Cytokine 66 (2014) 46–53

Contents lists available at ScienceDirect

Cytokine

journal homepage: www.journals .e lsevier .com/cytokine

The cytolethal distending toxin of Aggregatibacter actinomycetemcomi-tans inhibits macrophage phagocytosis and subverts cytokine production

1043-4666/$ - see front matter � 2014 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.cyto.2013.12.014

⇑ Corresponding author. Address: Department of Microbiology, Institute ofBiomedical Sciences, University of São Paulo, Ed. Biomédicas II, Av. Lineu Prestes,1374 São Paulo, SP 05508-900, Brazil. Tel.: +55 11 3091 7348; fax: +55 11 30917354.

E-mail address: esayuria@usp.br (E.S. Ando-Suguimoto).

Ellen Sayuri Ando-Suguimoto a,⇑, Maike Paulino da Silva a, Dione Kawamoto a, Casey Chen b,Joseph M. DiRienzo c, Marcia Pinto Alves Mayer a

a Department of Microbiology, Institute of Biomedical Sciences, University of São Paulo, São Paulo, Brazilb Division of Periodontology, Diagnostic Sciences and Dental Hygiene, Ostrow School of Dentistry of University of Southern California, USAc Department of Microbiology, School of Dental Medicine, University of Pennsylvania, PA, USA

a r t i c l e i n f o

Article history:Received 27 March 2013Received in revised form 19 November 2013Accepted 24 December 2013

Keywords:Aggregatibacter actinomycetemcomitansCytolethal distending toxinPhagocytosisMacrophage

a b s t r a c t

Aggregatibacter actinomycetemcomitans is an important periodontal pathogen that can participate inperiodontitis and other non-oral infections. The cytolethal distending toxin (Cdt) is among thevirulence factors produced by this bacterium. The Cdt is also secreted by several mucosa-associatedGram-negative pathogens and may play a role in perpetuating the infection by modulating theimmune response. Although the toxin targets a wide range of eukaryotic cell types little is knownabout its activity on macrophages which play a key part in alerting the rest of the immune systemto the presence of pathogens and their virulence factors. In view of this, we tested the hypothesisthat the A. actinomycetemcomitans Cdt (AaCdt) disrupts macrophage function by inhibiting phagocyticactivity as well as affecting the production of cytokines. Murine macrophages were co-cultured witheither wild-type A. actinomycetemcomitans or a Cdt� mutant. Viable counts and qPCR showed thatphagocytosis of the wild-type strain was significantly reduced relative to that of the Cdt� mutant.Addition of recombinant Aa(r)Cdt to co-cultures along with the Cdt� mutant diminished thephagocytic activity similar to that observed with the wild type strain. High concentrations ofAa(r)Cdt resulted in decreased phagocytosis of fluorescent bioparticles. Nitric oxide production wasmodulated by the presence of Cdt and the levels of IL-1b, IL-12 and IL-10 were increased. Productionof TNF-a did not differ in the co-culture assays but was increased by the presence of Aa(r)Cdt. Thesedata suggest that the Cdt may modulate macrophage function in A. actinomycetemcomitans infectedsites by impairing phagocytosis and modifying the pro-inflammatory/anti-inflammatory cytokinebalance.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The periodontal pathogen Aggregatibacter actinomycetemcomi-tans is associated with aggressive periodontitis. This bacterium, orits products contributes to destruction of the tooth-supportingalveolar bone by triggering a host immune response [1]. Further-more, there is evidence suggesting that the oral cavity is a microbialreservoir for various systemic infections. A. actinomycetemcomitanscan cross the mucosal epithelial barrier, accessing the bloodstream,and has also been detected in atheroma plaques, bacteremia, peri-carditis, septicemia, pneumonia and infectious arthritis [2]. It is

the most prevalent member of the HACEK group, comprised ofGram-negative facultative bacteria associated with endocarditis[3].

A. actinomycetemcomitans expresses a variety of potential viru-lence factors, including the cytolethal distending toxin (Cdt), whichis composed by three subunits: CdtA, CdtB and CdtC [4]. The enzy-matically active CdtB subunit, shares structural and functionalhomology with mammalian deoxyribonucleaseI (DNase I) [5].The CdtA and CdtC subunits are required for binding the holotoxinto the plasma membrane of target cells. CdtB-CdtC dimers entersthe cells [6] and CdtB then, translocates to the nucleus where it in-duces DNA lesions [7]. This DNA damage leads to cell cycle arrest atthe G2/M interphase of target cells [6,8–10]. In addition, CdtB con-tains a phosphatidylinositol-3,4,5-triphosphate phosphatase activ-ity, leading to cell cycle arrest and apoptosis in some types of cells[11]. Cdt is produced by a variety of Gram-negative mucosa-associated pathogens [12], and data suggest that this toxin mayaffect interactions between the bacterium and the host immune

E.S. Ando-Suguimoto et al. / Cytokine 66 (2014) 46–53 47

system in chronic diseases. The Campylobacter jejuni CdtB null mu-tant leads to a less persistent colonization in mice, and results in anattenuated inflammatory effect when compared to the wild type[13]. A recent study demonstrated that the exposure to sublethaldoses of Cdt lead to phenotypic propriety of malignancy of epithe-lial cells and fibroblasts, without alterations in cell cycle distribu-tion or decrease of cell viability [14]. Sublethal doses of Cdt wereshown to induce the production of IL-1b, IL-6 and IL-8 by humanmonocytes [15]. Furthermore, a recombinant Aa(r)Cdt up-regulated the expression of receptor activator of NFjB ligand(RANKL) in gingival/periodontal ligament and Jurkat T cells [16–18].

Phagocytic cells of the innate immune system, such as macro-phages, act to eliminate microbial pathogens from the bloodstreamand tissues. Once recognized, a microbe is engulfed by the phago-cyte, triggering the secretion of pro-inflammatory cytokines andchemokines, recruiting additional immune cells and activatingthe adaptive arm of the immune system. Macrophages activatedby microorganisms and their components can be differentiatedas M1 (classic activation), producing pro-inflammatory cytokines,such as IL-1, TNFa and IL-12, driving the response to lymphocytesT helper 1 (Th1) type, while the alternative activated macrophages(M2) induce the production of regulatory cytokine such as IL-10and lead to a Th2 response [19]. Evidence that phagocytic cellscould be targeted by Cdt came from studies showing that purifiedtoxin from Haemophilus ducreyi and C. jejuni induced apoptosis innon-proliferating dendritic cells (DCs) and macrophages [20,21].Moreover, recombinant Escherichia coli supernatant containingAa(r)Cdt inhibited the production of NO by murine macrophages[22] and proliferating and non-proliferating U937 macrophageswere distinctly affected by AaCdt [23]. However, the specific activ-ities of Cdt for these cell types were not fully understood. Thus, thisstudy aimed to test the hypothesis that Cdt would affect macro-phage activation, leading to changes in phagocytosis and produc-tion of NO and cytokines, consequently altering the hostresponse in A. actinomycetemcomitans infected cells.

2. Materials and methods

2.1. Cell culture

Raw 264.7 murine macrophages were grown in Dulbecco’smodified Eagle’s medium [DMEM (Gibco, Grand Island, NY)] sup-plemented with 10% fetal calf serum (FCS)and sodium bicarbonate(2.2 g/ml). Penicillin (1664 U/ml) and streptomycin (745 U/ml)were added to cultures and incubated at 37 �C with 5% CO2 in afully humidified atmosphere.

2.2. Bacteria strains and growth conditions

A. actinomycetemcomitans D7S-SA is a spontaneously occurringnon-fimbriated (smooth) derivative of the serotype a clinical isolateD7S1 [24,25]. Strain D7S-SA CHE001 is a Cdt isogenic mutant of D7S-SA obtained by replacing the polycistronic operon of cdtABC with thespectinomycin cassette aad9 [26]. Stock cultures were kept in 20%glycerol in a �80 �C freezer. Bacteria were cultured in tryptic soyagar or broth containing 0.6% w/v yeast extract, at 37 �C in 10%CO2. Bacteria were harvested from broth cultures were at the mid-logarithmic phase of growth. E. coli BL21 (DE3) (pET15bcdtA),E. coli BL21 (DE3) (pET15bcdtB) and E. coli BL21 (DE3) (pET15bcdtC)[27] were subcultured in Luria–Bertani broth or agar with the addi-tion of 100 lg/ml ampicillin (Sigma–Aldrich, St. Louis, MO).

2.3. Recombinant CdtA, CdtB and CdtC purification

His-tagged Aa(r)Cdt subunits were obtained from the E. colistrains by affinity chromatography on NI-NDA columns (Life

Technologies, Carlsbad, CA). Active Aa(r)Cdt was prepared by com-bining CdtA-His6, CdtB-His6 and CtdC-His6 in a 1:1:1 molar ratio asdescribed previously [27,28]. After shaking at 0 �C for 1 h, non-reconstituted subunits were removed using an Amicon centrifugalfilter unit with a molecular weight cutoff of 60 kDa (Millipore, Bil-erica, MA). Protein concentration was estimated using the Bradfordassay (Bio-Rad Laboratories, Hercules, CA).

2.4. Co-cultures of cells and bacteria

Raw 264.7 cells (2 � 105 cells/well) were grown with either A.actinomycetemcomitans strains D7S-SA or D7S-SA CHE001 at a mul-tiplicity of infection (MOI) 1:10 and 1:100 (macrophage:bacteria)in DMEM medium completed with 10% FCS and without antibiot-ics, in 24 wells plates, for 2 and 20 h. The same assay was per-formed with the addition of 30 lg/ml of Aa(r)Cdt to the wellscontaining A. actinomycetemcomitans D7S-SA CHE001 at a MOI of1:100. Bacteria that were not taken up by phagocytes were killedby incubation with 50 lg of gentamicin/well for 1 h (Sigma–Aldrich). Control wells contained only macrophages or bacteria.Experiments were performed three times with samples in tripli-cate wells.

2.5. Quantification of intracellular A. actinomycetemcomitans

Internalized bacteria were recovered after lysing cells with 0.1%Triton X-100 and quantified by counting colony forming units(CFU/ml). Intracellular A. actinomycetemcomitans levels were esti-mated by comparing the number of CFUs to an estimate of bacte-rial biomass using real-time PCR. Total DNA was obtained from thelysed cells using Wizard genomic DNA purification kit (Promega,Madison, WI). The amplification reaction was performed in a20 lL final volume containing 1 lL of template DNA, 0.5 lL eachA. actinomycetemcomitans specie-specific 16SrRNA primers(25 pmol/ll) [forward: 50GGCACGTAGGCGGACCTT30 and reverse:50ACCAGGGCTAAAGCCCAATC30] (Invitrogen, São Paulo, Brazil)and 10 lL of SYBR green (Biotools, Scriptools-Quantimix SYG kit,Madri, Spain). Reactions were performed using an initial denatur-ation step at 95 �C for 10 min followed by 40 cycles at 95 �C/15 s,65 �C/1 min, 86 �C/10 s, and then followed by two steps at 95 �C/15 s and 65 �C/1 min and a final step at 0.5–95 �C (ramp) for 15 sin a Step One Plus Real-Time PCR System (Applied Biosystem, Fos-ter City, CA). Quantification of A. actinomycetemcomitans was deter-mined by comparison with a standard curve obtained with16SrRNA of A. actinomycetemcomitans cloned in pCR2.1-TOPO vec-tor (Invitrogen, Carlsbad, CA). The number of A. actinomycetemcom-itans cells was calculated assuming that six copies of the 16S rRNAgene are present in the A. actinomycetemcomitans genome (A.actinomycetemcomitans HK1651 genome project) [25,29]. The val-ues obtained from each well where subtracted from those of wellswithout the addition of eukaryotic cells (bacteria only).

2.6. Cell viability

Integrity of the Raw 264.7 macrophages plasma membrane wasassessed by evaluating the levels of extracellular lactate dehydro-genase (LDH) using a CytoTox 96� Non-Radioactive assay (Prome-ga) according to the manufacturer’s instruction. Cytotoxicity wascalculated as [(experimental LDH�spontaneously released LDH)/(spontaneously released LDH)] � 100.

To determine the effect of Aa(r)Cdt on Raw 264.7 cell viability,1 � 105 cells/well were added to 96 well plates. Sublethal concen-trations of Aa(r)Cdt (6, 12 and 25 lg/ml) were added to each welland the plates were incubated for 48 h. Cell viability was deter-mined by MTT [3(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoli-um bromide] reaction (Sigma–Aldrich). To analyze the influence of

48 E.S. Ando-Suguimoto et al. / Cytokine 66 (2014) 46–53

Aa(r)Cdt on Raw 264.7 cell could lead to cell apoptosis, Annexin Vassay was done as described by manufacturer (FITC Annexin Vapoptosis detection kit, BD Pharmigen, USA). Briefly, after this per-iod of interaction, the wells were washed twice with cold 1x PBS,then the cells were trypsinized and 100 lL of 1� binding bufferwas added. The cells were transferred to an 1.5 ml tube then5 lL of annexin V FITC conjugated and 5 lL propidium iodide wereadded, and there was an incubation of 15 min at room temperaturein the dark. Two hundred microliters were added in 96 well platesand the cells were analyzed by Guava Easy cyteFlow Cytometer(Millipore, Billerica, MA, USA). Ten thousand events were analyzedin FL-1 (FITC) and FL-2 (PI) channels. As a positive control campto-thecin (Sigma Aldrich, Saint Louis, CA), 4 h before reading, wasadded in Raw 264.7 cells in a final concentration of 6 lM.

2.7. Phagocytosis assay

Phagocytic function of Raw 264.7 cells, after Aa(r)Cdt inocula-tion was evaluated using Vybrant phagocytosis assay kit (Invitro-gen). Briefly, 1 � 105 Raw 264.7 cells/well were cultivated in a96-well plate with the addition of Aa(r)Cdt(6, 12 and 25 lg/ml).After incubation for 48 h, fluorescein-labeled E. coli K-12 cells wereadded and plates were incubated for an additional 2 h. After wash-ings, to remove suspended bacteria, cells were stained with trypanblue and the fluorescence determined in a plate reader at wave-lengths of 480 (excitation) and 520 (emission) nm (SpectraMaxGemini XS – Molecular Devices, Sunnyvale, CA). Fluorescein-la-beled E. coli cells alone served as the positive control. Results wereexpressed as a percentage using the positive control as 100%.

2.8. Measurement of NO production and cytokine production

Nitrite concentration in the cell medium was determined afterco-cultures and Aa(r)Cdt assays, as described previously [30].Briefly, 50 lL of cell supernatant was incubated with an equal vol-ume of Griess reagent (1% sulfanilamide, 0.1% naphthyl ethylenedi-amine dihydrochloride, and 2.5% H3PO4) at room temperature for10 min. The absorbance was determined at 540 nm using a micro-plate reader (Bio-Rad), and nitrite concentration determined usinga sodium nitrite standard curve.

Levels of IL-1b, IL-12, TNF-a and IL-10 in cell supernatants fromco-culture and Aa(r)Cdt assays were determined by enzyme-linkedimmunosorbent assay (ELISA) using commercial kits (Peprotech,Rocky Hill, NJ). The peroxidase reaction was measured with o-phenylenediamine (OPD). Absorbance values, measured at400 nm, were read against a standard curve and data expressedin ng/ml.

Transcription profiles of il-1b, il-12p40, tnf-a, il-10 and inosgenes were analyzed after treatment of Raw 264.7 cells withAa(r)Cdt (25 lg/ml), as described above. Total RNA was obtainedusing a RNA extraction kit (Qiagen, Hilden, Germany) accordingto the manufacturer’s instructions. RNA quantity and integritywere determined with a spectrophotometer (Nanodrop ND-1000,Wilmington, DE). Contaminating genomic DNA was removed byDNase digestion. First strand synthesis was performed with 1 lgof RNA using RT2 First Strand Kit (Qiagen). Gene expression wasdetermined by real time PCR measuring markers in the nitric oxidesignaling pathway and common cytokines arrays (Qiagen). Exper-imental and control samples were analyzed in three independentexperiments. Expression profiles of the target genes were mea-sured relative to the mean cycle threshold (CT) values of five differ-ent calibrator genes (gusb, hprt, hsp90ab1, gapdh and actb) usingthe DDCT method.

2.9. Statistical analysis

Differences in cell viability, and NO and cytokines productionamong the groups were determined using ANOVA followed by Tu-key. For phagocytosis, the paired t-test was used. Values were con-sidered significant when p < 0.05. For gene transcription profiles heStudent’s t-test was used to assess differences between control andexperimental groups using mean CT values derived from the trip-licate samples. Differences in gene expression were considered tobe significant when p < 0.05 (at greater than or equal to a 4-foldchange in expression). For all analysis Graphpad Prism version4.0 was used (La Jolla, CA, USA).

3. Results

3.1. Cdt affects phagocytosis by Raw 264.7 macrophages

The percentage of viable internalized A. actinomycetemcomitansD7S-SA was significantly lower that of its isogenic Cdt-deficientstrain D7S-SA CHE001 after 2 h of co-culture with Raw 246.7 cells(Fig. 1A). These data were confirmed by qPCR (Fig. 1B).

No viable intracellular bacteria were observed after 20 h of co-culture with the wild-type and mutant A. actinomycetemcomitansstrains. However, a higher percentage of intracellular non-viablebacterial of the Cdt-deficient mutant cells were detected comparedto the wild-type (Fig. 1C).

Addition of Aa(r)Cdt decreased the phagocytic activity of mac-rophages co-cultured with the Cdt-deficient mutant for 2 h. Thenumber of intracellular bacteria was similar to that obtained withthe wild-type strain D7S-SA (Fig. 1A).

The effect of Aa(r)Cdt on the decreased phagocytic activity ofmacrophages was confirmed by using fluorescein-labeled E. coliand Aa(r)Cdt. Low concentrations of Aa(r)Cdt were shown to acti-vate macrophages and induce phagocytosis. However, as theAa(r)Cdt concentration increased, it resulted in decreased phago-cytic activity (Fig. 2).

3.2. Eukaryotic cell viability

Co-culture of Raw 264.7 cells with the Cdt-deficient strain (D7SSA CHE001) and wild-type strains for 2 h exhibited no difference inLDH release. However, LDH levels were higher when the cells wereco-cultured with the wild type A. actinomycetemcomitans after 20 h(Fig. 3).

The MTT assay was performed in order to confirm that Aa(r)Cdtwas used in sublethal doses in Raw 264.7 cells, there the cell via-bility was not affected in a significant manner compared to nega-tive control cells. The addition of Aa(r)Cdt resulted in about 20%cell death in all tested concentrations (6, 12 and 25 lg/ml). Thesedifferences in cell viability were not statistically significant whencompared to that of the negative control (Fig. 4). Moreover, noapoptosis was detected after 48 h of incubation (Fig. 5).

3.3. Modulation of NO production in the presence of Cdt

NO production was lower when cells were co-cultured with thewild-type (D7S-SA) strain, than with its Cdt-deficient strain, at aMOI 1:100 after 2 h and 20 h of interaction (Fig. 6A and B).Although incubation of Raw 264.7 cells with Aa(r)Cdt resulted inincreased levels of NO, these levels were lower with the additionof 25 lg/ml of the toxin than when with lower doses of Aa(r)Cdt(6 and 12 lg/ml) (Fig. 6C). The transcription profile of inos in-creased 3.02-fold (p = 0.03) after co-culture when 25 lg/ml ofAa(r)Cdt was added.

Fig. 1. Number of intracellular A. actinomycetemcomitans determined by viablebacteria counting after 2 h of co-culture (A) and qPCR after 2 (B) 20 h (C) of co-culture with Raw 264.7 cells. *Statistically significant difference, p < 0.05 (paired t-test).

Fig. 2. Phagocytosis of Escherichia coli bioparticle after incubation differentconcentrations of Aa(r)Cdt (lg/ml). *Statistically significant difference whencompared to control (p < 0.05).

Fig. 3. Raw 264.7 macrophages cytotoxicity (%) detected by LDH release after 2 (A)or 20 h (B) of interaction with A. actinomycetemcomitans D7S and CHE 001 (Dcdt) ina MOI of 1:10 and 1:100. *Statistically significant difference when compared withcontrol (p < 0.05).

E.S. Ando-Suguimoto et al. / Cytokine 66 (2014) 46–53 49

3.4. Cytokine production

Two hours of incubation of Raw 264.7 cells with A. actinomyce-temcomitans resulted in similarly low cytokine levels when thewild-type and Cdt-deficient strains were compared (data notshown). However, incubation of the wild-type strain D7S-SA withthe macrophages for 20 h resulted in higher levels of IL-1b, IL-12and IL-10 than with the Cdt-deficient strain. On the other hand,similar levels of TNF-a were observed in the supernatant of co-cultures with both A. actinomycetemcomitans strains (Table 1).The addition of different concentrations of Aa(r)Cdt to Raw 264.7macrophages resulted in increased production of IL-1b (25 lg/

ml), TNF-a (12 and 25 lg/ml Aa(r)Cdt), and of IL-12 and IL-10(12 lg/ml Aa(r)Cdt) (Fig. 7).

The analysis of mRNA expression demonstrated that the pres-ence of 25 lg/ml Aa(r)Cdt led to up regulation of transcription ofil-1b (63.78-fold; p = 0.025), tnf-a (3.06-fold; p = 0.009), il-12p40(15.85-fold; p = 0.09) and il-10 (794.30-fold; p = 0.034).

4. Discussion

Macrophages play a central role in inflammatory reactions ini-tiated by pathogenic bacteria. Many pathogens have evolved spe-cific mechanisms to avoid, alter, or disable the antimicrobial

Fig. 4. Effect of different concentrations of Aa(r)Cdt (6, 12 and 25 lg/ml) on Raw264.7 macrophages after 48 h of incubation on cell viability determined by MTT.

Table 1Levels of IL-1b, IL-12, TNF-a, and IL-10 in the supernatant of Raw 264.7 macrophagesculture after 20 h of interaction with A. actinomycetemcomitans D7S-SA and A.actinomycetemcomitans D7S CHE001 (MOI 1:100).

Cytokines (pg/ml) ± SD

IL-1b IL-12 TNF-a IL-10

Control 384 (±90) 0 (±0) 1113 (±636) 65 (±8)A.a D7S-SA 1349 (±231)* 169 (±23)* 5661 (±282) 184 (±72) *

A.a D7S-CHE001 312.5 (±140) 9.8 (±18) 5462 (±1092) 60 (±6)

Control: negative control was only Raw 264.7 cells.* Statistically significant difference between wild type A. actinomycetemcomitansD7S-SA and Cdt deficient strain (p < 0.05).

50 E.S. Ando-Suguimoto et al. / Cytokine 66 (2014) 46–53

effects of macrophages. Our previous data indicated that AaCdtwould modulate macrophages function by inhibiting NO produc-tion [31].

Fig. 5. Effect of different concentrations of Aa(r)Cdt (6, 12 and 25 lg/ml) on Raw 264.7stain analyzed by flow cytometer. Positive control: 6 lM of campthotecin.

Fig. 6. Nitric oxide production by Raw 264.7 cells: interaction with A. actinomycetemcomifor 2 (A) and 20 h (B). Raw 264.7 cells challenged with different concentrations of Aa(r)

The present study provides evidence that Cdt may impairphagocytosis by Raw 264.7 cells. Data were obtained not only bycomparing phagocytosis in co-cultures with the Cdt-deficient mu-tant, D7S-SA CHE001 and its wild-type strain, but also with theaddition of Aa(r)Cdt to the D7S-SA CHE001 strain co-culture assays.

cells, after 48 h of incubation on apoptosis, detected by a single positive annexin V

tans D7S-SA and A. actinomycetemcomitans D7S CHE001 (MOI 1:100)after incubationCdt (lg/ml, C). *Statistically significant difference (p < 0.05).

Fig. 7. Levels of IL-1b (A), TNF-a (B), IL-12(C) and IL-10 (D) in the supernatant of Raw 264.7 macrophages culture after addition of 6, 12 and 25 lg/ml of Aa(r)Cdt for 48 hControl: Negative control was only Raw 264.7 cells. *Statistically significant difference when compared with control cells (p < 0.05).

E.S. Ando-Suguimoto et al. / Cytokine 66 (2014) 46–53 51

Furthermore, the decreased phagocytosis obtained with Aa(r)Cdtwas observed not only with A. actinomycetemcomitans, but alsowhen E. coli fluorescent cells were used.

The impaired phagocytosis induced by Cdt after 2 h of co-culturewas not associated with loss in cell viability. A similar effect hadpreviously shown in epithelial cells in a study indicating that Cdttargets the ATM-Chk2 pathway, leading to cellular apoptosis [32].Furthermore, another factor of A. actinomycetemcomitans, distinctfrom Cdt, was suggested to be involved with a reversible DNA frag-mentation that does not lead to terminal apoptosis in oral epithelialcells challenged by the Cdt knockout strain [32]. Thus, our data indi-cated that Cdt was able to inhibit phagocytosis by macrophages at2 h, before inducing a loss in cell viability (20 h). However, de-creased macrophage viability, observed after 20 h of incubation ofthe co-culture with the Cdt-producing strain may have affectedthe bacteria viability since the damage of cellular membrane integ-rity would allow antibiotic killing in the intracellular compartment.

Cdt affected the macrophages production of NO and cytokinesin response to the infection. In the periodontal tissues, as in othertissues infected by microorganisms, the macrophages are activatedby the presence of LPS and production of IFN-c and TNF-a, leadingto the production of effector mediators such as inducible nitricoxide synthase (iNOS) [33]. Cells in periodontal tissues from pa-tients exhibiting periodontitis produce higher levels of iNOS thanlevels detected in periodontally healthy tissues, thus contributingto bacterial killing [34]. It is known that A. actinomycetemcomitansleads to the production of NO by macrophages [35]. However,infection with the Cdt-producing strain induced lower NO levelsthan with the Cdt-deficient strain. This corroborates our previousdata indicating NO inhibition by the Aa(r)Cdt in murine peritonealcells [22]. Despite the NO levels increased in presence of Aa(r)Cdt,these levels diminished when relatively high concentrations of Cdtwere added when compared to the lower concentrations of thetoxin. Under these conditions there was not a significant loss inmacrophages viability.

The host response against A. actinomycetemcomitans in peri-odontitis is characterized by the production of pro-inflammatorycytokines such as IL-1-b and IL-8 by epithelial gingival cells[36,37]. Previous data showed that AaCdt was able to induce thesynthesis of IL-1b, IL-6 and IL-8, but not of TNF-a, by human periph-eral blood mononuclear cells [15]. The present study indicated thatCdt producing A. actinomycetemcomitans induced an increased pro-duction of the pro-inflammatory cytokines IL-1b and IL-12 whencompared to AaCdt-deficient strain. A recent study demonstratedthat A. actinomycetemcomitans induced IL-1b production by NRLP3receptor of inflammasomes, but this activity was not related withCdt or Ltx production [38], suggesting that our observation of in-creased IL-1b production by the wild type A. actinomycetemcomin-tans compared to Cdt deficient strain is due to activation of otherpathways different from inflammasomes. Moreover, as inflamma-somes activate caspase 1 [38], consequently leading to secretionof mature IL-1b the non-activation of this pathway by the recombi-nant CDT could lead to minor secretion of mature IL-1b by macro-phages, despite the high mRNA levels of il1-b.

TNF-a production reached similar levels in the co-cultures withthe wild type or its Cdt-deficient strain. The regulatory cytokineIL-10 was also increased in the presence of Cdt. These results indi-cate that the presence of the Cdt induces the production of earlyresponse cytokine IL-1b, but not of TNF-a. The higher concentra-tions of Aa(r)Cdt decreased the production of IL-12 and IL-10 con-comitantly to an increase in IL-1b and TNF-a production. Theincrease of pro-inflammatory cytokines could lead to lower IL-10production. On the other hand, IL-10 production leads to lower lev-els of IL-12 [39]. Moreover, studies have demonstrated that TNF-ainhibits IL-12 production by macrophages and is independentof IL-10 challenged with Staphylococcus aureus and Listeriamonocytogenes [40,41]. Thus, our data suggest that TNF-a signalingselectively inhibited il-12 gene expression upon macrophage acti-vation as part of the scheme of cytokine feedback and self-limitingmodulations.

52 E.S. Ando-Suguimoto et al. / Cytokine 66 (2014) 46–53

The inhibitory effect of Cdt on phagocytosis was independent ofany other component specific to A. actinomycetemcomitans sincethe effect was also observed with the addition of Aa(r)Cdtto Raw264.7 cells that were co-cultured with fluorescent E. coli. This effectwas not due to cells death. Previous experiments done in our lab-oratory with different amounts of Aa(r)Cdt led to Raw 264.7 celldeath only when high concentrations of the toxin were added(unpublished data). The concentrations evaluated in this studywere sublethal for Raw 264.7 macrophages and no apoptosis wasdetected. Other data revealed that similar concentrations of theAa(r)Cdt were not toxic for human periodontal ligament fibro-blasts, but toxic for gingival epithelial cells [42].

The addition of Aa(r)Cdt to the Raw 264.7 cells resulted in rela-tively high expression of il-1b (63-fold change), although this cyto-kine release by the cells was not significantly affected. Sinceinflammasomes activate the caspase 1 [38] consequently leadingto secretion of mature IL-1b the non-activation of this pathwaycould lead to minor secretion of mature IL-1b by macrophages, de-spite the detection of high mRNA levels of il1-b were detected. Thesecond pro-inflammatory cytokine gene up-regulated by Aa(r)Cdtwas il-12 (15-fold change), followed by tnf-a (3-fold change). Theaddition of Aa(r)Cdt resulted in up-regulation of transcription ofil-10 (794 fold change), while was confirmed by high secreted lev-els of IL-10 in Aa(r)Cdt treated cells. These results suggest that Cdtcould induce the production of pro-inflammatory mediators, suchas IL-1b and IL-12, but not of TNF-a, and even higher levels ofIL-10, which are in accordance with our co-culture assay results.

Periodontitis caused by A. actinomycetemcomitans leads to se-vere destruction of periodontal tissues. Studies using animal mod-els, demonstrated that A. actinomycetemcomitans increases the PMNinfiltration, leading to production of inflammatory mediators andbone destruction [43–45]. Thus, Cdt seems to contribute to the pro-gression of periodontal disease, by the induction of IL-1b and IL-12production, suggesting the activation of macrophages to inflamma-tory (M1) type despite the decrease in their phagocythic activity.Active tissue destruction sites of periodontitis contain high concen-trations of IL-1b and macrophages are the main source of this cyto-kine which seems to be involved in osteoclasts formation andactivation [46,47]. IL-12 is a main mediator of the early innate im-mune response to intracellular microbes, able to stimulate IFN-cproduction by T cells and natural killer cells and promote Th1 re-sponses [48]. In an animal model of periodontal disease, IL-12was produced at the early stage of infection, and switched toIL-10production with the chronicity of the disease [44]. Previous datashowed that Cdt could act as a phosphatase by inhibiting PI 3Kpathway activation [11]. Inhibition of this pathway by PIP3 inhibi-tor in macrophages, induced the production of IL-10 and IL-12[49,50]. Moreover, the increased production of IL-10 induced bythe Cdt would lead to inhibition of the antimicrobial responseand increase pathogen load [51,52]. However, the deleterious ef-fects of pro-inflammatory cytokines such as IL-1b [44] would stillbe seen in infection by Cdt-producing pathogens. Furthermore,the reduction of NO production induced by Cdt would not only re-duce the host’s capacity to control the infection, but also interferewith the bone homeostasis [53]. Thus, Cdt seems to drive the re-sponse to a more chronic pattern, which could lead to the persistentdisease.

5. Conclusion

Taken together, our data indicated that the presence of Cdt maybe an important strategy used by A. actinomycetemcomitans tothwart an effective immune response during a infection, leadingto a decrease in phagocytosis. Additionally, A. actinomycetemcomi-

tans Cdt modulates the immune response indirectly by manipulat-ing cytokines production by macrophages.

Acknowledgements

This work was supported by FAPESP Grants: 09/54178-0 and09/54849-1, NIDCR Grant R01 DE12212 (CC) and an USPHS GrantDE012593 (JMD) from the National Institutes of Health.

References

[1] Graves D. Cytokines that promote periodontal tissue destruction. J Periodontol2008;79:1585–91.

[2] van Winkelhoff AJ, Slots J. Actinobacillus actinomycetemcomitansand Porphyromonas gingivalis in nonoral infections. Periodontol2000;1999(20):122–35.

[3] Paturel L, Casalta JP, Habib G, Nezri M, Raoult D. Actinobacillusactinomycetemcomitans endocarditis. Clin Microbiol Infect 2004;10:98–118.

[4] Mayer MP, Bueno LC, Hansen EJ, DiRienzo JM. Identification of a cytolethaldistending toxin gene locus and features of a virulence-associated region inActinobacillus actinomycetemcomitans. Infect Immun 1999;67:1227–37.

[5] Elwell CA, Dreyfus LA. DNase I homologous residues in CdtB are critical forcytolethal distending toxin-mediated cell cycle arrest. Mol Microbiol2000;37:952–63.

[6] Damek-Poprawa M, Jang JY, Volgina A, Korostoff J, DiRienzo JM. Localization ofAggregatibacter actinomycetemcomitans cytolethal distending toxin subunitsduring intoxication of live cells. Infect Immun 2012;80:2761–70.

[7] Lara-Tejero M, Galan JE. CdtA, CdtB, and CdtC form a tripartite complex that isrequired for cytolethal distending toxin activity. Infect Immun2001;69:4358–65.

[8] Guidi R, Guerra L, Levi L, Stenerlow B, Fox JG, Josenhans C, et al. Chronicexposure to the cytolethal distending toxins of Gram-negative bacteriapromotes genomic instability and altered DNA damage response. CellMicrobiol 2012.

[9] Hassane DC, Lee RB, Pickett CL. Campylobacter jejuni cytolethal distendingtoxin promotes DNA repair responses in normal human cells. Infect Immun2003;71:541–5.

[10] McSweeney LA, Dreyfus LA. Carbohydrate-binding specificity of the Escherichiacoli cytolethal distending toxin CdtA-II and CdtC-II subunits. Infect Immun2005;73:2051–60.

[11] Shenker BJ, Dlakic M, Walker LP, Besack D, Jaffe E, LaBelle E, et al. A novel modeof action for a microbial-derived immunotoxin: the cytolethal distending toxinsubunit B exhibits phosphatidylinositol 3,4,5-triphosphate phosphataseactivity. J Immunol 2007;178:5099–108.

[12] Jinadasa RN, Bloom SE, Weiss RS, Duhamel GE. Cytolethal distending toxin: aconserved bacterial genotoxin that blocks cell cycle progression, leading toapoptosis of a broad range of mammalian cell lineages. Microbiology2011;157:1851–75.

[13] Purdy D, Buswell CM, Hodgson AE, McAlpine K, Henderson I, Leach SA.Characterisation of cytolethal distending toxin (CDT) mutants ofCampylobacter jejuni. J Med Microbiol 2000;49:473–9.

[14] Guidi R, Guerra L, Levi L, Stenerlow B, Fox JG, Josenhans C, et al. Chronicexposure to the cytolethal distending toxins of Gram-negative bacteriapromotes genomic instability and altered DNA damage response. CellMicrobiol 2013;15:98–113.

[15] Akifusa S, Poole S, Lewthwaite J, Henderson B, Nair SP. RecombinantActinobacillus actinomycetemcomitans cytolethal distending toxin proteins arerequired to interact to inhibit human cell cycle progression and to stimulatehuman leukocyte cytokine synthesis. Infect Immun 2001;69:5925–30.

[16] Belibasakis GN, Brage M, Lagergard T, Johansson A. Cytolethal distending toxinupregulates RANKL expression in Jurkat T-cells. APMIS. APMIS: Acta Pathol.,Microbiol., Immunol. Scand. 2008;116:499–506.

[17] Belibasakis GN, Johansson A, Wang Y, Chen C, Kalfas S, Lerner UH. Thecytolethal distending toxin induces receptor activator of NF-kappaB ligandexpression in human gingival fibroblasts and periodontal ligament cells. InfectImmun 2005;73:342–51.

[18] Belibasakis GN, Johansson A, Wang Y, Chen C, Lagergard T, Kalfas S, et al.Cytokine responses of human gingival fibroblasts to Actinobacillusactinomycetemcomitans cytolethal distending toxin. Cytokine 2005;30:56–63.

[19] Mosser DM, Edwards JP. Exploring the full spectrum of macrophage activation.Nat Rev Immunol 2008;8:958–69.

[20] Hickey TE, Majam G, Guerry P. Intracellular survival of Campylobacter jejuni inhuman monocytic cells and induction of apoptotic death by cytholethaldistending toxin. Infect Immun 2005;73:5194–7.

[21] Xu T, Lundqvist A, Ahmed HJ, Eriksson K, Yang Y, Lagergard T. Interactions ofHaemophilus ducreyi and purified cytolethal distending toxin with humanmonocyte-derived dendritic cells, macrophages and CD4+ T cells. MicrobesInfect 2004;6:1171–81.

[22] Fernandes KP, Mayer MP, Ando ES, Ulbrich AG, Amarente-Mendes JG, Russo M.Inhibition of interferon-gamma-induced nitric oxide production in endotoxin-activated macrophages by cytolethal distending toxin. Oral MicrobiolImmunol 2008;23:360–6.

E.S. Ando-Suguimoto et al. / Cytokine 66 (2014) 46–53 53

[23] Rabin SD, Flitton JG, Demuth DR. Aggregatibacter actinomycetemcomitanscytolethal distending toxin induces apoptosis in nonproliferatingmacrophages by a phosphatase-independent mechanism. Infect Immun2009;77:3161–9.

[24] Wang Y, Liu A, Chen C. Genetic basis for conversion of rough-to-smooth colonymorphology in Actinobacillus actinomycetemcomitans. Infect Immun2005;73:3749–53.

[25] Chen C, Kittichotirat W, Chen W, Downey JS, Si Y, Bumgarner R. Genomesequence of naturally competent Aggregatibacter actinomycetemcomitansserotype a strain D7S-1. J Bacteriol 2010;192:2643–4.

[26] Nalbant A, Chen C, Wang Y, Zadeh HH. Induction of T-cell apoptosis byActinobacillus actinomycetemcomitans mutants with deletion of ltxA andcdtABC genes: possible activity of GroEL-like molecule. Oral MicrobiolImmunol 2003;18:339–49.

[27] Mao X, DiRienzo JM. Functional studies of the recombinant subunits of acytolethal distending holotoxin. Cell Microbiol 2002;4:245–55.

[28] Cao L, Volgina A, Korostoff J, DiRienzo JM. Role of intrachain disulfides in theactivities of the CdtA and CdtC subunits of the cytolethal distending toxin ofActinobacillus actinomycetemcomitans. Infect Immun 2006;74:4990–5002.

[29] Chen C, Kittichotirat W, Si Y, Bumgarner R. Genome sequence ofAggregatibacter actinomycetemcomitans serotype c strain D11S-1. J Bacteriol2009;191:7378–9.

[30] Ding AH, Nathan CF, Stuehr DJ. Release of reactive nitrogen intermediates andreactive oxygen intermediates from mouse peritoneal macrophages.Comparison of activating cytokines and evidence for independentproduction. J Immunol 1988;141:2407–12.

[31] Fernandes DM, Baldwin CL. Interleukin-10 downregulates protectiveimmunity to Brucella abortus. Infect Immun 1995;63:1130–3.

[32] Alaoui-El-Azher M, Mans JJ, Baker HV, Chen C, Progulske-Fox A, Lamont RJ,et al. Role of the ATM-checkpoint kinase 2 pathway in CDT-mediatedapoptosis of gingival epithelial cells. PLoS One 2010;5:e11714.

[33] Martinez FO, Sica A, Mantovani A, Locati M. Macrophage activation andpolarization. Front Biosci 2008;13:453–61.

[34] Lappin DF, Kjeldsen M, Sander L, Kinane DF. Inducible nitric oxide synthaseexpression in periodontitis. J Periodontal Res 2000;35:369–73.

[35] Sosroseno W, Bird PS, Seymour GJ. Nitric oxide production by a murinemacrophage cell line (RAW264.7 cells) stimulated with Aggregatibacteractinomycetemcomitans surface-associated material. Anaerobe2011;17:246–51.

[36] Laube DM, Dongari-Bagtzoglou A, Kashleva H, Eskdale J, Gallagher G, DiamondG. Differential regulation of innate immune response genes in gingivalepithelial cells stimulated with Aggregatibacter actinomycetemcomitans. JPeriodontal Res 2008;43:116–23.

[37] Umeda JE, Demuth DR, Ando ES, Faveri M, Mayer MP. Signaling transductionanalysis in gingival epithelial cells after infection with Aggregatibacteractinomycetemcomitans. Mol Oral Microbiol 2012;27:23–33.

[38] Belibasakis GN, Johansson A. Aggregatibacter actinomycetemcomitans targetsNLRP3 and NLRP6 inflammasome expression in human mononuclearleukocytes. Cytokine 2012;59:124–30.

[39] D’Andrea A, Ma X, Aste-Amezaga M, Paganin C, Trinchieri G. Stimulatory andinhibitory effects of interleukin (IL)-4 and IL-13 on the production of cytokinesby human peripheral blood mononuclear cells: priming for IL-12 and tumornecrosis factor alpha production. J Exp Med 1995;181:537–46.

[40] Ma X, Sun J, Papasavvas E, Riemann H, Robertson S, Marshall J, et al. Inhibitionof IL-12 production in human monocyte-derived macrophages by TNF. JImmunol 2000;164:1722–9.

[41] Zakharova M, Ziegler HK. Paradoxical anti-inflammatory actions of TNF-alpha:inhibition of IL-12 and IL-23 via TNF receptor 1 in macrophages and dendriticcells. J Immunol 2005;175:5024–33.

[42] Kanno F, Korostoff J, Volgina A, DiRienzo JM. Resistance of human periodontalligament fibroblasts to the cytolethal distending toxin of Actinobacillusactinomycetemcomitans. J Periodontol 2005;76:1189–201.

[43] Bezerra Bde B, Andriankaja O, Kang J, Pacios S, Bae HJ, Li Y, et al. A.actinomycetemcomitans-induced periodontal disease promotes systemic andlocal responses in rat periodontium. J Clin Periodontol 2012;39:333–41.

[44] Garlet GP, Cardoso CR, Silva TA, Ferreira BR, Avila-Campos MJ, Cunha FQ, et al.Cytokine pattern determines the progression of experimental periodontaldisease induced by Actinobacillus actinomycetemcomitans through themodulation of MMPs, RANKL, and their physiological inhibitors. OralMicrobiol Immunol 2006;21:12–20.

[45] Kang J, de Brito Bezerra B, Pacios S, Andriankaja O, Li Y, Tsiagbe V, et al.Aggregatibacter actinomycetemcomitans infection enhances apoptosis in vivothrough a caspase-3-dependent mechanism in experimental periodontitis.Infect Immun 2012;80:2247–56.

[46] Fukushima H, Jimi E, Okamoto F, Motokawa W, Okabe K. IL-1-induced receptoractivator of NF-kappa B ligand in human periodontal ligament cells involvesERK-dependent PGE2 production. Bone 2005;36:267–75.

[47] Matsuki Y, Yamamoto T, Hara K. Interleukin-1 mRNA-expressing macrophagesin human chronically inflamed gingival tissues. Am J Pathol1991;138:1299–305.

[48] Gately MK, Renzetti LM, Magram J, Stern AS, Adorini L, Gubler U, et al. Theinterleukin-12/interleukin-12-receptor system: role in normal and pathologicimmune responses. Annu Rev Immunol 1998;16:495–521.

[49] Hazeki K, Kametani Y, Murakami H, Uehara M, Ishikawa Y, Nigorikawa K, et al.Phosphoinositide 3-kinasegamma controls the intracellular localization of CpGto limit DNA-PKcs-dependent IL-10 production in macrophages. PLoS One2011;6:e26836.

[50] Hazeki K, Nigorikawa K, Hazeki O. Role of phosphoinositide 3-kinase in innateimmunity. Biol Pharm Bull 2007;30:1617–23.

[51] Lazarus JJ, Meadows MJ, Lintner RE, Wooten RM. IL-10 deficiency promotesincreased Borrelia burgdorferi clearance predominantly through enhancedinnate immune responses. J Immunol 2006;177:7076–85.

[52] Reed SG, Brownell CE, Russo DM, Silva JS, Grabstein KH, Morrissey PJ. IL-10mediates susceptibility to Trypanosoma cruzi infection. J Immunol1994;153:3135–40.

[53] Herrera BS, Martins-Porto R, Maia-Dantas A, Campi P, Spolidorio LC, Costa SK,et al. INOS-derived nitric oxide stimulates osteoclast activity and alveolar boneloss in ligature-induced periodontitis in rats. J Periodontol 2011;82:1608–15.