Otology & Neurotology

32:596Y601 2011, Otology & Neurotology, Inc.

Structural Tympanic Membrane Changes in Secretory

Otitis Media and Cholesteatoma
*Johan Knutsson, Dan Bagger-Sjoback, and kMagnus von Unge
*Center for Hearing and Communication Research, Karolinska Institute; Department of Otorhinolaryngology
and Center for Clinical Research, Vasteras Central Hospital; Department of Otorhinolaryngology,
Karolinska University Hospital; Department of Clinical Science, Intervention and Technology,
Karolinska Institute, Sweden; and kDepartment of Otorhinolaryngology, Akershus University Hospital
and Institute for Clinical Medicine, Akershus University Hospital, University of Oslo, Norway

Background: Otitis media may predispose for retraction

pathologic abnormality later in life. A weakening of the collagen
fiber bundles in the lamina propria of the tympanic membrane
(TM) is a prerequisite for the formation of a retraction pocket.
Various collagen types have different tensile strength. The
collagen-type distribution in the TM during otitis media and
cholesteatoma has not been reported before.
Materials and Methods: The collagen contents of TM biopsies
from child patients with longstanding secretory otitis media
without retraction pockets were compared with pars tensa cholesteatomas using immunohistochemical staining for collagen
Types I to IV. The histology was also investigated using transmission electron microscopy.
Results: The outer epithelium was in some biopsies thickened
with evidence of edema. The biopsies showed an intact lamina
propria with positive immunohistochemical staining for collagen Types I to III and showed normal collagen fiber bundles

on electron microscopy. The outer epithelium of the cholesteatomas showed marked thickness variations and signs of
edema. There was a presence of normal collagen fiber bundles in
smaller parts of all cholesteatomas, positive for collagen Types I
to II. In other parts, only scattered collagen fibers were found.
Conclusion: Tympanic membrane biopsies from patients with
longstanding secretory otitis media may show a thickening of
the outer epithelium. Collagen Types I to III are present in the
lamina propria, and no ultrastructural changes of the collagen
fiber bundles are observed. Collagen is found in cholesteatomas
in the remnants of the lamina propria, with positive staining
for collagen Types I and II, whereas Type III seems to be
lacking. Key Words: CholesteatomaVCollagen typeVElectron
microscopyVHumanVImmunohistochemistryVLamina
propriaVPars tensaVTympanic membrane.

Secretory otitis media (SOM) may be a risk factor for

the development of retraction pockets and cholesteatoma,
but the genesis of cholesteatoma has yet to be fully
clarified (1Y3). According to one leading theory, a retraction pocket formation precedes the development of
cholesteatoma (4). The latter is formed when the keratin
that is exfoliated from the squamous cell epithelium
surface becomes inadequately cleansed from the pocket,
leading to retention of this debris within the pocket.
Simultaneously, the epithelial layer turnover is accelerated with an increased expression of keratinocyte and
epidermal growth factor receptors (5Y7).

A weakening of the strength-bearing properties of the

tympanic membrane (TM) is assumed to be a prerequisite
for the formation of a retraction pocket, but the pathophysiology of this process is not fully understood (8,9).
The overall stiffness of the TM is reduced in the acute and
subacute stages of experimental otitis media (10,11).
Morphologic investigations using transmission electron
microscopy have not revealed any reduction in the
number of collagen fibers and no signs of deterioration
of the fibers in these diseases. The collagen fiber bundles
are often spatially separated from each other by edema,
but the stiffness losses cannot be fully attributed to this
finding (12).
The strength and tensile properties differ between the
various types of collagens. Collagen Type II is the most
prevalent in the TM lamina propria of man as well as of
rat (13Y15). There is also a significant presence of collagen Type III in the human TM, being most abundant
in the inner circular layer of the lamina propria (15). A

Otol Neurotol 32:596Y601, 2011.

Address correspondence and reprint requests to Johan Knutsson,

M.D., Ph.D., Department of Otorhinolaryngology, Vasters Central Hospital, SE-721 89 Vasters, Sweden; E-mail: johan.knutsson@ltv.se
The study was supported by the Centre for Clinical Research, Vasters
Central Hospital.

596

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

STRUCTURAL TM CHANGES IN SECRETORY OTITIS MEDIA

597

change in collagen-type composition, with a shift toward

the less deformation-resistant types, might explain the
change in the tensile properties after otitis media and
thereby set the stage for retraction pockets and consecutively cholesteatoma formation.
In the present study, the hypothesis was that TMs with
SOM and cholesteatoma exhibit a different collagen-type
composition compared with healthy TMs previously described (15). The present study used hematoxylin and
eosin staining for a histologic overview of the lamina
propria and the keratinizing epithelium in TM biopsies
obtained from patients with SOM without retraction
pockets and in surgically removed cholesteatomas. Immunohistochemistry was used to identify the different
collagen types. The morphology was further investigated
with the use of transmission electron microscopy (TEM).

Monoclonal mouse antibodies and polyclonal goat antibodies

against specific human collagen types were used as primary
antibodies for collagen Types I, II, III, and IV (Type I, COL1A1
[D-13]: sc-25974; Type III, COL3A1 [FH-7A]: sc-59817; Type
IV, COL4A [COL-94]: sc-59814 [Santa Cruz Biotechnology,
Inc., Santa Cruz, California, USA]; Type II, collagen II Ab-3
[clone 6B3; Labvision, Fremont, California, USA) diluted 1:100
in PBS.
Biotinylated antigoat IgG 5 Kg/ml (Vector Laboratories, Inc.,
Burlingame, California, USA) diluted 1:300 was used as secondary antibody for collagen Type I. Biotinylated antimouse
IgG 10 Kg/ml (Southern Biotech, Birmingham, Alabama, USA)
diluted 1:200 in PBS was used as secondary antibody for collagen Types II, III, and IV. The immunohistochemical procedure was performed according to a standard protocol using DAB
peroxidase substrate, as previously published (15). The sections
were evaluated using light microscopy (Zeiss Observer.Z1; Carl
Zeiss GmbH, Braunschweig, Germany).

MATERIALS AND METHODS

The specimens were immediately fixed in a solution of 2.5%

glutaraldehyde and 0.5% paraformaldehyde in phosphate buffer.
After rinsing in phosphate buffer, postfixation was performed
with 1% osmium tetroxide (Sigma Aldrich, Stockholm, Sweden)
in phosphate buffer and then rinsing and dehydration in graded
alcohol series and propylene oxide. The specimens were embedded in agar 100 resin (Link Nordiska, Lidingo, Sweden) and
incubated at 40-C during 24 hours and 60-C during 48 hours
for polymerization. Sectioning was performed with an ultrotome
to approximately 0.1 Km thickness, and the sections were dried
on formvar-covered copper grids. Staining was performed using
2% uranylacetate in 70% ethanol and Reynolds lead staining.
For visualization, a Jeol 1230 transmission electron microscope
(Jeol Ltd, Sollentuna, Sweden) was used, and the images were
digitally stored.

The parents of children with longstanding (96 mo) SOM and

with audiologically confirmed constant hearing impairment
scheduled for ventilation tube insertion were invited to enter
their child with written informed consent into this prospective
study. Tympanic membrane biopsies were harvested from 1 ear
in the procedure of ventilation tube insertion at our ENT department during a 12-month period. Ears with retraction pockets or
previous ear surgery were excluded. In all, 11 biopsies were
harvested from children aged between 3 and 7 years. A 1-mmdiameter biopsy was harvested with the use of a custom-made
microinstrument (R. Nagashima, Japan) after a myringotomy
was performed in the anteroinferior quadrant but before the
insertion of the ventilation tube. During the same 12-month
period, adult patients scheduled for eradicating surgery for
cholesteatoma emanating from the pars tensa were included.
The cholesteatoma matrix removed in the routine surgical procedure was collected. The protocols were approved by the local
ethics committees (KI forskningskommitte Nord 03-523 and
Regionala etikprovningsnamnden Stockholm 04-349). A total
of 6 biopsies and 6 cholesteatomas were prepared for light
microscopy and immunohistochemistry. A total of 5 biopsies
and 3 cholesteatomas were prepared for TEM.
For light microscopy, the specimens were fixed in formaldehyde (4%) immediately after removal, dehydrated in serial
alcohol concentrations, embedded in paraffin, and cut into
5-Km sections. The sections were placed on SuperFrost slides
(Thermo Scientific, Stockholm, Sweden), deparaffinized, and
rehydrated.

Transmission Electron Microscopy

RESULTS
One of the biopsies was lost in the paraffin embedding
process, and 2 were lost in the TEM preparations. Thus,
5 TM biopsies and 6 cholesteatomas were analyzed using

Hematoxylin and Eosin Staining

After a quick rinse in ionized water, the sections were stained
with fresh Mayer hematoxylin followed by washing in tap
water. Counterstaining was performed with eosin. After dehydration, the slides were mounted with cover glass.

Immunohistochemistry
Endogenous peroxidase activity was blocked by incubation
for 30 minutes in 0.3% hydrogen peroxide in distilled water,
and the sections were washed repeatedly in phosphate-buffered
saline (PBS, pH 7.4). A pepsin solution (Thermo Fisher Scientific,
Fremont, California, USA) was used for 10 minutes at 37-C for
enzyme-induced epitope retrieval.

FIG. 1. Hematoxylin and eosin staining of a TM biopsy from a

patient with longstanding SOM. EAC indicates external ear canal;
KE, keratinizing epithelium, tissue; LP, lamina propria; ME, middle
ear; SCT, inner subepithelial connective tissue layer.
Otology & Neurotology, Vol. 32, No. 4, 2011

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

598

J. KNUTSSON ET AL.

hematoxylin and eosin staining and immunohistochemistry, whereas 3 biopsies and 3 cholesteatomas were analyzed using TEM.
Hematoxylin and Eosin Staining
The biopsies showed a slightly thickened outer epithelium in some specimens, whereas it appeared normal
in others (Fig. 1). The lamina propria had a normal appearance without any obvious signs of deterioration.
The outer epithelium of the cholesteatomas showed
great thickness variations, measuring between 30 and
250 Km, excluding the desquamating keratin (Fig. 2A).

FIG. 3.

Positive staining for collagen Type II in a cholesteatoma.

The 4 layers (stratae basale, spinosum, granulosum, and

corneum) were clearly visible and, in many areas, markedly enlarged (Fig. 2B). The cells of the basal layer
showed notably large nuclei (Fig. 2C, thin arrow). All
cholesteatomas had some areas with somewhat preserved
fibrous lamina propria, although it was lacking in the
major portions of the specimens (Fig. 2C, thick arrow).
Immunohistochemistry
Each section had a corresponding negative control
mounted on the same glass slide, none of these showing
any staining at all. The biopsy sections showed staining
for collagen Types I to III in the lamina propria and
strongly for Type IV in the basal laminas. In the areas
with seemingly preserved lamina propria in the cholesteatomas, staining was positive for Types I and II but not
for III (Fig. 3). Staining for Type IV was positive and
constant in the basal laminas in all specimens.

FIG. 2. AYC, Hematoxylin and eosin staining of a cholesteatoma. A, Irregular variations of the epithelium thickness. B, All
layers of the normal outer epithelium are observed in the cholesteatoma although enlarged. C, Enlarged nuclei in the basal
layer cells (thin arrows). Possible remnants of collagen fibers
(thick arrow).

Transmission Electron Microscopy

In the biopsies, the outer epithelial layer appeared
normal or thickened to between 15 and 25 Km. The cells
of the basal layer were predominately elongated parallel
to the basal lamina, but in the thicker parts, the basal cells
were elongated standing almost perpendicularly to the
basal lamina (Fig. 4A). The lamina propria with its outer
radial and inner circular layers of collagen bundles appeared intact (Fig. 4B). In the case of a markedly thickened epithelium, there were signs of edema with the
epithelial cells seemingly separated by fluid in a similar
appearance as that in the cholesteatomas (Figs. 4C and 5A).
The cholesteatomas showed a similar edematous
separation of cells in the thickened epithelium and had
more cell layers (Fig. 5, A and B). The cells of the basal
layer were elongated perpendicularly to the basal lamina,
whereas the cells of the superficial layers were flat and
lacking nuclei (Fig. 5, A and B). Remnants of a lamina
propria consisting of intact appearing collagen fiber
bundles were identified in parts of the cholesteatomas
(Fig. 5C). In other parts, only scattered, disorganized
collagen fibers were found (Fig. 5D).

Otology & Neurotology, Vol. 32, No. 4, 2011

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

STRUCTURAL TM CHANGES IN SECRETORY OTITIS MEDIA

DISCUSSION
The patients charts were preoperatively checked, and
all patients had been observed by an otolaryngologist
regularly. The SOM diagnosis was confirmed at least at a
3-month interval. Episodes of improvement of the SOM

599

cannot, however, be ruled out because the disease often

has a variable course, but the parents denied episodes of
improved hearing within the 3-month interval preceding
the surgical intervention.
The mechanical damage to the samples produced by
the biopsy instrument somewhat obscured the assessment of the specimens, but reliable results could still
be obtained. Furthermore, it was difficult to orient the
biopsies in the embedding process because the biopsies
were so small, that is, only 1 mm in diameter. The reason
for not harvesting larger biopsies was that this could
inflict greater trauma to the TM than what would be
necessary for tube insertion. Thus, obtaining larger biopsies was considered unethical.
Increased Epithelial Activity
The SOM biopsies showed a normal or slightly
increased thickness of the keratinizing epithelium compared with the normal thickness of approximately 10 Km
(16,17). Similar results with a slight thickness increase in
the outer epithelium are found in Mongolian gerbils with
SOM (10). In the cholesteatoma specimens, the keratinizing epithelium was immensely thickened and with
great thickness variations, ranging from 30 to 250 Km in
the same specimen. An exact measurement of the thickness of a keratinizing epithelium is, however, difficult to
make because the keratin may desquamate during the
preparation resulting in reduced thickness. The increased
thickness found in the present study may be an expression
of increased cellular turnover.
Using TEM, the 4 epithelial layers were clearly visible
in the TM biopsies and the cholesteatomas. Normal outer
epithelium harbors basal layer cells elongated in the plane
of the basal lamina (16). In the present study, the basal
layer cells in the thickened parts were elongated with a
position almost perpendicular to the basal lamina. This
finding might reflect an increased cellular turnover because this cell arrangement allows for more cells present
per area. We have recently found a similar feature in
healthy human TMs in the area where the epithelial
migration originates (18). It can be speculated that the
areas of elongated cells in the cholesteatomas are important centers of keratinocyte proliferation.
Collagen Fiber Bundle Damage
The content of the different collagen types in SOM and
pars tensaYderived cholesteatomas was assessed in the
present study. Collagen is abundant in connective tissue.
FIG. 4. AYC, Transmission electron microscopy of TM biopsies
from patients with longstanding SOM. A, Outer keratinizing epithelium. Cells elongated parallel to the basal lamina in the thinner
part of the epithelium. In the thicker part, the basal cells are
elongated perpendicular to the basal lamina. oSCT indicates outer
subepithelial connective tissue. B, Intact lamina propria. Thickening of the inner subepithelial layer. iLP indicates lamina propria,
inner layer; iSCT, inner subepithelial connective tissue; oLP,
lamina propria, outer layer. C, Cells in the outer epithelium seemingly separated by fluid; KE, keratinizing epithelium, tissue; LP,
lamina propria.
Otology & Neurotology, Vol. 32, No. 4, 2011

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

600

J. KNUTSSON ET AL.
Types I to IV collagens are the most commonly found.
Type I provides resistance to force, whereas Type II provides shape and resistance to deformation. Types III and
IV provide elasticity and support. Type IV also functions
as a filtration barrier (19). A change in the collagen
composition of the lamina propria could explain the loss
of stiffness observed in experimental otitis media and
might set the stage for retraction pathologic findings
during unfavorable middle ear pressure situations (10,11).
The present immunohistochemical method was previously used for the detection of different collagen types
in healthy human TMs (15). The antibodies are highly
specific for each collagen type and have according to the
supplying companies less than 10% cross reactivity.
Therefore, the method seems quite reliable. Owing to the
low number of specimens, no semiquantification was,
however, performed as was done in the previous studies
of normal TMs.
Secretory otitis media in the Mongolian gerbil does not
seem to affect the lamina propria initially (10). Later, a
disintegration of the inner circular collagen bundles is
observed in ears with a highly viscous form of otitis
media. The lamina propria is, however, not affected in
low-viscosity SOM. In the present study, the TM biopsies
had no signs of fiber disorganization when observed
using TEM. It can be hypothesized that the human form
of SOM more resembles the gerbils low-viscosity type of
SOM. It must, however, be kept in mind that the present
TEM study is based on only 3 human TM biopsies
because of the difficulties in obtaining research material.
The collagen content of the healthy human TM was
previously investigated and semiquantified (15). In the
present study, the immunohistochemical staining for
different collagen types showed presence of Types I to III
in the lamina propria. The intention was to perform a
similar semiquantification as in the previous study. This
proved to be impossible owing to the mechanical damage
caused by the biopsy instrument.
Collagen fibers are present in cholesteatomas (20,21).
Areas of abundant collagen fiber bundles are found, but in
other areas, the collagen fibers are scarce. In the present
study, the cholesteatoma specimens showed presence of
collagen only in smaller parts, whereas in other parts,
light microscopy and TEM could not identify any organized collagen fiber bundles. In the areas where collagen
was present, the stainings were positive for Types I and II
but not for Type III. Type IV was found in the basal
lamina. The absence of Type III is interesting. This collagen type may play a role in the reparative process. Type
III is, together with Type I, the first collagen to be produced in the healing TM after a perforation (13). The
collagen content in the lamina propria becomes modified
FIG. 5. AYD, Transmission electron microscopy of cholesteatomas. A, Epithelial cells with nuclei (upper part of the image) and
flattened cells without nuclei (lower part of the image). B, Thickened epithelium with signs of edema. Cells of the basal layer
elongated perpendicularly to the basal lamina. C, Intact collagen
fiber bundles. D, Scattered collagen fibers (arrows).

Otology & Neurotology, Vol. 32, No. 4, 2011

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

STRUCTURAL TM CHANGES IN SECRETORY OTITIS MEDIA

later during the healing process with a shift toward Type
II. The absence of Type III in cholesteatomas might be
of importance for the reparative processes in a developing
cholesteatoma.
In summary, the present material does not reveal substantial morphologic similarities between the TM changes
in SOM and cholesteatoma. Nevertheless, the thickenings
encountered in the keratinizing epithelium in the biopsy
material could indicate an increased cellular proliferation
in SOM similar to what was observed in the cholesteatomas. It remains to be clarified whether the regulating factors are the same. The hypothesis of SOM being
an early precursor of cholesteatoma needs to be further
elucidated. Histologic studies of other possible forerunners of cholesteatoma, such as self-cleansing TM
retractions, would possibly further enlighten the pathophysiology of cholesteatoma.
CONCLUSION
Tympanic membrane biopsies from patients with longstanding SOM can show a thickening of the outer epithelium. Collagen Types I to III are present in the lamina
propria and collagen Type IV in the basal lamina.
The outer epithelium is markedly thickened in parts of
cholesteatoma specimens. Collagen is found in some
areas of cholesteatomas, with positive staining for Types I
and II, whereas Type III seems to be lacking.
Acknowledgment: The authors thank Paula Mannstrom for
excellent technical assistance.

REFERENCES
1. Tos M. Upon the relationship between secretory otitis in childhood
and chronic otitis and its sequelae in adults. J Laryngol Otol 1981;
95:1011Y22.
2. Tos M, Poulsen G. Attic retractions following secretory otitis. Acta
Otolaryngol 1980;89:479Y86.
3. Edelstein DR, Parisier SC, Ahuja GS, et al. Cholesteatoma in the
pediatric age group. Ann Otol Rhinol Laryngol 1988;97:23Y9.
4. Wells MD, Michaels L. Role of retraction pockets in cholesteatoma
formation. Clin Otolaryngol Allied Sci 1983;8:39Y45.

601

5. Barbara M, Raffa S, Mure` C, et al. Keratinocyte growth factor

receptor (KGF-R) in cholesteatoma tissue. Acta Otolaryngol 2008;
128:360Y4.
6. Bujia J, Kim C, Holly A, Sudhoff H, Ostos P, Kastenbauer E.
Epidermal growth factor receptor (EGF-R) in human middle ear
cholesteatoma: an analysis of protein production and gene expression. Am J Otol 1996;17:203Y6.
7. Kojima H, Shiwa M, Kamide Y, Moriyama H. Expression and localization of mRNA for epidermal growth factor and epidermal
growth factor receptor in human cholesteatoma. Acta Otolaryngol
1994;114:423Y9.
8. Ars B, De Craemer W, Ars-Piret N. The lamina propria and cholesteatoma. Clin Otolaryngol Allied Sci 1989;14:471Y5.
9. Ars BM. Tympanic membrane retraction pockets. Etiology, pathogeny, treatment. Acta Otorhinolaryngol Belg 1991;45:265Y77.
10. von Unge M, Bagger-Sjoback D. Tympanic membrane changes
in experimental otitis media with effusion. Am J Otol 1994;15:
663Y9.
11. von Unge M, Decraemer WF, Dirckx JJ, Bagger-Sjoback D. Shape
and displacement patterns of the gerbil tympanic membrane in experimental otitis media with effusion. Hear Res 1995;82:184Y96.
12. von Unge M, Decraemer WF, Bagger-Sjoback D, Van den Berghe D.
Tympanic membrane changes in experimental purulent otitis media.
Hear Res 1997;106:123Y36.
13. Stenfeldt K, Johansson C, Hellstrom S. The collagen structure of the
tympanic membrane: collagen Types I, II, and III in the healthy
tympanic membrane, during healing of a perforation, and during
infection. Arch Otolaryngol Head Neck Surg 2006;132:293Y8.
14. Knutsson J, Bagger-Sjoback D, von Unge M. Distribution of different collagen types in the rats tympanic membrane and its suspending structures. Otol Neurotol 2007;28:486Y91.
15. Knutsson J, Bagger-Sjoback D, von Unge M. Collagen type distribution in the healthy human tympanic membrane. Otol Neurotol
2009;30:1225Y9.
16. Hentzer E. Ultrastructure of the human tympanic membrane. Acta
Otolaryngol 1969;68:376Y90.
17. Kakoi H, Anniko M. Auditory epithelial migration. IV. Light- and
scanning electron-microscopic studies of the tympanic membrane
and external auditory canal in the mouse. ORL J Otorhinolaryngol
Relat Spec 1996;58:136Y42.
18. Knutsson J, von Unge M, Rask-Andersen H. Localization of
progenitor/stem cells in the human tympanic membrane. Audiol
Neurootol 2010;16:263Y9.
19. Kuhn K, Glanville RW. Molecular Structure and Higher Organization of Different Collagen Types. Biology of Collagen. London,
U.K.: Academic Press, 1980:1Y14.
20. Lim DJ, Saunders WH. Acquired cholesteatoma: light and electron
microscopic observations. Ann Otol Rhinol Laryngol 1972;81:1Y11.
21. Anniko M, Mendel L. Cholesteatoma. A clinical and morphological
analysis. Acta Otolaryngol 1981;91:275Y83.

Otology & Neurotology, Vol. 32, No. 4, 2011

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.