Hearing Research 196 (2004) 8–18

www.elsevier.com/locate/heares

RNA expression induced by cisplatin in an organof Corti-derived immortalized cell line

Maurizio Previati a,c,*, Irene Lanzoni c, Elisa Corbacella c, Sara Magosso c,Sarah Giuffr�e c, Francesca Francioso d, Diego Arcelli d, Stefano Volinia d,

Andrea Barbieri e, Stavros Hatzopoulos c, Silvano Capitani a,c, Alessandro Martini b,c

a Department of Morphology and Embryology, Human Anatomy Division, Ferrara University Via Fossato di Mortara 66, 44100 Ferrara, Italyb Department of Audiology, Ferrara University, Ferrara, Italy

c Center of Bioacustic, Ferrara University, Ferrara, Italyd Department of Morphology and Embryology, Histology Division, Ferrara University, Ferrara, Italy

e ISOF-CNR, Bologna Ferrara, Italy

Received 23 November 2003; accepted 6 April 2004

Available online 4 June 2004

Abstract

Cisplatin [cis-diamminedichloroplatinum(II)] (CDDP) is an organic compound that is widely used for the treatment of a large

number of tumors. Its clinical use is limited by the presence of some undesired side effects, like as oto- and nephro toxicity, whose

mechanisms of action are not understood. One of the possible CDDP toxicity mechanism seems to involve the generation of reactive

oxygen species (ROS), that can impair morphology and function of hair cells (HC) in the organ of Corti.

To test this hypothesis we evaluated the effect of CDDP treatment on RNA steady-state levels of 15,000 genes by microarray

analysis, using, as a experimental model, the OC-k3 cell line, obtained from the organ of Corti of transgenic mice and constitutively

expressing the large SV40 T antigen. We have found overexpression of several genes related to arachidonate mobilization including

phospholipase A2, group IV and V, phospholipase A2 activating protein and lysophospholipase I and III, as well as lipoxygenation

like arachidonate 12-lipoxygenase and arachidonate 5-lipoxygenase activating protein. In addition, we found significant tran-

scription of genes regulating cell respiration, including cyt c oxidase, as well as genes involved in xenobiotic detoxification and lipid

peroxidation such as cyt P450, and other oxidases including spermine oxidase and monoamine oxidase.

As a whole, overexpression of the group of different genes seems to indicate that an oxidative burst could take place during

cisplatin administration. We therefore searched for evidences of superoxide anion and hydrogen peroxide by means of electron

paramagnetic resonance (EPR) spectroscopy and flow cytometry, but failed to detect them. On the other hand, we found an increase

of malondialdehyde (MDA) synthesis and protein carbonylation products, indicating the occurence of lipid peroxidative degra-

dation. When we tested the effectiveness of butylated hydroxytoluene (BHT), dithiothreitol (DTT) and N-acetylcysteine (N-Ac) as

cytoprotectants, all of them reduced protein carbonylation to control levels and significantly protected OC-k3 from CDDP-induced

cell death, with an higher protection when using the lipophylic antioxidant BHT. The same antioxidants prevented also the onset of

protein carbonylation, which extent was decreased to basal levels.

These data indicate that CDDP is able to stimulate gene expression up to 12 h after the beginning of the treatment. This increase

in gene transcription involves a large number of genes potentially able to increase the level of cell ROS. Consistently, cells survival is

improved by cotreatment with antioxidants, in particular lipophilics.

� 2004 Elsevier B.V. All rights reserved.

* Corresponding author. Tel.: +39-0532-291193; fax: +39-0532-207351.

E-mail address: prm@unife.it (M. Previati).

Abbreviations: CDDP, cis-diamminedichloroplatinum(II); ROS, reactive oxygen species; HC, hair cells; EPR, electron paramagnetic resonance

spectroscopy; MDA, malondialdehyde; 4-HNE, 4-hydroxy-2-nonenal; BHT, butylated hydroxytoluene; DTT, dithiothreitol; N-Ac, N-acetylcysteine;

TBA, thiobarbituric acid; PI, propidium iodide; BCA, bicinchoninic acid assay; DMPO, 5,5-dimethylpyrroline-N-oxide; DHR, dihydrorhodamine;

PMA, phorbol myristate acetate; PMNC, polymorphonuclear cells

0378-5955/$ - see front matter � 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.heares.2004.04.009

M. Previati et al. / Hearing Research 196 (2004) 8–18 9

1. Introduction

Cisplatin is one of the most effective chemothera-

peutic drugs for a variety of human malignancies, in-

cluding ovarian, testicular, bladder, head and neck,esophageal, and small-cell lung cancer. Unfortunately

CDDP is limited in its use by its toxic side effects on the

kidney, such as renal insufficiency with tubular necrosis

and interstitial nephritis, and on the inner ear, including

high-frequency sensorineural hearing loss and periphe-

ral neuropathy. Ototoxic effects caused by CDDP are

often bilateral, occur more frequently after repeated

treatments, and generally manifest as tinnitus followedby hearing loss in the high frequency range (Rybak,

1981; Fausti et al., 1984). Morphological and functional

changes in the organ of Corti (Wright and Schaeffer,

1982; Kohn et al., 1988) include damage to outer hair

cells, that are preferentially involved, especially in the

basal turn of the cochlea (Stadnicki et al., 1975). Dam-

ages to inner hair cells (Barron and Daigneault, 1987),

supporting cells, stria vascularis (Kohn et al., 1988) andspiral ganglion has also been reported.

CDDP is usually regarded as a genotoxic agent in

force of its ability to bind to DNA and disrupt template

functions (Roberts et al., 1979); this damage to DNA

implies that replicating cells cannot complete cell divi-

sion and, in the attempt to repair DNA mismatches,

they stop cell cycle progression and subsequently die for

apoptosis, often p53 dependent (Perez, 1998). Duringthis time, protein and RNA synthesis as well as ATP

and NAD pools appear normal (Sorenson et al., 1990).

Genotoxicity alone is not the only mechanism by which

CDDP induces cell death. In fact, it also kills non rep-

licating cells; indeed only the 1% of the platinum bind to

DNA, while the remaining sticks to cellular proteins and

small molecules (Gonzalez et al., 2001).

In particular, hair cells are a postmitotic population, soother mechanisms different from interference with cell

cycle progression have to be involved in CDDP ototox-

icity induction. One of these mechanisms is believed to

consist in generation of ROS, which interfere with anti-

oxidant protection of the organ of Corti and result in

damage to hair cells. The use of antioxidants or protective

agents, such as melatonin, ascorbic acid, a-glutathione,N-acetylcysteine (Lopez-Gonzales et al., 2000), diet-hyldithiocarbamate, ebselen, lipoic acid, 4-methylthio-

benzoic acid (Rybak, 1981) and brain-derived

neurotrophic factor (BDNF) (Gabaizadeh et al., 1997),

would hypothetically protect hair cells against CDDP

damage and potentially also prevent hearing loss.

Oxidative stress has been associated with both ne-

crosis and apoptosis. Morphological and biochemical

characteristics distinguish these two forms of cell death.In necrosis, a decrease of membrane permeability oc-

curs, which results in swelling of organelles, loss of

membrane depolarization, and rupture of the plasma

membrane. During apoptosis, death promoting signals

activate a program of cell death through either new

protein or RNA synthesis (Raff et al., 1993; Steller et al.,

1995).

mRNA synthesis can be monitored by microarrays,an emerging technology based on DNA hybridization,

that provides the ability to comparatively analyze

mRNA expression of thousands of genes in parallel and

which can identify gene expression alterations when

target cells or tissues are compared with their normal

counterpart (Golub et al., 1999). Several studies have

already demonstrated the usefulness of this technique

for identifying novel disease related genes and in clas-sifying human disease at molecular level (Shulze et al.,

2001).

To better understand the biochemical mechanisms

underlying CDDP cytotoxic side effects, we have eval-

uated the effect of CDDP treatment on RNA steady-

state levels of 15,000 genes by means of microarray

analysis, and show an overexpression of several enzymes

involved in oxidative metabolism. In addition, we pro-vide evidence that confirms the existence of an impair-

ment of redox balance and in particular a selective

production of lipidic, but not hydrophilic ROS, and that

lipophilic and sulphydrilic antioxidants can improve

survival to CDDP and reduce the extent of protein

carbonylation to basal levels.

2. Materials and methods

2.1. Materials

Butylated hydroxytoluene (BHT), N-acetylcysteine

(N-Ac), dithiothreitol (DTT), propidium iodide (PI) and

all other common laboratory reagents were obtained

from Sigma Chemical Co (St. Louis, MO, USA). AsCDDP source we used Platinex from Bristol-Myers

Squibb. Dulbecco’s Modified Eagle Medium (DMEM),

fetal bovine serum (FBS) and recombinant mouse

gamma-interferon (IFN-c), were from Gibco- BRL

(Grand Island, NY, USA).

2.2. Cell culture

OC-k3 cells, obtained from the organ of Corti of

transgenic mice (ImmortomouseTM H-2Kb-tsA58,

Charles Rivers Laboratories, Wilmington, MA) (Kali-

nec et al., 1999), were cultured in the presence of 10%

CO2 at 33 �C, in DMEM supplemented with 10% FBS

and 50 U/ml of recombinant mouse IFN-c, without

antibiotics.

All drugs used to challenge cells were added directlyto the culture medium provided that the input volume

did not exceed 5% for water soluble agents and 0.5% for

water insoluble agents, which were dissolved in DMSO.

10 M. Previati et al. / Hearing Research 196 (2004) 8–18

2.3. Viability assay

Before treatment, cells were seeded in 6 well plates at

densities ranging from 150,000 to 200,000/well, left

overnight to adhere to the bottom and to acquire thenormal flattened morphology, and finally challenged

with drugs or vehicle. Cell viability was determined by a

PI exclusion assay. Briefly, the cells were treated for 6h,

after that medium was replaced with fresh medium, and

incubation allowed to continue in absence of cisplatin.

After 48 h from cisplatin challenge, cells were washed by

PBS and detached by trypsin/EDTA (for 1–2 min).

Proteolytic activity was blocked adding by fresh mediumcontaining 2 lg/ml PI. After 10 min of PI incubation the

cells in suspension were analyzed by flow citometry

(FACStar Becton–Dickinson, St. Jos�e, CA, USA). As

controls for cytometrical analysis we identified the range

of fluorescence intensity of viable cells in the untreated

population (PI-negative cells). To better identify non-

viable cells we permeabilized cells by adding for 20 min

0.1% NP40 (PI-positive cells, no more able to excludePI). For these experiments cells were considered PI-

labelled, so non viable, when displayed a fluorescence

intensity greater than the highest value of the PI-nega-

tive cells. They represented the sum of the PI-positive

cells plus cells with a fluorescence value intermediate

between PI-positive and PI-negative (PI-dim) (Zamai et

al., 2001). Data on the bar graph are expressed as per-

centage of PI-labelled cells over 10,000 cells analysed byflow cytometry in a single run.

2.4. Protein determination

Protein concentration was determined using the bi-

cinchoninic acid assay (BCA) (Pierce, Rockford, IL,

USA) (Smith et al., 1985) according to manufacturer’s

instructions, using bovine serum albumine (fraction V,Sigma Chemical Co, St. Louis, MO, USA) as a stan-

dard.

2.5. Microarray assay

Subconfluent OC-k3 cells were treated with 50 lMCDDP. After 3, 6, 12 and 24 h from CDDP addition

both treated and untreated samples were washed, tryp-sin detached (see above) and total RNA was extracted.

10 lg of total RNA were used for each sample to

synthesize cDNA by using Superscript II (Life Tech-

nologies, Inc.). Amino-allyl dUTP (Sigma) and mono-

reactive Cy3 and Cy5 esters were used for indirect

cDNA labeling. cDNA from untreated and CDDP-

treated cells was labeled with Cy3 and Cy5, respectively.

Identical amounts of labelled cDNA from untreated andCDDP-treated cells were mixed in DIG Easy hybrid-

ization solution (Roche) and spot at 37 �C overnight on

mouse is K DNA microarrays slides (Ontario Cancer

Institute, http://www.microarray.ca/). After washing

and drying, the hybridized slides were scanned using the

GenePix 4000A scanner (Axon Instruments, Foster

City, CA). Arabidopsis RNA was used as a reference for

RNA labeling.To correctly evaluate the samples, the analytical

system must be able to give identical emissions for the 2

fluorescence colors identical samples are analyzed on the

2 channels. This allows that any difference in fluores-

cence intensity between the 2 colors must be related to a

difference of cDNA mass between untreated and treated

cells. Several factors (e.g., difference in fluorochrome

emissions, different nucleotide incorporation, differentpower emission of the 2 laser used) can affect precision

and reliability. To allow for the best sample evaluation,

we performed data normalization by using GENEPIX

PRO 3.0 (Axon Instruments) and a GP3 post-processing

script (http://www.bch.msu.edu/~zacharet/microarray/

gp3.html). These programs removed spots showing no

signal or obvious defects, subtracted local backgrounds

and evaluated of the ratio between the total intensitiesassociated to Cy5 and Cy3, respectively, that ideally

should be identical. Consequentially, any ratio different

from 1 represented a normalization factor used to up-

grade the intensity values for each spot, before to per-

form data analysis.

Expression value for each gene were obtained cal-

culating the log2 ratios of fluorescence intensities from

the Cy5-specific channel versus the fluorescence fromthe Cy3-specific channel for each normalized spot. In

this way we calculate a fold increase of the expression

of mRNA from CDDP-stimulated cells relative to the

corresponding mRNA expression from untreated cells.

Thresholds for significant RNA expression changes

was established as three times the SD of the whole

assay.

For these study we performed three independent ex-periments, and only DNA spots present in at least 66%

of the replicates for each time point and with expression

ratios higher than the above-defined threshold, in at

least one array, were selected for the analysis. The sta-

tistical significance of the differential expression of any

gene was assessed by computing a p-value for each gene.

Any gene for which this p-value was less than 0.001 was

considered to be differentially expressed. No specificparametric form was assumed for the distribution of the

statistic test; consequently, to determine the p-value, we

used a permutation procedure in which the class labels

of the samples were permuted 50,000 times, and for each

permutation, two-sample t-statistics were computed for

each gene, obtaining a data score that was used to

represent the mRNA steady-state level variation for any

time point. The permutation p-value for a particulargene is the proportion of the permutations (out of

50,000) in which the permuted test statistic exceeds the

observed test statistic in absolute values.

M. Previati et al. / Hearing Research 196 (2004) 8–18 11

2.6. Evaluation of malondialdehyde formation and protein

carbonylation products

For these experiments, subconfluent OC-k3 cells were

treated for the indicated time with 50 or 100 lM CDDPcollected and subjected to the procedure specified below.

Cells (800 lg of protein) were suspended in 150 ll of0.4 M NaOH and 0.02% BHT, and incubated at 60 �Cfor 30 min. Subsequently, 300 ll of 0.66 N H2SO4 and

0.3 M Na2WO4 were added, and samples incubated on

ice for 10 min. After centrifugation to remove precipi-

tated material, supernatant was mixed 1:1 with 0.67%

(w/v) 2-thiobarbituric acid (TBA), and heated at 100 �Cfor 40 min. After cooling and removal of the insoluble

material by centrifugation, an aliquot of 10 ll was an-

alyzed by HPLC on a C18 column (Microsorb MV C18)

using as mobile phase 50 mM sodium acetate buffer/

methanol (5.5:4.5) and a fluorescence detector set at 250/

360 nm EXC/EM. Recovery tests were performed add-

ing known amounts of std MDA into samples before

alkaline hydrolysis, and evaluating external std curves.Peak area of recovered and external stds were used to

calculate MDA concentration in the sample. A MDA

std was prepared by hydrolyzing for 15 min at 37 �C a

solution of 1 mM 1,1,3,3-tetramethoxypropan 99% in

1% H2SO4.

Spectrophotometric assay of protein carbonylation

products was performed as in (Levine et al., 1990), on

samples containing 10–20 lg of protein. Data are pre-sented as the mean� SD of four replicate samples from

at least three separate experiments.

2.7. EPR assay

The spin trap 5,5-dimethylpyrroline-N-oxide(DMPO)

was purchased from Sigma–Aldrich Chem. Co. and was

vacuum distilled at 75 �C to remove paramagnetic im-purities. The EPR/spin trapping experiments were per-

formed on a Bruker EMX spectrometer, operating in

the X-band (Microwave frequency¼ 9.74 GHz, micro-

wave power¼ 20 mW, magnetic field modulation fre-

quency¼ 100 kHz, magnetic field modulation

amplitude¼ 1 G). Hyperfine coupling constants (hfcc)

have been calculated by best fit simulation of experi-

mental spectra using the NIEHS WinSim software(Duling, 1994). Assays were performed at room tem-

perature. Data are presented as plots showing in ab-

scissa the magnetic field expressed in Gauss (G), and in

ordinate the signal intensity, expressed in arbitrary

units.

2.8. Cytometrical hydrogen peroxide detection

Polymorphonuclear leukocyte cells (PMNC) were

purified from whole blood cells using Histopaque 1077

and 1119 (Sigma) according to manufacturer’s instruc-

tions. PMNC were stimulated by 20 nM phorbol myr-

istate acetate (PMA) up to 10 min, while OC-k3 were

treated with 50 lM CDDP up to 4 h. PMNC were

preincubated with 7.5 lM dihydrorhodamine (DHR)

(Molecular Probes, Eugene, Oregon) for 10 min beforePMA treatment while in OC-k3 DHR was added 10 min

before the end of CDDP treatment. The increase of

fluorescence on the green channel, due to the DHR

oxidation by H2O2, was evaluated by flow cytometry on

a FACStar (Becton Dickinson, St. Jos�e, CA, USA), and

represented on plots showing the cell number (Y-axis)

versus fluorescence intensity (X-axis, logarithmic).

2.9. Statistical analysis

Statistical analyses were performed using the Krus-

kal–Wallis test, followed by Tukey-Kramer, a multiple

comparison test. A p-value of <0.05 was considered

significant.

3. Results

We have previously reported that CDDP, after an

initial lag phase of 24 h, induces cell death in OC-k3

by apoptosis (Bertolaso et al., 2001). Since apoptosis

is an active process often dependent upon protein

synthesis, we decided to evaluate gene expression

patterns occurring after 3, 6, 12 and 24 h since CDDPadministration. We compared expression profiles of

treated samples with respect to controls, using a

cut-off value (greater than 3-fold for mRNA overex-

pression and lower than 3-fold for mRNA underex-

pression). We found that at the different treatment

times, mRNA steady state changed significantly.

Among the 15,000 ESTs present on the slides, at 3, 6,

12 and 24 h, the number of upregulated genes was1740, 1835, 1829 and 255, respectively, while the

downregulated genes were (3 h) 155, (6 h) 78, (12 h)

79 and (24 h) 871. In particular, the decrease of

mRNA synthesis occurring at 24 h was significantly

different with respect to the earlier times ðp < 0:05Þ.Among the large number of overexpressed genes, we

found that CDDP increased the transcription of some

classes of enzymes potentially ROS producers: spermineoxidase increased by 3.2-fold at 3 h and remained at

elevated (8–10-fold) at 6 and 12 h, while at 24 h its

mRNA level came back to amounts 0.5-fold higher than

untreated samples. Similarly, monoamine oxidase

showed an increase of mRNA steady state between 3.4

and 4-fold at 3, 6 and 12 h; at 24 h the mRNA expres-

sion was lower than in the untreated samples. Cyto-

chrome c oxidase is an enzyme responsible for theelectron transport from cyt c to oxygen and related to

ROS production. This enzyme is present in several iso-

forms, four of which (Vb, VIa, VIIa and VIIIa) showed

12 M. Previati et al. / Hearing Research 196 (2004) 8–18

increases from 1.4- to 13.6-fold higher than untreated

samples during the first 12 h of treatment. Similarly,

several oxygenases, including tryptophan 2,3-dioxygen-

ase, tyrosine and prolyne-hydroxylase and a tyrosin3-

tryptophan 5 monooxygenase activating protein showedincreases from 2.88 to 14.1 during the first 12 h of

treatment (Table 1).

In addition, CDDP significantly increased the ex-

pression of enzymes involved in arachidonate produc-

tion and oxygenation. In fact, both phospholipase A2,

group IVA and V showed increases in the range from

1.3- to 4.5-fold, while phospholipase A2 activating

protein increased its mRNA level of 27-fold at 6 and 12 hafter CDDP challenge. Also lysophospholipase I and III

increased; their mRNA steady-state level at least of 9-

and 1.2-fold, respectively, along the first 12 h from

stimulation. Interestingly, both phospholipase A2 group

IVA and lysophospholipase were the only 2 enzymes

which showed an higher mRNA level after 24 h from

stimulation (Table 1).

Consistently with overexpression of enzymes involvedin arachidonate mobilization, several enzymes involved

in arachidonate metabolism, including arachidonate 12-

lipoxygenase and a protein activator of arachidonate 5-

lipoxygenase, resulted to display a significantly higher

level of mRNA at 3, 6 and 12 h after CDDP adminis-

tration (Table 1). These increases were of approximately

1-fold for arachidonate 12-lipoxygenase, 6–7-fold for

the activating protein of arachidonate 5-lipoxygenaseand 15-fold for the epidermal isoform of arachidonate

lipoxygenase.

Another relevant enzyme involved in xenobiotic de-

toxification, lipid oxidation and ROS production is cyt

Table 1

Changes in mRNA steady-state levels after cisplatin administration

Spermine oxidase

Monoamine oxidase A

Cytochrome c oxidase, subunit Vb

Cytochrome c oxidase, subunit VI a

Cytochrome c oxidase subunit VIIa

Cytochrome c oxidase subunit VIIIa

Tyrosine hydroxylase

Prolyl 4-hydroxylase beta

Tryptophan 2,3-dioxygenase

Tyrosine 3-/tryptophan 5-monooxygenase activating protein

Lysophospholipase 1

Lysophospholipase 3

Phospholipase A2, group IVA

Phospholipase A2, group V

Phospholipase A2, activating protein

Arachidonate 5-lipoxygenase activating protein

Arachidonate 12-lipoxygenase

Arachidonate lipoxygenase, epidermal

P450 (cytochrome) oxidoreductase

Values represent the variation of mRNA steady-state levels at different tim

least in two microarray slides over three were used to calculate the mean of th

of the CDDP-treated sample and untreated sample, respectively.

P450, which increased of 14.3-fold at 3 h to 18-fold at 6

and 12 h, and dropped to 2.2-fold at 24 h.

The overexpression of so wide group of enzymes re-

lated to oxygen metabolism and lipid degradation and

oxygenation, led us to hypothesize that some forms ofoxidative burst can take place after CDDP administra-

tion to cells. In order to test this hypothesis, we studied

the production of oxygen radicals by different ap-

proaches: we initially attempted to directly detect su-

peroxide generation by spin trapping and EPR analysis,

and subsequently to identify several downstream prod-

ucts of radical production, including H2O2, MDA and

carbonylated protein species that originated by the re-action between residues on the protein surface and side

products of lipid peroxidation as MDA and 4-HNE.

EPR detection was performed on the cell medium

collected at different times after CDDP administration

using DMPO as a spin trap. When DMPO interacts

with a ROS, the unpaired electron of the radical can be

trapped by the spin trap. This link stabilizes the un-

paired electron, and its paramagnetic resonance can berevealed when an external magnetic field is applied. The

intensity of the paramagnetic resonance signal is directly

proportional to the radical concentration, while the

peak-to-peak distance (hyperfine coupling constant,

hfcc) and the relative peak intensity permit to identify

the radical species.

For these experiments, we used as positive control

PMNC, where the oxidative burst was induced by 20nM PMA (Fig. 1A). PMA was reported to be able to

trigger the physiological mechanism that in PMNC al-

lows the defence against invading pathogens and implies

extracellular liberation of ROS (Segal et al., 1993). Ef-

3 h 6 h 12 h 24 h

3.2 8.5 10.2 0.5

3.4 4 4 0.4

8.2 13.6 13.5 2.4

2.7 2.8 2.8 0.5

3.2 9.7 9.7 0.7

1.5 3.3 3.2 1.4

14.1 5.5 5.5 0.7

9 7.7 7.7 0.4

7.7 6.8 6.8 1.5

2.9 3.5 3.5 0.7

9.4 9.2 9.2 0.8

1.7 1.2 1.2 1

4.5 2.3 2.3 2.2

3 1.3 1.3 0.3

0.9 27.4 27.3 3

6.1 7.7 7.6 1.1

1.3 0.9 0.9 0.3

1.2 15.2 15.1 2

14.3 18.4 18.2 2.7

es of CDDP treatment, as evaluated by microarray. Genes present at

e log2 (ICy5/ICy3), where ICy5 and ICy3 are the fluorescence intensities

Fig. 1. EPR spectra of culture media from cells incubated in the presence of 20 mM DMPO and several pro-oxidant: (A) PMNC incubated with

vehicle (i) or 20 nM PMA for 10 min (ii). (B) OC-k3 incubated with vehicle (i) or 100 lM menadione for 4 h (ii). (C) OC-k3 incubated with vehicle (i)

or 50 lM CDDP for 4 h (ii). In abscissa is reported the magnetic field expressed in Gauss (G), and in ordinate the signal intensity, expressed in

arbitrary units.

M. Previati et al. / Hearing Research 196 (2004) 8–18 13

fectively, in PMNC we found a strong increase in the

paramagnetic signal of DMPO, as function of the ex-

ternal magnetic field, 10 min after stimulation. The hy-

perfine coupling constant of the signals were

aðNÞ ¼ 14:9 G, aðHÞ ¼ 14:9 G and the intensity ratio of

the four peaks is 1:2:2:1. These values were character-

istics for a OH radical trapped by the spin trap DMPO

in aqueous media. This indicated the generation of

14 M. Previati et al. / Hearing Research 196 (2004) 8–18

hydroxyl radical, which can be produced in presence of

hydrogen peroxide, a catalyst, and superoxide anion.

We further confirmed the presence of hydrogen peroxide

by a flow cytometry assay. In fact, PMNC incubated

with the non fluorescent probe DHR, and subsequentlychallenged with PMA, revealed an at least 10-fold in-

crease of DHR fluorescence intensity (Fig. 2A). By the

fact that DHR do not directly detect superoxide, but

rather reacts with hydrogen peroxide in the presence of

suitable catalysts, like as peroxidase, cytochrome c or

Fe2þ (Henderson et al., 1993; Royall et al., 1993), the

fluorescence intensity shift indicated a massive and fast

production of H2O2 which was reduced by DHR which,

Fig. 2. Flow cytometry analysis of PMNC and OC-k3 cells incubated

with 7.5 lM DHR and challenged by different drugs: (A) PMNC not

treated (line a) or treated for 10 min with 20 nM PMA (line b). (B) OC-

k3 cells not treated (line a) or treated for 4 h with 100 lM CDDP (line

b). In abscissa is reported the fluorescence intensity and in ordinate the

number of cells.

in turn, shifting to the oxidative state greatly increased

its fluorescence emission.

EPR analysis was also able to detect the presence of

H2O2 in OC-k3 when we used 100 lM menadione as

ROS generator (Fig. 1B). Similarly, flow cytometryanalysis was successful in detecting hydrogen peroxide

in OC-k3 treated with 100 lM menadione (data not

shown). This led us to conclude that it was so possible to

detect H2O2 in OC-k3, excluding the presence of rele-

vant matrix effects eventually related to this particular

cell type, which expresses a viral protein and is INF-ctreated. It is to note that, when we challenged OC-k3

with 50 lM CDDP, a concentration that induces ap-optosis in these cells, EPR spectrometer did not revealed

the presence of neither superoxide anions nor H2O2

(Fig. 1C). In a similar manner, flow cytometry analysis

did not show any oxidation of DHR when OC-k3 cells

where incubated with the apoptogenic concentration of

100 lM CDDP (Fig. 2B).

This suggested that, if CDDP implied the production

of hydrophilic ROS, their concentrations were below thesensitivity threshold of the EPR and cytometrical ana-

lytical systems and so not detectable, if not produced at

all.

Having not been successful in the detection of hy-

drophilic radicals by spin trap and fluorescent probe

oxidation, we decided to use an indirect assay for ROS

production, and searched for degradative products of

lipid peroxidation, such as MDA. This could be ad-vantageous because ROS-induced lipid peroxidation,

once started, is able to proceed autocatalytically, do not

stop if not in the presence of suitable amounts of radical

scavengers, and so is able to amplify the initial ROS

signal. We found that MDA production increased dur-

ing CDDP treatment, producing, in addition, a transient

peak at 2 h after CDDP challenge (Fig. 3).

Fig. 3. Evaluation of the MDA production by HPLC analysis in OC-

k3 cells after 50 lM CDDP treatment. At different times cells were

harvested and treated as in Section 2. In the Y-axis the MDA recovery

ðlM) is represented. *p < 0:05, **p < 0:01, ***p < 0:001: values ver-

sus control. The data are representative of three different experiments

giving similar results.

M. Previati et al. / Hearing Research 196 (2004) 8–18 15

In addition, MDA, like 4-HNE, another molecule

which can originate from lipid peroxidation, can react

with the aminic groups exposed on protein surfaces,

forming monofunctional or bifunctional carbonylic ad-

ducts. We verified it in our cell model, and found thatthe incubation of OC-k3 cells for 15 min with exogenous

MDA produced increasing amounts of protein car-

bonylation products significantly different from controls

(Fig. 4). Yet, we also detected protein carbonylation

products when we challenged cells with CDDP. These

products appeared to be roughly concentration depen-

dent and showed a peak at 3 h. Later, they decreased

with respect to the 3 h peak, but remained significantlyhigher than control (Fig. 5).

Fig. 4. Protein carbonylation depends upon exogenous MDA addition.

OC-k3 cells were treated with 1, 10 and 100 lMMDA for 15 min, after

which cells were collected, stained with 2-diphenylhydrazine acid,

precipitated, and the absorbance of the washed pellet evaluated.

Empty bar: control; light points bar: 1 lMMDA; diagonal stripes bar:

10 lM MDA; square bar: 100 lM MDA. Values on Y-axis represent

the absorbance at 360 nm, expressed in mOD. Data are the mean� SD

of three separate experiments, each performed in quadruplicate

*p < 0:05, **p < 0:01, ***p < 0:001: values versus control.

Fig. 5. Protein carbonylation evaluated in OC-k3 cells treated for 3, 6

and 16 h from addition of 50 or 100 lM CDDP. Cells were collected,

stained with 2-diphenylhydrazine, acid precipitated, and the absor-

bance of the dinitrophenol adduct in the washed pellet evaluated

spectrophometrically. Empty bar: control; light points bar: 3 h; bar:

diagonal stripes 6 h; square bar: 16 h. Values on Y-axis represent the

absorbance at 360 nm, expressed in mOD. Data are the mean�SD of

three separate experiments, each performed in quadruplicate.

*p < 0:05, **p < 0:01, ***p < 0:001: values versus control.

Finally, to determine if the production of lipid ROS

can actually have a role in inducing apoptosis or, on the

contrary is only an epiphenomenon not related to cell

death, we tested the cytoprotective activity of different

radical scavengers. So we stained cell cultures with PIand evaluated the number of PI labelled cells by flow

cytometry. We measured the fluorescence intensity of

cells in untreated or NP40-lysed samples, respectively,

and identified as dead cells all the cells showing a fluo-

rescence intensity higher than the viable group. We

found that in not treated cells the number of dead cells

was the 6� 1% of the total population, while in the

presence of 50 lM CDDP cell death was increased to50.2� 8%. When, together with CDDP, we added 100

lM BHT, a lipophilic molecule with a high partition

coefficient for the organic phase, it gave the highest level

protection, with a significant reduction ðp < 0:001Þ of

CDDP cytotoxicity to a level not significantly different

from the control (10.2� 2%). Also 100 lM DTT, a

sulphidryl group-containing molecule, gave significant

protection ðp < 0:001Þ, reducing during the CDDP co-treatment the number of dead cells to 11.6� 6% of the

total population. 500 lM N-Ac, a commonly used an-

tioxidant, showed the lowest significant protection

ðp < 0:05Þ, decreasing the number of dead cells up to

21.8� 3% of total cell numbers (Fig. 6).

The protection level of the three antioxidants was

found to roughly correlate to their scavenging activity.

In fact, when we determined the level of carbonylation

Fig. 6. PI supravital labelling of detached OC-k3. Cells were treated for

6 h with 50 lM CDDP and, where indicated, cotreated with 100 lMBHT, 100 lM DTT, 500 lM N-Ac, detached by proteolytic treatment,

immediately labelled with 2 lg/ml PI and analyzed by flow cytometry.

Data represent the percentage, over 10,000 analyzed events, of the cells

displaying a fluorescence intensity greater than viable cell population,

and are the mean of three separate experiments, each performed at

least in quintuplicate. Empty bar: control; light points bar: 50 lMCDDP; diagonal stripes bar: 100 lM BHT; square bar: 100 lM DTT;

black squares: 500 lM N-Ac. *p < 0:05, **p < 0:01, ***p < 0:001:

values versus control. �p < 0:05, ��p < 0:01, ���p < 0:001: values versus

50 lM cisplatin.

Fig. 7. Protein carbonylation evaluated in OC-k3 cells treated for 6 h

with 50 lM CDDP and, where indicated, cotreated with 100 lM BHT,

100 lM DTT, 500 lM N-Ac. Cells were collected, stained with 2-

diphenylhydrazine, acid precipitated, and the absorbance of the dini-

trophenol adduct in the washed pellet evaluated spectrophometrically.

Empty bar: control; light points bar: 50 lM CDDP; diagonal stripes

bar: 100 lM BHT; square bar: 100 lM DTT; black squares: 500 lMN-Ac. *p < 0:05, **p < 0:01, ***p < 0:001: values versus control.

�p < 0:05, ��p < 0:01, ���p < 0:001: values versus 50 lM cisplatin.

16 M. Previati et al. / Hearing Research 196 (2004) 8–18

adducts in samples cotreated with 50 lM CDDP andantioxidants, we found that 50 lM CDDP alone in-

duced an increase of absorbance from 267� 36 mOD/

well of the control to 499� 26 mOD/well, with an in-

crease of 86.9% ðp < 0:001Þ. Cotreatment with BHT,

DTT and N-Ac resulted in similar protein carbonylation

levels (279� 42, 295� 24, 350� 48 mOD/well, respec-

tively) that were all significantly different from cisplatin

ðp < 0:001Þ and not significantly different either fromcontrol or among them (see Fig. 7).

4. Discussion

In this work, we used as cell model a cell line derived

from the organ of Corti (Kalinec et al., 1999) to study

the role of gene expression in the cytotoxicity that oc-curred after CDDP administration, with particular at-

tention to oxidative enzymes.

Initially, we investigated the gene expression profiles

by cDNA microarray analysis of mRNA extracts from

cells treated with CDDP for times ranging from 3 to 24 h.

The results show that approximately the 11–12% of the

screened genes increase significantly their expression up

to 12 h after CDDP administration, while less than 1%were repress in this time range. After 24 h from treat-

ment we assist to a drastic and significant reduction of

overexpressed gene to 1.6%, while downregulated in-

crease to 6%.

In the present report we restrict our attention to some

gene groups that have been upregulated in the first 12 h

of treatment and are related to ROS production. We

find that after CDDP challenge OC-k3 cells overexpressa large group of oxidases and oxygenases. Among oth-

ers, several isoforms of cytochrome c oxidase increased

significantly. This enzyme is known to catalyze the

transfer of electrons from cyt CC to oxygen, with pro-

duction of water; during this physiological process the

transfer of an additional unpaired electron generates the

superoxide molecule, in the terms of 0.5–2% of the water

produced.Yet, cytochrome P450 oxidoreductase is known to be

part of the microsomal monooxygenase, multienzyme

system present both in endoplasmic reticulum and in

mitochondria. This complex, the principal function of

which is to oxygenate hydrophobic exogenous com-

pounds or endogenous substrates (e.g., xenobiotics in

liver, cholesterol in adrenal cortex) is also believed to be

one of the major sources of cellular ROS (Bondy et al.,1994; Patten et al., 1995). This ROS generation seems to

be an important feature of this enzyme complex, because

its rate has been shown to be higher with respect to the

rate of substrate consumption. Additionally, it has

shown to be the principal cause of cytotoxicity after

ethanol and arachidonic acid challenge in HepG2 liver

cells (Chen et al., 1997a,b; Cederbaum et al., 1998).

In addition to these enzymes, CDDP induces theupregulation of a set of enzymes that in principle can

increase the rate of lipid radical production: in fact we

noticed a higher expression of phospholipase A2, group

IVA and V, phospholipase A2 activating protein, lyso-

phospholipase 1 and 3. These proteins can de-esterifiy

the fatty acids bound to phosholipids and lysophos-

pholipids; these fatty acids and arachidonate in partic-

ular can in turn act directly as apoptotic inducers (Caoet al., 2000). In addition, they are physiological

substrates of lipoxygenases including arachidonate 12-

lipoxygenase, or 5-lipoxygenase activating protein, the

mRNA synthesis of which increases during CDDP

treatment. In the renal proximal tubule, where the S2

and S3 segments are target sites for CDDP and other

drugs, xenobiotic metabolic activation occurs via pros-

taglandin synthetase and lipoxygenase systems. Thesemediate co-oxidation of arachidonate and xenobiotics

and are overexpressed in models of chemically induced

renal toxicity (Hawksworth et al., 1996). It is reasonable

that an overactivation of these lipoxygenases could also

lead to cell toxicity in our inner ear-derived cell model.

In fact, these findings are consistent with the cell’s

overexpression of a set of defensive enzymes in order to

counteract the effects of xenobiotic compound. As a sideeffect, this cascade could induce the overproduction of

radicals of different origin.

To verify this possibility we monitored ROS pro-

duction by several different approaches, including spin

trapping, fluorescent probe oxidation and lipid peroxi-

dation product detection. As positive controls for radi-

cal production we employed PMNC treated with PMA

and OC-k3, an otocyst-derived cell line treated withmenadione, a redox-cycling agent able to generate su-

peroxide anions. The results indicate that, while mena-

dione on OC-k3 cells and PMA on PMNC produces

M. Previati et al. / Hearing Research 196 (2004) 8–18 17

measurable amounts of hydrogen peroxide with different

kinetics in the time range explored, no detectable pro-

duction of oxygen superoxide or hydrogen peroxide is

possible after CDDP treatment of OC-k3 cells in a pe-

riod up to 4 h.Is possible that the amount of ROS produced by

CDDP are below the detection level of our assays. We

therefore searched for indicators of degradative pro-

cesses involving lipid peroxides, and in particular for the

presence of MDA. In principle, also a small amount of

ROS can originate a relevant amount of lipid peroxides

in force of an autocatalytic mechanism. In addition, we

used a HPLC approach instead of spectrophotometricanalysis because, as already reported in literature, the

chromatographic step led to separation of MDA from

other thiobarbituric reacting substances, allowing for

greater specificity in the detection of MDA, and reduc-

ing or eliminating the frequent pitfalls that affect the

spectrophotometric assay (Esterbauer et al., 1990; Dra-

per et al., 1990).

With this approach, we found an increase in MDAproduction with a peak at 2 h, followed by a constant

increase up to 48 h. This suggests the presence of a

radical-mediated degradative damage of cell mem-

branes, that occurs up to the onset of cell death, but that

becomes significant after 2 h of CDDP incubation. This

early production of lipid-derived aldheydes could not

merely be a marker of cell degeneration but have a

trigger role for the further cytotoxicity steps. In fact,MDA, like the similar compound 4-HNE, another

degradative products of lipid peroxidation, have been

reported to have cytotoxic, genotoxic and chemotactic

properties, both inhibitory and stimulatory effects on a

number of enzymes and contribute to the pathogenesis

of many diseases and to induce apoptotic death (Dalle-

Donne et al., 2003: Byoung et al., 2001).

In addition, they are highly reactive aldheydes andcan react with some protein residues, in particular

amino groups, generating protein carbonylation prod-

ucts. We have investigated this point and found that

addition of exogenous MDA actually gave rise to a

proportional increase in protein carbonylation products.

Furthermore, when we added CDDP to cells, we found

an increase in protein carbonylation product recovery

that become significant at 3 h, successively to CDDPinduced MDA production. This fact further supports

the existence of CDDP-induced lipid peroxide degra-

dative effects.

Additionally, further verification of the role of lipid

peroxidation in CDDP cytotoxicity come from antioxi-

dant experiments: not only do the use of antioxidants

prevent the CDDP-induced protein carbonylation, but

also reduce cell death due to CDDP. Is to note that alipophilic antioxidant such as BHT gives better results

than N-Ac both in preventing protein carbonylation

increase and cell death. The finding that DTT is highly

efficient in preventing cisplatin toxicity is not contra-

dictory because we have to consider that DTT can uti-

lize a mechanism of action different from scavenger

activity. In fact it possess two –SH groups that can not

only furnish reductive power but could also efficientlyand covalently react to the CDDP molecule, as previ-

ously shown for methionine and glutathione (Eastman,

1987).

The finding that a radical scavenger is able to easily

dissolve in the organic phase and can counteract the

cytotoxic effect of CDDP better than hydrophilic com-

pounds, suggests that, although radical production can

occur everywhere in the cell, the site of action that isrelevant to cytotoxicity can be strictly compartmental-

ized in the lipid phase. Is to note that, in the absence of

chain-breaking scavenger, the radical damage at the le-

vel of polyunsaturated lipid can be autocatalitically

propagated to neighbouring lipid molecules, and repre-

sents a potentially good source of cell cytotoxicity.

Taken together, these data support the hypothesis

that the CDDP mechanism of action can affect geneexpression, and in particular that genes involved in the

oxidative metabolism of lipids can have a key role in

generating oxygen radicals and lipid degradative prod-

ucts able to play a key role in the induction of pro-

grammed cell death.

Acknowledgements

The present study was supported by grants from

CNR (Progetto Finalizzato Biotecnologie) and from

MURST (COFIN 2002 to M Previati, and COFIN 2001

to S. Capitani) as well as from the University of Ferrara

(ex 60% funds) to S. Capitani and A. Martini.

E. Corbacella is attending a Ph.D. in Neurobiological

and Electrophysiological Sciences. Sarah Giuffr�e andSara Magosso are attending a Ph. D in Biomedical,

Endocrinological and Neurophysiological Sciences.

References

Barron, S.E., Daigneault, E.A., 1987. Effect of CDDP on hair cell

morphology and lateral wall Na,K-ATPase activity. Hear Res. 26,

131–137.

Bertolaso, L., Martini, M., Bindini, D., Lanzoni, I., Parmeggiani, A.,

Vitali, C., Kalinec, G., Kalinec, F., Capitani, S., Previati, M., 2001.

Apoptosis in OC-k3 immortalized cell line treated with different

agents. Audiology 40, 327–335.

Bondy, S.C., Naderi, S., 1994. Contribution of hepatic cytochrome

P450 systems to the generation of reactive oxygen species.Biochem.

Pharmacology 48, 155–159.

Byoung, J.S., Yunjo, S., Myung-Ae, B., Jae-Eun, P., Jie, W., Kyu-

Shik, J., 2001. Apoptosis of PC12 cells by 4-hydroxy-2-nonenal is

mediated through selective activation of the c-Jun N-terminal

protein kinase pathway. Chemico-Biol. Interact. 130–132, 943–954.

Cao, Y., Pearman, A.T., Zimmerman, G.A., McIntyre, T.M., Prescott,

S.M., 2000. Intracellular unesterified arachidonic acid signals

apoptosis. PNAS 97 (21), 11280–11285.

18 M. Previati et al. / Hearing Research 196 (2004) 8–18

Cederbaum, A.I., 1998. Ethanol-related cytotoxicity catalyzed by

CYP2E1-dependent generation of reactive oxygen intermediates in

transduced HepG2 cells. Biofactors 8, 93–96.

Chen, Q., Cederbaum, A.I., 1997a. Menadione cytotoxicity to Hep G2

cells and protection by activation of nuclear factor-kappaB. Mol.

Pharmacol. 52, 648–657.

Chen, Q., Galleano, M., Cederbaum, A.I., 1997b. Cytotoxicity and

apoptosis produced by arachidonic acid in Hep G2 cells overex-

pressing human cytochrome P4502E1. J. Biol. Chem. 272, 14532–

14541.

Dalle-Donne, I., Giustarini, D., Colombo, R., Rossi, R., d Milzani, A.,

2003. Protein carbonylation in human diseases. Trends Mol. Med.

9, 169–176.

Draper, H.H., Hadley, M., 1990. Malondialdehyde determination as

index of lipid peroxidation. Methods Enzymol. 186, 421–431.

Duling, D.R., 1994. Simulation of Multiple Isotropic Spin Trap EPR

Spectra. J. Magn. Reson. 104, 105–110.

Esterbauer, H., Cheeseman, K.H., 1990. Determination of aldehydic

lipid peroxidation products: malonaldehyde and 4-hydroxynone-

nal. Methods Enzymol. 186, 407–421.

Eastman, A., 1987. Crosslinking of glutathione to DNA by cancer

chemotherapeutic platinum coordination complexes. Chem. Biol.

Interact. 61, 241–248.

Fausti, S.A., Schechter, M.A., Rappaport, B.Z., Frey, R.H., Mass,

R.E., 1984. Early detection of CDDP ototoxicity. Selected case

reports. Cancer 53, 224–231.

Gabaizadeh, R., Staecker, H., Liu, W., Van De Water, T., 1997.

BDNF protection of auditory neurons from CDDP involves

changes in intracellular levels of both reactive oxygen species and

glutathione. Brain Res. Mol. Brain. Res. 50, 71–78.

Golub, T.R., Slonim, D.K., Tamayo, P., Huard, C., Gaasenbeek, M.,

Mesirov, J.P., Coller, H., Loh, M.L., Downing, J.R., Caligiuri,

M.A., Bloomfield, C.D., Lander, E.S., 1999. Molecular classifica-

tion of cancer: class discovery and class prediction by gene

expression monitoring. Science 286, 531–537.

Gonzalez, V.M., Fuertes, M.A., Alonso, C., Perez, J.M., 2001. Is

cisplatin-induced cell death always produced by apoptosis? Mol.

Pharmacol. 59, 657–663.

Hawksworth, G.M., McCarthy, R., McGoldrick, T., Stewart, V.,

Tisocki, K., Lock, E.A., 1996. Site specific drug and xenobiotic

induced renal toxicity. Arch. Toxicol. Suppl. 18, 184–192.

Henderson, L.M., Chappell, J.B., 1993. Dihydrorhodamine 123: a

fluorescent probe for superoxide generation? Eur. J. Biochem. 217,

973–980.

Kalinec, F., Kalinec, G., Boukhvalova, M., Kachar, B., 1999.

Establishment and characterization of conditionally immortalized

organ of corti cell lines. Cell Biol. Int. 23 (3), 175–184.

Kohn, S., Fradis, M., Pratt, H., Zidan, J., Podoshin, L., Robinson, E.,

Nir, I., 1988. CDDP ototoxicity in guinea pigs with special

reference to toxic effects in the stria vascularis. Laryngoscope 98,

865–871.

Levine, R.L., Garland, D., Oliver, C.N., Amici, A., Climent, I., Lenz,

A.-G., Ahn, B.-W., Shaltiel, S., Stadtman, E.R., 1990. Determina-

tion of carbonyl content in oxidatively modified proteins. Methods

Enzymol. 186, 464–478.

Lopez-Gonzalez, M.A., Guerrero, J.M., Rojas, F., Delgado, F., 2000.

Ototoxicity caused by CDDP is ameliorated by melatonin and

other antioxidants. J. Pineal Res. 28, 73–80.

Patten, C.J., Koch, P., 1995. Baculovirus expression of human P450

2E1 and cytochrome b5: spectral and catalytic properties and effect

of b5 on the stoichiometry of P450 2E1-catalyzed reactions. Arch.

Biochem. Biophys. 317, 504–513.

Perez, R.P., 1998. Cellular and molecular determinants of cisplatin

resistence. Eur J Cancer. 34, 1535–1544.

Raff, M.C., Barres, B.A., Burne, J.F., Coles, H.S., Ishizaki, Y.,

Jacobson, M.D., 1993. Programmed cell death and the control of

cell survival: lessons from the nervous system. Science 262, 695–

700.

Royall, J.A., Ischiropoulos, H., 1993. Evaluation of 20,70-dichloroflu-

orescin and dihydrorhodamine123 as fluorescent probes for intra-

cellular H2O2 in cultured endothelial cells. Arch. Biochem.

Biophys. 302, 348–355.

Rybak, L.P., 1981. cis-Platinum associated hearing loss. J. Laryngol.

Otol. 95, 745–747.

Roberts, J.J., Thomson, A.J., 1979. The mechanism of action of

antitumor platinum compounds. Prog. Nucleic Acid Res. Mol.

Biol. 22, 71–133.

Schulze, A., Downward, J., 2001. Navigating gene expression using

microarrays – a technology review. Nat. Cell Biol., E190–E195.

Segal, A.W., Abo, A., 1993. The biochemical bases of the NADPH

oxidase in phagocytes. Trends Biochem. Sci. 18, 43–47.

Smith, P.K., Krohn, R.I., Hermanson, G.T., Mallia, A.K., Gartner,

F.H., Provenzano, M.D., Fujimoto, E.K., Goeke, N.M., Olson,

B.J., Klenk, D.C., 1985. Measurement of protein using bicinchon-

inic acid. Anal. Biochem. 150, 76–85, Erratum in Anal. Biochem.

1987 163, 279.

Sorenson, C.M., Barry, M.A., Eastman, A., 1990. Analysis of events

associated with cell cycle arrest at G2 phase and cell death induced

by CDDP. J. Natl. Cancer. Inst. 82, 749–755.

Stadnicki, S.W., Fleischman, R.W., Schaeppi, U., Merriam, P., 1975.

cis-Dichlorodiammineplatinum (II) (NSC-119875): hearing loss

and other toxic effects in rhesus monkeys. Cancer Chemother. Rep.

59, 467–480.

Steller, H., 1995. Mechanisms and genes of cellular suicide. Science

267, 1445–1449.

Wright, C.G., Schaefer, S.D., 1982. Inner ear histopathology

in patients treated with cis-platinum. Laryngoscope 92,

1408–1413.

Zamai, L., Canonico, B., Luchetti, F., Ferri, P., Melloni, E., Guidotti,

L., Cappellini, A., Cutroneo, G., Vitale, M., Papa, S., 2001.

Supravital exposure to propidium iodide identifies apoptosis in

adherent cells. Citometry 44, 57–64.