Large intersubject variabilities in acoustic injury are known to occur in both humans and animals; however, the mechanisms underlying such differences are poorly understood. The olivocochlear efferent system has been hypothesized to play a significant role in protecting the cochlea from noise overexposure. In this study, we demonstrate that a newly developed test for determining average efferent system strength can predict intersubject variations in acoustic injury. In addition, the intersubject variability in cochlear expression of the α9 subunit of the nicotinic acetylcholine receptor was found to be proportional to an animals average efferent strength. Therefore, the inter-animal variability in the α9-containing acetylcholine receptor expression may be one mechanism contributing to the inter-animal variability in acoustic injury.
- hair cells
- α9 subunit of the nicotinic acetylcholine receptor
- distortion-product otoacoustic emissions
- noise-induced hearing loss
- noise susceptibility
A current dilemma in investigating acoustic trauma is the intersubject variability to noise-induced hearing loss. This well recognized variability in reaction to noise exposures is so pronounced that common vernacular has been developed in at least two languages to refer to ears that are less susceptible to noise damage as “steel” (German) or “tough” (English) ears, whereas more susceptible ears are termed “glass” (German) or “tender” (English) ears. It is estimated that acoustic overexposures may be responsible for at least 10 million cases of noise-induced hearing loss in the United States alone.
One protective mechanism against noise damage is the acoustic reflex, which uses the actions of the two middle ear muscles, the tensor tympani and stapedius, to physically dampen sound-induced oscillations of the eardrum and basilar membrane. In addition, the cochlea receives feedback information that originates from neurons in the superior olivocochlear brainstem region. The efferent olivocochlear projection from these brainstem neurons has been hypothesized to improve signal detection in the presence of noise (Winslow and Sachs, 1987, 1988), to contribute to protection from acoustic overexposures that would result in temporary hearing loss (Rajan and Johnstone, 1988), and, more recently, to contribute to protection from a permanent noise-induced hearing loss (Maison and Liberman, 2000). A number of findings suggest that the variation in the robustness or strength of this efferent reflex may underlie the well recognized variation in acoustic susceptibility. These observations include previous demonstrations: (1) the reflex strength of the olivocochlear system, as measured with either transient otoacoustic emissions or distortion-product otoacoustic emissions (DPOAEs), varies considerably among normal-hearing humans (Veuillet et al., 1991; Collet et al., 1992); and (2) chronically de-efferented ears are more vulnerable to noise-induced permanent threshold shifts (Kujawa and Liberman, 1997;Zheng et al., 1997), and de-efferentation causes the intersubject variability in noise susceptibility to disappear (Patuzzi and Thompson, 1991). In addition, a recent report by Maison and Liberman (2000) in alert guinea pigs using DPOAE measures showed that the amount of olivocochlear reflex strength present at one single frequency was able to predict susceptibility to acoustic injury.
It has long been known that the intracochlear application of acetylcholine (ACh) mimics the effects of electrically stimulating the olivocochlear efferent fibers (Bobbin and Konishi, 1971; Kujawa et al., 1993). Moreover, the α9 subunit of the nicotinic ACh receptor (α9s nAChR) was cloned and shown to be expressed by the targets of olivocochlear efferent fibers, i.e., the outer hair cells of the cochlea (Elgoyhen et al., 1994; Glowatzki et al., 1995; Morley et al., 1998; Simmons and Morley, 1998). In outer hair cells of the cochlea, the α9 subunit of the nicotinic acetylcholine receptor may act in concert with the α10 subunit, the newest member of the nicotinic ACh receptor family (Elgoyhen et al., 2001; Lustig et al., 2001); however, the α9 subunit is essential for the best-studied cholinergic effect on cochlea function, because α9 null mutant animals appear functionally de-efferented because they lack suppression of electrically evoked olivocochlear responses (Vetter et al., 1999). In hair cells, ligand-gated Ca2+ entry through the α9/α10 receptor (Fuchs and Murrow, 1992) is coupled to K+ efflux and intracellular hyperpolarization (Housley and Ashmore, 1991). This hyperpolarization of outer hair cells affects their electromotile responses and decreases their mechanical responses to sounds and elevates the auditory thresholds (Wiederhold and Kiang, 1970; Winslow and Sachs 1987,1988).
This current study used a newly developed test to assess the average strength of an animal's efferent response (Luebke et al., 2001). These experiments are designed to determine whether the cholinergic efferent system plays a role in protecting the cochlea from acoustical injury. Baseline distortion-product otoacoustic emissions were monitored, and average efferent strengths were determined. Animals were subsequently exposed to noise, and the amount of noise damage was determined by noting the difference in distortion-product otoacoustic emissions and the difference in the distortion-product otoacoustic emission thresholds. In addition, the average efferent strengths were calculated and compared with cochlear α9s nicotinic acetylcholine receptor expression.
MATERIALS AND METHODS
Animals. Twenty-one pigmented guinea pigs (strain 2/Ncr; Charles River Laboratories, Wilmington, MA), weighing 300–400 gm, were used as experimental subjects. All procedures were approved and monitored by the Institutional Animal Care and Use Committee of the University of Miami.
α9 nAChR antibody generation. An antibody (MU43f) against the proposed intracellular loop (between transmembranes III and IV) of the guinea pig α9 nAChR was generated in rabbits (Covance, Princeton, NJ) against a synthetic peptide (SKPKTARNKDL) conjugated to keyhole-limpet-hemocyanin. The MU43f antibody was tested by both ELISA analysis and Western analysis and is specific for the α9 nAChR in both tissue and heterologous expression studies, and this peptide sequence is not contained in any other neuronal nicotinic AChR subunits (α2–10, β2–4) discovered to date.
Immunohistochemistry. The samples were fixed in 4% paraformaldehyde for 2 hr, and either whole mounts or 20 μm sections were blocked with 10% normal goat serum in PBS. Sections were incubated in the anti-α9 MU43f antibody at 1:3000 in fresh blocking solution overnight at 4°C, rinsed, and incubated in rhodamine-conjugated donkey anti-rabbit (Chemicon, Temecula) at 1:150 in 0.01 m PBS for 48 hr at 4°C. Endorgans were mounted using Vectashield on slides with spacers inserted so as not to crush the tissue. Slides were stored horizontally at 4°C in the dark and were imaged using a Zeiss (Oberkochen, Germany) LSM 510 confocal microscope at the University of Illinois at Chicago confocal microscope facility.
Western blot analysis. Guinea pig tissues were homogenized in 500 μl of radioimmunoprecipitation analysis buffer (1m NaCl, 1 m Tris, 0.1% NP-40, 0.05% deoxycholate, and 0.01% SDS) and incubated overnight at 4°C in the presence of a protease inhibitor cocktail (Boehringer Mannheim, Mannheim, Germany). Equal volumes of guinea pig cochlear proteins (40 μl, ∼30 μg of protein) or 30 μg of other control tissue proteins (e.g., brain, skeletal muscle, or pituitary) were separated by SDS-PAGE on a 4–15% gradient gel. The gel was electroblotted to an Immobilon P membrane (Millipore, Bedford, MA), and Western blot was performed as described in the fast-blot protocol for Immobilon P membranes. Membranes were incubated for 60 min in PBS–1% nonfat milk with primary antibody [MU43f anti-α9 nAChR antibody at 1:500 dilution or anti-oncomodulin monoclonal antibody at 1:10 dilution (gift of Dr. M. Henzl, University of Missouri School of Medicine, St. Louis, MO)]. The membranes were then washed and incubated in the appropriate secondary antibody (anti-rabbit for α9 nAChR or anti-mouse for oncomodulin) conjugated to horseradish peroxidase for 30 min (Jackson ImmunoResearch, West Grove, PA). In the final step, the membranes were washed, incubated in chemiluminescent substrate (Pierce, Rockford, IL), and exposed to film (Kodak LS; Eastman Kodak, Rochester, NY). Prestained molecular weight markers were also run in a parallel lane to allow size estimations of the protein bands (Novex, Wadsworth, OH). The resulting autoradiographs were scanned, and band densities were determined using BioMax 1D image analysis software, version 1.6 (Eastman Kodak).
Oocyte membrane protein isolation. Membranes fromXenopus laevis oocytes were isolated according to the method developed by Parker et al. (1998). Briefly, oocytes were homogenized in buffer (in mm: 50 Tris, 120 NaCl, 5 KCl, 1 MgCl2, and 2.5 CaCl2, pH 7.0) using a Brinkman Instruments (Westbury, NY) model PT 10/35 homogenizer. Homogenates were centrifuged at 4°C at 2000 × g for 10 min. The supernatant was removed and then centrifuged at 4°C at 45,0000 × gfor 20 min. The supernatant was discarded, and the membrane pellet was resuspended in buffer. Approximately 125–150 oocytes yielded 20 μg of membrane protein.
DPOAE measurements. Guinea pigs were sedated with 40 μg/kg ketamine HCl, and DPOAEs at 2f 1-f 2were measured using software described previously (Martin et al., 1999), which included ear speakers (ER-2; Etymotic Research, Elkgrove Village, IL) to produce the f 1 andf 2 primaries and a commercially available emissions measuring system with an acoustic probe encased microphone assembly (ER-10B+; Etymotic Research) to measure the emitted responses. Both stimulus generation and response acquisition were under computer control. DPOAEs were obtained from both ears as standard distortion product grams using primary-tone levels ranging from 45 to 75 dB sound pressure level (SPL), in 5 dB steps, at geometric-mean (GM) frequencies of thef 1 andf 2 tones [GM = (f 1 *f 2)1/2] ranging, in 0.1-octave steps, from 1.4 to 17.8 kHz, in whichf 2/f 1= 1.2.
Average efferent determination. The magnitude of the efferent effect varies greatly across frequencies. To test efferent after-effects throughout a broad range of frequencies, the DPOAE differences were computed over a two-octave frequency range, i.e., from 2.8 to 11.3 kHz. Briefly, the baseline DPOAEs were measured in the test ear using monaural stimulation (L 1 =L 2 = 70 dB SPL) before the onset of an efferent-based adaptive response to minimize any efferent effect (Liberman et al., 1996). A second test was then performed over the same frequency region with L 1 =L 2 = 70 dB SPL but, instead, with binaural stimulation with longer primary tones continuously presented for 1 sec before measuring the DPOAEs to maximize the efferent-induced effects. To allow the efferent-based DPOAE measurements to reset, a 2.5 sec interblock interval was allotted for both the baseline monaural and the binaural tests. The absolute value of each of the differences between the baseline and binaural stimulation DPOAE levels were then summed and averaged over the two-octave range by dividing this sum by the number of tested frequencies, with the result representing the average efferent strength for the test ear in units of decibels. The statistical package (StatView, version 4.5; Abacus Concepts, Berkeley, CA) was used to determine linear regression coefficients andp values.
Average acoustic reflex determination. The difference between the f 1 tones during baseline monaural and binaural stimulation was computed, summed, and averaged over the same frequency range as the average efferent measure to provide a measure of the average acoustic reflex, which reflected a change in the ear-canal impedance that would result if the 1 sec binaural primary-tone stimulation activated the middle ear muscles.
Noise exposure. One day after prenoise exposure, control DPOAEs were obtained, and guinea pigs were placed inside a sound-reverberant chamber in which they were allowed access to food and water ad libitum and exposed for 4 hr to a 109 dB SPL octave band of noise, ranging from 2 to 4 kHz. After a 1 week recovery period, postexposure DPOAEs were measured as before to compare preexposure with postexposure DPOAE levels.
Average DPOAE loss and DPOAE threshold shift determinations.To quantify the loss in DPOAEs attributable to the noise overexposure, another measure was defined termed the average DPOAE loss, which represented the average difference between the pre-DPOAEs and post-DPOAEs (elicited at 65 dB) summed over the frequencies from 2.8 to 11.3 kHz, i.e., over the same frequency region used to determine the average efferent and average acoustic reflex measures. In addition, the change in isoresponse DPOAE contour “thresholds” were determined at a 3 dB SPL criterion. Average DPOAE threshold shifts were determined between the pre-DPOAE and post-DPOAE thresholds over the frequencies from 2.8 to 11.3 kHz.
The fast adaptive efferent test developed by Liberman et al. (1996) to test olivocochlear function in the cat was modified for application to the guinea pig using both monaural and binaural constant-tone stimulation. An example of the results obtained by this technique is presented in Figure1 A for eliciting the DPOAE at a frequency of 5.6 kHz stimulation. This plot shows that, similar to the cat, the efferent-induced reduction in DPOAEs was greatest with binaural than with monaural stimulation. The greater DPOAE adaptation with binaural stimulation is presumably attributable to more olivocochlear neurons being stimulated with binaural stimulation than with monaural sound stimulation. The binaural efferent effect on DPOAEs was stable within the same test session in which retests were performed every 5 min over a 50 min period, with variations of <0.2 dB. Moreover, the amount of adaptation was the same when an animal was retested 2 d later (data not shown). Figure1 B shows the effect of binaural efferent stimulation on DPOAE levels as a function of frequency, for the same guinea pig as shown in Figure 1 A for a single frequency of 5.6 kHz. Figure 1 C shows the equation used to determine the average efferent strength for all animals tested. More details on this testing protocol can be found by Luebke et al. (2001).
The two plots of Figure2 A show the difference in decibels between the baseline and the effect of binaural stimulation for one animal with a large average efferent strength [guinea pig (gp) 87L; filled squares] and one animal with a small average efferent strength (gp 89L; filled triangles). The differences between the f 1 tones with baseline and binaural stimulation were also computed over the same frequency range to provide a measure of the average acoustic reflex, which reflects an impedance change in the ear canal if the middle ear muscles are activated by the longer tones. However, ketamine anesthesia significantly raised the acoustic reflex threshold and prevented the reflex from occurring, as shown in Figure 2 B. Therefore, in the sedated guinea pig, the acoustic reflex did not contribute significantly to the average efferent measure.
To determine whether the average efferent strength measure can be used to predict susceptibility to noise overexposure and whether it can account for the intersubject variability in their susceptibility to noise damage, average efferent strengths were computed in a group of guinea pigs that were then subjected to the free-field octave band noise exposure, i.e., a 109 dB SPL exposure from 2 to 4 kHz for 4 hr, delivered within a reverberant sound chamber. After a 1 week recovery period, the difference between their preexposure and postexposure DPOAEs was computed, and this difference was then summed across the tested frequency range and divided by the number of tested frequencies (n = 11) to yield a number reflecting each guinea pig's DPOAE loss. Figure 3 Ashows the prenoise versus postnoise exposure DPOAEs for the animal shown previously to have a large average efferent strength in Figure2 A. It is evident that noise exposure did not significantly alter the DPOAEs in this subject. In contrast, exposure to the same level and duration of noise significantly reduced the DPOAEs in a second guinea pig shown previously to have a small average efferent strength (Fig. 2 B). Figure 3 Bshows DPOAE responses as a function of sound pressure level (i.e., growth response functions for both prenoise and postnoise exposure) for two frequencies (2 and 4 kHz) for the two guinea pigs shown in Figure3 A. As can be appreciated, the loss of gp 89 (weak efferent strength) existed for both threshold and suprathreshold sounds, whereas gp 87 (strong efferent strength) exhibited very little difference between the prenoise and postnoise exposure DPOAEs.
To quantify the loss in DPOAEs attributable to the noise overexposure, another measure termed the average DPOAE loss was used. The DPOAE loss value represents the average difference between the pre-DPOAEs and post-DPOAEs summed over the same frequency region used to determine the average efferent and acoustic reflexes. When the average DPOAE loss value was calculated for the animals shown in Figure 3, the guinea pig with a strong average efferent strength exhibited an average DPOAE loss of only 3.0 dB (Fig. 3 A), whereas the guinea pig with a weak average efferent strength displayed an average DPOAE loss of 29.52 dB (Fig. 3 B). For all animals, the noise-induced reduction in DPOAE levels in the form of the average DPOAE loss was analyzed atL 1 =L 2 = 65 dB SPL, which allowed the postnoise emissions to be above noise-floor levels. When the average DPOAE threshold shifts were determined for all animals tested, there was a range of DPOAE threshold shifts of ∼35 dB. Again, as shown in Figure 4 B, there was an inverse correlation between animals with weak average efferent strengths (<1 dB) exhibiting the great DPOAE threshold shifts (>20 dB) and animals with greater efferent strengths exhibiting lower threshold shifts (<5 dB).
The average efferent strength and the average DPOAE losses were inversely correlated for all cochleas (n = 11) tested, allowing a prediction of susceptibility to noise exposure (r = 0.78; p < 0.005; as shown in Fig.4). Furthermore, the average efferent strength and the average DPOAE threshold shifts were also inversely correlated for all cochleas (n = 11) tested but to a lesser extent (r = 0.69; p < 0.02). As shown previously in Figure 3, the animal with the greatest loss of DPOAEs after noise exposure also had the lowest average efferent value (gp 89L; filled triangles). In contrast, the animal that showed the least loss of DPOAEs after noise exposure exhibited the highest average efferent value (gp 87L; filled squares).
The presence of α9s nAChR on outer hair cells receiving efferent innervation suggests that acetylcholine signaling might be mediating the average efferent measure. Western blot analysis using an antibody against the α9s nAChR was performed to determine whether the average efferent strength of an intact cochlea reflected the amount of α9snAChR present in tissue extracts of that cochlea.
The polyclonal antibody (MU43f) was specific for the α9s nAChR subunit from guinea pig, and, as shown in Figure 5 A, this antibody specifically recognized α9s protein near the base of outer hair cells and a lesser amount of α9s protein immunostaining is also found in inner hair cells, consistent with in situ hybridization findings in the rat (Morley et al., 1998). We also found using immunohistochemistry that this antibody recognizes α9s nAChR in the vestibular system. In the vestibular system, type I hair cell calyces and the bases of type II hair cells contain α9s nAChR protein, consistent with the known location of cholinergic efferent endings (A. Luebke, P. Maroni, S. Guth, and A. Lysakowski, unpublished results).
As shown in Figure 5 B, brain and muscle tissues did not react with this antibody and only recognized protein found in the cochlea, vestibular system, and pituitary, which are the known locations for the α9 nAChR subunit (Elgoyhen et al., 1994). In addition, preabsorption with the synthetic peptide eliminated the antibody reactivity (data not shown), as did incubation with the preimmune sera. The antibody did recognize α9sprotein from α9 cRNA-injected Xenopus laevisoocytes and showed no spurious bands for uninjected oocytes, as shown in Figure 5 C. Using the MU43f antibody, the amount of α9s nAChR protein present in the cochlea was found to be highly variable among animals, as shown in Figure5 D. However, the amount of α9s nAChR protein between the two ears of the same animal was equivalent, as shown in Figure 5 E.
The amount of α9 nAChR protein present in the cochlea correlated with the magnitude of the average efferent strength, as shown in Figure5 F, in which the magnitude of the average efferent strength is plotted against the amount of α9 nAChR present in the corresponding cochlea. To control for differential protein recovery or gel-loading differences, these blots were stripped and then reprobed with a monoclonal antibody generated against oncomodulin. Oncomodulin is a calcium-binding protein present only in outer hair cells of the guinea pig cochlea (Sakaguchi et al., 1998). Thus, normalizing the blots to the amount of oncomodulin present in each cochlea serves as a control for equal number of outer hair cells loaded into each lane. When the α9s nAChR protein amount was normalized by the amount of oncomodulin present in the cochlea, the correlation between α9 nAChR protein and the average efferent strength measure yielded a correlation coefficient of 0.94 (p < 0.002; n = 6 animals), as shown in Figure 5 F. Thus, the amount of α9 nAChR present in a given animal's cochlea correlates with the average efferent strength for that particular guinea pig.
The present study demonstrates that inter-animal variation in α9 nAChR cochlear protein expression correlates with the inter-animal variability in the average efferent strength index. The variability in the average efferent strength index also correlates with the susceptibility to noise overexposure. Therefore, we conclude that the variation in α9s nAChR expression in adult animals correlates with differences in susceptibility to noise-induced hearing loss. The current study cannot determine whether α9s nAChR expression is mediating susceptibility or whether its expression has changed parallel to other factors. Therefore, it is not clear what causes the variability in the α9s nAChR protein expression observed in the guinea pig strain tested in the current study. Additionally, because the functional assessment of efferent reflex strength is based on olivocochlear feedback, it is not yet known whether the increased α9 nAChR protein levels reflect an increase in olivocochlear fibers projecting to the cochlea. In fact, when the α9 nAChR was knocked-out in 129 SvEv/CBA mice, those animals showed no efferent-related response when assayed using direct electrical stimulation of the efferent system (Vetter et al., 1999) and presumably would have an absent average efferent strength index. Whereas this α9s nAChR knock-out mouse could be used to address the relationship between α9s nAChR expression and noise susceptibility, the mouse strain used for the knock-out, and most knock-out animals generated to date, used embryonic stem cells from the 129 SvEv mouse for the genetic recombination step. Because the 129/SvEv mouse strain shows exceptional resistance to acoustic injury, a straightforward test of the protective role of α9s nAChR in the knock-out mouse is confounded (Yoshida et al., 2000).
The acoustic reflex was greatly reduced in the sedated guinea pigs of the present study. Thus, the variability in activation of the acoustic reflex did not contribute significantly to the observed efferent strength variability. In accordance with these data, when Patuzzi and Thompson (1991) severed the olivocochlear bundle in guinea pigs, they noted that not only did the animals become more susceptible to acoustic injury but that the typical variability in their noise susceptibility also decreased.
In the present study, the average efferent strength was found to account for ∼61% of the inter-animal variability in their response to acoustic injury. This measure of efferent strength was especially effective at predicting acoustic injury for highly susceptible and weakly susceptible individuals. Perhaps the strength of each animal's acoustic reflex can account for the remaining variability in susceptibility, because the average efferent strength measurements performed in this study were determined under ketamine sedation and are not contaminated by the acoustic reflex measurement (Luebke et al., 2001). However, Maison and Liberman (2000) found that a single-frequency DPOAE measure tested on alert guinea pigs was also only 61% effective in predicting acoustic injury. Their single-frequency efferent measure was detecting a medial efferent and possibly an acoustic response, yet it also could not account for 100% of the inter-animal variability to acoustic injury. Perhaps there are other factors that we have not yet determined (i.e., strength of the lateral efferent pathway, hair cell stereocilia robustness, hair cell motility gain, etc.) that can account for the remaining inter-animal variability to acoustic injury.
The average efferent strength measure can be applied to human testing and could prove to be a useful functional test to determine both noise susceptibility and an objective diagnostic aid for brainstem-based auditory disorders. Currently, there are no objective tests for brainstem auditory impairments, yet many children may be afflicted with this problem each year, which clearly has the potential to impair their ability to integrate auditory information in noisy environments (Winslow and Sachs, 1987), thus impeding their progress in school. Because the average efferent measure specifically tests the integrity of the pathway from the brainstem to the cochlea, it can provide a reliable indicator of the functional status of that portion of the central auditory nervous system.
This work was supported by United States Public Health Service Grant R01 DC03086 and the Chandler Chair fund of the University of Miami. We thank Amanda Lowrey and Dr. Jiao He for technical assistance. We also acknowledge the skillful assistance and support of Dr. Anna Lysakowski and the confocal microscopy center of the University of Illinois at Chicago.
Correspondence should be addressed to Dr. Anne E. Luebke, University of Miami School of Medicine, Miami, FL 33136. E-mail:.