transporter, resulting in decreased cisplatin uptake and reduced cisplatin-induced toxicity.⁷⁶

Two other copper transporters, ATP7A and ATP7B, also have been linked to cisplatin binding within the cell and export out of the cell.⁸² As with CTR1, increased expression of ATP7A or ATP7B endows increased resistance to cisplatin-induced toxicity in tumor cell lines.^80,83 Following cellular uptake, ATP7B binds to cisplatin and is transported into intracellular vesicles.⁸⁴ Kidney tubule cells also express polyspecific transporters that are able to transport a variety of drugs and toxins (typically organic cations and weak bases) across the cell membrane.⁸⁵ One such transporter, OCT2, has been associated with cisplatin excretion in the kidneys.^86,87 OCT2 is necessary for cisplatin uptake and cisplatin-induced apoptosis in vitro.⁸⁶ CTR1, ATP7A, ATP7B, and OCT2 are all highly expressed at the choroid plexus, a part of the blood-brain barrier that regulates ion and substrate trafficking into and out of cerebrospinal fluids.^88–90 These transporters are likely to be expressed in other blood-neural barriers like the blood-labyrinth barrier.

Several platinum-based derivatives of cisplatin (oxaliplatin, carboplatin) have been developed to reduce their oto- and nephrotoxic side effects. The reduced ototoxicity appears to be due to the replacement of chloride ions, which are integral for DNA adduct formation, thus leading to reduced cisplatin DNA adduct formation in inner ear tissues. These chemical modifications may reduce the uptake and trafficking of these drugs into the nucleus cochlear and kidney cells.^91,92

SOLVENTS

Organic solvents are consumed in alcohol-containing drinks, or inhaled from paints, adhesives, nail polish, and perhaps most significantly from heating (e.g., kerosene) and automotive (jet fuel, gasoline) fuels. However, by themselves, they generally do not impair auditory function. Toxic exposure to organic solvents typically leads to severe physiological dysfunction, including pulmonary edema, neural dysfunction, cell lysis, and renal and hepatic failure before directly affecting hearing.⁹³

More pertinently, exposure to organic solvents in noisy conditions, particularly in military and chemical plant environments, enhances the risk of ototoxicity.⁹⁴ Exposure to toxic levels of solvents alone primarily induces loss of outer hair cells in the outermost rows.^95,96 However, exposure to solvents alone in the presence of noise enhances the loss of the innermost row of outer hair cells, suggesting that solvents potentiate the mechanisms of noise-induced hearing loss, rather than initiate additional cytotoxic mechanisms.

Inhalation or consumption of organic solvents leads to distribution throughout the body. This is countered by metabolism and clearance of the toxic solvents via the liver and kidneys. However, once solvents have entered the body, their lipophilic nature enables them to diffuse across cellular membranes, including the blood-labyrinth and blood-brain barrier with ease.⁹³

TRANSIENT TOXICITY

Several compounds induce transient ototoxicity (i.e., reversible loss of auditory function, including loop diuretics, salicylate [aspirin], and quinines). Treatment with loop diuretics (e.g., furosemide, bumetanide) to reduce edema or renal insufficiency in patients with impaired kidney function results in increased urine output. This is accomplished by loop diuretic inhibition of the sodium-potassium-chloride (NKCC) cotransporter in the distal tubule of the nephron, raising the osmotic content of urine, preventing water resorption and increasing urine volume. Systemic administration of loop diuretics also inhibits NKCC transporters at the basolateral membrane of marginal cells in the stria vascularis, resulting in impaired potassium (K⁺) trafficking and sodium (Na⁺) cycling. This results in a drop in the EP and consequent loss of hearing sensitivity.^97,98 As the drug is cleared from the body, the EP recovers, and hearing sensitivity is restored.⁹⁹ Treatment with loop diuretics and aminoglycosides simultaneously will cause severe ototoxicity, due to the enhanced entry of aminoglycosides into the cochlear tissues and fluids.^23,24,56,100 Once loop diuretic inhibition is removed and the EP restored, aminoglycosides within the endolymphatic scala media would be electrophoretically driven into hair cells and supporting cells at concentrations sufficient to induce hair cell and organ of Corti degeneration.

Quinine is an effective antimalarial drug (and leads to toxicity in the Plasmodium falciparum parasite) and has analgesic and anti-inflammatory properties. However, it also has transient high-frequency ototoxicity.¹⁰¹ Quinine has acute physiological effects on hair cell transduction, blocking the MET channel and drug uptake.¹⁰² The mechanism of ototoxicity appears to be associated with loss of the membrane potential and consequent reduction in the motility of sensory outer hair cells in the cochlea.¹⁰³ Reports of permanent ototoxicity due to quinine are rare.

Salicylate is a nonsteroidal anti-inflammatory drug with a long history in fever suppression (through a preparation of willow leaves) dating back to 1550 BC. However, the first reports of salicylate-induced auditory dysfunction (tinnitus, temporary threshold shifts) did not occur until the active component, salicin, was first purified, and large-scale consumption of salicylate occurred in the 19th century. Today, salicylate, in the form of aspirin, is the most frequently consumed drug, with no confirmed reports of permanent ototoxicity.¹ Significantly, aspirin has been shown to have prophylactic effects against heart attacks, stroke, and blood clot formation (thrombosis).¹⁰⁴ In addition, aspirin has a protective affect against gentamicin- and cisplatin-induced ototoxicity, presumptively through an antioxidant mechanism.^105–109

PREVENTING TRAFFICKING OF OTOTOXIC DRUGS INTO THE COCHLEA

In individual adults, acquired hearing loss is estimated to have economic costs of ~$297,000 (in year 2000 dollars per person). This is due to loss of income, reduced productivity and career progression, coupled with increased expenses for rehabilitation, prosthetic devices, education, and accessibility.¹¹⁰ In infants, uncorrected auditory deficits can result in a delay in speech acquisition and educational development, which hinder psychosocial interactions and future employment.^111–116 Deafness has an economic cost greater than $1 million over the lifetime of each prelingually deafened child.¹¹⁰ Thus, preventing ototoxicity is of great socioeconomic importance.

The intracellular manifestations of drug-induced toxicity can be reduced by using a variety of antioxidants and inhibitors of hair cell death pathways.^{107,108,117–122} Some have proven effective in maintaining acute hair cell survival in vitro. Others have proven effective in attenuating functional hearing loss caused by drug-induced toxicity in animal trials. In humans, the primary goal of protecting hearing during drug therapy is to prevent the loss of sensitivity to tones below 8 kHz, the frequencies that are important for speech discrimination.

An alternative approach to preventing drug-induced ototoxicity is to prevent ototoxin trafficking into the hair cells. The apical surfaces of mammalian hair cells currently associated with aminoglycoside uptake are bathed in endolymph, which has low extracellular Ca²⁺ concentrations that is essential for sensitive auditory function.^36,123 However, low extracellular calcium levels also potentiate aminoglycoside uptake through aminoglycoside-permeant ion channels, such as the MET channel.^36,38 One solution is to modify the molecular structure of aminoglycosides and increase its diameter to greater than the pore size of the MET channel (~1.2 nm¹⁰²) so that the modified aminoglycosides can no longer permeate through the MET channel. The modified aminoglycosides would still need to be bactericidal to be an effective therapeutic tool for treating bacterial sepsis.

Another approach would be to identify, and block, the mechanisms by which drugs are trafficked across the blood-labyrinth barrier prior to uptake by hair cells. It was recently hypothesized that there is a trans-strial pathway for trafficking aminoglycosides across the cellular blood-endolymph barriers between the vasculature and the cochlear fluids (i.e., endothelial and epithelial [marginal] cells in the stria vascularis). Several transient receptor potential channels are located in the strial endothelial, intermediate and marginal cells, including TRPV1 (vanilloid family member 1), and TRPV4,^124–126 and are thought to allow aminoglycosides to permeate through their channel pores. In addition, an active transporter of aminoglycosides is hypothesized to reside at the interface between the marginal and intermediate cells where the ~+100 mV EP isgenerated.^54,98 A candidate transporter for cisplatin, CTR1, also has been localized in the stria vascularis,⁷⁷ although the trafficking of quinine, or the otoprotectant salicylate, into the inner ear remains unknown.

Thus, the stria vascularis may represent a convenient systemic location for blocking drug trafficking into the cochlea. In addition, this strial location has the advantage of allowing systemic administration of cotherapeutics that efficaciously inhibit drug transport across the blood-endolymph barrier to sufficiently low levels that ensure the cellular and functional survival of sensory hair cells. If blockers of drug trafficking into the cochlea are identified, it is likely that pharmacotherapeutic intervention to prevent drug-induced ototoxicity will consist of both drug uptake inhibitors and ameliorators of drug-induced toxicity (e.g., antioxidants).

Learning Outcomes:

As a result of this activity, the participant will be able to (1) describe how ototoxic drugs enter the cochlea and hair cells to induce their cytotoxic effect, and (2) describe the importance of the blood-labyrinth barrier in cochlea function and otoprotection.