Context-dependent activation kinetics elicited by soluble versus outer membrane vesicle-associated heat-labile enterotoxin

Abstract

Enterotoxigenic Escherichia coli (ETEC) is the leading cause of traveler's diarrhea and children's diarrhea worldwide. Among its virulence factors, ETEC produces heat-labile enterotoxin (LT). Most secreted LT is associated with outer membrane vesicles that are rich in lipopolysaccharide. The majority of prior studies have focused on soluble LT purified from ETEC periplasm. We investigated the hypothesis that the extracellular vesicle context of toxin presentation might be important in eliciting immune responses. We compared the polarized epithelial cell responses to apically applied soluble LT and LT-containing vesicles (LT(+) vesicles) as well as controls using a catalytically inactive mutant of LT and vesicles lacking LT. Although vesicle treatments with no or catalytically inactive LT induced a modest amount of interleukin-6 (IL-6), samples containing catalytically active LT elicited higher levels. A combination of soluble LT and LT-deficient vesicles induced significantly higher IL-6 levels than either LT or LT(+) vesicles alone. The responses to LT(+) vesicles were found to be independent of the canonical LT pathway, because the inhibition of cyclic AMP response element (CRE)-binding protein (CREB) phosphorylation did not lead to a decrease in cytokine gene expression levels. Furthermore, soluble LT caused earlier phosphorylation of CREB and activation of CRE compared with LT(+) vesicles. Soluble LT also led to the activation of activator protein 1, whereas LT(+) vesicle IL-6 responses appeared to be mediated by NF-κB. In summary, the results demonstrate that soluble LT and vesicle-bound LT elicit ultimately similar cytokine responses through distinct different activation pathways.

Fig. 1.

Treatments containing catalytically active LT elicit high levels of TNF-α and IL-6 expression in polarized intestinal epithelial cells, and induction is independent of flagellin. (A and B) Polarized T84 cells were treated with 200 pM or 1 nM standardized amounts (see Materials and Methods) of LT, LT⁺ OMVs, ΔLT OMVs, LT/ΔLT OMVs, S63K LT, or S63K OMVs as indicated, and the TNF-α (A) and IL-6 (B) gene expression levels were measured at 6 h using RT-PCR. (C) Polarized T84 cells were incubated with samples standardized to 1 nM LT for 10 h, and IL-6 levels were measured in supernatants collected from the apical compartment using ELISA. (D) Flagellin was detected in preparations of LT⁺ OMVs, ΔLT OMVs, LT/ΔLT OMVs, and S63K OMVs by immunoblotting with anti-FliC. The preparations were concentrated to allow for detection of flagellin, but the concentrations remained proportional to those used in the treatments described above.

Fig. 2.

TNF-α and IL-6 induction in response to LT⁺ OMVs is independent of CREB phosphorylation. (A and B) Polarized T84 cells were pretreated with 100 μM Rp-camps or 10 μM H89 for 1 h and incubated with samples standardized to 200 pM LT. After 6 h, the levels of TNF-α (A) and IL-6 (B) gene induction were evaluated using RT-PCR. Asterisks indicate significant differences from the respective untreated sample: *, P < 0.05; **, P < 0.01; ***, P < 0.005. (C) Nuclear extracts were prepared from polarized T84 cells incubated with samples standardized to 1 nM LT for 4 h (upper panel) or 6 h (lower panel), and immunoblotting was performed to determine the phosphorylation status of CREB. TBP was used as a loading control. The image shown is representative of three independent experiments. (D) Densitometry analysis of three independent immunoblots for each time point. Asterisks indicate significant differences from the respective mock control at the corresponding time: *, P < 0.05; **, P < 0.01; ***, P < 0.005. (E) HEK293T cells were cotransfected with a firefly luciferase reporter fused to a CRE promoter and a Renilla luciferase reporter fused to a constitutively active promoter. Transfected cells were then incubated with samples standardized to 1 nM LT for 4 h and 6 h, and the firefly luciferase activity was measured and normalized to Renilla luciferase values. Asterisks indicate significant differences from the respective mock control at the corresponding time: *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Fig. 3.

MEK1/2 and p38 play similar roles in the induction of TNF-α in response to both soluble LT and LT⁺ OMVs but do not play a role in IL-6 induction. Polarized T84 cells were pretreated for 1 h with inhibitors of MEK1/2 (10 μM PD98059) or p38 (10 μM SB203580) before being incubated with samples standardized to 200 pM LT for 6 h. (A and B) TNF-α levels were measured using RT-PCR after pretreatment with inhibitors of MEK1/2 (A) and p38 (B). (C and D) IL-6 levels were measured using RT-PCR after pretreatment with inhibitors of MEK1/2 (C) and p38 (D). Asterisks indicate significant differences from the corresponding untreated sample: *, P < 0.05; **, P < 0.01.

Fig. 4.

AP-1 plays a role in the induction of TNF-α and IL-6 in response to soluble LT but not LT⁺ OMVs. (A and B) Polarized T84 cells were pretreated for 1 with a JNK inhibitor (10 μM SP600125), followed by incubation with samples standardized to 200 pM LT for 6 h. The levels of TNF-α (A) and IL-6 (B) were measured using RT-PCR. Asterisks indicate significant differences from the corresponding untreated sample: *, P < 0.05; **, P < 0.01; ***, P < 0.005. (C) Polarized T84 cells were incubated with samples standardized to 1 nM LT for 6 h, and nuclear extracts were prepared. Nuclear protein (5 μg) was incubated with a radiolabeled oligonucleotide that corresponded to the DNA-binding region of activated AP-1, and an EMSA was performed to visualize binding. Densitometric values of the shifted DNA band are shown.

Fig. 5.

OMVs but not soluble LT induce IL-6 through NF-κB. Polarized T84 cells were pretreated for 30 min with an NF-κB inhibitor (100 μM PDTC), followed by incubation with samples standardized to 200 pM LT for 6 h. The levels of TNF-α (A) and IL-6 (B) were measured using RT-PCR. Asterisks indicate significant differences from the corresponding untreated sample: *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Fig. 6.

Overview of the different pathways through which soluble LT and LT⁺ OMVs elicit TNF-α and IL-6 responses in human intestinal epithelial cells. See Discussion for details.