The B Subunit of an AB5 Toxin Produced by Salmonella enterica Serovar Typhi Up-Regulates Chemokines, Cytokines, and Adhesion Molecules in Human Macrophage, Colonic Epithelial, and Brain Microvascular Endothelial Cell Lines

Abstract

The principal function of bacterial AB5 toxin B subunits is to interact with glycan receptors on the surfaces of target cells and mediate the internalization of holotoxin. However, B subunit-receptor interactions also have the potential to impact cell signaling pathways and, in so doing, contribute to pathogenesis independently of the catalytic (toxic) A subunits. Various Salmonella enterica serovars, including Salmonella enterica serovar Typhi, encode an AB5 toxin (ArtAB), the A subunit of which is an ADP-ribosyltransferase related to the S1 subunit of pertussis toxin. However, although the A subunit is able to catalyze ADP-ribosylation of host G proteins, a cytotoxic phenotype has yet to be identified for the holotoxin. We therefore examined the capacity of the purified B subunit (ArtB) from S. Typhi to elicit cytokine, chemokine, and adhesion molecule responses in human macrophage (U937), colonic epithelial (HCT-8) cell, and brain microvascular endothelial cell (HBMEC) lines. Secretion of the chemokines monocyte chemotactic protein 1 (MCP-1) and interleukin 8 (IL-8) was increased in all three tested cell lines, with macrophage inflammatory protein 1α (MIP-1α), MIP-1β, and granulocyte colony-stimulating factor (G-CSF) also significantly increased in U937 cells. ArtB also upregulated the cytokines tumor necrosis factor alpha (TNF-α) and IL-6 in HBMECs and HCT-8 cells, but not in U937 cells, while intercellular adhesion molecule 1 (ICAM-1) was upregulated in HCT-8 and U937 cells and vascular cell adhesion molecule 1 (VCAM-1) was upregulated in HBMECs. Thus, ArtB may contribute to pathogenesis independently of the A subunit by promoting and maintaining a strong inflammatory response at the site of infection.

INTRODUCTION

Bacterial AB5 toxins are so termed because they comprise a catalytic A subunit noncovalently linked to a pentameric B subunit. They exert their effects in a two-step process: first, the B subunit pentamer binds to specific glycan receptors on the cell surface, triggering uptake of the holotoxin; this is followed by inhibition or corruption of essential host functions, mediated by the enzymatic activity of the A subunit. AB5 toxins are critical weapons in the armory of virulence factors deployed by major bacterial pathogens, which collectively kill over a million people each year (1). The AB5 toxins characterized to date are classified into four families according to A subunit sequence homology and catalytic activity, as well as the structural organization of the holotoxin (1). The A subunits of both the cholera toxin (Ctx) family (comprising Ctx and the enterotoxigenic Escherichia coli labile enterotoxin [LT]) and the pertussis toxin (Ptx) family catalyze the ADP-ribosylation of G_sα and G_iα proteins in the host cell cytosol, disrupting their signal transduction pathways. The A subunits of the Shiga toxin (Stx) family have RNA N-glycosidase activity and inhibit eukaryotic protein synthesis by cleaving a specific adenine base from 28S rRNA. The fourth and most recently discovered AB5 toxin family is subtilase cytotoxin (SubAB), produced by a subset of Shiga-toxigenic E. coli (STEC) strains (2). Its A subunit is a highly specific subtilase-like serine protease that cleaves the essential endoplasmic reticulum (ER) chaperone BiP/GRP78, thereby inducing a massive ER stress response and triggering cellular apoptosis (3–6).

The B subunits of the Ctx and Stx families (CtxB and StxB, respectively) recognize glycans displayed by host glycolipids (gangliosides, including GM1, and glycosphingolipids, such as Gb3 and Gb4) (1). The Ptx family is unique among AB5 toxins, as its B subunit is not a homopentamer, comprising 4 nonidentical subunits (S2, S3, S4, and S5) in a 1:1:2:1 stoichiometry. The S2 and S3 subunits both bind to sialylated glycoproteins rather than glycolipids (1). The B subunit of SubAB (SubB) is a homopentamer, like CtxB and StxB, but it binds to glycoproteins like Ptx. SubB and the C-terminal portion of Ptx S2 exhibit strong structural homology, in spite of relatively low amino acid sequence identity (18% over a 95-amino-acid [aa] overlap) (7). Interestingly, SubB displays a high degree of specificity for glycans terminating in N-glycolylneuraminic acid (Neu5Gc), a sialic acid that is not synthesized in humans due to a deletion mutation in the cmah gene. Humans should be genetically resistant to the SubAB toxin. However, assimilation of Neu5Gc from dietary sources enables the expression of high-affinity receptors on the cell surface, thereby conferring susceptibility to the lethal effects of SubAB (7, 8).

In our original description of SubAB (2), we also reported the presence of homologues of SubB encoded on the genomes of both Yersinia pestis (57% amino acid identity) and Salmonella enterica serovar Typhi (51% identity). BLAST searches of current genome sequences revealed SubB homologues in other pathogenic Yersinia species and S. enterica serovars, including Salmonella enterica serovar Paratyphi, Salmonella enterica serovar Typhimurium, Salmonella enterica serovar Montevideo, and Salmonella enterica serovar Arizonae. All these proteins have now been classified as superfamily PRK15265 (subtilase cytotoxin B subunit-like proteins). The Salmonella genes are parts of operons that also contain homologues of the Ptx catalytic S1 subunit. The putative S. Typhimurium AB5 toxin has been named ArtAB (for ADP-ribosylating toxin) (9), and recent studies have confirmed that, like Ptx S1, S. Typhimurium ArtA is an ADP-ribosyltransferase capable of modifying host cell G proteins (10). However, to date, no A subunit-dependent cytotoxic phenotype has been identified for ArtAB. Interestingly, the ArtAB homologue from S. Typhi (also referred to as PltAB, for pertussis-like toxin) has been reported to associate with cytolethal distending toxin and to deliver it from an intracellular compartment to target cells via autocrine and paracrine pathways (11). Although both A and B subunits were required for this property, it was not dependent on A subunit ADP-ribosyltransferase activity (11).

Although the principal function of AB5 toxin B subunits is to interact with receptors on the surfaces of target cells and to mediate internalization of holotoxin, interactions with surface molecules have the potential to impact a variety of signaling pathways. In so doing, the B subunits may directly contribute to pathogenesis by triggering host responses independently of their respective catalytic (“toxic”) A subunits. In the present study, we have examined the capacity of purified S. Typhi ArtB to elicit cytokine, chemokine, and adhesion molecule responses in human macrophage and colonic epithelial and brain microvascular endothelial cell lines.

MATERIALS AND METHODS

Purification of ArtB and ArtAB.

The artB open reading frame (ORF) was first amplified by PCR using an S. enterica serovar Typhi T2 chromosomal-DNA template and primers STyArtBF (5′-GCGCGCGAATTCTTTTAAACGTACAGGAGAGTAG-3′) and STyArtBHis6R (5′-GCAGCAAAGCTTTTAGTGGTGGTGGTGGTGGTGCTTG-3′). The resultant PCR product, which incorporates EcoRI and HindIII restriction sites, as well as fusing a region encoding a His₆ tag on the 3′ terminus of artB, was then cloned into pBAD18 (12), so that the artB ORF is under the control of the vector ara promoter. The complete artBA ORF was also cloned into pBAD18 by fusing the native artA ORF immediately downstream of artB-His₆ in the above construct (unlike most other AB5 toxin operons, artB precedes artA in the S. Typhi chromosome). These constructs were then transformed into an E. coli BL21(DE3) lpxM mutant (13) that produces a penta-acylated lipopolysaccharide (LPS) that has very low endotoxic activity. The recombinant bacterium was grown in 500 ml Terrific Broth (14) at 37°C to late logarithmic phase, diluted 50:50 with fresh medium supplemented with 0.2% arabinose to induce artB or artBA expression, and then incubated overnight at 26°C. Cells were harvested by centrifugation, resuspended in 20 ml loading buffer (50 mM sodium phosphate, 300 mM NaCl, 20 mM imidazole, pH 8.0), and lysed in a French pressure cell. Cell debris was removed by centrifugation at 20,000 × g for 30 min at 4°C. The supernatant was then loaded onto a 2-ml column of Ni-nitrilotriacetic acid (NTA) resin, which had been preequilibrated with 20 ml loading buffer. The column was then washed with 40 ml loading buffer, and bound proteins were eluted with a 30-ml gradient of 0 to 500 mM imidazole in loading buffer; 3-ml fractions were collected and analyzed by SDS-PAGE, followed by staining with Coomassie blue or Western blotting with monoclonal anti-His₆. ArtB migrates as a single 16-kDa species on SDS-PAGE, while ArtAB has an additional 26-kDa species corresponding to ArtA; densitometric analysis of ArtAB gels was consistent with 1:5 stoichiometry. Fractions containing >95% pure ArtB or ArtAB were pooled. Under the above conditions, ArtB and ArtAB retain their native folding and conformation.

Cell culture and toxin treatment.

Human brain microvascular endothelial cells (HBMEC) were grown at 37°C in 5% carbon dioxide in culture medium (RPMI 1640 medium [GIBCO 11875] and F-12 Nutrient Mixture [GIBCO 11765] [1:1], supplemented with 10 mM HEPES, 2 mM l-glutamine, 1 mM sodium pyruvate, 10% heat-inactivated fetal calf serum [FCS], 50 IU of penicillin, and 50 μg/ml of streptomycin). U937 (human macrophages) and HCT-8 (human colonic epithelial cells) were grown at 37°C and 5% CO₂ in RPMI 1640 medium supplemented with 10 mM HEPES, 2 mM l-glutamine, 1 mM sodium pyruvate, 10% heat-inactivated FCS, 50 IU of penicillin, and 50 mg/ml streptomycin. For the toxin treatment, cells were seeded into 24-well tissue culture plates; 1 million U937 cells per well or confluent monolayers of HBMEC (2 × 10⁵/well) or HCT-8 cells (5 × 10⁵/well) were then exposed to ArtB at the indicated concentration in 300 μl of culture medium containing 1% FCS for the indicated times.

Examination of apoptosis and necrosis.

Apoptosis and necrosis were detected by differential staining with annexin V (which stains apoptotic and necrotic cells) and propidium iodide (PI) (which stains necrotic cells only) using a Vybrant Apoptosis Assay Kit 2 (Invitrogen V13241) according to the manufacturer's instructions. Briefly, 10⁶ U937 cells were treated for 24 h with 10 μg/ml ArtB or ArtAB and transferred from the wells of the tissue culture tray into fluorescence-activated cell sorter (FACS) tubes (Falcon 352008). The cells were washed three times with 3 ml annexin V binding buffer and resuspended in 50 μl of annexin V working reagent, consisting of 1/25 annexin V-Alexa 488 (which emits at 519 nm) and 1 μg/ml PI (which emits at 617 nm) in annexin V binding buffer. The cells were incubated at room temperature in the dark for 15 to 20 min and then washed and analyzed immediately on a FACscan flow cytometer.

Measurement of secreted chemokines, cytokines, and adhesion molecules.

Secreted chemokines, cytokines, and adhesion molecules in cell culture supernatants were determined using Cytometric Bead Array Kits (Bender Medsystems) according to the manufacturer's instructions. The Human Chemo/Cytokine 9plex kit included 6 chemokines—granulocyte colony-stimulating factor (G-CSF), interleukin 8 (IL-8), monocyte chemotactic protein 1 (MCP-1), monokine induced by gamma interferon (MIG), macrophage inflammatory protein1α (MIP-1α), and MIP-1β—and 3 cytokines—IL-6, tumor necrosis factor -alpha (TNF-α), and IL-1β. The Human Adhesion 6plex kit included the adhesion molecules E selectin, intercellular adhesion molecule 1 (ICAM-1), ICAM-3, platelet endothelial cell adhesion molecule 1 (PECAM-1), P selectin, and vascular cell adhesion molecule 1 (VCAM-1). The cytometric bead array technique utilizes beads of different sizes coated with specific capturing antibodies, which are labeled with an internal fluorescent dye (which emits at 700 nm) at discrete intensities. Up to 20 bead populations distinguished by internal dye intensity and bead size allow the simultaneous quantification of up to 20 analytes in a small-volume sample. During incubation of the beads with culture supernatant, different analytes were captured by their corresponding beads. The bead mixtures were then mixed with phycoerythrin (PE)-conjugated adhesion molecule-detecting antibodies or biotin-conjugated chemokine/cytokine-detecting antibodies, followed by streptavidin-PE to form sandwich complexes. Following incubation and washing, fluorescence data were acquired with a BD FACSCanto (serial number V0130) or BD LSR II with high-throughput sampler (serial number H1169) plus BD FACSDiva Software (version 5.0.3) and analyzed with WEASLE v2.6 software. The fluorescence intensity of PE (which emits at 575 nm) is proportional to the amount of analytes present in the test samples. The data are presented as means and standard errors (SE), and differences were analyzed using Student's t test.

Measurement of cell-bound chemokines, cytokines, and adhesion molecules.

Cell-bound chemokines, cytokines, or adhesion molecules were directly probed by incubating ArtB-treated cells (fixed with 1% paraformaldehyde and permeabilized with 0.05% Triton X-100 in phosphate-buffered saline [PBS]) with PE-conjugated adhesion molecule-detecting antibodies or biotin-conjugated chemokine/cytokine-detecting antibodies, followed by streptavidin-PE. At the end of the experiments, after 3 PBS washes, the cells were analyzed with a BD FACSCanto, and the data were acquired with BD FACSDiva software (version 5.0.3) and analyzed with WEASLE v2.6.

Measurement of mRNA levels of chemokines/cytokines and adhesion molecules.

HCT-8 cells were seeded in 6-well tissue culture plates at 3.6 × 10⁶ cells/well and incubated overnight (to confluence). The HCT-8 cells were then washed with PBS and incubated for 4 h in the presence of 10 μg/ml ArtB or vehicle control in serum- and antibiotic-free medium. U937 cells were plated into 6-well tissue culture trays at 2.65 × 10⁶ cells/well in serum- and antibiotic-free medium and then incubated for 4 h in the presence of 10 μg/ml ArtB or vehicle control. RNA was then extracted from the HCT-8 or U937 cells using RNeasy minicolumns and QIAshredder columns according to the manufacturer's instructions (Qiagen, Valencia, CA). RNasin RNase inhibitor (Promega, Madison, WI) was added to samples. Contaminating DNA was digested with RNase-free DNase I (Roche Applied Science, Sydney, Australia) according to the manufacturer's instructions. The absence of DNA contamination in all RNA preparations was confirmed by reverse transcription (RT)-PCR analysis using primers specific for the gene encoding the housekeeping enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH), as previously described (15). The RNA was then recleaned using RNeasy minicolumns according to the manufacturer's instructions (Qiagen). A total of 1 μg RNA was synthesized into cDNA using the RT² first-strand kit according to the manufacturer's instructions (SABiosciences). A subset of human chemokines/cytokines or adhesion molecule mRNAs (as mentioned above) were quantitated using PCR arrays or real-time RT-PCR and run in a LightCycler 480 (Roche) according to the manufacturer's instructions in 96-well plates (PAHS-225F; SABiosciences) or on a Rotorgene RG-2000 Cycler (Corbett Research). Each treatment shown is the result of three independent experiments. The results were calculated using the RT² Profiler PCR array data analysis template v3.3, which uses the comparative cycle threshold (2^−ΔΔCT)) method and Student's t test for data analysis (SABiosciences). The data are presented as means and standard deviations (SD).

RESULTS

ArtB and ArtAB-induced apoptosis and necrosis in U937 cells.

As an initial experiment, general cytotoxicity was assessed by determining rates of apoptosis and necrosis in U937 cells exposed to 10 μg/ml of purified ArtB or ArtAB for 24 h by flow cytometry (Fig. 1). At this time point, approximately 21% of ArtB- and 12% of ArtAB-treated cells were apoptotic (annexin V positive/PI negative), and 14% of ArtB-treated cells and 0.5% of ArtAB-treated cells were necrotic (annexin V positive/PI positive), compared to approximately 7.5% apoptotic and 0.1% necrotic for untreated control cells. The higher potency of ArtB than ArtAB may be attributable to the higher actual dose in molar terms and also serves to underscore the phenotypic dominance of the B subunit. Thus, subsequent experiments were focused on the effects of ArtB alone. Decreases in cell size and complexity were also seen in HCT-8 cells treated with 10 μg/ml ArtB for 24 or 48 h relative to control cells, as judged by FACS analysis (see Fig. S1 in the supplemental material). In view of this evidence of general cytotoxicity, 10 μg/ml and 24 h were set as the maximum dose and time tested in most experiments.

Fig 1.

Effect of ArtB on apoptosis and necrosis. U937 cells were incubated with ArtB or ArtAB (10 μg/ml) for 24 h. The cells were then labeled live with annexin V and PI and analyzed by FACS. (A) Dot plots of annexin V versus PI relative fluorescence intensities (RFI). Cells in the lower left quadrant (annexin V and PI double negative) were classified as living; cells in the lower right (LR) quadrant (only annexin V positive) were classified as apoptotic; cells in the upper right (UR) quadrant (annexin V and PI double positive) were considered necrotic. (B) Bar chart showing percent apoptosis and necrosis derived from 10,000 cell counts; the data are representative of two independent experiments.

ArtB-induced changes in chemokine and cytokine secretion.

The effect of treatment with 10 μg/ml ArtB for up to 24 h on secretion of selected chemokines (G-CSF, IL-8, MCP-1, MIG, MIP-1α, and MIP-1β) and cytokines (IL-6, TNF-α, and IL-1β) by HCT-8 and U937 cell and HBMEC lines was then investigated using FACS bead array assays. At 1 h for HCT-8 and U937 cells, there was no significant difference in the secretion of the tested chemokines or cytokines relative to vehicle-treated controls, while for HBMEC, a slight increase in TNF-α and MCP-1 was detected (see Fig. S2 in the supplemental material). However, at 4 h, for all three tested cell lines, IL-8 secretion was significantly elevated (see Fig. S2), while the secretion of IL-6 and MCP-1 increased in HBMEC and the secretion of MIP-1α, MIP-1β, and G-CSF increased in U937 cells (see Fig. S2). At 24 h, there was significant elevation of secretion of the chemokines IL-8 and MCP-1 in all three tested cell lines; secretion of the chemokines MIP-1α, MIP-1β, and MIG was significantly increased in U937 cells only, while secretion of the proinflammatory cytokines TNF-α and IL-6 was elevated in HBMEC and HCT-8 cells, but not in U937 cells (Fig. 2).

Fig 2.

ArtB-induced changes in secreted chemokines/cytokines at 24 h. HCT-8 or U937 cells or HBMEC were treated with 10 μg/ml ArtB for 24 h or untreated. Levels of secreted chemokines or cytokines (as indicated) were assayed using a cytometric bead array (see Materials and Methods). The data are presented as RFI for each chemokine/cytokine (means and standard errors of the mean [SEM] from at least three experiments). ***, P < 0.001; **, P < 0.01; and *, P < 0.05 relative to control cells (Student's unpaired two-tailed t test).

Since chemokine and cytokine responses were maximal at 24 h post-ArtB treatment, dose response at that time point was investigated (Fig. 3). HBMEC were the most responsive cell line, exhibiting highly significant (P < 0.001) elevations in IL-8, MCP-1, and IL-6 secretion at 1.25 μg/ml, the lowest ArtB concentration tested. U937 cells exhibited significant elevation of IL-8, MCP-1, and MIP-1α at 10 μg/ml ArtB and above, while HCT-8 cells exhibited significant elevation of IL-8 and MCP-1 secretion at and above 2.5 and 5 μg/ml ArtB, respectively.

Fig 3.

ArtB dose-dependent changes in secreted chemokines/cytokines. HCT-8 or U937 cells or HBMEC were treated with the indicated doses of ArtB for 24 h, and secreted chemokines or cytokines were assayed using a cytometric bead array (see Materials and Methods). The data are presented as RFI for each chemokine/cytokine (means and SEM from at least three experiments). ***, P < 0.001; **, P < 0.01; and *, P < 0.05 relative to control cells (Student's unpaired two-tailed t test).

ArtB-induced changes in cell-bound chemokine and cytokine levels.

Cell-bound chemokine and cytokine levels were also assayed. Initially, HCT-8 cells were treated with 10 μg/ml ArtB for either 1, 4, or 24 h, and increases in expression of G-CSF, IL-8, MCP-1, MIG, MIP-1α, MIP-1β, IL-6, and TNF-α were observed at all time points (see Fig. S3A in the supplemental material). Responses at 24 h were dose dependent for all these chemokines/cytokines, with the maximum responses elicited at 10 μg/ml, except for IL-6 and TNF, which showed further increases at 20 μg/ml (Fig. 4A). For U937 cells, treatment with 10 μg/ml ArtB resulted in higher surface expression of IL-8, MCP-1, MIP-1α, MIP-1β, and G-CSF from 1 h and MIG from 4 h (see Fig. S3B in the supplemental material). Dose-response experiments after 24 h treatment indicated that U937 cells were somewhat less sensitive to ArtB than HCT-8 cells, with maximal responses occurring at 20 μg/ml in all cases (Fig. 4B). Consistent with findings for secreted cytokines, there were no increases in cell-bound IL-6 or TNF-α in ArtB-treated U937 cells (results not shown).

Fig 4.

ArtB dose-dependent changes in cell-bound chemokines/cytokines in HCT-8 (A) and U937 (B) cells. HCT-8 or U397 cells were treated with the indicated doses of ArtB for 24 h. Cell-bound chemokines or cytokines were measured by direct probing with biotin-conjugated specific antibodies, followed by streptavidin-PE, and analyzed by FACS. The data are the mean chemokine or cytokine RFI, which were acquired from at least 5,000 cells. The results are representative of two independent experiments.

Transcriptional responses to ArtB.

To determine whether the effects of ArtB and ArtAB on chemokine/cytokine expression at the protein level were also reflected at the transcriptional level, mRNA levels were measured by quantitative real-time reverse transcription-PCR arrays 4 h after treatment of either HCT-8 or U937 cells with 10 μg/ml ArtB (Fig. 5A). For HCT-8 cells, IL-8 and TNF-α mRNA levels were upregulated 18.6-fold and 2.4-fold, respectively, relative to untreated control cells (P < 0.05 in both cases). For U937 cells, MCP-1 and IL-8 mRNAs were upregulated 6.4-fold and 132-fold, respectively, relative to untreated control cells (P < 0.05 in both cases). RT-PCR was also used to compare the effects of ArtB and the ArtAB holotoxin on IL-1β, TNF-α, and IL-8 mRNA levels in U937 cells at 4 h (Fig. 5B). While both elicited significant gene upregulation at 10 μg/ml, ArtB was slightly more potent than ArtAB, consistent with the earlier findings for cytotoxicity (Fig. 5B).

Fig 5.

ArtB-induced changes in chemokine/cytokine mRNA in HCT-8 and U937 cells. HCT-8 or U937 cells were treated with 10 μg/ml ArtB for 4 h or left untreated (A); alternatively, U937 cells were treated with 10 μg/ml ArtB or ArtAB for 4 h (B). Total RNA was then extracted from the cells, and chemokine/cytokine mRNA (as indicated) was quantitated by PCR arrays (A) or real-time RT-PCR (B) (see Materials and Methods). The results are expressed as the fold change in [mRNA] relative to levels in control cells. The data are shown as the means and SD of three independent experiments; ***, P < 0.001; **, P < 0.01; and *, P < 0.05 relative to control cells (Student's unpaired two-tailed t test).

ArtB-induced changes in adhesion molecules.

FACS bead arrays were also used to measure levels of the adhesion molecules ICAM-1, E selectin, ICAM-3, PECAM-1, VCAM-1, and P selectin in the culture supernatants of the various cell types treated with 10 μg/ml ArtB for 1, 4, and 24 h. At 1 h for all three cell lines, there was no significant difference in the secretion of the tested adhesion molecules relative to vehicle-treated controls (see Fig. S4 in the supplemental material). However, at 4 h for HCT-8 cells, the secretion of all the tested adhesion molecules (except ICAM-1) was elevated, whereas for U937 cells, no significant increases were observed, while secretion of both ICAM-1 and VCAM-1 was increased at 4 h in HBMEC (Fig. S4). At 24 h, there was significant elevation of the secretion of ICAM-1 in both HCT-8 and U937 cells, while the secretion of VCAM-1was elevated in HBMEC only (Fig. 6A).

Fig 6.

ArtB-induced changes in secreted adhesion molecules. (A) HCT-8 or U937 cells or HBMEC were treated with 10 μg/ml ArtB for 24 h or left untreated. (B) HCT-8 or U937 cells or HBMEC were treated with the indicated dose of ArtB for 24 h. Levels of secreted adhesion molecules (as indicated) were determined using a cytometric bead array (see Materials or Methods). The data are presented as RFI for each adhesion molecule (means and SEM from at least three experiments). **, P < 0.01, and *, P < 0.05 relative to control cells (Student's unpaired two-tailed t test).

Dose-response analysis at 24 h (Fig. 6B) showed that HCT-8 cells were more sensitive to ArtB than U937 cells in terms of ICAM-1 response, with significant elevation seen at doses as low as 2.5 μg/ml in the former. VCAM-1 responses of HBMEC were also highly responsive to ArtB and were near maximal at 2.5 μg/ml.

Expression of cell-bound ICAM-1 and VCAM-1 in ArtB-treated HCT-8 and U937 cells was also examined. At 10 μg/ml, ArtB treatment increased cell-associated expression of ICAM-1, but not VCAM-1, in both cell types at 24 h (Fig. 7A), which is consistent with the levels observed in culture supernatant (see above). Dose-response experiments (Fig. 7B) indicated that cell-associated ICAM-1 expression was maximal in HCT-8 cells at 5 μg/ml, while expression in U937 cells increased up to 20 μg/ml, the maximum dose tested.

Fig 7.

ArtB-induced changes in cell-bound adhesion molecules. (A) HCT-8 or U937 cells were treated with ArtB (10 μg/ml) for 1, 4, or 24 h or left untreated. (B) HCT-8 or U937 cells were treated with the indicated dose of ArtB for 24 h. The cell-bound adhesion molecules ICAM-1 and VCAM-1 were measured by direct probing with PE-conjugated adhesion molecule antibodies and analyzed by FACS. The data are the mean RFI, which were acquired from at least 5,000 cells. The results are representative of two independent experiments.

ICAM-1, VCAM-1, E selectin, and P selectin mRNA levels were also measured in HCT-8 and U937 cells treated with 10 μg/ml ArtB for 4 h (Fig. 8). ICAM-1 was the only adhesion molecule to exhibit significantly increased transcription (1.3-fold and 3.8-fold, respectively; P < 0.05 in both cases). These findings are consistent with the data obtained at the protein level.

Fig 8.

ArtB-induced changes in adhesion molecule mRNA. HCT-8 or U937 cells were treated with ArtB (10 μg/ml) for 4 h or left untreated. Total RNA was then extracted from cells, and adhesion molecule mRNAs (as indicated) were quantitated using PCR arrays (see Materials and Methods). The results are expressed as the fold change in [mRNA] relative to levels in control cells. The data are shown as the means of three independent experiments; *, significant difference relative to control cells (P < 0.05).

DISCUSSION

In the present study, we have shown broad effects of ArtB on the production of chemokines, proinflammatory cytokines, and adhesion molecules in distinct cell types. Secretion of the chemokines MCP-1 (CCL2) and IL-8 (CXCL8) was increased in all 3 tested cell lines, with MIP-1α (CCL3), MIP-1β (CCL4), and G-CSF also significantly increased in U937 cells. ArtB also upregulated the cytokines TNF-α and IL-6 in HBMEC and HCT-8 cells, but not in U937 cells, while ICAM-1 was upregulated in HCT-8 and U937 cells and VCAM-1 was upregulated in HBMEC. These findings regarding levels of expressed proteins were largely consistent with data obtained at the transcriptional level. However, there were some exceptions; TNF-α mRNA was upregulated in HCT-8 cells at 4 h without significant elevation in soluble TNF-α in supernatants at the same time point. This may be related to posttranslational modification necessary to generate soluble TNF-α, which was indeed detected at elevated levels at 24 h. Also, IL-1β transcripts were elevated in U937 cells, yet IL-1β protein was not produced at levels above controls at any time point, again suggesting possible alterations in posttranslational processing. Nevertheless, the overall net impact of the various changes induced by ArtB would be to promote and maintain a strong inflammatory response, with MCP-1 recruiting monocytes, memory T cells, and macrophages and IL-8 recruiting neutrophils to the site of infection or tissue damage. Neutrophil proliferation and release from the bone marrow would be promoted by G-CSF and then activated by MIP-1α and MIP-1β produced by macrophages. Inflammation would also be promoted by TNF-α and IL-6. These cytokine and chemokine responses are likely to have triggered the observed upregulation of ICAM-1 and VCAM-1, which would promote transmigration of leukocytes into underlying tissues.

Other AB5 toxins, most notably Ctx, LT, and Ptx, have potent immunomodulatory properties, including strong adjuvant effects (16, 17). Some of these effects are dependent on the ADP-ribosyltransferase activity of the respective A subunit, but significant immunomodulation effects, particularly in lymphocytes, can also be elicited either by mutant holotoxins devoid of catalytic activity or by purified B subunit pentamers (16–19). For CtxB and LTB, effects include altered antigen processing and presentation by macrophages, induction of cytokine secretion by monocytes, stimulation of proliferation of B and CD4⁺ T cells, and induction of apoptosis in CD8⁺ T cells (17). Such effects were not observed when cells were treated with LTB with a G33D mutation that abrogated the capacity to bind the ganglioside receptor GM1, suggesting that B subunit-GM1 interaction was required for the signaling events that result in immunomodulation (17). However, LTB derivatives with either H57A or H57S mutations retained high-affinity binding to GM1 and underwent normal uptake and intracellular trafficking, but were also defective in immunomodulation (20, 21). Thus, binding to GM1 alone is insufficient, and additional interactions between the GM1-bound B pentamer and other surface molecules are required for immunomodulation.

The capacity of AB5 toxin B subunits to elicit signaling responses in epithelial or endothelial cells, however, has received less attention. Treatment of T84 (intestinal epithelial) cells with Ctx or LT induced expression of several cytokines/chemokines, including IL-6, IL-10, IL-1α, and IL-1β, as well as low levels of IL-8. However, this was totally dependent on the ADP-ribosyltransferase activity of the respective A subunit, and catalytically inactive holotoxin mutants or isolated B subunits were inactive (22). LT-induced reduction in trans-epithelial resistance in polarized T84 cells also required functional holotoxin (23). In the present study, treatment of human colonic epithelial (HCT-8) cells with ArtB elicited significant IL-6, IL-8, TNF-α, and MCP-1 secretion, as well as ICAM-1 expression, at 24 h. Interestingly, at 4 h, expression of other adhesion molecules (E selectin, ICAM-3, PECAM-1, VCAM-1, and P selectin) was also significantly elevated. It has been established that both leukocyte chemoattractants (IL-8 and MCP-1) and activators (TNF-α and IL-6), as well as adhesion molecules, play essential roles in the initial recruitment of leukocytes to the site of infection (24). In particular, MCP-1 has been shown to play a regulatory role during innate immune responses to Salmonella infection (25). MCP-1^−/− mice show decreased survival after administration of purified Salmonella LPS, with accompanying elevated levels of TNF-α and IL-6 in the serum (25). Similarly, increased mortality was also observed in normal mice treated with anti-MCP-1 antibodies prior to endotoxin exposure, while administration of MCP-1 during endotoxin challenge protected the mice from death (26). Administration of MCP-1 to Salmonella-infected mice also results in enhanced macrophage activation and accumulation (27). Adhesion molecules are also critically involved in leukocyte-epithelial/endothelial cell interactions (28–32). The expression of ICAM-1 is strongly upregulated by inflammatory mediators and plays a key role in acquired immunity to salmonellosis (33). Thus, ArtB not only has the potential to significantly impact innate immune responses at the intestinal epithelium, the initial site of interaction between S. Typhi and the host, but also may have systemic effects.

HBMEC were particularly sensitive to ArtB, with significant elevations in IL-6, IL-8, and MCP-1 secretion at 1.25 μg/ml, the lowest dose tested, while VCAM-1 expression was significantly elevated at 2.5 μg/ml. VCAM-1 is an endothelial cell membrane glycoprotein, which can be upregulated by proinflammatory cytokines (e.g., TNF-α and IL-1), and has been implicated in leukocyte/endothelial cell interactions in inflammation (34). Therefore, ArtB may play a role in promoting endothelial cell damage and leukocyte extravasation and may induce further tissue or organ damage in the later stages of infection. To our knowledge, the effects of isolated B subunits of Ctx, LT, or Ptx on human endothelial cells have not been investigated. However, purified B subunits of Stx1 and Stx2 have been reported to induce secretion of von Willebrand factor by human endothelial cells and to contribute to thrombotic microangiopathy in mice (35). Interestingly, this appeared to involve activation of distinct signaling pathways by Stx1B and Stx2B, dependent upon protein kinase Cα and protein kinase A, respectively (36). Thus, Stx B subunits may play a direct role in the endothelial injury that is the hallmark of the systemic complications of STEC disease in humans.

At present, the role of ArtAB in the pathogenesis of salmonellosis remains obscure, and to date, no ADP-ribosyltransferase-dependent cytotoxic effects have been reported. However, it is known to be expressed in vivo, as antibodies to the A subunit have been detected in sera from patients with S. Typhi bacteremia (37). Spanò et al. (11) have reported that ArtAB (referred to as PltAB in that study) functions as a delivery vehicle for the catalytic subunit of cytolethal distending toxin (CdtB), whose cognate binding subunits (CdtA and CdtC) are missing in S. Typhi. This property required both the A and B subunits of ArtAB, but not A subunit catalytic activity. This might be explained if the interaction with CdtB occurred via ArtA, or a domain spanning the ArtA-ArtB interface, while the ArtB pentamer mediated membrane binding and translocation. The findings of the present study, however, raised the possibility that ArtB could also function in pathogenesis in its own right by triggering signaling responses that result in secretion of proinflammatory cytokines and chemokines and adhesion molecule expression, both in the intestinal epithelium and systemically. This finding may also have significance for infections caused by Yersinia spp., whose genomes encode homologues of ArtB but lack a cotranscribed cognate catalytic subunit.