Oxidative stress has been shown to play an important role in the pathogenesis of a variety of cochlear disorders including ototoxicity and noise trauma (Clerici and Yang, 1996; Clerici et al., 1996; Kopke et al., 1997; Ohinata et al., 2003). Outer hair cells appear particularly vulnerable to an oxidative stress. This may be due to inherent differences in antioxidant levels within apical versus basal outer hair cells. Noise trauma causing permanent threshold shifts has been shown to result in a four-fold increase in perilymph levels of reactive oxygen species (ROS) (Ohlemiller and Dugan 1999; Ohlemiller et al., 1999). Animal models of cochlear implantation also suggest that oxidative stress may mediate delayed damage after implantation. The measurement of isoprostane levels using an immunohistochemical assay after noise trauma demonstrated accumulation of oxidative damage in the outer hair cells (Ohinata, 2000). Treatment of animals with antioxidants has been shown to protect hearing and prevent cell death (Ohinata, 2003). Models of cochlear ischemia reperfusion injury, which can be seen during acoustic neuroma surgery, suggest that the hydroxyl radical is responsible for damage to outer hair cells (Ohlemiller et al., 1999). Thus, several types of injury may be mediated by free radicals in the cochlea. The mechanism by which oxidative stress damages tissue is incompletely understood. Recent studies have suggested that noise trauma is capable of causing both apoptosis and necrosis (Hu et al., 2000, 2002, 2003; Yamashita et al., 2003; Yang et al., 2003). Several different techniques are available to study the interaction of oxidative stress and cellular function. These are difficult to carry out within the cochlea because of limits imposed by complex histological processing and tissue availability. In vitro systems can provide a means of examining cochlear tissue and allow an examination and of hair cell function over a predetermined period of time (Malgrange et al., 2002). The effects of oxidative stress in hair cells can been observed either by augmenting levels of pro-oxidant in the culture medium or inhibiting the production of endogenous antioxidants (Kopke et al., 1997). Dehne et al. (2000) used an adult guinea pig organ of Corti culture system to examine the effects of high concentrations of hydrogen peroxide on hair cell survival (Dehne et al., 2000). Addition of hydrogen peroxide to a tissue culture medium containing a transition metal results in production of hydroxyl radical through the Fenton reaction. This in turn reacts with proteins, lipids, and nucleic acid to produce a variety of different types of damage ranging from production of toxic proapoptotic fatty acids such as 4 hydroxy nonenol to mutation of mitochondrial and nuclear DNA. This study was undertaken to examine the effects of a physiologically relevant dose of a pro-oxidant on hair cell survival and to determine if there were long-term effects of oxidative stress. Use of an in vitro system has the advantages of allowing determination of dose response curves for a set amount of pro-oxidant and allows one to follow the effects of oxidative stress on living tissue over a time course. We chose to use hydrogen peroxide as a pro-oxidant because of its established use in tissue culture (Grasl-Kraupp et al., 1995; Dehne et al., 2000; Jin et al., 2001). We found that even minimal exposure to pro-oxidants resulted in significant changes within hair cells. Oxidative stress in noncytotoxic doses appears to predispose hair cells to widespread cell death if re-exposed to another low dose of pro-oxidant. This appears to occur due to changes in mitochondrial membrane potential and changes in mitochondrial gene expression.
MATERIALS AND METHODS
All studies were approved by the Animal Care and Use Committee Univ. Maryland. P3 CBA × C57Bl/6 mice raised in a breeding colony were painlessly euthanized by cold CO2 inhalation and decapitated. The cochleae were removed and placed in sterile phosphate buffered saline (PBS). The cartilaginous otic capsule was removed and the organ of Corti dissected away from the stria vascularis. All cultures were maintained in Dulbeco Modified Eagle Medium supplemented with 10 g/mL of N1 (Gibco), 5.5 L/mL of 30% glucose, and 100 U/mL of penicillin. P3 explants were cultured on Millicell-CM 0.4 μm culture plate inserts (Millipore) with five explants/membrane and a total volume of 1500 L of medium. Insert diameter was 30 mm and six-well culture dishes were used. Long-term cultures were maintained for up to 14 days with medium being exchanged every 3 days. Hair cell viability was checked using differential interference contrast (DIC) microscopy on a daily basis.
Hair Cell Counts
Explants were rinsed in PBS and fixed for 1 hr in 4% PBS buffered paraformaldehyde. Explants were again rinsed and permeabilized with 0.3% triton X 100 in PBS for 10 min at room temperature. The tissue was washed three times in PBS and then incubated with PBS + 1/500 phalloidin FITC or phalloidin TRITC (Sigma) for 20 min at room temperature. The explants were washed three times in PBS and mounted with an antifade medium between two cover slips, allowing examination of both sides of the explant. Only fully intact explants were counted. Three separate regions in the apical, mid, and basal turn were examined with a 60× oil immersion lens with a fluorescent microscope. All inner and outer hair cells were counted in each high power field. Hair cell survival was determined by combining the average hair cell counts for each region and taking a ratio of peroxide treated to control culture hair cell survival.
Cultures were examined with a Nikon fluorescent microscope with a 60× oil immersion lens. Images were captured with a SPOT 2 cooled CCD color digital camera and saved in our image data storage system. Positive staining was quantified objectively using IP Lab image analysis software.
Oxidative Stress Exposure
Cultures were prepared as described above. Explants were exposed to 0.01 to 5 mM H2O2 for 1 hr. Peroxide medium was produced using DMEM N1 + glucose to carry out serial dilutions of hydrogen peroxide 30% stock (Sigma). Hydrogen peroxide stock was kept at 4°C in the dark and was regularly checked for potency by spectrophotometry. Cultures were immediately exposed to the diluted peroxide medium once it was made up. After exposure, cultures were washed in medium to remove excess peroxide and new DMEM N1+ glucose medium was replaced. Cultures were maintained at 37°C with 5% CO2. After 5 days, in vitro cultures were washed in PBS and hair cells visualized by phalloidin staining as described above. Based on the hair cell survival data, 0.01 mM H2O2 for 1 hr exposure was used for all further experiments.
At set time points (12, 24, 48, 72, and 120 hr post 0.01 mM H2O2 for peroxide treatment), the specimens were harvested and stained for caspase activity using CaspaTag™ pan caspase Activity Kit (Serologicals® Corp.). About 10 μL of carboxyfluorscein benzyloxycarbonyl-valyalanylaspartic acid fluoromethyl ketone (FAM-zVAD-FMK) was added to every 300 L of medium. Cultures were protected from light and returned to the incubator for 1 hr. Medium was then removed and cultures were washed four times for 20 min each with PBS. The specimens were then fixed and counterstained with phalloidin TRITC. Hair cells undergoing apoptosis were determined by counting caspase positive cells that showed staining of stereocilia in three high power fields per cochlear region as described above. Results were compared with total hair cell survival data at 5 days post-peroxide. Positive controls for pan caspase activation were obtained by treating P3 cultures with neomycin 0.001M for 6 hr and performing the CaspaTag™ assay as described above (data not shown).
Double Oxidative Stress Treatment
P3 cultures were exposed to 0.01 mM H2O2 for 1 hr and cultured for 48 hr as described above. Cultures were then again exposed to 0.01 mM H2O2 for 1 hr and either stained for caspase activity 3 hr post the second peroxide exposure or cultured for a further 48 hr and stained for the presence of hair cells using phalloidin-FITC.
Evaluation of Mitochondrial Membrane Potential
Organ of Corti explants was harvested and exposed to 0.01 mM hydrogen peroxide as described above. At 24, 48, 72, and 120 hr post-peroxide, cultures were exposed to 5,5 ′6,6′-tetraethylbenzimidazolyl-carbocyanine iodide (Depsipher, Trevigen) for 30 min at 37°C as per the manufacturer's directions. This potential sensitive dye is excitable and fluoresces at 510/527 nm (Green) when it is in the monomeric form (mitochondrial transition pore open) and when inside an intact mitochondrion aggregates to excite and emit at 585/590 nm (Red). Imaging was done immediately after 30 min dye-loading time and the ratio of aggregate/monomer determined at base, mid, and apex turn for 20 outer hair cells in each explant (n = 20). Values normalized by expressing results as a base/apex ratio.
Evaluation of Cytochrome Oxidase Function
To determine if oxidative stress affected the mitochondrial function, a histochemical analysis of cytochrome oxidase function was carried out. Cultures were exposed to 0.01 mM H2O2 for 1 hr and then cultured for 24, 48, 72, and 120 hr. Cytochemical staining was carried out by combining 10 mg of 0.1% DAB (Sigma, St. Louis), 10 mg of 0.1% cytochrome c (horse heart) (Sigma, St. Louis), and 2 mg of 0.02% catalase (Sigma, St. Louis) are added to 10 mL of PBS. The solution was vortexed for 2 min until the solvents are completely dissolved, filtered, and stored in the the dark. At the designated time points, incubation was interrupted and the media removed from the specimens. The samples were then resuspended in 1 mL of the staining solution and returned to the incubator for 60 min.
After 1 hr, the staining reaction was terminated by washing the samples three times with ddH2O. The samples were fixed in 4% paraformaldehyde and mounted to slides for analysis and photography with DIC. The procedure was repeated for both the treatment and control groups at all five time points and photographs obtained.
Inhibition of Mitochondrial Function With Chloramphenicol
Organ of Corti cultures was prepared as described above and treated with chloramphenicol (Sigma, St. Louis) 1 mg/mL for 48 hr. Cultures were exposed to 0.01 mM H2O2 for 1 hr and then cultured for 24, 48, 72, and 120 hr. Cultures were assayed for caspase activation and stained for the presence of stereocilia with phalloidin as described above.
Evaluation of Mitochondrial Gene Expression and DNA Deletion
Organ of Corti cultures was prepared as described above, exposed to 0.01 mM H2O2 for 1 hr and then cultured for 72 hr as described above. The explants were washed in RNAse-free PBS and mRNA was extracted. Semiquantitiative RT-PCR was carried out using the Light Cycler RNA amplification kit, Syber Green I (Roche Applied Science) for the COX IV subunit (CAACAACCCCGTATTAACCG; TTGCT TGATTTAGTCGGCCT). Amplification of G6P and 18SRNA served as controls. An additional set of explants was used to assay for mitochondrial deletions. Explants were exposed 0.01 mM H2O2 for 1 hr and then cultured for 120 hr as described above. Explants were washed and DNA was extracted. Mitochondrial DNA deletion assay was carried out by amplifying mitochondrial DNA using three primer sets at different positions on the mitochondrial DNA (L1 8858 TCTATTCATC GTCTCGG AAG; L2 12883 TACCATTC CTAACAGGGTTC; H 13354 TTTATGGGTG TAATGCG GTG). Primer sets L2 and H amplify a control segment of mitochondrial DNA. Amplification of L1/H segments of mitochondrial DNA occur only if the “common” deletion associated with aging has occurred, brining the L1 and H primer sets closer to one another.
All results were evaluated using the Student t test and ANOVA testing where appropriate using SPSS™ statistics software (Chicago, IL)
Effect of Low Dose Oxidative Stress on Organ of Corti Cultures
P3 organ of Corti cultures was prepared and then treated with concentrations of H2O2 ranging from 5 to 0.01 mM for 1 hr. At doses above 0.05 mM, outer hair cell survival was seen only at the apex, with the greatest trend toward oxidative injury at the basal turn. Higher concentrations of peroxide resulted in complete destruction of the explant (Fig. 1B). Use of a 0.01 mM concentration of H2O2 resulted in loss of hair cells in the basal turn and occasional loss in the middle turns (Fig 1B). At the lowest concentration of peroxide used, outer hair cells showed a 25% loss in the basal third of the cochlea (Fig 1B). Survival of inner hair cells was more robust with over 92% of inner hair cells surviving in all turns at 5 days postinjury (Fig 1A).
Induction of Apoptosis After Oxidative Stress
P3 cultures were exposed to 0.01 mM hydrogen peroxide, which in previous experiments was shown to cause loss of 25% of hair cells at the basal turn of the explants. At 12, 24, 48, 72, 96, and 120 hr post-treatment, cultures were labeled with FAM-zVAD-FMK, fixed, and then counterstained with phalloidin-TRITC. Figure 2 demonstrates caspase (+) outer hair cells (Fig. 2A) that can be seen 12 hr after peroxide treatment. As seen in the matching phalloidin stained image (Fig. 2B), this represents a caspase positive cell in the outer hair cell region. Control cultures (Fig. 2C,D) show no caspase (+) cells. The inset demonstrates a high magnification merged image of a culture that has been exposed to 0.01 mM hydrogen peroxide. The image shows green (caspase positive) indicating the hair cells undergoing apoptosis. Inner hair cells undergoing apoptosis were seen only rarely, irrespective of position of time post-peroxide exposure.
Effect of Repeat Oxidative Stress on Hair Cell Survival
To determine the effect of recurrent oxidative stress, organ of Corti explants were exposed twice to 0.01 mM H2O2 for 1 hr. Each individual peroxide exposure was not enough to induce significant cell death (Fig. 1) but sequential dosing resulted in complete destruction of almost all auditory hair cells by 72 hr in vitro (Fig. 3).
Effect of Low Dose Oxidative Stress on Mitochondrial Membrane Potential (ΔΨ)
Exposure of hair cells to 0.01 mM H2O2 for 1 hr resulted in a significant reduction of mitochondrial membrane potential at the mid and basal turns. Imaging of the peroxide-exposed cultures demonstrates a visible decrease in the aggregated form of 5,5′6,6′-tetraethylbenzimidazolyl-carbocyanine iodide within the hair cells. The monomeric form of the dye can be seen as green fluorescence within both control and peroxide-exposed hair cells. (Fig. 4). When expressed as a ratio to apical cells, decrease in the mitochondrial membrane potential can be seen within 24 hr after exposure. The decrease in ⊗¬ is maintained throughout the duration of the experiment (Fig. 5).
Low Dose Oxidative Stress Impairs Mitochondrial Function and Gene Expression
To determine the effect of low-dose oxidative stress on mitochondrial gene expression, explants were treated with hydrogen peroxide. Subsequently, mitochondrial gene expression was assayed by semiquantitative RT-PCR and enzymes histochemical staining for cytochrome oxidase (Cox), which is dependent on mitochondrial gene expression, was carried out. As can be seen in Fig. 6, oxidative stress results in loss of Cox activity in the outer hair cells. This occurs within 24 hr after exposure to peroxide and is maintained through 72 hr in vitro. Semiquantitative RT-PCR for COX subunit IV demonstrated a 1.3-fold decrease in expression compared to control cultures. Using a mitochondrial DNA deletion assay, deletions of the mitochondrial genome were detectable by 5 days after peroxide exposure.
Inhibition of Mitochondrial Function Predisposes Hair Cells to Stress-Induced Apoptosis
Mitochondrial protein synthesis was impaired by pretreating organ of Corti cultures with chloramphenicol. Subsequent challenge of the explants with 0.01 mM hydrogen peroxide resulted in widespread activation of apoptosis and loss of hair cells (Fig. 7).
Oxidative stress and cellular damage due to ROS are thought to underlie a variety of pathogenic mechanisms in the cochlea including ototoxicity, noise trauma, and ischemic injury (Clerici, 1996; Kopke et al., 1997; Ohlemiller and Dugan, 1999; Ohlemiller et al., 1999; Ohinata et al., 2003). Using a long-term neonatal organ of Corti culture model (Malgrange et al., 2002), we have developed an in vitro model of oxidative injury that uses a low dose of hydrogen peroxide to induce hair cell injury. The dose of peroxide was chosen based on the observation that up to 250 M H2O2 can be generated within a cell (Arnaiz et al., 1999), thus the lower concentrations of peroxide (0.01mM) used in the experiment lie within a physiologic range. One of the goals of the experiment was to define a dose response curve for oxidative injury, so that a dose of pro-oxidant could be defined that injuries a defined population of cells while sparing other hair cells. This in turn allows for the evaluation of the nonlethal consequences of oxidative injury in surviving hair cells. At the lowest dose of hydrogen peroxide used, we found that basal turn outer hair cells were most vulnerable to damage while inner hair cells and the outer hair cells of the mid and apical turns were largely spared. As observed in a multitude of studies, basal turn hair cells are more vulnerable to oxidative damage and this effect is preserved in this neonatal in vitro model. Differential levels of antioxidant levels as observed by Sha et al., (2001) may explain this effect. Loss of hair cells after a low-level peroxide challenge occurred predominantly in a delayed fashion rather than the immediate loss of hair cell viability described after higher doses of peroxide (Dehne et al., 2000).
In our current study, low concentration hydrogen peroxide (0.01mM) treatment resulted in only a very low number of outer hair cells undergoing apoptosis. The distribution of apoptotic cells is consistent with previous data, demonstrating a vulnerability of the basal turn of the cochlea to damage that occurs early after the initial exposure to peroxide. When comparing this data to the survival data generated by counting hair cell stereocilia, only 50% of missing hair cells can clearly be accounted for by apoptotic death when correcting for control cultures. This could be accounted for in several different fashions. The period that caspase is active could be limited and therefore we may be missing individual hair cells undergoing apoptosis. Alternately, hair cells may be undergoing apoptosis via caspase independent mechanisms. Interestingly, even the hair cells that survive appear altered after being exposed to minimal oxidative stress. Double treatment of cultures with low-dose peroxide resulted in a second wave of apoptosis, in this case also affecting inner hair cells. This culture system spatially and temporally models a variety of physiologically occurring oxidative injury in the cochlea and could model the effect of delayed or continuing ROS production in the inner ear after injury (Hu et al., 2002; Yamashita et al., 2003). An important observation in this model is that double peroxide treatment results in a higher number of apoptotic cells and overall greater tissue damage than would be expected if effect of two individual peroxide doses caused a simple additive effect. An initial oxidative insult may therefore sensitize hair cells to further cell death.
The exact mechanism of oxidative damage in the inner ear has not been clearly defined. Studies of ischemia reperfusion injury suggest that hydroxyl radical plays an important role in the initiation of cell damage (Ohlemiller, 1999). This could either be generated in vivo through conversion of H2O2 by the Fenton reaction or through reactions with peroxynitrate, production of which is associated with sound trauma (Clerici et al., 1999). Besides inducing apoptosis, oxidative stress can also induce mitochondrial damage leading to abnormal cellular metabolism (Hoyt et al., 1997; Fiskum et al., 1999). This in turn could lead to failure of energy production and necrosis. The cell death due to necrosis observed in these studies and in recent studies of noise trauma could be potentially explained by this mechanism. Recent studies have shown dysfunction of succinate dehydrogenase after sound trauma (Hu et al., 2003), suggesting that mitochondrial failure may be central to oxidative stress-induced hair cell death. Based on studies in other systems, we examined the function of hair cell mitochondria in oxidative stress. Even a single low dose of peroxide results in the reduction of the mitochondrial membrane potential. The loss of mitochondrial membrane potential represents a loss of cellular bioenergetics. This is reflected in the reduction of cytochrome oxidase activity in hair cells after exposure to oxidative stress. Initial studies appear to show that there is a corresponding reduction in mitochondrial-derived COX mRNA. At later time points, mitochondrial DNA deletions are detectable by PCR suggesting that even nonlethal oxidative stress has the potential to permanently alter mitochondria.
As the population of damaged mitochondria increase within a cell, mitochondrial protein synthesis decreases and ultimately oxidative phosphorylation is compromised. We reduced mitochondrial protein synthesis in hair cells by culturing organ of Corti explants with chloramphenicol. The dose of chloramphenicol was chosen, so there was a 40–50% reduction in COX histochemical stain (data not shown). When chloramphenicol-pretreated explants were challenged with a single low dose of hydrogen peroxide, widespread hair cell loss resulted similar to double peroxide-treated cultures.
A single exposure to oxidative stress, such as sound trauma or an ischemia reperfusion injury could cause, may therefore cause permanent changes to the hair cells that could at later time points make them more susceptible to damage. It is likely that this is mediated by damage to the mitochondria. Further studies will be needed to determine what proportion of the cell's mitochondrial need to be damaged before making a hair cell more susceptible to cell death. In addition, we will have to determine if these mechanisms are valid in vivo.
We have developed an in vitro model of hair cell oxidative stress using a long term P3 organ of Corti culture system. A variety of effects can be demonstrated using this culture system. Oxidative stress is induced by addition of a physiologically relevant dose of hydrogen peroxide to the culture system. Higher doses of peroxide induce widespread hair cell loss, whereas lower doses of hydrogen peroxide produce minimal loss of hair cells via apoptosis. Hair cells that are preserved appear to have alteration in mitochondrial function that predisposes them to further injury.