Murine TOLL-like Receptor 4 Confers Lipopolysaccharide Responsiveness as Determined by Activation of NFκB and Expression of the Inducible Cyclooxygenase*

Genetic evidence indicating that TOLL-like receptor 4 (Tlr4) is the lipopolysaccharide (LPS) receptor in mice was reported. However, biochemical evidence that murine Tlr4 confers LPS responsiveness has not been convincingly demonstrated. Inducible cyclooxygenase (COX-2) is selectively expressed in LPS-stimulated macrophages in part mediated through the activation of NFκB. Thus, we determined whether murine Tlr4 confers LPS responsiveness as evaluated by the activation of NFκB and COX-2 expression. Transfection of a murine macrophage-like cell line (RAW264.7) with the constitutively active form (ΔTlr4) of Tlr4 is sufficient to activate NFκB and COX-2 expression. However, the truncated form (ΔTlr4(P712H)) of the missense mutant Tlr4(P712H) found in LPS-hyporesponsive mouse strain (C3H/HeJ) inhibits LPS-induced NFκB activation and COX-2 expression. The inability of ΔTlr4(P712H) to activate NFκB and induce COX-2 expression is rescued by a constitutively active adapter protein myeloid differentiation factor 88 (MyD88), which interacts directly with the cytoplasmic domain of Tlr proteins. Furthermore, MyD88 is co-immunoprecipitated with the wild-type ΔTlr4 but not with the ΔTlr4(P712H) mutant. Together, these results indicate that Tlr4 confers LPS responsiveness in RAW264.7 cells and suggest that hyporesponsiveness of C3H/HeJ mice to LPS is attributed to the disruption of Tlr4-mediated signaling pathways that results from the inability of the mutant Tlr4(P712H) to interact with MyD88.

Genetic evidence indicating that TOLL-like receptor 4 (Tlr4) is the lipopolysaccharide (LPS) receptor in mice was reported. However, biochemical evidence that murine Tlr4 confers LPS responsiveness has not been convincingly demonstrated. Inducible cyclooxygenase (COX-2) is selectively expressed in LPS-stimulated macrophages in part mediated through the activation of NFB. Thus, we determined whether murine Tlr4 confers LPS responsiveness as evaluated by the activation of NFB and COX-2 expression. Transfection of a murine macrophage-like cell line (RAW264.7) with the constitutively active form (⌬Tlr4) of Tlr4 is sufficient to activate NFB and COX-2 expression. However, the truncated form (⌬Tlr4(P712H)) of the missense mutant Tlr4(P712H) found in LPS-hyporesponsive mouse strain (C3H/HeJ) inhibits LPS-induced NFB activation and COX-2 expression. The inability of ⌬Tlr4(P712H) to activate NFB and induce COX-2 expression is rescued by a constitutively active adapter protein myeloid differentiation factor 88 (MyD88), which interacts directly with the cytoplasmic domain of Tlr proteins. Furthermore, MyD88 is co-immunoprecipitated with the wild-type ⌬Tlr4 but not with the ⌬Tlr4(P712H) mutant. Together, these results indicate that Tlr4 confers LPS responsiveness in RAW264.7 cells and suggest that hyporesponsiveness of C3H/HeJ mice to LPS is attributed to the disruption of Tlr4-mediated signaling pathways that results from the inability of the mutant Tlr4(P712H) to interact with MyD88.
The pathogenesis of Gram-negative septic shock is presumed to be because of excess stimulation of host cells by bacterial lipopolysaccharide (LPS) 1 endotoxin (1)(2)(3). Such stimulation leads to the expression and release of a plethora of pro-inflammatory marker gene products and lipid mediators, which in turn can initiate a chain of events leading to systemic toxicity (4,5). However, the initial recognition of LPS by cells of the innate immune system is required to defend the host from a Gram-negative infection before it becomes widely disseminated (6,7).
Identifying the downstream signaling pathways derived from LPS stimulation is of fundamental importance to understanding the cellular mechanism of Gram-negative septic shock. CD14, a glycosylphosphatidylinositol-linked membrane protein that is widely expressed in mononuclear cells, is considered a high affinity receptor for LPS (8,9). However, CD14 lacks a cytoplasmic domain, and there has been a puzzling question as to how CD14 transmits extracellular signals into downstream cytoplasmic signaling pathways. Recently, it was demonstrated that human TOLL-like receptor-2 (Tlr2) mediates the LPS-induced cellular signaling pathway (10,11). Human Tlr4 can constitutively activate NFB but fails to confer LPS responsiveness (10). However, results from studies using Tlr2-or Tlr4-deficient mice and from in vitro transfection studies suggest that Tlr4 recognizes LPS, but Tlr2 recognizes Granpositive bacterial cell wall components (41,42). The human TOLL protein is a transmembrane protein with an extracellular domain consisting of leucine-rich repeats and a cytoplasmic domain homologous to that of the IL-1 receptor (12).
Genetic evidence indicating that murine Tlr4 is the LPS receptor was demonstrated using two mouse strains (C3H/HeJ and C57BL/10ScCr), which are hyporesponsive to LPS (13). The former strain has Tlr4 with a missense mutation to replace proline with histidine at position 712, whereas the latter strain is homozygous for a null mutation of Tlr4. This genetic evidence is confirmed by another independent investigation (14). It is also demonstrated that the activation of NFB and the expression of certain NFB-induced gene products in LPSstimulated macrophages derived from the C3H/HeJ mouse strain are impaired (15,16). In addition, the overexpression of mutant Tlr4 derived from C3H/HeJ mice into human dermal endothelial cells results in the inhibition of LPS-induced NFB activation (17). However, biochemical evidence that murine Tlr4 indeed confers LPS responsiveness has not been conclusively demonstrated.
Activation of TOLL proteins and IL-1 receptor induces recruitment of the adapter molecule, myeloid differentiation factor 88 (MyD88) (18 -20), which in turn leads to the activation of NFB and the expression of NFB-induced gene products (10,11,21). Results from our previous studies indicate that LPS induces selective expression of the mitogen-induced cyclooxygenase (COX-2) in murine macrophages (22). LPS activates NFB through TOLL-like receptors in macrophages (23)(24)(25)(26). However, the role of NFB in LPS-induced COX-2 expression in macrophages is not clearly established; there were contradicting reports regarding the role of NFB in LPS-induced COX-2 expression in macrophages (27,28). COX-2 is shown to be overexpressed in sites of inflammation and in tissues of many types of tumors (29 -33). Elucidating the signaling pathways is the key to understanding why COX-2 is overexpressed in such pathological states and can provide crucial information for identifying the potential targets for pharmacological and dietary modulation.
In this study, we have addressed three important issues in elucidating LPS-stimulated signaling pathways in murine macrophages. First, we determined whether the activation of NFB is sufficient and required for LPS-induced COX-2 expression. Second, we determined whether the activation of Tlr4 confers LPS responsiveness as evaluated by the activation of NFB and the expression of COX-2. Third, we determined why the Tlr4 mutant with a missense mutation at position 712 fails to transmit the LPS-induced signal to downstream signaling pathways.

EXPERIMENTAL PROCEDURES
Cell Culture-RAW264.7 cells (murine macrophage-like cell line, ATCC TIB-71) were cultured in LPS-free Dulbecco's modified Eagle's medium containing 10% (v/v) heat-inactivated fetal bovine serum (Intergen) and 10 units/ml penicillin and 100 g/ml streptomycin (Life Technologies, Inc.) at 37°C in a 5% CO 2 air environment. Human embryonic kidney cells (293T cells) were provided by Sam Lee (Beth Israel Hospital, Boston, MA) and cultured in the same medium used for the RAW264.7 cells.
Expression plasmids for the wild-type NFB-inducing kinase (NIK),  Luciferase Reporter Gene Array-RAW264.7 cells were plated in sixwell plates (5 ϫ 10 5 cells/well) and transfected with total 5 g of DNA plasmids including HSP-70-␤-galactosidase plasmid as an internal control using SuperFect transfect reagent (Qiagen) according to the manufacturer's instruction. Relative luciferase activity was determined by normalization with ␤-galactosidase activity as described in our previous study (36).

LPS-induced Expression of COX-2 Is Suppressed by Inhibition of NFB in RAW264.7 Cells-
To determine whether LPS-induced expression of COX-2 is mediated through the activation of NFB, we investigated whether or not the inhibition of LPSinduced activation of NFB leads to the suppression of COX-2 expression. Cells were co-transfected with the luciferase reporter plasmid for NFB or COX-2 promoter and with an expression plasmid containing a dominant-negative mutant of NIK or IB␣ cDNA. The results show that LPS-induced COX-2 expression is significantly inhibited by the co-transfection of cells with a dominant-negative mutant of NIK or IB␣ (Fig. 2). These results indicate that the activation of NFB is required for the full expression of COX-2 in LPS-stimulated RAW264.7 cells.

Activation of NFB by NIK Leads to Expression of COX-2, and This Expression Is Suppressed by a Dominant-negative
Mutant of IB␣ in RAW264.7 Cells-If direct activation of NFB without involvement of LPS leads to the expression of COX-2, this result would further support that NFB is required for the full expression of COX-2 induced by LPS. The results show that direct activation of NFB by transfection of cells with the wild-type NIK plasmid leads to the expression of COX-2. Furthermore, the co-transfection of cells with NIK and a dominant-negative mutant of IB␣ plasmid significantly suppressed the NIK-induced COX-2 expression (Fig. 3). Together, these results indicate that the activation of NFB is required for the full expression of COX-2 in LPS-stimulated macrophages. There were two conflicting reports regarding the role of NFB in LPS-induced COX-2 expression in RAW264.7 cells. The requirement for NFB was demonstrated in our previous study using the pharmacological inhibitors of IB␣ degradation or nuclear translocation of NFB (27). However, the results from another study using the luciferase-reporter gene assay showed that NFB may not be required for the full expression of COX-2 in LPS-stimulated RAW264.7 cells (28).
It is not clear what causes the discrepancy between these two studies (27,28) in requirements of NFB for the full expression of COX-2. We used the murine COX-2 luciferase construct containing 3.2-kilobase upstream promoter sequences, whereas the COX-2 construct used in another study showing the results that were different from our study contains only 700-base pair upstream promoter sequences (28). Although the B binding site in the murine COX-2 promoter is located within Ϫ400 kilobases, it is possible that other enhancer elements located further upstream of the 5Ј-flanking region of COX-2 gene are required for the full expression of COX-2 in LPSstimulated RAW264.7 cells.
The Constitutively Active Tlr4(⌬Tlr4) but Not the Truncated Mutant (⌬Tlr4(P712H)) of Tlr4 Found in LPS-hyporesponsive Mouse Strain (C3H/HeJ) Activates NFB and Induces COX-2 Expression in RAW264.7 Cells-To determine whether or not the activation of Tlr4 is sufficient to activate NFB and induce the expression of COX-2, cells were co-transfected with the truncated mouse Tlr4 construct (⌬Tlr4) lacking the leucine-rich repeat extracellular domain and the NFB-or COX-2-luciferase reporter gene construct. The gain of function by the truncated TOLL proteins lacking the extracellular domain was demonstrated both in Drosophila and in human cells (21,39). The results show that transfection of cells with the constitutively active ⌬Tlr4 leads to the activation of NFB and COX-2 expression (Fig. 4). The same truncated ⌬Tlr4(P712H) with the missense mutation at position 712 is unable to activate NFB and COX-2 expression. Thus, these results indicate that the activation of Tlr4 is sufficient to activate NFB and induce COX-2 expression in RAW264.7 cells.
Kirschning et al. (10) showed that human Tlr2, when cotransfected with CD14 into 293T cells, conferred LPS inducibility of NFB. However, the overexpression of human Tlr4 constitutively activated the NFB reporter gene, and the treatment of these cells with LPS did not enhance the reporter gene activity. Based on these results, they concluded that human Tlr2 but not Tlr4 confers LPS responsiveness. Our results show that transfection of RAW264.7 cells with the truncated ⌬Tlr4 but not the full-length wild-type murine Tlr4 elicits a constitutive activation of the NFB reporter gene (Fig. 4).
The Constitutively Active ⌬Tlr4-induced Activation of NFB and COX-2 Expression Are Suppressed by the Inhibition of NFB with a Dominant-negative Mutant of NIK or IB␣ in RAW264.7 Cells-To determine whether the ⌬Tlr4induced expression of COX-2 is mediated through activation of NFB, we investigated whether the inhibition of NFB by a dominant-negative mutant of NIK or IB␣ results in the suppression of ⌬Tlr4-induced COX-2 expression. The results show that inhibition of ⌬Tlr4-induced NFB activation by the dominant-negative mutant of NIK or IB␣ leads to significant suppression of COX-2 expression (Fig. 5), indicating that ⌬Tlr4-induced COX-2 expression is at least in part mediated through NFB.
The Truncated Mutant (⌬Tlr4(P712H)) of Tlr4(P712H) Found in LPS-hyporesponsive Mouse Strain (C3H/HeJ) Inhibits LPS-induced Activation of NFB and COX-2 Expression in RAW264.7 Cells-To establish whether activation of Tlr4 confers LPS responsiveness, we next determined whether the expression of the mutant Tlr4(P712H) found in C3H/HeJ mouse strain inhibits LPS-induced activation of NFB and COX-2 expression. The results show that the truncated mutant ⌬Tlr4(P712H) inhibits whereas the truncated wild-type ⌬Tlr4 enhances LPS-induced activation of NFB and COX-2 expression (Fig. 6). LPS responsiveness is slightly enhanced in cells transfected with the full-length Tlr4 as compared with the vector-transfected cells. These results suggest that ⌬Tlr4 (P712H) acts as a dominant-negative mutant. Together, these results indicate that the activation of Tlr4 confers LPS responsiveness in RAW264.7 cells. However, our results do not permit ruling out the possibility that other TOLL-like receptors also mediate LPS responsiveness in RAW264.7 cells.
The Adapter Protein MyD88 Is Co-immunoprecipitated with the Wild-type ⌬Tlr4 but Not with the Mutant ⌬Tlr4(P712H) in 293T Cells-Next, we investigated why the missense mutation to replace histidine with proline at position 712 of Tlr4 resulted in loss of function. It was demonstrated that MyD88 is an adapter protein directly interacting with the cytoplasmic TOLL/IL-1R homology (TIR) domain of human Tlr4 (40) and considered as one of the most upstream components of the human Tlr4-mediated signaling cascade. This TIR domain is also present near the N-terminal death domain of MyD88. Such a sequence homology is also present among murine counterparts. Proline at position 712 lies within this TIR domain of murine Tlr4, which is critical for binding the MyD88 adapter protein. Thus, we determined whether the mutation at position 712 interferes with the binding of MyD88 to Tlr4, thereby resulting in failure of the signal transmission.
Human embryonic kidney cells (293T cells) were co-transfected with an epitope Flag-tagged MyD88 and HA-tagged wild-type ⌬Tlr4 or the mutant ⌬Tlr4(P712H) cDNA. When cell lysates from these cells were immunoprecipitated with anti-HA antibody and immunoblotted using anti-Flag or Tlr4 antibodies, MyD88 was co-immunoprecipitated with the wild-type HA-⌬Tlr4 (Fig. 7, lane 1) but not with the mutant HA-⌬Tlr4(P712H) (Fig. 7, lane 2). These results suggest that the mutant Tlr4(P712H) is unable to interact with MyD88 and thus fails to activate downstream signaling pathways.
The Constitutively Active Form of the Adapter Protein MyD88 Rescues Inability of the Dominant-negative Mutant Tlr4 to Activate the Downstream Signaling Pathway in RAW264.7 Cells-The adapter molecule MyD88 is known to be an immediate downstream signaling molecule interacting directly with the TIR domain of Tlr4 (18 -20). Proline at position 712 is located in this TIR domain. Therefore, it would be interesting to determine whether the substitution of proline with histidine resulting from the missense mutation interferes with the binding of MyD88 to the TIR domain of Tlr4. If the failure of the mutant Tlr4 to activate downstream signaling pathways is because of its inability to recruit the adapter molecule MyD88, then transfecting cells with a constitutively active form of MyD88 should restore signal transmission. Indeed, the cotransfection of RAW264.7 cells with the dominant-negative mutant (⌬Tlr4(P712H)) and a constitutively active form of MyD88 lacking TOLL/IL-1R domain (MyD88(⌬TOLL)) results in the restoration of NFB activation and COX-2 expression (Fig. 8). Taken together these results suggest that hyporesponsiveness of the mouse strain (C3H/HeJ) to LPS is because of the disruption of Tlr4-mediated signaling pathways resulting from the inability of the mutant (Tlr4(P712H)) to recruit the downstream signaling molecule MyD88.
In summary, the results presented here indicate that activation of Tlr4 confers LPS responsiveness and that disruption of Tlr4-mediated signaling pathways leads to hyporesponsiveness to LPS in murine macrophage-like cell line (RAW264.7) as determined by the activation of NFB and the expression of COX-2. Transcription factor NFB regulates the expression of a diverse array of genes including inflammation marker gene products that are involved in innate immune responses and pathogenesis of Gram-negative septic shock. Therefore, our results underscore the importance of Tlr-mediated signaling pathways in these processes.