Drug-induced hearing loss is typically characterized by dose dependency, where increasing cumulative exposure to ototoxins increases the degree of subsequent hearing loss. Furthermore, the onset of drug-induced hearing loss typically begins at higher frequency regions of the cochlea and, with time and/or increasing exposure to ototoxins, spreads to the lower frequencies in the more apical regions of the cochlea.
Many factors are thought to converge to enhance the sensitivity of cochlear sensory cells, particularly the outer hair cells (OHCs), to ototoxic drugs in the basal higher frequency region of the cochlea. These include (but are not limited to) some or all of the following:
1. basal OHCs have more stereocilia, and consequently more mechanoelectrical transduction (MET) channels that are permeant to aminoglycosides and facilitate cisplatin uptake;
2. basal OHCs have MET channels that have greater conductance than apical MET channels, allowing greater uptake of aminoglycosides; and
3. apical OHCs have a larger volume (and size) than basal OHCs, with increased anti-oxidative capacity before saturation.
Other questions regarding aminoglycoside-induced cochlear dysfunction revolve around the molecular mechanisms involved in trafficking systemic ototoxins from the bloodstream across the blood-labyrinth barrier, and whether spiral ganglion (auditory) neuronal death is a primary consequence of ototoxicity or a secondary one following sensory cell death.
Another intriguing and long-standing observation from many preclinical ototoxicity studies is that the frequency range of elevated auditory thresholds is greater than that observed for loss of cochlear sensory cells, particularly OHCs. In other words, there are still elevated auditory thresholds in cochlear regions where almost all sensory cells survive treatment with ototoxic drugs, suggesting that ototoxins permanently damage hair cell transduction of auditory stimuli into neuroelectrical action potentials prior to hair cell death (Sci Transl Med. 2015;7:298ra118 http://bit.ly/2maaWOO; Hear Res. 1992;61[1-2]:117 http://bit.ly/2ma1PO8; Hear Res. 2001;158[1-2]:165 http://bit.ly/2maejFz). A recent study by Jochen Schacht and colleagues at the University of Michigan and Universität Zürich may hold one mechanistic explanation of this lack of apparent correlation between elevated auditory thresholds in regions of almost total hair cell survival (Cell Death Dis. 2015;6:e1763 http://bit.ly/2ma3WBo).
Schacht examined how aminoglycosides trigger protein mistranslation and misfolding in auditory neurons, as well as the link to neuropathy, often also known as cochlear synaptopathy (Cell Death Dis. 2015 http://bit.ly/2ma3WBo). Protein mistranslation leads to improper folding of secondary and tertiary protein structures, which can result in loss of protein activity or dysfunctional integration of intracellular signaling cascades and/or structures. This disruption is one of the major causes of broad-spectrum bactericidal efficacy of these antibiotics. In eukaryotic cells, aminoglycoside-induced protein mistranslation and misfolding also induce stress in the endoplasmic reticulum (ER) that can ultimately trigger cell death via one or more pathways. In this study, the authors also utilized the power of generalizable in vitro data extrapolated into organ culture and in vivo studies to test and validate a specific line of hypotheses.
Using cytoplasmic assays (lysed cells) or intact cultured cells, the authors demonstrated that aminoglycosides misread or mistranslated a variety of messenger RNA (mRNA) transcripts, including mutated mRNA with premature stop codons that rescued protein functionality in a dose-dependent manner. Misread proteins induced upregulation of a wide range of cytoplasmic- and ER-associated chaperone, and/or protein degradation, indicating an increased capacity to respond to the stress induced by misread or unfolded proteins. Analysis of protein expression upregulated by gentamicin revealed a specific protein, X-box binding protein-1 (XBP1), which is essential to the unfolded protein response induced by increased levels of misread proteins and to cellular survival. Unsurprisingly, XBP1-null embryos do not survive, and mice, heterozygous for the XBP1 gene, have reduced levels of this critical protein (haploinsufficiency). Other aminoglycosides that do not cause misreading of mRNA during ribosomal translation (e.g., hygromycin), or inhibition of protein translation (by cyclohexamide) did not upregulate XBP1 expression. This strongly indicated that XBP1 expression is dependent on the presence of misread proteins, and not on inhibition of protein synthesis.
Remarkably, cochlear coil explants (which include the organ of Corti and spiral ganglion neurons, SGNs, i.e., the auditory nerve) incubated with gentamicin displayed markers for ER stress in SGNs, but not hair cells (even after exposure for long periods). A positive control (tunicamycin) rapidly induced expression of ER stress markers in both hair cells and SGNs. This outcome was unexpected given that aminoglycosides appear to primarily diminish OHC function and survival, relative to SGNs. The authors then extrapolated from their in vitro and explant data to develop an in vivo hypothesis that gentamicin preferentially affects SGNs in mice with insufficient expression of XBP1.
MORPHOLOGICAL & PHYSIOLOGICAL OBSERVATIONS
Gentamicin was delivered via intra-tympanic injection to avoid the complications of systemic gentamicin administration, as well as to facilitate a more timely outcome (Hear Res. 2001 http://bit.ly/2maejFz). In both wild-type and XBP1-haploinsufficient mice, OHCs survived gentamicin treatment (except in the extreme base of the cochlea, closest to the site of gentamicin administration) without significant differences between the mouse strains.
However, SGN densities were significantly reduced in XBP1-haploinsufficient mice, but not in wildtype littermates. These data were corroborated by a significant (50%) reduction in the number of synapses between inner hair cells (IHCs) and afferent dendrites of SGNs that were not replicated in wild-type littermates. The authors then confirmed that these cochlear morphological observations had physiological correspondence.
The overall survival of OHCs in both wild-type and XBP1-haploinsufficient mice was physiologically corroborated by non-significant changes in distortion product otoacoustic emissions, characteristic of OHC integrity and functionality. However, auditory brainstem responses displayed significant high frequency (32 kHz) threshold shifts, characteristic of degraded IHC synaptic integrity with afferent dendrites and diminished auditory nerve functionality. These threshold shifts in XBP1-haploinsufficient mice were abrogated by the administration of a molecular chaperone on misfolded proteins that rescues cells from ER stress.
These studies confirm prior indirect evidence of the involvement of the endoplasmic reticulum in aminoglycoside-induced cytotoxicity in cochlear cells, and, crucially, the preferential effect in SGNs relative to OHCs, as illustrated by the XPB1-haploinsufficient mice. The intravenous delivery of gentamicin over several days also produces a similar differential effect on SGNs and synaptic density relative to IHC and OHC survival, which remains to be determined. Furthermore, it's unknown if the loss of SGNs alone is sufficient to account for the apparent larger loss (50%) of synaptic densities in IHCs. There is published evidence that gentamicin can degrade IHC synaptic densities, which may partially recover over time (Mol Neurobiol. 2013;48:647 http://bit.ly/2ma5GuC; Mol Neurobiol. 2015;52:1680 http://bit.ly/2maglFx).
This paper demonstrates three remarkable advances to understanding ototoxicity:
1. individual aminoglycoside compounds can induce ER stress in specific cell types, but not others, suggesting multiple modes of cellular toxicities in different cell types;
2. SGNs are potentially sensitive to ototoxicity even when IHCs survive, providing corresponding and corroborating evidence to that reported by Corfas et al., showing SGN dendrite survival in the absence of IHCs (J Neurosci. 2012;32:405 http://bit.ly/2ma4k2O); and
3. drug-induced cochlear synaptopathy and neuropathy may underlie the substantially larger (high) frequency range of elevated auditory thresholds relative to the narrower frequency range of cochlear OHC loss in the basal region of the cochlea.
Journal Club Highlight
XBP1 mitigates aminoglycoside-induced endoplasmic reticulum stress and neuronal cell death
Oishi N, Duscha S, Boukari H, et al. Cell Death Dis. 2015;6:e1763