The Laryngoscope
Lippincott Williams & Wilkins, Inc.
© 2004 The American Laryngological,
Rhinological and Otological Society, Inc.
Cytokine Expression in Experimental
Chronic Otitis Media With Effusion in Mice
Kazuhiko Maeda, MD; Takashi Hirano, MD; Issei Ichimiya, MD; Yuichi Kurono, MD; Masashi Suzuki, MD;
Goro Mogi, MD
Objectives: Although otitis media with effu-
sion (OME) is still a common disease in children
and adults, the pathogenesis is not yet fully un-
derstood. We studied the effects of intratympanic
injection with endotoxin purified from nontype-
able Haemophilus influenzae on the characteris-
tics of middle ear effusion (MEE). Methods: Mu-
rine model of OME was developed by eustachian
tube (ET) blockage followed by intratympanic in-
oculation with endotoxin (endotoxin group) or sa-
line (control group). The mice were decapitated
and histological changes and the production of
inflammatory cytokines in MEEs were examined 3
days, 2 weeks, and 2 months after injection. Re-
sults: All mice showed OME until 2 months after
ET blockage. Most MEEs in the control group
were serous, and mucoid or pultaceous MEEs
were found only in the endotoxin group. Subepi-
thelial space of middle ear mucosa was severely
thickened with the infiltration of a large number
of mononuclear cells in the endotoxin group. The
levels of tumor necrosis factor-␣ (TNF-␣) in MEEs
were significantly higher in the endotoxin group
than in the control group at all time points. Fur-
ther, in situ hybridization showed that TNF-␣
messenger RNA was expressed not only by leuko-
cytes and macrophages in MEEs but mononu-
clear cells present in the subepithelial space of
middle ear mucosa. Conclusions: These results in-
dicate that ET blockage is essential for the induc-
tion of serous MEE and additional administration
of endotoxin is associated with the production
of mucoid MEE accompanied by histological
changes with inflammatory cell infiltration and
cytokine production in the tympanic cavity. Key
Words: Endotoxin, otitis media with effusion,
mouse, TNF-␣.
Laryngoscope, 114:1967–1972, 2004
From the Departments of Otolaryngology, Faculty of Medicine, Oita
University (K.M., T.H., I.I., M.S., G.M.), Oita, and Kagoshima University (Y.K.),
Kagoshima, Japan.
Editor’s Note: This Manuscript was accepted for publication March
24, 2004.
Send Correspondence to Masashi Suzuki, MD, Department of Oto-
laryngology, Faculty of Medicine, Oita University, 1-1, Idaigaoka, Hasama-
cho, Oita, Japan. E-mail: suzukim@med.oita-u.ac.jp
DOI: 10.1097/01.mlg.0000147930.29261.51
Laryngoscope 114: November 2004
INTRODUCTION
Otitis media with effusion (OME) is one of the most
common diseases in pediatric and geriatric populations
and is characterized by persistent mucoid or serous mid-
dle ear effusion (MEE). Although microbial infection and
eustachian tube (ET) dysfunction are well known to be
involved in the pathogenesis of OME, the factors respon-
sible for the persistence of this disease and the production
of serous and mucoid MEE are not yet fully understood.
Nontypeable Haemophilus influenzae (NTHi) and
Streptococcus pneumoniae (Spn) are frequently cultured
from MEE of patients with OME and considered the chief
pathogens of OME. Although both bacteria contain sev-
eral inflammatory agents, antigens of NTHi cause greater
inflammatory responses than are caused by Spn.1 One of
these antigens, endotoxin, consists of lipo-oligosaccharide
complex with outer membrane proteins of NTHi as well as
the other Gram-negative bacteria and is a potent inducer
of inflammation and immunological responses. In fact,
administration of endotoxin alone can induce OME in
experimental animals.2 Moreover, endotoxin is detected in
MEE in a high percentage of patients with OME. DeMaria et
al.3 established the presence of endotoxin in 80% of MEE
specimens of patients with chronic OME and in 67% of
culture-negative MEE specimens. Findings indicate that en-
dotoxin is one of the chief pathogenic factors of OME.
Endotoxin-induced inflammation is caused by the
production of a variety of proinflammatory cytokines such
as tumor necrosis factor-␣ (TNF-␣) and interleukin-1␤
(IL-1␤), and the concentrations of endotoxin in MEE cor-
relate significantly with the concentrations of these cyto-
kines.4 In addition, the presence of TNF-␣ and IL-1␤ in
MEE is considered to be responsible for the mucosal dam-
age, bone erosion, fibrosis, and hearing loss seen in some
cases of chronic OME, suggesting that these proinflamma-
tory cytokines are responsible for the persistence of OME
in children.5 Thus, it is important to clarify the roles of
endotoxin and these cytokines in the pathogenesis of
chronic OME. However, there has been little research into
the production of proinflammatory cytokines in the middle
ear cavity during chronic OME. In the present study, we
generated a murine model of chronic OME by ET blockage
and intratympanic injection of endotoxin, and we studied
the pathogenesis of chronic OME and factors associated
Maeda et al.: Cytokine in Otitis Media With Effusion
1967
with the production of mucoid MEE by examining his-
topathological changes and expression of proinflammatory
cytokines in the middle ear mucosa.
MATERIALS AND METHODS
Endotoxin
Endotoxin was purified according to the method of Westphal
et al.6 from NTHi isolated in the Department of Otolaryngology of
Oita University from the nasopharynx of a patient with OME.
Animals
One hundred twenty male BALB/c mice at 6 weeks of age
were used for all experiments. The mice were maintained under
specific pathogen-free conditions. The experimental protocol was
approval by the Committee on Animal Experiments of Oita Med-
ical University (Oita, Japan).
Induction of Otitis Media With Effusion
With the animal under anesthesia induced with 0.6 mg/mL
ketamine hydrochloride the skin on the right side of the subman-
dibular area was incised, and the tympanic bulla was exposed.
The pharyngeal part of the ET near the tympanic bulla was cut,
and gelatin sponge was inserted into the ET. Two holes were then
made with a 27-gauge needle in the right-side tympanic bulla,
and 10 ␮L of 2.0 mg/mL endotoxin was injected into the bulla
(endotoxin group [n ⫽ 60]). Some animals were had injection of 10
␮L physiological saline instead of endotoxin into the bulla as
control animals (control group [n ⫽ 60]). After injection with
endotoxin, mice were monitored by otomicroscopic observation to
confirm the presence of MEE and tympanic membrane changes.
Obstruction of the ET by granulation tissue at 2 months after
injection with endotoxin is shown in Figure 1.
Histological Preparation
Mice were decapitated under deep anesthesia at 3 days, 2
weeks, and 2 months after intratympanic injection of endotoxin or
saline and were fixed by intracardiac perfusion with 10% neutral-
buffered formalin, periodate lysine paraformaldehyde, and 4% para-
Fig. 1. Obstruction of the eustachian tube (ET). (A) Normal mouse.
(B) Eustachian tube obstruction by granulation tissue (arrows) at 2
months (H&E stain, original magnification ⫻100).
Laryngoscope 114: November 2004
1968
formaldehyde solution for histological evaluation. Then the heads
were removed, decalcified, and embedded in paraffin for hematoxy-
lin and eosin (H&E) staining. Some tissues were embedded in OCT
compound (Sakura Finetechnical Co., Ltd., Tokyo, Japan) for immu-
nohistochemical analysis and in situ hybridization.
Histological Evaluation
Six-micrometer serial paraffin sections containing the tym-
panic bulla were prepared. Sections of the middle ear mucosa
around the tympanic orifice were stained with H&E, and his-
topathological changes were observed microscopically.
Immunohistochemical Analysis for Tumor
Necrosis Factor-␣
Ten-micrometer-thick frozen sections of the OCT-embedded
samples were treated with 3% hydrogen peroxide in absolute
methanol for 20 minutes and were exposed to 5% normal goat
serum in phosphate-buffered saline (PBS) for 30 minutes at room
temperature. Then these sections were incubated with biotinyl-
ated goat anti-mouse TNF-␣ antibody (Pierce Biotechnology, Bos-
ton, MA) diluted 1:3000 in 1% BSA–PBS for 12 hours at room
temperature. After a rinse with PBS, sections were incubated with
ABC reagent (Vector Laboratories, Burlingame, CA) for 1 hour and
were developed in 0.05% 3,3'-diaminobenzidine 0.01% H2O2 sub-
strate medium in 0.1 mol/L phosphate buffer for 8 minutes.
In Situ Hybridization
Fluorescein isothiocyanate–labeled DNA oligonucleotide
(HybriProbe, Biognostik, GmbH, Göttingen, Germany) was used
as a probe to detect specific TNF-␣ messenger RNA (mRNA)
sequences. Frozen, 10-␮m-thick sections were incubated with
probe hybridization solution for 16 hours at 30°C in a moist
chamber. After hybridization, sections were washed with Tris-
buffered saline (TBS) (50 mmol/L Tris-HCl, 150 mmol/L NaCl, pH
7.6) containing 0.1% Triton X-100 and were immersed in 15%
normal rabbit serum diluted in TBS, 3% BSA, and 0.1% Triton
X-100 for 10 minutes at room temperature. After draining of the
solution, the sections were incubated with rabbit F(ab’) anti–
fluorescein isothiocyanate alkaline phosphatase– conjugated an-
tibody (Novocastra Lab, Newcastle, on Tyne, UK) for 30 minutes
at room temperature. The color reaction was developed with
5-bromo-4-chloro-3-indolylphosphate and nitroblue tetrazolium
with 1 mmol/L levamisole hydrochloride at room temperature.
Cytological Study and Cytokine Enzyme-Linked
Immunosorbent Assay
Samples of MEE obtained by myringotomy at the time of
decapitation were centrifuged at 120g for 10 minutes. Superna-
tants were stored at ⫺80°C for cytokine assays. The pellets were
resuspended in physiological saline to make smears and were
examined by cytological studies. The concentration of IL-1␤,
TNF-␣, interferon-␥ (IFN-␥), and interleukin-4 (IL-4) were deter-
mined with commercially available ELISA kits (Quantikine, R&D
Systems, Minneapolis, MN) according to the manufacturer’s in-
structions. Five samples each from the endotoxin and the control
groups were assayed in duplicate, and the results were averaged.
Statistics
Unpaired Student t test was used to evaluate the significant
differences in the concentrations of cytokines in MEE samples. A
P value of less than .05 was considered to be significant.
RESULTS
Middle Ear Effusion and Cytological Study
The production of MEE was found in all mice of
endotoxin group, and 75% (15 of 20) of MEE samples were
Maeda et al.: Cytokine in Otitis Media With Effusion
mucoid at 3 days after injection (Fig. 2). At 2 weeks after
injection, 75% (15 of 20) were serous, and mucoid effusions
had decreased to 25% (5 of 20). At 2 months, 50% (10 of 20)
of effusions were found to be pultaceous. Other effusions
were seromucoid (5% [1 of 20]) and serous (45% [9 of 20]).
Cytological examination revealed that polymorphonuclear
leukocytes (PMNs) had markedly infiltrated the middle
ear cavity by 3 days after injection with endotoxin. Then,
macrophages and mononuclear cells were predominant in
the cells infiltrating into the middle ear cavity at 2 weeks
as well as at 2 months after injection (Fig. 3).
In the control group, all mice had serous MEE at 3
days (20 of 20) and 2 weeks (20 of 20) after injection with
saline. At 2 months, 85% (17 of 20) of MEE samples were
serous, and 15% (3 of 20) were seromucoid MEE samples.
There were few inflammatory cells in the middle ear cav-
ity at any time point in the control group.
Hematoxylin and Eosin Staining
In the control group, the middle ear mucosa was
covered by a single layer of ciliated cells, and a portion was
composed of nonciliated epithelium at 2 months after in-
jection with saline. The subepithelial space appeared
slightly thickened, with only slight infiltration by inflam-
matory cells at all time points (Fig. 4A–C).
In contrast, in the endotoxin group the subepithelial
space was severely thickened with edema, and PMNs in-
filtrated the middle ear mucosa and cavity by 3 days after
injection (Fig. 4D). At 2 weeks after injection, thickening
of the subepithelial space attributable to edema was ob-
served, and a small number of mononuclear cells infil-
trated the middle ear mucosa (Fig. 4E). Macrophages and
mononuclear cells were also present in the middle ear
cavity. At 2 months after injection, subepithelial space
thickening was observed, and a large number of mononu-
clear cells were present in the subepithelial space of the
Fig. 2. Quality of middle ear effu-
sion (MEE). Otitis media with effu-
sion was induced in all mice (120
of 120). In the control group, all
mice showed serous MEE at 3
days (20 of 20) and 2 weeks (20 of
20) after injection. At 2 months,
85% (17 of 20) of MEE samples
were serous, and 15% (3 of 20)
were seromucoid. In the endo-
toxin group, 75% (15 of 20) of
MEE samples were mucoid, and
25% (5 of 20) were serous at 3
days after injection. At 2 weeks,
75% (15 of 20) were serous, and
mucoid effusions decreased to
25% (5 of 20). At 2 months, 50%
(10 of 20) were pultaceous, 5% (1
of 20) were mucoid, and 45% (9
of 20) were serous.
Laryngoscope 114: November 2004
middle ear mucosa (Fig. 4F). The middle ear mucosa was
also altered to a nonciliated, pseudostratified epithelium. In
the middle ear cavity, cellular infiltration was decreased, but
macrophages and mononuclear cells were present.
Immunohistochemical Analysis for Tumor
Necrosis Factor-␣
In the endotoxin group, most PMNs in the middle ear
cavity were stained with anti-TNF-␣ antibody at 3 days
after injection with endotoxin (Fig. 5A). At 2 weeks, mac-
rophages were predominant in the inflammatory cells in
the middle ear cavity and stained with anti-TNF-␣ antibody
(Fig. 5B). Although the number of macrophages was de-
creased at 2 months, the macrophages and some lympho-
cytes migrating into the subepithelial spaces of the middle
ear mucosa were still stained with anti-TNF-␣ antibody (Fig.
5C). Specimens from the control group showed no staining
with anti-TNF-␣ antibody at any time point of observation.
In Situ Hybridization
In the endotoxin group, TNF-␣ mRNA was expressed
in PMNs in the middle ear cavity at 3 days after injection
(Fig. 6A), in macrophages at 2 weeks (Fig. 6B), and in both
macrophages and lymphocytes present in the subepithe-
lial spaces of the middle ear mucosa at 2 months (Fig. 6C).
In the control group, no inflammatory cells in the middle
ear mucosa expressed TNF-␣ mRNA.
Cytokine Enzyme-Linked Immunosorbent Assay
At 3 days after injection with endotoxin, the concen-
tration of IL-1␤ in MEE increased significantly compared
with that of the control group (1013 ⫾ 778 vs. 3.6 ⫾ 7.2
pg/mL [P ⬍ .05]) (Fig. 7). Then the level of IL-1␤ remark-
ably declined, and there were no significant differences in
IL-1␤ levels between the control and endotoxin groups at
2 weeks and at 2 months after injection. In contrast,
Maeda et al.: Cytokine in Otitis Media With Effusion
1969

Fig. 3. Inflammatory cell infiltration of the middle ear cavity in the
endotoxin group. Polymorphonuclear leukocytes infiltrated the mid-
dle ear cavity at 3 days after injection (A). At 2 weeks (B) and 2
months (C), inflammatory cell infiltrates were composed primarily of
macrophages and mononuclear cells (H&E stain, original magnifi-
cation ⫻400).
TNF-␣ levels in the endotoxin group were statistically
higher than those in the control group at all time points
after injection (3 d, 190 ⫾ 78.1 vs. 9.70 ⫾ 11.3 pg/mL [P ⬍
.05]; 2 wk, 71.6 ⫾ 10.2 vs. 1.36 ⫾ 2.11 pg/mL [P ⬍ .05]; 2
mo, 84.9 ⫾ 51.4 vs. 19.9 ⫾ 28.9 pg/mL [P ⬍ .05]). The
concentrations of IFN-␥ in MEE also increased after the
injection with endotoxin and peaked at 2 weeks, but the
differences between endotoxin and control groups were
not significant. Interleukin-4 was not identified in any of
the effusion samples. Cytokine levels in sera were below
the limits of detection in these assays.
DISCUSSION
Chronic OME persisting for 2 months was achieved
in mice both by ET blockage combined with intratympanic
injection of endotoxin and by ET blockage alone. However,
degeneration of the middle ear mucosa with mononuclear
cell infiltration was found in the endotoxin group but not
in the control group. In a human study, Kitajiri et al.7
showed the histopathological appearance of OME in in-
Laryngoscope 114: November 2004
1970
Fig. 4. Histological changes in the middle ear mucosa: control
group at 3 days (A), 2 weeks (B), and 2 months (C) after injection.
Endotoxin group at 3 days (D), 2 weeks (E), and 2 months (F). (A and
B) The middle ear mucosa was covered by a single layer of ciliated
cells. (C) A part of the middle ear mucosa was composed of non
ciliated epithelium. (D) Polymorphonuclear leukocytes have infil-
trated the middle ear mucosa. (E) A small number of mononuclear
cells has infiltrated the middle ear mucosa. (F) Thickening of the
subepithelial space observed, and a large number of mononuclear
cells is present in the subepithelial space of the middle ear mucosa.
The middle ear mucosa is altered to a non ciliated, pseudostratified
epithelium (H&E stain, original magnification ⫻200).
fants to be characterized by the formation of pseudostrati-
fied, cuboidal, or cylindrical epithelium in the middle ear
mucosa. In addition, mononuclear cell infiltration was
found in the ET of patients with OME, particularly in the
pharyngeal half of the cartilaginous portion of the ET, and
was increased by repeated infection.8 These pathological
findings in humans are similar to those observed in our
endotoxin-treated mice, suggesting that inflammatory
changes in the middle ear mucosa affected by chronic OME
might involve inflammatory factors such as endotoxin.
In the present study, mucoid MEE was frequently
found in the endotoxin group, even in the early stage of
OME. In contrast, most MEE samples observed in the
control group were of the serous type. Kuijpers et al.9
investigated the effect of ET obstruction on the middle ear
of rats raised under conventional, specific pathogen-free,
and germ-free conditions and found that ET occlusion
caused serous MEE, and transformation of the epithelium
into mucus-producing cells was observed only in animals
that developed a middle ear infection after ET occlusion.
These results suggest that ET blockage is essential for the
induction of MEE, especially serous MEE, and the addi-
tional inflammatory stimulation of the middle ear mucosa
might be associated with the production of mucoid MEE.
Ondrey et al.10 showed that MEE samples obtained from
adult patients with OME are usually of the serous type,
whereas those from infants are usually of the mucoid type.
Moreover, endotoxin was found in 69% of MEE samples
from children but in only 17% of MEE samples from
adults.11 Thus, endotoxin plays a role in producing mucoid
MEE in humans as well as in animal models.
Recent studies of otitis media have highlighted the
importance of two cytokines, TNF and IL-1, as potent
Maeda et al.: Cytokine in Otitis Media With Effusion
 that inoculation of TNF-␣ into the middle ear cavity, fol-
lowed by ET blockage, enhanced mucin accumulation in
the middle ear mucosa. Moreover, TNF-␣ markedly in-
creased Muc2 mucin mRNA expression in middle ear ep-
ithelium in a time- and dose-dependent manner.18 Thus,
TNF-␣ may be associated with the histopathological
changes in the middle ear mucosa and the production of
mucoid MEE in persistent chronic OME.
In humans, TNF-␣ is thought to be produced mainly
by macrophages or mast cells infiltrating into the middle
ear mucosa and MEE.16 However, the source of TNF-␣ in
animal models of OME remains unclear. In the present
study, for the first time, we showed that TNF-␣ is pro-
duced by PMNs and macrophages in the early stage of
OME and by mononuclear cells migrating into the subep-
ithelial spaces of the middle ear mucosa in the chronic
stage. Further, the concentrations of IFN-␥ in MEE were
higher, although not significantly, in the endotoxin group
than in the control group. Interleukin-4 was not detected

Fig. 5. Immunohistochemical findings for tumor necrosis factor-␣

(TNF-␣) in the endotoxin group at 3 days (A), 2 weeks (B), and 2
months (C) after injection. (A) Polymorphonuclear leukocytes in the
middle ear cavity are stained with anti–TNF-␣ antibody. (B) Macro-
phages in the cavity are stained with anti–TNF-␣ antibody. (C) Some
lymphocytes that migrated into the subepithelial space of the mid-
dle ear mucosa are stained with anti–TNF-␣ antibody.

inflammatory mediators induced by endotoxin exposure

and microbial invasion.2,4,5,12 We previously reported that
IL-1␤ was produced in the inflamed middle ear mucosa in
response to stimulation with endotoxin and that it may
play an important role in the induction of OME.13 In the
present study, IL-1␤ expression was significantly in-
creased in the early stages of OME, whereas TNF-␣ ex-
pression was augmented in both early and chronic stages
of the disease. Yellon et al.5 and Himi et al.12 examined
the concentrations of these cytokines in the MEE of patients
with chronic OME and showed that TNF-␣ correlated with
persistence of the disease. Recombinant human TNF, but
not IL-1, induced experimental OME in guinea pig, and the
development of OME was inhibited by TNF-binding pro- Fig. 6. In situ hybridization for tumor necrosis factor-␣ (TNF-␣)
tein.14,15 These findings suggest that TNF-␣ is an important messenger RNA (mRNA) in the endotoxin group at 3 days (A), 2
factor responsible for the persistence of chronic OME. weeks (B), and 2 months (C) after injection. (A) Tumor necrosis
factor-␣ mRNA is expressed in polymorphonuclear leukocytes in
TNF-␣ is thought to be associated with tissue dam- the middle ear cavity. (B) Tumor necrosis factor-␣ mRNA is ob-
age, fibrosis, and bone resorption in the middle ear of served in macrophages in the middle ear cavity. (C) Lymphocytes in
patients with OME.16 Recently, Kawano et al.17 showed the subepithelial space of the middle ear express TNF-␣ mRNA.

Laryngoscope 114: November 2004 Maeda et al.: Cytokine in Otitis Media With Effusion
1971

in the MEE of either group. Interferon-␥ is known to
promote the production of TNF-␣ by macrophages,
whereas IL-4 decreases this production.19,20 These findings
might explain the consistently high production of TNF-␣
during the entire course of OME in the endotoxin group.
CONCLUSION
The present study showed that chronic OME with
mucoid MEE can be induced in mice by a combination of ET
blockage and intratympanic injection of endotoxin. The re-
sults suggest that ET blockage is essential for the production
of serous MEE and that TNF-␣ expression induced by endo-
toxin is responsible for the production of mucoid MEE and
the histological changes of middle ear mucosa characteristic
of chronic OME. The murine model developed in the present
study should be useful for further investigation of the mech-
anisms regulating production of mucoid MEE and the per-
sistence of OME frequently observed in children.
BIBLIOGRAPHY
1. Miller MB, Koltai PJ, Hetherington SV. Bacterial antigens
and neutrophil granule proteins in middle ear effusions.
Arch Otolaryngol Head Neck Surg 1990;116:335–337.
2. DeMaria TF, Murwin DM. Tumor necrosis factor during ex-
perimental lipopolysaccharide-induced otitis media. La-
ryngoscope 1997;107:369 –372.
3. DeMaria TF, Prior RB, Briggs BR, Lim DJ, Birck HG. Endo-
toxin in middle-ear effusions from patients with chronic
otitis media with effusion. J Clin Microbiol 1984;20:15–17.
4. Willett DN, Rezaee RP, Billy JM, Tighe MB, DeMaria TF.
Relationship of endotoxin to tumor necrosis factor-␣ and
interleukin-1␤ in children with otitis media with effusion.
Ann Otol Rhinol Laryngol 1998;107:28 –33.
5. Yellon RF, Leonard G, Marucha PT, et al. Characterization of
cytokines present in middle ear effusions. Laryngoscope
1991;101:165–169.
6. Westphal O, Jann K, Himmelspach K. Chemistry and immu-
nochemistry of bacterial lipopolysaccharides as cell wall
antigens and endotoxins. Prog Allergy 1983;33:9 –39.
7. Kitajiri M, Sando I, Hashida Y, Doyle WJ. Histopathology of
otitis media in infants with cleft and high-arched palates.
Ann Otol Rhinol Laryngol 1985;94:44 –50.
Laryngoscope 114: November 2004
1972
Fig. 7. Mean quantities of
interleukin-1␤ (IL-1␤), tumor ne-
crosis factor-␣ (TNF-␣), and
interferon-␥ (IFN-␥). In the endo-
toxin group the level of IL-1␤ at 3
days was statistically higher than
that in the control group and de-
clined significantly after 2 weeks.
Tumor necrosis factor-␣ levels in
the endotoxin group decreased
gradually compared with those of
the control group after increasing
initially after injection with endo-
toxin. There were no significant
differences in IFN-␥ levels be-
tween the control and endotoxin
groups.
8. Kamimura M, Sando I, Balaban CD, Haginomori S. Mucosa-
associated lymphoid tissue in middle ear and eustachian
tube. Ann Otol Rhinol Laryngol 2001;110:243–247.
9. Kuijpers W, van der Beek JM, Willart EC. The effect of
experimental tubal obstruction on the middle ear: prelim-
inary report. Acta Otolaryngol 1979;87:345–352.
10. Ondrey FG, Juhn SK, Adams GL. Early-response cytokine
expression in adult middle ear effusions. Otolaryngol Head
Neck Surg 1998;119:342–345.
11. Iino Y, Kaneko Y, Takasaka T. Endotoxin in middle ear
effusions tested with Limulus assay. Acta Otolaryngol
1985;100:42–50.
12. Himi T, Suzuki T, Kodama H, Takezawa H, Kataura A. Immuno-
logic characteristics of cytokines in otitis media with effusion.
Ann Otol Rhinol Laryngol Suppl 1992;157:21–25.
13. Watanabe T, Hirano T, Suzuki M, Kurono Y, Mogi G. Role of
interleukin-1␤ in a murine model of otitis media with ef-
fusion. Ann Otol Rhinol Laryngol 2001;110:574 –580.
14. Catanzaro A, Ryan A, Batcher S, Wasserman SI. The re-
sponse to human rIL-1, rIL-2, and rTNF in the middle ear
of guinea pigs. Laryngoscope 1991;101:271–275.
15. Ball SS, Prazma J, Dais CG, Triana RJ, Pillsbury HC. Role of
tumor necrosis factor and interleukin-1 in endotoxin-
induced middle ear effusions. Ann Otol Rhinol Laryngol
1997;106:633– 639.
16. Ophir D, Hahn T, Schattner A, Wallach D, Aviel A. Tumor
necrosis factor in middle ear effusions. Arch Otolaryngol
Head Neck Surg 1988;114:1256 –1258.
17. Kawano H, Haruta A, Tsuboi Y, et al. Induction of mucous cell
metaplasia by tumor necrosis factor alpha in rat middle ear:
the pathological basis for mucin hyperproduction in mucoid
otitis media. Ann Otol Rhinol Laryngol 2002;111:415– 422.
18. Lin J, Haruta A, Kawano H, et al. Induction of mucin gene
expression in middle ear of rats by tumor necrosis factor-
alpha: potential cause for mucoid otitis media. J Infect Dis
2000;182:882– 887.
19. Held TK, Weihua X, Yuan L, Kalvakolanu DV, Cross AS.
Gamma interferon augments macrophage activation by
lipopolysaccharide by two distinct mechanisms at the sig-
nal transduction level and via an autocrine mechanism
involving tumor necrosis factor alpha and interleukin-1.
Infect Immun 1999;67:206 –212.
20. te Velde AA, Huijbens RJ, Heije K, de Vries JE, Figdor CG.
Interleukin-4 (IL-4) inhibits secretion of IL-1␤, tumor ne-
crosis factor-␣, and IL-6 by human monocytes. Blood 1990;
76:1392–1397.
Maeda et al.: Cytokine in Otitis Media With Effusion