Toll-Like Receptor 2 Recognizes Orientia tsutsugamushi and Increases Susceptibility to Murine Experimental Scrub Typhus

Abstract

Scrub typhus is a potentially lethal infection that is caused by the obligate intracellular bacterium Orientia tsutsugamushi The roles of Toll-like receptor 2 (TLR2) and TLR4 in innate recognition of O. tsutsugamushi have not been elucidated. By overexpression of TLR2 or TLR4 in HEK293 cells, we demonstrated that TLR2, but not TLR4, recognizes heat-stable compounds of O. tsutsugamushi that were sensitive to treatment with sodium hydroxide, hydrogen peroxide, and proteinase K. TLR2 was required for the secretion of tumor necrosis factor alpha (TNF-α) and interleukin-6 (IL-6) by dendritic cells. In an intradermal mouse infection model, TLR2-deficient mice did not show impaired control of bacterial growth or reduced survival. Moreover, after intraperitoneal infection, TLR2-deficient mice were even more resistant to lethal infection than C57BL/6 wild-type mice, which showed stronger symptoms and lower survival rates during the convalescent phase. Compared to the time of reduction of bacterial loads in TLR2-deficient mice, the reduction of bacterial loads in infected organs was accelerated in wild-type mice. The higher mortality of wild-type mice was associated with increased concentrations of serum alkaline phosphatase but not aspartate aminotransferase. The transcription of mRNA for TNF-α and IL-6 decreased more rapidly in peritoneum samples from wild-type mice than in those from TLR2-deficient mice and was therefore not a correlate of increased susceptibility. Thus, although TLR2 is an important mediator of the early inflammatory response, it is dispensable for protective immunity against O. tsutsugamushi Increased susceptibility to O. tsutsugamushi infection in TLR2-competent mice rather suggests a TLR2-related immunopathologic effect.

FIG 1

O. tsutsugamushi is recognized by TLR2 but not TLR4. HEK293 cells were transiently transfected with plasmids expressing human TLR2 or TLR4 and were infected with live O. tsutsugamushi (O. tsu live) or stimulated with heat (56°C)-inactivated O. tsutsugamushi (O. tsu 56°C). S. Minnesota LPS, Pam3C, and TNF-α were used as controls. After 18 h, IL-8 concentrations from cell culture supernatants were measured by a sandwich cytokine ELISA. (A) TLR2-transfected cells responded to live O. tsutsugamushi, heat (56°C)-inactivated O. tsutsugamushi, and Pam3C in a dose-dependent fashion. (B) TLR4-transfected cells responded to S. Minnesota LPS only. Shown are combined results from two independent experiments (2 to 3 replicates, mean ± SEM). med., medium. ***, P < 0.001 by one-way ANOVA; *, P < 0.05 by one-way ANOVA; ns, no significant difference.

FIG 2

The TLR2 ligand of O. tsutsugamushi is sensitive to proteinase K, H₂O₂, and NaOH. (A to D) HEK293 cells transfected with human TLR2 were stimulated with heat-inactivated and crude lysates of O. tsutsugamushi pretreated with the indicated compound. IL-8 concentrations were measured 18 h later by ELISA. Representative results from one of two independent experiments are shown (n = 4 replicates, mean ± SEM). ***, P < 0.001 by one way-ANOVA; ns, no significant difference.

FIG 3

Production of IL-6 and TNF-α by BMDCs in response to O. tsutsugamushi requires TLR2. (A, B) BMDCs (2 × 10⁵) were infected or stimulated with O. tsutsugamushi or treated with LPS (1 μg/ml), Pam3C (1 μM), or controls. At 24 h p.i., the levels of IL-6 (A) and TNF-α (B) in the supernatants were measured by ELISA (n = 4 replicates per experiment). Shown are pooled results from two independent experiments (mean ± SEM). *, P < 0.05 by one way-ANOVA; **, P < 0.01 by one way-ANOVA; ***, P < 0.001 by one way-ANOVA.

FIG 4

tlr2^−/− mice are able to restrict bacterial growth after dermal infection with O. tsutsugamushi. C57BL/6 and tlr2^−/− mice were infected with 5,000 SFU of O. tsutsugamushi, which was injected into the right hind footpad. Bacterial loads in popliteal lymph node, spleen, and lung were measured by traD qPCR at the indicated time points. Pooled results from two independent experiments (n = 6, mean ± SEM) are shown.

FIG 5

TLR2-competent mice show higher levels of susceptibility to a lethal outcome, which is associated with a more efficient reduction of bacterial loads and increased serum ALP levels but not prolonged inflammation. C57BL/6 and tlr2^−/− mice were infected i.p. with 5,000 SFU of O. tsutsugamushi. (A) Clinical scores were determined at the indicated time points p.i. (pooled results from two independent experiments, n = 9, mean ± SEM). (B) Kaplan-Meier curves show the survival of i.p. infected C57BL/6 and tlr2^−/− mice. Data from three independent experiments were pooled (n = 15). (C) AST and ALP concentrations (n = 5 to 8) in serum samples obtained at the indicated times were measured using Reflotron colorimetric analysis (pooled results from two independent experiments, n = 5 to 8, mean ± SEM). (D) Serum concentrations of IL-6 (left) and TNF-α (right) were measured by use of a cytokine bead array on days 7, 14, and 21 p.i. (n = 3, mean ± SEM). (E) Bacterial loads in the lung, spleen, and peritoneum were measured by traD qPCR on days 12, 15, and 18 p.i. (n = 7 to 8, data pooled from two independent experiments, mean ± SEM). (F) RNA was extracted from peritoneum samples obtained on days 12, 15, and 18 p.i., and the relative expression of il-6 and tnf-a mRNA was measured by quantitative real-time PCR (n = 4, means ± SEMs). *, P < 0.05 by two way-ANOVA (A and C to E); **, P < 0.01 by two way-ANOVA (A and C to E); ***, P < 0.001 by two way-ANOVA (A and C to E); *, P < 0.05 by log-rank (Mantel-Cox) test (B).