Contents lists available at ScienceDirect

Respiratory Investigation

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8

http://dx.doi.org/102212-5345/& 2015 T

nCorresponding a1These authors

journal homepage: www.elsevier.com/locate/resinv

Original article

Cigarette smoke augments MUC5AC production viathe TLR3-EGFR pathway in airway epithelial cells

Kuninobu Kanaia,1, Akira Koaraib,n,1, Yutaka Shishikurab,Hisatoshi Sugiurab, Tomohiro Ichikawaa, Takashi Kikuchia,Keiichiro Akamatsua, Tsunahiko Hiranoa, Masanori Nakanishia,Kazuto Matsunagaa, Yoshiaki Minakataa, Masakazu Ichinoseb

aThird Department of Internal Medicine, Wakayama Medical University, Wakayama, JapanbDepartment of Respiratory Medicine, Tohoku University Graduate School of Medicine, 1-1 Seiryo-machi, Aoba-ku,Sendai 980-8574, Japan

a r t i c l e i n f o

Article history:

Received 10 September 2014

Received in revised form

5 January 2015

Accepted 21 January 2015

Available online 12 March 2015

Keywords:

COPD

Exacerbation

Mucin

Oxidative stress

Toll-like receptor 3

.1016/j.resinv.2015.01.007he Japanese Respiratory

uthor. Tel.: þ81 22 717 85contributed equally to th

a b s t r a c t

Background: Viral infections are a major cause of chronic obstructive pulmonary disease

(COPD) exacerbations. Toll-like receptor 3 (TLR3) reacts with double-stranded RNA (dsRNA)

and participates in the immune response after viral infection. In the present study, we

examined whether cigarette smoke, which is involved in the pathogenesis of COPD,

enhances mucin production via the TLR3-epidermal growth factor receptor (EGFR) pathway

in airway epithelial cells.

Methods: We studied the effects of cigarette smoke extract (CSE) on signal transduction and

the production of mucin 5AC (MUC5AC) in NCI-H292 cells and differentiated primary

human bronchial epithelial cells stimulated with a synthetic dsRNA analogue, poly-

inosinic-polycytidylic acid [poly(I:C)], used as a TLR3 ligand.

Results: CSE significantly potentiated the production of MUC5AC in epithelial cells stimu-

lated with poly(I:C). Antibodies to EGFR or EGFR ligands inhibited CSE-augmented MUC5AC

release in poly(I:C)-treated cells. Treatment with poly(I:C) or CSE alone increased the

phosphorylation of EGFR and extracellular signal-regulated kinase (ERK). However, after

poly(I:C) stimulation, CSE did not enhance EGFR phosphorylation, but did augment ERK

phosphorylation. EGFR inhibitors and an ERK inhibitor inhibited the augmented release of

MUC5AC. In addition, treatment with N-acetylcysteine, an antioxidant, inhibited the CSE-

augmented phosphorylation of ERK and MUC5AC.

Conclusions: These data show that cigarette smoke increases TLR3-stimulated MUC5AC

production in airway epithelial cells, mainly via ERK signaling. The effect might be

mediated in part by oxidative stress. Modulation of this pathway might be a therapeutic

target for viral-induced mucin overproduction in COPD exacerbation.

& 2015 The Japanese Respiratory Society. Published by Elsevier B.V. All rights reserved.

Society. Published by Elsevier B.V. All rights reserved.

39; fax: þ81 22 717 8549.is work.

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8138

1. Introduction

Chronic obstructive pulmonary disease (COPD) is currently thefourth leading cause of death and a major cause of chronicmorbidity and mortality worldwide [1]. COPD, a chronicinflammatory disease characterized by a specific pattern ofinflammation involving neutrophils, macrophages, and CD8 Tlymphocytes, is caused by long-term exposure to noxiousgasses such as cigarette smoke [1,2]. Goblet cell hyperplasiaand mucus hypersecretion are common features of the dis-ease, especially during exacerbation. Mucociliary clearance is apotent, innate defense system against exogenous insults,including bacteria and viruses. However, excessive mucusproduction causes airway obstruction, which can lead toexacerbation of COPD. It is also associated with diseasemorbidity and mortality [3–6]. However, the mechanism forenhanced mucus production has yet not been fully elucidated.

At least 19 different mucin genes have been identified inhumans. Mucin 5AC (MUC5AC) is a major component of themucus produced by airway epithelial cells [7]. Various factorsinduce airway mucus production, including proinflammatorycytokines [8,9] and bacterial exoproducts [10,11]. Oxidative stressand cigarette smoke, which contains many particles [12] andoxidants, including hydrogen peroxide [13], enhance airwaymucin production [14–16], mainly via the epidermal growthfactor receptor (EGFR) and mitogen-activated protein kinase(MAPK) (including extracellular signal-regulated kinase [ERK])signaling pathways [15,17].

Viral infection is a major cause of COPD exacerbations [18,19].Toll-like receptors (TLRs), which recognize pathogen-associatedmolecular patterns, have a key role in the innate immunesystem [20]. TLR3, which reacts with virus-derived double-stranded RNA, is thought to play a key role in virus-inducedimmune reactions [21]. TLR3 is mainly detected on endosomeswithin airway epithelial cells, dendritic cells, and macrophages[20–22]; its activation leads to the release of proinflammatorymediators and type I interferons [20]. Viral infection and virus-derived dsRNA also induce airway mucus production andenhance MUC5AC expression via TLR3 and the activation ofEGFR and ERK [23,24]. Previously, we and others have shown thatoxidative stress and cigarette smoke augment the TLR3-mediated response in airway epithelial cells [25,26]. Cigarettesmoke also reportedly enhances MUC5AC induction in responseto proinflammatory stimuli [27]. However, the effect of cigarettesmoke on the induction of mucin production by viral-deriveddsRNA via the TLR3-EGFR pathway in airway epithelial cellsremains unclear.

Abbreviations: ALI, air–liquid interface; COPD, chronic obstructive p

stranded RNA; ELISA, enzyme-linked immunosorbent assay; EGFR,

regulated kinase; FBS, fetal bovine serum; GAPDH, glyceraldehyde-3

cell; LPS, lipopolysaccharide; MAPK, mitogen-activated protein kina

C), polyinosinic-polycytidylic acid; ROS, reactive oxygen species; siR

TGF, transforming growth factor; TLR, Toll-like receptor; TNF, tumE-mail addresses: kanaik@wakayama-med.ac.jp (K. Kanai), koarai

Yutaka.Sisikura@mc2.seikyou.ne.jp (Y. Shishikura), sugiura@rm.med.1kawa@wakayama-med.ac.jp (T. Ichikawa), kikuchi@wakayama-med.tsuna@wakayama-med.ac.jp (T. Hirano), masa-n@wakayama-med.ackazmatsu@wakayama-med.ac.jp (K. Matsunaga), minakaty@wakayamichinose@rm.med.tohoku.ac.jp (M. Ichinose).

Here, we used a synthetic dsRNA analogue, polyinosinic-polycytidylic acid [poly(I:C)], as a TLR3 ligand to mimic viralinfection in order to determine whether (a) cigarette smokeaffects poly(I:C)-induced MUC5AC production in airwayepithelial cells and (b) cigarette smoke modulates poly(I:C)-induced TLR3-EGFR signaling.

2. Materials and methods

Detailed methods are provided in the online Supplementarydata.

2.1. Materials

Bafilomycin was purchased from Alexis Biochemicals (Lausen,Switzerland). Poly(I:C) (polyinosinic acid/polycytidylic acid,sodium salt, double-stranded), AG1478, U0126, BIBX1382,TAPI-1, GM6001, and cycloheximide were from Calbiochem(La Jolla, CA). The mouse anti-EGFR neutralizing antibody waspurchased from Calbiochem (La Jolla, CA); other neutralizingantibodies (anti-transforming growth factor [TGF]-α, anti-amphiregulin) were from R&D Systems, Inc. (Minneapolis,MN). Lipopolysaccharide (LPS), imiquimod, and ssRNA40 werefrom InvivoGen (San Diego, CA). N-acetylcysteine (NAC) waspurchased from Sigma-Aldrich (St. Louis, MO).

2.2. Preparation of epithelial cells

NCI-H292 cells, a human pulmonary mucoepidermoid carci-noma cell line, were cultured in RPMI-1640 medium supple-mented with 10% fetal bovine serum (FBS). Cells were grownto 80% confluence in 6-well, 24-well, or 96-well plates andwere maintained in FBS-free medium for 24 h before stimula-tion. Primary normal human bronchial epithelial cells(HBECs) from three different donors were purchased fromLonza (Walkersville, MD) and ScienCell Research Laboratories(Carlsbad, CA). Air–liquid culture of human primary bronchialepithelial cells was performed using Cloneticss B-ALI™ Air–Liquid Interface (ALI) medium. The ALI state was maintainedfor 7–10 days; previous studies have shown this duration isrequired for mucociliary differentiation [28,29].

To investigate the effect of poly(I:C) on the cells,supernatants were harvested 24 h (unless otherwise indicated)after treatment with poly(I:C). Compounds or neutralizing

ulmonary disease; CSE, cigarette smoke extract; dsRNA, double-

epidermal growth factor receptor; ERK, extracellular signal-

-phosphate dehydrogenase; HBEC, human bronchial epithelial

se; MUC5AC, mucin 5AC; NAC, N-acetylcysteine; poly(I:

NA, small interfering RNA; TACE, TNF-α-converting enzyme;

or necrosis factor.@rm.med.tohoku.ac.jp (A. Koarai),tohoku.ac.jp (H. Sugiura),ac.jp (T. Kikuchi), akamatsu@wakayama-med.ac.jp (K. Akamatsu),.jp (M. Nakanishi),a-med.ac.jp (Y. Minakata),

Rel

ativ

e qu

antit

y of

MU

C5A

C m

RN

A

0

10

20

30

40

50

poly(I:C)[ g/ml]

0 0.3 3 10 30

***

1

**

***

MU

C5A

C p

rote

in[F

old

indu

ctio

n]

0

1

2

3

4

time [h] 0 4 8 16 24

*

***

***

MU

C5A

C p

rote

in[%

]

0

50

100

150

0 10 30 100 0 10 30 100Bafilomycin[nM]

vehicle poly(I:C) 10 g/ml

***

+++

TLR3 ~110 kD

positivecontrol

NCI-H292 cell

21 3

TLR3 (ALEXA488) Nucleus (PI)

overlay Negative control

MU

C5A

C p

rote

in[F

old

indu

ctio

n]

0

1

2

3

4

poly(I:C)[ g/ml]

0 0.3 10 301 3

*****

μ

μμ

Fig. 1 – Expression of TLR3 on NCI-H292 cells and the effect of the TLR3 ligand poly(I:C) on MUC5AC production. (A) TLR3immunoreactivity in NCI-H292 cells (A, upper left panel, green), immunofluorescence of the nucleus (A, upper right panel,red), an overlaid image (A, lower left panel), and a negative control (A, lower right panel). The results are representative of atleast three independent samples. Original magnification: �400. (B) TLR3 in NCI-H292 cells was detected at 110 kD inimmunoblot experiments with whole cell extracts in three different experiments. Recombinant human TLR3 was used as thepositive control. PI¼propidium iodide. (C–F) Effect of the TLR3 ligand poly(I:C) on MUC5AC release and the effect of bafilomycinon poly(I:C)-induced MUC5AC release in NCI-H292 cells. (C, D) Cells were treated with various concentrations of poly(I:C). After24 h, whole cells and supernatants were harvested and assayed. The expression of MUC5ACwas assessed with real-time PCR,and the release of MUC5AC was assessed with ELISA. (E) Cells were treated with 10 μg/ml poly(I:C). At various time points afterthe incubation, supernatants were harvested and assayed. (F) The cells were treated with 10 μg/ml poly(I:C) or vehicle in thepresence of various concentrations of bafilomycin, a TLR3 inhibitor that blocks endosomal Hþ-ATPase activity. After 24 h,supernatants were harvested and assayed. The data were expressed as the mean7SEM for three to four separateexperiments. npo0.05, nnpo0.01,nnnpo0.001 compared with values for control cells. þþþpo0.001 compared with the values forpoly(I:C)-treated control cells. (For interpretation of the references to colour in this figure legend, the reader is referred to theweb version of this article.)

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8 139

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8140

antibodies were added to the medium at various concentra-tions 60min before cigarette smoke extract (CSE) or poly(I:C)treatment. To stimulate cells in the ALI state, 100 μl of mediumcontaining CSE or/and poly(I:C) was added to the apical side ofdifferentiated epithelial cells.

Rel

ativ

e qu

antit

y of

MU

C5A

C m

RN

A

0

20

40

60

CSE [%] 0 1 5 10 15 20

***

**

0

2

4

6

8

***

*** ***

***

++

CSE (-)

CSE (+)CSE (-)CSE (+)

poly(I:C) (-)poly(I:C) (+)poly(I:C) (+)

poly(I:C) (-)

MU

C5A

C p

rote

in[F

old

indu

ctio

n]

+++++++

++++++

+++++++++

+++++++++++++

time [h] 0 4 8 16 24

Rel

ativ

e qu

antit

y of

MU

C5A

C m

RN

A

0

20

40

60

80

100CSE 0%CSE 5%

poly(I:C)[ g/ml]

0 0.3 3 10 301

p < 0.05

p < 0.05

p < 0.01p < 0.01

p < 0.01

p < 0.001

***

*****+ +

+++

+++

+++

Rel

ativ

e qu

antit

y of

MU

C5A

C m

RN

A

0

2

4

6

8

10

poly(I:C) [μg/ml] 0 0CSE [%] 0 0

105 5

10

N.S.p < 0.001

p < 0.001

p < 0.001N.S.

μ

2.3. Preparation of CSE

CSE was prepared using the method of Wirtz and Schmidt[30,31]. To estimate its effect, CSE was added to the medium15 min before poly(I:C) treatment [31].

control 5% CSE

poly(I:C) 5% CSE+ poly(I:C)

MU

C5A

C p

rote

in[F

old

indu

ctio

n]

0

1

2

3

CSE [%] 0 1 5 10 15 20

*****

MU

C5A

C p

rote

in[F

old

indu

ctio

n]

0

1

2

3

poly(I:C) [μg/ml] 0 0CSE [%] 0 0

105 5

10

N.S.p < 0.05

p < 0.05

N.S.N.S.

MU

C5A

C p

rote

in[F

old

indu

ctio

n]

0

2

4

6CSE 0%CSE 5%

poly(I:C)[ g/ml]

0 0.3 10 30

+++

+N.S.

N.S.

N.S.

p < 0.01

***

1 3

+++

p < 0.05

**N.S.

μ

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8 141

2.4. Enzyme-linked immunosorbent assay (ELISA)

MUC5AC was measured using ELISA-based methods withsome modification [27,32,33]. The results were calculated asthe fold induction over control. TGF-α and amphiregulin weremeasured using ELISA (R&D Systems) according to the man-ufacturer's instructions.

2.5. Immunocytochemistry

Immunocytochemistry was performed as previouslydescribed [25,26]. A rabbit polyclonal anti-TLR3 antibody(Abcam plc, Cambridge, UK), rabbit IgG antibody (negativecontrol for the TLR3 antibody; Dako, Glostrup, Denmark), ormouse monoclonal anti-MUC5AC antibody (Clone 45M1;Thermo Fisher Scientific, Fremont, CA) was used as theprimary antibody and detected with an Alexa Fluor 488-conjugated goat anti-rabbit antibody (Life Technologies,Grand Island, NY).

2.6. Immunoblotting

Whole cell extracts were prepared, and target proteins weredetected using SDS-PAGE/western blot as previouslydescribed [25,26] with rabbit polyclonal antibodies againstTLR3 (Abcam plc), phospho-EGFR, EGFR, phospho-ERK1/2(Tyr1068), or ERK1/2 (Cell Signaling Technology, Danvers,MA). Recombinant human TLR3 (R&D systems) was used asa positive control.

2.7. Real-time RT-PCR

MUC5AC, TGF-α, and amphiregulin mRNA in epithelial cellswas measured with real-time RT-PCR and analyzed using thecomparative threshold cycle method with normalization toglyceraldehyde-3-phosphate dehydrogenase (GAPDH).

2.8. Statistical analysis

Data were expressed as the mean7SEM. GraphPad Prism(GraphPad Software Inc., San Diego, CA) was used for statis-tical tests. Experiments with multiple comparisons wereevaluated using one way analysis of variance followed byBonferroni's test or Dunnett's test to adjust for multiplecomparisons. A two-tailed Student's t-test was used for singlecomparisons. Significance was defined as po0.05.

Fig. 2 – Effect of cigarette smoke extract (CSE) on MUC5AC produtreated with various concentrations of CSE. After 24 h, whole ceMUC5ACwas assessed with real-time PCR, and the release of MU5% CSE or vehicle 15 min before treatment with various concenassayed. (E) Cells were treated with 5% CSE or vehicle 15 min beafter the incubation, supernatants were harvested and assayed.immunoreactivity in NCI-H292 cells. Pretreatment with 5% CSE inC). Original magnification: �200. (G, H) Differentiated human br15 min before treatment with 10 μg/ml poly(I:C). After 24 h, sampthe mean7SEM for three to four separate experiments. nnpo0.01cells. þpo0.05, þþþpo0.001 compared with values for CSE-treate

3. Results

3.1. Expression of TLR3 on HBECs and the effect of theTLR3 ligand poly(I:C) on MUC5AC production

In a previous study, we demonstrated TLR3 expression onprimary HBECs and Beas2B cells [25]. Here, we used NCI-H292cells, a human pulmonary mucoepidermoid carcinoma cellline, to evaluate TLR3-induced mucin production. NCI-H292cells have been used as a model for mucin production innormal HBECs because they share key components of thesignaling pathways for mucin production with normal cellsand express much higher amounts of MUC5ACmRNA than donormal cells [34,35]. First, we examined TLR3 expressionusing immunocytochemistry and immunoblotting to confirmwhether NCI-H292 cells express TLR3. TLR3 was detected inthe cells (Fig. 1A and B). Next, we investigated the effect of asynthetic dsRNA analogue, poly(I:C), used as a TLR3 ligand tomimic viral infection, on MUC5AC production in NCI-H292cells. Poly(I:C) significantly increased the expression ofMUC5AC in a dose-dependent manner and the release ofMUC5AC from cells in a dose- and time-dependent manner(Fig. 1C–E). TLR3 reportedly resides in endosomes andrequires an acidic environment for activation. To confirmthat TLR3 mediated the effect of poly(I:C) (as previouslyshown by depletion of TLR3 using siRNA [24]), we usedbafilomycin, an inhibitor of endosomal acidification. Bafilo-mycin inhibited poly(I:C)-induced MUC5AC release in a dose-dependent manner (Fig. 1F).

3.2. Effect of cigarette smoke on the production ofMUC5AC in HBECs after poly(I:C) stimulation

To determine the effect of cigarette smoke on MUC5ACproduction, cells were treated with CSE, which significantlyincreased MUC5AC expression in NCI-H292 cells. However,the effect of CSE on the intracellular production and releaseof MUC5AC was small (Fig. 2A, B and Fig. S1A in the onlineSupplementary data). Because treatment with less than 10%CSE did not affect MUC5AC release, we used 5% CSE in thisstudy. Pretreatment with 5% CSE potentiated MUC5ACexpression and MUC5AC release in the presence of poly(I:C)in a time- and concentration-dependent manner (Fig. 2C–E).In addition, potentiation of intracellular MUC5AC productionwas detected with ELISA and immunocytochemistry (Fig. 2F

ction in poly(I:C)-treated cells. (A, B) NCI-H292 cells werells and supernatants were harvested. The expression ofC5AC was assessed with ELISA. (C, D) Cells were treated withtrations of poly(I:C). After 24 h, samples were harvested andfore treatment with 10 μg/ml poly(I:C). At various time points(F) Panels show representative photographs of MUC5ACcreased MUC5AC immunoreactivity in the presence of poly(I:onchial epithelial cells were treated with 5% CSE or vehicleles were harvested and assayed. The data were expressed as,nnnpo0.001 compared with values for vehicle-treated controld control cells or each group. N.S.¼not significant.

0

50

100

150

200

250

poly(I:C) [μg/ml] 0 100CSE [%] 5 00

MU

C5A

C p

rote

in[%

]

105

105

105

IgG 50

TGF50

Amp50

antibody [μg/ml]

IgG 50

IgG 50

IgG 50

***

+++***

+++

+

0

100

200

300

poly(I:C) [μg/ml] 0 10 10

TAPI-1 [μM] 0 10

10

0

0CSE [%] 0 0 5 55

0 0 30

105

MU

C5A

C p

rote

in[%

] +++***

+++ +++

***

0

50

100

150

200

poly(I:C) [μg/ml] 0 10 10

GM6001 [μM] 0 3

10

0

0CSE [%] 0 0 5 55

0 10

105

MU

C5A

C p

rote

in[%

]

30

105

50

105

0

+++***

***

+++

++++++ +++

TGF-

[pg/

ml]

0

2

4

6

8

10

poly(I:C) [μg/ml] 0 0

CSE [%] 0 010

5 510

p < 0.001N.S.p < 0.001

p < 0.05

control 5%CSE poly10 5%+poly 100

50

100

150

200

250

poly(I:C) [μg/ml] 0 0

CSE [%] 0 010

5 510

p < 0.001

Am

phire

gulin

[pg/

ml]

N.S.p < 0.001

p < 0.01

MU

C5A

C p

rote

in[%

]

0

100

200

300

400 anti-mice IgG2Aanti-EGFR IgG

CSE [%] 0 5 50poly(I:C)

[μg/ml]0 0 1010

***

***+++

p < 0.001

p < 0.001

α

α

Fig. 3 – Effect of cigarette smoke on the release of EGFR ligands from NCI-H292 cells after poly(I:C) stimulation. (A, B) Aneutralizing antibody against EGFR or an EGFR ligand (TGF-α or amphiregulin) or a negative control antibody (anti-mouseIgG2A or anti-goat antibody) was added to NCI-H292 cells 1 h before CSE treatment. Cells were then cultured in the presence of10 μg/ml poly(I:C). After 24 h, supernatants were harvested, and the release of MUC5AC was assessed with ELISA. (C, D)Various concentrations of a specific TNF-α-converting enzyme (TACE) inhibitor, TAPI, or a general inhibitor ofmetalloprotease, GM6001, were added 1 h before CSE treatment. Cells were then cultured in the presence of 10 μg/ml poly(I:C).After 24 h, supernatants were harvested, and the release of MUC5AC was assessed with ELISA. (E, F) NCI-H292 cells weretreated with 5% CSE or vehicle 15 min before treatment with 10 μg/ml poly(I:C). After 24 h, supernatants were harvested, andthe release of TGF-α and amphiregulin was assessed with ELISA. The data were expressed as the mean7SEM for three to fourseparate experiments. nnnpo0.001 compared with the values for vehicle-treated cells; þpo0.05, þþþpo0.001 compared withthe values for CSE- and poly(I:C)-treated cells. N.S.¼not significant.

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8142

and Fig. S1B in the online Supplementary data). Furthermore,the potentiating effect of CSE on MUC5AC expression andMUC5AC release in the presence of poly(I:C) was confirmed indifferentiated HBECs, which mimic in vivo features (Fig. 2Gand H).

3.3. Effect of cigarette smoke on the release of EGFRligands from NCI-H292 cells after poly(I:C) stimulation

To investigate the mechanism of CSE-augmented MUC5ACrelease in HBECs, the effects of CSE on the release of EGFR

pEG

FR/E

GFR

[rel

ativ

e in

tens

ity]

0

1

2

3

4

poly(I:C) [ g/ml] 0 1010 100CSE [%] 0 00 55

105

time [h] 0 241 1 4

052

054

1002

1051

pER

K/E

RK

[rel

ativ

e in

tens

ity]

0

10

20

30

poly(I:C) [ g/ml] 0 1010 100CSE [%] 0 00 55

105

time [h] 0 241 1 4

052

054

1002

1051

pERK

ERK

poly(I:C) [μg/ml]

CSE [%]

time [h] 0 1 2 4 1 2 4 1 2 4

0 5 5 5 0 0 0 5 5 5

0 0 0 0 10 10 10 10 10 10

pEGFR

EGFR

poly(I:C) [μg/ml]

CSE [%]

time [h] 0 1 2 4 1 2 4 1 2 4

0 5 5 5 0 0 0 5 5 5

0 0 0 0 10 10 10 10 10 10

μμ

μ

***

*****

p < 0.001p < 0.001 N.S.

***

***

p < 0.001N.S. p < 0.001

***

**

Fig. 4 – Effect of CSE on the phosphorylation of EGFR and ERKin poly(I:C)-treated NCI-H292 cells. (A,C) Cells were treatedwith 5% CSE or vehicle 15 min before treatment with 10 μg/mlpoly(I:C). At various time points, whole cell lysates wereobtained. EGFR and ERK1/2 phosphorylation was evaluatedwith immunoblotting. (B,D) Band intensities were assessedwith densitometry. The relative intensity was calculated bydividing the phosphorylated EGFR or ERK1/2 band intensityby the EGFR or ERK1/2 band intensity. npo0.05, nnpo0.01,nnnpo0.001 compared with values for vehicle-treated cells at0 min. The data were expressed as the mean7SEM for eightor nine separate experiments. N.S.¼not significant.

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8 143

ligands were evaluated in NCI-H292 cells. Increased sheddingof EGFR pro-ligands induced by tumor necrosis factor (TNF)-α-converting enzyme (TACE) reportedly affects EGFR activationand mucin induction [24]. Consistently, in the current study,the release of the EGFR ligands TGF-α and amphiregulinincreased in NCI-H292 cells treated with poly(I:C). Further-more, pretreatment with neutralizing antibodies againstEGFR or EGFR ligands or with TACE inhibitors (a TACE-specific inhibitor, TAPI, or a general inhibitor of metallopro-tease, GM6001) inhibited poly(I:C)-induced MUC5AC release(data not shown). Pretreatment with neutralizing antibodiesto EGFR or EGFR ligands, particularly amphiregulin, inhibitedCSE-augmented MUC5AC release in poly(I:C)-treated cells(Fig. 3A and B). Furthermore, pretreatment with TAPI orGM6001 inhibited CSE-augmented MUC5AC release in poly(I:C)-treated cells (Fig. 3C and D). In addition, pretreatment with5% CSE potentiated the release of TGF-α and amphiregulin inthe presence of poly(I:C) at 24 h (Fig. 3E and F). To clarify theinvolvement of newly synthesized EGFR ligands in MUC5ACproduction after poly(I:C) treatment, we evaluated the effectof cycloheximide, which inhibits protein translation, onMUC5AC expression. Pretreatment with cycloheximide sig-nificantly inhibited poly(I:C)-stimulated MUC5AC expressionat 6 h (Fig. S2 in the online Supplementary data).

3.4. EGFR-ERK signalling pathways in CSE-stimulatedMUC5AC production

The effects of CSE on the EGFR-ERK signaling pathways werealso evaluated in poly(I:C)-treated NCI-H292 cells. Thesesignaling pathways reportedly affect poly(I:C)-inducedMUC5AC expression [24], which we confirmed by demonstrat-ing the inhibitory effects of EGFR inhibitors (AG1478 andBIBX1382) and an ERK1/2 inhibitor (U0126) on ERK1/2 phos-phorylation and MUC5AC production (data not shown). Treat-ment with poly(I:C) or CSE alone increased EGFR and ERK1/2phosphorylation after 4 h in a time-dependent manner(Fig. 4A–D). Contrary to our expectations, pretreatment withCSE did not enhance EGFR phosphorylation in poly(I:C)-treated NCI-H292 cells, but did augment ERK1/2 phosphoryla-tion in a time-dependent manner (Fig. 4A–D). Pretreatmentwith an EGFR inhibitor, AG1478, almost completely inhibitedpoly(I:C)-induced ERK phosphorylation, with or without CSEtreatment (Fig. 5A and B). Pretreatment with an EGFR inhi-bitor, AG1478 or BIBX1382, or an ERK1/2 inhibitor, U0126,inhibited CSE-augmented MUC5AC expression and CSE-augmented MUC5AC release in a dose-dependent mannerin poly(I:C)-treated cells (Fig. 5C–G). In addition, pretreatmentwith bafilomycin inhibited poly(I:C)-induced ERK1/2 phos-phorylation (data not shown) and CSE-augmented MUC5ACrelease in poly(I:C)-treated cells (Fig. 5H).

3.5. Involvement of oxidative stress in CSE-augmentedMUC5AC release and ERK1/2 phosphorylation

To clarify the involvement of oxidative stress in CSE-augmented MUC5AC release, the effect of NAC, an antiox-idant, was evaluated. Pretreatment with NAC decreased poly(I:C)-induced MUC5AC release and ERK phosphorylat-ion (Fig. 6A–C). Furthermore, NAC pretreatment reduced

CSE-augmented MUC5AC release and ERK phosphorylationto the same degree observed in poly(I:C)-treated cells withoutCSE treatment (Fig. 6A–C).

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8144

3.6. Cell viability

The effects of the compounds, neutralizing antibodies, CSE,and poly(I:C) on cell viability were assessed with the

0

50

100

150

200

250

poly(I:C) [μg/ml] 0 10 10

AG1478 [μM] 0 0.1

10

0

0CSE [%] 0 0 5 55

0 0 0.5

105

MU

C5A

C p

rote

in[%

]

2

105

1

105

***

+++***

+++++++++

+++

0

50

100

150

200

poly(I:C) [μg/ml] 0 10 10

BIBX1382 [μM] 0 0.1

10

0

0CSE [%] 0 0 5 55

0 0.3

105

MU

C5A

C p

rote

in[%

]

1

105

3

105

0

+++***

+++++++++

+++

***

Rel

ativ

e qu

antit

y of

MU

C5A

C m

RN

A [%

]

0

50

100

150

200AG 1478 0 MAG 1478 0.5 M

***

***++

p < 0.001

p < 0.001

CSE [%] 0 5 50poly(I:C)

[μg/ml]0 0 1010

pERK

ERK

poly(I:C) [μg/ml]

CSE [%]AG1478 [μM] 0 0.5 0 0.5 0 0.5

0 0 0 0 5 5

0 0 10 10 10 10

μμ

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide(MTT) assay [25]. Cell viability was482.5% at the highest doseof the compound or neutralizing antibody after CSE and poly(I:C) treatment (data not shown).

0

100

200

300

poly(I:C) [μg/ml] 0 10 10

U0126 [μM] 0 0.2

10

0

0CSE [%] 0 0 5 55

0 0 2

105

MU

C5A

C p

rote

in[%

]

10

105

20

105

+++***

+++++++++

+++

***

0

50

100

150

200

poly(I:C) [μg/ml] 0 10 10 100CSE [%] 0 0 5 55

105

MU

C5A

C p

rote

in[%

]

105

105

Bafilomycin [nM] 0 300 0 10 30 100

++***

+++

++

***

Rel

ativ

e qu

antit

y of

MU

C5A

C m

RN

A [%

]

0

50

100

150

200 U0126 0 MU0126 2uM

CSE [%] 0 5 50poly(I:C)

[μg/ml]0 0 1010

***

***

++

p < 0.001

p < 0.001

*** ***

0

1

2

3

4

5

*

***

p < 0.01

p < 0.001

p < 0.001

pER

K/E

RK

[rel

ativ

e in

tens

ity]

poly(I:C)[ g/ml] 0 0 10 10 10 10CSE[%] 0 0 0 0 5 5

AG1478[μM] 0 0.5 0 0.5 0 0.5

μ

μ

pERK

ERK

poly(I:C) [μg/ml]

CSE [%]

NAC [mM] 0 10 0 10 0 10

0 0 0 0 5 5

0 0 10 10 10 10

0

1

2

3

4

5

6

7 ***

***

p < 0.05

p < 0.05

p < 0.001

pER

K/E

RK

[rel

ativ

e in

tens

ity]

poly(I:C) [μg/ml] 0 10 10

NAC [mM] 0 0

10

0

0CSE [%] 0 0 0 50

10 10 10

105

0

1

2

3

4

5

6

***

***

p < 0.01

N.S

p < 0.001

MU

C5A

C[h

old

incr

ease

]

p < 0.001

poly(I:C) [μg/ml] 0 10 10

NAC [mM] 0 0

10

0

0CSE [%] 0 0 0 50

10 10 10

105

Fig. 6 – Effect of N-acetylcysteine (NAC) on CSE-augmentedMUC5AC release and ERK1/2 phosphorylation in poly(I:C)-treated NCI-H292 cells. NAC was added 1 h before CSEtreatment, and cells were cultured in the presence of 10 μg/ml poly(I:C). (A) After 24 h, supernatants were harvested,and the release of MUC5AC was assessed using ELISA. (B)After 4 h, whole cell lysates were obtained. Phosphorylationof ERK1/2 was evaluated with immunoblotting. (C) Bandintensity was assessed with densitometry. The relativeintensity was calculated by dividing the phosphorylatedERK1/2 band intensity by the ERK1/2 band intensity. Thedata were expressed as the mean7SEM for four to eightseparate experiments. nnnpo0.001 compared with controlvalues. N.S.¼not significant.

Fig. 5 – Effect of EGFR, ERK1/2, and TLR3 inhibitors on CSE-augmpoly(I:C)-treated NCI-H292 cells. AG1478, an EGFR inhibitor, or vcultured in the presence of 10 μg/ml poly(I:C). (A) After 4 h, whoevaluated with immunoblotting. (B) Band intensities were assesby dividing the phosphorylated ERK1/2 band intensity by the ERmean7SEM for five separate experiments. npo0.05, nnnpo0.001BIBX1382 (EGFR inhibitors), U0126 (ERK1/2 inhibitor), bafilomycinor vehicle was added 1 h before CSE treatment. Cells were thenextracts were assayed for MUC5AC expression with real-time PCwith ELISA (E–H). The data were expressed as the mean7SEM fwith values for vehicle-treated cells; þþpo0.01, þþþpo0.001 com

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8 145

4. Discussion

In the current study, we confirmed that a synthetic dsRNAanalogue, poly(I:C), used as a TLR3 ligand to mimic viralinfection, induced MUC5AC production via an EGFR-ERKpathway in NCI-H292 cells. Pretreatment with CSE poten-tiated poly(I:C)-induced MUC5AC production in NCI-H292cells and differentiated primary human airway epithelialcells, suggesting that cigarette smoke potentiates dsRNA-induced MUC5AC production in human airway epithelialcells. We also showed that neutralizing antibodies (to EGFR,TGF-α, or amphiregulin), a specific TACE inhibitor (TAPI-1),and a metalloprotease inhibitor (GM6001) inhibited CSE-potentiated MUC5AC release in poly(I:C)-treated cells, sug-gesting that cigarette smoke augments the production ofMUC5AC induced by dsRNA via EGFR ligands, which are shedby TACE and metalloprotease. Pretreatment with CSE did notenhance EGFR phosphorylation, but did augment ERK phos-phorylation after poly(I:C) stimulation. The augmentedMUC5AC release was inhibited by EGFR inhibitors (AG1378and BIBX1382) and an ERK inhibitor (U0126) (Fig. 7). Thesedata suggest that cigarette smoke potentiates TLR3-stimulated MUC5AC production in airway epithelial cells,mainly through the increased phosphorylation of ERK.

During COPD exacerbation, excessive mucus productionobstructs the airways, which can lead to enhanced inflam-mation and further exacerbation [1]. Although the mechan-ism underlying the enhanced mucus production has not beenfully elucidated, our findings might help explain it, at least invirus-induced exacerbation. When viral infections occur,

TLR3ERK1/2

Bafilomycin

MUC5ACNucleus

EGFR

Endosome

P

P

U0126

AG1478BIBX1382

N-acetylcysteine

TAPI-1GM6001 TACE Cigarette smoke

Reactive oxygenspecies

N-acetylcysteine

TGF-αAmphiregulin

Pro-TGF-αPro-Amphiregulin

dsRNA (poly I:C)

Fig. 7 – Schematic representation of the effect of cigarettesmoke on the production of MUC5AC induced by TLR3stimulation.

ented ERK1/2 phosphorylation and MUC5AC production inehicle was added 1 h before CSE treatment. Cells were thenle cell lysates were obtained. ERK1/2 phosphorylation wassed with densitometry. The relative intensity was calculatedK1/2 band intensity. The data were expressed as thecompared with the values for control cells. AG1478 or(TLR3 inhibitor that blocks endosomal Hþ-ATPase activity),

cultured in the presence of 10 μg/ml poly(I:C). After 24 h, cellR (C, D) and supernatants were assayed for MUC5AC releaseor three to four separate experiments. nnnpo0.001 comparedpared with values for CSE- and poly(I:C)-treated cells.

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8146

TLR3-mediated mucin production increases in airways; themucin production could be augmented in smokers and inCOPD, a smoking-related disease. Regarding TLR3 activationin response to cigarette smoke, genetic depletion of TLR3 in amurine model inhibited smoking-enhanced airway inflam-mation and remodeling after influenza virus infection [36].The cigarette smoke-induced augmentation of airway inflam-mation under TLR3 stimulation was confirmed by anothergroup [37]. Although these studies lack information aboutmucin production, mucus overproduction triggered by cigar-ette smoke exposure under TLR3 stimulation might affect thepathophysiology of enhanced airway inflammation.

To elucidate the effect of cigarette smoke on MUC5ACproduction, we used CSE, a widely accepted model system forevaluating the in vitro effects of cigarette smoke [15,26,27].However, the 24-h period of exposure to 5% CSE might nothave been optimal for mimicking natural cigarette smokeexposure in smokers. There is scant literature on the validityof a comparison between CSE and in vivo exposure tocigarette smoke. However, the concentration of cigarettesmoke in the in vivo exposure is likely diluted, and highconcentrations of CSE are less likely to be relevant in vivo[38,39]. In the current study, the 15–20% CSE needed to induceMUC5AC protein in the cell lysate or supernatant was less thanthe concentration needed to induce MUC5AC mRNA expres-sion. This might mean that cigarette smoke also affectedMUC5AC production via post-translational modulation.

Cigarette smoke contains many particles and oxidants[12,13], but the mechanism by which cigarette smoke potenti-ates MUC5AC production under TLR3 stimulation is unclear.Oxidative stress might be involved, given that oxidativestress reportedly induces MUC5AC production in NCI-H292cells [16,40] and that pretreatment with the antioxidant NACreverses CSE-augmented MUC5AC production induced by LPSor TNF-α [27]. In the current study, pretreatment with NACinhibited CSE-potentiated MUC5AC production and ERK phos-phorylation under TLR3 stimulation, consistent with previousresults. These results suggest that oxidative stress partlymediates the potentiating effect of CSE on TLR3-stimulatedMUC5AC production. NAC pretreatment also diminishedMUC5AC production and ERK phosphorylation in poly(I:C)-treated cells without CSE treatment in the current study. Theresult suggests that poly(I:C) induces reactive oxygen species(ROS) generation, which could contribute to ERK phosphor-ylation and MUC5AC production mediated by TACE activa-tion, in a pathway involving EGFR ligand shedding and EGFRactivation [34,41–43]. In addition, poly(I:C) has been reportedto induce ROS in NCI-H292 cells via Duox2, an NADPH oxidasefamily member that is a major source of ROS in airwayepithelial cells [44]. The same study demonstrated that NACtreatment inhibited poly(I:C)-induced ROS generation [44],supporting the notion that ROS generation is involved inthe mechanism leading to ERK phosphorylation and MUC5ACproduction under TLR3 stimulation.

MUC5AC induction by cigarette smoke reportedly dependson EGFR activation [14]. Increased shedding of the EGFR pro-ligand by TACE, which is activated by cigarette smoke, alsoaffects EGFR activation and mucin induction [15]. In thecurrent study, we demonstrated that neutralizing antibodies(to EGFR, TGF-α, or amphiregulin), a specific TACE inhibitor

(TAPI-1), or a general metalloprotease inhibitor (GM6001)partly inhibited CSE-augmented MUC5AC release, consistentwith previous results. These results suggest that the aug-mented release of EGFR ligands partly mediates the poten-tiating effect of CSE on TLR3-stimulated MUC5AC production.In addition, Baginski et al. have reported that cigarette smokeincreases the production of EGFR ligands (TGF-α and amphir-egulin) induced by bacterial exoproducts or proinflammatorystimulation at 24 h [27], which supports the current findingthat CSE augmented the release of EGFR ligands (TGF-α andamphiregulin) under TLR3 stimulation at 24 h. In the currentstudy, pretreatment with cycloheximide, which inhibits pro-tein translation, inhibited poly(I:C)-stimulated MUC5ACexpression at 6 h, suggesting the involvement of newlysynthesized EGFR ligands in MUC5AC production after poly(I:C) treatment. However, the inhibitory effect of cyclohex-imide on poly(I:C)-stimulated MUC5AC expression might bedue to other newly synthesized proteins or a non-specificeffect of cycloheximide. The extent to which newly synthe-sized EGFR ligands are involved in poly(I:C)-stimulatedMUC5AC production at 24 h remains unknown; further stu-dies are needed to clarify this point.

Our results suggest that an increase in EGFR ligands isrelated to the effect of CSE on poly(I:C)-induced MUC5ACproduction. However, 5% CSE did not increase EGFR ligandrelease. Rather, 5% CSE induced EGFR phosphorylation; theeffect was significantly stronger than that of poly(I:C), with orwithout CSE treatment. These results suggest that 5% CSE-induced EGFR phosphorylation is independent of EGFRligands. Previous studies have suggested that both ligand-dependent and ligand-independent EGFR activation path-ways lead to mucin production [45,46]. Cigarette smoke andoxidative stress have been reported to lead directly totyrosine phosphorylation and EGFR activation without EGFRligand binding, consistent with the present results [14,47].In addition, CSE did not enhance EGFR phosphorylation inpoly(I:C)-treated cells, which is inconsistent with the CSE-augmented release of EGFR ligands. The amount of EGFRligands released in response to CSE might be insufficient topotentiate EGFR phosphorylation. Alternatively, other tyro-sine residues in EGFR, not detected by the phospho-EGFreceptor (Tyr1068) antibody used here, might be phosphory-lated and involved in signal activation after CSE and poly(I:C)stimulation.

EGFR inhibitors almost completely abolished CSE-enhancedMUC5AC expression and secretion in poly(I:C)-treated cells andabolished the increase in ERK phosphorylation. These resultsconfirm that EGFR functions upstream of ERK signaling, whichis crucial for MUC5AC production [24]. However, CSE did notenhance EGFR phosphorylation in poly(I:C)-treated cells. Con-versely, CSE increased poly(I:C)-induced ERK phosphorylation,and an ERK inhibitor abolished CSE-augmented MUC5ACrelease. These results suggest that the synergistic augmentationof ERK phosphorylation by CSE in poly(I:C)-treated cells does notoccur through EGFR. CSE might potentiate ERK phosphorylationthrough an EGFR-independent pathway in poly(I:C)-treated cells.Reportedly, CSE not only induces ERK phosphorylation [48,49],but might enhance ERK phosphorylation in response to proin-flammatory cytokines [27]. Pace et al. have evaluated the effectsof LPS (a TLR4 ligand) and CSE co-treatment on ERK

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8 147

phosphorylation in 16-HBE cells. However, CSE did not augmentLPS-induced ERK phosphorylation [50], which is inconsistentwith our results. This discrepancy might derive from differencesin the stimulated TLR or in the experimental conditions,including the CSE concentration and time points evaluated. Inthe current study, NAC pretreatment significantly reduced theCSE-potentiated phosphorylation of ERK under TLR3 stimula-tion, suggesting that oxidative stress partly mediates the poten-tiating effect of CSE on the TLR3-stimulated phosphorylation ofERK. However, the mechanism remains unclear, and furtherstudies are needed.

This study has several limitations. First, poly(I:C) might actas a secretagogue for mucin. We mainly evaluated MUC5ACin the cell supernatant and found that secretion increased.However, we also evaluated MUC5AC mRNA and proteinlevels using real-time PCR, western blot, and immunocyto-chemistry to confirm the production. Second, the 1.6-foldincrease in MUC5AC production induced by exposure tocigarette smoke is less than that reported in previous studies[14,15]. The discrepancy might reflect different experimentalconditions, including the method of making CSE and itsdegree of dilution.

5. Conclusions

We have shown that cigarette smoke potentiates TLR3-stimulated MUC5AC production in airway epithelial cells byincreasing ERK signaling. The results imply that cigarettesmoke increases mucus production in the airways of smokersand COPD patients during viral infection. The pathwaymight be a therapeutic target for the treatment of COPDexacerbations.

Conflict of interest

The authors have no conflicts of interest to declare.

Contributions

KK, AK, and YS analyzed the data and drafted the manu-script. KK, AK, HS, and MI contributed to the conception anddesign of the original study and contributed substantially tothe manuscript. TI, TK, KA, TH, MN, KM, and YM assistedwith data analysis and interpretation and supervised thestatistical analysis. All authors approved the final version forpublication.

Acknowledgments

We thank Mr. Brent Bell for reading the manuscript. We alsoacknowledge Ms. Rika Horiuchi and Ms. Yasuko Matsumotofor technical assistance, mainly with immunostaining andwestern blotting.

This work was supported by Grant-in-Aid for ScientificResearch (C) from the Japan Society for the Promotion ofScience (JSPS KAKENHI Grant no. 24591135).

Appendix A. Supplementary data

Supplementary data associated with this article can be foundin the online version at http://dx.doi.org/10.1016/j.resinv.2015.01.007.

r e f e r e n c e s

[1] Global initiative for chronic obstructive lung disease[Internet]. Global strategy for the diagnosis, management,and prevention of chronic obstructive pulmonary disease(updated Dec 2011). [cited 2013 Feb 01]. Available from:⟨http://www.goldcopd.org/guidelines-global-strategy-for-diagnosis-management.html⟩.

[2] Barnes PJ. Mediators of chronic obstructive pulmonarydisease. Pharmacol Rev 2004;56:515–48.

[3] Vestbo J, Prescott E, Lange P. Association of chronic mucushypersecretion with FEV1 decline and chronic obstructivepulmonary disease morbidity. Copenhagen City Heart StudyGroup. Am J Respir Crit Care Med 1996;153:1530–5.

[4] Vestbo J, Hogg JC. Convergence of the epidemiology andpathology of COPD. Thorax 2006;61:86–8.

[5] Speizer FE, Fay ME, Dockery DW, et al. Chronic obstructivepulmonary disease mortality in six U.S. cities. Am Rev RespirDis 1989;140:S49–55.

[6] Lange P, Nyboe J, Appleyard M, et al. Relation of ventilatoryimpairment and of chronic mucus hypersecretion tomortality from obstructive lung disease and from all causes.Thorax 1990;45:579–85.

[7] Hovenberg HW, Davies JR, Carlstedt I. Different mucins areproduced by the surface epithelium and the submucosa inhuman trachea: identification of MUC5AC as a major mucinfrom the goblet cells. Biochem J 1996;318(Pt 1):319–24.

[8] Perrais M, Pigny P, Copin MC, et al. Induction of MUC2 andMUC5AC mucins by factors of the epidermal growth factor(EGF) family is mediated by EGF receptor/Ras/Raf/extracellular signal-regulated kinase cascade and Sp1. J BiolChem 2002;277:32258–67.

[9] Song KS, Lee WJ, Chung KC, et al. Interleukin-1 beta andtumor necrosis factor-alpha induce MUC5AC overexpressionthrough a mechanism involving ERK/p38 mitogen-activatedprotein kinases-MSK1-CREB activation in human airwayepithelial cells. J Biol Chem 2003;278:23243–50.

[10] Li JD, Feng W, Gallup M, et al. Activation of NF-κB via a Src-dependent Ras-MAPK-pp90rsk pathway is required forPseudomonas aeruginosa-induced mucin overproduction inepithelial cells. Proc Natl Acad Sci USA 1998;95:5718–23.

[11] Shao MX, Ueki IF, Nadel JA. Tumor necrosis factor alpha-converting enzyme mediates MUC5AC mucin expression incultured human airway epithelial cells. Proc Natl Acad SciUSA 2003;100:11618–23.

[12] Church DF, Pryor WA. Free-radical chemistry of cigarettesmoke and its toxicological implications. Environ HealthPerspect 1985;64:111–26.

[13] Pryor WA, Stone K. Oxidants in cigarette smoke. Radicals,hydrogen peroxide, peroxynitrate, and peroxynitrite. Ann NY Acad Sci 1993;686:12–27.

[14] Takeyama K, Jung B, Shim JJ, et al. Activation of epidermalgrowth factor receptors is responsible for mucin synthesisinduced by cigarette smoke. Am J Physiol Lung Cell MolPhysiol 2001;280:L165–72.

[15] Shao MX, Nakanaga T, Nadel JA. Cigarette smoke inducesMUC5AC mucin overproduction via tumor necrosis factor-alpha-converting enzyme in human airway epithelial (NCI-

r e s p i r a t o r y i n v e s t i g a t i o n 5 3 ( 2 0 1 5 ) 1 3 7 – 1 4 8148

H292) cells. Am J Physiol Lung Cell Mol Physiol 2004;287:L420–7.

[16] Takeyama K, Dabbagh K, Jeong Shim J, et al. Oxidative stresscauses mucin synthesis via transactivation of epidermalgrowth factor receptor: role of neutrophils. J Immunol2000;164:1546–52.

[17] Takeyama K, Dabbagh K, Lee HM, et al. Epidermal growthfactor system regulates mucin production in airways. ProcNatl Acad Sci USA 1999;96:3081–6.

[18] Wedzicha JA. Role of viruses in exacerbations of chronicobstructive pulmonary disease. Proc Am Thorac Soc2004;1:115–20.

[19] Sethi S, Mallia P, Johnston SL. New paradigms in thepathogenesis of chronic obstructive pulmonary disease II.Proc Am Thorac Soc 2009;6:532–4.

[20] Akira S, Uematsu S, Takeuchi O. Pathogen recognition andinnate immunity. Cell 2006;124:783–801.

[21] Guillot L, Le Goffic R, Bloch S, et al. Involvement of Toll-likereceptor 3 in the immune response of lung epithelial cells todouble-stranded RNA and influenza A virus. J Biol Chem2005;280:5571–80.

[22] Johnsen IB, Nguyen TT, Ringdal M, et al. Toll-like receptor 3associates with c-Src tyrosine kinase on endosomes toinitiate antiviral signaling. EMBO J 2006;25:3335–46.

[23] Londhe V, McNamara N, Lemjabbar H, et al. Viral dsRNAactivates mucin transcription in airway epithelial cells. FEBSLett 2003;553:33–8.

[24] Zhu L, Lee PK, Lee WM, et al. Rhinovirus-induced majorairway mucin production involves a novel TLR3-EGFR-dependent pathway. Am J Respir Cell Mol Biol 2009;40:610–9.

[25] Koarai A, Sugiura H, Yanagisawa S, et al. Oxidative stressenhances Toll-like receptor 3 response to double-strandedRNA in airway epithelial cells. Am J Respir Cell Mol Biol2010;42:651–60.

[26] Hudy MH, Traves SL, Wiehler S, et al. Cigarette smokemodulates rhinovirus-induced airway epithelial cellchemokine production. Eur Respir J 2010;35:1256–63.

[27] Baginski TK, Dabbagh K, Satjawatcharaphong C, et al.Cigarette smoke synergistically enhances respiratory mucininduction by proinflammatory stimuli. Am J Respir Cell MolBiol 2006;35:165–74.

[28] Chibana K, Trudeau JB, Mustovitch AT, et al. IL-13 inducedincreases in nitrite levels are primarily driven by increases ininducible nitric oxide synthase as compared with effects onarginases in human primary bronchial epithelial cells. ClinExp Allergy 2008;38:936–46.

[29] Jiang J, Malavia N, Suresh V, et al. Nitric oxide gas phaserelease in human small airway epithelial cells. Respir Res2009;10:3.

[30] Wirtz H, Schmidt M. Acute influence of cigarette smoke onsecretion of pulmonary surfactant in rat alveolar type II cellsin culture. Eur Respir J 1996;9:24–32.

[31] Koarai A, Yanagisawa S, Sugiura H, et al. Cigarette smokeaugments the expression and responses of Toll-like receptor3 in human macrophages. Respirology 2012;17:1018–25.

[32] Mata M, Sarria B, Buenestado A, et al. Phosphodiesterase 4inhibition decreases MUC5AC expression induced byepidermal growth factor in human airway epithelial cells.Thorax 2005;60:144–52.

[33] Cortijo J, Mata M, Milara J, et al. Aclidinium inhibitscholinergic and tobacco smoke-induced MUC5AC in humanairways. Eur Respir J 2011;37:244–54.

[34] Shao MX, Nadel JA. Dual oxidase 1-dependent MUC5ACmucin expression in cultured human airway epithelial cells.Proc Natl Acad Sci USA 2005;102:767–72.

[35] Tadaki H, Arakawa H, Mizuno T, et al. Double-stranded RNAand TGF-α promote MUC5AC induction in respiratory cells.J Immunol 2009;182:293–300.

[36] Kang MJ, Lee CG, Lee JY, et al. Cigarette smoke selectivelyenhances viral PAMP- and virus-induced pulmonary innateimmune and remodeling responses in mice. J Clin Invest2008;118:2771–84.

[37] Kimura G, Ueda K, Eto S, et al. Toll-like receptor 3 stimulationcauses corticosteroid-refractory airway neutrophilia andhyperresponsiveness in mice. Chest 2013;144:99–105.

[38] Culpitt SV, Rogers DF, Shah P, et al. Impaired inhibition bydexamethasone of cytokine release by alveolar macrophagesfrom patients with chronic obstructive pulmonary disease.Am J Respir Crit Care Med 2003;167:24–31.

[39] Oltmanns U, Chung KF, Walters M, et al. Cigarette smokeinduces IL-8, but inhibits eotaxin and RANTES release fromairway smooth muscle. Respir Res 2005;6:74.

[40] Casalino-Matsuda SM, Monzon ME, Forteza RM. Epidermalgrowth factor receptor activation by epidermal growth factormediates oxidant-induced goblet cell metaplasia in humanairway epithelium. Am J Respir Cell Mol Biol 2006;34:581–91.

[41] Koff JL, Shao MX, Ueki IF, et al. Multiple TLRs activate EGFRvia a signaling cascade to produce innate immune responsesin airway epithelium. Am J Physiol Lung Cell Mol Physiol2008;294:L1068–75.

[42] Myers TJ, Brennaman LH, Stevenson M, et al. Mitochondrialreactive oxygen species mediate GPCR-induced TACE/ADAM17-dependent transforming growth factor-α shedding.Mol Biol Cell 2009;20:5236–49.

[43] Yu H, Zhou X, Wen S, et al. Flagellin/TLR5 responses inducemucus hypersecretion by activating EGFR via an epithelialcell signaling cascades. Exp Cell Res 2012;318:723–31.

[44] Yu M, Lam J, Rada B, et al. Double-stranded RNA inducesshedding of the 34-kDa soluble TNFR1 from human airwayepithelial cells via TLR3-TRIF-RIP1-dependent signaling:roles for dual oxidase 2- and caspase-dependent pathways.J Immunol 2011;186:1180–8.

[45] Burgel PR, Nadel JA. Roles of epidermal growth factorreceptor activation in epithelial cell repair and mucinproduction in airway epithelium. Thorax 2004;59:992–6.

[46] Kang JH, Lee EH, Park SW, et al. MUC5AC expression throughbidirectional communication of Notch and epidermal growthfactor receptor pathways. J Immunol 2011;187:222–9.

[47] Goldkorn T, Balaban N, Matsukuma K, et al. EGF-receptorphosphorylation and signaling are targeted byH2O2 redox stress. Am J Respir Cell Mol Biol 1998;19:786–98.

[48] Gensch E, Gallup M, Sucher A, et al. Tobacco smoke controlof mucin production in lung cells requires oxygen radicalsAP-1 and JNK. J Biol Chem 2004;279:39085–93.

[49] Yu H, Li Q, Kolosov VP, et al. Regulation of cigarette smoke-mediated mucin expression by hypoxia-inducible factor-1αvia epidermal growth factor receptor-mediated signalingpathways. J Appl Toxicol 2012;32:282–92.

[50] Pace E, Ferraro M, Uasuf CG, et al. Cilomilast counteracts theeffects of cigarette smoke in airway epithelial cells. CellImmunol 2011;268:47–53.