Sensorineural Hearing Loss Cells in Patients with Autoimmune ...

The Journal of Immunology

Increased Frequencies of Cochlin-Specific T Cells in Patients with Autoimmune Sensorineural Hearing Loss1 Moo-Jin Baek,2*† Hyun-Min Park,3*† Justin M. Johnson,* Cengiz Z. Altuntas,*§ Daniel Jane-wit,* Ritika Jaini,* C. Arturo Solares,† Dawn M. Thomas,‡ Edward J. Ball,‡ Nahid G. Robertson,¶ Cynthia C. Morton,¶ Gordon B. Hughes,† and Vincent K. Tuohy4*§ Autoimmune sensorineural hearing loss (ASNHL) is the most common cause of sudden hearing loss in adults. Although autoimmune etiopathogenic events have long been suspected in ASNHL, inner ear-specific Ags capable of targeting T cell autoreactivity have not been identified in ASNHL. In this study, we show by ELISPOT analysis that compared with normal hearing age- and sex-matched control subjects, ASNHL patients have significantly higher frequencies of circulating T cells producing either IFN-␥ (p ⴝ 0.0001) or IL-5 (p ⴝ 0.03) in response to recombinant human cochlin, the most abundant inner ear protein. In some patients, cochlin responsiveness involved both CD4ⴙ and CD8ⴙ T cells whereas other patients showed cochlin responsiveness confined to CD8ⴙ T cells. ASNHL patients also showed significantly elevated cochlin-specific serum Ab titers compared with both normal hearing age- and sex-matched control subjects and patients with noise- and/or age-related hearing loss (p < 0.05 at all dilutions tested through 1/2048). Our study is the first to show T cell responsiveness to an inner ear-specific protein in ASNHL patients, and implicates cochlin as a prominent target Ag for mediating autoimmune inner ear inflammation and hearing loss. The Journal of Immunology, 2006, 177: 4203– 4210.

A

utoimmune sensorineural hearing loss (ASNHL)5 is characterized as a bilateral rapidly progressive adult hearing impairment that responds therapeutically to corticosteroids (Ref. 1 and Health Information on Sudden Deafness, National Institute on Deafness and Other Communication Disorders, 具www.nidcd.nih.gov/health/hearing/sudden.asp典). This treatment response distinguishes ASNHL from most other hearing disorders and implies a great potential for further contemporary medical intervention. Despite its name, there is a limited understanding of possible inner ear-specific Ags that may target autoreactivity in ASNHL. Although sera from ASNHL patients contain Abs capable of immunostaining human inner ear tissues (2– 4), the Ag specificity

*Department of Immunology, Lerner Research Institute, †Head and Neck Institute, and ‡Allogen Laboratories, Cleveland Clinic, Cleveland, OH 44195; §Department of Biology, Cleveland State University, Cleveland, OH 44115; and ¶Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115 Received for publication September 28, 2005. Accepted for publication June 26, 2006. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by grants from the Deafness Research Foundation (New York, NY); the Samuel Rosenthal Foundation and the Milton and Charlotte Kramer Foundation (Cleveland, OH); the Triple-T Foundation (Chardon, OH); and National Institutes of Health Grants DC-006422 (to V.K.T.) and DC-003402 (to C.C.M.). 2 Current address: Department of Otolaryngology-Head and Neck Surgery, Inje University, Pusan, Paik Hospital, Kaekum-Dong, Pusanjin-gu, Pusan City, South Korea. 3 Current address: Mirae ENT Clinic, 617 Shinsa Dong, Kangnam Gu, Seoul, South Korea. 4 Address correspondence and reprint requests to Dr. Vincent K. Tuohy, Department of Immunology, Cleveland Clinic, Lerner Research Institute, NB30, 9500 Euclid Avenue, Cleveland, OH 44195. E-mail address: [email protected] 5 Abbreviations used in this paper: ASNHL, autoimmune sensorineural hearing loss; HSP, heat shock protein; dB, decibel; WDS, word discrimination score; OHL, other hearing loss; PPD, purified protein derivative; PTA, pure tone average; PVDF, polyvinylidene difluoride.

Copyright © 2006 by The American Association of Immunologists, Inc.

of the autoantibodies is currently unclear. Heat shock protein 70 (HSP70) has been implicated as a possible target Ag in ASNHL (5–7) but its ability to target inner ear-specific pathology is questionable in light of its expression in a variety of noninner ear tissues and its failure to target hearing loss in immunized mice (8). Collagen II has also been proposed as a target Ag in ASNHL (9, 10). However, the ubiquitous expression of collagen II in numerous organs makes it an improbable candidate for targeting the inner ear-specific autoimmune features prevalent in patients with “primary” ASNHL who show no signs of additional inflammatory disorders or systemic autoimmunity. Most other putative autoimmune diseases are linked to targeted recognition of major differentiation proteins. For example, proinflammatory T cell responses to myelin proteolipid protein, myelin basic protein, and myelin oligodendrocyte glycoprotein have been associated extensively with multiple sclerosis (11), and high serum titers specific for the acetylcholine receptor have been prominently linked to the symptoms of myasthenia gravis (12). Thus, it is reasonable to consider that ASNHL may be due to autoimmune responses targeted against inner ear-specific differentiation proteins. One prominent candidate protein for targeting inner ear-specific autoimmunity in ASNHL is cochlin, primarily because it is the most abundant protein expressed in inner ear tissues (13) and because its expression in adult humans is substantially confined to the cochlear and vestibular labyrinth (14, 15). Cochlin is a product of the COCH gene mapped in humans to chromosome 14q12-q13 (16), and its mutations cause DFNA9, an autosomal dominant, nonsyndromic, progressive sensorineural hearing loss with vestibular pathology (14 –18). Cochlin is an integral part of the inner ear extracellular matrix and its expression is substantially confined to the fibrocyte regions of the spiral limbus and spiral ligament, inner ear regions showing histologic abnormalities in humans with COCH mutations (14, 15). Cochlin immunogenicity has already been implicated in hearing loss. Cochlin coprecipitates with the choline transporter-like target 0022-1767/06/$02.00

4204

INCREASED FREQUENCIES OF COCHLIN-SPECIFIC T CELLS IN ASNHL

Ag recognized by the KHRI-3 IgG1 mAb that induces hearing loss in animals (19, 20). Moreover, our prior work has shown that immunization of SWXJ mice with the p131–150 peptide derived from mouse cochlin results in a significant loss of hearing that may be passively transferred into naive recipients with peptide-activated CD4⫹ T cells (21, 22). Cochlin antigenicity has also been shown in ASNHL since many patients have elevated serum levels of cochlin Abs (23). Despite the ability of cochlin to target hearing loss in animals and the presence of serum cochlin Abs in ASNHL patients, there is currently no evidence implicating cochlin-specific T cell responsiveness in ASNHL. In this study, we show by ELISPOT analysis that ASNHL patients have significantly increased frequencies of circulating cochlin-specific IFN-␥-producing CD4⫹ and/or CD8⫹ T cells. In addition, our data indicate that ASNHL patients also have significantly higher frequencies of regulatory IL-5-producing T cells that may be targeted for expansion and immune-deviating treatment of ASNHL. The higher frequencies of cochlin-specific T cells in ASNHL were accompanied by significantly higher cochlin serum Ab titers in ASNHL patients compared with both normal hearing age- and sex-matched control subjects and patients with noise- and/or age-related hearing loss. Thus, our study implicates inner ear-specific autoreactivity in ASNHL and provides a rational basis for developing contemporary immunomodulatory strategies for treating ASNHL, a hearing disorder proven to be responsive to therapeutic intervention.

ramped up in a linear fashion to 95% over 45 min. Detection was performed with a Beckman System Gold 166 detection module reading at 280 nm. Peak cochlin elution occurred at an acetonitrile concentration of 61%. Purified cochlin fractions were collected and lyophilized overnight on a Savant ModulyoD-115 system (Thermo Electron). After purification by HPLC, the purified recombinant human cochlin was resuspended in double-distilled deionized H2O, and the final concentration was determined using a Bio-Rad Protein Assay.

Materials and Methods Production of recombinant human cochlin

Selection of control study subjects with other hearing loss (OHL) abnormalities

Human cochlin cDNA generated as previously described (17), was inserted into pQE82L (Qiagen) for producing a His-tagged fused protein. XL1-Blue Escherichia coli (Stratagene) were transformed and screened for expression with HRP-conjugated His Ab (Qiagen). High level expression colonies were selected and the plasmid was maxiprepped and sequenced for verifying proper orientation and alignment. His-tagged cochlin was purified under denaturing conditions on a Ni-NTA agarose column (Qiagen). A total of 10 ␮l of samples in denaturing SDS-PAGE buffer were loaded on a 15% Tris-HCL gel (Bio-Rad) and blotted onto Immobilon-P polyvinylidene difluoride (PVDF) membrane (Millipore) and stained with HRP-conjugated His Ab (Qiagen). Detection was performed with the ECL Western Blotting Analysis system (Amersham Biosciences) and exposure to Biomax MR film (Kodak). Molecular weight was determined by Kaleidoscope Prestained Standards (Bio-Rad). Cochlin purity was optimized by HPLC purification of the Ni-NTA product using a Beckman System Gold 126 solvent module (Beckman Coulter) with a Vydac C4 semipreparative column (Grace Vydac). Initial acetonitrile concentration was 5% and was

In addition to ASNHL subjects, we evaluated cochlin serum Ab titers in patients with bilateral sensorineural hearing loss not associated with rapidly progressive inner ear disease or with any immune or autoimmune disorder (Table II). These OHL control study subjects included patients with noise-induced hearing loss and patients with presbycusis (age-related hearing loss). Inclusion criteria were defined as patients with normal hearing in both ears early in life, recent gradual progressive hearing loss equally in both ears, ⬎60 years of age for presbycusis, ⬎40 years of age for noise-induced hearing loss, normal otoscopic examination with audiogram showing bilateral symmetric sensorineural hearing loss unless mild asymmetry is due to noise exposure greater to one ear. Exclusion criteria were defined as patients with a family history of genetic hearing loss, ototoxic medication, corticosteroids taken in recent 3 mo for any reason or for hearing loss at any time previously, rapid progression of hearing loss occurring within 3 mo or less, history of systemic immune disease, prior or concomitant diagnosis of other ear disease, abnormal otoscopy, or audiogram showing conductive or mixed hearing loss in

Selection of ASNHL study subjects All protocols were reviewed and approved by the institutional review board of the Cleveland Clinic Foundation, and all study subjects provided informed consent before inclusion in the study. Main entry criteria for selection of ASNHL study subjects were those defined recently by Harris et al. (1). Inclusion criteria were defined as patients with bilateral sensorineural hearing loss of at least 30 decibels (dB), and progression of loss in at least one ear within 3 mo as measured by a 10 dB pure tone worsening over three consecutive test frequencies and/or a 16% decrease in word discrimination score (WDS) using a 25-word list. Exclusion criteria were defined as patients with congenital or genetic disease or patients with acquired otologic disease other than idiopathic or sudden hearing loss as well as patients with any additional autoimmune abnormalities as determined by clinical and medicinal history. Thus, our ASNHL study subjects represent a patient population believed to have an organ-specific “primary” autoimmune disorder involving only inner ear abnormalities (Table I). Although our study and a previously published national study (1) did not restrict participation to any ethnic group(s), all patients in our study and 90% in the national study were Caucasian. The basis for this high Caucasian frequency is currently unclear. Normal hearing ageand sex-matched control study subjects were selected based on no prior clinical or medicinal history of either hearing loss or other immune abnormalities (Table I).

Table I. ASNHL patients and age- and sex-matched normal hearing control study subjects Hearing Lossb (dB PTA) ASNHL Patients

P1f P2f P3 P4 P5 P6 P7 P8 a

Hearing Drop Withinc

Disease Durationd

Steroid Responsee

Sex/Age

68-kDa Aba

Right

Left

Right

Left

Right

Left

Right

Left

Control Subjects

Sex/Age

F/73 F/69 F/53 F/54 F/51 M/41 F/57 F/49

Negative Negative Negative Negative Negative Negative Negative ND

48.8 52.5 48.8 25 43.8 38 60 60

33.8 65 45 81 26.3 NR 22 68

3 mo SP 1 mo 1 wk 1 mo 2 mo 3 mo 1 mo

1 mo 1 wk 3 mo 1 wk 1 mo SP 1 mo SP

4 year 19 year 1 year 1 mo 5 mo 6 year 15 year 12 year

2 mo 2 mo 2 year 5 year 1 year 39 year 1 mo 22 year

Negative Negative ⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹ Negative ⫹

Negative Negative ⫹ ⫹⫹ ⫹⫹⫹ Negative ⫹⫹⫹ ⫹

C1 C2 C3 C4 C5 C6 C7 C8

F/73 F/69 F/52 F/54 F/51 M/41 F/57 F/49

A 68-kDa Ab represents anti-HSP70. Hearing loss indicates the degree of each patient’s dB hearing level determined as the PTA of 500, 1000, and 2000 Hz at the time of ELISPOT testing. NR indicates no response. c Time over which hearing loss occurred in weeks, months, or slowly progressive (SP). d Disease duration indicates the period between onset of hearing loss and ELISPOT testing in weeks, months, or years. e Degree of hearing improvement measured by dB PTA after steroid therapy: ⫺ none; ⫹ 10 –20 dB increment; ⫹⫹ 20 –30 dB; ⫹⫹⫹ restored normal hearing. f Two (25%) of eight of our ASNHL patients failed to show corticosteroid responsiveness similar to the 20% failure rate observed by Harris et al. (1). b


4205

Table II. Patients with OHL abnormalities

Patient

Age/Sex

History of Noise

Hearing Lossa (dB PTA)

WDS (%)b

Most Likely Cause of Hearing Lossc

OHL1 OHL2 OHL3 OHL4 OHL5 OHL6 OHL7 OHL8

46/M 53/F 55/F 56/M 57M 62M 67M 65M

Yes Yes Yes Yes Yes Yes Yes No

12 12 20 10 20 46 NAd 32

96 96 100 96 96 84 NAd 90

Noise Noise Noise Noise Noise Noise and age Noise and age Agee

a Hearing loss indicates the degree of each patient’s dB hearing level determined as the PTA of 500, 1000, and 2000 Hz at the time of phlebotomy. b The WDS measures the percentage of words that can be correctly identified when presented at a comfortable loudness level. One PTA and one WDS score is given for each patient because a symmetric hearing loss equal in both ears was required for each control, i.e., the hearing impairment is the same for both ears. c Usually a PTA ⬎ 20 is considered impaired in an adult. The reason three of our OHL controls have PTA ⬍ 20 dB is because both noise and age affect primarily the higher frequencies above 2000 Hz; the subjective impairment would not be evident in the PTA measurement but might still bother the patient enough to come in for testing. There is no reporting convention that averages test results in only the higher frequencies above the speech range. The PTA is the only reporting convention although it sometimes includes 3000 Hz as well. Generally, a WDS better than 80% is considered normal. Very often WDS is preserved despite high frequency hearing loss from noise, age, or both. This explains the apparent normal WDS scores as well. Once the audiogram review shows the high-tone loss objectively, the diagnosis of noise-induced loss is based solely on the patient history (i.e., military service, recreational noise, or work-related noise of significant duration and/or intensity). d NA means audiogram not available. A retrospective review of patient charts did not identify any record of audiogram and the diagnosis was based on written records of the audiologist and clinician. e Age-related loss was based on a history that did not suggest an alternative cause of hearing loss and the patient’s age was 60 or more.

one or both ears or substantial asymmetry not explained by noise exposure greater to one ear.

ELISPOT assays ELISPOT assays were performed as previously described (24). Briefly, PBMC were separated by centrifugation on Ficoll-Paque (Amersham Biosciences) and the leukocyte fraction was washed and resuspended at 3 ⫻ 106 cells/ml HL-1 medium (Cambrex) supplemented with L-glutamine, penicillin, streptomycin, HEPES buffer (Invitrogen Life Technologies), and 5% autologous serum. Cells were added at 3 ⫻ 105 cells/well in ELISPOT plates (Polyfiltronics) in wells precoated with IFN-␥ capture Ab (no. M-700A; Endogen) or IL-5 capture Ab (no. 18551D; BD Biosciences). Recombinant human cochlin was added to wells at a final concentration of 100 ␮g/ml in a final volume of 200 ␮l. Positive control wells contained 10 ␮g/ml human CD3 mAb (BD Biosciences), a dose that has reliably provided us with distinguishable nonconfluent spots. Negative control wells contained no Ag. Ag-specificity control wells contained a 1/1000 dilution of tuberculin-purified protein derivative (PPD; Evans Medical). All cochlin and control assays were performed in triplicate wells. After 24 (IFN-␥) or 48 (IL-5) h of culture at 37°C in humidified air containing 5% CO2, cells were removed, and wells were treated with secondary biotinylated mouse anti-human IFN-␥ (no. M701; Endogen) or mouse anti-human IL-5 (no. 18522D; BD Biosciences). After overnight incubation and washing, spots were visualized by sequential treatment with peroxidase-conjugated streptavidin (DakoCytomation), filtered 3-amino-9-ethylcarbazole substrate, and a 1/2000 dilution of 30% H2O2. The reaction was halted after 10 min by repeated washing with doubledistilled deionized H2O.

ELISPOT analysis ELISPOTs were detected using an automated Series-1 Immunospot Satellite Analyzer (Cellular Technology) with proprietary software designed to distinguish real spots from artifact. Digitized images of the wells were analyzed for detecting concentrated red color spots in which the spot density exceeds background by a factor individually calculated for each plate based on the appearance of spots in positive and negative control wells. Parameters for automated spot counting have been established to distinguish contiguous and overlapping spots, and spot size criteria and circularity were used to exclude noise caused by nonspecific Ab binding.

Purification of human CD4⫹ and CD8⫹ T cells PBMC from ASNHL patients were enriched to ⬃90% CD4⫹ T cells by negative selection following treatment with anti-CD8 microbeads and double passage through a MidiMACS magnetic column (Miltenyi Biotec). CD8⫹ T cells were similarly enriched by double passage negative selection of anti-CD4 microbead-treated PBMC. Purity of the negatively selected T cell subpopulations was determined by flow cytometry analysis using CD4FITC and CD8-PerCP-conjugated Abs (BD Biosciences). The negatively selected enriched populations were tested in ELISPOT assays as described above.

Cochlin Ab titers Sera from ASNHL and control subjects were tested by direct ELISA for cochlin Ab titers. Recombinant human cochlin was plated at 10 ␮g/ml on 96-well Nunc-immuno plates, MaxiSorp (Nalge Nunc International), and sera were added in duplicate wells at dilutions ranging from 1/4 through 1/2048. Presence of bound Ab was determined using a peroxidase-goat anti-human IgG H and L chain (Zymed Laboratories) followed by sequential treatment with ABTS substrate and H2O2 (Sigma-Aldrich). PBS was used as a control substitute for the secondary detection Ab. The reaction was stopped after 30 min by adding SDS/dimethylformamide, and absorbance at 405 nm was measured using a Wallac 1420 VICTOR2 Multilabel ELISA reader (PerkinElmer).

HLA typing Low- to intermediate-level HLA class I and class II typing was accomplished by sequence-specific oligonucleotide probing using commercial kits (LABType, One ␭). Briefly, purified DNA was amplified using locusspecific biotin-coupled primers followed by hybridization with motifspecific oligonucleotides coupled to flow cytometry beads. Specific hybridization to arrays of motif-specific beads was determined by flow cytometry (Luminex 100; Luminex). HLA assignment was made on the basis of hybridization patterns using a program provided by the kit manufacturer. Higher resolution HLA-DRB1* typing was accomplished by direct DNA sequencing of PCR-amplified products using primers and conditions as previously described (25, 26). Sequencing of PCR products was performed using DNA polymerase and base-specific fluorescence-labeled dideoxynucleotide termination reagents (dye terminators; Applied Biosystems). Analysis was conducted by capillary gel electrophoresis (AB 3730 DNA Analyzer; Applied Biosystems) and obtained sequences were compared with known alleles for HLA assignment (Table III).

4206


Table III. Frequencies of HLA class I and class II types in ASNHL patients Patient

A*

A*

B*

B*

P1 P2 P3b P4 P5 P6 P7 P8

02 02 01 01 23 03 02 02

03

27 07 08 08 07 35 37 13

51 51 51 13 1516 44 44 18

24 24 33 29 25

Cw*

Cw*

DRB1*

DRB1*

DQB1*

DQB1*

DPB1*

DPB1*

01 07 02 06 08 04 05 06

02 14 07 07 14 16 06 12

0101 1301 03 0301 1503 0701 0407 1501

0401 1501 07 0701

0301/09 0602/19a ND 02 0602/19 02 0301/09/13 02

0501 0603a ND

0301 0401 0301c 0201 0101 0101d 0301e 0201

0401 0402 0401c

1302 0701 0701

0604 0303/12 0603

1801 0402d 0401e 5901

a

Rare allele combinations of DQB1*0611,0614 not excluded. b HLA-DRB1* typing for P3 was performed at low resolution due to sample limitation. c Rare allele combinations of DPB1*0502,7801; *7201/9901; *2301,9201; *5601,8001 not excluded. d Rare allele combinations of DPB1*7501,8901 not excluded. e Rare allele combinations of DPB1*2301,9201 not excluded.

Statistical analysis The unpaired Student t test was used to analyze differences in ELISPOT frequencies and Ab titers between ASNHL and control study subjects.

Results Production of recombinant human cochlin We have previously shown that ASNHL subjects have increased frequencies of IFN-␥-producing T cells capable of responding to human inner ear homogenate (24). Despite its ubiquitous use in ASNHL studies, inner ear homogenate is essentially an uncharacterized assortment of a variety of Ags, each at undefined and different concentrations. In addition, homogenates contain factors that are not at all specific to the inner ear making interpretation of tissue-specific responsiveness quite tenuous. To address these concerns directly, we generated E. coli derived recombinant human cochlin from human cDNA (Ref. 17; Fig. 1). Increased frequencies of cochlin-responsive T cells in ASNHL We next used our recombinant human cochlin to determine frequencies of IFN-␥-producing and IL-5-producing T cells in PBMC from several ASNHL patients and from age- and sex-matched control subjects with no history of hearing loss (Table I). We focused our analysis on these two cytokines because IFN-␥ represents a proinflammatory Th1-type cytokine while IL-5 represents an antiinflammatory or regulatory Th2-like cytokine produced predominantly by T cells. Thus, the relative frequencies of T cells expressing each of these cytokines indicate the predominance of inflammatory vs regulatory autoreactivity. We found that ASNHL patients had substantially higher frequencies of cochlin-specific IFN-␥-secreting T cells in their PBMC when compared with ageand sex-matched control subjects (Fig. 2, A and B). ASNHL patients showed a mean of 472 ⫾ 64SE/1 ⫻ 106 PBMC cochlinspecific IFN-␥-producing T cells compared with 118 ⫾ 64SE/1 ⫻ 106 PBMC in control subjects. The difference in frequencies of cochlin-specific IFN-␥-secreting T cells between eight ASNHL patients (mean age 55.88 ⫾ 10.49 SD) and eight normal control subjects (mean age 55.75 years ⫾ 10.54 SD) was highly significant ( p ⫽ 0.0001; Fig. 3A), and all ASNHL patients showed IFN-␥ T cell frequencies higher than the highest frequencies observed in any of the control subjects. Frequencies of cochlin-specific IL-5-secreting T cells in ASNHL patients vs normal controls were also significantly higher ( p ⫽ 0.03) but were substantially much lower than frequencies of IFN-␥-producing T cells (Figs. 2, C and D; Fig. 3D). ASNHL patients showed a mean of 69 ⫾ 25SE/1 ⫻ 106 PBMC cochlinspecific IL-5-producing T cells compared with 8.4 ⫾ 2.3SE/1 ⫻ 106 PBMC in control subjects. Nonspecific immune responsive-

ness was not significantly different between ASNHL patients and normal control subjects as determined by frequencies of IFN-␥secreting T cells ( p ⫽ 0.14) and IL-5-secreting T cells ( p ⫽ 0.48) in response to activation with 10 ␮g/ml human anti-CD3 (Fig. 3, B and E) and by frequencies of IFN-␥-secreting T cells ( p ⫽ 0.83) and IL-5-secreting T cells ( p ⫽ 0.31) in response to activation with a 1/1000 dilution of the irrelevant Ag, tuberculin PPD (Fig. 3, C and F). CD4⫹ and CD8⫹ T cells from ASNHL patients respond to cochlin We next examined whether IFN-␥ responsiveness to cochlin was mediated by CD4⫹ and/or CD8⫹ T cells. PBMC from ASNHL patients were incubated with either anti-CD8 or anti-CD4 microbeads (Miltenyi Biotec) and passed twice through a MACS LS magnetic column. Flow cytometry analysis showed that the negatively selected CD3⫹ T cell populations were consistently enriched to ⬃90% CD4⫹ T cells (in anti-CD8-depleted PBMC) or CD8⫹ T cells (in anti-CD4-depleted PBMC; Fig. 4A). ELISPOT analysis showed that cochlin-induced IFN-␥ production occurred

FIGURE 1. Production of purified recombinant human cochlin. Histagged human cochlin from transformed E. coli was purified on a Ni-NTA agarose column, loaded onto a 15% Tris-HCL gel in SDS-PAGE buffer, blotted onto a PVDF membrane, and stained with HRP-conjugated His Ab. Lane 1 of Western blot shows crude culture sample, induced with isopropyl ␤-D-thiogalactoside and allowed to express for 4.5 h at 37°C with shaking. Lane 2 shows purified product corresponding to the predicted molecular mass of His-tagged human cochlin.


4207

FIGURE 2. Increased frequencies of cochlin responsive IFN-␥- and IL-5-secreting T cells in ASNHL patients. A, PBMC from eight ASNHL patients were tested by ELISPOT for frequencies of IFN-␥-secreting T cells in response to 100 ␮g/ml recombinant human cochlin. ASNHL patients showed highly significant (p ⫽ 0.0001) increased frequencies compared with (B) age- and sex-matched control subjects. Frequencies of IL-5-producing T cells were also significantly higher (p ⫽ 0.03) in (C) ASNHL patients compared with (D) controls. As described previously (24), IFN-␥ ELISPOTS were determined at 24 h whereas IL-5 spots were determined at 48 h. n.d., Not determined. Error bars indicate ⫾SE.

in both CD4⫹ and CD8⫹ T cell populations in some patients (Fig. 4B; ASNHL no. P6) but was confined in other patients to the CD8⫹ T cell population (Fig. 4B; ASNHL no. P7). ASNHL patients have elevated cochlin Ab titers Sera from ASNHL patients, from patients with noise- and/or agerelated hearing loss (Table II), and from normal hearing age- and sex-matched control subjects were tested by direct ELISA for binding to recombinant human cochlin. At all serum dilutions from 1/32 through 1/2048, ASNHL patients showed significantly elevated titers to cochlin when compared with control OHL subjects with noise and/or age-related hearing loss or to normal hearing control subjects (Fig. 5). Surprisingly, the cochlin serum Ab titers

of OHL subjects were also significantly higher than those of ageand sex-matched controls. Differences in titers between all three study groups were significant at all dilutions tested from 1/32 to 1/2048 with p values ranging from p ⬍ 0.05 to p ⬍ 0.0005. HLA typing of ASNHL patients Class I HLA-B*13 and CW*06 and class II DRB1*07 types were found, respectively, in two of eight (25%), three of eight (37.5%), and five of eight (62.5%) of our ASNHL patients (Table III), but are found, respectively, in only 5.4, 13.6, and 25.5% of the normal American Caucasian population (Review Population Study, USA Caucasian American Bethesda, 具www.allelefrequencies.net典). These differences did not reach statistical significance and the study of additional

FIGURE 3. PBMC from ASNHL patients have significantly increased frequencies of cochlin-responsive T cells. A, The difference in frequencies of cochlin-specific IFN-␥-secreting T cells was highly significant (p ⫽ 0.0001) between ASNHL and control subjects. D, The difference in frequencies of cochlin-specific IL-5-producing T cells was also significant (p ⫽ 0.03). Differences in nonspecific responses were not significant as determined by measuring (B) IFN-␥ (p ⫽ 0.14) or (E) IL-5 (p ⫽ 0.48) ELISPOT frequencies in response to antiCD3 or by measuring (C) IFN-␥ (p ⫽ 0.83) or (F) IL-5 (p ⫽ 0.31) T cell frequencies in recall responses to PPD. Error bars indicate ⫾SE.

4208


FIGURE 4. Cochlin responses in ASNHL patients is mediated by CD4⫹ and/or CD8⫹ T cells. A, Upper, Untreated PBMC from ASNHL patients (solid line) were enriched to 91% CD4⫹ T cells (broken line) following treatment with anti-CD8 microbeads and double passage through a magnetic column. A, Lower, Untreated PBMC from ASNHL patients (solid line) were similarly enriched for CD8⫹ T cells (broken line) by double passage negative selection of anti-CD4 microbead-treated PBMC. B, Some ASNHL patients showed responsiveness to cochlin in both CD4⫹ and CD8⫹ T cells (patient no. P6) whereas others showed responses confined to CD8⫹ T cells (patient no. P7). Each response pattern is representative of three patients showing the CD4/CD8 dual responsiveness and two patients showing responsiveness confined to CD8⫹ T cells. Error bars indicate ⫾SE.

patients is needed. Higher frequencies of these HLA types in ASNHL subjects were not observed in studies by other groups (27–33).

Discussion Our data implicate cochlin-specific IFN-␥-producing T cells in the etiopathogenesis of ASNHL and support the view that cochlin, as the most abundant protein of the inner ear, serves as a prominent candidate Ag for targeting inner ear inflammation and autoimmune-mediated hearing loss. Our data also confirm the presence of cochlin-specific IgG Ab in ASNHL and suggest that both ELIS-

FIGURE 5. Sera from ASNHL patients show elevated cochlin Ab titers compared with control subjects. Sera from ASNHL and control subjects were tested by direct ELISA for cochlin Ab titers. Significant differences with p values ranging from p ⬍ 0.05 to p ⬍ 0.0005 occurred at all mean titers shown between ASNHL and each control study group as well as between the OHL patients and normal hearing control subjects. Error bars show ⫾SE.

POT determination for IFN-␥-producing T cells and cochlin Ab titers may be useful as diagnostic support assays for ASNHL. Current diagnostic support for ASNHL focuses on detection of Abs to the HSP70 heat shock protein. In light of the lack of inner ear specificity of HSP70 and the inability of HSP70 to mediate inner ear pathology in animals, the appearance of such elevated serum Abs may represent an epiphenomenon that accompanies ASNHL symptoms rather than a causative event that initiates or mediates ASNHL. In any case, the potential for using cochlin ELISPOT and Ab responses for ASNHL diagnostic support remains appealing particularly since all of the ASNHL patients in our study had higher frequencies of IFN-␥-producing T cells and higher Ab titers than the highest control subject tested. It remains to be determined whether cochlin autoreactivity assays may be useful in discerning ASNHL from other hearing disorders including age-related hearing deficiency and noise-induced hearing loss. Our data also show that CD4⫹ and CD8⫹ T cells are involved in cochlin recognition and in some cases only CD8⫹ T cell recognition of cochlin was evident. This implies that CD8⫹ T cells may be sufficient to account for inner ear autoimmunity. Indeed, several studies have implicated higher frequencies of a variety of HLA class I alleles in ASNHL including A2, A25, B18, B35, B39(B16), B41, Bw16, Cw4, Cw5, and Cw7 (27–30). The increased frequencies of numerous different class I alleles in ASNHL implies that multiple proteins and their derived peptides may be involved in inner ear-specific self-recognition. Although none of the above-mentioned HLA class I Ags was overrepresented in our study, we did find that B*13 and Cw*06 were found, respectively, in two of eight (25%) and three of eight (37.5%) of our ASNHL patients, compared with 5.4 and 13.6%, respectively, in the American Caucasian population (具www.allelefrequencies.net典). These increases did not reach statistical significance and should be confirmed by the study of additional patient populations. Overrepresentation of numerous class II alleles has also been implicated in ASNHL including increased frequencies of DRB1*03,

The Journal of Immunology DRB1*14, DRB3*01, DQB1*02, and DPB1*04 and decreased frequencies of DQB1*03 (31–33). Although none of these HLA class II types was overrepresented in our study, we did find that five of eight (62.5%) of our ASNHL patients were DRB1*07 (DRB1*0701) for those typed at high resolution, compared with a normal American Caucasian frequency of 25.5% (具www.allelefrequencies.net典). This difference did not reach statistical significance and should be confirmed by additional studies. The differences in HLA frequencies in ASNHL patients compared with normals suggested by various studies may be due to several factors, including differences in different ethnic or racial groups (German (27), British (28), Polish (29, 30), Belgian (31), Korean (32, 33)). Alternatively, the differences may reflect inherent diversity of restriction elements involved in ASNHL and a fundamental promiscuity of inner ear epitopes used in different ASNHL patient populations. Corticosteroid treatment is currently the only proven effective therapy in ASNHL and responsiveness to corticosteroids has long served as a convincing argument for viewing ASNHL as an autoimmune disorder. However, corticosteroid treatment inherently has limited long-term benefit and its systemic side effects are fundamentally intolerable. Thus, a treatment approach combining corticosteroids with other immunomodulatory agents may prove effective in preventing progression of ASNHL as has been shown in other T cell-implicated autoimmune disorders (34). It is evident from our study that ASNHL patients have increased frequencies of cochlin-specific IFN-␥-producing T cells, but also have increased numbers of cochlin-responsive IL-5-producing T cells. Although the absolute numbers of these regulatory Th2-like cells are much lower than their proinflammatory counterparts, their increased presence in ASNHL provides a basis for enhancing or complementing this native regulatory response. Several contemporary treatment strategies for autoimmune disease may be useful in this regard including treatment with IFN-␤, statins, as well as targeted inhibition of TNF, and Ag-based therapies including those involving altered variants of targeted self-peptides or TCR vaccination (35). Our data indicating increased cochlin Ab titers in ASNHL patients confirm those of Boulassel et al. (23). However, we also found that OHL control subjects with noise- and/or age related hearing loss have significantly higher cochlin Ab titers compared with normal hearing control subjects. The observed intermediate cochlin autoantibody titers in the OHL cohort implies that an adaptive inner ear autoimmune response may be implicated in noiseand/or age-related hearing loss. Indeed, elevated serum Abs against other autoantigens, most notably HSP60 and HSP70, have been reported in patients with noise-induced hearing loss (36, 37). Although it is appealing to consider that noise-induced inflammation may trigger pathogenic inner ear self-recognition events that lead to ASNHL, a direct link between acoustic trauma and the development of ASNHL has not been shown. Thus, it remains unclear whether the intermediate cochlin autoantibody levels in OHL patients have any pathogenic autoimmune implications. In summary, our study shows that cochlin autoreactivity is significantly enhanced in patients with ASNHL. This cochlin responsiveness involves IFN-␥-producing CD4⫹ and CD8⫹ T cells as well as elevated cochlin-specific serum Abs. Detection of cochlin autoreactivity may ultimately prove to be diagnostically supportive and the low level but significant production of cochlin-specific IL-5-producing regulatory T cells may serve as a platform for the development of contemporary immunomodulatory adjunct therapies for one of the few hearing disorders shown to be responsive to therapeutic intervention.

4209

Disclosures The authors have no financial conflict of interest.

References 1. Harris, J. P., M. H. Weisman, J. M. Derebery, M. A. Espeland, B. J. Gantz, A. J. Gulya, P. E. Hammerschlag, M. Hannley, G. B. Hughes, R. Moscicki, et al. 2003. Treatment of corticosteroid-responsive autoimmune inner ear disease with methotrexate: a randomized controlled trial. J. Am. Med. Assoc. 290: 1875–1883. 2. Arnold, W., R. Pfaltz, and H. J. Altermatt. 1985. Evidence of serum antibodies against inner ear tissues in the blood of patients with certain sensorineural hearing disorders. Acta Otolaryngol. 99: 437– 444. 3. Arnold, W., and C. R. Pfaltz. 1987. Critical evaluation of the immunofluorescence microscopic test for identification of serum antibodies against human inner ear tissue. Acta Otolaryngol. 103: 373–378. 4. Soliman, A. M. 1989. Experimental autoimmune inner ear disease. Laryngoscope 99: 188 –193. 5. Billings, P. B., E. M. Keithley, and J. P. Harris. 1995. Evidence linking the 68 kilodalton antigen identified in progressive sensorineural hearing loss patient sera with heat shock protein 70. Ann. Otol. Rhinol. Laryngol. 104: 181–188. 6. Bloch, D. B., J. E. San Martin, S. D. Rauch, R. A. Moscicki, and K. J. Bloch. 1995. Serum antibodies to heat shock protein 70 in sensorineural hearing loss. Arch. Otolaryngol. Head Neck Surg. 121: 1167–1171. 7. Billings, P. B., S. O. Shin, and J. P. Harris. 1998. Assessing the role of anti-hsp70 in cochlear impairment. Hear. Res. 126: 210 –213. 8. Trune, D. R., J. B. Kempton, C. R. Mitchell, and S. H. Hefeneider. 1998. Failure of elevated heat shock protein 70 antibodies to alter cochlear function in mice. Hear. Res. 116: 65–70. 9. Yoo, T. J., K. Tomoda, J. M. Stuart, T. A. Cremer, A. S. Townes, and A. H. Kang. 1983. Type II collagen-induced autoimmune sensorineural hearing loss and vestibular dysfunction in rats. Ann. Otol. Rhinol. Laryngol. 92: 267–271. 10. Yoo, T. J., Y. Yazawa, K. Tomoda, and R. Floyd. 1983. Type II collagen-induced autoimmune endolymphatic hydrops in guinea pig. Science 222: 65– 67. 11. Steinman, L. 1996. Multiple sclerosis: a coordinated immunological attack against myelin in the central nervous system. Cell 85: 299 –302. 12. Vincent, A. 2002. Unravelling the pathogenesis of myasthenia gravis. Nat. Rev. Immunol. 2: 797– 804. 13. Ikezono, T., A. Omori, S. Ichinose, R. Pawankar, A. Watanabe, and T. Yagi. 2001. Identification of the protein product of the Coch gene (hereditary deafness gene) as the major component of bovine inner ear protein. Biochim. Biophys. Acta 1535: 258 –265. 14. Robertson, N. G., B. L. Resendes, J. S. Lin, C. Lee, J. C. Aster, J. C. Adams, and C. C. Morton. 2001. Inner ear localization of mRNA and protein products of COCH, mutated in the sensorineural deafness and vestibular disorder, DFNA9. Hum. Mol. Genet. 10: 2493–2500. 15. Grabski, R., T. Szul, T. Sasaki, R. Timpl, R. Mayne, B. Hicks, and E. Sztul. 2003. Mutations in COCH that result in non-syndromic autosomal dominant deafness (DFNA9) affect matrix deposition of cochlin. Hum. Genet. 113: 406 – 416. 16. Manolis, E. N., N. Yandavi, J. B. Nadol, Jr., R. D. Eavey, M. McKenna, S. Rosenbaum, U. Khetarpal, C. Halpin, S. N. Merchant, G. M. Duyk, et al. 1996. A gene for non-syndromic autosomal dominant progressive postlingual sensorineural hearing loss maps to chromosome 14q12–13. Hum. Mol. Genet. 5: 1047–1050. 17. Robertson, N. G., A. B. Skvorak, Y. Yin, Y., S. Weremowicz, K. R. Johnson, K. A. Kovatch, J. F. Battey, F. R. Bieber, and C. C. Morton. 1997. Mapping and characterization of a novel cochlear gene in human and in mouse: a positional candidate gene for a deafness disorder, DFNA9. Genomics 46: 345–354. 18. Robertson, N. G., L. Lu, S. Heller, S. N. Merchant, R. D. Eavey, M. McKenna, J. B. Nadol, Jr., R. T. Miyamoto, F. H. Linthicum, Jr., J. F. Lubianca Neto, et al. 1998. Mutations in a novel cochlear gene cause DFNA9, a human nonsyndromic deafness with vestibular dysfunction. Nat. Genet. 20: 299 –303. 19. Nair, T. S., D. M. Prieskorn, J. M., Miller, D. F. Dolan, Y. Raphael, and T. E. Carey. 1999. KHRI-3 monoclonal antibody-induced damage to the inner ear: antibody staining of nascent scars. Hear. Res. 129: 50 – 60. 20. Nair, T. S., K. E. Kozma, N. L. Hoefling, P. K. Kommareddi, Y. Ueda, T. W. Gong, M. I. Lomax, C. D. Lansford, S. A. Telian, B. Satar, et al. 2004. Identification and characterization of choline transporter-like protein 2, an inner ear glycoprotein of 68 and 72 kDa that is the target of antibody-induced hearing loss. J. Neurosci. 24: 1772–1779. 21. Solares, C. A., A. E. Edling, J. M. Johnson, M.-J. Baek, K. Hirose, G. B. Hughes, and V. K. Tuohy. 2004. Murine autoimmune hearing loss mediated by CD4⫹ T cells specific for inner ear peptides. J. Clin. Invest. 113: 1210 –1217. 22. Billings, P. 2004. Experimental autoimmune hearing loss. J. Clin. Invest. 113: 1114 –1117. 23. Boulassel, M. R., J. P. Tomasi, N. Deggouj, and M. Gersdorff. 2001. COCH5B2 is a target antigen of anti-inner ear antibodies in autoimmune inner ear diseases. Otol. Neurotol. 22: 614 – 618. 24. Lorenz, R. R., C. A. Solares, P. Williams, J. Sikora, C. M. Pelfrey, G. B. Hughes, and V. K. Tuohy. 2002. Interferon-␥ production to inner ear antigens by T cells from patients with autoimmune sensorineural hearing loss. J. Neuroimmunol. 130: 173–178. 25. Kotsch, K., J. Wehling, and R. Blasczyk. 1999. Sequencing of HLA class II genes based on the conserved diversity of the non-coding regions: sequencing based typing of HLA-DRB genes. Tissue Antigens 53: 486 – 497. 26. Paul, P., J. Apgar, and E. J. Ball. 2003. HLA-DRB1* intron-primed sequencing for haploid genotyping. Clin. Chem. 49: 692– 694.

4210


27. Gross, M., and A. Arndt-Hanser. 1982. HLA-antigens and sensorineural deafness. Laryngol. Rhinol. Otol. 61: 316 –318. 28. Bowman, C. A., and R. A. Nelson. 1987. Human leukocytic antigens in autoimmune sensorineural hearing loss. Laryngoscope 97: 7–9. 29. Flasza, J., G. Turowski, and E. Reron. 1997. Class I HLA antigens in patients with sensorineural hearing loss: preliminary report. Otolaryngol. Pol. 51: 216 –222. 30. Kedzierska, A., J. Flasza, G. Turowski, and E. Reron. 2001. Immunogenetic significance in sensorineural hearing loss (SNHL) in the light of our own investigations. Otolaryngol. Pol. 55: 317–322. 31. Cao, M. Y., J. Thonnard, N. Deggouj, M. Gersdorff, M. Philippe, J. C. Osselaer, and J. P. Tomasi. 1996. HLA class II-associated genetic susceptibility in idiopathic progressive sensorineural hearing loss. Ann. Otol. Rhinol. Laryngol. 105: 628 – 633. 32. Yeo, S. W., K. H. Chang, B. D. Suh, T. G. Kim, and H. Han. 2000. Distribution of HLA-A, -B and -DRB1 alleles in patients with sudden sensorineural hearing loss. Acta Otolaryngol. 120: 710 –715.

33. Yeo, S. W., S. N. Park, Y. S. Park, B. D. Suh, H. Han, H. B. Choi, and T. G. Kim. 2001. Different distribution of HLA class II alleles according to response to corticosteroid therapy in sudden sensorineural hearing loss. Arch Otolaryngol. Head Neck Surg. 127: 945–949. 34. Steinman, L. 2004. Immune therapy for autoimmune diseases. Science 305: 212–216. 35. Feldmann, M., and L. Steinman. 2005. Design of effective immunotherapy for human autoimmunity. Nature 435: 612– 619. 36. Yang, M., J. Zheng, Q. Yang, H. Yao, Y. Chen, H. Tan, C. Jiang, F. Wang, M. He, S. Chen, et al. 2004. Frequency-specific association of antibodies against heat shock proteins 60 and 70 with noise-induced hearing loss in Chinese workers. Cell Stress Chaperones 9: 207–213. 37. Yuan, J., M. Yang, H. Yao, J. Zheng, Q. Yang, S. Chen, Q Wei, R. M. Tanguay, and T. Wu. 2005. Plasma antibodies to heat shock protein 60 and heat shock protein 70 are associated with increased risk of electrocardiograph abnormalities in automobile workers exposed to noise. Cell Stress Chaperones 10: 126 –135.