[CANCER RESEARCH 64, 9217–9224, December 15, 2004]

Caspase Inhibitors, but not c-Jun NH2-Terminal Kinase Inhibitor Treatment,Prevent Cisplatin-Induced Hearing Loss

Jing Wang,1 Sabine Ladrech,1 Remy Pujol,1 Philippe Brabet,1 Thomas R. Van De Water,2 and Jean-Luc Puel1

1Institut National de la Sante et de la Recherche Medicale–UMR 583 and Universite de Montpellier 1, Physiopathologie et therapie des deficits sensoriels et moteurs,Montpellier, France; and 2University of Miami Ear Institute, Department of Otolaryngology, University of Miami School of Medicine, Miami, Florida

ABSTRACT

Cisplatin (CDDP) is a highly effective chemotherapeutic agent but withsignificant ototoxic side effects. Apoptosis is an important mechanism ofcochlear hair cell loss following exposure to an ototoxic level of CDDP.This study examines intracellular pathways involved in hair cell deathinduced by CDDP exposure in vivo to develop effective therapeutic strat-egies to protect the auditory receptor from CDDP-initiated hearing loss.Guinea pigs were treated with systemic administration of CDDP. Cochlearhair cells from CDDP-treated animals exhibited classic apoptotic alter-ations in their morphology. Several important signaling events that reg-ulate the death of CDDP-injured cochlear hair cells were identified. CDDPtreatment induced the activation and redistribution of cytosolic Bax andthe release of cytochrome c from injured mitochondria. Activation ofcaspase-9 and caspase-3, but not caspase-8, was detected after treatmentwith CDDP, and the cleavage of fodrin by activated caspase-3 was ob-served within damaged hair cells. Intracochlear perfusions with caspase-3inhibitor (z-DEVD-fmk) and caspase-9 inhibitor (z-LEHD-fmk) preventhearing loss and loss of sensory cells, but caspase-8 inhibitor (z-IETD-fmk) and cathepsin B inhibitor (z-FA-fmk) do not. Although the stress-activated protein kinase/c-Jun NH2-terminal kinase (JNK) signaling path-way is activated in response to CDDP toxicity, intracochlear perfusion ofD-JNKI-1, a JNK inhibitor, did not protect against CDDP ototoxicity butinstead potentiated the ototoxic effects of CDDP. The results of the presentstudy show that blocking a critical step in apoptosis may be a usefulstrategy to prevent harmful side effects of CDDP ototoxicity in patientshaving to undergo chemotherapy.

INTRODUCTION

Cisplatin (cis-diamine-dichloroplatinum II; CDDP) is a highly ef-fective and widely used anticancer agent (1). The risk of ototoxic andnephrotoxic side effects commonly hinders the use of higher dosesthat could maximize its antineoplastic effects (2). CDDP has beenshown to induce auditory sensory cell apoptosis in vitro (3, 4) and invivo (5, 6). Devarajan et al. (7) recently have reported CDDP-inducedapoptosis in an immortalized cochlear cell line. In this model, CDDPtoxicity was associated with an increase in caspase-8 activity, fol-lowed by truncation of Bid, translocation of Bax, release of cyto-chrome c, and activation of caspase-9. This suggests that death re-ceptor and mitochondrial pathways are involved in CDDP-inducedapoptosis of hair cells. However, results obtained in vitro may not bedirectly applicable in vivo.

The present study explores the mechanisms of CDDP-inducedototoxicity in vivo and develops an efficacious strategy for protectionagainst CDDP-induced hearing loss. In addition to death receptor andmitochondrial pathways, we also examine the c-Jun NH2-terminalkinase (JNK) signaling pathway (8) on CDDP-induced apoptosis incochlear sensory cells.

MATERIALS AND METHODS

Animals. The care and use of animals followed the animal welfare guide-lines of the Institut National de la Sante et de la Recherche Medicale and wasapproved by the French Ministere de l’Agriculture et de la Foret. Adult female,pigmented guinea pigs (250 to 300 g; Charles River Laboratories, Wilmington,MA) were used in this study.

Preparation of Otoprotective Molecules for Perfusion. The specific in-hibitors of caspase-3 (z-DEVD-fmk), caspase-8 (z-IETD-fmk), caspase-9 (z-LEHD-fmk), and cathepsin (z-FA-fmk; Calbiochem, La Jolla, CA) were 83-mmol/L, 85-mmol/L, 81-mmol/L, and 72-mmol/L stock solution, respectively,in 100% DMSO (Merck, Whitehouse Station, NJ), and D-JNKI-1 was a1-mmol/L stock solution in 0.1 mol/L PBS at pH 7.2.

Before perfusion, stock solutions of the putative otoprotective compoundswere diluted in artificial perilymph to final concentrations of 100 �mol/L forcaspases and cathepsin inhibitor and 10 �mol/L for D-JNKI-1. Before surgery,the osmotic minipump (1 �L/h, over 7 days; Alza Corp., Palo Alto, CA) wasfilled under sterile conditions with 200 �L of artificial perilymph or artificialperilymph containing one of the putative otoprotective agents.

Minipump Implantation and Electrophysiologic Recordings. Experi-ments were designed to record compound action potential (CAP) in awakeanimals from a plug fixed on the head of the animal during minipumpimplantation. Briefly, the animals were anesthetized with an intraperitonealinjection of a 6% solution of sodium pentobarbital (Sanofi, New York, NY;dose, 0.3 mL/kg). Each bulla was opened under sterile conditions, and arecording electrode was positioned on the round window membrane. A hole0.25 mm in diameter was gently drilled in the base of the right cochlea forinstallation of the perfusion catheter.

CAPs of the auditory nerve were elicited by tone bursts of alternating phase(1-ms increase/decrease; 8-ms duration) and applied to the ear at a rate of 10/sfrom 0 to 100 dB sound pressure level in 5-dB steps in a free field via a JBL075 earphone (JBL, Northridge, CA). Cochlear responses were amplified(gain, 2000), averaged (128 samples), and stored on a Pentium PC computeroperating at 100 MHz (Dell Dimension; Dell, Austin, TX). CAP recordingswere measured peak to peak between the negative depression N1 and thesubsequent positive wave P1. The threshold of the CAP was defined as theintensity in decibel sound pressure level needed to elicit a measurable response(�5 �V).

Terminal Deoxynucleotidyl Transferase-Mediated Nick End LabelingAssay. ApopTag in situ apoptosis detection kit (Intergen, Purchase, NY) wasused. The incorporated digoxigenin nucleotides were immunostained with aFITC-conjugated antidigoxigenin antibody. The tissues were counterstainedwith propidium iodide (PI). FITC and PI fluorescent signals were observedwith an MRC 1024 laser scanning confocal microscope (Bio-Rad, Hercules,CA).

Hoechst Staining. Cell nuclei morphology was assessed by Hoechst 33342dye (Sigma Chemical Co., St. Louis, MO). Deparaffinized cochlear sectionswere stained with a 1:100 dilution of rhodamine-conjugated phalloidin (R-415;Molecular Probes, Eugene, OR) for 1 hour at room temperature. The sectionsthen were counterstained with Hoechst 33342 (0.002%, w/v) at room temper-ature for 10 minutes and examined on a Leica DMRB microscope (blocks offilters N2.1 and A4; Leica Microsystems, Inc., Wetzlar, Germany).

Immunocytochemistry. Cytochrome c was detected in cryostat sections(10 �m) with mouse monoclonal anti–cytochrome c antibody (1:200; Pharm-ingen, San Jose, CA) and FITC-conjugated secondary antibody (1:200 dilu-tion; goat antimouse IgG antibody B13113C; BioSys, Ann Arbor, MI). Baxwas detected with rabbit polyclonal antibody to the NH2-terminus of Bax(1:750; Upstate Biotechnology, Lake Placid, NY) that recognizes only theactivated form of Bax and an Alexa 488–labeled secondary antibody (1:1000dilution; goat antirabbit IgG antibody; Molecular Probes). Immunostained

Received 5/5/04; revised 10/6/04; accepted 10/11/04.Grant support: Association pour la Recherche sur le Cancer (ARC) and Ligue Contre

le Cancer (Comite de l’Herault).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 with18 U.S.C. Section 1734 solely to indicate this fact.

Requests for reprints: Jean-Luc Puel, INSERM U. 583, 80 rue Augustin Fliche,34295 Montpellier, France. Phone: 33-499-636-009; Fax: 33-499-636-020; E-mail:puel@montp.inserm.fr.

©2004 American Association for Cancer Research.

9217

Research. on July 6, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

specimens were counterstained with PI. The subcellular distribution of Baxwas confirmed by double labeling with two primary antibodies [i.e., a rabbitpolyclonal antibody to the NH2-terminus of Bax and a mouse monoclonalantibody to mitochondrial heat shock protein (HSP) 70, the specific mitochon-drial resident protein of the hsp70 family; clone, JG1; 1:500; Affinity BioRe-agents, Golden, CO]. Secondary antibodies were Alexa 488–labeled goatantirabbit antibody and Alexa 568–labeled goat antimouse antibody (1:500;Molecular Probes).

Activated caspase-3 and cleavage of fodrin were detected using a TyramideSignal Amplification kit (TSA Fluorescence System kit; NEN Life ScienceProducts, Boston, MA). Cryostat sections were incubated with two primaryantibodies [i.e., a monoclonal antibody against calbindin (1:600 dilution;Sigma) and a rabbit polyclonal antibody against active caspase-3 (1:6000;BioVision Incorporated, Mountain View, CA)] or a rabbit polyclonal antibodyagainst cleaved �-fodrin (Asp1185). This latter antibody recognizes fodrinfragments that are cleaved exclusively by caspases (1:400; Cell Signaling, SanDiego, CA). Secondary antibodies were biotinylated antirabbit antibody (1:300; P.A.R.I.S., Compiegne, France) and Cy3-conjugated goat antimouse IgGantibody (1:500; Jackson ImmunoResearch Labs, West Grove, PA). Fluores-cent tags were visualized with a confocal microscope (Bio-Rad). In controlspecimens without primary antibodies, neither FITC nor Cy3 fluorescent tagswere observed.

In vivo Substrate Detection of Activated Caspases. Caspase activity wasexamined in vivo using the fluorescent substrates fam-LETD-fmk (caspase-8substrate), fam-LEHD-fmk (caspase-9 substrate), and fam-DEVD-fmk(caspase-3 substrate) obtained from Intergen. Fluorescence substrate (2.5�)was perfused directly into the cochleae of untreated and CDDP-treated ani-mals. The method of perfusion used has been described previously (9). Briefly,after the cochlea was exposed ventrally, a 1-hour perfusion of the scalatympani at 2 �L/min was performed through a hole made in the lateralcochlear wall and allowed to flow out of the cochlea through a hole in the scalavestibuli. After the perfusion of the substrate, wash buffer was perfused for 1hour. Cochleae then were fixed with a final perfusion of fixative for 45minutes. The cochleae were removed and double labeled with rhodamine-conjugated phalloidin.

Western Blot Analyses. Tissue proteins were homogenized in samplebuffer (10) separated on 10% SDS-PAGE and transferred to nitrocellulosemembranes. Blots were incubated overnight at 4°C with anti-JNK polyclonalantibody or anti–phosphoThr183/Tyr185-JNK polyclonal antibody (1:1000;Cell Signaling), anti–c-Jun polyclonal antibody, or anti–phosphoSer73-Junpolyclonal antibody (1:1000; Upstate Biotechnology), followed by incubationwith peroxidase-conjugated secondary antibody (Roche Applied Science, In-dianapolis, IN). Protein-antibody complexes were revealed with the WesternLightning Chemiluminescence Reagent (PerkinElmer Inc., Wellesley, MA).Image scanning of Western blot analyses was used to quantify phosphorylationand expression levels of JNK and c-Jun.

Transmission Electron Microscopy Assessment of Hair Cell Integrity.The transmission electron microscopy procedure has been described previ-ously (11). Briefly, animals were decapitated during deep anesthesia, and theircochleae were prepared using our standard protocol. Semithin and ultrathinradial sections of the plastic embedded organ of Corti were cut from the basaland middle turns and observed using a Hitachi 7100 electron microscope(Hitachi, Tokyo, Japan) at Centre Regional d’Imagerie Cellulaire de Montpel-lier.

Statistics. All of the quantitative results are expressed as mean � SE. Allof the statistics were calculated using Sigmaplot 2000 for Windows (version6.1; Microsoft, Redmond, WA). All of the comparisons between means wereperformed using Student’s paired two-tail t tests or a nonparametric Wilcoxonrank test.

RESULTS

This study examined the mechanisms by which CDDP damageinduces apoptosis of affected hair cells and evaluated the otoprotec-tive effects of five protective molecules on preservation of structuralintegrity and function in CDDP-treated guinea pig cochleae.

Functional and Morphologic Assessment of Hair Cell Integrity.In the present study, we used two types of CDDP administration [i.e.,a single high dose of CDDP (10 mg/kg, intraperitoneally) and amultiple dose of CDDP (2 mg/kg/d, intraperitoneally, for 5 days)] thatcorrespond to the high therapeutic doses of CDDP used to managehuman cancers (12). When compared with multiple doses of CDDP,a single high-dose administration resulted in a similar pattern ofhearing loss that occurred more rapidly (i.e., 6 versus 3 days, respec-tively) but remained stable for at least 6 days (Fig. 1A). To reduce theanimal mortality that results from single high-dose administration ofCDDP, quantification of cell death of hair cells and long-term pro-tection were studied after multiple-dose CDDP injections. Even withthis precaution, 6 animals within the group of 42 animals treated withmultiple doses of CDDP died during the course of treatment. Nineanimals within this multiple-dose group did not show a clear patternof hearing loss. Because of this interindividual variability with respectto CDDP-induced hearing loss, we ensured that CDDP ototoxicity hadoccurred by recording CAP audiograms in each CDDP animal fromthe nonperfused contralateral cochlea. Only animals with a hearingloss of �30 dB at 20 kHz by day 6 in this contralateral ear wereincluded in this study (i.e., 27 of 42 animals).

In the contralateral ears, 3 days after the beginning of CDDPtreatment (day 3), only the highest frequencies tested (i.e., 16 to 26kHz) showed measurable hearing losses. During the continuing courseof the CDDP treatment (i.e., between days 3 and 5), hearing lossescontinued to develop, progressing to the middle (i.e., 6 to 16 kHz) andthen the lower (�6 kHz) frequencies. One day after the end oftreatment (day 6), significant hearing losses were observed across allof the frequencies, with higher frequencies being most affected (Fig.1A).

Comparison of surface preparations of untreated cochleae withcochleae subjected to systemic CDDP treatment (Fig. 1B) revealedthat many of the outer hair cells (OHCs) were missing from theCDDP-exposed cochleae. Counting of remaining hair cells was per-formed on surface preparations of the organ of Corti stained with arhodamine-phalloidin conjugate. These counts showed a large loss ofOHCs from the basal turns of the CDDP-treated, nonperfused con-tralateral cochleae (n � 4), with a typical gradation of base to apexpattern of hair cell loss and a typical gradient of OHC loss thatprogressed from the first to the third row of OHCs (Fig. 1A, inset). Afew areas of inner hair cell (IHC) loss were observed located in themost basal portion of CDDP-treated cochleae (Fig. 1A, inset).

DNA Fragmentation and Nuclear Morphology. To determinethe nature of cell death, we performed terminal deoxynucleotidyltransferase-mediated nick end labeling (TUNEL) on untreated (n � 2)and CDDP-treated (n � 6) cochleae. TUNEL-positive cells were notobserved in any of the cochleae from untreated control animals (Fig.1C). In contrast, the basal part of cochleae from CDDP-treated ani-mals showed TUNEL-positive labeling of a variety of cell typeswithin the organ of Corti, especially OHCs and occasionally IHCs andsupport cells (Fig. 1C).

Cell nuclei within sensory cells of cochleae from untreated controlanimals stained uniformly with Hoechst dye (Fig. 1D). After CDDPtreatment, characteristic pathologic changes in the nuclear stainingwere observed in the basal and middle turns of cochleae (i.e., theappearance of karyopyknosis in the nuclei of some OHCs and IHCs;Fig. 1D).

Bax Redistribution. We examined the activation and subcellularlocalization of Bax by using an antibody directed to the NH2-terminusof Bax. Cochleae from untreated control animals (n � 2) exhibited adiffuse, pattern of light immunostaining barely above a backgroundlevel, indicating an absence of Bax activation (Fig. 2A). Two to 4hours after a single high-dose injection of CDDP (n � 3), the pattern

9218

CASPASE INHIBITORS PREVENT CDDP OTOTOXICITY

Research. on July 6, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

of Bax immunostaining of the OHCs of the basal and middle cochlearturns changed to a dense, punctate pattern of immunostaining con-sistent with the activation and redistribution of Bax from the cytosolto the mitochondria of affected sensory cells (Fig. 2A). Double label-ing of cells showed that activation of Bax colocalized with mt HSP 70,the specific mitochondrial resident protein of the hsp70 family, intreated animals but not in control untreated cochleae (Fig. 2C).

Release of Cytochrome c. Subcellular localization of cytochromec was examined by immunostaining. In the cochleae of untreated,control animals (n � 2), immunolabeling with an anti–cytochrome cantibody showed an unequal, punctate distribution of reaction productwithin the cytoplasm (Fig. 3A). This pattern of immunostaining isconsistent with mitochondrial localization of cytochrome c. Six hoursafter a single, high-dose injection of CDDP (n � 3), the pattern of haircell and support cell cytoplasm immunostaining for cytochrome cchanged from punctate to a diffuse and weak pattern of immuno-staining with reaction product uniformly distributed throughout thecytoplasm of the affected cells (Fig. 3B). The pattern of immuno-staining in the CDDP specimens was no longer consistent with mi-tochondrial localization of cytochrome c in the auditory hair cells andthe organ of Corti support cells.

Specific Caspases. Activation of caspases was examined usingfluorescent substrate detection tests in untreated control cochleae andcochleae observed at 6 and 12 hours after a single high-dose injectionof CDDP. In untreated control cochleae (n � 6; Fig. 4A), few haircells contained the green fluorescent labeling that indicated cleavedcaspase-8, -9, and -3 substrates. A slight but not significant increase(P � 0.05) of caspase-8–, -9–, and -3–positive auditory hair cells was

noted at 6 hours after CDDP treatment (Fig. 4B; n � 9). Similarly, aslight but not significant increase (P � 0.05) of caspase-8–positivecells was noted at 12 hours (Fig. 4A and B; n � 3). In contrast,high-dose CDDP-exposed cochleae at 12 hours after treatmentshowed a robust and significant increase in the number of caspase-9–positive OHCs (Fig. 4A and B; P � 0.001; n � 3), whereas theproportion of caspase-9–positive IHCs remained approximately thesame as measured at 6 hours. Finally, a significantly increased per-centage of caspase-3–positive IHCs and OHCs was evident at 12hours after high-dose CDDP treatment (Fig. 4A and B; P � 0.01;n � 3).

Caspase-3 activation also was determined by immunolabeling usingan antibody specific for activated caspase-3. Activated caspase-3 wasnever detected in the cochlear cells from untreated control animals(Fig. 4C; n � 3). By contrast, marked caspase-3 activation wasobserved in some OHCs and a few IHCs of the basal and middle turns(Fig. 4C; n � 3) of CDDP-treated animals at 12 hours after high-doseCDDP.

Cleavage of Fodrin by Caspases. Using an antibody specific forthe Mr 150,000 NH2-terminal fragment of cleaved fodrin, we lookedfor the presence of cleaved fodrin in high-dose CDDP-exposed coch-lear cells. Cleaved fodrin was never detected in the cochlear cells fromuntreated control animals (Fig. 4D; n � 2). In contrast, there wasmarked immunostaining for cleaved fodrin in the region of the cutic-ular plates of some OHCs and a few IHCs located in the basal andmiddle cochlear turns of CDDP-treated animals observed between 6and 48 hours after injection of a single high dose of CDDP (Fig. 4D;n � 6).

Fig. 1. A, averaged CAP audiograms after singledose and multiple doses of CDDP. On day 3, thesingle-dose CDDP treatment induces a similar pat-tern of hearing loss as the multiple dose adminis-tration of CDDP (day 6). Inset, A cytocochleogramderived from nonperfused cochleae of day 6CDDP-treated animals (n � 4) shows extensiveloss of OHCs in all three rows in the basal region ofthe cochlea. B, surface preparations of organs ofCorti stained with rhodamine-labeled phalloidin(red). All of the OHCs (O) and IHCs (I) are presentin the untreated control cochlea (p � pillar cells).In contrast, only three OHCs (curved white arrows)remain in the area viewed of this CDDP-damagedcochlea. C, TUNEL staining of transverse tissuesections. DNA strand breaks are stained fluorescentgreen by FITC-labeled dUTP; nuclei are counter-stained red with PI; TUNEL-labeled nuclei appearyellow. Note the TUNEL-positive nuclei present inthree OHCs and a single IHC (white arrows) and inseveral supporting cells in CDDP-treated specimen.D, staining of nuclear chromatin with Hoechst33342 dye. In untreated control cochleae, all of thehair cells have a normal uniform pattern of nuclearchromatin staining. In CDDP-treated cochleae,OHCs show shrunk, pyknotic nuclei (white ar-rows). In both insets, representative Hoechst-stained nuclei are shown. Scale bars: B, 15 �m, Cand D, 20 �m, insets, 2.5 �m.

9219

CASPASE INHIBITORS PREVENT CDDP OTOTOXICITY

Research. on July 6, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

JNK Signaling Pathway. To identify involvement of the JNKsignal pathway in CDDP-induced apoptosis of cochlear hair cells invivo, activation of JNK was measured by immunolabeling of Westernblot analyses obtained from untreated cochleae (n � 4) and singlehigh-dose (10 mg/kg) CDDP-treated cochleae removed at 3, 6, 12, 24,and 48 hours after CDDP treatment (n � 4 per time point). High-dose

CDDP treatment induced strong phosphorylation of JNK (p-JNK) firstdetected at 3 hours after treatment, which persisted through 48 hours,with a peak level of p-JNK protein reached at 6 hours after injection.In contrast, there was no change in the expression levels of total JNKprotein observed during the 48-hour course of CDDP treatment (Fig.5A). The level of p-JNK in CDDP-treated cochleae compared with

Fig. 2. Confocal images from transverse sections ofthe basal cochlear turns from organs of Corti labeled witheither PI (red, in A and B) and antibody against Bax-NT(green, in A and B) or with antibody against HSP 70 (red,in C) and antibody against Bax-NT (green, in C). A. Anuntreated, control cochlea shows a diffuse and light pat-tern of Bax immunostaining of the organ of Corti. Incontrast, 4 hours after a single injection of CDDP, anintense, punctate pattern of staining is seen with theBax-NT antibody in the OHCs (white arrowhead). B.CDDP-treated cochlea perfused with z-DEVD-fmk orwith D-JNKI-1 shows nearly the same pattern of Bax-NTimmunostaining 4 hours after treatment as the CDDP-treated, unperfused cochlea (A, right). C. Double labelingwith Bax and HSP 70 in an untreated control cochleashows a punctate staining (red) for HSP 70 and a light,diffuse pattern for Bax (green). The same HSP 70 stain-ing is seen in CDDP-treated cochlea. The dense andpunctate Bax staining colocalizes with HSP 70 (yellow)in OHCs; tc, tunnel of Corti. Scale bars: 15 �m in A andB and 10 �m in C.

Fig. 3. Confocal images from transverse sections ofthe basal cochlear turns from organs of Corti labeled withPI (red) and antibody against cytochrome c (green, inA–D). A. An untreated, control cochlea shows an intense,punctate pattern of cytochrome c staining in the IHC andthe OHCs. B. At 6 hours after treatment, the CDDP-exposed cochlea has a diffuse, pale pattern of immuno-staining for cytochrome c. C and D. CDDP-treated co-chlea perfused with z-DEVD-fmk (C) or with D-JNKI-1(D) shows a similar diffuse pattern of cytochrome cstaining 6 hours after treatment as the CDDP-treated,unperfused cochlea (B); tc, tunnel of Corti; bar, 15 �m.

9220

CASPASE INHIBITORS PREVENT CDDP OTOTOXICITY

Research. on July 6, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

untreated cochleae (Fig. 5A) reached a peak at 6 hours after high-doseCDDP exposure. Similar results were obtained for phospho-c-Jun(p-c-Jun) and expression levels of c-Jun proteins. A single high doseof CDDP induced distinct bands of p-c-Jun labeling that increased inintensity during the 3- to 48-hour period following CDDP treatment.No change in the expression level of c-Jun protein was observedduring the course of CDDP treatment (Fig. 5B). The level of p-c-Junin CDDP-treated cochleae compared with untreated cochleae (Fig.5B) reached a peak of expression at 12 hours after high-dose CDDP.

Caspase-3 Inhibitor II and Caspase-9 Inhibitor I, but not Ca-thepsin B Inhibitor I and Caspase-8 Inhibitor III, Prevent Hear-ing Loss and Loss of Sensory Cells. Substrate detection and immu-nolabeling results indicate that procaspase-9 and procaspase-3, but notprocaspase-8, were activated in the cochlear hair cells of CDDP-treated animals. To further investigate involvement of caspase-8 inCDDP-induced apoptosis of cochlear cells, we perfused a caspase-8inhibitor III (z-IETD-fmk) into the scala tympani (n � 5). There wasno protective effect on CDDP-induced hearing loss in animals treatedwith 100 �mol/L of z-IETD-fmk (Fig. 6A). The roles of caspase-9 andcaspase-3 were investigated by perfusing the caspase-9 inhibitor I(z-LEHD-fmk) and caspase-3 inhibitor II (z-DEVD-fmk) into thescala tympani of CDDP-treated animals. Either 100 �mol/L of z-LEHD-fmk (n � 5; Fig. 6A) or 100 �mol/L of z-DEVD-fmk (n � 6;Fig. 6A) inhibitors prevented most of the CDDP-initiated hearing loss.Although 100 �mol/L of z-DEVD-fmk prevented CDDP-inducedhearing loss, it prevented neither the subcellular redistribution of Bax(Fig. 2B; n � 2) nor the mitochondrial release of cytochrome c intothe cells’ cytoplasm (Fig. 3C; n � 2). However, it did significantlysuppress the occurrence of TUNEL-positive hair cells within 48 hours

after the onset of CDDP treatment (Fig. 6B; P � 0.001; n � 3).Quantitative evaluation performed at the end of CDDP treatment (day6) of the z-DEVD-fmk–perfused cochleae revealed that there wereonly a few OHCs lost from the basal turn and that there was noapparent damage to the middle or apical turns of these z-DEVD-fmk/CDDP-treated cochleae (Fig. 6C; n � 3). The otoprotective effect ofz-DEVD-fmk against CDDP-induced ototoxicity was confirmed byultrastructural analysis of the organ of Corti (Fig. 6D), in which nodegeneration-related cellular abnormalities were observed in CDDP-treated cochleae (n � 2) protected by z-DEVD-fmk.

Because z-DEVD-fmk also might inhibit cathepsins, we performedadditional experiments using the cathepsin inhibitor (z-FA-fmk). Noprotective effect against CDDP ototoxicity was observed with perfu-sion of 100 �mol/L z-FA-fmk (Fig. 6A; n � 5), suggesting that theprotective effect of z-DEVD-fmk was caused by inhibition of caspaseactivation rather than an effect on cathepsins.

A Peptide Inhibitor of JNK (D-JNKI-1) Potentiated the Oto-toxic Action of CDDP. JNK and c-Jun activation increased over timefollowing exposure to high-dose CDDP treatment (Fig. 5A and B).Thus, we blocked the JNK pathway by using D-JNKI-1, a cell-permeable peptide that competitively blocks the ability of JNK tophosphorylate its targets. Intracochlear perfusion of 10 �mol/L D-JNKI-1 solution did not rescue the cochlea from CDDP-initiatedototoxic damage (n � 6); rather, it potentiated CDDP-initiated hearingloss between 10 and 26 kHz (Fig. 5C, left). Statistical analysisrevealed that the potentiation of CDDP-induced hearing loss by D-JNKI-1 was significant at the frequencies of 20 and 26 kHz (P � 0.05)when compared with CDDP-treated animals. Morphologic evaluationof the organ of Corti indicated that the combination of D-JNKI-1 with

Fig. 4. A, confocal images of surface preparations of the basal turn from organs of Corti showing in vivo detection of activated caspases (green) in control (left) and CDDP-damaged(right) cochleae. Rhodamine-conjugated phalloidin (red) shows actin labeling in the cuticular plates and stereocilia of the hair cells. CDDP treatment causes no activation of caspase-8.By contrast, 12 hours after injection of CDDP, a robust activation of caspase-9 and -3 is seen in many OHCs and in some IHCs. B, quantitative data for caspase-8, -9, or -3 activationin hair cells from CDDP-exposed cochleae. The number of hair cells containing activated caspase-8, -9, or -3 is counted in the 1-mm length of basal turns of CDDP-exposed cochlearducts for different time points (before and 6 and 12 hours after treatment). The results are expressed as the mean percentage of caspase-positive cells � SE (bars) for n � 4 independentexperiments. A significant increase of activated caspase-9 in OHCs and activated caspase-3 in the OHCs and IHCs appears at 12 hours after CDDP treatment (��, P � 0.01;���, P � 0.001). C, transverse sections of control and CDDP-treated organ of Corti, double labeled with calbindin (red) and with the antibody specific of the activated form of caspase-3(green). Caspase-3 immunoreactivity (arrowheads) is seen 12 hours after CDDP treatment in the OHCs (O1, O2, and O3). D, transverse sections double labeled with calbindin (red)and the antibody specific of the cleaved form of �-fodrin (green). A robust immunostaining for the cleaved form of �-fodrin (green) is seen 48 hours after CDDP treatment in thecuticular plates (arrowheads) of the OHCs and the IHC; bars, A � 20 �m, C and D � 10 �m.

9221

CASPASE INHIBITORS PREVENT CDDP OTOTOXICITY

Research. on July 6, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

CDDP treatment caused greater damage than CDDP treatment alone(Fig. 5C, right; n � 3). The basal turns of the D-JNKI-1–perfused,CDDP-treated cochleae also lost a few IHCs. Perfusion with D-JNKI-1 prevented neither the subcellular redistribution of Bax (Fig.2B; n � 2) nor the release of cytochrome c from the mitochondria(Fig. 3D; n � 2) that occurs following CDDP treatment. Ultrastruc-tural examination showed that D-JNKI-1–perfused cochleae (n � 2)displayed extensive damage and degeneration of the hair cells andmore extensive damage to other cochlear-related structures (Fig. 5D).The most frequent pathology observed within the somas of OHCs waslarge cytoplasmic vacuoles and chromatin condensations within theirnuclei, and a few OHCs showed degenerative changes indicative ofcellular necrosis.

DISCUSSION

The pattern of hair cell loss correlated with the electrophysiologicdetermination of a high-frequency hearing loss. CDDP-induced apo-ptosis of OHCs, IHCs, and nonsensory cells of the organ of Corti asshown by TUNEL labeling, Hoechst staining, and ultrastructuralanalysis correlates well with the findings of recent reports (5, 13, 14).Identification of intracellular cell death signaling pathways activatedin CDDP-stressed hair cells may offer the possibility of developing astrategy for otoprotection against CDDP-induced hearing loss.

Death Receptor Pathway. One major signaling pathway that leadsto the apoptosis of damaged cells is the Fas/APO-1–dependent (celldeath receptor) pathway. Activated Fas/APO-1 receptors interactcausing the activation of receptor-associated procaspase-8, which canactivate downstream effector caspases (15). In contrast to the in vitrostudy showing a transient activation of caspase-8 in an immortalizedcell line (7), the present study did not observe a significant increase incaspase-8 activation in CDDP-treated guinea pig cochleae. It is worthnoting that this immortalized auditory cell line represents OHC pre-cursors and not adult-like OHCs (7). Thus, the different patterns ofcaspase-8 activation observed in vitro (7) and in our in vivo studymight be caused by the differences in these two experimental models.Additional functional data reported in our study reveal that local scalatympani perfusion of z-IETD-fmk, a caspase-8 inhibitor, was ineffec-tive in preventing either CDDP-induced hair cell death or hearing loss.This result suggests that in vivo CDDP-induced apoptosis of cochlearcells is caspase-8 independent.

Mitochondrial Pathway. Another cell death pathway that partic-ipates in apoptosis of stress-damaged hair cells is the mitochondriapathway (16). Mitochondria are a target of reactive oxygen speciesand a source for the generation of additional reactive oxygen species.In situ-generated reactive oxygen species can cause cytochrome crelease in primary cultures of cerebellar granule neurons (17). Cyto-

Fig. 5. A and B, representative Western blot analyses (left) using anti-bodies to p-JNK and JNK (A) and p-c-Jun and c-Jun (B). Histograms (right)represent p-JNK/JNK (A) and p-c-Jun/c-Jun (B) ratios in CDDP-treatedcochleae expressed relative to the same ratio in untreated cochleae (i.e.,0-hour specimens). Values are presented as mean � SD from three separateexperiments. C, audiograms from D-JNKI-1–perfused right cochleae andnonperfused left cochleae in the same CDDP-treated animals (left; n � 6).Hearing losses were similar between 2 and 10 kHz but were greater in theD-JNKI-1–perfused cochleae between 10 and 26 kHz (P � 0.05). Note theextensive loss of OHCs and of a few IHCs in the basal and middle cochlearturns (right; n � 3). D, Transmission electron micrographs of the organ ofCorti from the basal turn of a CDDP-treated cochlea perfused with D-JNKI-1.OHCs and the IHC are absent and have been replaced by expansions of theirsupporting cells, the Deiters’ cell (D1, D2, and D3) and the phalangeal (ph)and border (b) cells. The curved arrow points to an apoptotic residue; ip,inner pillar cell; bar, 5 �m.

9222

CASPASE INHIBITORS PREVENT CDDP OTOTOXICITY

Research. on July 6, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

chrome c release into the cytosol is required for the formation of theapoptosome and the resultant activation of procaspase-9.

Consistent with other studies (3, 7, 18, 19), we show that anupstream initiator caspase (i.e., caspase-9) and a downstream effectorcaspase (i.e., caspase-3) were activated in the CDDP-damaged OHCsand some IHCs located in the basal turn of the cochlea. The intraco-chlear perfusion of caspase-3 inhibitor (z-DEVD-fmk) and caspase-9inhibitor (z-LEHD-fmk) dramatically reduced the ototoxic effects ofCDDP, as evidenced by a lack of DNA fragmentation with almost noapoptotic cell death of hair cells or other cell types within the cochleaand almost no loss of hearing in the CDDP-treated animals. z-DEVD-fmk is a potent, cell-permeable, and irreversible inhibitor ofcaspase-3. We cannot exclude that other members of the caspase-3subfamily are not involved in CDDP ototoxicity because it is knownthat caspase-2 and -7 also are inhibited by z-DEVD-fmk (20). Mandicet al. (21) reported on the possible involvement of calcium-dependentprotease calpains in CDDP toxicity in a human melanoma cell line,suggesting a role for calpains in CDDP ototoxicity. Although exper-iments yet have to be done in vivo, the lack of a protective effect ofcalpain inhibitors on auditory hair cell and neurons in culture arguesagainst this hypothesis (4). The present results suggest that CDDPototoxicity is mediated through a molecular pathway that involves themitochondria and the sequential activation of initiator and effectorcaspases.

JNK Signaling Pathway. The JNK signaling transduction path-way that phosphorylates the NH2-terminal region of c-Jun (22, 23) isinvolved in the control of apoptosis in certain models of oxidativestress damage. A protective role for c-Jun NH2-terminal phosphoryl-ation against DNA damage-induced apoptosis is supported by theresults of recent studies (24, 25). These studies have shown that theJNK signal cascade is activated by CDDP-induced DNA damage andis necessary for the repair of DNA-CDDP adducts and for cellviability. Therefore, to identify the involvement of this signal pathwayin CDDP-induced apoptosis of cochlear cells in vivo, levels of acti-

vated JNK were measured. Our results show that CDDP treatmentinduced a significant increase in activated JNK and that inhibition ofthis signal pathway by scala tympani perfusion of D-JNKI-1, a cell-permeable peptide that blocks JNK-mediated activation of c-Jun (26),prevented neither the CDDP-initiated activation and subcellular re-distribution of Bax nor the mitochondrial release of cytochrome c.Contrarily, inhibition of this signal cascade by D-JNKI-1 increased thesensitivity of cochlear hair cells to damage by CDDP. This resultsuggests that the JNK pathway is not involved in the CDDP-inducedhair cell death but may have a role in DNA repair and maintenance ofCDDP-damaged sensory cells.

Clinical Implications. As yet, there is no reliable test to predictwhich of those patients treated with CDDP will develop clinicallysignificant ototoxicity. It now is possible to introduce chemopro-tectants, antiapoptotic agents, and other compounds across the roundwindow membrane of the cochlea in humans (27, 28). If successful,the application of an otoprotective strategy for patients who begin toshow signs of ototoxic side effects should greatly enhance the qualityof life for patients who have to undergo high-dose CDDP chemo-therapy.

ACKNOWLEDGMENTS

We thank Pr. J.-L. Pujol for helpful comments on chemotherapy/oncologyand Dr. C. Bonny for the gift of D-JNKI-1. We also thank Dr. R. V. LloydFaulconbridge, Dr. J. Ruel, and Dr. M. Lenoir for their helpful discussionconcerning experiments.

REFERENCES

1. Trimmer EE, Essigmann JM. Cisplatin. Essays Biochem 1999;34:191–11.2. Humes HD. Insights into ototoxicity. Analogies to nephrotoxicity. Ann N Y Acad Sci

1999;884:15–8.3. Liu W, Staecker H, Stupak H, Malgrange B, Lefebvre P, Van De Water TR. Caspase

inhibitors prevent cisplatin-induced apoptosis of auditory sensory cells. Neuroreport1998;9:2609–14.

Fig. 6. A, audiograms from perfused cochleae and con-tralateral, nonperfused cochleae in CDDP-treated animals.Hearing losses are similar between the z-IETD–perfused,z-FA–perfused cochleae, and contralateral, nonperfused co-chleae. In contrast, only high-frequency hearing losses (i.e.,between 20 and 26 kHz) are seen in the z-DEVD–perfusedand z-LEHD ears; lower frequency hearing losses are totallyprevented. B, counting of TUNEL-positive hair cells in a1-mm length of surface preparation of the organ of Corti fromcontrol untreated (n � 3) and 2 days after treatment inCDDP-treated (n � 3) cochleae and z-DEVD–perfused,CDDP-treated cochleae (n � 3). Note the significant increaseof TUNEL-positive hair cells in the CDDP-exposed cochleaeand the minimal number in the z-DEVD-fmk–perfused,CDDP-treated cochleae (���, P � 0.001). C, a cytocochleo-gram from z-DEVD-fmk–perfused cochleae (n � 3). Thecytocochleogram derived from z-DEVD-fmk–perfused,CDDP-treated cochleae shows only minimal OHC loss and noIHC loss. D, transmission electron micrographs from thebasal turn of a z-DEVD-fmk–perfused cochlea from a CDDP-treated animal. The IHC (I) and the three OHCs (O1, O2, andO3) are well preserved, and their stereocilia (arrows) have anormal appearance. D1–3, Deiter’s cells; op, outer pillar cell;ip, inner pillar cell; ph, phalangeal cell; b, border cell; bar, 5�m.

9223

CASPASE INHIBITORS PREVENT CDDP OTOTOXICITY

Research. on July 6, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

4. Cheng AG, Huang T, Stracher A, et al. Calpain inhibitors protect auditory sensorycells from hypoxia and neurotrophin-withdrawal induced apoptosis. Brain Res 1999;850:234–43.

5. Alam SA, Ikeda K, Oshima T, et al. Cisplatin-induced apoptotic cell death inMongolian gerbil cochlea. Hear Res 2000;141:28–38.

6. Watanabe K, Hess A, Michel O, Yagi T. Nitric oxide synthase inhibitor reduces theapoptotic change in the cisplatin-treated cochlea of guinea pigs. Anticancer Drugs2000;11:731–5.

7. Devarajan P, Savoca M, Castaneda MP, et al. Cisplatin-induced apoptosis in auditorycells: role of death receptor and mitochondrial pathways. Hear Res 2002;174:45–54.

8. Ip YT, Davis RJ. Signal transduction by the c-Jun N-terminal kinase (JNK) frominflammation to development. Curr Opin Cell Biol 1998;10:205–19.

9. Puel JL, Saffiedine S, Gervais d’Aldin C, Eybalin M, Pujol R. Synaptic regenerationand functional recovery after excitotoxic injury in the guinea pig cochlea. C R AcadSci III 1995;318:67–75.

10. Laemmli UK. Cleavage of structural proteins during the assembly of the head ofbacteriophage T. Nature 1970;227:680–5.

11. Puel JL, Pujol R, Tribillac F, Ladrech S, Eybalin M. Excitatory amino acids antag-onists protect cochlear auditory neurons from excitotoxicity. J Comp Neurol 1994;341:241–56.

12. Waters GS, Ahmad M, Katsarkas A, Stanimir G, McKay J. Ototoxicity due tocis-diamminedichloroplatinum in the treatment of ovarian cancer: influence of dosageand schedule of administration. Ear Hear 1991;12:91–102.

13. Wang J, Lloyd Faulconbridge RV, Fetoni A, Guitton MJ, Pujol R, Puel JL. Localapplication of sodium thiosulfate prevents cisplatin-induced hearing loss in the guineapig. Neuropharmacology 2003;45:380–93.

14. Huang T, Cheng AG, Stupak H, et al. Oxidative stress-induced apoptosis of cochlearsensory cells: otoprotective strategies. Int J Devel Neurosci 2000;18:259–70.

15. Muzio M, Stockwell BR, Stennicke HR, Salvesen GS, Dixit VM. An inducedproximity model for caspase-8 activation. J Biol Chem 1998;273:2926–30.

16. Li P, Nijhawan D, Budihardjo I, et al. Cytochrome c and dATP-dependent formationof Apaf-1/Caspase-9 complex initiates an apoptotic protease cascade. Cell 1997;91:479–89.

17. Atlante A, Calissano P, Bobba A, Azzariti A, Marra E, Passarella S. Cytochrome c isreleased from mitochondria in a reactive oxygen species (ROS)-dependent fashionand can operate as a ROS scavenger and as a respiratory substrate in cerebellarneurons undergoing excitotoxic death. J Biol Chem 2000;275:37159–66.

18. Watanabe K, Inai S, Jinnouchi K, Baba S, Yagi T. Expression of caspase-activateddeoxyribonuclease (CAD) and caspase 3 (CPP32) in the cochlea of cisplatin (CDDP)-treated guinea pigs. Auris Nasus Larynx 2003;30:219–25.

19. Zhang M, Liu W, Ding D, Salvi R. Pifithrin-� suppresses p53 and protects cochlearand vestibular hair cells from cisplatin-induced apoptosis. Neurosci 2003;120:191–205.

20. Janicke RU, Ng P, Sprengart ML, Porter AG. Caspase-3 is required for �-fodrincleavage but dispensable for cleavage of other death substrates in apoptosis. J BiolChem 1998;273:15540–5.

21. Mandic A, Hansson J, Linder S, Shoshan MC. Cisplatin induces endoplasmic retic-ulum stress and nucleus-independent apoptotic signaling. J Biol Chem 2003;278:9100–6.

22. Hibi M, Lin A, Smeal T, Minden A, Karin M. Identification of an oncoprotein- andUV-responsive protein kinase that binds and potentiates the c-Jun activation domain.Genes Dev 1993;7:2135–48.

23. Xia Z, Dickens M, Raingeaud J, Davis RJ, Greenberg ME. Opposing effects of ERKand JNK-p38 MAP kinases on apoptosis. Science 1995;270:1326–31.

24. Potapova O, Haghighi A, Bost F, et al. The Jun kinase/stress-activated protein kinasepathway functions to regulate DNA repair and inhibition of the pathway sensitizestumor cells to cisplatin. J Biol Chem 1997;272:14041–4.

25. Potapova O, Basu S, Mercola D, Holbrook NJ. Protective role for c-Jun in the cellularresponse to DNA damage. J Biol Chem 2001;276:28546–53.

26. Bonny C, Oberson A, Negri S, Sauser C, Schorderet DF. Cell-permeable peptideinhibitors of JNK: novel blockers of �-cell death. Diabetes 2001;50:77–82.

27. Thomsen J, Charabi S, Tos M. Preliminary results of a new delivery system forgentamicin to the inner ear in patients with Meniere’s disease. Eur Arch Otorhino-laryngol 2000;257:362–5.

28. Seidman MD, Van De Water TR. Pharmacologic manipulation of the labyrinth withnovel and traditional agents delivered to the inner ear. Ear Nose Throat J 2003;276–300.

9224

CASPASE INHIBITORS PREVENT CDDP OTOTOXICITY

Research. on July 6, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

2004;64:9217-9224. Cancer Res   Jing Wang, Sabine Ladrech, Remy Pujol, et al.   Inhibitor Treatment, Prevent Cisplatin-Induced Hearing Loss

-Terminal Kinase2Caspase Inhibitors, but not c-Jun NH

  Updated version

  http://cancerres.aacrjournals.org/content/64/24/9217

Access the most recent version of this article at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/64/24/9217.full#ref-list-1

This article cites 27 articles, 9 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/64/24/9217.full#related-urls

This article has been cited by 14 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  SubscriptionsReprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://cancerres.aacrjournals.org/content/64/24/9217To request permission to re-use all or part of this article, use this link

Research. on July 6, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from