Corticosteroid Response and Supporting Cell Antibody in Autoimmune Hearing Loss FREE

McCabe¹ introduced the concept of autoimmune sensorineural hearing loss (AISNHL) to explain a pattern of idiopathic rapidly progressive hearing loss with a “period of progression . . . over weeks or months, not hours nor days nor years,” involving both ears either synchronously or sequentially, with or without vestibular symptoms. Multiple studies have attempted to confirm the existence of such an entity, better characterize its clinical picture, and develop methods for diagnosis. The clinical picture of AISNHL overlaps with other pathologic conditions. Therefore, a specific test for AISNHL would be of great clinical and research interest since AISNHL is a potentially treatable form of sensorineural hearing loss. Harris and Sharp² demonstrated that 32% of their patients with presumed AISNHL had antibodies to a 68-kDa protein in bovine inner-ear extract. Hughes et al³ reported similar findings in 86% of patients with idiopathic, progressive, bilateral sensorineural hearing loss. Similarly, Moscicki et al⁴ found antibody to a 68-kDa protein in 42 (58%) of 72 patients with active idiopathic, progressive, bilateral sensorineural hearing loss.

Our interest in antibody-mediated hearing loss developed as part of a study to identify proteins uniquely expressed in the inner ear. The KHRI-3 monoclonal antibody was developed from mice immunized with guinea pig organ of Corti.⁵ KHRI-3 binds to supporting cells in the organ of Corti with a characteristic punctate “wineglass” staining pattern and identifies a 68- to 72-kDa protein in inner-ear extracts.⁵ Mice carrying the KHRI-3 hybridoma develop high-frequency hearing loss and loss of outer hair cells in the basal turn of the cochlea.⁶ Similarly, KHRI-3 antibody infused into the guinea pig cochlea binds to supporting cells in vivo and results in hair cell loss and hearing loss, effects that were not seen in animals infused with isotonic sodium chloride solution or control antibody.⁷ Thus, antibody to supporting cells can damage the inner ear and affect its function. Patients with suspected AISNHL have antibodies to a 68- to 72-kDa protein and to inner-ear supporting cells that are similar to the KHRI-3 hybridoma antibody.⁸ Furthermore, serum with antibody to inner-ear supporting cells also identifies a 68- to 72-kDa protein immunoprecipitated from guinea pig inner-ear extracts by the KHRI-3 antibody,⁸ suggesting that the KHRI-3 monoclonal antibody and human antibodies recognize the same inner-ear supporting cell antigen (IESCA). Therefore, the human antibodies may also cause damage to the organ of Corti and induce hearing loss.

The objective of this study was to test the hypothesis that antibody to the 68- to 72-kDa inner-ear protein and/or to inner-ear supporting cells correlates with clinical response to corticosteroid therapy in patients with suspected AISNHL.

Patients with suspected autoimmune inner-ear disease were recruited from collaborating otology practices in Pennsylvania, Michigan, and Indiana. The University of Michigan Institutional Review Board approved the protocol for the study, and all patients gave written informed consent. A complete medical history, otologic examination, audiologic evaluation with pure-tone and speech audiometry, and tympanometry (American National Standards Institute, standard S3.6, 1986) were obtained. Appropriate imaging and laboratory studies were obtained to exclude other inner-ear disorders. Informed consent and control serum samples were obtained from 20 normal volunteers ranging in age from 20 to 52 years who had no hearing-related complaints. Serum samples were analyzed for a 68- to 72-kDa band on Western blot (WB) of guinea pig inner-ear extract and for antibodies to the IESCA by means of immunofluorescence (IF) on guinea pig organ of Corti surface preparations.

This analysis includes 63 patients who met the inclusion criteria of active, rapidly progressive, unilateral or bilateral sensorineural hearing loss documented by a pretreatment audiogram, treatment with a standard course of corticosteroids, and posttreatment audiologic assessment within 6 weeks of initial presentation. Active hearing loss was defined as a pure-tone threshold of 30-dB hearing level or greater at any single frequency with evidence of 10-dB or greater deterioration in any frequency within 3 months before the serum draw. Hearing improvement was defined as a greater than 10-dB threshold improvement at 2 consecutive frequencies and/or a 20% increase in speech discrimination score as previously defined by Moscicki et al.⁴

All patients were treated according to the clinical judgment of their physician without knowledge of the antibody assay results. Most of the patients (83% [52/63]) received prednisone, 1 mg/kg of body weight per day (maximum of 60 mg) for a minimum period of 7 days, followed by a tapering schedule. Patients who had an initial improvement in hearing that deteriorated as the corticosteroid dosage tapered were maintained on a corticosteroid regimen until successful weaning was possible. Eleven patients (17%) received methylprednisolone, 24 mg, as a loading dose that was tapered by 4 mg daily for 6 days. Nine of these patients showed a considerable improvement and no further treatment was given. The other 2 (patients 10 and 18) received additional treatment. Patient 18 received methylprednisolone, 60 mg/d for 7 days followed by a rapid taper, and patient 10 had a 30-day burst at 60 mg/d followed by a slow taper. Both improved after this more intense therapy.

The University of Michigan Committee on Use and Care of Animals approved all studies, and the Unit for Laboratory Animal Medicine provided veterinary care and housing. Guinea pigs (250-300 g) were anesthetized and decapitated. Tissue extracts were prepared as described previously.⁸ The organ of Corti, including the modiolus but without the spiral ligament, together with vestibular tissues from the ampulla, saccule, and utricle, was harvested and solubilized on ice in lysis buffer (1% nonionic detergent p-40 in phosphate-buffered saline [PBS], pH 7.2) containing protease inhibitors (1-mmol/L phenylmethylsulfonyl fluoride; leupeptin, 1 μg/mL; antipain, 2 μg/mL; benzamidine, 10 μL/mL; aprotinin, 10 kU/mL; chymostatin, 1 μg/mL; pepstatin, 1 μg/mL), homogenized, and allowed to stand at 4°C for 30 minutes. After low-speed centrifugation (1000 rpm for 3 minutes) to remove bone, the lysate was mixed 1:3 with 4× sample buffer to give these final concentrations: 0.0625-mol/L Tris hydrochloride, pH 6.8, 2% sodium dodecyl sulfate, 10% glycerol, 5% 2-mercaptoethanol, and 0.005% bromophenol blue. The samples were boiled for 2 minutes, loaded onto a 7.7 × 12.8-cm 7% polyacrylamide gel, and subjected to electrophoresis for 2 hours at 100 V per gel. Electrophoretic transfer to a nitrocellulose membrane was carried out at 25 V per gel overnight. The membranes were cut into strips, soaked in 5% milk in PBS–0.1% Tween to block nonspecific binding, and incubated for 2 hours at room temperature with human serum diluted 1:50 in the 5% milk solution. The strips were washed 3 times for 10 minutes each and incubated for 2 hours at room temperature with secondary antibody (goat anti–human IgG/IgM) conjugated with horseradish peroxidase (Jackson ImmunoResearch Laboratories Inc, West Grove, Pa), diluted to 1:500. The blots were washed 3 times and developed in 4-chloro-1-napthol (0.5 mg/mL) in methanol-PBS, pH 7.6 (1:5) containing 0.05% hydrogen peroxide for 1.5 to 2 minutes. Colored molecular weight markers (Amersham, Arlington Heights, Ill) were used to construct a semilog plot of relative molecular mass vs electrophoretic migration distance for each gel. Serum samples were scored for staining of a protein migrating at 68 to 72 kDa.

Guinea pig inner ear was fixed in freshly prepared 2% paraformaldehyde for 2 hours and then dissected free of bone, the spiral ligament, and the tectorial membrane. The modiolus and the organ of Corti were halved, incubated in 3% goat serum in PBS for 1 hour, and washed, and each half was incubated overnight at 4°C in patient or control serum diluted 1:50 in PBS. The specimens were washed, incubated for 45 minutes at 23°C with rhodamine-conjugated anti–human goat IgG/IgM heavy and light chain–specific antibody (Jackson ImmunoResearch Laboratories Inc), and diluted 1:200 in PBS (pH 7.4). After 3 washes the organ of Corti was dissected free of the modiolus, mounted in glycerol vinyl alcohol aqueous mounting medium (Zymed Laboratories, San Francisco, Calif), and examined by confocal microscopy. A “wineglass” staining pattern^{5 - 8} was considered positive for antibodies to IESCA.

Given the dichotomous nature of the study measures and outcomes, data were analyzed by means of the χ² test and the Fisher exact test as appropriate. A 2-sided α level of less than .05 was considered statistically significant.

Sixty-three of 159 consecutive patients recruited in a 5-year period by the collaborating practices met the inclusion criteria for active, unilateral or bilateral sensorineural hearing loss, with no other cause of hearing loss. Each had serum drawn immediately before the onset of corticosteroid therapy and had appropriate audiologic follow-up. There were 33 males and 30 females (Table 1) with a mean age of 47 years (range, 4-84 years; SD, 17.8). Eight (13%) had a history of one or more systemic autoimmune diseases, which included Cogan syndrome, systemic lupus erythematosus, unclassified autoimmunity, Raynaud disease, Hashimoto thyroiditis, rheumatoid arthritis, and Wegener granulomatosis (Table 1).

The left ear was affected in 25 patients, the right was affected in 12, and 26 had hearing loss in both ears (Table 1). Although hearing loss was the primary otologic symptom, 52 (83%) of the 63 patients also complained of tinnitus, 36 (57%) had aural fullness, and 36 (57%) had vestibular symptoms. Eight (13%) had rapid-onset hearing loss with progression over hours or days.

In the study sample, the mean ± SD of the 4-frequency (0.5, 1, 2, and 4 kHz) pure-tone average of all ears with active hearing loss was 52.7 ± 25.5 dB, and the mean speech discrimination score was 69.4% ± 29.9%. After treatment, the 4-frequency average for all ears with active hearing loss was 43.7 ± 23.5 dB and the mean speech discrimination score was 78.1% ± 28.0%.

Of the 63 patients, serum from 43 (68%) was WB positive and serum from 47 (75%) was IF positive. Serum from 37 patients (59%) was positive in both assays, in 10 it was negative in both, and in 16 results were discordant (positive in one and negative in the other) (Table 1).

Serum samples from University of Michigan hearing loss (UMHL) patients UMHL-6, UMHL-9, and UMHL-33 (Figure 1) are representative examples of serum samples that exhibited the typical wineglass staining pattern of the outer pillar cell phalangeal processes that is reminiscent of the KHRI-3 monoclonal antibody staining.⁸ These serum samples also stained a protein band migrating in the 68- to 72-kDa region, and all 3 patients responded to therapy (Table 1). Examples of serum samples with concordant and discordant results in the 2 assays are shown in Figure 2 and Figure 3. In part A of Figure 2 and lane A of Figure 3, results are shown for patient UMHL-80, who had rapidly progressive bilateral hearing loss and failed to respond to treatment (Table 1). This patient had very strong staining in both assays, which is consistent with a high level of circulating antibody. Part B (Figure 2) and lane B (Figure 3) show results with serum from patient UMHL-82, who had unilateral sudden-onset hearing loss. Her serum contained antibody to supporting cells, but little or no detectable antibody to the 68- to 72-kDa protein bands. She responded favorably to treatment (Table 1). Part C and lane C show results from patient UMHL-48, who had rapidly progressive bilateral hearing loss but did not respond to treatment. Her serum did not stain supporting cells but did stain a 68- to 72-kDa band on WB. This patient’s serum contained antibody to stereocilia (Figure 2C). Stereocilia antibodies are often present in normal mouse and human serum, as well as in patients’ serum, but their significance is unknown. Patient UMHL-54 had rapidly progressive bilateral hearing loss and responded to treatment (Table 1) but had no detectable antibody by either assay (part D of Figure 2 and lane D of Figure 3). None of the 20 normal serum samples stained supporting cells, and only 1 stained a 68- to 72-kDa protein on WB.

Hearing improvement was documented in 21 (49%) of 43 patients who were WB positive and 7 (35%) of 20 WB-negative patients (P = .30) (Table 2). Of the 28 patients who improved, 25 (89%) were positive for antibody to supporting cells by the IF assay and only 3 (11%) were IF negative. Of the 47 patients who were IF positive, 53% (25/47) had hearing improvement. In contrast, only 3 (19%) of the 16 patients who were IF negative improved (relative rate, 2.8) (P = .02) (Table 2). Patients whose acute-phase serum samples were IF positive were nearly 3 times more likely to experience improved hearing with corticosteroid treatment than patients whose serum samples were IF negative. Thus, antibody to supporting cells is significantly associated with hearing improvement after corticosteroid treatment. When results from both tests were combined, 19 (51%) of the 37 patients who were positive by both assays showed improvement, compared with 1 (10%) of 10 who were negative by both tests. The combined results of both tests also were significantly associated with response to therapy (P = .03).

All 8 patients with systemic autoimmune disease (Table 1) had antibody to inner-ear antigens, 6 (75%) by WB and 8 (100%) by IF. Of these 8 patients, 6 (75%) demonstrated hearing improvement after treatment.

Forty-nine patients were tested for erythrocyte sedimentation rate, rheumatoid factor, and antinuclear antibody; of these, 32 (65%) had results within normal limits and 17 (35%) had abnormal results. Of the 17 with abnormal test results, 11 (65%) had improved hearing and 6 (35%) did not; this was not significantly different from the group with normal values for erythrocyte sedimentation rate, rheumatoid factor, and antinuclear antibody, of whom 17 (53%) of 32 improved and 15 (47%) did not (P = .43). There was no association of antibody to inner-ear antigens by age group, sex, or ABO blood type (data not shown).

A somewhat higher proportion of patients with unilateral hearing loss (18/37 [49%]) showed improvement after corticosteroid treatment compared with 10 (38%) of the 26 with bilateral hearing loss. This observation is consistent with the earlier stage of disease in unilaterally affected patients who have not yet progressed to involvement of both ears and thus may have acquired less damage to the inner ear. It is of interest that, of the 8 patients with rapid-onset hearing loss (marked as “sudden” in (Table 1), all but 1 had antibody to inner-ear antigens, detected as either the 68- to 72-kDa band on WB (4 patients) or antibody to supporting cells (6 patients) or both. One of these patients (UMHL-28, (Table 1) also had Raynaud disease, a systemic autoimmune disease. Six of the 8 with very-rapid-onset hearing loss improved after corticosteroid treatment. Of these, 5 had antibody to supporting cells and 3 had antibody to the 68- to 72-kDa band on WB. There was no improvement in the single patient with sudden-onset hearing loss without antibody.

Antibodies to inner-ear antigens other than the 68- to 72-kDa protein have been described in serum from patients with AISNHL. Cao et al⁹ failed to detect the 68- to 72-kDa band, reporting instead bands of 58 and 30 kDa with guinea pig inner-ear extracts. Joliat et al¹⁰ also found antibodies to a 30-kDa collagen antigen. Boulassel et al¹¹ suggested that the cochlin protein may be the 58-kDa target of autoimmune antibodies in some patients with autoimmune inner-ear disease. We also noted that some patients show staining of a protein in the cochlin 60-kDa range, as in the case shown in Figure 3, lane A. Boulassel et al¹² also investigated other proteins stained by patients’ serum and identified a 30-kDa protein as myelin P0 protein, a 42-kDa protein as β-actin, and a 68-kDa protein as a novel protein.

This study focuses on the 68- to 72-kDa region because consistent reports and our own experiments demonstrated antibody to a 68-kDa protein in serum of patients with suspected AISNHL,^{2 - 4 ,8 ,13 - 14} and because similar antibodies were detected in animal models of antibody-mediated sensorineural hearing loss.^{5 ,15} Furthermore, 68- to 72-kDa bands are often prominent and are sometimes the only band(s) observed,^{4 ,8} and patients with antibodies to a band of 68 to 72 kDa are reported to be more likely to respond to corticosteroids.⁴ Also, patient serum with antibody to a 68- to 72-kDa band also stains protein precipitated by KHRI-3 antibody,⁸ which itself binds to supporting cells and can cause hearing loss in animal models.^{6 - 7} Normal serum rarely stains a 68-kDa protein,^{2 - 3} and we have not observed staining of supporting cells with serum from normal donors. Most of the 63 patients in the present study had antibodies to inner-ear antigens; 68% showed staining of a 68- to 72-kDa protein on WB and 75% had staining of inner-ear supporting cells with the wineglass staining pattern. Collectively, the evidence implicates a 68- to 72-kDa supporting cell antigen in most cases of AISNHL.

The clinical picture of AISNHL is far from being clearly defined. In the current study, we found antibodies to IESCA detected by IF in all 8 patients with systemic autoimmune disease. Of these, 6 had improved hearing after treatment. This finding recalls the original observation of McCabe¹ that autoimmune patients with hearing loss often exhibit improved hearing with immunosuppressive therapy. McCabe¹ included patients with systemic autoimmune disease and acknowledged that unilateral disease often progressed to affect both ears but excluded patients with sudden hearing loss that developed during periods of less than weeks. Similarly, others⁴ excluded both sudden and unilateral hearing loss from AISNHL, yet many patients who present with unilateral loss have progression to the contralateral ear later. Furthermore, we have within our patient populations patients who have developed sudden, sometimes severe, hearing loss in one ear and later progress to worse hearing first in that ear and then later exhibit a second episode of sudden hearing loss affecting the opposite ear. Thus, we elected to include these patients in our search for autoantibodies and corticosteroid-responsive disease because the definitions used by others have somewhat arbitrarily excluded these cases, which may represent an earlier and more treatable stage of disease.

In fact, our group’s previous study⁸ and current data show that patients with sudden onset and unilateral progressive hearing loss frequently have antibodies to inner-ear antigens. In the current study, 28 (75%) of 37 patients with unilateral hearing loss had a positive IF test result, which is the same proportion found in the bilateral cases (19 [73%] of 26). In addition, antibodies to IESCA are present in a comparable proportion of patients with sudden (all unilateral) hearing loss (6 [75%] of 8 patients) and rapidly progressive hearing loss (41 [75%] of 55). The average rate of hearing improvement among study subgroups was 51%. Since it is widely believed that 30% of patients with idiopathic, unilateral, or sudden-onset hearing loss are expected to improve spontaneously, we assessed the likelihood of improvement being associated with a positive IF test result in all subjects as well as in the bilateral group or the unilateral group separately. For all subjects, the odds ratio for improvement and a positive IF test result was 4.92, or, stated differently, patients who improve are nearly 5 times more likely to have antibodies to supporting cells. Although we lose statistical significance when we divide the groups into unilateral or bilateral hearing loss, the odds ratios for the association in each group were 4.67 and 5.4, respectively. Furthermore, for the unilateral group, the P value by Fisher exact test reached .07, indicating a very strong association that approached statistical significance. The striking similarity of odds ratios between the stratified groups strongly supports the association between IF positivity and hearing improvement in both groups. Finally, for the IF test in the combined groups, we observed a 53% improvement rate with 95% confidence intervals of 39% to 67%, indicating that the improvement rate was well above the one third expected by chance alone. Thus, our data indicate that many patients with sudden hearing loss (6 of 8 in our series) have an autoimmune cause that includes antibody to supporting cells as an important predictive factor of corticosteroid response. Thus, although rapidly progressive unilateral and sudden hearing loss (developing over hours or days) has been excluded from AISNHL,^{1 ,4} our laboratory and clinical evidence strongly supports the inclusion of sudden and unilateral hearing loss in AISNHL. In the group with sudden hearing loss, 7 of 8 patients had antibody to inner-ear antigens, detected either as the 68- to 72-kDa protein (1 patient) or as IESCA (3 patients) or both antigenic substrates (3 patients), and 75% (6/8) improved with corticosteroid treatment. Thus, we conclude that many patients with sudden-onset hearing loss have a corticosteroid-responsive autoimmune cause. These observations reemphasize the importance of developing a reliable means for identifying AISNHL and demonstrate the limitations of relying on clinical features alone.

Less than 20% of patients with a negative IF test result for IESCA showed improvement in hearing, whereas more than half of patients with a positive IF test result improved. Among the 28 patients with improved hearing after treatment, almost 90% had IESCA antibodies by IF. Thus, the IF test may have value for predicting patients most likely to improve with corticosteroid treatment. The WB test result alone was not predictive of response to treatment. Half of WB-positive and more than a third of WB-negative patients had some degree of recovery. It is not surprising that antibodies to IESCA correlated with treatment outcome better than the WB results. The WB test detects antibodies to 68- to 72-kDa proteins, of which there could be many in the guinea pig cochlear tissue. A positive WB result can reflect the presence of antibodies to irrelevant proteins of a similar weight, weakening specificity. In this regard, heat shock protein 70 (HSP70) also migrates in this region and was thought to be the target of autoimmune hearing loss antibodies.^{16 - 17} A positive WB for HSP70 was reported to predict corticosteroid response¹⁸ in patients with AISNHL, but many WB-negative patients also responded. More recent studies exclude HSP70 as a specific target of autoantibodies in AISHNL.^{19 - 21} In contrast to the findings for HSP70, the IF test for antibodies to the IESCA identifies a limited and possibly unique binding pattern on supporting cells⁸ and a cellular localization already linked to antibody-induced hearing loss in an animal model.^{7 ,22} In this study, 7 (25%) of the 28 patients who improved with corticosteroid treatment were negative by WB, but 6 (86%) of those 7 were IF positive, thus identifying a group of 6 responders without detectable antibodies to the 68- to 72-kDa band. This suggests that the antibody titers were adequate for detection by IF but not by WB. Some WB-negative serum samples become positive for a 68- to 72-kDa band when tested on concentrated inner-ear antigen precipitated with KHRI-3–coupled beads. This suggests that low-titer antibody may be missed on a standard WB assay.

Hearing improvement was achieved in half of the patients with inner-ear–reactive antibodies. In patients with high antibody titer, irreversible damage may have already occurred by the time treatment is started. Some of the patients with rapid onset and profound hearing deficits had very strong staining intensity consistent with high antibody titers. This subset of antibody-positive patients may not show improvement. However, prompt treatment in such cases may prevent additional damage. A recent study by Slattery et al²³ reported that only 35% of patients with idiopathic sudden hearing loss treated with corticosteroids exhibited a response. Thus, it appears that the presence of an IF-positive test result may identify patients most likely to benefit from corticosteroid treatment.

In contrast to our findings, Moscicki et al⁴ identified antibodies to a bovine inner-ear 68-kDa protein in 31 patients with bilateral rapidly progressive hearing loss among whom 27 (87%) responded to treatment. The differences between their findings and those of the present study may be related to several factors. A considerably higher dose of corticosteroids was used in the previous study, likely resulting in more intense immunosuppression. Alternatively, a higher proportion of the patients in their study may have been at a stage of the disease before irreversible damage occurred. It is also possible that there is a difference in the severity of hearing loss between their study and the present one, making the response to corticosteroid therapy more likely. These points can best be tested in larger cooperative studies with inclusion of more patients so that response as a function of degree of hearing loss can be assessed accurately.

For consistency, we used Moscicki and coworkers’ criteria⁴ to define active sensorineural hearing loss and hearing improvement. However, we have reservations regarding its clinical application and significance. A hearing threshold of 30 dB at a single frequency may have limited clinical significance, particularly at 250 or 8000 Hz. In addition, these criteria classify patients with hearing loss with a threshold elevation greater than 10 dB at just 1 frequency. It is theoretically possible to have cases where it is impossible for the patient to “recover” according to this criterion, since a greater than 10-dB improvement is required at 2 consecutive frequencies. Instead of judging loss or improvement at single frequencies, an alternative may be the pure-tone average or speech reception threshold to provide a broader picture of the patients’ hearing changes, although this would be less sensitive to more modest improvements.

The IESCA assay is not yet a tool for clinical testing. It is labor intensive, requires a large number of expensive animals, and is technically challenging. Thus, it remains at this time primarily a research method to investigate the mechanism of immune-mediated hearing loss. Second- or third-generation assays based on standardized antigenic substrates must be developed to make this a practical test for guiding clinical decision making and patient treatment. In this regard, our group recently identified the KHRI-3 antigen as choline transporter–like protein 2.²⁴ We cloned the complementary DNA and plan to express choline transporter–like protein 2 as a potential test for antibodies in patients’ serum. We also discovered several other proteins that are coprecipitated with choline transporter–like protein 2 that may make up a cell surface complex of proteins whose function may be blocked by antibody binding. Thus, antibodies to other proteins (see Figure 3) may also affect the function of this complex and be pathogenic to the organ of Corti.

Correspondence: Thomas E. Carey, PhD, Cell Biology and Immunology Laboratory, Department of Otolaryngology/Head and Neck Surgery, 6020 Kresge Hearing Research Institute, University of Michigan, 1301 E Ann St, Ann Arbor, MI 48109-0506 (careyte@umich.edu).

Submitted for Publication: April 27, 2004; final revision received February 11, 2005; accepted March 11, 2005.

Funding/Support: This research was supported in part by The Ruth and Lynn Townsend Family Fund for Autoimmune Sensorineural Hearing Research at the University of Michigan and the Deafness Research Foundation, McLean, Va; grants R01 DC02272 and R01 DC03686 from the National Institute on Deafness and Other Communication Disorders, National Institutes of Health, Bethesda, Md; University of Michigan Rheumatic Diseases Core Center grant P30 AR048310; and University of Michigan Kresge Hearing Research Core grant P30 DC05188 from the National Institutes of Health. Dr Zeitoun is the recipient of a TWJ Foundation fellowship.