reduced expression of sialoglycoconjugates in the outer hair cell glycocalyx after systemic...

www.elsevier.com/locate/heares

Hearing Research 205 (2005) 68–82

Reduced expression of sialoglycoconjugates in the outer haircell glycocalyx after systemic aminoglycoside administration

J.C.M.J. de Groot *, E.G.J. Hendriksen, G.F. Smoorenburg

Hearing Research Laboratories, Department of Otorhinolaryngology, University Medical Center Utrecht,

Room G.02.531, P.O. Box 85.500, NL-3508 GA Utrecht, The Netherlands

Received 2 May 2004; accepted 3 March 2005

Available online 5 May 2005

Abstract

In this study we investigated the effect of systemic aminoglycoside administration on the expression of sialoglycoconjugates in the

outer hair cell (OHC) glycocalyx of the adult guinea pig. Sialoglycoconjugates were visualized by means of ultrastructural lectin

cytochemistry, using Limax flavus agglutinin (LFA) and wheat germ agglutinin (WGA) as probes. Labelling densities were deter-

mined for the apical membranes (including the stereocilia and stereociliary cross-links) and basolateral membranes of OHCs in

the respective (basal, middle and apical) cochlear turns from animals that had been treated with gentamicin or neomycin for 5

or 15 consecutive days. Our results indicate that: (1) sialoglycoconjugate expression in the OHC glycocalyx demonstrates an intra-

cochlear gradient decreasing towards the apical turn; (2) OHCs demonstrate a polarity in sialoglycoconjugate expression, in that the

basolateral membranes contain more sialoglycoconjugates per surface area than the apical membranes; (3) aminoglycoside admin-

istration results in reduced expression of sialoglycoconjugates in the OHC glycocalyx; in this respect, basal-turn OHCs are more

susceptible than those in the middle and apical turns; (4) reduction in sialoglycoconjugate expression after aminoglycoside admin-

istration is more prominent in the basolateral membranes; and (5) the difference in ototoxic potencies between gentamicin and neo-

mycin is not reflected at the level of sialoglycoconjugate expression. The present data support our earlier hypothesis that

aminoglycosides, already at an early phase of intoxication, interfere with the function of the endoplasmic reticulum and/or the Golgi

apparatus, implying that these organelles play a crucial role in the initial phase of aminoglycoside-induced OHC degeneration.

� 2005 Elsevier B.V. All rights reserved.

Keywords: Cochlea; Outer hair cells; Glycocalyx sialoglycoconjugates; Aminoglycoside ototoxicity; Gentamicin; Neomycin

1. Introduction

Cochleotoxicity is one of the most prevalent adverse

effects that are encountered during or after prolonged

treatment with aminoglycoside antibiotics. Clinically,

0378-5955/$ - see front matter � 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.heares.2005.03.002

Abbreviations: GlcNAc, N-acetyl-D-glucosamine; LD, labelling

density; LFA, Limax flavus agglutinin; NAcGlucase, b-N-acetylglucos-

aminidase; NANA, N-acetylneuraminic acid; NANase, neuraminidase;

OHC, outer hair cell; PBS, phosphate-buffered saline; SEM, standard

error of the mean; WGA, wheat germ agglutinin* Corresponding author. Tel.: +31 30 2506481; fax: +31 30 2541922.

E-mail address: [email protected] (J.C.M.J. de

Groot).

aminoglycoside-induced cochleotoxicity is manifestedas a permanent sensorineural hearing loss with an initial

high-frequency deficit that may continue to progress to

the lower frequencies. Hearing loss results from a pro-

gressive and irreversible loss of cochlear hair cells and

subsequent degeneration of the organ of Corti.

The histopathological lesions observed in the organ of

Corti following aminoglycoside administration have been

well documented in variousmammalian species:Morpho-logical studies have demonstrated that the outer hair cells

(OHCs) are more vulnerable than the inner hair cells, and

that the OHCs show signs of degeneration well before

morphological changes are observed in any other cell

mailto:[email protected]

J.C.M.J. de Groot et al. / Hearing Research 205 (2005) 68–82 69

type, within or outside the organ of Corti (for reviews, see

Lim, 1986; Govaerts et al., 1990; Garetz and Schacht,

1996; Forge andSchacht, 2000).OHCs undergo a number

of distinct morphological alterations in response to

chronic aminoglycoside administration. An increase in

the number of secondary lysosomes in the infracuticularregion, vesiculation and dilatation of the subsurface cis-

ternae of the endoplasmic reticulum as well as the Golgi

saccules, formation of Hensen�s bodies, and disorganiza-

tion of the glycocalyx lining the apical and basolateral

membranes are changes that are ultrastructurally ob-

served in OHCs during an early phase of intoxication

(Lim, 1986; De Groot and Veldman, 1988; De Groot

et al., 1991).The cellular mechanisms underlying the cytotoxic ef-

fect of aminoglycosides are incompletely understood.

Since the effects aminoglycosides have on the metabolic

pathways in OHCs are so manifold (cf., Lim, 1986;

Rybak, 1986; Govaerts et al., 1990; Garetz and Schacht,

1996; Forge and Schacht, 2000), the precise trigger for

(outer) hair cell degeneration remains elusive. However,

it is now well established that aminoglycosides, aftertheir endocytotic uptake, are sequentially transported

to endosomes and then lysosomes via vesicular transport

(Lim, 1986; De Groot et al., 1990; Hiel et al., 1992;

Dulon et al., 1993; Fikes et al., 1994; Aran et al., 1995;

Hashino and Shero, 1995; Hashino et al., 1997; Steyger

et al., 2003). Also, it has been shown that cellular uptake

and lysosomal accumulation of aminoglycosides largely

precedes, even by days, the development of functionaland cellular damage in hair cells (Hiel et al., 1993; Dulon

et al., 1993; Hashino and Shero, 1995; Hashino et al.,

1997; Imamura and Adams, 2003a). It is thought that

this delay in the emergence of the drug�s cytotoxic effectis related to the storage capacity of the lysosomes, and

that lysosomal disruption with release of the aminogly-

coside molecules into the cytosol could be a direct trigger

for (outer) hair cell degeneration (Lim, 1986; De Grootet al., 1990; Hashino et al., 1997). Once spilt into the cyto-

sol, the drug could react with potential cytosolic targets

or interfere with crucial signalling pathways and, hence,

initiate the chain of molecular events that ultimately lead

to (outer) hair cell degeneration.Oxidative stress resulting

from an overproduction of reactive oxygen species

(Takumida et al., 1999; Sha and Schacht, 2000; Bertolaso

et al., 2001, 2003; Dehne et al., 2002), up-regulation ofcalcium-dependent proteases (Ding et al., 2002), and

activation of intracellular signalling pathways that are

involved in apoptosis (Ylikoski et al., 2002; Kalinec

et al., 2003; Bodmer et al., 2003) are but some of the

mechanisms that have recently been implicated in amino-

glycoside-induced (outer) hair cell degeneration.

However, this ‘‘lysosomal-accumulation-and-disrup-

tion’’ model cannot, in itself, explain the subcellularchanges that take place during an early phase of

intoxication. De Groot and Veldman (1988) observed

that the colloidal thorium reactivity of the glycocalyx

lining the apical and basolateral surfaces of the OHCs

is diminished as early as 24 h after the first application

of gentamicin, and is even completely abolished after

five consecutive days of chronic administration. At

this point of time, however, the lysosomes display anormal ultrastructural appearance and immunogold

labelling for gentamicin is not yet scattered through-

out the cytosol (De Groot et al., 1990; Hashino and

Shero, 1995; Hashino et al., 1997; Imamura and

Adams, 2003a). These observations suggest that ami-

noglycoside molecules, after their endocytotic uptake,

are not only targetted to the lysosomes but also to

other intracellular organelles, from where they caninterfere with hair cell physiology. Likely, but hitherto

ignored, candidates are the endoplasmic reticulum

and the Golgi apparatus, which play a pivotal role

in the de novo synthesis and subsequent modification

(e.g., glycosylation) of proteins and lipids that are

destined to the cell surface. The oligosaccharide side-

chains of these membrane glycoconjugates form the

glycocalyx which covers the basolateral and apical mem-branes, including the stereocilia and the stereociliary

cross-links.

In previous publications, we have attributed the loss

of colloidal thorium reactivity of the OHC glycocalyx

after aminoglycoside administration to aberrant syn-

thesis or glycosylation of membrane glycoconjugates

(De Groot and Veldman, 1988; De Groot et al.,

1990). It should be noted that aberrant glycosylationor altered expression patterns of cellular glycoconju-

gates are common findings in cells and tissues undergo-

ing pathological changes (Damjanov, 1987; Roth,

1993), and sialic acids are the most ubiquitous oligo-

saccharides in the glycocalyx of OHCs (Tachibana

et al., 1987, 1990; Gil-Loyzaga and Brownell, 1988;

Katori et al., 1996). Furthermore, it is evident from

enzymatic digestion experiments that colloidal thoriumreactivity of the OHC glycocalyx is mainly due to the

presence of sialic acid residues (Van Benthem et al.,

1992, 1993). In the light of these observations, it is log-

ical to hypothesize that aminoglycosides interfere with

the function of the endoplasmic reticulum and/or the

Golgi apparatus and, hence, with the expression of

sialoglycoconjugates in OHCs.

In order to verify this hypothesis, we have quanti-fied the effect of chronic systemic administration

of two aminoglycosides with different cochleotoxic

potencies (gentamicin and neomycin) upon OHC

glycocalyx reactivity for Limax flavus agglutinin

(LFA), which has a specific affinity for the sialic acid

N-acetylneuraminic acid, and wheat germ agglutinin

(WGA), which reacts with the monosaccharides N-

acetyl-D-glucosamine and N-acetylneuraminic acid.For this purpose, we used ultrathin cryosections of

microdissected organ of Corti samples taken from

70 J.C.M.J. de Groot et al. / Hearing Research 205 (2005) 68–82

three different locations (basal, middle and apical

turn) along the basilar membrane of the guinea pig

cochlea.

2. Materials and methods

2.1. Experimental groups

Adult, female albino guinea pigs (Dunkin-Hartley

strain; weight 250–300 g) were purchased from Harlan

Laboratories (Horst, the Netherlands). They were

housed in the Animal Care Facility of the Central

Laboratory Animal Institute of Utrecht University.Animals had free access to both food and water

and were kept under standard laboratory conditions.

All experimental protocols and procedures used in

this study were reviewed and approved by the Univer-

sity�s Committee on Animal Research (DEC-UMC

#89055). Animal care was under supervision of the

Central Laboratory Animal Institute of Utrecht

University.One group of animals (n = 4) was given daily intra-

peritoneal injections of a sterile aqueous solution of gen-

tamicin sulphate (Garamycin�; Schering-Plough BV,

Amstelveen, The Netherlands) at a dose of 100 mg/kg,

either for 5 (n = 2) or 15 consecutive days (n = 2). A sec-

ond group of animals (n = 4) received daily intraperito-

neal injections of a solution of neomycin sulphate

(Sigma, St. Louis, USA) in sterile physiological saline(pH 7.4) at a dose of 100 mg/kg, either for 5 (n = 2) or

15 consecutive days (n = 2). The animals were sacrificed

24 h after the last injection. A third group of animals

(n = 4) was not treated with aminoglycosides and served

as control group.

2.2. Tissue processing

Animals were anaesthesized with a sublethal dose

of sodium pentobarbital (Nembutal�) and decapitated.

The bullae were removed and the cochleas were

immediately fixed by intralabyrinthine perfusion with

2% glutaraldehyde in 0.1 M sodium cacodylate buffer

(pH 7.4), followed by immersion in the same fixative

for an additional 2 h at 4 �C. Postfixation with OsO4

was omitted to obviate a detrimental effect on lectin-binding sites. Cochleas were then rinsed twice for

15 min in 0.1 M sodium cacodylate buffer (pH 7.4)

and microdissected. After detaching the organ of Corti

from the modiolus and the lateral wall, the respective

(basal, middle and apical) cochlear turns were

separately collected and subdivided into smaller

segments.

Subdivided turns were decalcified overnight in 10%EDTA Æ 2Na (pH 7.4) at 4 �C, rinsed in 0.1 M sodium

cacodylate buffer (pH 7.4) and infiltrated with molten

gelatin through a graded series (1%, 2%, 5%, and 10%

in phosphate-buffered saline (PBS), pH 7.4, at 37 �C)followed by final embedding in molten gelatin (10% in

PBS, pH 7.4, at 37 �C) and solidification on ice. Solidi-

fied tissue blocks were trimmed to an appropriate size

and infused overnight with 1.6 M sucrose containing20% poly(vinylpyrrolidone) and 44 mM Na2CO3 at pH

7.4 (Tokuyasu, 1989), mounted on specimen stubs and

frozen in liquid nitrogen.

Ultrathin transverse cryosections of the organ of

Corti were cut with a Diatome� diamond knife in a

Reichert Ultracut-E ultracryotome. During cryosec-

tioning, the temperatures of the knife, sample and

cryochamber were all set at �100 �C, the cutting speedat 50 mm/s and the section thickness at �80 nm.

Cryosections were collected on a drop of 2.3 M

sucrose with a stainless-steel wire loop and, after

thawing at ambient temperature, transferred to piolo-

form-coated single-slot copper grids. Subsequently, the

grids were laid, face-down, on the surface of gelatin

plates and stored overnight in a humid incubation

chamber at 4 �C.

2.3. Ultrastructural lectin cytochemistry

Prior to on-grid staining, grids were retrieved by

gradually melting the gelatin plates. The grids were

then floated, face-down, on drops of PBS for 5 min

at 37 �C, to remove any remaining gelatin and su-

crose. All subsequent steps were carried out in a hu-mid incubation chamber at ambient temperature.

The grids were pre-incubated on drops of 0.1% acety-

lated bovine serum albumin (AurIon, Wageningen,

The Netherlands) in dilution buffer (PBS containing

0.1 mM MgCl2, 0.1 mM MnCl2, and 0.1 mM CaCl2)

for 5 min, in order to prevent non-specific staining.

Next, the grids were incubated for 2 h on drops of lec-

tin–colloidal gold (10 nm) conjugates (EY Laborato-ries, San Mateo, USA) diluted with dilution buffer

to a final lectin concentration of 1 lg/ml. Lectins used

were Limax flavus agglutinin (LFA; specific affinity for

sialic acids, i.e., N-acetylneuraminic acid) and wheat

germ agglutinin (WGA; specific affinities for N-ace-

tyl-D-glucosamine and N-acetylneuraminic acid). After

rinsing in PBS (2 · 5 min), the grids were floated on

drops of 2% glutaraldehyde in PBS for 15 min to se-cure the lectin–colloidal gold conjugates to the sec-

tions, and then rinsed in distilled water (2 · 5 min).

Finally, the sections were contrast-stained with an

alkaline mixture of oxalic acid and uranyl acetate –

resulting in a negative image of the membranes –

and embedded in a mixture of methyl cellulose and

uranyl acetate (Griffiths et al., 1984). The sections

were examined and photographed in a JEOL1200CX transmission electron microscope operating

at 80 kV.


2.4. Specificity and cytochemical controls

Two kinds of controls were performed. First, to test

the binding specificities of each lectin, ultrathin cryosec-

tions of the basal cochlear turn from non-treated ani-

mals were incubated with a solution consisting oflectin–colloidal gold conjugate (final lectin concentra-

tion: 1 lg/ml) pre-absorbed to the corresponding inhib-

itory monosaccharide (400 mM). N-acetylneuraminic

acid (Sigma, St. Louis, USA) and N-acetyl-D-glucosa-

mine (Sigma, St. Louis, USA) were used as inhibitory

monosaccharides.

Second, to differentiate WGA binding to N-acetyl-

D-glucosamine from binding to N-acetylneuraminicacid, ultrathin cryosections of basal, middle and apical

cochlear turns from non-treated animals were treated

with neuraminidase (type V, isolated from Clostridium

perfringens; Sigma, St. Louis, USA) and b-N-acetylglu-

cosaminidase (isolated from Canavalia ensiformis;

Sigma, St. Louis, USA) prior to incubation with the

lectin–colloidal gold conjugate. Enzymatic digestion

was performed by floating the grids on drops of theenzyme solutions for 45 min at ambient temperature.

2.5. Determination of labelling densities

To estimate the labelling density (i.e., number of gold

particles per surface area of membrane profile; LD),

gold particles were counted on electron micrographs

printed at a fixed magnification of 72,500·. A randomlyplaced single-lattice grid (test-line distance: 2 mm) was

used to evaluate the boundary length of gold-labelled

profiles of the plasma membrane. Labelling densities

were calculated by the formula:

LD ¼ Q½gold�=ðI � d � t½section�Þ;

where Q[gold] is the number of gold particles, I is thenumber of intersections, d is the actual distance

(27.5 nm) between the grid lines, and t[section] is the thick-

ness (80 nm) of the cryosection (Griffiths, 1993). Gold

particles were counted over the total boundary length

of the apical (including the stereocilia and stereociliary

cross-links) and basolateral membrane profiles of the

OHCs. We did not discriminate between OHC-1,

OHC-2, and OHC-3. These measurements were donefor each lectin (LFA and WGA), each cochlear turn (ba-

sal, middle and apical turns), and for each group (no

treatment, 5 days gentamicin, 15 days gentamicin, 5

days neomycin, and 15 days neomycin) as well as for

the specificity and cytochemical controls. Data of the

individual OHCs (n = 5–10) in each group were aver-

aged and the results expressed as the mean number of

gold particles/lm2 ± SEM. All data were statisticallyanalysed using a non-parametric Mann–Whitney U-test

for independent samples; a P-value of <0.05 was consid-

ered to be the criterion for statistical significance.

3. Results

3.1. General pattern of lectin reactivity in the normal

organ of Corti

WGA and LFA reacted abundantly with the basolat-eral membranes of the OHCs (Fig. 1) as well as with the

apical membranes, including the stereocilia and stereo-

ciliary cross-links (Fig. 2). Inner hair cells demonstrated

WGA- and LFA-binding sites only along their apical

membranes. Reactivity for both lectins was also ob-

served on the tectorial membrane and along the luminal

membranes of most supporting cells in the organ of Cor-

ti (i.e., Deiters� cells, Hensen�s cells, inner and outer pil-lar cells), the apical membranes of the inner sulcus cells

and Claudius� cells, and the plasma membrane of the

tympanal cells lining the undersurface of the basilar

membrane. In addition, the tectorial membrane and

the extracellular matrix of the basilar membrane demon-

strated an abundance of WGA- and LFA-binding sites.

Intracellular binding of WGA and LFA was present

in both the OHCs and IHCs as well as the supportingcells. It was present in vesicular and tubular structures,

the cisternae of the endoplasmic reticulum and stack-

like tubular structures resembling Golgi saccules as well

as the mitochondria (Fig. 1(a)). The nucleus and the

cytoplasmic matrix as well as the cuticular plate were

nearly always devoid of reactivity.

3.2. Binding specificities of WGA and LFA

In order to test the binding specificities of each lectin,

the labelling densities after pre-absorption to the corre-

sponding inhibitory monosaccharides were determined

for the apical and basolateral membranes of basal-turn

OHCs (Table 1).

WGA labelling densities were calculated at

1201 ± 186 particles/lm2 for the basolateral membranesand at 899 ± 209 particles/lm2 for the apical mem-

branes. After pre-absorption to N-acetyl-D-glucosamine,

the decrease in reactivity amounted to 91% for the baso-

lateral membranes (LD: 103 ± 16 particles/lm2;

P = 0.014) and 94% for the apical membranes (LD:

54 ± 6 particles/lm2; P = 0.014). Pre-absorption to N-

acetylneuraminic acid resulted in a near-complete loss

of reactivity in the basolateral membranes (LD:60 ± 19 particles/lm2; �95%; P = 0.014) and the apical

membranes (LD: 36 ± 14 particles/lm2; �96%;

P = 0.014). These results demonstrate that pre-absorp-

tion with these monosaccharides blocks the reactive sites

on the WGA molecule and thus prevents WGA binding

to reactive sites in the tissue.

LFA labelling densities were 2450 ± 188 particles/

lm2 for the basolateral membranes and 1167 ± 176 par-ticles/lm2 for the apical membranes. After pre-absorp-

tion to N-acetylneuraminic acid, reactivity was reduced

Fig. 1. High-magnification view of basal-turn OHCs in cochleas from non-treated control animals. (a) The LFA-colloidal gold conjugate reacts

abundantly with the basolateral membrane as well as with the subsurface cisternae of the endoplasmic reticulum (ER) and cytoplasmic vesicles and

tubular structures. (b) Digestion with neuraminidase prior to on-grid staining results in complete abolishment of LFA reactivity in the glycocalyx

lining the basolateral surface of the OHC as well as in the endoplasmic reticulum (ER), mitochondria (Mi) and Hensen bodies (HB).

Fig. 2. High-magnification electron micrograph demonstrating WGA

reactivity in the glycocalyx lining the stereocilia and apical membrane

of an OHC in the basal turn of a cochlea from a non-treated control

animal. Inset: A group of transversely cut stereocilia which contain

numerous gold particles indicating the presence of WGA-binding sites

(basal turn, non-treated control animal).


by 54% for the basolateral membranes (LD:

1140 ± 16 particles/lm2), while it was reduced by 69%

for the apical membranes (LD: 360 ± 8 particles/lm2).

Contrary to the WGA-binding experiments, the reduc-tion in LFA labelling density was less than anticipated,

but still statistically significant (basolateral membranes:

P = 0.045; apical membranes: P = 0.04). This incom-

plete reduction might be explained by the use of too

low a concentration (i.e., 400 mM) of the inhibitory

monosaccharide. Pre-absorption to N-acetyl-D-glucosa-

mine did not show a statistically significant decrease in

LFA labelling, neither in the basolateral membranes(LD: 1907 ± 144 particles/lm2; P = 0.096) nor in the

apical membranes (LD: 1190 ± 29 particles/lm2;

P = 0.77). These results show that LFA does not specif-

ically bind to N-acetyl-D-glucosamine and only exhibits

binding specificity for N-acetylneuraminic acid. Taken

together, the specific affinities of both WGA and LFA

were as stated in the information sheets provided by

the supplier.

3.3. Enzymatic digestion prior to lectin incubation

In order to distinguish between WGA binding to

N-acetylneuraminic acid and binding to N-acetyl-D-glu-

cosamine, parallel sections treated with neuraminidase

Table 1

Binding specificities of LFA and WGA as demonstrated in ultrathin cryosections of basal cochlear turns from non-treated animals. Sections were

incubated with lectins pre-absorbed to their respective inhibitory monosaccharides, N-acetylneuraminic acid (NANA) and N-acetyl-D-glucosamine

(GlcNAc). Labelling densities are expressed as the mean number of gold particles/lm2 ± S.E.M.: numbers between brackets represent sample sizes

Lectin alone Lectin + NANA Lectin + GlcNAc

LFA

OHC (basolateral membranes) 2450 ± 188 1140 ± 16 (*) 1907 ± 144 (NS)

[n = 6] [n = 2] [n = 2]

OHC (apical membranes) 1167 ± 176 360 ± 8 (*) 1190 ± 29 (NS)

[n = 7] [n = 2] [n = 2]

WGA

OHC (basolateral membranes) 1201 ± 186 60 ± 19 (*) 103 ± 16 (*)

[n = 8] [n = 3] [n = 3]

OHC (apical membranes) 899 ± 209 36 ± 19 (*) 54 ± 6 (*)

[n = 8] [n = 3] [n = 3]

NS: Not significant with non-parametric Mann–Whitney U-test for independent samples.* Statistically significant at P < 0.05.


and b-N-acetylglucosaminidase, respectively, prior to

incubation with the lectin–gold conjugates were com-

pared to sections treated with the lectin–gold conjugates

alone (Tables 2 and 3).

WGA labelling densities did not demonstrate a statis-

tically significant decrease after digestion with b-N-acet-

ylglucosaminidase, neither in the basolateral nor in the

apical membranes, in either of the cochlear turns. Incontrast, digestion with neuraminidase produced a sig-

nificant reduction of WGA reactivity, both in the baso-

lateral (�75%; P = 0.014) and apical membranes

(�85%; P = 0.036) of basal-turn OHCs. A smaller

reduction was measured in the middle and apical turns

(Table 3). In the middle turn, the reduction amounted

to 45% for the basolateral membranes, which did not

reach statistical significance (P = 0.059), and 58% forthe apical membranes, statistically significant at

Table 2

LFA labelling densities for the apical and basolateral membranes of OHCs in

non-treated animals. Sections were treated with either neuraminidase (NAN

with LFA. Labelling densities are expressed as the mean number of gold par

LFA alone

Basal turn

OHC (basolateral membranes) 2450 ± 188

[n = 6]

OHC (apical membranes) 1167 ± 176

[n = 7]

Middle turn


[n = 9]


[n = 7]

Apical turn


[n = 8]


[n = 8]

NS: Not significant with non-parametric Mann–Whitney U-test for indepen* Statistically significant at P < 0.05.

P = 0.008. In the apical turn, the reduction was 63%

for the basolateral membranes (P = 0.002) and 85%

for the apical membranes (P = 0.017). These results sug-

gest that WGA reactivity in both the basolateral and

apical membranes of the OHCs, irrespective of the co-

chlear turn examined, is mainly due to the presence of

N-acetylneuraminic acid.

LFA labelling densities were unaffected by digestionwith b-N-acetylglucosaminidase, neither in the basolat-

eral nor the apical membranes. After digestion with

neuraminidase, however, LFA labelling was completely

absent, both in the basolateral membranes (�100%;

P = 0.02; see Fig. 1(b)) and the apical membranes

(�98%; P = 0.017) of basal-turn OHCs (Table 2). A sim-

ilar reduction in labelling densities was found for the

middle and apical turns. These results demonstrate thatterminal sialic-acid residues, i.e., N-acetylneuraminic

ultrathin cryosections of basal, middle and apical cochlear turns from

ase) or b-N-acetylglucosaminidase (NAcGlucase) prior to incubation

ticles/lm2 ± S.E.M.; numbers between brackets represent sample sizes

LFA + NANase LFA + NAcGlucase

10 ± 7 (*) 1786 ± 430 (NS)

[n = 3] [n = 3]

19 ± 7 (*) 955 ± 454 (NS)

[n = 3] [n = 3]

51 ± 16 (*) 1123 ± 163 (NS)

[n = 5] [n = 3]

53 ± 18 (*) 864 ± 153 (NS)

[n = 3] [n = 4]

96 ± 36 (*) 1399 ± 62 (NS)

[n = 3] [n = 2]

56 ± 13 (*) 848 ± 102 (NS)

[n = 3] [n = 2]

dent samples.

Table 3

WGA labelling densities for the apical and basolateral membranes of OHCs in ultrathin cryosections of basal, middle and apical cochlear turns from

non-treated animals. Sections were treated with either neuraminidase (NANase) or b-N-acetylglucosaminidase (NAcGlucase) prior to incubation

with WGA. Labelling densities are expressed as the mean number of gold particles/lm2 ± S.E.M.; numbers between brackets represent sample sizes

WGA alone WGA + NANase WGA + NAcGlucase

Basal turn


[n = 8] [n = 3] [n = 4]


[n = 8] [n = 2] [n = 4]

Middle turn

OHC (basolateral membranes) 733 ± 112 402 ± 89 (NS) 559 ± 177 (NS)

[n = 9] [n = 6] [n = 3]


[n = 7] [n = 4] [n = 2]

Apical turn


[n = 8] [n = 7] [n = 3]


[n = 7] [n = 3] [n = 3]

NS: Not significant with non-parametric Mann–Whitney U-test for independent samples.* Statistically significant at P < 0.05.


acid, are present in the basolateral and apical mem-

branes of the OHCs. Our data are well in line with ear-

lier publications (Tachibana et al., 1990; Van Benthem

et al., 1992, 1993), but contradict the conclusion by

Plinkert et al. (1992) that sialic acids are not present in

the OHC glycocalyx. It should be noted, however, that

the latter authors omitted heavy metal ions during their

lectin incubations, whereas traces of heavy metalsare needed to maintain active binding sites on lectins

(Leathem, 1986).

3.4. Membrane polarity in WGA and LFA distribution

Most epithelial cells demonstrate a polarity in lipid

and protein composition between the apical (mostly

luminal) and basolateral domains of their plasma mem-brane (cf., Simons and Fuller, 1985). Cochlear hair cells

demonstrate a pronounced polarity in organelle distri-

bution and cell surfaces, suggesting that the apical mem-

branes of OHCs also might differ from their basolateral

counterparts with regard to sialoglycoconjugate expres-

sion. Therefore, we have compared the labelling densi-

ties for WGA and LFA in the basolateral and apical

membranes of OHCs in normal, non-treated organs ofCorti.

The labelling density for WGA in the basolateral

membranes of basal-turn OHCs was calculated to be

1201 ± 186 particles/lm2 and that for the apical mem-

branes to be 899 ± 209 particles/lm2; these data are sta-

tistically not significantly different (P = 0.21). In the

middle turn, the labelling for WGA in the basolateral

membranes (LD: 733 ± 112 particles/lm2) was signifi-cantly higher (but only just) than that for the apical

membranes (LD: 454 ± 77 particles/lm2; P = 0.05). In

the apical turn, the labelling density for WGA in the

basolateral membranes was determined at 714 ± 64 par-

ticles/lm2, statistically not significantly different from

that calculated for the apical membranes (LD:

535 ± 54 particles/lm2; P = 0.06; see also Table 3,

WGA alone).

The labelling density for LFA in the basolateral

membranes of basal-turn OHCs was determined at2450 ± 188 particles/lm2, which is significantly higher

than the values obtained for the apical membranes

(LD: 1,167 ± 176 particles/lm2; P = 0.004). In the mid-

dle turn, the basolateral membranes contain signifi-

cantly more LFA-reactive sites (LD: 1481 ± 144

particles/lm2) than the apical membranes (LD:

907 ± 63 particles/lm2; P = 0.003). Also, in the apical

turns there is a significantly higher number of LFA-binding sites in the basolateral membranes (LD:

1164 ± 132 particles/lm2) than in the apical membranes

(LD: 759 ± 83 particles/lm2; P = 0.046; see also Table

2, LFA alone).

These results demonstrate that the basolateral mem-

branes of the OHCs, irrespective of the cochlear turn

studied, contain more LFA-reactive sites per membrane

surface area than the apical membranes. AlthoughWGA binding is mainly due to the presence of N-acetyl-

neuraminic acid (vide supra), membrane polarity be-

tween the apical and basolateral membranes is less

outspoken when using WGA as marker. It is also curi-

ous to note that, although both WGA and LFA possess

specific affinities for N-acetylneuraminic acid and,

hence, similar labelling densities were anticipated, LFA

labelling densities in the basolateral membranes were al-ways significantly higher than those for WGA: LFA/

WGA ratios were 1.6 (apical turn, P = 0.012) and 2

Fig. 3. Effect of gentamicin on WGA labelling densities in the

basolateral and apical membranes of OHCs in the basal (a), middle

(b) and apical (c) turns from non-treated control animals (0) and

animals treated for either 5 days (5) or 15 days (15). Asterisk (*)

denotes statistical significance at P < 0.05.


(basal turn, P = 0.002; middle turn, P = 0.002). In the

apical membranes, the results were more varied and sta-

tistical significance was only reached in the middle turn

(P = 0.009).

3.5. Longitudinal gradient in WGA and LFA distribution

It is a well-established fact that OHCs from the basal

and apical cochlear turns differ in structure, innervation,

and physiological responses. Also, there is evidence that

OHCs demonstrate a gradation in the distribution of

cellular macromolecules along the basilar membrane.

In order to resolve whether or not sialoglyoconjugate

expression in the OHC glycocalyx demonstrates a longi-tudinal gradient, we have compared the labelling densi-

ties for WGA and LFA in both membrane domains of

OHCs in the different cochlear turns from normal,

non-treated animals: (1) basal versus middle turn; (2) ba-

sal versus apical turns; and (3) middle versus apical

turns.

WGA labelling densities of the basolateral mem-

branes were only significantly different (P = 0.036) be-tween the basal and apical turns. No evidence of a

statistically significant difference between the cochlear

turns was found when comparing the labelling densities

of the apical membranes. These results demonstrate that

the expression of WGA-binding sites in the basolateral

membranes displays an intracochlear gradient, decreas-

ing towards the apical turn. Such a gradient is not

observed in the apical membranes.LFA labelling densities of the basolateral membranes

were significantly different for basal versus middle turns

(P = 0.005) and basal versus apical turns (P = 0.002),

but not for middle versus apical turns (P = 0.15). For

the apical membranes there was only a statistically sig-

nificant difference for basal versus apical turns

(P = 0.049). These results show that the LFA-binding

sites in the basolateral and apical membranes exhibitan intracochlear gradient, decreasing towards the apical

turn.

3.6. Effect of gentamicin on WGA labelling

WGA labelling densities for the basolateral mem-

branes did not show any significant decrease as com-

pared to the non-treated animals, neither after 5 daysnor after 15 days of continuous gentamicin treatment

in the middle and apical turns (Figs. 3(b) and (c)). How-

ever, in the basal turn (Fig. 3(a)) a significant decrease in

WGA labelling of 43% occurred within 5 days (without

gentamicin: 1,201 ± 186 particles/lm2; 5 days gentami-

cin: 688 ± 83 particles/lm2; P = 0.032), and remained

present after 15 days, even progressed to a significant

loss of 70% (LD: 364 ± 31 particles/lm2; P = 0.002).WGA labelling densities for the apical membranes in

basal-turn OHCs (Fig. 3(a)) demonstrated a 45% reduc-

tion after 5 days of gentamicin treatment, but statistical

significance was not reached (without gentamicin:

899 ± 209 particles/lm2; 5 days gentamicin: 494 ± 81

particles/lm2; P = 0.098). At this point of time, neither

in the middle nor apical turn did gentamicin treatment

result in a reduction in WGA-binding for the apicalmembranes (Figs. 3(b) and (c)). After 15 days of

treatment, WGA labelling was significantly reduced in

the basal turn (without gentamicin: 899 ± 209 particles/

lm2; 15 days gentamicin: 387 ± 65 particles/lm2;

�57%; P = 0.04) as well as in the middle turn (without

gentamicin: 454 ± 77 particles/lm2; 15 days gentamicin:

304 ± 21 particles/lm2; �33%; P = 0.025), but not in the

apical turn.


3.7. Effect of gentamicin on LFA labelling

LFA labelling densities for the basolateral mem-

branes of basal-turn OHCs (Fig. 4(a)) already showed

a significant reduction within 5 days (without gentami-

cin: 2450 ± 88 particles/lm2; 5 days gentamicin:1070 ± 73 particles/lm2; �56%; P = 0.001), which pro-

gressed after 15 days (LD: 695 ± 146 particles/lm2;

�72%; P = 0.004). In the middle turn (Fig. 4(b)), no sig-

nificant change in LFA labelling was obvious after 5

days, but a significant decrease was seen after 15 days

(without gentamicin: 1481 ± 144 particles/lm2; 15 days

gentamicin: 526 ± 71 particles/lm2; �65%; P = 0.001).

In the apical turns, no changes were obvious, neitherafter 5 days nor after 15 days (Fig. 4(c)).

Fig. 4. Effect of gentamicin on LFA labelling densities in the


(b) and apical turns (c) from non-treated control animals (0) and

animals treated for either 5 days (5) or 15 days. Asterisk (*) denotes

statistical significance at P < 0.05.

LFA labelling densities in the apical membranes of

OHCs in the basal turn (Fig. 4(a)) were less than in their

non-treated counterparts, both after 5 and 15 days of

treatment, but this reduction did not reach statistical

significance at the P < 0.05 level. A reduction of the

LFA labelling densities for the apical membranes wasnot observed in the middle and apical turns, neither

after 5 days nor after 15 days (Figs. 4(b) and (c)).

3.8. Effect of neomycin on WGA labelling

WGA labelling densities for the basolateral mem-

branes of basal-turn OHCs (Fig. 5(a)) demonstrated a

significant decrease after 5 days (without neomycin:1201 ± 186 particles/lm2; 5 days neomycin: 612 ± 21

particles/lm2; �49%; P = 0.015) and 15 days of neomy-

cin treatment (LD: 575 ± 110 particles/lm2; �52%;

P = 0.04). In the middle and apical turns no significant

changes were obvious (Figs. 5(b) and (c)).

WGA labelling densities for the apical membranes

were reduced within 5 days of neomycin treatment,

but this reduction was not statistically significant. After15 days, in the basal turn (Fig. 5(a)), a significant loss of

WGA-binding sites was measured (without neomycin:

899 ± 209 particles/lm2; 15 days neomycin: 394 ± 38

particles/lm2; �56%; P = 0.028). In the middle and api-

cal turns, neither after 5 days nor after 15 days, no

changes in WGA labelling densities were observed (Figs.

5(b) and (c)).

3.9. Effect of neomycin on LFA labelling

The basolateral membranes demonstrated a signifi-

cant reduction in LFA labelling in basal-turn OHCs

(Fig. 6(a)) within 5 days (without neomycin:

2450 ± 188 particles/lm2; 5 days neomycin: 613 ± 69

particles/lm2; �75%; P = 0.004) as well as after 15 days

(LD: 811 ± 37 particles/lm2; �67%; P = 0.011). In themiddle turn (Fig. 6(b)) a significant decrease occurred

after 5 days (without neomycin: 1481 ± 144 particles/

lm2; 5 days neomycin: 819 ± 63 particles/lm2; �45%;

P = 0.003). After 15 days the observed reduction of

LFA-binding sites was not statistically significant at

the P < 0.05 level. In the apical turn (Fig. 6(c)), no

significant changes were obvious, neither after 5 days

nor after 15 days.LFA labelling densities for the apical membranes in

the basal-turn OHCs (Fig. 6(a)) demonstrated a slight

reduction after 5-day treatment with neomycin, but sta-

tistical significance was not reached. However, after 15

days the reduction was statistically significant (without

neomycin: 1167 ± 176 particles/lm2; 15 days neomycin:

669 ± 104 particles/lm2; �43%; P = 0.022). In the mid-

dle and apical turns, neither after 5 days nor after 15days, no significant changes were observed (Figs. 6(b)

and (c)).

Fig. 5. Effect of neomycin on WGA labelling densities in the


(b) and apical turns (c) from non-treated control animals (0) and

animals treated for either 5 days (5) or 15 days (15). Asterisk (*)

denotes statistical significance at P < 0.05.

Fig. 6. Effect of neomycin on LFA labelling densities in the basolateral

and apical membranes of OHCs in the basal (a), middle (b) and apical

turns (c) from non-treated control animals (0) and animals treated for

either 5 days (5) or 15 days (15). Asterisk (*) denotes statistical

significance at P < 0.05.


3.10. Effect of aminoglycosides on OHC membrane

polarity

It has been suggested that disorganization of the gly-

cocalyx covering the apical membrane of hair cells is thefirst sign of aminoglycoside-induced fusion of the stere-

ocilia (Takumida et al., 1988, 1989a,b,c,d; Zhang, 1991).

In order to investigate whether the apical and basolat-

eral membranes of OHCs are equally affected by amino-

glycoside administration, we have compared the

labelling densities for WGA and LFA in the apical

and basolateral membranes in all experimental groups.

Our data suggest that the reduction in sialoglycocon-jugate expression, especially when visualized with LFA,

is more prominent in the basolateral membranes than in

the apical membranes. Also, the effect in the apical

membranes seems to lag behind; the effect on the apical

membranes is statistically significant after 15 days (Figs.

3(a), 5(a), 6(a)), whereas the reduction in the basolateral

membranes is already significant after 5 days of contin-uous treatment (Figs. 3(a)–6(a)).

3.11. Difference in cochleotoxic potencies between

gentamicin and neomycin

Since it is well established that the two aminoglyco-

sides applied exhibit differences in cochleotoxic potency

(cf., Rybak, 1986; Govaerts et al., 1990), we studiedwhether this difference is reflected in labelling densities.

In the basolateral membranes of basal-turn OHCs


significantly lower labelling densities for LFA were pres-

ent after 5-day administration of neomycin as compared

to gentamicin (P = 0.04), which was anticipated as neo-

mycin is the more cochleotoxic drug. However, none of

the other cochlear turns showed such a shift in labelling

densities for LFA or WGA, neither after 5 days norafter 15 days of consecutive administration.

4. Discussion

4.1. Deleterious effect of aminoglycoside administration

upon OHC glycocalyx reactivity

The effect of aminoglycoside administration on the

hair cell glycocalyx has been previously studied, both

in the organ of Corti (De Groot and Veldman, 1988;

Takumida et al., 1989c; Rueda et al., 1991) and the ves-

tibular end organs (Takumida et al., 1988, 1989a,b,d;

Zhang, 1991). These studies demonstrate that aminogly-

cosides provoke disorganization of the OHC glycocalyx,

already at an early phase of intoxication. In histologicalsections, this disorganization is usually manifested as a

reduction, or even complete abolishment, of glycocalyx

reactivity after pre-embedding incubation with colloidal

thorium (De Groot and Veldman, 1988), ruthenium red

(Takumida et al., 1989c) and colloidal iron (Rueda

et al., 1991) – each reacting with different molecular

moieties of the glycocalyx – or after post-embedding

detection with lectins (Rueda et al., 1991). Enzymaticdigestion experiments have demonstrated that colloidal

thorium reactivity of the OHC glycocalyx is mainly

due to the presence of sialic acid residues (Van Benthem

et al., 1992, 1993). Furthermore, aminoglycoside admin-

istration is known to result in a decrease in glycocalyx

reactivity for ruthenium red in vestibular hair cells

(Takumida et al., 1988, 1989a,b,d; Zhang, 1991). It

has been suggested that disorganization of the glycoca-lyx may lead to disruption of the stereociliary cross-links

and that it is the first sign of stereocilia fusion. However,

it should be emphasized that stereocilia fusion is a rela-

tively late event.

The aim of this study was to obtain quantitative data

to verify our earlier observation that the amount of sia-

lic acid residues (i.e., sialoglycoconjugates) in the glyco-

calyx of OHCs is diminished after gentamicinadministration (De Groot and Veldman, 1988). There-

fore, we studied the effect of two different aminoglyco-

sides (gentamicin and neomycin) upon OHC

glycocalyx reactivity for LFA and WGA, lectins that

are known to react with the sialic acid N-acetylneuram-

inic acid, using ultrathin cryosections of microdissected

organ of Corti.

The rationale for this choice is that, in order to detectcellular epitopes at the ultrastructural level, high spatial

resolution and optimal antigen preservation are manda-

tory (De Groot et al., 1994). The major disadvantage of

pre-embedding detection of lectin-binding sites is that

penetration of lectin–colloidal gold conjugates into the

fluid-filled spaces of the organ of Corti is poor, which

may result in low binding to reactive sites in the basolat-

eral membranes of the OHCs. On the other hand, post-embedding detection involves dehydration in organic

solvents and embedding in either epoxy (Epon, Spurr�slow-viscosity resin) or methacrylate resins (LR White,

LR Gold, Lowicryl). In our experience such an ap-

proach, especially in cochlear tissues, frequently results

in: (1) loss of reactivity due to extraction, alteration or

masking of reactive groups; (2) a high noise-to-signal ra-

tio; (3) extraction of soluble cell-matrix constituents;and (4) low membrane contrast. Moreover, artefactual

variations in reactivity between different cell organelles

might occur, and because penetration of lectin–colloidal

gold conjugates into the resin-supplemented matrix is

generally poor, reactivity is low and restricted to the sec-

tion�s surfaces. We have, therefore, opted for the prepa-

ration of ultrathin cryosections because this approach,

albeit more tedious and time-consuming, circumventsmany of the limitations presented by resin-based

methods.

The results of the present study show that OHCs

from normal, non-treated cochleas demonstrate a polar-

ity in sialoglycoconjugate expression, in that the baso-

lateral membranes contain more sialoglycoconjugates

per surface area than the apical membranes, similar to

epithelial cells (cf., Simons and Fuller, 1985). Also, itis evident from our quantitative data that sialoglycocon-

jugate expression in the OHC glycocalyx demonstrates a

longitudinal gradient decreasing towards the apical turn

of the cochlea. This finding is in line with the earlier

observation that OHCs demonstrate a gradation in the

distribution of other cellular macromolecules, such as

glycogen (Postma et al., 1978) and the cytoskeletal pro-

tein F-actin (Thorne et al., 1987).The presented data confirm the earlier observation

that aminoglycoside administration results in disorgani-

zation of the OHC glycocalyx (De Groot and Veldman,

1988; Takumida et al., 1989c; Rueda et al., 1991), and

that aminoglycoside-induced disorganization of the

OHC glycocalyx is due to a loss of terminal sialic-acid

residues. Also, basal-turn OHCs are more susceptible

than those in the middle and apical turns. It is a well-known fact that the loss of OHCs after aminoglycoside

administration follows a similar pattern (for a review,

see Forge and Schacht, 2000). Any conclusions on

whether or not the radial gradient in OHC loss is also

reflected at the level of sialoglycoconjugate expression

cannot be drawn, unfortunately, as the data of the three

rows of OHCs were pooled.

Remarkably, the difference in ototoxic potencies be-tween gentamicin and neomycin is not reflected at the le-

vel of sialoglycoconjugate expression. Apparently, this


difference is determined by other factors or dependent

on other cellular processes. It should be noted that dif-

ferences in ototoxic potency are usually reflected in the

degree of OHC loss. Also, they may manifest themselves

as a shift in the time window, during which OHC loss is

first observed. In the present study, however, the degreeof OHC loss in the respective cochlear turns was not

determined. The number of gold particles per surface

area of plasma membrane profile were counted in OHCs

that were still discernible as such. Hence, OHC loss does

not turn into a confounding factor in our study.

4.2. Is the reduced expression of sialoglycoconjugates in

the OHC glycocalyx due to an intracellular response to

aminoglycoside administration?

It has been demonstrated that aminoglycosides and

lectins interact at the cell surface of hair cells in organo-

typic explants of postnatal mice organ of Corti, through

competitive binding or steric hindrance (Richardson

et al., 1989; Kossl et al., 1990), and this might explain

the reduced expression of sialoglycoconjugates in theOHC glycocalyx that is observed after aminoglycoside

administration. However, several observations make

this explanation less plausible.

Firstly, it has been demonstrated that colloidal tho-

rium reactivity of the glycocalyx lining the apical and

basolateral surfaces of OHCs is diminished 24 h after

a single injection of gentamicin (De Groot and Veld-

man, 1988). After prolonged administration for 5 con-secutive days, colloidal thorium reactivity of the OHC

glycocalyx is even completely abolished, whereas reac-

tivities for other cytochemical probes such as cationized

ferritin and OsO4/K4Ru(CN)6 are not affected (De

Groot and Veldman, 1988). Furthermore, enzymatic

digestion experiments have demonstrated that colloidal

thorium reactivity of the hair cell glycocalyx is mainly

due to the presence of sialic acid residues (Van Benthemet al., 1992, 1993). These results correspond to the data

obtained in the present study and cannot be explained

by competitive binding or steric hindrance.

Secondly, in an earlier study, we used an immunogold

technique to detect gentamicin in ultrathin sections of the

guinea-pig organ of Corti (De Groot et al., 1990). Spe-

cific labelling for gentamicin was not present on the baso-

lateral and apical membranes of OHCs, neither 24 hfollowing a single dose nor after prolonged treatment

for 5, 10 or 15 consecutive days. This absence of plasma

membrane-bound gentamicin molecules may be ex-

plained by the rapid endocytotic uptake of the drug

(Richardson and Russell, 1991; Steyger et al., 2003),

which continues during the 24-h period between the final

injection of the drug and fixation of the cochlea. In the

present study, guinea pigs were treated with aminoglyco-sides for either 5 or 15 consecutive days, and the cochleas

were fixed 24 h after the final injection of the drug.

Thirdly, in this study, the reduction in labelling den-

sities after aminoglycoside administration demonstrates

a longitudinal gradient. Our data clearly show that

treatment with aminoglycosides results in a decrease in

WGA and LFA labelling densities in basal-turn OHCs.

Should the suggestion that reduced expression is due tocompetitive binding or steric hindrance be true, then one

would anticipate that the labelling densities for either

lectin be equally reduced in all cochlear turns, already

within 5 days of treatment. However, this is not the case:

Labelling densities in the middle and apical turns were

not significantly reduced after 5-day treatment with

either aminoglycoside (except for the unexplained reduc-

tion in LFA labelling densities in the basolateral mem-branes of middle-turn OHCs after 5-day neomycin

administration).

Fourthly, in the present study, the reduction in label-

ling densities after aminoglycoside administration ap-

pears to be time dependent. This is especially evident

for WGA labelling densities in the basolateral mem-

branes of basal-turn OHCs after gentamicin administra-

tion (see Fig. 3(a)). The reduction in WGA-binding sitesafter 5-day treatment amounts to 43%, but after 15 days

it has increased to 70%. Should the reduction in labelling

densities be due to competitive binding or steric hin-

drance, then one would have expected a more equally

distributed reduction in labelling densities.

Fifthly, in the present study, we observed that the

labelling densities in the basolateral membranes of the

OHCs are affected by the aminoglycoside treatment al-ready after 5 days, whereas reduction of labelling densi-

ties in the apical membranes were statistically significant

only at day 15. In the case of competitive binding or ste-

ric hindrance, a more equally distributed reduction in

labelling densities would have been anticipated.

4.3. Are aminoglycosides present within the OHC at the

time at which expression of sialoglycoconjugates is

downregulated?

It is generally accepted that two events can be distin-

guished in the cellular action of aminoglycosides. Ini-

tially, the drug interacts directly with the (outer) hair

cell glycocalyx by binding to negatively charged surface

moieties of the plasma membrane (i.e., glycocalyx).

After its endocytotic uptake, the drug is sequentiallytransported to endosomes and then lysosomes via vesic-

ular transport. Endocytotic uptake is rapid and genta-

micin can be detected in hair cells from bullfrog

saccular explants already within 30 min (Steyger et al.,

2003). In addition, it has been observed that application

of neomycin, already within 5 min, results in the forma-

tion of membranous blisters on the apical surfaces of

OHCs in organotypic explants of rat organ of Corti(Richardson and Russell, 1991). The observation that

these membranous blisters are not present when the


drug is applied at a temperature of 4 �C, is additional

evidence for endocytotic uptake of aminoglycosides.

In an earlier study, we used an immunogold tech-

nique to detect gentamicin in ultrathin sections of the

guinea-pig organ of Corti (De Groot et al., 1990), but

could not detect any immunogold labelling on the baso-lateral and apical membranes of OHCs, neither 24 h fol-

lowing a single dose nor after prolonged treatment (5, 10

or 15 consecutive days). In the same study, specific label-

ling for gentamicin was observed in OHCs, mainly in the

lysosomes in the infracuticular region, in all cochleas

receiving the drug for 5, 10 or 15 consecutive days. Also,

immunogold labelling was seen in cytoplasmic vesicles

and tubules as well as the cisternae of the endoplasmicreticulum. Similar results have been reported in avian

cochlear hair cells (Fikes et al., 1994; Hashino et al.,

1997) and amphibian saccular hair cells (Steyger et al.,

2003).

We, therefore, think that it is conceivable that the ob-

served reduction in membrane sialoglycoconjugate

expression is associated with effects on intracellular

organelles. On previous occasions, we already suggestedthat aminoglycosides are transported in a retrograde

manner from the endocytotic pathway through the

endoplasmic reticulum and the Golgi apparatus in

OHCs (De Groot and Veldman, 1988; De Groot et al.,

1990). Recently, this retrograde transport has been dem-

onstrated to exist in the proximal tubule cells from rat

and porcine kidneys (Sundin et al., 2001; Sandoval

and Molitoris, 2004).

4.4. Arguments in favour of the suggestion that the

endoplasmic reticulum and/or the Golgi apparatus

are the primary cellular targets in OHCs

Disorganization of the OHC glycocalyx takes place

before disruption of the lysosomes and subsequent re-

lease of the aminoglycoside molecules into the cytosol.Previously, we have attributed this disorganization to

aberrant synthesis and/or glycosylation of membrane

glycoconjugates in OHCs, by assuming that the internal-

ized drug, after vesicular transport, enters the endoplas-

mic reticulum and/or the Golgi apparatus (De Groot

and Veldman, 1988; De Groot et al., 1990). In addition,

aberrant glycosylation of membrane glycoproteins is a

common finding in cells and tissues undergoing patho-logical changes (Damjanov, 1987; Roth, 1993). Our data

suggest that aminoglycosides interfere with the expres-

sion of sialoglycoconjugates in the glycocalyx of the

OHCs. Therefore, we think that it is conceivable that

the endoplasmic reticulum and/or the Golgi apparatus

are the primary cellular targets of the drug, especially

since arguments in favour of this hypothesis are

numerous.Firstly, gentamicin is already observed in the subsur-

face cisternae of the endoplasmic reticulum of guinea pig

OHCs within 5 days of continuous treatment with gen-

tamicin (De Groot et al., 1990). Fikes et al. (1994) found

labelling for gentamicin in the endoplasmic reticulum of

both the short and tall hair cells in the papilla basilaris

of the chick embryo, although it could not be detected

in the postnatal cochlea.Secondly, vesiculation and dilatation of the subsur-

face cisternae of the endoplasmic reticulum as well as

the formation of Hensen�s bodies are seen within 5 days

(De Groot et al., 1991).

Thirdly, colloidal thorium reactivity of the glycocalyx

diminishes already 24 h after a first application of genta-

micin; in the subsequent 5 days reactivity is completely

abolished (De Groot and Veldman, 1988). In line withthis finding is the recent observation (Imamura and

Adams, 2003b) that immunostaining for plasma-

membrane Ca2+-ATPase in the apical membranes (i.e.,

stereocilia) of OHCs is dramatically decreased after

systemic gentamicin administration.

Fourthly, cytochemical activity for the lysosomal en-

zyme acid phosphatase is observed within the subsurface

cisternae of the endoplasmic reticulum after 5 days ofgentamicin treatment (De Groot et al., 1994).

Fifthly, freeze-fracture studies demonstrate that 1–

24 h after the first application of gentamicin, redistribu-

tion of intramembrane particles takes place in the lateral

(subsurface) cisternae of the endoplasmic reticulum

(McDowell et al., 1989).

Sixthly, OHCs in organotypic cultures demonstrate

apical membranes with reduced particle densities ofthe intramembrane particles in response to neomycin,

which suggests that one effect of the drug is to cause

the insertion of new, but abnormal, membrane into

the apical surface. Thus, neomycin may interfere with

some aspect(s) of glycolipid metabolism (Richardson

and Russell, 1991; Forge and Richardson, 1993).

The endoplasmic reticulum is the site of synthesis of

glycolipids destined to the cell surface, whereas theGolgi apparatus is responsible for cytoplasmic sorting.

In this respect, it should be noted that recent investi-

gations have demonstrated that defects in cytoplasmic

sorting signals are linked to human disease (for a re-

view, see Stein et al., 2002). Finally, lysosomes retain

their normal appearance and contain gold labelling

indicating the presence of gentamicin, within 5–7 days

after continuous treatment; during this period of timethere are no indications for lysosomal disruption and

release of gentamicin into the cytosol (De Groot

et al., 1990; Hashino and Shero, 1995; Hashino

et al., 1997).

These observations indicate that the endoplasmic

reticulum and/or the Golgi apparatus are likely to play

a central role in aminoglycoside-induced OHC degener-

ation. It has been demonstrated that a variety of toxicinsults, including inhibitors of glycosylation, chemical

toxicants and oxidative stress can all cause so-called


‘‘endoplasmic reticulum stress’’ and ultimately lead to

cell death (Kaufman, 1999). Rao et al. (2001, 2002) have

demonstrated that any cellular insult that causes pro-

longed endoplasmic reticulum stress (which is the case

in aminoglycoside-induced cytotoxicity) may initiate

programmed cell death through a mitochondrial andApaf-1-independent, intrinsic pathway. Recently,

Bobbin et al. (2003) have observed that perilymphatic

infusion of thapsigargin, a drug that inhibits sarco-

endoplasmic reticulum Ca2+-ATPases, results in sup-

pression of the cochlear potentials and complete loss

of OHCs. In conclusion, it could be possible that endo-

plasmic reticulum stress is involved in aminoglycoside-

induced OHC degeneration.

Acknowledgements

The authors are indebted to the Heinsius-Houbolt

Fund, the Netherlands, for their continuous financial

support. Also, we thank Dr. Marjolein van Ruijven

for her assistance in processing the data for statisticalanalysis.

References

Aran, J.-M., Chappert, C., Dulon, D., Erre, J.-P., Aurousseau, C.,

1995. Uptake of amikacin by hair cells of the guinea pig cochlea

and vestibule and ototoxicity: comparison with gentamicin. Hear.

Res. 82, 179–183.

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

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

Apoptosis in the OC-k3 immortalized cell line treated with different

agents. Audiology 40, 327–335.

Bertolaso, L., Bindini, D., Previati, M., Falgione, D., Lanzoni, I.,

Parmeggiani, A., Vitali, C., Corbacella, E., Capitani, S., Martini,

A., 2003. Gentamicin-induced cytotoxicity involves protein kinase

C activation, glutathione extrusion and malondialdehyde produc-

tion in an immortalized cell line from the organ of Corti. Audiol.

Neuro-Otol. 8, 38–48.

Bobbin, R.P., Parker, M., Wall, L., 2003. Thapsigargin suppresses

cochlear potentials and DPOAEs and is toxic to hair cells. Hear.

Res. 184, 51–60.

Bodmer, D., Brors, D., Pak, K., Bodmer, M., Ryan, A.F., 2003.

Gentamicin-induced hair cell death is not dependent on the

apoptosis receptor Fas. Laryngoscope 113, 452–455.

Damjanov, I., 1987. Biology of disease. Lectin cytochemistry and

histochemistry. Lab. Invest. 57, 5–20.

De Groot, J.C.M.J., Veldman, J.E., 1988. Early effects of gentamicin

on inner ear glycocalyx cytochemistry. Hear. Res. 35, 39–46.

De Groot, J.C.M.J., Meeuwsen, F., Ruizendaal, W.E., Veldman, J.E.,

1990. Ultrastructural localization of gentamicin in the cochlea.

Hear. Res. 50, 35–42.

De Groot, J.C.M.J., Huizing, E.H., Veldman, J.E., 1991. Early

ultrastructural effects of gentamicin cochleotoxicity. Acta Otolar-

yngol. (Stockh.) 111, 273–280.

De Groot, J.C.M.J., Hendriksen, E.G.J., Veldman, J.E., 1994. A new

cytochemical approach for studying aminoglycoside cochleotoxic-

ity. In: Mogi, G., Veldman, J.E., Kawauchi, H. (Eds.), Immuno-

biology in Otorhinolaryngology. Progress of a Decade. Kugler

Publications, Amsterdam/New York, pp. 313–319.

Dehne, N., Rauen, U., De Groot, H., Lautermann, J., 2002.

Involvement of the mitochondrial permeability transition in

gentamicin ototoxicity. Hear. Res. 169, 47–55.

Ding, D., Stracher, A., Salvi, R.J., 2002. Leupeptin protects cochlear

and vestibular hair cells from gentamicin ototoxicity. Hear. Res.

164, 115–126.

Dulon, D., Hiel, H., Aurousseau, C., Erre, J.-P., Aran, J.-M., 1993.

Pharmacokinetics of gentamicin in the sensory hair cells of the

organ of Corti: rapid uptake and long-term persistence. C. R.

Acad. Sci. (Paris) Life Sci. 316, 682–687.

Fikes, J.D., Render, J.A., Reed, W.M., Bursian, S., Poppenga, R.H.,

Sleight, S.D., Yoshioka, T., 1994. Distribution of gentamicin to the

cochlea of the chicken embryo. Toxicol. Pathol. 22, 15–22.

Forge, A., Richardson, G., 1993. Freeze-fracture analysis of apical

membranes in cochlear cultures: differences between basal and

apical-coil outer hair cells and effects of neomycin. J. Neurocytol.

22, 854–867.

Forge, A., Schacht, J., 2000. Aminoglycoside antibiotics. Audiol.

Neuro-Otol. 5, 3–22.

Garetz, S.L., Schacht, J., 1996. Ototoxicity: of mice and men. In: Van

de Water, Th.R., Popper, A.N., Fay, R.R. (Eds.), Handbook of

Auditory Research, Clinical Aspects of Hearing Research, vol. 7.

Springer, New York, pp. 116–154.

Gil-Loyzaga, P., Brownell, W.E., 1988. Wheat germ agglutinin and

Helix pomatia agglutinin lectin binding on cochlear hair cells. Hear.

Res. 34, 149–156.

Govaerts, P.J., Claes, J., Van de Heyning, P.H., Jorens, P.h.G.,

Marquet, J., De Broe, M.E., 1990. Aminoglycoside-induced

ototoxicity. Toxicol. Lett. 52, 227–251.

Griffiths, G., 1993. Fine Structure Immunocytochemistry. Springer,

Berlin, pp. 371–445.

Griffiths, G., McDowall, A., Back, R., Dubochet, J., 1984. On the

preparation of cryosections for immunocytochemistry. J. Ultra-

struct. Res. 89, 65–78.

Hashino, E., Shero, M., 1995. Endocytosis of aminoglycoside antibi-

otics in sensory hair cells. Brain Res. 704, 135–140.

Hashino, E., Shero, M., Salvi, R.J., 1997. Lysosomal targeting and

accumulation of aminoglycoside antibiotics in sensory hair cells.

Brain Res. 777, 75–85.

Hiel, H., Bennani, H., Erre, J.-P., Aurousseau, C., Aran, J.-M., 1992.

Kinetics of gentamicin in cochlear hair cells after chronic

treatment. Acta Otolaryngol. (Stockh.) 112, 272–277.

Hiel, H., Erre, J.-P., Aurousseau, C., Bouali, R., Dulon, D., Aran, J.-

M., 1993. Gentamicin uptake by cochlear hair cells precedes

hearing impairment during chronic treatment. Audiology 32, 78–

87.

Imamura, S.-I., Adams, J.C., 2003a. Distribution of gentamicin in the

guinea pig inner ear after local or systemic application. JARO 4,

176–195.

Imamura, S.-I., Adams, J.C., 2003b. Changes in cytochemistry of

sensory and nonsensory cells in gentamicin-treated cochleas. JARO

4, 196–218.

Kalinec, G.M., Webster, P., Lim, D.J., Kalinec, F., 2003. A cochlear

cell line as an in vitro system for drug ototoxicity screening. Audiol.

Neuro-Otol. 8, 177–189.

Katori, Y., Tonosaki, A., Takasaka, T., 1996. WGA lectin-binding

sites of the apical surface of Corti epithelium: enhancement by

back-scattered electron imaging in guinea-pig inner ear. J. Electron

Microsc. 45, 207–212.

Kaufman, R.J., 1999. Stress signalling from the lumen of the

endoplasmic reticulum: coordination of gene transcriptional and

translational controls. Genes Dev. 13, 1211–1233.

Kossl, M., Richardson, G.P., Russell, I.J., 1990. Stereocilia bundle

stiffness: effects of neomycin sulphate, A23187 and Concanavalin

A. Hear. Res. 44, 217–230.

Leathem, A., 1986. Lectin histochemistry. In: Polak, J.M., Van

Noorden, S. (Eds.), Immunocytochemistry. Modern Methods


and Applications, second ed. John Wright & Sons, Bristol, pp.

167–187.

Lim, D., 1986. Effects of noise and ototoxic drugs at the cellular level

in the cochlea: a review. Am. J. Otolaryngol. 7, 73–99.

McDowell, B., Davies, S., Forge, A., 1989. The effect of gentamicin-

induced hair cell loss on the tight junctions of the reticular lamina.

Hear. Res. 40, 221–232.

Plinkert, P.K., Plinkert, B., Zenner, H.P., 1992. Carbohydrates in the

cell surface of hair cells from the guinea pig cochlea. Eur. Arch.

Otorhinolaryngol. 249, 67–73.

Postma, D.S., Logue, S., Pecorak, J.B., Prazma, J., 1978. Histochem-

istry of glycogen in the inner ear. Histochem. J. 10, 53–61.

Rao, R.V., Hermel, E., Castro-Obregon, S., Del Rio, G., Ellerby,

L.M., Ellerby, H.M., Bredesen, D.E., 2001. Coupling endoplasmic

reticulum stress to the cell death program. Mechanism of caspase

activation. J. Biol. Chem. 276, 33869–33874.

Rao, R.V., Castro-Obregon, S., Frankowski, H., Schuler, M., Stoka,

V., Del Rio, G., Bredeseen, D.E., Ellerby, H.M., 2002. Coupling

endoplasmic reticulum stress to the cell death program. An

Apaf-1-independent intrinsic pathway. J. Biol. Chem. 277,

21836–21842.

Richardson, G.P., Russell, I.J., 1991. Cochlear cultures as a model

system for studying aminoglycoside-induced ototoxicity. Hear.

Res. 53, 293–311.

Richardson, G.P., Russell, I.J., Wasserkort, R., Hans, M., 1989.

Aminoglycoside antibiotics and lectins cause irreversible increases

in the stiffness of cochlear hair-cell stereocilia. In: Wilson, J.P.,

Kemp, D.T. (Eds.), Cochlear Mechanisms. Structure, Function,

and Models. Plenum Press, New York, pp. 57–65.

Roth, J., 1993. Cellular sialoglycoconjugates: a histochemical perspec-

tive. Histochem. J. 25, 687–710.

Rueda, J., Estaban, M., Gutierrez, A., Merchan, J.A., 1991. Glyco-

conjugates change in the organ of Corti of amikacin-treated rats.

In: Lim, D.J. (Ed.), Abstract Book of the Fourteenth Midwinter

Meeting of the Association for Resesearch in Otolaryngology, p.

15.

Rybak, L.P., 1986. Drug ototoxicity. Annu. Rev. Pharmacol. Toxicol.

26, 79–99.

Sandoval, R.M., Molitoris, B.A., 2004. Gentamicin trafficks retro-

grade through the secretory pathway and is released in the cytosol

via the endoplasmic reticulum. Am. J. Physiol. Renal. Physiol. 286,

F617–F624.

Sha, S.-H., Schacht, J., 2000. Antioxidants attenuate gentamicin-

induced free radical formation in vitro and ototoxicity in vivo: D-

methionine is a potential protectant. Hear. Res. 142, 34–40.

Simons, K., Fuller, S.D., 1985. Cell surface polarity in epithelia. Annu.

Rev. Cell Biol. 1, 243–288.

Stein, M.-P., Wandinger-Nes, A., Roitbak, T., 2002. Altered traffick-

ing and epithelial cell polarity in disease. Trends Cell Biol. 12, 374–

381.

Steyger, P.S., Peters, S.L., Rehling, J., Hordichok, A., Dai, C.F., 2003.

Uptake of gentamicin by bullfrog saccular hair cells. JARO 4, 565–

578.

Sundin, D.P., Sandoval, R., Molitoris, B.A., 2001. Gentamicin inhibit

renal protein and phospholipid meatbolism in rats: implications

involving intracellular trafficking. J. Am. Soc. Nephrol. 12, 114–

123.

Tachibana, M., Morioka, H., Machino, M., Yoshimatsu, M., Miz-

ukoshi, O., 1987. Wheat germ agglutinin binding sites in the organ

of Corti as revealed by lectin-gold labelling. Hear. Res. 27, 239–

244.

Tachibana, M., Morioka, H., Bagger-Sjoback, D., Wersall, J., 1990.

Electron microscopic demonstration of sialic acid residues in the

organ of Corti with lectin from Limax flavus. Eur. Arch. Otorh-

inolaryngol. 247, 240–243.

Takumida, M., Harada, Y., Wersall, J., Bagger-Sjoback, D., 1988. The

glycocalyx of inner ear sensory and supporting cells. Acta

Otolaryngol. (Stockh.) Suppl. 458, 84–89.

Takumida, M., Wersall, J., Bagger-Sjoback, D., 1989a. Initial changes

in the sensory hair-cell membrane following aminoglycoside

administration in a guinea pig model. Arch. Otorhinolaryngol.

246, 26–31.

Takumida, M., Urquiza, R., Bagger-Sjoback, D., Wersall, J., 1989b.

Effect of gentamicin on the carbohydrates of the vestibular end

organs: an investigation by the use of FITC-lectins. J. Laryngol.

Otol. 103, 357–362.

Takumida, M., Wersall, J., Bagger-Sjoback, D., Harada, Y., 1989c.

Observation of the glycocalyx of the organ of Corti: an investiga-

tion by electron microscopy in the normal and gentamicin-treated

guinea pig. J. Laryngol. Otol. 103, 133–136.

Takumida, M., Bagger-Sjoback, D., Harada, Y., Lim, D., Wersall, J.,

1989d. Sensory hair fusion and glycocalyx changes following

gentamicin exposure in the guinea pig vestibular organs. Acta

Otolaryngol. (Stockh.) 107, 39–47.

Takumida, M., Popa, R., Anniko, M., 1999. Free radicals in the

guinea pig inner ear following gentamicin exposure. ORL 61, 63–

70.

Thorne, P.R., Carlisle, L., Zajic, G., Schacht, J., Altschuler, R.A.,

1987. Differences in the distribution of F-actin in outer hair cells

along the organ of Corti. Hear. Res. 30, 253–266.

Tokuyasu, K.T., 1989. Use of poly(vinylpyrrolidine) and poly(vinyl

alcohol) for cryoultramicrotomy. Histochem. J. 21, 163–171.

Van Benthem, P.P.G., Albers, F.W.J., De Groot, J.C.M.J., Veldman,

J.E., Huizing, E.H., 1992. Glycocalyx heterogeneity in normal and

hydropic cochleas of the guinea pig. Acta Otolaryngol. (Stockh.)

112, 976–984.

Van Benthem, P.P.G., De Groot, J.C.M.J., Albers, F.W.J., Veldman,

J.E., Huizing, E.H., 1993. Structure and composition of stereocilia

cross-links in normal and hydropic cochleas of the guinea pig. Eur.

Arch. Otorhinolaryngol. 250, 73–77.

Ylikoski, J., Xing-Qun, L., Virkkala, J., Pirvola, U., 2002. Blockade of

c-Jun N-terminal kinase pathway attenuates gentamicin-induced

cochlear and vestibular cell death. Hear. Res. 163, 71–81.

Zhang, J., 1991. Electron microscopic studies on the glycocalyx of the

utricular macula following amikacin in guinea pigs. Zhonghua Er

Bi Yan Hou Ke Za Zhi 26, 3–5 (Abstract).

reduced expression of sialoglycoconjugates in the outer hair cell glycocalyx after systemic...

Documents