Toll-Like Receptor 4 Mediates Tolerance in Macrophages Stimulated with Toxoplasma gondii-Derived Heat Shock Protein 70

ABSTRACT Peritoneal macrophages (PMs) from toll-like receptor 4 (TLR4)-deficient and wild-type (WT) mice were responsive to recombinant Toxoplasma gondii-derived heat shock protein 70 (rTgHSP70) and natural TgHSP70 (nTgHSP70) in NO release, but those from TLR2-, myeloid differentiation factor 88 (MyD88)-, and interleukin-1R-associated kinase 4 (IRAK4)-deficient mice were not. Polymyxin B did not inhibit PM activation by TgHSP70 and nTgHSP70 from WT and TLR4-deficient mice, while it inhibited PM activation by lipopolysaccharide. Pretreatment of PMs from WT but not from TLR4-deficient mice with rTgHSP70 resulted in suppression of NO release on restimulation with rTgHSP70. Similarly, pretreatment of PMs from WT but not TLR4-deficient mice with nTgHSP70 resulted in suppression of NO release on restimulation with nTgHSP70. Polymyxin B did not inhibit rTgHSP70- and nTgHSP70-induced tolerance of PMs from TLR4-deficient mice. Furthermore, PMs from WT mice increased suppressor of cytokine-signaling-1 (SOCS-1) expression after restimulation with rTgHSP70, while those from TLR4-deficient mice did not. Phosphorylation of JNK and I-κBα occurred in rTgHSP70-induced tolerance of PMs from TLR4-deficient mice, but not in that from WT mice. These data indicated that TgHSP70 signaling mechanisms were mediated by TLR2, MyD88, and IRAK4, but not by TLR4. On the other hand, signaling of TgHSP70-induced tolerance was mediated by TLR4, and the expression of SOCS-1 suppressed the TLR2 signaling pathway.

Members of the heat shock protein (HSP) family have been shown to have important functions as (i) intracellular detergents for aggregated and denatured molecules formed as a result of exposure of cells to physical stressors and (ii) molecular chaperones in peptide and protein transport between cell organelles (13,34,42,43). Of the HSP family members, HSP70 has been shown to be a major immunodominant antigen in bacterial and parasite infections, as well as the preferred target of humoral and cell-mediated immune responses to infection (12,30,49). We have previously demonstrated a potential role of human heat shock cognate protein 71 in the antigen presentation process of Toxoplasma gondii-infected cells to CD4 ϩ cytotoxic T lymphocytes in humans (46).
The gene for HSP70 of T. gondii (TgHSP70) has been cloned (26,48), and we have reported anti-TgHSP70 antibody formation (26,27) and anti-HSP70 autoantibody formation (5) in T. gondii-infected mice. T. gondii in gamma interferon (IFN-␥) knockout mice infected with T. gondii expressed TgHSP70 mRNA 2 days before death, a finding suggesting the role of TgHSP70 as a danger signal during lethal, acute T. gondii infection (24,47). Interestingly, TgHSP70 was shown to cause deterioration of the host defense by down-regulating NO release by peritoneal macrophages (PMs) in the T. gondii-infected host (24,47).
NO has been identified as an important signaling molecule involved in regulating a wide range of biological activities in the neural, vascular, and immune systems (23). NO and its metabolites control a number of host defense functions regulated by activated macrophages, including antimicrobial and tumoricidal activities, implicated in the pathogenesis of tissue damage associated with acute and chronic inflammation (4,19,21). T. gondii-infected inducible nitric oxide synthase (iNOS) knockout mice or iNOS inhibitor L-N G -monomethyl argininetreated wild-type (WT) mice, displayed increased parasite numbers in the brain (25,32), and they succumbed to acute infection from a high dose of T. gondii. Also, toll-like receptor 2 (TLR2)-and myeloid differentiation factor 88 (MyD88)deficient mice that fail to produce NO cannot survive, but TLR4-deficient and WT mice that produce NO can survive acute infection with high-dose T. gondii (25).
TLRs play a critical role in the innate immune responses in mammals (15,37). TLR4 and TLR2 are essential for recognition of lipopolysaccharide (LPS) and lipoteichoic acids, respectively (15,37). The current understanding of TLR2-and TLR4-associated signaling complexes is based on the signal transduction cascade shared by interleukin-1 receptor (IL-1R) and other TLRs where, upon specific ligand binding, the receptor associates with the intracellular adaptor protein MyD88, which recruits IL-1R-associated kinases (IRAKs) to the receptor complex (20,44). Among these IRAK proteins, IRAK4 appears to be critical for the transduction of IL-1R/ TLR signals (35,36).
Prior exposure to LPS induces a transient state of cellular hyporesponsiveness to subsequent stimulation with LPS known as endotoxin tolerance (10,45). This mechanism may protect the host from developing a shock syndrome caused by hyperactivation of monocytes and macrophages with persistent bac-teria and LPS. Animals pretreated with LPS showed reduced febrile responses and mortality rates after a second challenge with LPS. Some articles have reported that suppressor of cytokine-signaling 1 (SOCS-1) is a negative regulator of LPSinduced macrophage activation to protect the host from harmful overresponses to LPS (18,28). On the other hand, it has recently been postulated that suppressed IL-12 production by monocytes and dendritic cells associated with endotoxin tolerance may result in an inability to respond appropriately to secondary infections in survivors of sepsis (16).
This report for the first time demonstrates that TgHSP70induced NO release is mediated via the TLR2/MyD88 signal transduction pathway and that prior exposure to TgHSP70 induces a hyporesponse to subsequent stimulation with TgHSP70 by expression of SOCS-1 via TLR4.
Preparation of PMs. Peritoneal exudate cells (PECs) from TLR2-, TLR4-, MyD88-, and IRAK4-deficient and WT mice were harvested and added to 96-well plates at a density of 5 ϫ 10 5 cells/100 l/well and incubated for 2 h at 37°C in RPMI 1640 containing 10% fetal calf serum. Nonadherent cells were then removed by gentle washing (3,17). Adherent cells were collected with a cell lifter. The adherent cells were stained with fluorescein isothiocyanate-conjugated anti-mouse F4/80 antibody and then analyzed by FACScan (Becton Dickinson, Tokyo, Japan). Over 97% of these adherent cells were macrophages.
Determination of NO. Serial dilutions of rTgHSP70, rTgHSP30, or rSAG1 (2 g/ml); nTgHSP70 (2 g/ml); or LPS (30 ng/ml) (Escherichia coli O55-B5; Sigma, St. Louis, MO) were simultaneously added to the culture of 2 ϫ 10 5 PMs and were present for the entire cultivation time of 24 h. Polymyxin B (10 g/ml; 10 l; Sigma) or TgNCR C2 (10 l) was added to the culture with rTgHSP70, nTgHSP70, or LPS and was present for the entire cultivation time of 24 h posttreatment by Griess reaction (24).
Tolerance experiments. To induce tolerance of rTgHSP70, nTgHSP70, and LPS, PMs from mice were stimulated for 24 h in the presence of rTgHSP70 (2 g/ml), nTgHSP70 (2 g/ml), or LPS (30 ng/ml). Then, the cells were washed twice and restimulated for an additional 24 h with the same concentration of rTgHSP70, nTgHSP70, or LPS and assayed for NO release.
Western blotting. The expression levels of phosphorylated JNK and I-B were ascertained by Western blotting using monoclonal antibodies (MAbs) specific for phosphorylated JNK (anti-p-JNK MAb) (sc-6254; Santa Cruz Biotechnology, Inc., Santa Cruz, CA), I-B (anti-I-B MAb) (sc-1643; Santa Cruz Biotechnol-ogy, Inc.), or phosphorylated I-B (anti-p-I-B MAb) (sc-8404; Santa Cruz Biotechnology, Inc.). In brief, PM lysates were dissociated by boiling them in a sodium dodecyl sulfate (SDS) sample buffer and run under reducing conditions using 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (26). After electrophoresis, the separated proteins were electroblotted onto nitrocellulose membranes (Amersham International Plc., Buckinghamshire, England) as previously described. Blots were blocked with 10% milk in Tris-buffered saline containing Tween 20; probed with anti-p-JNK MAb, anti-I-B␣ MAb, or antip-I-B␣ MAb in Tris-buffered saline containing Tween 20 overnight; incubated with biotinylated rabbit anti-mouse immunoglobulin G antibody (Jackson Im-munoResearch Laboratories, West Grove, PA) at 1:2,000 for 1 h; and incubated with horseradish peroxidase-conjugated streptavidin (ZYMED Laboratory, San Francisco, CA) at 1:2,000 for 20 min. Protein bands were visualized with an ECL detection system (Amersham International Plc.) according to the manufacturer's specifications.
Statistical analysis. The significance of differences between groups was determined by Student's t test. A P value of Ͻ0.05 was taken as significant.

VOL. 73, 2005
TgHSP70-MEDIATED TOLERANCE VIA TLR4 4637 and rSAG1. rTgHSP70 induced the release of NO, and the expression of iNOS was dependent on TLR2 and MyD88, but not on TLR4 ( Fig. 2A and C). On the other hand, PMs from TLR2-deficient and WT mice, but not from TLR4-and MyD88-deficient mice, responded to LPS challenge with NO production and iNOS mRNA expression ( Fig. 2B and D). To determine whether the contaminating LPS in rTgHSP70 was responsible for its NO production, we used polymyxin B and TgNCR C2 in cultures of PMs stimulated with rTgHSP70, nTgHSP70, and LPS ( Fig. 2E to G). Polymyxin B did not inhibit rTgHSP70 activation of PMs, in contrast to its effect on LPS, and TgNCR C2 inhibited rTgHSP70 activation of PMs. TLR4-dependent NO tolerance induced by TgHSP70. PMs from WT mice released NO by primary stimulation with LPS. Then, the cells were washed and restimulated with the same concentration of LPS, and a more than 90% reduction in the level of NO release was observed ( Fig. 3A and Table 1). Similarly, PMs from WT mice released NO by primary stimulation with rTgHSP70, but the level of NO release was reduced by more than 90% upon secondary stimulation with the same concentration of rTgHSP70. Thus, like LPS tolerance, a secondary rTgHSP70 challenge induced tolerance in PMs from WT mice. On the other hand, when PMs from TLR4-deficient mice were pretreated with rTgHSP70, no reduction was observed in the level of NO release upon secondary stimulation with rTgHSP70. Furthermore, PMs from WT mice released NO by primary stimulation with nTgHSP70, but the level of NO release was reduced by more than 90% upon secondary stimulation with the same concentration of nTgHSP70 (Fig.  3B). On the other hand, when PMs from TLR4-deficient mice were pretreated with nTgHSP70, no reduction was observed in the level of NO release upon secondary stimulation with nT-gHSP70. Thus, TLR4-mediated signaling activation by secondary TgHSP70 challenge induces tolerance of NO release in mouse PMs by primary TgHSP70 challenge through TLR2.
To determine whether the contaminating LPS in rTgHSP70 was responsible for its tolerance, we used polymyxin B in cultures of PMs stimulated with rTgHSP70, nTgHSP70, and LPS. Polymyxin B did not inhibit rTgHSP70 and nTgHSP70 activation of PMs, in contrast to its effect on LPS (Fig. 3A and B).

Cross-reactive NO tolerance by TgHSP70 and LPS.
We next analyzed whether rTgHSP70 treatment induced cross-reactive tolerance to LPS and vice versa. rTgHSP70 pretreatment resulted in a significant decrease in NO release in response to LPS, and LPS pretreatment resulted in a significant decrease in NO release in response to rTgHSP70 in PMs from WT mice (Fig. 3A).
NO release from PMs of TLR2-deficient mice stimulated by rTgHSP70/LPS (primary stimulation/secondary stimulation) was induced only by the secondary stimulation of LPS via TLR4. Primary stimulation by rTgHSP70 had no effect on NO release by TLR2-deficient mice. NO release from PMs of TLR4-deficient mice stimulated by LPS/rTgHSP70 was induced only by the secondary stimulation of rTgHSP70 via TLR2. Primary stimulation by LPS did not affect NO release by TLR4-deficient mice. As PMs were washed after primary stimulation, NO release was not observed from PMs either from TLR2-deficient mice receiving LPS/rTgHSP70 stimulation or from TLR4-deficient mice receiving rTgHSP70/LPS stimulation. Polymyxin B did not affect rTgHSP70 tolerance of PMs, in contrast to its effect on LPS.
These results indicate that rTgHSP70 treatment cross-reactively induced tolerance to LPS-induced NO production, and vice versa.
TgHSP70-stimulated TNF-␣ expression of PMs via TLR2. PMs from WT mice responded to produce TNF-␣ expression by primary rTgHSP70 challenge (Fig. 4). PMs from TLR4deficient and WT mice, but not from TLR2-and MyD88deficient mice, responded to produce TNF-␣ expression by primary rTgHSP70 challenge. To determine whether the contaminating LPS in rTgHSP70 was responsible for its TNF-␣ expression, polymyxin B was added to cultures of PMs stimulated with rTgHSP70. Polymyxin B did not inhibit the rT-gHSP70 activation of PMs.
Ϫ ND a The table summarizes the data in Fig. 2 and 3. b ϩ, NO synthesis positive; Ϫ, NO synthesis negative. c ϩ, occurrence of tolerance induction; Ϫ, no occurrence of tolerance induction. ND, not detected. In these groups, NO synthesis did not occur by primary stimulation with TgHSP70 and LPS.

TLR4-dependent TNF-␣ expression tolerance induced by TgHSP70.
PMs from WT mice expressed TNF-␣ by primary stimulation with rTgHSP70, but the level of TNF-␣ expression decreased upon secondary stimulation with the same concentration of rTgHSP70 (Fig. 4B). On the other hand, when PMs from TLR4-deficient mice were pretreated with rTgHSP70, no reduction was observed in the level of TNF-␣ expression upon secondary stimulation with rTgHSP70. Thus, TLR4-mediated signaling activation by secondary TgHSP70 challenge induces the tolerance of TNF-␣ expression of PMs by primary TgHSP70 challenge through TLR2.
To determine whether the contaminating LPS in rTgHSP70 was responsible for its TNF-␣ expression tolerance, polymyxin B was added to cultures of PMs stimulated with rTgHSP70. Polymyxin B did not inhibit the rTgHSP70 activation of PMs.
SOCS-1-mediated tolerance by TgHSP70 through TLR4. Next, the role of SOCS-1 in TgHSP70-mediated signaling activation for NO release was examined in the PMs from mice. A FIG. 5. Suppression of TgHSP70 response by SOCS-1 in mouse PMs. PMs from TLR4-, TLR2-, and MyD88-deficient and WT mice were stimulated with rTgHSP70 (2 g/ml) or LPS (30 ng/ml). Ten micrograms per milliliter of polymyxin B was added to the culture with rTgHSP70 or LPS and was present for the entire cultivation time of 12 h posttreatment; then, the PMs were restimulated after being washed for an additional 8 h with the same concentration of polymyxin B-treated rTgHSP70 and LPS. Total RNA was isolated, and the SOCS-1 and IFN-␤ mRNA expression levels from PMs from mice were investigated by RT-PCR. The data are representative of three independent experiments. stim, stimulation.

VOL. 73, 2005
TgHSP70-MEDIATED TOLERANCE VIA TLR4 4639 high level of SOCS-1 expression was detected in the PMs from TLR2-and MyD88-deficient and WT mice after rTgHSP70/ rTgHSP70, rTgHSP70/LPS, LPS/rTgHSP70, and LPS/LPS stimulation (Fig. 5). On the other hand, SOCS-1 was not expressed at a high level in PMs from TLR4-deficient mice after rTgHSP70/rTgHSP70, rTgHSP70/LPS, LPS/rTgHSP70, and LPS/LPS stimulation, suggesting that a TLR4-dependent signal was required for SOCS-1 expression and TgHSP70-induced tolerance. Furthermore, a high level of IFN-␤ expression was detected in TgHSP70-pretreated PMs from TLR2and MyD88-deficient and WT mice after rTgHSP70/rT-gHSP70, rTgHSP70/LPS, LPS/rTgHSP70, and LPS/LPS stimulation, whereas IFN-␤ was not expressed at a high level in PMs from TLR4-deficient mice. Tolerance was shown to be maintained for up to 24 h (data not shown). Thus, these data indicated that SOCS-1 and IFN-␤ induced TgHSP70 tolerance via TLR4.

Lack of phosphorylation of JNK and I-B␣ in rTgHSP70induced tolerance.
To examine whether degradation, as well as phosphorylation, of JNK and I-B␣ occurred in rTgHSP70induced tolerance, we performed immunoblotting analysis of JNK and I-B␣ in PMs after primary and secondary stimulation with rTgHSP70 (Fig. 6). A high level of phosphorylation of JNK and I-B␣ was detected in the PMs from TLR4-deficient and WT mice after primary stimulation with rTgHSP70. In contrast, the levels of phosphorylation of JNK and I-B␣ decreased in the TgHSP70-pretreated PMs from WT mice after primary and secondary stimulation with rTgHSP70. On the other hand, when PMs from TLR4-deficient mice were pretreated with rTgHSP70, no reductions in the levels of phosphorylation of JNK and I-B␣ were observed upon secondary stimulation with rTgHSP70. These results suggest that the absence of phosphorylation of JNK and I-B␣ contributes to the downregulated NO release and TNF-␣ mRNA expression in rTgHSP70-tolerant cells.

DISCUSSION
In this study, we showed that TgHSP70-induced NO release is dependent on TLR2, MyD88, and IRAK4, but not on TLR4 (Fig. 7). Furthermore, TLR4-mediated signaling activation by TgHSP70 challenge induced the tolerance of NO release in mouse PMs by TLR2-mediated TgHSP70 challenge. On the other hand, rhHSP70-1, a highly purified rhHSP70, did not induce TNF-␣ release from murine macrophages, and the TNF-␣-inducing activity of rhHSP70-2, a less purified rhHSP70, was entirely due to the contaminating LPS (11,40). In the present study, however, polymyxin B did not inhibit rTgHSP70 and nTgHSP70 activation of PMs. TgHSP70-induced NO release was dependent on TLR2, whereas LPSinduced NO release was dependent on TLR4.
There was an impression that greater NO release occurred in WT mice than in TLR4-deficient mice when stimulated with TgHSP70 and also that greater NO release occurred in WT mice than in TLR2-deficient mice when stimulated with LPS. The precise mechanisms to cause such lower levels of NO release by TLR4-and TLR2-deficient mice compared with WT mice are not known.
Recent evidence supports the hypothesis that HSP activates the innate immune system via TLRs, the sensors of innate immunity. TLRs are transmembrane proteins with an extracellular domain consisting of leucine-rich repeats involved in recognition of microbial components. Asea  Aosai et al. reported that TgHSP70 induced spleen B-cell proliferation in TLR2-and MyD88-deficient mice, but not in TLR4-deficient mice, indicating the involvement of TLR4 in TgHSP70-induced proliferative responses of B cells (1). These data and the present study indicated that TgHSP70-induced tolerance in NO production of PMs by secondary challenge, as well as TgHSP70-induced proliferation of B cells, was mediated via TLR4, but TgHSP70-induced NO production in PMs by primary challenge was mediated via TLR2 and MyD88.
Previous studies have reported that production of NO was induced by LPS even in LPS-tolerant cells (9,22). That result seems contradictory to our findings and those of others (29,31); however, in those studies, macrophages had been pretreated with a low concentration of LPS (Ͻ20 ng/ml). Several articles had previously reported that induction of LPS tolerance was not severe when stimulated with a low concentration of LPS (29,31). In the present study, we used a 30-ng/ml concentration of LPS for LPS tolerance by PMs.
In our study, PMs from WT mice expressed SOCS-1 and IFN-␤ after stimulation with rTgHSP70, while those from TLR4-deficient mice did not. LPS has been demonstrated to induce the expression of IFN-␤ in macrophages (33). The induction of SOCS-1 mRNA by LPS requires autocrine/paracrine factors that include IFN-␤ (7). In this report, we suggested that primary TgHSP70 stimulation initiated the expression of SOCS-1, which could not inhibit NO production by primary TgHSP70 stimulation. SOCS-1 activated by primary TgHSP70 stimulation via TLR4 inhibits the NO synthesis signaling activated by secondary stimulation of TgHSP70 via TLR2. Kinjyo et al. reported that SOCS-1 is a negative regulator of TLR signaling, suggesting that SOCS-1 might be involved in the induction of endotoxin tolerance (18). SOCS-1 overexpression suppressed LPS-induced NF-B transcriptional activity in RAW cells, as well as in fibroblasts expressing both TLR4 and MD2 (18). Therefore, the requirement for SOCS-1 suggests that the mechanism for LPS signal suppression by SOCS-1 is different from that of the cytokine-dependent JAK/ STAT pathway. SOCS-1 also inhibited NF-B activation by IRAKs and TNF receptor-associated factor 6 (TRAF6) but not by I-B kinase-␤ (IKK-␤) when these molecules were overexpressed in fibroblasts, indicating that the inhibition by SOCS-1 occurs downstream of TRAF6 and upstream of IKK-␤ (18). Overall, these results indicated that the TgHSP70-induced NO release was mediated via the TLR2/MyD88 signal transduction pathway and that prior exposure to TgHSP70 induced a tolerance of NO production for subsequent stimulation with TgHSP70 via TLR4.