Efferent-mediated adaptation of the DPOAE as a predictor of aminoglycoside toxicity

www.elsevier.com/locate/heares

Hearing Research 201 (2005) 99–108

Efferent-mediated adaptation of the DPOAE as a predictorof aminoglycoside toxicity

Karin Halsey a, Asa Skjonsberg b,c, Mats Ulfendahl b,c, David F. Dolan a,*

a Department of Otolaryngology, Kresge Hearing Research Institute, 1301 East Ann Street, Ann Arbor, MI 48109-0506, USAb Center for Hearing and Communication Research, Karolinska Institutet, SE-171 76 Stockholm, Sweden

c Deptartment of Otolarygology, Karolinska Hospital, Solna, SE-171 76 Stockholm, Sweden

Received 22 June 2004; accepted 16 September 2004

Available online 30 October 2004

Abstract

Rapid efferent adaptation of the distortion product otoacoustic emission (DPOAE) predicts susceptibility to noise-induced dam-

age, and is linked to the concentration of the efferent receptor (a9). Maximum adaptation occurs at intense primary levels, rapidly

switching from positive to negative orientation in a very narrow (2 dB) range of F1 and F2 levels.

Aminoglycosides are commonly used antibiotics, with the undesirable side-effect of ototoxicity. Susceptibility to hair cell damage

from the aminoglycoside gentamicin can be quite variable, even within a single strain and species of animal. Since one of gentam-

icin�s first sites of action in the outer hair cell (OHC) is at the efferent receptor, it is possible that efferent activity could be a predictor

of susceptibility to gentamicin induced damage.

Significant sex-related differences were found in two strains of guinea pigs when treated with gentamicin. Female guinea pigs were

more susceptible both to systemic effects and to specific ototoxic effects.

Efferent-mediated DPOAE adaptation served as a predictor of sensitivity to aminoglycoside damage, predicting both number of

days before onset of deafness in male animals, and predicting final threshold shifts from gentamicin doses which produced variable

results.

� 2004 Elsevier B.V. All rights reserved.

Keywords: Hearing; Gender; Aminoglycoside; Efferent; Adaptation; DPOAE

1. Introduction

Aminoglycosides are commonly used antibiotics,

with the undesirable side-effect of ototoxicity. Suscepti-

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

doi:10.1016/j.heares.2004.09.010

Abbreviations: ABR, auditory brainstem response; DPOAE, dis-

tortion product oto-acoustic emission; dB, decibel; SPL, sound pres-

sure level; OHC, outer hair cell; L1, level of frequency one; L2, level of

frequency two; F1, frequency one; F2, frequency two; SPF, specific

pathogen free; TDT, Tucker–Davis technologies; FFT, fast Fourier

transform* Corresponding author. Tel.: +1 734 7639704/7648110; fax: +1 734

6158111/7640014.

E-mail address: [email protected] (D.F. Dolan).

bility to hair cell damage from the aminoglycoside gen-

tamicin can be quite variable, even within a single strain

and species of animal (for example, see Imamura and

Adams, 2003). Since one of gentamicin�s first sites of ac-tion in the outer hair cell (OHC) is at the efferent recep-tor (Blanchet et al., 2000), it is possible that efferent

activity could be a predictor of susceptibility to gentam-

icin induced damage.

Efferent-mediated adaptation of the distortion prod-

uct otoacoustic emission (DPOAE) has been demon-

strated in cats (Liberman et al., 1996), guinea pigs

(Kujawa and Liberman, 2001), rabbits (Luebke et al.,

2002), mice (Sun and Kim, 1999), and humans (Kim

mailto:[email protected]

100 K. Halsey et al. / Hearing Research 201 (2005) 99–108

et al., 2001). It is a non-invasive measure of efferent

function that has been shown to be a predictor of sen-

sitivity to noise-induced trauma (Maison and Liber-

man, 2000). We undertook a project to determine

whether it could also serve as a predictor for aminogly-

coside-induced hearing loss. Maximum change ofDPOAE levels with efferent activation is not the only

indicator of efferent function; a measure of the rate

of efferent activation can be obtained by measuring

the time constant of the DPOAE timetrace. For con-

sistency, rate of adaptation was measured for the pri-

mary tone level conditions (L1/L2) conditions that

induced the maximum positive adaptation.

Our early results were confounded by what we even-tually discovered to be sex-related differences in how

guinea pigs reacted to gentamicin. Female animals were

more sensitive; not only to ototoxic effects, but also to

adverse systemic effects impacting animal health. Sex-re-

lated differences in response to aminoglycoside treat-

ments have been reported before, primarily in rats, but

with males being more sensitive to adverse effects (Mills

et al., 1999; Goodrich and Hottendorf, 1995).For both male and female animals, for gentamicin

dosages causing moderate threshold shifts, efferent-me-

diated DPOAE adaptation served as a predictor of sen-

sitivity to aminoglycoside-induced threshold shifts. In

males, it also predicted the number of days an animal

could be dosed before developing a hearing loss.

Fig. 1. Demonstration of adaptation calculation: (a) adaptation

strength for a given response curve is defined as the DPOAE amplitude

at steady state (average of four adjacent points) subtracted from the

DPOAE amplitude at the onset of the primary tones. (b) Adaptation

plotted as a function of L2, for L1 held at 88 dB SPL.

2. Materials and methods

Specific pathogen free (SPF) male and female

pigmented outbred guinea pigs (initial body weight

200–250 g) were purchased from Elm Hill Breeding Lab-

oratories (Chelmsford, MA, USA). They were housed in

a traditional SPF room in individual polycarbonate

cages with free access to water and standard guineapig diet (PMI Nutrition International, Brentwood,

MO, USA). Additionally, pigmented guinea pigs were

obtained from Karolinska Institute (Stockholm,

Sweden) as part of a companion study, and data on

these Swedish wildtype animals, submitted to an identi-

cal protocol, are also presented here.

2.1. Auditory brainstem responses

ABRs were recorded in an electrically and acousti-

cally shielded chamber (Acoustic Systems, Austin, TX,

USA). Animals were anesthetized with ketamine 58.8

mg/kg (Fort Dodge Animal Health), xylazine 2.4 mg/kg

(Ben Venue Laboratories) and acepromazine 1.2 mg/kg

(Boehringer Ingelheim Vetmedica, Inc.), and body tem-

perature was maintained with heating pads and heatlamps. At the time of the baseline ABR, chronically

indwelling recording electrodes were aseptically im-

planted in the skull at vertex (1 cm posterior to bregma),

reference (1 cm lateral to bregma, ipsilateral to the test

ear) and ground (2 cm anterior to bregma) sites. These

K. Halsey et al. / Hearing Research 201 (2005) 99–108 101

electrodes were used for response recording in all ABRs.

Tucker–Davis Technologies (TDT) System II hardware

and SigGen/Biosig software (TDT, Alachua, FL,

USA) were used to present the stimulus and record re-

sponses. Tones were delivered through a Beyer driver

(Beyer Dynamic Inc., Farmingdale, NY; Aluminum-shielded enclosure made in house), with the speculum

placed just inside the tragus. Stimulus presentation

was 15 ms tone bursts, with 1 ms rise/fall times, pre-

sented 10/s. Up to 1024 responses were averaged for

each stimulus level. Responses were collected for stimu-

lus levels in 10 dB steps at higher stimulus levels, with

additional 5 dB steps near threshold. Thresholds were

interpolated between the lowest stimulus level where aresponse was observed, and 5 dB lower, where no re-

Fig. 2. (a) Adaptation plotted in a three-dimensional ‘‘grid’’ over an

L1/L2 plane, from data collected in 1 dB steps. This data provides

information about the location of peak adaptation in the L1/L2 plane,

but the 1 dB resolution is insufficient to capture the detail of

the adaptation curve, as demonstrated by (b). (b) Demonstrates the

importance of step-size in measuring adaptation magnitude. The

overlaid adaptation in the 0.4 dB ‘‘grid’’ results in a maximum

adaptation measure of 41.6 dB; 28.8 dB greater than measured using

the 1.0 dB ‘‘grid’’ in (a).

sponse was observed. Baseline and final (one week after

the termination of dosing) ABRs were tested at 2, 8 and

16 kHz. In addition, daily 16 kHz ABR screenings were

monitored in awake, gently restrained animals from the

initiation of dosing until a sustained threshold shift was

observed.

2.2. Efferent mediated adaptation

This response was recorded prior to initiation of gen-

tamicin dosing. Animals were anesthetized as described

above. The stimuli were generated and the response data

collected using TDT System II hardware and a MAT-

LABTM script written in-house. Stimulus tones F1

(8000 Hz) and F2 (9600 Hz) were presented with a F2/

F1 = 1.2 ratio, and the distortion product (2F1�F2) re-

corded at 6400 Hz. Responses for 12–14 F1 levels and

12 F2 levels for each F1 level (for a total of 144–168

L1/L2 combinations) were measured. For each level

combination, primaries were presented in 1 s bursts,

with 10 ms on and off ramps. A 2 s pause followed every

primary tone presentation. Responses were collectedwith a sampling rate of 50 kHz. A fast Fourier trans-

form (FFT) was performed on the response waveform

with an analysis window of 25 ms, and the sound level

of the distortion product was obtained for each window.

Responses to four 1 s identical stimulus presentations

were averaged.

2.3. Gentamicin administration

Subcutaneous injections of gentamicin (American

Pharmaceutical Partners, Inc., Schaumberg, IL, USA)

Fig. 3. A single-exponential curve is used to fit the curve exhibiting

maximum adaptation, to solve for the time constant s: f(t) = f0 + ae�t/s

where f0 is the DPOAE level steady state, a is the adaptation

magnitude, and s is the time constant.


were given once daily for 14 consecutive days at 160,

145, 130, 120, or 100 mg/kg dosages. The animals�body weight and condition were monitored and sup-

portive care (nutritional supplements given orally [Nu-

tri-Cal, EVSCO Pharmaceuticals] and physiological

saline injections subcutaneously [Abbott Laboratories])administered if necessary to maintain animal condition

if the animals lost weight or were noticeably

dehydrated.

On the 10th day of dosing, a (0.75–1.0) ml blood sam-

ple was drawn from the saphenous vein, and a blood ser-

um chemistry panel was submitted for animals in the

100, 130 and 145 mg/kg dosage groups, along with sam-

ples from control (undosed) animals of the same strain.The assays were performed by the Unit for Laboratory

Fig. 4. Mean and standard deviation of baseline adaptation magnitude (a), ti

(c) for each treatment group. There were no significant differences among gr

Animal Medicine�s Animal Diagnostic Laboratory

(University of Michigan, Ann Arbor, MI, USA) using

a VetTest blood chemistry analyzer (IDEXX Laborato-

ries, Inc.). A general health panel was evaluated that

included albumin, alkaline phosphatase, alanine amino-

transferase, amylase, calcium, cholesterol, creatinine,glucose, blood urea nitrogen, total bilirubin, total pro-

tein, inorganic phosphate, and globulin.

2.4. Analysis

Group comparisons were student unpaired t-tests,

performed in Excel. Linear regression and exponential

curve-fitting analysis was done using SigmaPlotsoftware.

me constant of maximum positive adaptation (b), and ABR thresholds

oups.

Table 1

Differences between males and females in response to gentamicin are

clear in the survival rate of animals on study

Animal group Number of animals

starting protocol

Number of animals

completing protocol

Females 160 mg/kg 11 0

Males 160 mg/kg 15 4a


Males 145 mg/kg 9 9


Males 130 mg/kg 7 7


Female wildtypes 3 3

Male wildtypes 4 4

A large percentage of females at higher dosages did not complete the

protocol. They either died suddenly from what may have been paral-

ysis due to neuromuscular blockade, shock, or they had to be eutha-

nized as a result of deteriorating condition due to apparent renal

failure.a Due to the high mortality rate of the 160 mg/kg group, dosing was

terminated early for seven males. So, while they were not eliminated

due to health reasons, they did not complete the protocol�s 14 days of

dosing.


All procedures were reviewed and approved by the

University of Michigan�s University Committee on the

Use and Care of Animals.

Fig. 5. When bloodwork was assayed for measures included in the general

animals at the same dosage group (where applicable) and of non-dosed co

creatinine (b), which are indicators of poor renal function, and of amylase (e)

dosed animals had lowered levels of alkaline phosphatase (f), a test in which el

be a sign of vitamin C deficiency, or possibly of stress. There were also some

alanine aminotransferase (g), and calcium (d).

3. Results

3.1. Efferent-mediated DPOAE adaptation

Efferent-mediated fast adaptation was visible as a

change in 2F1�F2 intensity over approximately the first300–500 ms of the DPOAE response (Fig. 1(a)). Adap-

tation strength for any given L1 and L2 combination

was defined as the DPOAE level at steady state sub-

tracted from the DPOAE level at the onset of the

primary tones (Fig. 1(b)). An example of a three-

dimensional plot of the adaptation strength over the

‘‘grid’’ of one-dB L1/L2 steps is shown in Fig. 2(a),

showing the stereotyped adaptation curve. A fine ‘‘grid’’of 0.4 dB steps was collected and overlayed to get a

more detailed and accurate measure of maximum adap-

tation (Fig. 2(b)). Initially, we used a resolution of 0.2

dB steps, but found the 0.4 dB resolution provided very

similar results for maximum adaptation. For each ani-

mal, the locations of the positive and negative adapta-

tion peaks in the L1/L2 plane were noted, and each

animal�s adaptation magnitude was defined as the larg-est ‘‘negative adaptation’’ subtracted from the largest

‘‘positive adaptation.’’ A single exponential function

was used to model the maximum positive adaptation

health profile, female guinea pigs had levels differing from both male

ntrols. Most notably, they had elevated blood urea nitrogen (a) and

, which is an indicator of impaired pancreatic function. All gentamicin-

evated levels indicate poor liver function, and lowered blood levels may

significant differences found in blood levels of inorganic phosphate (c),

Fig. 6. In the groups where there were sufficient surviving females for

comparison to identically treated males (145 mg/kg and the Swedish

wildtypes), females had significantly higher threshold shifts. In most

cases at those frequencies, females had thresholds high enough to

exceed the limits of our system (105 dB SPL).


curves and solve for the time constant s, as shown in the

example in Fig. 3. Adaptation magnitude was similar in

all groups, with the male and female Swedish wildtypes

having slightly higher values (but statistically insignifi-

cant, Fig. 4(a)). Time constants for the maximum adap-

tation curves were likewise similar in all groups (Fig.4(b)).

3.2. Baseline ABRs

Baseline ABR thresholds were very similar for all

groups, with the Swedish wildtypes having a slight

(but statistically insignificant) trend towards lower

ABR thresholds (Fig. 4(c)).

3.3. Sex-related differences in response to gentamicin

With the first group of animals tested, dosed at 160

mg/kg, problems immediately became apparent. A large

number of animals died acutely, with little warning, with

symptoms consistent with aminoglycoside-induced neu-

romuscular blockade paralysis (Sokoll and Gergis,1981). All females died, and some of the males died as

well. Animals surviving for longer periods of time also

began showing symptoms of renal toxicity including

anorexia, dehydration, and deteriorating physical condi-

tion. That dosing trial was ended early due to the high

mortality rate, with only four males completing the en-

tire protocol. We suspected there were sex-related differ-

ences at this time, but confirmed the differences withlater trials at lower dosages. Table 1 summarizes the sur-

vival rates of all groups of animals. When bloodwork

was assayed for measures included in the general health

profile, female guinea pigs had levels differing from both

male animals at the same dosage group (when applica-

ble) and from controls. The results are summarized in

Fig. 5. Notably, all gentamicin-dosed animals had ele-

vated blood urea nitrogen (145 mg/kg females vs. femalecontrols p = 0.006, 130 mg/kg females vs. female con-

trols p = 0.001, 100 mg/kg females vs. female controls

p < 0.001, 145 mg/kg males vs. male controls

p = 0.005, 130 mg/kg males vs. male controls was not

significant, Fig. 5(a)) with females more affected than

males at the same dose (145 mg/kg p = 0.013, 130

mg/kg p = 0.001). Creatinine levels were similarly af-

fected (145 mg/kg females vs. female controls p =0.002, 130 mg/kg females vs. female controls p < 0.001,

100 mg/kg females vs. female controls p < 0.001, 145

mg/kg males and 130 mg/kg males vs. male controls

were not significant, Fig. 5(b)) with females more

affected than males at the same dose (145 mg/kg

p = 0.002, 130 mg/kg p < 0.001). Elevated blood urea

nitrogen and creatinine are indicators of poor renal

function. Amylase, which is an indicator of impairedpancreatic function, was also elevated in gentamicin-

dosed animals (145 mg/kg females vs. female controls

p = 0.001, 130 mg/kg females vs. female controls

p < 0.001, 100 mg/kg females vs. female controls

p = 0.009, 145 mg/kg males and 130 mg/kg males vs.

control males were not significant, Fig. 5(e)) with fe-

males more affected than males at the same dose (145

mg/kg p = 0.004, 130 mg/kg p < 0.001). All gentamicin-dosed animals had lowered levels of alkaline phospha-

tase (145 mg/kg females vs. female controls p = 0.023,

130 mg/kg females vs. female controls p = 0.016, 100

mg/kg females vs. female controls p = 0.033, 145 mg/

kg males vs. male controls p = 0.038, 130 mg/kg males

vs. male controls was not significant, Fig. 5(f)) with fe-

males more affected than males at the same dose (145

mg/kg p = 0.007, 130 mg/kg not significant). Alkalinephosphatase is a test in which elevated levels indicate

poor liver function, but lowered blood levels may be a

sign of vitamin C deficiency (Mahmoodian et al.,

1996) or possibly of stress (Degkwitz, 1982). Since, like

humans, guinea pigs do not manufacture vitamin C, it

is necessary to supply it through their diet. Animals on

gentamicin regimes may not have been eating normally,

and therefore not getting sufficient quantities of vitaminC. There were also some significant differences found in

blood levels of inorganic phosphate (Fig. 5(c), males 145

mg/kg vs. control males p = 0.024, males 130 mg/kg vs.

control males p = 0.005, females 100 mg/kg vs. control


females p = 0.005), alanine aminotransferase (Fig. 5(g),

males 130 mg/kg vs. control males p = 0.034), and cal-

cium (Fig. 5(d), males 145 mg/kg vs. control males

p = 0.021, females 100 mg/kg vs. control females

p = 0.014), but these results were not consistent in dosed

vs. undosed groups or in male vs. female groups, and itwas harder to draw generalizations from the data. No

significant differences were found in the albumin, choles-

terol, glucose, total bilirubin, total protein, and globulin

assays (data not shown). There were no significant dif-

ferences in any of the assays between male and female

control animals. The blood samples producing these re-

sults were drawn on dosing day 10, which is prior to on-

set of hearing loss in most animals, and before themajority of animals showed any overt signs of illness.

In another observed sex-related difference, the female

animals that did complete the protocol (the 145 mg/kg

group and the Swedish wildtype females) had signifi-

cantly higher thresholds than male animals (145 mg/kg

females vs. males: 2 kHz p = 0.016, 8 kHz p = 0.003,

16 kHz p < 0.001, Swedish wildtype females vs. males:

2 kHz p = 0.005, 8 kHz p = 0.050, 16 kHz p = 0.017,Fig. 6). Due to the significantly different responses of fe-

males, and the poor survival rates of females at higher

dosages, only data from male animals will be shown

Fig. 7. Gentamicin-treated guinea pigs showed dose-dependent eleva-

tions in thresholds, with the greatest variability at 2 and 8 kHz.

for the 130 and 145 mg/kg dosage groups, and female

data will be shown only for the 100 mg/kg dosage for

the analysis in the rest of this paper.

3.4. Responses to gentamicin

Animals showed the expected dose-related elevations

in ABR thresholds measured one week after the termi-

nation of gentamicin dosing (Fig. 7). While the Swedish

wildtype males may not be directly comparable to the

other animals in this figure due to background strain dif-

ferences (Sinswat et al., 2000), they are included for

Fig. 8. (a) Daily ABR screenings indicated a rapid rise in 16 kHz

threshold around or shortly after dosing day 10, occurring earliest and

most rapidly in the 160 mg/kg group. (b) Demonstrates the greatly

reduced variability in deafness onset in the 160 mg/kg group (black)

compared to other groups. The 100 mg/kg female group had a

statistically different day of deafness when compared to the male

animals dosed at 160 mg/kg (p < 0.001) and the male animals dosed at

145 mg/kg (p = 0.002).


reference. Daily ABR screenings show that threshold

elevations began around or shortly after day 10, and

are most rapid in the highest dosage group (Fig. 8(a)).

For each animal, the day of deafness onset was defined

as the first day a sustained threshold shift from baseline

of at least 20 dB was observed (Fig. 8(b)). The 100 mg/kg female group had a statistically different day of deaf-

ness when compared to the male animals dosed at 160

mg/kg (p < 0.001) and the male animals dosed at 145

mg/kg (p = 0.002). None of the other group differences

were statistically significant.

Fig. 9. The baseline measure of adaptation magnitude, in select groups,

gentamicin. In general, animals with a larger efferent effect were more sensit

mg/kg R = 0.22, Swedish wildtypes R = 0.8, 100 mg/kg females R = 0.29. (b

R = 0.98, which was statistically significant p = 0.02, 100 mg/kg females R =

wildtypes R = 0.89, 100 mg/kg females R = 0.27. Swedish wildtype animals sh

significant at 8 kHz (p = 0.02). 160 mg/kg (data not shown) and 145 mg/kg

variability in the thresholds shifts.

3.5. Threshold shifts

Adaptation magnitude, measured at baseline, was in

some cases correlated to final post-gentamicin ABR

thresholds (Fig. 9(a)–(c)). This correlation was seen in

the lower dosage groups at 2 kHz (Fig. 9(a): 2 kHz:130 mg/kg R = 0.13, 145 mg/kg R = 0.22, Swedish wild-

types R = 0.8, 100 mg/kg females R = 0.29), 8 kHz (Fig.

9(b): 8 kHz: 130 mg/kg R = 0.61, 145 mg/kg R = 0.12,

Swedish wildtypes R = 0.98, which was statistically sig-

nificant p = 0.02, 100 mg/kg females R = 0.57), and 16

predicted threshold shifts resulting from systemic administration of

ive to ototoxic gentamicin effects. (a) 2 kHz: 130 mg/kg R = 0.13, 145

) 8 kHz: 130 mg/kg R = 0.61, 145 mg/kg R = 0.12, Swedish wildtypes

0.57. (c) 16 kHz: 130 mg/kg R = 0.45, 145 mg/kg R = 0.24, Swedish

owed a positive correlation at all frequencies, and the correlation was

had poor correlations especially at 16 kHz where there was very little


kHz (Fig. 9(c): 16 kHz: 130 mg/kg R = 0.45, 145 mg/kg

R = 0.24, Swedish wildtypes R = 0.89, 100 mg/kg fe-

males R = 0.27), but not in the highest dosage group

of 160 mg/kg, where final ABR threshold shifts were

tightly grouped for the four animals completing the dos-

ing protocol (regressions not shown, but see Fig. 7). Thistight grouping of threshold shifts around 50–65 dB (and

therefore poor correlation) was also apparent in the 145

mg/kg group at 8 and 16 kHz, and may indicate regions

of the cochlea where most or all outer hair cells are miss-

ing or damaged by the ototoxic drug (Dallos and Harris,

1978). Otherwise, correlations were strongest at 8 kHz

(and to some extent 2 kHz), where final thresholds were

more variable than at 16 kHz for most groups.

3.6. Time constants

Time constant of the maximum adaptation curve was

not a predictor of threshold shifts or of the animals� dayof deafness (data not shown).

3.7. Deafness onset

Deafness onset was inversely correlated with adapta-

tion magnitude (Fig. 10) in the 145 mg/kg male animals

(R = 0.44) and 130 mg/kg male animals (R = 0.43) dos-

age groups. Animals with larger adaptation magnitude

became deaf earlier than animals with smaller adapta-

tion magnitude. At 160 mg/kg, animals that completed

the protocol (n = 4) had such tightly grouped deafnessonsets (regression data not shown, but see Fig. 8(b)) that

no correlation was seen. Only two of the four male

Swedish wildtype animals met the criteria for deafness

Fig. 10. Using a ‘‘day of deafness’’ criteria of a sustained threshold

shift of 20 dB or greater, efferent-mediated DPOAE adaptation

magnitude was inversely correlated with the onset of deafness in male

animals dosed at 130 (R = 0.43) and 145 mg/kg (R = 0.44). Animals

with greater magnitude developed a threshold shift earlier than

animals with smaller adaptation magnitude.

onset, so deafness onset correlation data was not plotted

for that group. A correlation was not seen in the female

group dosed at 100 mg/kg.

4. Discussion

There were significant sex-related differences in re-

sponses to gentamicin in guinea pigs. Female guinea pigs

were less likely to complete the protocol (higher mortal-

ity rates or euthanasia rates due to illness) at the same

dose, they had significantly different blood chemistry

levels from male animals and from controls, and female

animals that did complete the protocol had significantlylarger ABR threshold shifts than male animals at the

same dosage (when applicable). These results differ from

published results from other species (primarily rats),

where there was either no difference between sexes, or

males were found to be more susceptible (Mills et al.,

1999; Goodrich and Hottendorf, 1995).

For male animals, adaptation magnitude served as a

predictor of the number of days an animal could bedosed with gentamicin before developing a hearing loss

and of their final thresholds. Male animals with larger

adaptation magnitude tended to become deaf earlier

than animals with smaller adaptation magnitude. This

correlation was fairly weak, and was not seen in female

animals.

Animals with larger adaptation magnitudes tended to

have larger threshold shifts. This effect became morepronounced at an even lower dosage of gentamicin,

where there was a great deal more variability among ani-

mals. Logically, correlations were poor if threshold

shifts were clustered around 0 dB (for example wildtype

males and the male 130 mg/kg dosage group at 2 kHz)

indicating that that region of the cochlea was unaffected

by the aminoglycoside treatment, or if threshold shifts

were clustered around 55–60 dB (for example the 16kHz shifts for the 145 and 130 mg/kg groups, or the

145 mg/kg group at 8 kHz), indicating most or all

OHC in that region were affected (Dallos and Harris,

1978). In dosages where a broad range of threshold

shifts were observed, correlations improved. While

trends were consistent in all groups, and especially

strong at 8 kHz (the tested frequency closest to the fre-

quency where efferent-mediated DPOAE adaptationwas measured), the only statistically significant correla-

tion occurs at 8 kHz in the Swedish wildtype male group

(p = 0.02).

Since adaptation magnitude has previously been cor-

related with number of efferent receptors at the OHC

(Luebke and Foster, 2002), and one of the first sites of

action of gentamicin in the hair cell is at the efferent

receptor (Blanchet et al., 2000) blocking its function(Lima da Costa et al., 1997; Smith et al., 1994; Yoshida

et al., 1999), it may be that cells with a greater number


of efferent terminals are more susceptible to damage

from the aminoglycoside. This is also consistent with

the finding that animals that have been surgically de-

efferented are less susceptible to aminoglycoside damage

(Capps and Duvall, 1977). Whether these conclusions

can be generalized to humans remains to be seen. It isapparent that aminoglycoside ototoxicity is a multi-

component process (Rothlin et al., 2000; Wu et al.,

2002), and no single factor is likely to account for all

the variability in sensitivity seen in a given population.

Acknowledgements

This research was supported by RO1 DC004194, P30

DC05188 and PO1 DC00078. The authors wish to thank

Lisa Kabara and Laura Grant for invaluable assistance

with animal care and data collection.

References

Blanchet, C., Erostegui, C., Sugasawa, M., Dulon, D., 2000. Gentam-

icin blocks Ach-evoked K+ current in guinea-pig outer hair cells by

impairing Ca2+ entry at the cholinergic receptor. Journal of

Physiology 525 (3), 641–654.

Capps, M.J., Duvall, A.J., 1977. Ototoxicity and the olivocochlear

bundle. Laryngoscope 87, 1100–1108.

Dallos, P., Harris, D., 1978. Properties of auditory nerve responses in

absence of outer hair cells. Journal of Neurophysiology 41, 365–

383.

Degkwitz, E., 1982. Activity of alkaline phosphatase in the serum of

normal and ascorbic-acid-deficient guinea pigs. Zeitschrift fur

Ernahrungswissenschaft 21 (1), 51–56.

Goodrich, J.A., Hottendorf, G.H., 1995. Tobramycin gender-related

nephrotoxicity in Fisher but not Sprague–Dawley rats. Toxicology

Letters 75, 127–131.

Imamura, S., Adams, J., 2003. Distribution of gentamicin in the guinea

pig inner ear after local or systemic application. JARO 4 (2), 176–

195.

Kim, D.O., Dorn, P.A., Neely, S.T., Gorga, M.P., 2001. Adaptation of

distortion product otoacoustic emission in humans. JARO 2 (1),

31–40.

Kujawa, S.G., Liberman, M.C., 2001. Effects of olivocochlear

feedback on distortion product otoacoustic emissions in guinea

pig. JARO 2 (3), 268–278.

Liberman, M.C., Puria, S., Guinan, J.J., 1996. The ipsilaterally evoked

olivocochlear reflex causes rapid adaptation of the 2f1�f2 distor-

tion product otoacoustic emission. JASA 99 (6), 3572–3584.

Lima da Costa, D., Chibois, A., Erre, D.P., Blanchet, C., Charlet de

Sauvage, R., Aran, J.M., 1997. Fast, slow, and steady-state effects

of contralateral acoustic activation of the medial olivocochlear

efferent system in awake guinea pigs: action of gentamicin. Journal

of Neurophysiology 78 (4), 1826–1836.

Luebke, A.E., Foster, P.K., 2002. Variation in inter-animal suscepti-

bility to noise is associated with a9 acetylcholine receptor subunit

expression level. Journal of Neuroscience 22 (10), 4241–4247.

Luebke, A.E., Foster, P.K., Stagner, B.B., 2002. A multifrequency

method for determining cochlear efferent activity. JARO 3 (1), 16–

25.

Mahmoodian, F., Gosiewska, A., Peterkofsky, B., 1996. Regulation

and properties of bone alkaline phosphatase during vitamin C

deficiency in guinea pigs. Archives of Biochemistry & Biophysics

336 (1), 86–96.

Maison, S.F., Liberman, M.C., 2000. Predicting vulnerability to

acoustic injury with a noninvasive assay of olivocochlear reflex

strength. Journal of Neuroscience 20 (12), 4701–4707.

Mills, C.D., Loos, B.M., Henley, C.M., 1999. Increased susceptibility

of male rats to kanamycin-induced cochleotoxicity. Hearing

Research 128 (1–2), 75–79.

Rothlin, C.V., Katz, E., Verbitsky, M., Vetter, D.E., Heinemann, S.F.,

Elgoyhen, A.B., 2000. Block of the a9 nicotinic receptor by

ototoxic aminoglycosides. Neuropharmacology 39, 2525–2532.

Sinswat, P., Wu, W.J., Sha, S.H., Schacht, J., 2000. Protection from

ototoxicity of intraperitoneal gentamicin in guinea pig. Kidney

International 58 (6), 2525–2532.

Smith, D.W., Erre, J.P., Aran, J.M., 1994. Rapid reversible elimination

of medial olivocochlear efferent function following single injections

of gentamicin in the guinea pig. Brain Research 625 (2), 243–248.

Sokoll, M.D., Gergis, S.D., 1981. Antibiotics and neuromuscular

function. Anesthesiology 55, 148–159.

Sun, X.-M., Kim, D.O., 1999. Adaptation of 2F1�F2 distortion

product otoacoustic emission in young-adult and old CBA and C57

mice. JASA 105, 3399–3409.

Wu, W.J., Sha, S.H., Schacht, J., 2002. Recent advances in under-

standing aminoglycoside ototoxicity and its prevention. Audiology

& Neuro-Otology 7 (3), 171–174.

Yoshida, N., Liberman, M.C., Brown, M.C., Sewell, W.F., 1999.

Gentamicin blocks both fast and slow effects of olivocochlear

activation in anesthetized guinea pigs. Journal of Neurophysiology

82 (6), 3168–3174.

Efferent-mediated adaptation of the DPOAE as a predictor of aminoglycoside toxicity

Documents