In innate immunity, microbial components stimulate macrophages to produce antimicrobial substances, cytokines, other proinflammatory mediators, and IFNs via TLRs, which trigger signaling pathways activating NF-κB, MAPKs, and IFN response factors. We show in this study that, in contrast to its activating role in T cells, in macrophages the protein phosphatase calcineurin negatively regulates NF-κB, MAPKs, and IFN response factor activation by inhibiting the TLR-mediated signaling pathways. Evidence for this novel role for calcineurin was provided by the findings that these signaling pathways are activated when calcineurin is inhibited either by the inhibitors cyclosporin A or FK506 or by small interfering RNA-targeting calcineurin, and that activation of these pathways by TLR ligands is inhibited by the overexpression of a constitutively active form of calcineurin. We further found that IκB-α degradation, MAPK activation, and TNF-α production by FK506 were reduced in macrophages from mice deficient in MyD88, Toll/IL-1R domain-containing adaptor-inducing IFN-β (TRIF), TLR2, or TLR4, whereas macrophages from TLR3-deficient or TLR9 mutant mice showed the same responses to FK506 as those of wild-type cells. Biochemical studies indicate that calcineurin interacts with MyD88, TRIF, TLR2, and TLR4, but not with TLR3 or TLR9. Collectively, these results suggest that calcineurin negatively regulates TLR-mediated activation pathways in macrophages by inhibiting the adaptor proteins MyD88 and TRIF, and a subset of TLRs.
In innate immunity, microbial components such as LPS, lipoprotein, peptidoglycan, unmethylated CpG DNA motifs, flagellin, and viral dsRNA stimulate macrophages to produce proinflammatory cytokines such as TNF-α, IL-1, IL-6, and IL-12, as well as chemokines, inducible NO synthase, other antimicrobial responses, and coreceptor molecules. The interactions of the microbial components with TLRs lead to the activation of the transcription factor NF-κB (1, 2, 3), which is required for the induction of these effector functions. The TLR-activated signaling pathway is evolutionarily conserved, because activation of antimicrobial gene expression in insects by bacteria or fungi occurs via a homologous pathway triggered by Toll receptors that leads to the activation of the dorsal/rel family of NF-κB homologues (4).
Bacterial components activate specific TLRs on the plasma membrane, as follows: lipid A component of bacterial endotoxin LPS activates TLR4 (5); lipoprotein and peptidoglycan activate TLR2 (6); bacterial CpG DNA activates TLR9 (7); flagellin activates TLR5 (8); viral dsRNA activates TLR3 (9); and ssRNA and synthetic nucleotide derivatives activate TLR7 and 8 (10, 11). TLR10 has been identified in humans, but its ligand specificity is unknown (3, 12). TLR11, which is only expressed in mice, but not in humans, is reported to be important in the immune response against the uropathogenic bacteria (13). TLRs are expressed extra- or intracellularly. Although some TLRs (TLRs 1, 2, 4, 5, and 6) are expressed on the cell surface, others (TLRs 3, 7, 8, and 9) are found in intracellular compartments such as endosomes (14). Although the receptor for each ligand is specific, the downstream events are mediated by two shared signaling pathways. The first, which is MyD88 dependent, is activated by all TLRs, except TLR3, and also by IL-1R (15, 16). Activated receptors induce the formation of a complex of the adaptor protein MyD88 (17), which recruits IL-1R-associated kinase (IRAK)4-4, which in turn binds and activates IRAK-1, which then undergoes autophosphorylation (18, 19). Following activation, phosphorylated IRAK-1 binds TNFR-associated factor (TRAF)6 (20, 21), which in turn forms a complex at the plasma membrane with TGF-β-activated kinase (TAK)1, TAK1-binding protein (TAB)1, and TAB2, which induces the phosphorylation of TAB2 and TAK1. Activated TAK1 phosphorylates and activates the IκB kinases (IKKs). The NF-κB p50:p65 complex is normally sequestered in an inactive form in the cytoplasm through interaction with IκBs. Activated IKK phosphorylates IκBs, which are then ubiquitinated and degraded by proteasomes (22, 23, 24). The free NF-κB complex translocates to the nucleus and activates the transcription of target genes. TAK1 also activates pathways leading to the activation of ERK, JNK, and p38 MAPK pathways (25), which activate downstream transcription factors that contribute to inflammatory gene expression.
The second pathway, which is MyD88 independent, is triggered by TLR4 (which also activates the MyD88-dependent pathway) and TLR3, leading to the activation of the transcription factor IFN response factor (IRF)3 and IFN-β production (9, 26, 27), which contributes to innate antiviral responses. The Toll/IL-1R domain-containing adaptor-inducing IFN-β (TRIF) was identified as an adaptor for TLR3 and TLR4 for the MyD88-independent pathway (28, 29).
Calcineurin is a serine/threonine phosphatase that consists of a catalytic subunit, CnA, and a regulatory subunit, CnB. It participates in a variety of biological responses, including lymphocyte activation, neuronal development, muscle remodeling, memory, and heart valve morphogenesis (30). In T lymphocytes, calcineurin positively regulates the activation of NF-AT. Following binding of the TCR to peptide:MHC complexes, the increase in intracellular free Ca2+ leads to calcineurin activation through the binding of Ca2+ and calmodulin. Calcineurin dephosphorylates the inactive cytosolic form of the transcription factor NF-ATc, which then translocates to the nucleus and binds to the promoter/enhancer region of critical T cell response genes such as IL-2. In B cells, calcineurin activity is required for immune responses in vivo (31). Calcineurin is the target of the immunosuppressive drugs FK506 (tacrolimus) and cyclosporin A (CsA); these drugs block calcineurin activity by forming inhibitory complexes with FK506-binding proteins and cyclophilins, respectively.
Information about effects of FK506 or CsA on the activation of NF-κB has been limited and unclear. These immunosuppressive drugs have been reported to reduce the activation of NF-κB in T cells (32, 33). However, we have shown (34) that FK506 and CsA activate NF-κB and induce cytokine expression in nonactivated macrophages and several other nonlymphoid cell types. This study examines the mechanism by which calcineurin inhibitors activate TLR signaling and the target of negative regulation by calcineurin in the TLR activation pathways. We demonstrate that inhibition of calcineurin activates both the MyD88-dependent and the MyD88-independent pathways, reflecting calcineurin’s inhibitory effects on receptor-proximal signaling events through its interactions with certain TLRs and the adaptors MyD88 and TRIF.
Materials and Methods
Mice, cell culture, and transfections
Thioglycolate-elicited peritoneal macrophages from wild-type mice, mice deficient in MyD88, TRIF, TLR2, TLR3, or TLR4, or from TLR9 mutant TLR9CpG1/CpG1 mice (provided to J. Han by B. Beutler, The Scripps Research Institute, La Jolla, CA) were aseptically harvested. The protocol for the use of animals was approved by the Institutional Animal Care and Use Committees at Stanford University and The Scripps Research Institute.
The mouse macrophage cell line RAW 264.7 was maintained in complete RPMI 1640 medium (supplemented with 10% FBS and penicillin-streptomycin). For the luciferase reporter gene assay, cells were transiently transfected with the corresponding expression vectors using DEAE-dextran, following the manufacturer’s instructions (Promega). For the transfection of small interfering (siRNA), 40 nM control or targeting siRNA was transfectected into RAW 264.7 cells using Lipofectamine 2000 reagent (Invitrogen Life Technologies). The human embryonic kidney cell line 293T was maintained in complete DMEM (supplemented with 10% FBS and penicillin-streptomycin), and transiently transfected using Lipofectamine 2000.
Plasmids and reagents
Constitutively active calcineurin expression vector pEGFP-C1-(ΔCnA-(ΔCaM-ΔAI)) was obtained from M. Kubo (Science University of Tokyo, Tokyo, Japan). Wild-type and dominant-negative forms of MyD88 expression vectors were obtained from A. Mantovani (Mario Negri Institute, Milan, Italy). TLR4 and TLR3 expression vectors were obtained from R. Medzhitov (Yale University, New Haven, CT). Wild-type IRAK-1, wild-type IRAK-4, and two different forms of dominant-negative IRAK-4 (N-IRAK-4) (aa 1–191) and IRAK-4 KK213AA expression vectors were cloned into pFLAG-C expression vector (obtained from S. Her, Stanford University, Stanford, CA). FLAG-tagged human TRIF was cloned into pcDNA expression vector (Invitrogen Life Technologies). The pEGFP-C1 vector was purchased from BD Clontech; pNF-κB-luciferase vector was purchased from Stratagene; pSV-β-galactosidase vector was purchased from Promega. Control and CnA (the catalytic subunit of calcineurin, α isoform)-targeting siRNA were purchased from Santa Cruz Biotechnology.
Calcineurin inhibitor FK506 was provided by Fujisawa Pharmaceutical or purchased from LC Laboratories. LPS (Escherichia coli O111:B4) was obtained from Sigma-Aldrich and List Biological Laboratories; soluble peptidoglycan (Staphylococcus aureus) was purchased from Fluka; calcineurin inhibitor CsA and poly(I:C) were purchased from Sigma-Aldrich. Mouse rTNF-α was purchased from Roche Molecular Biochemicals; human rIL-1β was purchased from R&D Systems; and unmethylated CpG DNA and Pam3CSK4 were purchased from InvivoGen. FK506, cyclosporine, and all TLR ligands were tested for contamination of endotoxin by Limulus amebocyte lysate end-point test and found to have no or insignificant levels.
Abs to p65, p50, IκB-α, IκB-β, IKK-α, MEK-1, phospho-ERK-1, ERK-1, JNK, p38, IRAK-1, IRF3, calcineurin (CN-A), TLR4, and MyD88 were purchased from Santa Cruz Biotechnology. Anti-IKK-β Ab was purchased from Upstate Biotechnology; Abs to phopho-JNK, phospho-p38, and phospho-MEK-1/2 were purchased from Cell Signaling Technology; anti-GADPH Ab was purchased from Chemicon International. Abs to AU-1 and GFP were purchased from Covance, and FLAG-M2 Ab was purchased from Sigma-Aldrich.
Reporter gene assays
Transfected cells were lysed and assayed for luciferase activity using the Luciferase Assay Kit, according to the manufacturer’s protocol (Promega). Transfection efficiency was monitored by assaying β-galactosidase activity by using β-Galactosidase Assay Kit (Roche Molecular Biochemicals).
Preparation of nuclear extracts and EMSA
Nuclear extracts were prepared, as previously described (34). Radiolabeling of NF-κB oligonucleotides with [γ-32P]ATP, EMSA, and supershift assays was performed, according to the manufacturer’s protocols (Promega).
Preparation of cell lysates and Western blot
Preparation of cell lysates and Western blot analysis were performed, as previously described (35).
IKK in vitro kinase assay
IKK in vitro kinase assays were performed, as described (36). Briefly, cells were washed twice with ice-cold PBS and lysed in lysis buffer, and 300 μg of cell lysate was incubated with 4 μg of anti-IKK-α or IKK-β Abs at 4°C for 2 h, and 30 μl of protein G-agarose beads (Santa Cruz Biotechnology) were added and incubated at 4°C for an additional 3 h. The immunoprecipitate was washed three times with lysis buffer and twice with kinase buffer containing 25 mM Tris-Cl (pH 7.5), 5 mM β-glycerophosphate, 2 mM DTT, 0.1 mM sodium orthovanadate, 10 mM MgCl2, 1 mM PMSF, 1 μg/ml leupeptin, 2 μg/ml aprotinin, and 5 μg/ml pepstatin. The immunoprecipitate was incubated in 20 μl of kinase buffer supplemented with 5 μCi of [γ-32P]ATP and 5 μg of GST-IκB-α (1–54) at 30°C for 30 min, and the reaction was terminated by adding 4× SDS sample buffer. The samples were separated by SDS-PAGE on 10% acrylamide gels. The gel was cut in half, as follows: the upper part (above molecular mass of 47.5 kDa) was transferred to polyvinylidene difluoride membrane for the Western blot, and the lower part was dried and subjected to autoradiography.
Native PAGE for analysis of IRF3 activation
Native PAGE was performed, as described (37).
RNA preparation and RT-PCR analysis
Total RNA was isolated using RNeasy miniprep kit (Qiagen), and RT-PCR was performed. The primer sets used were the following: for mouse IFN-β, forward, CCACAGCCCTCTCCATCAACTATAAGC, and reverse, AGCTCTTCAACTGGAGAGCAGTTGAGG; for mouse IFN-γ-inducible protein-10 (IP-10), forward, CATCAGCACCATGAACCCAAGC, and reverse, CCTCACTCCAGTTAAGGAGCC; for the mouse housekeeping gene GAPDH, forward, TCGTGGAGTCTACTGGCGT, and reverse, GCCTGCTTCACCACCTTCT. The PCR products were analyzed on a 1.2% agarose gel.
Coimmunoprecipitation and immunoblotting
For analysis of interactions between calcineurin and MyD88, TRIF, TLR2, TLR3, TLR4, and TLR9, 293T cells were transfected with expression vectors and harvested after 24 h, and cell lysates were immunoprecipitated using Abs and analyzed by SDS-PAGE and immunoblotting.
For the analysis of the dissociation of calcineurin from TLR4 and MyD88, RAW 264.7 cells were treated with FK506 (10 μg/ml) for 0 or 30 min, and cell lysates were subjected to immunoprecipitation using anti-TLR4 or MyD88 Abs. Briefly, cell lysates were incubated with 5 μg of Ab for 1 h at 4°C, and further incubated with protein G-agarose for 2 h at 4°C. The immunoprecipitate was washed and subjected to SDS-PAGE and immunoblotting.
Statistical significance was analyzed by Student’s t test.
Calcineurin inhibitors activate NF-κB through the phosphorylation and degradation of IκBs induced by activated IκB kinases
Our previous studies showed that treatment with FK506 activates NF-κB in mouse peritoneal macrophages, mouse myelomonocytic cell line WEHI-3, and L cell fibroblasts (34). To investigate the mechanism of calcineurin regulation of the NF-κB activation pathway, cells of the mouse macrophage cell line RAW 264.7 were treated with LPS, FK506, or CsA for 4 h, and nuclear extracts were prepared for EMSA. Treatment with LPS or calcineurin inhibitors activated p50:p65 NF-κB complexes and induced the degradation of IκB-α and IκB-β (Fig. 1, A and B). These results confirm in this cell line that calcineurin-specific inhibitors FK506 and CsA activate NF-κB, consistent with the inhibition by calcineurin of the NF-κB activation pathway.
The degradation of phosphorylated IκB-α and the activation of NF-κB induced by LPS or calcineurin inhibitors were inhibited by the inclusion of the proteasome inhibitor N-acetyl-Leu-Leu-Nle-CHO (data not shown), reflecting the role of proteasomes in pIκB-α degradation. To test whether FK506 treatment activates IKK-α and/or -β, in vitro kinase assays were performed in RAW 264.7 cells. The kinase activities of both IKK-α and IKK-β were induced by LPS and also by FK506 (Fig. 1 C). These results are consistent with induction of IκΒ degradation and NF-κB activation and indicate that the activation of NF-κB by calcineurin inhibitors is mediated by IKK complex activation, leading to the phosphorylation and degradation of IκBs.
Production of the inflammatory cytokine TNF-α was examined. Peritoneal macrophages were treated with several concentrations of FK506 for 24 h, and culture supernatants were collected to measure TNF-α levels by ELISA (Fig. 1,D). Treatment with FK506 or LPS induced the production of TNF-α, consistent with the activation of NF-κB, a critical transcription factor for the expression of TNF-α in macrophages. We further examined whether FK506 augments the production of TNF-α when added with LPS (Fig. 1 D). Stimulation of macrophages by LPS induced production of TNF-α in a dose-dependent manner, and FK506 augmented the production of TNF-α at a lower dose of LPS treatment (i.e., 0.01 μg/ml). FK506 did not significantly increase production of TNF-α at the higher concentration of LPS (0.1 μg/ml), probably due to the higher potency of LPS at this concentration. Thus, inhibiting calcineurin adds to the signaling induced by LPS that leads to TNF-α production.
Calcineurin inhibitor induces the IRF3 activation and IFN-β expression
Whether FK506 activates the MyD88-independent pathway leading to IFN-β expression was also investigated. RT-PCR analysis showed that LPS, poly(I:C), and FK506 all induce the expression of IFN-β and chemokine IP-10 (Fig. 1,E). To confirm the activation of the MyD88-independent pathway, lysates from RAW 264.7 cells treated with all three stimulants were tested for the presence of the activated dimeric form of IRF3. Anti-IRF3 immunoblot revealed increased dimeric IRF3 in lysates from cells treated with FK506 as well as LPS and poly(I:C) (Fig. 1 F), suggesting that calcineurin also negatively regulates the IFN signaling pathway.
Calcineurin inhibitors also activate TLR-mediated MAPK pathways
Because the TLR signaling cascade also leads to the activation of the ERK, p38, and JNK MAPK pathways (25), whether FK506 activates MAPK pathways was examined. Treatment of RAW 264.7 cells with either LPS or FK506 induced the phosphorylation of the upstream kinase MEK-1/2 and of the ERK, JNK, and p38 kinases (Fig. 1 G), and as will be shown below, FK506 also activated MAPKs in mouse peritoneal macrophages. These results demonstrate that calcineurin inhibitors activate MAPK pathways and indicate that calcineurin negatively regulates the common TLR-proximal step(s) leading to both NF-κB and MAPK activation.
Knockdown of calcineurin enhances the activation of TLR-mediated pathway
To confirm that calcineurin functions as an inhibitor of TLR signaling pathways, which become activated when calcineurin is not active, experiments were performed using siRNA to knockdown calcineurin levels. RAW 264.7 cells were transfected with control or calcineurin-targeting siRNA (targeting catalytic subunit of calcineurin A α isoform) and cultured for 3 days. Knockdown of calcineurin was confirmed by immunoblot (Fig. 2,A). Similar to the effects of inhibiting calcineurin activity with CsA or FK506, the basal level of active NF-κB was higher in the knockdown cells. As shown in Fig. 2,B (top), higher levels of active NF-κB were present in nuclear extracts from unstimulated calcineurin siRNA-transfected cells than from unstimulated control siRNA-transfected cells. In addition, the level of active NF-κB induced by LPS, peptidoglycan, or CpG DNA was higher in the siRNA-transfected calcineurin-knockdown cells (Fig. 2,B, bottom). Induction of IFN-β mRNA and TNF-α secretion in response to LPS or poly(I:C) was also enhanced in the calcineurin-knockdown cells (Fig. 2, C and D). As was true for the activation of NF-κB itself, the low basal level of TNF-α production was higher in the knockdown cells. In addition, levels of IFN-β and TNF following stimulation with TLR ligands LPS, poly(I:C), or CpG DNA were higher in calcineurin siRNA-targeted cells. In contrast, cells in which calcineurin was knocked down by siRNA were not activated by FK506 (Fig. 2 E). This confirms that FK506 activates cells by inhibiting calcineurin, not through nonspecific or TLR-activating effects. These results support the conclusion that calcineurin inhibits the activation of TLR signaling pathways in resting cells and that inhibiting calcineurin mimics and synergizes with TLR ligands in activating downstream signaling pathways.
Calcineurin inhibits the TLR-mediated NF-κB activation pathway, but not the TNF-α receptor-mediated activation pathway
The finding that calcineurin inhibitors induce NF-κB activation indicates that in resting macrophages calcineurin inhibits the NF-κB activation pathway. Therefore, NF-κB activation by specific stimuli might be blocked or overcome by active calcineurin. To test this, cells were transfected with a vector encoding a constitutively active form of the calcineurin A subunit, whose regulatory regions (calmodulin-binding and autoinhibitory regions) have been deleted (ΔCnA-(ΔCaM-ΔAI)) (38). To assay NF-κB activation, cells were cotransfected with a pNF-κB-luc reporter vector and a β-galactosidase expression vector as the transfection control. Transfected cells were stimulated by FK506, various bacterial components, poly(I:C), IL-1β, or TNF-α (which binds to TNFR and activates NF-κB through a distinct set of receptor-proximal signaling proteins that also lead to IKK activation). In control vector-transfected cells, the various bacterial components, FK506, poly(I:C), IL-1β, and TNF-α, all activated NF-κB, but in constitutively active calcineurin vector-transfected cells, NF-κB activation by the bacterial components, FK506, poly(I:C), or IL-1β, was inhibited (Fig. 3,A). Importantly, however, NF-κB activation by TNF-α was not affected (Fig. 3 A). The TLR-specific microbial components and IL-1β all activated NF-κB in a dose-dependent manner, and activation at all concentrations was inhibited by calcineurin (data not shown). These results indicate that calcineurin specifically inhibits the IL-1R/TLR-proximal part of this pathway that is not shared by the TNF-α-mediated NF-κB activation pathway that is dependent on TRAF2 (39, 40) (i.e., upstream of the TRAF6/TAK1/TAB complex). Conversely, FK506 and CsA appear to activate NF-κB by blocking calcineurin’s inhibition of the receptor-proximal segment of the IL-1R/TLR pathway.
Calcineurin regulates receptor-proximal steps in TLR-mediated NF-κB activation pathway
After ligand binding, the TLRs (except TLR3) recruit the adaptor protein MyD88, which in turn recruits IRAK-4, which recruits and activates IRAK-1, resulting in its hyperphosphorylation and degradation (19, 41). We tested whether the activation and phosphorylation of IRAK-1 are inhibited by the expression of constitutively active calcineurin. The 293T cells were transfected with MyD88-AU-1, FLAG-IRAK-1, and constitutively active calcineurin expression vectors. Immunoblots showed that the phosphorylation of IRAK-1 following activation by MyD88 overexpression was inhibited by the overexpression of constitutively active calcineurin in a dose-dependent manner (Fig. 3 B). It is unlikely that IRAK-1 itself is the direct target of negative regulation by calcineurin, because calcineurin failed to dephosphorylate in vivo hyperphosphorylated IRAK-1 immunoprecipitated from lysates of LPS-activated cells and because the association of IRAK-1 and calcineurin could not be detected (data not shown).
Upstream of the IKK complex, the proteins involved in the TLR-proximal signaling cascade include MyD88, IRAK-4, IRAK-1, and TRAF6. As one approach to testing whether these proteins participate in the activation of NF-κB induced by FK506, cells were transfected with vectors encoding dominant-negative or mutant forms of these signaling proteins, and cotransfected with the NF-κB-luc reporter vector. In the cells transfected with any of these dominant-negative or mutant expression vectors, NF-κB activation by either LPS or FK506 was inhibited (Fig. 3 C). These results support the conclusion that calcineurin negatively regulates the NF-κB activation pathway at early, receptor-proximal step(s).
Experiments were performed to determine which receptor-proximal step or component is negatively regulated by calcineurin. The dependence of FK506-induced NF-κB activation on MyD88 was assessed by testing whether FK506 activates NF-κB in macrophages from MyD88 knockout mice. In MyD88−/− cells, the level of activation of NF-κB by FK506 or by the MyD88-dependent ligand CpG DNA was reduced compared with the level in the wild-type cells, but activation by the MyD88-independent ligand poly(I:C) was not (Fig. 3 D). Residual NF-κB activation in MyD88−/− cells by FK506 presumably reflects its activation of the MyD88-independent pathway (see below). These results indicate that calcineurin acts on or above MyD88 to block downstream signaling events leading to NF-κB activation.
Targets of negative regulation by calcineurin
The targets of negative regulation by calcineurin were further investigated using peritoneal macrophages from mice deficient in MyD88, TRIF, TLR2, TLR3, or TLR4, or from TLR9 mutant mice. Compared with wild-type macrophages incubated with FK506, FK506-treated cells from mice deficient in TLR4, TLR2, MyD88, or TRIF showed reduced/delayed degradation of IκB-α and phosphorylation of p38, JNK, and ERK MAPKs. In contrast, cells from TLR3-deficient or TLR9 mutant mice responded normally to FK506 (Fig. 4,A). Levels of TNF-α secretion by FK506-treated cells from various deficient/mutant mice paralleled the degree of IκB-α degradation and of MAPK phosphorylation. Induction of TNF-α secretion by FK506 was reduced in macrophages from mice deficient in MyD88, TRIF, TLR2, or TLR4, whereas TNF-α secretion by TLR3-deficient or TLR9 mutant cells was the same as that of wild-type cells (Fig. 4 B). Together with the finding that constitutively active calcineurin blocks activation by ligands for all TLR tested, including TLR3 and TLR9, these results strongly suggest that the receptor-proximal adaptor proteins MyD88 and TRIF, and TLR2 and TLR4, but not TLR3 and TLR9, are targets of negative regulation by calcineurin.
Calcineurin associates with MyD88, TRIF, and TLRs
To test whether calcineurin interacts with MyD88, TRIF, and TLRs, expression constructs encoding tagged TLR4, MyD88, and constitutively active calcineurin were transiently expressed in 293T cells, and possible protein interactions were assessed by immunoprecipitation and immunoblotting. GFP-tagged constitutively active calcineurin was detected in AU-1-tagged MyD88 immunoprecipitates from cells cotransfected either with MyD88 alone or with MyD88 and TLR4 (Fig. 5 A) (note that the amount of GFP-calcineurin in the immunoprecipitate is greater in cells cotransfected with FLAG-TLR4 (see below)).
Coimmunoprecipitation analysis was also used to test whether calcineurin interacts with TLR2, TLR4, and TRIF, but not with TLR3 and TLR9, as suggested by the functional studies with cells from defective/mutant mice. GFP-tagged calcineurin was coimmunoprecipitated from lysates of transfected 293T cells with FLAG-TRIF, but not with FLAG-IRAK-4, and with FLAG-TLR2 and FLAG-TLR4, but not with FLAG-TLR3 or FLAG-TLR9 (Fig. 5, B and C). These results support the findings that MyD88, TRIF, TLR2, and TLR4, but not TLR3 or TLR9, are targets of negative regulation by calcineurin. The finding that calcineurin associates with TLR4 also explains the higher levels of calcineurin in MyD88 immunoprecipitates when TLR4 was cotransfected (Fig. 5 A).
Interaction of endogenous calcineurin with TLR4 and MyD88 was examined. Calcineurin was coimmunoprecipitated with TLR4 or MyD88 from untreated cell lysates, indicating the interaction of endogenous TLR4 and MyD88 with calcineurin (Fig. 5 D). In contrast, little or no calcineurin was detected in the TLR4 or MyD88 immunoprecipitates from FK506-treated cell lysates, suggesting that the calcineurin inhibitor FK506 induces the dissociation of calcineurin from TLR4 and MyD88, which most likely leads to the activation of TLR signaling pathway.
Together these findings demonstrate that calcineurin negatively regulates both MyD88-dependent and MyD88-independent TLR signaling pathways, and that calcineurin appears to act by interacting with TLR2 and TLR4 as well as with the receptor-proximal adaptor proteins MyD88 and TRIF.
Ligand binding to the IL-1R and TLRs leads to the activation of NF-κB and MAPKs, critical events for the induction of antimicrobial mediators, proinflammatory cytokines, and other effector functions. We previously reported (34) that calcineurin inhibitors activate NF-κB in mouse peritoneal macrophages, the WEHI-3 mouse myelomonocytic cell line, primary mouse astrocytes, and the L929 fibroblast cell line, and more recently have extended these findings to the mouse macrophage cell line RAW 264.7, mouse bone marrow-derived macrophages and dendritic cells and splenic dendritic cells, and the human monocytic cell lines U937 and THP-1. Importantly, treatment of calcineurin inhibitors alone did not induce the activation of NF-κB in mouse T cells (34) or the human Jurkat T cell line (data not shown). In T cells, TNF-α is induced by activation of NF-AT (42), which is inhibited by calcineurin inhibitors, whereas in macrophages and dendritic cells it is induced by NF-κB activation. As we reported earlier (34), NF-κB is activated by FK506 and CsA, but not by the immunosuppressant rapamycin, which, like FK506 and cyclosporin, binds immunophilins and inhibits their peptidyl-prolyl isomerase activity, but unlike FK506 and CsA, rapamycin does not inhibit calcineurin. That FK506 activates macrophages by inhibiting calcineurin and not by a nonspecific calcineurin-independent activity was shown by the failure of FK506 to activate cells in which calcineurin was knocked down by siRNA.
Because calcineurin is involved in embryonic development (30), no calcineurin-negative mice have been generated. Calcineurin Aα knockout mice show a variety of abnormalities and a shortened life span (43), and retain partial phosphatase activity due to calcineurin Aβ, resulting in the partial impairment of T cell functions (44, 45). Calcineurin Aβ-deficient mice fail to generate mature T cells (46). Therefore, calcineurin inhibitors FK506 and CsA have been an important approach for dissecting the function of calcineurin in innate immunity. The activation of NF-κB by these calcineurin inhibitors or by knocking down calcineurin with siRNA, together with the inhibition of NF-κB activation by expression of constitutively active calcineurin, strongly suggests that calcineurin is a negative regulator of antimicrobial and proinflammatory gene activation in macrophages and other cell types that contribute to innate immunity.
The findings reported in this study demonstrate that the mechanisms of NF-κB activation by calcineurin inhibitors are similar to those of the well-characterized NF-κB activation pathways triggered by microbial components (schematized in Fig. 6). As with stimulation by TLR ligands, the phosphorylation and degradation of IκB-α were shown to be required for the activation of NF-κB by calcineurin inhibitors. Although it was previously reported that FK506 activates NF-κB through the phosphorylation and degradation of IκB-α in L929 fibroblast cells (36), in that study IKK-α or IKK-β activation was reported not to be involved. In our studies, FK506 activated both IKK-α and IKK-β, leading to activation of NF-κB through the phosphorylation and proteasome-mediated degradation of IκB proteins.
Calcineurin negatively regulates not only the MyD88-dependent, but also the MyD88-independent pathways. Constitutively active calcineurin inhibits NF-κB activation by poly(I:C), whereas, conversely, FK506 activates IRF3 and induces IFN-β and IP-10 expression, and siRNA-mediated calcineurin knockdown increases IFN-β mRNA expression. That macrophages from TRIF-deficient mice show delayed/reduced degradation of IκB and activation of MAPKs following FK506 treatment compared with cells from wild-type mice is also consistent with a role for calcineurin in inhibiting the MyD88-independent signaling pathway.
Results from a number of functional studies indicate that calcineurin regulates early, receptor-proximal step(s) in the TLR signaling pathway. First, because overexpression of constitutively active calcineurin blocked the activation of NF-κB in macrophages by microbial components, IL-1β, or FK506, but not by TNF, calcineurin must negatively regulate the common portion of the IL-1R/TLR-mediated NF-κB activation cascade not shared by the TNF-mediated pathway, i.e., upstream of the TRAF6/TAK1/TAB complex (2). A similar conclusion can be drawn from the finding that calcineurin inhibitors also activate MAPK pathways, which are also activated by TAK1. Third, FK506-induced activation of NF-κB was inhibited by the expression of dominant-negative forms of the receptor-proximal components MyD88, IRAK-4, and IRAK-1. Fourth, the phosphorylation of IRAK-1 induced by the overexpression of MyD88 was inhibited by the expression of constitutively active calcineurin. Finally, that calcineurin acts at receptor-proximal steps of the TLR activation pathway was most convincingly demonstrated by the finding that activation by FK506 of downstream signaling pathways leading to IκB-α degradation and MAPK phosphorylation is delayed or impaired in macrophages from mice deficient in MyD88, TRIF, TLR2, or TLR4 compared with macrophages from wild-type mice. Interestingly, FK506 activation of downstream pathways is not impaired in macrophages from TLR3-deficient or TLR9 mutant mice.
These observations suggest that the targets of negative regulation by calcineurin are the adaptor proteins MyD88 and TRIF, and some TLRs as follows: TLR2 and TLR4 (which are expressed on the cell surface), but not TLR3 or TLR9 (which are not on the cell surface, but in endosomal membranes). Whether this apparent specificity of calcineurin interaction reflects the cellular localization of the TLR or other TLR differences, such as in their cytoplasmic domains, remains to be determined. The ability of constitutively active calcineurin to inhibit NF-κB activation by TLR3 and TLR9 ligands (poly(I:C) and CpG DNA, respectively) presumably reflects inhibitory interactions of calcineurin with the adaptors TRIF and MyD88.
Consistent with the functional evidence that calcineurin negatively regulates receptor-proximal steps in the TLR activation pathway, calcineurin was shown to associate both with TLR2 and TLR4, but not with TLR3 or TLR9, and with the proximal adaptor proteins MyD88 and TRIF.
Our findings (Ref. 34 and this study) that treatment of macrophages and some other cell types with calcineurin inhibitors activates NF-κB and MAPK indicate that these resting cells have basal levels of calcineurin activity that help to maintain the TLR-mediated NF-κB and MAPK activation pathways in an “off” position that is reversed by calcineurin inhibitors (Fig. 6). The presence of basal calcineurin activity is supported by nuclear localization studies with nonactivated mouse peritoneal macrophages and WEHI-3 myelomonocytic cells that showed detectable levels of NF-AT in the nucleus (34) (nuclear localization of NF-AT is dependent on its dephosphorylation by calcineurin). Exposing these cells to calcineurin inhibitors resulted in a decrease in NF-AT and a concomitant increase in the activation and nuclear localization of NF-κB. We found that the free calcium level in resting macrophages is 130–150 nM (unpublished observation, T. Chan, P. Jones, R. Luik, and R. Lewis), and reports in the literature also indicate that resting macrophages have higher levels of intracellular free calcium (100–200 nM) than do resting T cells (47, 48). This appears to be sufficient to support a basal level of calcineurin activity in macrophages that normally keeps the signaling pathways off until calcineurin activity is blocked or bypassed following activation by IL-1R/TLR ligands.
The mechanism(s) by which active calcineurin negatively regulates the activation of the IL-1R/TLR pathway at the level of the receptors and their proximal adaptor proteins is not known. Because calcineurin is a serine/threonine phosphatase, one possible mechanism is that calcineurin removes a serine- or threonine-associated phosphate group that is essential for activation. No serine/threonine phosphorylation of TLRs, MyD88, or TRIF has been described. However, evidence has been published that TLR4 activation by LPS results in MyD88 tyrosine phosphorylation, allowing recruitment of PI3K to MyD88 and activation of Akt, which contributes to the activation of NF-κB and expression of IL-1β (49). Kinase-mediated serine/threonine phosphorylation and calcineurin-mediated dephosphorylation of the receptors and proximal adaptors could play a role in regulating the activation status of these proteins.
Several cytoplasmic proteins that function as calcineurin inhibitors have been identified, and in some cases they are components of complexes with receptors and other signaling components that regulate the activation of signaling pathways. For example, in neurons, calcineurin is associated with A-kinase anchoring protein (AKAP), which inhibits its phosphatase activity. Protein kinases A and C are also associated with AKAP in these complexes (50, 51). Thus, complexes of AKAP, calcineurin, and protein kinases constitute a signal regulation complex that may reversibly modulate the phosphorylation status and activity of signaling molecules.
Calcineurin inhibitors FK506 and CsA are potent immunosuppressants, reflecting calcineurin’s key role in T cell activation due to its calcium flux-activated dephosphorylation and activation of NF-AT. Treatment with either inhibitor reduces immune-mediated rejection following organ transplantation. However, they also cause side effects, including nephrotoxicity, hypertension, and increased risk of cardiovascular events (52, 53). The pathogenesis of nephrotoxicity and the other side effects induced by calcineurin inhibitors are complex and incompletely understood, but among the factors implicated are renal and systemic vasoconstriction, increased release of endothelin-1, and increased expression of TGF-β1. Endothelin-1 transcription is controlled by NF-κB in vascular endothelium (54), and endothelin-1 itself is proinflammatory, because it activates NF-κB in macrophages (55). Interstitial fibrosis caused by CsA is associated with increased expression of TGF-β1 (56), and TGF-β1 is responsive to induction by IL-1 in macrophages, which activates NF-κB and several other transcription factors (57). IL-6 has been reported to contribute to nephrotoxicity (58, 59), and FK506 induces IL-6 production in fibroblasts through the activation of NF-κB (60). A role for FK506-induced NF-κB activation in inducing nephrotoxicity was shown by a study in rats of chronic FK506 nephropathy, in which the NF-κB inhibitor pyrrolidine dithiocarbamate reduced FK506-induced monocyte/macrophage infiltration, tubular injury, and intestinal fibrosis, which are hallmarks of nephrotoxicity (61). These observations suggest that the activation of NF-κB and subsequent production of inflammatory cytokines and other mediators may contribute to nephrotoxicity and other side effects from the use of calcineurin inhibitors, perhaps through the mechanisms described in this study, and that use of specific inhibitors of the NF-κB activation pathway may reduce the side effects caused by FK506 and CsA treatment.
Systemic activation of inflammatory mediators in patients treated with calcineurin inhibitors, which might be expected based on our in vitro findings, has not been reported. This appears to reflect a well-known alternative consequence of the activation of the TLR pathway and NF-κB: negative feedback regulation. As has been well described with LPS (endotoxin) tolerance (37), after activating an initial response, exposure to LPS or other TLR ligands induces negative feedback pathways that inhibit NF-κB activation and innate and inflammatory responses following a second exposure. We have found that exposure of murine macrophages and dendritic cells to FK506 in vitro or in vivo induces tolerance to subsequent challenge with either LPS or FK506 in vitro (C. Jennings, B. Kusler, and P. Jones, manuscript in preparation). In addition, we have found that, as has been observed with LPS, initial injection of FK506 provides some protection against in vivo LPS-induced endotoxin shock. Thus, in vivo treatment with calcineurin inhibitor immunosuppressants may induce a state of ongoing suppression of innate as well as adaptive immune responses, contributing to the overall state of immunosuppression and increased susceptibility to infections.
FK506 used in some experiments was generously provided by Fujisawa Healthcare. We thank many laboratories that generously provided reagents, Bruce Beutler for providing mice, and Martha Cyert for helpful discussions.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by funds from Stanford University (to P.P.J.).
Abbreviations used in this paper: IRAK, IL-1R-associated kinase; AKAP, A-kinase anchoring protein; CsA, cyclosporin A; IKK, IκB kinase; IP-10, IFN-γ-inducible protein-10; IRF, IFN response factor; siRNA, small interfering RNA; TAB, TGF-β-activated kinase 1-binding protein; TAK, TGF-β-activated kinase; TRAF, TNFR-associated factor; TRIF, Toll/IL-1R domain-containing adaptor-inducing IFN-β.