Biochimica et Biophysica Acta 1822 (2012) 1581–1589

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbad is

Clostridium perfringens alpha-toxin induces the release of IL-8 through a dualpathway via TrkA in A549 cells

Masataka Oda a,⁎⁎, Ryota Shiihara a, Yuka Ohmae a, Michiko Kabura a, Teruhisa Takagishi a,Keiko Kobayashi a, Masahiro Nagahama a, Masahisa Inoue b, Tomomi Abe b, Koujun Setsu b, Jun Sakurai a,⁎a Department of Microbiology Faculty of Pharmaceutical Sciences, Tokushima Bunri University, Yamashiro-cho, Tokushima, 770-8514, Japanb Functional Morphology, Faculty of Pharmaceutical Sciences, Tokushima Bunri University, Yamashiro-cho, Tokushima, 770-8514, Japan

⁎ Corresponding author. Tel.: +81 88 622 9611; fax⁎⁎ Corresponding author. Tel.: +81 88 602 8485; fax

E-mail addresses: sakurai@ph.bunri-u.ac.jp (J. Sakur(M. Oda).

0925-4439/$ – see front matter. Published by Elsevier Bdoi:10.1016/j.bbadis.2012.06.007

a b s t r a c t

a r t i c l e i n f o

Article history:Received 14 March 2012Received in revised form 8 May 2012Accepted 12 June 2012Available online 18 June 2012

Keywords:Alpha-toxinA549 cellTrkAIL-8ERK1/2p38 MAPK

A characteristic feature of gas gangrenewith Clostridium perfringens (C. perfringens) is the absence of neutrophilswithin the infected area and the massive accumulation of neutrophils at the vascular endothelium around themargins of the necrotic region. Intravenous injection of C. perfringens alpha-toxin into mice resulted in the accu-mulation of neutrophils at the vascular endothelium in lung and liver, and release of GRO/KC, a member of theCXC chemokine family with homology to human interleukin-8 (IL-8). Alpha-toxin triggered activation of signaltransduction pathways causingmRNA expression and production of IL-8, which activatesmigration and bindingof neutrophils, in A549 cells. K252a, a tyrosine kinase A (TrkA) inhibitor, and siRNA for TrkA inhibited the toxin-induced phosphorylation of TrkA and production of IL-8. In addition, K252a inhibited the toxin-induced phos-phorylation of extracellular regulated kinase 1/2 (ERK1/2) and p38 mitogen-activated protein kinase (MAPK).PD98059, an ERK1/2 inhibitor, depressed phosphorylation of ERK1/2 and nuclear translocation of nuclear factorkappa B (NF-κB) p65, but SB203580, a p38 MAPK inhibitor, did not. On the other hand, PD98059 and SB203580suppressed the toxin-inducedproduction of IL-8. Treatment of the cellswith PD98059 resulted in inhibition of IL-8 mRNA expression induced by the toxin and that with SB203580 led to a decrease in the stabilization of IL-8 mRNA. These results suggest that alpha-toxin induces production of IL-8 through the activation of two separatepathways, the ERK1/2/NF-κB and p38 MAPK pathways.

Published by Elsevier B.V.

1. Introduction

Clostridium perfringens alpha-toxin has been shown to be a virulencefactor required for the development of gas gangrene [1,2]. Gas gangreneis reported to occur without septicemia and is a highly lethal infection.

A characteristic feature of infection with C. perfringens is an ab-sence of neutrophils within the area of necrosis associated withC. perfringens growth and the accumulation of neutrophils at the bor-ders of the region showing necrosis.We reported that exposure of rab-bit neutrophils to alpha-toxin induced firm adhesion of the cells tofibrinogen and fibronectin, suggesting that the toxin stimulates bind-ing of neutrophils to the vascular endothelium [3]. IL-8, which belongsto the CXC chemokine family, is a pivotal chemotactic cytokine re-sponsible for the recruitment of neutrophils and macrophages [4].Bryant and Stevens reported that alpha-toxin induced IL-8 synthesisin cultured human umbilical vein endothelial cells, and demonstratedin vitro that both alpha- and theta-toxin have profound effects onpolymorphonuclear leukocyte functions of adherence and chemotaxis

: +81 88 655 3051.: +81 88 655 3051.ai), masa@ph.bunri-u.ac.jp

.V.

[5]. Carveth et al. reported that IL-8/CXCL8 promoted adhesion of neu-trophils to plastic, extracellular matrix proteins [6], suggesting that IL-8 induces binding of neutrophils to vascular endothelium or tissues.Furthermore, phospholipase C (PLC) from Pseudomonas aeruginosa in-duced release of IL-8 and neutrophil infiltration into the lung, involv-ing pulmonary inflammation [7]. Therefore, it is possible that theunique behavior of neutrophils in gas gangrene caused by C. per-fringens is dependent on IL-8 released from macrophages and/or en-dothelial cells activated by the toxin. However, little is known aboutthe accumulation of neutrophils at the margins of the infected area.

We have reported that alpha-toxin induces formation of diac-ylglycerol (DG) through the activation of endogenous PLC via a pertus-sis toxin-sensitive GTP-binding protein (Gi) and phosphorylation of3-phosphoinositide-dependent kinase 1 (PDK1) via activation of theTrkA receptor, so that the production of DG and activation of PDK1lead to a synergic elevation of protein kinase C θ (PKCθ) activity, andthat the activation of PKCθ stimulates generation of superoxide throughMAPK-associated signaling events in rabbit neutrophils [3,8]. We alsoreported that anti-tumor necrosis factor (TNF)-alpha antibody inhibitedthe death of mice induced by alpha-toxin, and that TNF-alpha-deficientmice were resistant to alpha-toxin, suggesting that the lethal effect isclosely related to the toxin-induced release of TNF-alpha into the blood-stream. Therefore, this lethal disease is thought to be caused by the

1582 M. Oda et al. / Biochimica et Biophysica Acta 1822 (2012) 1581–1589

toxin which induces production and activation of an endogenousinflammatory mediator, TNF-alpha [9]. In addition, we reported thatalpha-toxin-induced release of TNF-alpha is dependent on the activa-tion of ERK1/2 signal transduction via TrkA in mouse neutrophils andmacrophages and that erythromycin, which inhibits the toxin-inducedphosphorylation of TrkA and release of inflammatory cytokines includ-ing TNF-alpha, specifically blocks the toxin-induced events includingdeath through the activation of neutrophils and macrophages [9]. Littleis known about the mechanism of IL-8 release induced by the toxin.

To understand the recruitment of neutrophils at the margins ofan infected wound, the toxin-induced production of IL-8, a potentneutrophil chemoattractant, in a human lung adenocarcinoma epi-thelial cell line (A549 cells) was investigated. We demonstratedthat alpha-toxin induced production of IL-8 through activation ofboth an ERK1/2 to NF-κB system via endogenous PLC and a p38MAPK system via TrkA.

2. Materials and methods

2.1. Cell culture and reagents

A549 cellswere cultured at 37 °C ingrowthmedium(DMEMwith10%horse serum). K252a, AG1478, PD98059, SB203580, and BAY11-7082were from Calbiochem, USA. SP600125 was from TOCRIS, USA. All otherdrugs were of analytical grade.

2.2. Purification of alpha-toxin

A recombinant form of the plasmid pHY300PLK harboring thestructural gene of wild-type alpha-toxin was introduced into Bacillussubtilis ISW1214 by transformation. The expression and purificationof the recombinant alpha-toxin toxin were performed as describedpreviously [10].

2.3. Histopathological evaluation

ICRmice (25–30 g, SLC, Japan)were used. Alpha-toxin (50 ng) or sa-line (each 100 μl) was intravenously injected in the tail vein of mice.The animals were sacrificed 3 h after the injection of the toxin. Thelung tissues were fixed in a 10% neutralized buffered formalin solution(pH 7.4). After fixation, the tissues were dehydrated through gradedethanol, and paraffin sections were prepared by the routine method.These tissue sections were stained with hematoxylin or eosin for lightmicroscopy. Neutrophils were identified morphologically.

2.4. GRO/KC ELISA

Immunoreactive GRO/KC was quantified in serum with a double-antibody ELISA kit using rGRO/KC as a standard (IBL, Japan). Thisassay had a lower limit of detection of 23 pg/ml.

2.5. IL-8 ELISA

Immunoreactive IL-8 was quantified in cell culture supernatantswith a double-antibody ELISA kit using rIL-8 as a standard (R&Dsystems, USA). The assay had a lower limit of detection of 5 pg/ml.

2.6. Western blotting

A549 cells were incubated with 1.0 μg/ml of alpha-toxin for 0, 1, 5,15, 30, or 60 min. They were then solubilized by incubation for 10 minon ice in 50 mMTris–HCl (pH 7.6), 150 mMNaCl, 5 mMEDTA, 10% glyc-erol, 1% TritonX-100, 10 mM sodium pyrophosphate, 1 mM Na3VO4,10 mMNaF, 1 mMPMSF, and leupeptin (10 μg/ml). After centrifugation(15 min, 15,000×g), samples (20 mg protein) of supernatants weresubjected to SDS-PAGE and Western blotting using anti-phospho-TrkA

(Cell Signaling Technology, USA), anti-phospho-NF-κB p65 (Cell Signal-ing Technology, USA), anti-phospho-MAPKs (Cell Signaling Technology,USA) and anti-β-actin (Santa Cruz, USA) antibodies. Detectionwas con-ducted using an enhanced chemiluminesence kit (Millipore, USA). Aquantitative analysis of bands was performed by densitometry (LAS-4000, Fujifilm, Japan).

2.7. Transfection with siRNA for TrkA

A549 cells were seeded (2.0 x 105/ml of DMEM/dish) in 35-mmculture dishes and grown overnight; 10 nM siRNA for TrkA (siTrkA),or 10 nM negative control siRNA (NC-siRNA) duplex was introducedinto cells using HiPerFect Transformation reagent (Qiagen, USA),according to the manufacturer's recomendations. A mock transfectionwithout siRNA was also performed. These siRNA were purchasedfrom Qiagen. Silencing of the TrkA was determined by Western blot-ting using anti-TrkA antibody (Cell Signaling Technology, USA).

2.8. Cy3-labeled alpha-toxin

Alpha-toxin (3.0 mg/ml) was labeled with a Cy3-labeling kit (GEhealthcare, USA) following the manufacturer's directions. The releaseof IL-8 from A549 cells treated with Cy3-labeled alpha-toxin was thesame as that of non-labeled toxin.

2.9. Alexa488-labeled anti-TrkA antibody

The anti-TrkA antibody (Cell Signaling Technology, USA) waslabeled with an Alexa488-labeling kit (Invitrogen, USA).

2.10. Immunofluorescence staining

A549 cells (1.0×105/ml) seeded on a glass-bottomed dish (MatTek,USA) were incubated with alpha-toxin (1.0 μg/ml) at 37 °C for variousperiods, fixed with 2% paraformaldehyde in PBS at room temparaturefor 15 min, and then washed three times with PBS. Cells permeabilizedwith PBS containing 0.1% TritonX-100 for 5 min were treated withblocking buffer for 30 min. They were incubated with the primary anti-body for 1 h, followed by, at room temparature for 1 h, the appropriatesecondary antibody, such as anti-rabbit antibody, conjugated withAlexa Fluor 546. The nuclei were stained with Hoechest 33342. Stainedcells were visualized using confocal microscopy (Nikon, A1R, Japan), orBIOREVO (BZ-9000, Keyence, Japan) with an associated analysis soft-ware package. The fluorescence intensity in visual fields was measuredbyDynamic Cell Count application of BIREVO (BZ-H1C, Keyence, Japan).Values were calculated as the average for 5 fields.

2.11. Detection of IL-8 mRNA

Total RNA was isolated from proliferating cells or tissue samples(3.5 mm-diameter) using a Fast pure RNA kit (Takara, Japan). Quantita-tive RT-PCR for the mRNA of IL-8 was performed using PrimeScript RT-PCR kits (Takara, Japan). The primers: 5′-TCTGCAGCTCTGTGTGAAGG-3′(sense) and 5′-TGAATTCTCAGCCCTCTTCAA-3′ (antisense) for amplifyingfragments of IL-8; and 5′-GTGGGGCGCCCCAGGCACCA-3′ (sense) and 5′-CTCCTTAATGTCACGCACGATTTC-3′ (antisense) for amplifying fragmentsof a housekeeping gene,β-actin,were purchased fromTakara. ThemRNAlevel of β-actin was used as an internal control.

2.12. Administration of anti-GRO/KC antibody

The mice were intravenously injected with alpha-toxin or a mix-ture of anti-GRO/KC (IBL, Japan) and alpha-toxin, and sacrificed 3 hlater. The lung and liver were fixed in a 10% neutralized buffered for-malin solution (pH 7.4), the tissues were dehydrated through gradedethanol, and paraffin sections were prepared by the routine method.

1583M. Oda et al. / Biochimica et Biophysica Acta 1822 (2012) 1581–1589

These tissue sections were stained with hematoxylin or eosin for lightmicroscopy.

2.13. Statistics

Resultswere expressed as themean±SEM.n equals the sample size.Statistical comparisonswere performed using anunpaired t-test or one-way analysis of variance (ANOVA) with Bonferroni correction. p Valuesless than 0.01 were considered statistically significant.

3. Results

3.1. Alpha-toxin-induced infiltration of and release of IL-8 fromneutrophils

Konig et al. reported that PLC from P. aeruginosa induced neutrophilinfiltration into the lung and release of IL-8 [7]. We investigated ifalpha-toxin causes infiltration of neutrophils into lung and liver of mice.

Fig. 1. Infiltration of neutrophils into the lung and release of GRO/KC in serum of micetreated with alpha-toxin. A) Mice were administrated with 50 ng of alpha-toxin/mouse.After 3 h, lung tissue was fixed in formalin and sections were stained with hematoxylinand eosin. The number of neutrophils in blood vessels of the lung was counted usinglight microscopy. The neutrophils of 50 vessels were counted from the images. Theaverage number of neutrophils per mm2 in the vessels was morphologically analyzed.Values represent the mean±SEM; n=3; *Pb0.01, compared with the number of neutro-phils in the lungs of mice treated with alpha-toxin. B) Mice were treated with variousconcentrations of alpha-toxin. After 3 h, GRO/KC in serumwas measured by ELISA. Valuesrepresent the mean±SEM; n=3 independent experiments.

The mice were intravenously administrated with 50 ng of alpha-toxin/mouse. The injection led to the accumulation of neutrophils on the vascu-lar endothelium in the lung (Fig. 1A). The accumulationwas also detectedin the liver (data not shown). In addition, alpha-toxin dose-dependentlyinduced the release of GRO/KC, the rodent homologue for human IL-8which activates the migration of neutrophils, in serum (Fig. 1B). On theother hand, intravenous injection of the toxin into mice preinjectedwith anti-GRO/KC antibody resulted in a significant reduction inthe accumulation of neutrophils (Fig. 1A). It therefore appears thatGRO/KC plays a role in the accumulation of neutrophils on the vas-cular endothelium.

We examined whether the toxin induces release of IL-8 from A549cells. A549 cells were incubated with alpha-toxin (0.1–1.0 μg/ml) at37 °C for 5 h. The toxin caused the release of IL-8 in a dose-dependent manner, as shown in Fig. 2A. The release of IL-8 reacheda maximum within 5 h and then later sharply decreased (Fig. 2B).The H148G variant, which does not have PLC activity, did not inducethe release of IL-8 from the cells (Fig. 2A and B), showing the enzy-matic activity of alpha-toxin to be essential for the release.

Fig. 2. Alpha-toxin-induced release of IL-8 from A549 cells A) A549 cells (1.0×107/ml).were incubated with alpha-toxin (black circles) or H148G (white circles) at 37 °C for5 h. B) The cells were incubated with 1.0 μg/ml alpha-toxin (black circles) or H148G(white circles) the toxin at 37 °C for various periods. The amount of IL-8 in culture super-natants was determined by ELISA. Values represent the mean±SEM; n=4 independentexperiments.

Fig. 4. Effect of siRNA for TrkA on release of IL-8 induced by alpha-toxin. A549 cells weretransfectedwith siTrkA or NC-siRNA (10 nM). A) The expression of TrkA in the cells treatedwith siRNA was detected by Western blotting using anti-TrkA antibody and anti-β-actinantibody. B) The intact cells (black columns), NC-siRNA-treated cells (white columns), orsiTrkA-treated cells (gray columns) were incubated with 1.0 μg/ml alpha-toxin at 37 °C

1584 M. Oda et al. / Biochimica et Biophysica Acta 1822 (2012) 1581–1589

3.2. Participation of TrkA in release of IL-8 by alpha-toxin from A549 cells

We have reported that alpha-toxin bound to TrkA and inducedphosphorylation of the receptor in various cells [8,9,11]. Phosphorylat-ed TrkA in A549 cells incubated with 1.0 μg/ml of alpha-toxin at 37 °Cfor various periods was analyzed by SDS-PAGE and Western blottingusing anti-phospho-TrkA antibody. Phosphorylated TrkAwith a molec-ular mass of about 140 kDa was detected in A549 cells (Fig. 3). Fig. 3Ashows that the phosphorylation of TrkA reached a maximum within1 min and then gradually decreased until 60 min. A549 cells werepretreated with various amounts of K252a at 37 °C for 60 min, and in-cubated with alpha-toxin (1.0 μg/ml) at 37 °C for 5 h. K252a dose-dependently inhibited the release of IL-8 induced by the toxin(Fig. 3B). Next, we used a siRNA to knockdown TrkA expression in tar-get cells. Transfection of A549 cells with siTrkA (TrkA-specific siRNA)resulted in a significant reduction in the expression of TrkA, whereasthat with NC-siRNA (a negative control siRNA) had little effect, asshown in Fig. 4A. The transfection with siTrkA drastically reduced therelease of IL-8 by alpha-toxin (Fig. 4B) and the binding of alpha-toxin(Fig. 4C), compared to that with NC-siRNA.

Furthermore, to test whether alpha-toxin binds to TrkA in A549cells, the cells were incubated with Cy3-labeled alpha-toxin andAlexa488-conjugated anti-TrkA antibody at 37 °C for 1 h. As shownin Fig. 5A, the Cy3-labeled alpha-toxin merged with the fluorescenceof the Alexa488-labeled anti-TrkA antibody on membranes of thecells. The percentage of colocalization was approximately 70%(Fig. 5B). It therefore was confirmed that alpha-toxin binds toTrkA in A549 cells.

Fig. 3. Relationship between release of IL-8 and phosphorylation of TrkA induced by alpha-toxin. A) A549 cells (1.0×107/ml) were incubated with alpha-toxin (1.0 μg/ml) at 37 °C forvarious periods. The lysates were subjected to SDS-PAGE, followed by immunoblottingwith antibody against phosphorylated TrkA. B) A549 cells (1×107 /ml) were pretreatedwith various amounts of K252a at 37 °C for 1 h, and incubated with (black columns)1.0 μg/ml alpha-toxin or without (white columns) the toxin at 37 °C for 5 hr. IL-8 in theculture supernatants was determined by ELISA. Values represent the mean±SEM; n=4independent experiments.

for 3 h. IL-8 in the culture supernatants was determined by ELISA. Values represent themean±SEM; n=5 independent experiments. C) These siRNA-treated cells were incubatedwith Cy-3-labeled alpha-toxin (1.0 μg/ml) at 37 °C for 15 min. The cells were fixed in 4%paraformaldehyde and stainedwith hoechst33342 andanalyzed usingfluorescencemicros-copy. The fluorescent intensity in the visual fields was measured as described in MaterialsandMethods. The binding of alpha-toxin to the intact cells was set as themaximal response(100%), against which all other results were compared. Values represent the mean±SEM;n=3 independent experiments.

3.3. Relationship between the toxin-induced release of IL-8 and activationof MAPK pathways

Several studies have reported that the release of IL-8 induced byvarious toxins such as LPS and Helicobacter pylori VacA is linked tothe activation of MAPK pathways [12–14]. We investigated the effectof alpha-toxin on MAPK pathways. When A549 cells were incubatedwith 1.0 μg/ml of the toxin at 37 °C, phosphorylation of ERK1/2 andp38 MAPK reached a maximum within 15 min and graduallydecreased until 60 min, but that of SAPK/JNK was not observed(Fig. 6A). Furthermore, K252a inhibited phosphorylation of ERK1/2and p38 MAPK in a dose-dependent manner (Fig. 6B and C),suggesting TrkA to be located upstream of ERK1/2 and p38 MAPK.Next, we examined the effect of PD98059, an ERK1/2 cascade inhibi-tor, and SB203580, a p38 MAPK inhibitor, on the toxin-induced re-lease of IL-8 and phosphorylation of these proteins. PD98059 andSB203580 dose-dependently inhibited the release of IL-8 (Fig. 6Dand E).

3.4. Alpha-toxin-induced translocation of NF-κB to the nucleus andinhibition of the toxin-induced release of IL-8 by BAY11-7082

Activation of the IL-8 promoter is known to require activation of atranscription factor, NF-κB [15,16]. Therefore we examined if alpha-toxin stimulates IL-8 production through an NF-κB-mediated path-way in A549 cells. NF-κB is known to be phosphorylated before itstranslocation to the nucleus. Fig. 7A shows that phosphorylationof NF-κB in the cells treated with alpha-toxin reached a maximumwithin 15 min and then decreased. BAY11-7082, an inhibitor kappa

Fig. 5. Localization of alpha-toxin and TrkA in A549 cells. A549 cells were incubated with or without Cy3-coupled alpha-toxin (1.0 μg/mL) and Alexa488-conjugated anti-TrkA an-tibody at 37 °C for 1 h. The cells were fixed in 4% paraformaldehyde and stained with hoechst 33342. Alpha-toxin, TrkA, and nuclear staining were analyzed using confocal lasermicroscopy. A white arrow indicates the colocalization of alpha-toxin and TrkA. The data represents the mean for three independent experiments. Scale bar: 10 μm B) The percent-age of colocalization was calculated for each combination of fluorescence markers by analysis of each confocal plane of >20 cells.

1585M. Oda et al. / Biochimica et Biophysica Acta 1822 (2012) 1581–1589

B-alpha (IκBα) inhibitor, inhibited the toxin-induced IL-8 secretion ina dose-dependent manner as shown in Fig. 7B. Next, we investigatedthe effect of PD98059 and SB203580 on the translocation of NF-κB tothe nucleus in the cells treated with the toxin. Fig. 7C and D showsthat treatment of A549 cells with alpha-toxin resulted in the translo-cation of NF-κB to the nucleus, as visualized by immunostaining withanti-NF-κB p65 antibody. Treatment of PD98059-pretreated cellswith the toxin inhibited the transduction, but the treatment withSB203580 had no effect (Fig. 7C and D).

3.5. The relationship between MAPK systems and production ofIL-8 induced by alpha-toxin in A549 cells

To investigate the relationship between the activation of MAPK sys-tems and expression of IL-8 mRNA induced by alpha-toxin, A549 cellswere pretreatedwith 10 μMPD98059 or SB203580, and later incubatedwith 1.0 μg/ml alpha-toxin. As shown in Fig. 8A, PD98059 almostcompletely inhibited alpha-toxin-induced expression of IL-8 mRNAunder the conditions. Treatment with SB203580 led to a decrease ofabout 40% in the expression of mRNA in the cells treated with thetoxin (Fig. 8A). It is generally thought that mRNA has a short half life.It has been reported that p38MAPKenhanced IL-1-induced IL-8 produc-tion by stabilizing IL-8 mRNA [17,18]. Shimohata et al. reported that

SB203580 decreased the stabilization of IL-8 mRNA following Vibrioparahaemolyticus infection [19]. Therefore, we examined the effect ofthe inhibitor on IL-8 mRNA in the cells treated with alpha-toxin. Thecells were pretreated with alpha-toxin for 1 h to allow transcription ofIL-8 mRNA and subsequently treated with SB203580 in the presenceof actinomycin D (5 μg/ml). As shown in Fig. 8B, the mRNA level inthe toxin-treated cells was about 50% of that in the cells treated withthe toxin in the absence of the inhibitor, and a gradual decay in the ex-pression was observed until 6 h. The level of IL-8 mRNA in the toxin-pretreated cells treated with the p38 MAPK inhibitor reached a mini-mum (about 50% of that in the absence of the inhibitor) after 2 h andremained low for 6 h. It therefore appears that IL-8 mRNA in the cellstreated with SB203580 is significantly unstable.

Next, we investigated the effect of 10 μM PD98059 or SB203580on the intracellular distribution of IL-8 using fluorescence microsco-py. The level of IL-8 in the cytosol of A549 cells treated with alpha-toxin reached a maximum within 2 h and then decreased in a time-dependent manner (Fig. 8C). When the PD98059-pretreated cellswere treated with the toxin, IL-8 was not observed in the cytosol(Fig. 8C). However, when the SB203580-pretreated cells were treatedwith the toxin for 2 h, the amount of IL-8 in the cytosol was about 30%of that in the cytosol of A549 cells treated with alpha-toxin (Fig. 8C).The result was consistent with that shown in Fig. 6C.

Fig. 6. Phosphorylation of various kinases of MAPK cascades induced by alpha-toxin in A549 cells. A) A549 cells (1.0×107/ml) were incubated with alpha-toxin (1.0 μg/ml) at 37 °C forvarious periods. The lysateswere subjected to SDS-PAGE, followed by immunoblottingwith antibodies against phosphorylated ERK1/2, phosphorylated p38MAPK, phosphorylated SAPK/JNK, and β-actin. The data represents themean for three independent experiments. B, C) A549 cells (1.0×107/ml)were preincubatedwith (white circles) 10 nMK252a orwithout (blackcircles) the blocker at 37 °C for 60 min, and then washed. The cells were incubated with 1.0 μg/ml alpha-toxin at 37 °C for indicated periods. The phospho-ERK1/2 (B) and -p38 MAPK(C) extracted by 10% trichloroacetic acid was subjected to SDS-PAGE and immunoblotting using specific antibody. The phosphorylated level of ERK1/2 and p38 MAPK in untreatedcells was set to 1. Values represent the mean±SEM; n=5 independent experiments. D, E) A549 cells (1.0×107 /ml) were preincubated with various concentrations of PD98059 orSB203580 at 37 °C for 60 min, and then washed. The treated cells were incubated with 1.0 μg/ml alpha-toxin at 37 °C for 5 h. The level of IL-8 (empty columns) and phosphorylation(black circles) of various kinases was determined as described in Materials and Methods. The phosphorylation of ERK1/2 and p38 MAPK induced by alpha-toxin was set as the maximalresponse (100%), against which all other results were compared. Values represent the mean±SEM; n=3 independent experiments.

1586 M. Oda et al. / Biochimica et Biophysica Acta 1822 (2012) 1581–1589

4. Discussion

Several studies have reported that neutrophils accumulated on thevascular endothelium at the margins of gas gangrene caused byC. perfringens [20,21]. The present study showed that alpha-toxininduced the accumulation of neutrophils on the vascular endotheliumof organs, especially lung and liver, and release of GRO/KC in mice,and that the injection of anti-GRO/KC significantly reduced the accu-mulation of neutrophils, indicating that the toxin-induced release of

GRO/KC plays a role in the accumulation. We also found that thetoxin induced the release of IL-8 in A549 cells. The CXC chemokineIL-8/CXCL8 is known to be a potent chemoattractant and induce neu-trophil chemotaxis toward sites of infection. Therefore, it appears thatthe toxin-induced release of IL-8 plays a pivotal role in the pathogen-esis of the infectious disease.

We have reported that alpha-toxin-induced activation of endoge-nous PLC and sphingomyelinase via a pertussis toxin-sensitive GTP-binding protein is closely related to the hemolysis of rabbit and

Fig. 7. Relationship between activation of NF-κB and release of IL-8 induced by alpha-toxin. A) A549 cells (1×107 /mL) were incubated with alpha-toxin (1.0 μg/mL) at 37 °C forvarious periods. The lysates were subjected to SDS-PAGE, followed by immunoblotting with antibodies against phosphorylated NF-κB, and β-actin. Signal intensities from immu-noblots were recorded densitometrically For quantification, NF-κB signals applying the phosphorylation-sensitive antibody were normalized to the NF-κB signals applying theantibody of NF-κB. The phosphorylation level of untreated cells was set to 1. Values represent the mean±SEM; n=3; *Pb0.005, compared with the phosphorylation of NF-κBin untreated cells. B) The cells were pretreated with various concentrations of BAY11-7082 at 37 °C for 1 h, and then incubated with (black columns) 1.0 μg/mL alpha-toxin or with-out (white columns) the toxin at 37 °C for 3 h. C) The cells were pretreated with 10 μM PD98059 or 10 μM SB203580 at 37 °C for 60 min, and then incubated with alpha-toxin(1.0 μg/mL) at 37 °C for 15 min. The cells were fixed by paraformaldehyde and stained with SYTO 16 and antibody against NF-κB. The nucleus and NF-κB were viewed with confocalmicroscopy. The data represents the mean for three independent experiments. Scale bar: 10 μm. D) The percentage of colocalization was calculated for each combination offluorescence markers by analysis of each confocal plane of>50 cells. Values represent the mean±SEM; n=3; *Pb0.003, compared with the colocalization of NF-κB with nucleiin untreated cells.

1587M. Oda et al. / Biochimica et Biophysica Acta 1822 (2012) 1581–1589

sheep erythrocytes, respectively, and that the toxin independentlytriggered two pathways, the formation of diacylglycerol through theactivation of endogenous PLC and the activation of PDK1 throughTrkA in rabbit neutrophils, eliciting the generation of superoxidethrough ERK1/2-associated signaling events [8]. We also reportedthat the toxin-induced release of TNF-alpha was dependent on the ac-tivation of ERK1/2 signal transduction via the phosphorylation ofTrkA in neutrophils and macrophages and that inhibition of thetoxin-induced phosphorylation of TrkA by erythromycin led to the ar-rest of the toxin-induced event [9]. Therefore, TrkA is a key signalingmolecule in the action induced by the toxin. The present study shows1) alpha-toxin specifically binds to and phosphorylates TrkA in A549cells, 2) treatment of A549 cells with K252a and transfection of thecells with siRNA for TrkA resulted in reduced release of IL-8 by

alpha-toxin, confirming that alpha-toxin induces the release of IL-8through TrkA in A549 cells.

It is known that the infection of cultured epithelial cells withSalmonella enterica serovar [22], Salmonella typhimurium [23], H. pylori[12,14,24], Enterohemorrhagic Escherichia coli [25] or Enteropatho-genic Escherichia coli [26] results in the release of IL-8 through theactivation of MAPK pathways. The present study showed that alpha-toxin induced the release of IL-8 through the activation of ERK1/2and p38 MAPK pathways via TrkA in A549 cells. It has been reportedthat neurite-outgrowth through the activation of TrkA induced bynerve growth factor is linked to activation of the MAPK signalingsystem [27–31]. Our previous study concerning signal transductionmechanisms for alpha-toxin-induced superoxide generation hasshown that activation of TrkA, PDK1, and PKCθ occurs upstream of

1588 M. Oda et al. / Biochimica et Biophysica Acta 1822 (2012) 1581–1589

that of the MEK-ERK1/2 pathway [8]. K252a inhibited the toxin-induced release of IL-8 and the phosphorylation of ERK1/2 and p38MAPK in A549 cells, indicating that the phosphorylation of ERK1/2and p38 MAPK occurs downstream of the activation of TrkA. The pro-cess of IL-8 release induced by the toxin resembles that of the previ-ous results. In the present study, we obtained evidence that alpha-toxin induced the release of IL-8 through the activation of ERK1/2and p38 MAPK pathways via TrkA in A549 cells.

Alpha-toxin caused the phosphorylation of NF-κB. The toxin-induced release of IL-8 was inhibited by BAY11-7082, an IκBα inhib-itor. It therefore appears that IL-8 production is linked to phosphory-lation of NF-κB. PD98059 and SB203580 inhibited the toxin-inducedrelease of IL-8. However, PD98059 inhibited the toxin-induced phos-phorylation of NF-κB and translocation of NF-κB to the nucleus, butSB203580 did not. In addition, PD98059 inhibited the expression ofIL-8 mRNA in the cells treated with alpha-toxin, but SB203580 didnot. These observations suggest that IL-8 mRNA expression inducedby the toxin is related to translocation of phosphorylated NF-κB tothe nucleus through activation of ERK1/2 via TrkA, but not to activa-tion of p38 MAPK. Next, fluorescence microscopic analysis revealedthat IL-8 was not present in the cytoplasm of PD98059-pretreatedcells treated with the toxin, and that the amount of IL-8 in the cyto-plasm of SB203580-pretreated cells treated with the toxin wasapproximately 30% of that in control cells treated with alpha-toxin.Furthermore, as shown in Fig. 8B, IL-8 mRNA induced by the toxinwas labile in SB203580-treated cells. Several studies have reportedthat p38 MAPK plays an important role in the stabilization of IL-8mRNA [17–19,32]. Therefore, it appears that p38 MAPK contributedto the stabilization of IL-8 mRNA in the cells treated with the toxin.In addition, several workers have reported that SAPK/JNK contributesto expression of IL-8 [33–36]. However, treatment of A549 cells withthe toxin resulted in little phosphorylation of SAPK/JNK, suggestingthat expression of IL-8 in the cells treated with the toxin is indepen-dent of the activation of SAPK/JNK. Goyal et al. reported that expres-sion of activated JNK was not seen in INT-407 cells treated withgalactose- specific adhesion, which induced secretion of IL-8 [37].Furthermore, Shimohata et al. reported that the ERK1/2 pathwayactivates the transcription of IL-8, and that the p38 MAPK pathwaystabilizes IL-8 mRNA, following V. parahaemolyticus infection [19].The ERK1/2 and p38 MAPK pathways are known to be activated byother bacterial infections, such as Yersinia enterocolitica [38]. It is in-teresting that alpha-toxin-triggered IL-8 production is dependent onexpression of IL-8 mRNA through the ERK1/2 pathway and stabiliza-tion of IL-8 mRNA through the p38 MAPK pathway in A549 cells,showing that the toxin activates the two pathways at the same time.

Fig. 8. Effect of PD98059 and SB203580 on expression of IL-8 mRNA induced by alpha-toxin in A549cells. A) A549 cells (1.0×107 /ml) were pretreated with 10 μM PD98059 or10 μM SB203580 at 37 °C for 1 h, and then incubated with alpha-toxin (1.0 μg/ml) at37 °C for 2 h. IL-8 mRNA expression was determined by RT-PCR. IL-8 mRNA was normal-ized with mRNA levels of the housekeeping gene β-actin. Signal densities from RT-PCRwere recorded densitometrically. The IL-8 mRNA level of untreated cells was set to 1.Values represent themean±SEM; n=6; *Pb0.01, **Pb0.005, comparedwith the expres-sion of IL-8 mRNA in control cells treated with alpha-toxin. B) A549 cells (1.0×107 /ml)were treated with alpha-toxin for 2 h to allow transcription of mRNA. Cells were thentreated with actinomycin D (5 mg/ml), and DMSO (white circles) or SB203580 (black cir-cles). At indicated times, total RNA was isolated, reverse transcribed and analyzed by RT-PCR. Valueswere calculated as the fold-increase in total IL-8mRNA, compared to the totalamount of IL-8 mRNA before the addition of actinomycin D. Values represent the mean±SEM; n=5; *Pb0.01, compared with the expression of IL-8 mRNA in control cells treatedwith alpha-toxin. C) A549 cells (1.0×107/ml) were pretreated with 10 μM SB203580 or10 μM PD98059 at 37 °C for 1 h, and then incubated with alpha-toxin (1.0 μg/ml) at37 °C for various periods. The cells were fixed by paraformaldehyde and stained with an-tibody against IL-8. IL-8 was viewedwith fluorescencemicroscopy. The fluorescent inten-sity in the visual fields was measured as described in Materials and Methods. Symbols:control, white squares; alpha-toxin, white circles; PD98059 plus alpha-toxin, blacksquares; SB203580 plus alpha-toxin, black circles. Values represent the mean±SEM;n=3 independent experiments.

In conclusion, alpha-toxin triggered signal transduction pathwayscausing mRNA expression and production of IL-8, which activated mi-gration and binding of neutrophils, in A549 cells. The toxin-inducedrelease of IL-8 required at least two distinct protein kinase cascades.The toxin activates ERK1/2 and p38 MAPK cascades via TrkA, and in-duces expression and stabilization of IL-8 mRNA, respectively.

Acknowledgements

The authors thank Ayaka Suzue and Kaori Tominaga for technicalassistance. This research was supported in part by a Grant-in-aid forScientific Research from the Ministry of Education, Culture, Sports,

1589M. Oda et al. / Biochimica et Biophysica Acta 1822 (2012) 1581–1589

Science and Technology of the Japanese Government, and SENRYAKUprogram (2009–2010) of MEXT.

References

[1] D.L. Stevens, The pathogenesis of clostridial myonecrosis, Int. J. Med. Microbiol.290 (2000) 497–502.

[2] R.W. Titball, Bacterial phospholipases, Trends Microbiol. 5 (1997) 265.[3] S. Ochi, T. Miyawaki, H. Matsuda, M. Oda, M. Nagahama, J. Sakurai, Clostridium

perfringens alpha-toxin induces rabbit neutrophil adhesion, Microbiology 148(2002) 237–245.

[4] N. Mukaida, Pathophysiological roles of interleukin-8/CXCL8 in pulmonarydiseases, Am. J. Physiol. Lung Cell. Mol. Physiol. 284 (2003) L566–L577.

[5] A.E. Bryant, D.L. Stevens, Phospholipase C and perfringolysin O from Clostridiumperfringens upregulate endothelial cell-leukocyte adherence molecule 1 and inter-cellular leukocyte adherence molecule 1 expression and induce interleukin-8 syn-thesis in cultured human umbilical vein endothelial cells, Infect. Immun. 64(1996) 358–362.

[6] H.J. Carveth, J.F. Bohnsack, T.M. McIntyre, M. Baggiolini, S.M. Prescott, G.A.Zimmerman, Neutrophil activating factor (NAF) induces polymorphonuclearleukocyte adherence to endothelial cells and to subendothelial matrix proteins,Biochem. Biophys. Res. Commun. 162 (1989) 387–393.

[7] B. Konig, M.L. Vasil, W. Konig, Role of haemolytic and non-haemolytic phospholi-pase C from Pseudomonas aeruginosa in interleukin-8 release from human mono-cytes, J. Med. Microbiol. 46 (1997) 471–478.

[8] M. Oda, S. Ikari, T. Matsuno, Y. Morimune, M. Nagahama, J. Sakurai, Signal transduc-tionmechanism involved in Clostridium perfringens alpha-toxin-induced superoxideanion generation in rabbit neutrophils, Infect. Immun. 74 (2006) 2876–2886.

[9] M. Oda, A. Kihara, H. Yoshioka, Y. Saito, N. Watanabe, K. Uoo, M. Higashihara, M.Nagahama, N. Koide, T. Yokochi, J. Sakurai, Effect of erythromycin on biologicalactivities induced by Clostridium perfringens alpha-toxin, J. Pharmacol. Exp.Ther. 327 (2008) 934–940.

[10] M. Nagahama, Y. Okagawa, T. Nakayama, E. Nishioka, J. Sakurai, Site-directed mu-tagenesis of histidine residues in Clostridium perfringens alpha-toxin, J. Bacteriol.177 (1995) 1179–1185.

[11] M. Oda, Y. Saito, Y. Morimune, M. Nagahama, J. Sakurai, Induction of neurite-outgrowth in PC12 cells by alpha-toxin from Clostridium perfringens, Biochem.Biophys. Res. Commun. 411 (2011) 241–246.

[12] J. Hisatsune, M. Nakayama, H. Isomoto, H. Kurazono, N. Mukaida, A.K.Mukhopadhyay, T. Azuma, Y. Yamaoka, J. Sap, E. Yamasaki, K. Yahiro, J. Moss, T.Hirayama, Molecular characterization of Helicobacter pylori VacA induction ofIL-8 in U937 cells reveals a prominent role for p38MAPK in activating transcrip-tion factor-2, cAMP response element binding protein, and NF-kappaB activation,J. Immunol. 180 (2008) 5017–5027.

[13] H. Yang, D.W. Young, F. Gusovsky, J.C. Chow, Cellular events mediated bylipopolysaccharide-stimulated toll-like receptor 4. MD-2 is required for activa-tion of mitogen-activated protein kinases and Elk-1, J. Biol. Chem. 275 (2000)20861–20866.

[14] A. Bhattacharyya, S. Pathak, S. Datta, S. Chattopadhyay, J. Basu, M. Kundu, Mitogen-activated protein kinases and nuclear factor-kappaB regulate Helicobacterpylori-mediated interleukin-8 release from macrophages, Biochem. J. 368(2002) 121–129.

[15] R. Medzhitov, P. Preston-Hurlburt, C.A. Janeway Jr., A human homologue of theDrosophila Toll protein signals activation of adaptive immunity, Nature 388(1997) 394–397.

[16] A.T. Gewirtz, A.S. Rao, P.O. Simon Jr., D. Merlin, D. Carnes, J.L. Madara, A.S. Neish,Salmonella typhimurium induces epithelial IL-8 expression via Ca(2+)-mediatedactivation of the NF-kappaB pathway, J. Clin. Invest. 105 (2000) 79–92.

[17] H. Holtmann, R. Winzen, P. Holland, S. Eickemeier, E. Hoffmann, D. Wallach, N.L.Malinin, J.A. Cooper, K. Resch, M. Kracht, Induction of interleukin-8 synthesis in-tegrates effects on transcription and mRNA degradation from at least three differ-ent cytokine- or stress-activated signal transduction pathways, Mol. Cell. Biol. 19(1999) 6742–6753.

[18] R. Winzen, M. Kracht, B. Ritter, A. Wilhelm, C.Y. Chen, A.B. Shyu, M. Muller, M.Gaestel, K. Resch, H. Holtmann, The p38 MAP kinase pathway signals forcytokine-induced mRNA stabilization via MAP kinase-activated protein kinase 2and an AU-rich region-targeted mechanism, EMBO J. 18 (1999) 4969–4980.

[19] T. Shimohata, M. Nakano, X. Lian, T. Shigeyama, H. Iba, A. Hamamoto, M. Yoshida,N. Harada, H. Yamamoto, M. Yamato, K. Mawatari, T. Tamaki, Y. Nakaya, A.

Takahashi, Vibrio parahaemolyticus infection induces modulation of IL-8 secretionthrough dual pathway via VP1680 in Caco-2 cells, J. Infect. Dis. 203 (2011)537–544.

[20] A.E. Bryant, C.R. Bayer, M.J. Aldape, R.J. Wallace, R.W. Titball, D.L. Stevens, Clostridiumperfringens phospholipase C-induced platelet/leukocyte interactions impedeneutrophil diapedesis, J. Med. Microbiol. 55 (2006) 495–504.

[21] A.E. Bryant, R.Y. Chen, Y. Nagata, Y. Wang, C.H. Lee, S. Finegold, P.H. Guth, D.L.Stevens, Clostridial gas gangrene. I. Cellular and molecular mechanisms of micro-vascular dysfunction induced by exotoxins of Clostridium perfringens, J. Infect. Dis.182 (2000) 799–807.

[22] T.L. Mynott, B. Crossett, S.R. Prathalingam, Proteolytic inhibition of Salmonellaenterica serovar typhimurium-induced activation of the mitogen-activated pro-tein kinases ERK and JNK in cultured human intestinal cells, Infect. Immun. 70(2002) 86–95.

[23] S. Hobbie, L.M. Chen, R.J. Davis, J.E. Galan, Involvement of mitogen-activated pro-tein kinase pathways in the nuclear responses and cytokine production inducedby Salmonella typhimurium in cultured intestinal epithelial cells, J. Immunol.159 (1997) 5550–5559.

[24] B. Ritter, P. Kilian, M.R. Reboll, K. Resch, J.K. DiStefano, R. Frank, W. Beil, M.Nourbakhsh, Differential effects of multiplicity of infection on Helicobacterpylori-induced signaling pathways and interleukin-8 gene transcription, J. Clin.Immunol. 31 (2011) 60–68.

[25] S. Dahan, V. Busuttil, V. Imbert, J.F. Peyron, P. Rampal, D. Czerucka,Enterohemorrhagic Escherichia coli infection induces interleukin-8 productionvia activation of mitogen-activated protein kinases and the transcription factorsNF-kappaB and AP-1 in T84 cells, Infect. Immun. 70 (2002) 2304–2310.

[26] N. Semiramoth, A. Gleizes, I. Turbica, C. Sandre, R. Gorges, I. Kansau, A. Servin, S.Chollet-Martin, Escherichia coli type 1 pili trigger late IL-8 production byneutrophil-like differentiated PLB-985 cells through a Src family kinase- andMAPK-dependent mechanism, J. Leukoc. Biol. 85 (2009) 310–321.

[27] T. Mutoh, A. Tokuda, G. Guroff, N. Fujiki, The effect of the B subunit of choleratoxin on the action of nerve growth factor on PC12 cells, J. Neurochem. 60(1993) 1540–1547.

[28] Y. Yamazaki, Y. Horibata, Y. Nagatsuka, Y. Hirabayashi, T. Hashikawa,Fucoganglioside alpha-fucosyl(alpha-galactosyl)-GM1: a novel member of lipidmembrane microdomain components involved in PC12 cell neuritogenesis,Biochem. J. 407 (2007) 31–40.

[29] H. Watanabe, T. Yokozeki, M. Yamazaki, H. Miyazaki, T. Sasaki, T. Maehama, K.Itoh, M.A. Frohman, Y. Kanaho, Essential role for phospholipase D2 activationdownstream of ERK MAP kinase in nerve growth factor-stimulated neurite out-growth from PC12 cells, J. Biol. Chem. 279 (2004) 37870–37877.

[30] S. Fujioka, K. Masuda, M. Toguchi, Y. Ohoka, T. Sakai, T. Furuyama, S. Inagaki, Neu-rotrophic effect of Semaphorin 4D in PC12 cells, Biochem. Biophys. Res. Commun.301 (2003) 304–310.

[31] S. Kao, R.K. Jaiswal, W. Kolch, G.E. Landreth, Identification of the mechanismsregulating the differential activation of the mapk cascade by epidermal growthfactor and nerve growth factor in PC12 cells, J. Biol. Chem. 276 (2001)18169–18177.

[32] E. Hoffmann, O. Dittrich-Breiholz, H. Holtmann, M. Kracht, Multiple control ofinterleukin-8 gene expression, J. Leukoc. Biol. 72 (2002) 847–855.

[33] S.S. Kang, S.S. Woo, J. Im, J.S. Yang, C.H. Yun, H.R. Ju, C.G. Son, E.Y. Moon, S.H. Han,Human placenta promotes IL-8 expression through activation of JNK/SAPK andtranscription factors NF-kappaB and AP-1 in PMA-differentiated THP-1 cells, Int.Immunopharmacol. 7 (2007) 1488–1495.

[34] I.S. Cheon, S.S. Woo, S.S. Kang, J. Im, C.H. Yun, D.K. Chung, D.K. Park, S.H. Han,Peptidoglycan-mediated IL-8 expression in human alveolar type II epithelial cells re-quires lipid raft formation and MAPK activation, Mol. Immunol. 45 (2008)1665–1673.

[35] J. Zhong, J.M. Kyriakis, Dissection of a signaling pathway by which pathogen-associated molecular patterns recruit the JNK and p38 MAPKs and trigger cyto-kine release, J. Biol. Chem. 282 (2007) 24246–24254.

[36] Y.M. Kim, W. Reed, W. Wu, P.A. Bromberg, L.M. Graves, J.M. Samet, Zn2+−inducedIL-8 expression involves AP-1, JNK, and ERK activities in human airway epithelialcells, Am. J. Physiol. Lung Cell. Mol. Physiol. 290 (2006) L1028–L1035.

[37] A. Goyal, M. Konar, A. Setia, A. Narang, S. Ghosh, Galactose specific adhesin ofenteroaggregative E. coli induces IL-8 secretion via activation of MAPK andSTAT-3 in INT-407 cells, Biochim. Biophys. Acta 1800 (2010) 574–579.

[38] G.A. Grassl, M. Kracht, A. Wiedemann, E. Hoffmann, M. Aepfelbacher, C. vonEichel-Streiber, E. Bohn, I.B. Autenrieth, Activation of NF-kappaB and IL-8 byYersinia enterocolitica invasin protein is conferred by engagement of Rac1 andMAP kinase cascades, Cell. Microbiol. 5 (2003) 957–971.