Src family kinases are involved in a plethora of aspects of cellular signaling. We demonstrate in this study that the Src family kinase Lyn negatively regulates TLR signaling in murine bone marrow-derived macrophages (BMMΦs) and in vivo. LPS-stimulated Lyn−/− BMMΦs produced significantly more IL-6, TNF-α, and IFN-α/β compared with wild type (WT) BMMΦs, suggesting that Lyn is able to control both MyD88- and TRIF-dependent signaling pathways downstream of TLR4. CD14 was not involved in this type of regulation. Moreover, Lyn attenuated proinflammatory cytokine production in BMMΦs in response to the TLR2 ligand FSL-1, but not to ligands for TLR3 (dsRNA) or TLR9 (CpG 1668). In agreement with these in vitro experiments, Lyn-deficient mice produced higher amounts of proinflammatory cytokines than did WT mice after i. v. injection of LPS or FSL-1. Although Lyn clearly acted as a negative regulator downstream of TLR4 and TLR2, it did not, different from what was proposed previously, prevent the induction of LPS tolerance. Stimulation with a low dose of LPS resulted in reduced production of proinflammatory cytokines after subsequent stimulation with a high dose of LPS in both WT and Lyn−/− BMMΦs, as well as in vivo. Mechanistically, Lyn interacted with PI3K; in correlation, PI3K inhibition resulted in increased LPS-triggered cytokine production. In this line, SHIP1−/− BMMΦs, exerting enhanced PI3K-pathway activation, produced fewer cytokines than did WT BMMΦs. The data suggest that the Lyn-mediated negative regulation of TLR signaling proceeds, at least in part, via PI3K.
Recognition of conserved molecular structures of pathogens by pattern recognition receptors, including TLRs on or in cells of the innate immune system (e.g., monocytes and macrophages [MΦs]), provides a basis of antimicrobial defense. Sensing of bacterial markers, such as LPS (endotoxin) or lipopeptides by TLR4 or TLR2, respectively, triggers the production of endogenous mediators, including various cytokines, reactive oxygen species, and MHC and costimulatory molecules required for efficient Ag presentation (1, 2).
It is known that excessive production of proinflammatory cytokines by monocytes and MΦs in response to multiple waves of pathogenic stimuli can promote the development of septic shock, which still remains a major cause of mortality worldwide (3, 4). Important molecular mechanisms to avoid such detrimental inflammatory reactions are negative feedback regulation and the so-called endotoxin or microbial tolerance. Prior exposure to sublethal doses of LPS or certain other microbial components leads to a transient state of hyporesponsiveness to subsequent challenges with microbial constituents at lethal doses (4, 5). Various mechanisms have been proposed for endotoxin tolerance. Among them are the downregulation of the LPS-receptor TLR4 (6) and decreased build-up of signal-promoting protein complexes at the TLR4—for example, decreased recruitment of the adaptor protein MyD88 to TLR4 or suppressed interaction between MyD88 and the kinase IRAK-1 (7, 8). Furthermore, a shift from transcriptionally competent NFκB heterodimers (p50/p65) to transcriptionally inactive homodimers (p50/p50) was associated with the tolerance state (9). In addition, the upregulation of inhibitory proteins, such as IRAK-M, SOCS-1, and SHIP1, induced by low-dose microbial stimulation has been implicated in the generation of the hyporesponsive state (10–13). Finally, loss of tyrosine phosphorylation of TLR4, probably via inhibition of the Src family kinase (SFK) Lyn, has been proposed (14).
SFKs are involved in the signal transduction of many different receptor systems, leading to a wide array of responses such as proliferation, differentiation, and mediator secretion (15). In addition, it is evident now that redundancy and hierarchy exist in this kinase family. The latter is best exemplified by the fact that the family member Lyn is able to negatively control other family members by regulating their phosphorylation at the C-terminal inhibitory tyrosine residue (16). It does so by phosphorylating the transmembrane adaptor protein PAG/Cbp, which leads to the recruitment of C-terminal Src kinase, which in turn phosphorylates the C-terminal tyrosine residues in SFK members (16). Thus, it has been shown in different cell types addressing various receptor systems that Lyn deficiency results in hyperactivity of the respective system under investigation. For example, Lyn-deficient neutrophils exert a hyperadhesive phenotype mediated by integrins (17), and Lyn-deficient mast cells show enhanced Ag-triggered degranulation, cytokine secretion, and cytokine-induced proliferation (18, 19). In vivo, deficiency of Lyn leads to a severe lupus-like autoimmunity (20). This autoimmunity can result from loss of a feedback inhibitory pathway in which Lyn phosphorylates CD22, which then dampens B cell Ag receptor signaling via phosphatase recruitment. Loss of Lyn also compromises an inhibitory pathway involving FcγRIIB (CD32) and SHIP1, an SH2-domain–containing inositol 5-phosphatase (21). Phosphorylation of a tyrosine residue in the cytoplasmic part of CD32 by Lyn enables recruitment of SHIP1, thus causing downregulation of PI3K- and Erk-mediated signaling pathways (21, 22). Furthermore, direct phosphorylation and activation of SHIP1 by Lyn has been reported with a loss of SHIP1 activity in Lyn-deficient mast cells (19).
Lyn is activated after LPS stimulation (23), and more interestingly, tyrosine phosphorylation of TLR4 by Lyn was reported to be important for TLR4 function (14). In this study, we have investigated the role of Lyn in TLR signaling in MΦs in vitro and in mice in vivo. We show that Lyn attenuates LPS- and lipopeptide-induced secretion of different cytokines in MΦs and in vivo. Although Lyn is involved in dampening the cytokine TLR response, Lyn-deficient MΦs and mice still develop microbial tolerance. Our data further indicate that Lyn exerts its negative function via activation of the PI3K pathway and that SHIP1 promotes TLR-mediated activation of macrophages. In conclusion, Lyn is an important negative regulator of TLR-induced cytokine response in MΦs and in vivo.
Materials and Methods
S-LPS from Salmonella abortus equi and R-form LPS from Salmonella minnesota mutant R595 were extracted and purified as described (24). Synthetic lipopeptide (FSL-1) was obtained from Echaz Microcollections (Tübingen, Germany), and synthetic CpG 1668 (TLR grade) and polyinosinic-polycytidilic acid (dsRNA; TLR grade) were obtained from Alexis Biochemicals (Enzo Life Sciences, Lörrach, Germany). Recombinant LPS binding protein (LBP) was obtained from Biometec (Greifswald, Germany). Human serum albumin was purchased from Octapharma (Langenfeld, Germany). The PI3K inhibitor Wortmannin was purchased from Calbiochem (Darmstadt, Germany) and p38 inhibitor, BIRB0796 from Division of Signal Transduction Therapy, College of Life Sciences, University of Dundee (Dundee, Scotland, U.K.). Polyclonal rabbit anti-Lyn (clone no. 44) and anti-p110δ (H-219) Abs were purchased from Santa Cruz Biotechnology (Heidelberg, Germany), anti–P-PKB (S473) and anti–P-p38 (T180/Y182) from Cell Signaling Technology (Frankfurt, Germany), and anti-p85 from Upstate/Biomol (Hamburg, Germany).
Wild type (WT) C57BL/6, Lyn+/+ and Lyn−/− (20) as well as SHIP1+/+ and SHIP1−/− mice (25) (C57BL/6 × 129/Sv) were bred under specific pathogen-free conditions in the animal facilities of the Max-Planck-Institute for Immunobiology. Littermates of both sexes, at the age of 6–8 wk were used.
For cultivation of bone marrow-derived MΦs (BMMΦs) bone marrow cells were flushed from mouse femurs and cultivated at a concentration of 1 × 105 cells per milliliter in hydrophobic Teflon film bags (Heraeus, Hanau, Germany) for 10 d. The culture medium contained 70% high-glucose formulation of DMEM (Life Technologies, Paislay, U.K.), 10% FCS, 5% horse serum, 1 mM sodium pyruvate (Life Technologies), 60 μM 2-ME (Roth, Karlsruhe, Germany), 100 U/ml penicillin (Life Technologies), 100 μg/ml streptomycin (Life Technologies) and 15% L-cell–conditioned medium containing CSF. L-Cell–conditioned medium containing GM-CSF was prepared as described previously (26). RAW264.7 macrophages were cultured, in 37°C at 5% CO2, in DMEM supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM l-glutamine.
Cytokine induction in vitro
MΦs were harvested by centrifugation at 4°C, washed, and resuspended in serum-free DMEM. Thereafter, they were placed (105 cells per 200 μl) in 96-well plates (Nunc, Roskilde, Denmark) and cultured for 24 h at 37°C in a humidified atmosphere containing 8% CO2. After exchanging the medium, cells were stimulated with the indicated amounts of LPS, FSL-1, CpG 1668, or dsRNA in triplicates, and culture supernatants were collected after 4 h for TNF-α, after 6 h for IFN-α/β, and after 24 h for IL-6 measurements. Supernatants were stored at −80°C until use.
Cytokine induction in vivo
Mice were injected i.v. with different amounts of R-LPS or S-LPS in 0.15 M glucose solution (0.2 ml per 20 g body weight) or lipopeptides in PBS (0.2 ml per 20 g body weight). Mice were exsanguinated under isoflurane anesthesia, and heparinized blood for cytokine determination was collected. Plasma TNF-α was detected 1 h after injection, and IL-6 and IFN-α/β were detected 2 h after injection.
Determination of IL-6, TNF-α,and IFN-α/β
IL-6 levels in culture supernatants and in plasma were measured by ELISA using the MP5-20F3 rat anti-mouse IL-6 Ab (BD Pharmingen, San Diego, CA) as the capturing Ab and the MP5-32C11 biotinylated rat anti-mouse IL-6 Ab (Pharmingen) as the detection reagent for IL-6, according to the supplier’s instructions. The detection limit of the assay was 30 pg/ml. TNF-α levels in culture supernatants and plasma were determined in a cytotoxicity test using a TNF-α sensitive L929 cell line in the presence of actinomycin D (27). The detection limit of the assay was 5.0 pg TNF-α/ml supernatant and 50 pg/ml plasma. IFN-α/β levels in culture supernatants and in plasma were measured using an L929 cell line transfected with an IFN-sensitive response element luciferase construct (provided by B. Beutler, The Scripps Research Institute, La Jolla, CA) as described (28).
Western blotting and immunoprecipitation
For Western blotting, MΦs were resuspended in serum-free DMEM (1 × 107 cells/ml) and stimulated in Teflon bags for the indicated time points at 37°C in a humidified atmosphere containing 8% CO2. Stimulation was stopped by adding ice-cold PBS containing 1 mM NaVO4. Cells were washed and lysed in RIPA buffer (20 mM Tris pH 7,5, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 1 mM EDTA, 0.1% SDS, phosphatase inhibitor mixture I + II [Sigma-Aldrich, St. Louis, MO], protease inhibitor mixture [Sigma-Aldrich], and PMSF [Sigma-Aldrich]). Cell lysates (50 μg protein per lane) were subjected to SDS-PAGE under reducing conditions and transferred to nitrocellulose membranes (Amersham Life Science, Buckinghamshire, U.K.). Nonspecific binding was blocked by overnight incubation of membranes in TBS containing 5% powdered milk. Membranes were incubated overnight at 4°C with first Ab and for 1 h at room temperature with the enzyme-coupled secondary Ab, both diluted in TBS containing 0.5% BSA. After washing, specific proteins were detected with ECL reagents (ECL, Amersham Pharmacia Biotech, Zürich, Switzerland). For immunoprecipitation, 107 cells were lysed with IP-lysis buffer (phosphorylation solubilization buffer (29) supplemented with 0.5% NP-40, 0.5% deoxycholate, 1% phosphatase inhibitor mixture I, 1% phosphatase inhibitor mixture II, 1% protease inhibitor mixture and PMSF) for 45 min on 4°C. Cell lysates were precleared by incubation with 10 μl protein G-Sepharose beads (PharmaciaBiotech, Freiburg, Germany) for 1 h at 4°C, and subsequently incubated overnight with 5 μg/ml anti-Lyn or anti-PKB Ab at 4°C on a rotary wheel. For the pull-down, 10 μl protein G-sepharose beads were added for an additional hour. The beads were washed twice in IP-lysis buffer and immunoprecipitated proteins were analyzed by Western blotting.
Lyn negatively regulates proinflammatory cytokine secretion in macrophages
In different murine and human macrophage models, phosphorylation and activation of the Src kinase Lyn in response to LPS has been demonstrated (23, 30, 31). To elucidate the role of Lyn in TLR4-mediated effector functions, BMMΦs were grown in vitro from the bone marrows of WT and Lyn-deficient (Lyn−/−) mice and stimulated with increasing concentrations of S. minnesota Re-LPS. Both cell types expressed comparable amounts of the TLR4/MD-2 complex on the cell surfaces when analyzed by FACS (data not shown). However, Lyn−/− BMMΦs produced significantly higher levels of the proinflammatory cytokines IL-6 and TNF-α, compared with WT cells (Fig. 1A, 1B), indicating that Lyn acts as a negative regulator of R-LPS–induced signaling. Recently, we demonstrated that recognition of R-chemotypes of LPS by BMMΦs in the absence of LBP is independent of mCD14, whereas recognition of S-LPS is dependent on the presence of both mCD14 and LBP (32). Thus, using S-LPS we investigated whether Lyn is also influencing activation of BMMΦs via the mCD14-dependent route, given that GPI-anchored mCD14 and palmitoylated/myristoylated Lyn have been reported to interact functionally (23). In agreement with our previous results (32), responses of WT and Lyn−/− BMMΦs to S-LPS were dramatically reduced compared with R-LPS (Fig. 1A, 1C). Nevertheless, IL-6 production at high S-LPS concentrations in Lyn−/− BMMΦs was slightly higher than in WT cells (Fig. 1C). The addition of recombinant LBP to cell cultures resulted in increased sensitivity to S-LPS in both cell types—again, Lyn−/− BMMΦs responded significantly stronger than did WT BMMΦs (Fig. 1C). This finding demonstrates that the Src kinase Lyn is capable of negatively controlling both mCD14-independent and mCD14-dependent LPS-induced activation of the proinflammatory MΦ response.
TRIF-dependent pathway in macrophages is attenuated by Lyn
Production of proinflammatory cytokines, like IL-6 and TNF-α, is under the control of the adaptor proteins MyD88 and TRIF, whereas transcriptional regulation of IFN-α/β expression depends exclusively on the adaptor protein TRIF (33). To elucidate whether Lyn is capable of regulating TRIF-dependent activation, WT and Lyn−/− BMMΦs were stimulated with R-LPS and IFN-α/β secretion was measured. As was the case for the proinflammatory cytokines, Lyn−/− BMMΦs secreted enhanced levels of IFN-α/β compared with WT cells (Fig. 1D). These data indicate that Lyn is negatively controlling the TRIF signaling pathway downstream of TLR4/MD-2.
Lyn attenuates LPS effects in vivo
Lyn deficiency results in enhanced cytokine responses to LPS in BMMΦs. Next, we were interested in the role of Lyn in the course of the respective in vivo responses. WT and Lyn−/− mice were injected i.v. with R-LPS. The production of IL-6 and TNF-α as well as TRIF-dependent IFN-α/β cytokines was analyzed in serum at the indicated time points. Corroborating our in vitro observations, LPS induced augmented production of IL-6 (Fig. 2A), TNF-α (Fig. 2B), and IFN-α/β (Fig. 2C) in Lyn−/− mice compared with WT mice. This finding indicates that the SFK Lyn plays an important negative regulatory role downstream of the LPS receptor, TLR4, in MΦs and in vivo.
Lyn suppresses TLR2-mediated cytokine production in MΦs and in vivo
In a study analyzing invasion of alveolar epithelial cells by Pseudomonas aeruginosa, interaction of Lyn and TLR2 as well as a positive role of Lyn in this process has been observed (34). Having seen the negative effects of Lyn on TLR4-mediated cytokine production, we were interested in whether Lyn might play a comparable role in the context of a different TLR—TLR2. Moreover, because TLR2 signaling is exclusively dependent on the adapter protein MyD88, we sought to elucidate whether Lyn is also involved in regulating MyD88-dependent cytokine production. WT and Lyn−/− BMMΦs were stimulated with increasing concentrations of the TLR2 agonist FSL-1, a synthetic analog of a mycoplasma lipoprotein (35), and levels of secreted IL-6 and TNF-α were measured. As was already observed after LPS stimulation, Lyn-deficient BMMΦs secreted substantially higher levels of IL-6 (Fig. 3A) and TNF-α (Fig. 3B) in response to FSL-1 compared with WT BMMΦs. When injected into mice, FSL-1 induced stronger accumulation of IL-6 (Fig. 3C) and TNF-α (Fig. 3D) in the sera of Lyn-deficient compared with WT mice. These data demonstrate that Lyn is capable of exerting a negative role downstream of both TLR4 and TLR2, controlling both MyD88- and TRIF-dependent cytokine production.
Intracellular TLRs do not make use of Lyn
After having shown that Lyn acts as a negative regulator of TLR2 and TLR4 signaling, the question arose whether Lyn might exert a comparable role in regard to signal transduction from TLRs located in the endosome, like TLR9 and TLR3, the receptors for CpG DNA and dsRNA, respectively (1). However, stimulation of WT and Lyn-deficient BMMΦs with CpG 1668 or dsRNA did not reveal significant differences concerning production and secretion of the proinflammatory cytokines, IL-6 and TNF-α (Fig. 4). These data indicate that, unlike signaling from the surface receptors TLR2 and TLR4, the signaling from intracellular TLRs is not negatively regulated by the Src kinase Lyn.
TLR ligand-mediated tolerance in Lyn-deficient macrophages and mice
Prior exposure to small amounts of LPS or lipopeptide induces a state of cell refractoriness to subsequent challenge with higher doses of the same stimulus known as endotoxin/lipopeptide tolerance (4). It has been shown on a molecular level that induction of LPS tolerance suppresses LPS-mediated TLR4 tyrosine phosphorylation and recruitment of Lyn to TLR4 (14). Thus, in the light of our in vitro and in vivo data, we sought to address the role of Lyn in the development of tolerance. WT and Lyn-deficient BMMΦs were treated with 0.1 μg R-LPS/ml or vehicle (PBS) and 24 h later challenged with 1.0 μg R-LPS/ml. Subsequently, IL-6 (Fig. 5A) and TNF-α concentrations (Fig. 5B) in the supernatants were determined. Both WT and Lyn−/− cells appeared to be completely susceptible to tolerization. Moreover, as already shown in Fig. 1, Lyn−/− BMMΦs produced significantly more cytokines than WT cells under intolerant conditions. Next, the ability of LPS to induce tolerance in the absence of Lyn was analyzed in vivo. WT and Lyn−/− mice were treated with 0.2 μg R-LPS per 20g body weight or vehicle (PBS) and challenged 24 h later with 2 μg R-LPS per 20g body weight. Plasma for TNF-α determination was collected 1h after challenge. In agreement with previous results (Fig. 2B), LPS injection into PBS-treated Lyn−/− mice caused significantly elevated plasma levels of TNF-α compared with PBS-treated WT mice (Fig. 5C). Moreover, like the LPS-pretreated BMMΦs in vitro (Fig. 5A), the LPS-pretreated mice developed tolerance regardless of presence or absence of Lyn. However, contrary to BMMΦs, final cytokine levels in tolerant Lyn−/− mice were higher than in tolerant WT mice (Fig. 5C). These data suggest that in comparison with other negatively acting proteins, the deficiency of which results in complete loss of endotoxin tolerance (like IRAK-M and SOCS-1) (10–12), Lyn appears not to be involved in the induction of tolerance, but rather regulates the signaling capacity of TLRs.
The Lyn/PI3K module negatively regulates TLR signaling in BMMΦs
Lyn has been reported to be involved in the activation of SHIP1 (19), which has been suggested to either restrict or promote MΦ activation through TLR4 (13, 36, 37). To elucidate whether activated SHIP1 might mediate the inhibitory activity of Lyn in the cellular system under investigation, WT and SHIP1−/− BMMΦs were stimulated with increasing concentrations of R-LPS, and the cytokine response was measured. In agreement with Fang et al. (36), but contrary to Sly et al. (13), SHIP1-deficient BMMΦs produced significantly less IL-6 and TNF-α compared with WT cells (Fig. 6A). Comparable data were obtained by analyzing IL-6 and TNF-α production in response to the TLR2 ligand FSL-1 (Fig. 6B). These data suggest that SHIP1 is, unlike Lyn, an enhancer rather than an attenuator of TLR signaling in BMMΦs. Moreover, they suggest that the negative regulation of TLR responses by Lyn in BMMΦs is not mediated through SHIP1.
SHIP1 is hydrolyzing the phospholipid second messenger, phosphatidylinositol-3,4,5-trisphosphate (PIP3), which is the product of the PI3K-catalyzed reaction. Activated PI3K negatively regulates cytokine induction downstream of TLRs (38). In agreement, inhibition of PI3K by Wortmannin resulted in significantly increased IL-6 and TNF-α secretion from LPS-stimulated WT BMMΦs (Fig. 6C). Moreover, Lyn interacted with the regulatory subunit of PI3K (p85) and the catalytic subunit (p110δ) in unstimulated as well as LPS-stimulated MΦs (Fig. 6D). Therefore, our data suggested that the contrariwise effects of Lyn and SHIP1 on TLR signaling can result from their distinct roles in the PI3K signaling pathway.
Polumuri et al. (38) reported that increased activation of PI3K-dependent PKB leads to attenuated p38 MAPK activation and cytokine secretion in MΦs. Thus, LPS-stimulated SHIP1-deficient BMMΦs should exert increased PKB and decreased p38 phosphorylation/activation compared with the corresponding WT cells. Alternatively, LPS treatment of Lyn-deficient BMMΦs should result in attenuated PKB and augmented p38 phosphorylation/activation compared with WT cells. In this study, we found that PI3K-dependent PKB phosphorylation in response to R-LPS was stronger in SHIP1-deficient and attenuated in Lyn-deficient BMMΦs compared with cells from WT mice (Fig. 6E). Moreover, the absence of SHIP1 and Lyn exhibited an opposite effect on LPS-induced p38 phosphorylation: it was decreased in SHIP1-deficient and augmented in Lyn-deficient BMMΦs (Fig. 7A). These data suggest that Lyn, in contrast to SHIP1, is involved in the activation of PI3K/PKB, which in turn acts in a suppressive manner in the course of TLR signaling in BMMΦs by controlling p38 activation. In corroboration, inhibition of PI3K by Wortmannin in WT BMMΦs resulted in increased and prolonged LPS-triggered p38 phosphorylation (Fig. 7B). Moreover, the enhancing effect of Wortmannin on LPS-induced IL-6 production was more pronounced in SHIP1-deficient than in Lyn-deficient BMMΦs (Fig. 7C). Inhibition of p38 by BIRB0796 caused complete suppression of LPS-induced IL-6 and TNF-α production (Fig. 7D). Our data indicate that Lyn controls the PI3K-mediated attenuation of TLR-induced cytokine production.
The role of the SFK Lyn in LPS stimulation of various cell types has been of long-standing interest. Fifteen years ago, activation of Lyn as a CD14-associated protein in response to LPS was reported by Horak et al. (23). In addition, LPS stimulated activation of two other SFKs, Hck and Fgr (23). Analysis of MΦs from Hck−/−Fgr−/−Lyn−/− mice revealed strong reduction in tyrosine-phosphorylated proteins; however, nitrite production and cytokine secretion in response to LPS were normal or enhanced (39). Further analysis to identify the kinase responsible for enhanced responses is lacking. Recently, Lyn was shown to phosphorylate TLR4, and it was suggested that Lyn-mediated TLR4 phosphorylation is involved in the process of endotoxin tolerance (14).
In this study, we demonstrate that Lyn negatively regulates MyD88- and TRIF-dependent cytokine secretion in LPS-stimulated BMMΦs as well as in vivo (Figs. 1, 2). Lipopeptide-stimulated cytokine production in MΦs and in vivo was comparably attenuated by Lyn (Fig. 3), suggesting that it represents a common control element for signal transduction events induced by different types of plasma membrane TLRs. Indeed, signaling from intracellular TLRs was not affected by Lyn-deficiency (Fig. 4). This attenuation affects TLR4 and TLR2 signaling at an early phase and functions at the first encounter to TLR stimuli, because induction of LPS tolerance was not qualitatively influenced by Lyn deficiency (Fig. 5). Mechanistically, Lyn exerts part of its suppressive role by the functional interaction with PI3K; correlating with that, SHIP1, an important counter-player of PI3K, acts as a promoting element of TLR signaling in BMMΦs (Fig. 6).
Lyn interacts with PI3K in unstimulated and LPS-stimulated BMMΦs (Fig. 6). Such an interaction has been reported already in LPS-treated human monocytes by Reiner et al. (31). Furthermore, Lyn was demonstrated to bind via its SH3 domain to the regulatory subunit of PI3K, p85, thus activating the lipid kinase (40). Because Lyn is membrane bound via its palmitoyl and myristoyl anchors, it would aid PI3K in translocating to its lipid substrate, phosphatidylinositol-4,5-bisphosphate. p85 is the regulatory subunit of class IA PI3K. However, class IB PI3K playing a role in this suppressive signaling module cannot be fully excluded for two reasons. First, the PI3K inhibitor Wortmannin does not discriminate between class IA and IB PI3Ks. Second, a G protein-coupled receptor, the chemokine receptor CXCR4, has been found to interact with TLR4 (41), and thus Gq protein-mediated activation of class IB PI3K might be considered. Further investigations are needed to clarify this subject. The observed suppressive effect of Lyn on both MyD88- and TRIF-dependent pathways raised the interesting question of whether there is a common inhibitory module for the production of proinflammatory cytokines and type I IFNs, or whether there are separate, pathway-specific modules comprising of Lyn, PI3K, and pathway-specific factors.
Interestingly, a functional role of PI3K in the course of TLR signaling has been reported by others; however, positive and negative regulatory mechanisms have been proposed. PI3K positively regulated NFκB activation in response to TLR2 ligands in THP-1 and HEK293 cells (42), and involvement of PI3K in the TLR2/4-mediated induction of several cytokine and chemokine mRNAs was demonstrated in human dendritic cells (43). Moreover, in human neutrophils PI3K promoted TLR4-induced activation of JNK and thus, expression of MCP 1 (44). In regard to a negative role of PI3K, induction of NO synthesis (45, 46) and expression of cyclooxygenase 2 in response to LPS (47) have been reported to be under the control of PI3K in different types of MΦs. Furthermore, PI3K, via negative regulation of p38 MAPK, attenuates IL-12 production in dendritic cells stimulated by TLR2/4/9 ligands (48). In this line, costimulation of MΦs via TLR4 and FcγR resulted in reduced secretion of IL-12 compared with MΦs stimulated with LPS alone. This was due to increased activation of PI3K and PKB leading to attenuated p38 MAPK activation in costimulated cells (38). In this study, we observed enhanced p38 phosphorylation as well as TNF-α and IL-6 responses in LPS-stimulated macrophages treated with the panPI3K-specific inhibitor Wortmannin (Figs. 6 and 7). Recently, Sly et al. (49) reported on the differential effect of isoform-specific PI3K inhibitors on LPS-induced IL-6 production in MΦs. In particular, p110β appeared to negatively regulate this response. This finding might suggest that, in BMMΦs differentiated in our study, LPS-induced effects of PI3K are particularly dependent on the p110β isotype. An additional model was proposed recently by Vogel et al. Because the adaptor protein Mal contains a functionally important phosphatidylinositol-4,5-bisphosphate–binding domain (50), they suggest limitation of this phospholipid by PI3K action and thus, attenuation of Mal/MyD88-mediated signaling (51). This model, however, would not explain the effect of SHIP1 deficiency on LPS-induced signaling, because SHIP1 acts downstream of PI3K.
A prominent counter-player of PI3K signaling in hemopoietic cells is the lipid phosphatase SHIP1 (52). Provided that PI3K plays a negative regulatory role in TLR signaling in MΦs, SHIP1 deficiency should result in enhanced PKB phosphorylation and reduced p38 MAPK activation, coupled with decreased TLR-mediated cytokine production. This was found in our study in BMMΦs and in an earlier study by Fang et al. (36). Moreover, low-dose glucocorticoid conditioning of myeloid progenitors resulted in MΦs with stronger SHIP1 expression paired with enhanced TLR4-induced cytokine production (37, 53). In addition to the importance of the PI3K product, phosphatidylinositol-3,4,5-trisphosphate, in attenuating LPS-triggered MΦ activation, Cao et al. (54) demonstrated a positive regulatory role for the inositol 3-phosphatase PTEN. It should be stressed at this point that an inhibitory role of SHIP1 in LPS signaling in MΦs (RAW264.7 and BMMΦs) has also been reported by two groups (13, 49, 55). The reasons for these discrepancies are not clear yet; however, they might imply differences in cell types as well as differentiation and culture conditions.
Intriguingly, both Lyn and SHIP1 negatively regulate PKB activation in human monocytes stimulated with M-CSF (56). In addition, in mast cells activated via the high-affinity receptor for IgE, both Lyn and SHIP1 suppress PKB activation (19, 57). Interestingly, in this study suppression of PKB activation by SHIP1, but not by Lyn, was observable in TLR-stimulated BMMΦs. On the contrary, Lyn was required for PKB phosphorylation in LPS-stimulated BMMΦs (Fig. 6). The reasons for this behavior await further clarification.
Several tyrosine kinases long known for their implications in signal transduction downstream of ITAM-containing receptors, such as Fc and B cell Ag receptors, have been demonstrated to be involved in TLR-mediated cell activation. Thus far, Lyn seems unique by exerting the described negative regulatory function. In contrast, the tyrosine kinases Syk and Btk have been described to play positive roles in monocytes/MΦs stimulated with microbial components. Syk interacts with and phosphorylates TLR4 and controls production of IL-10 and IL-12p40 (58), whereas Btk, by positively regulating activation of p38 MAPK, promotes secretion of TNF-α (59, 60).
Lyn-deficient mice have been reported to have circulating autoreactive Abs and an increased susceptibility to severe glomerulonephritis caused by deposition of IgG immune complexes in the kidney (20, 61). In this regard, it is interesting that enforced TLR4 signaling in mice overexpressing TLR4, or expressing gp96 on cell surfaces was found to result in the production of anti-dsDNA Abs and the immune complex-mediated glomerulonephritis (62). Lyn−/− mice also show signs of perturbed myelopoiesis and erythropoiesis, succumb to a severe asthma-like syndrome, and, of great interest, develop monocyte/MΦ tumors (63–65). Because tumor growth is a multifactorial process that, in addition to mutations in oncogenes and tumor suppressor genes, requires specific conditions that provide a supportive physiologic environment (66), enhanced production of tumor-promoting cytokines (e.g., IL-6 and TNF-α) in response to microbial stimuli might promote neoplastic growth in Lyn−/− animals.
We thank Kerstin Fehrenbach, Tanja Nöcker, Nadja Goos, and Petra Lüderitz for expert technical assistance, and Drs. G. Krystal, M.L. Hibbs, and V.L. Tybulewicz for providing SHIP1- and Lyn-deficient mice.
Disclosures The authors have no financial conflicts of interest.
Abbreviations used in this paper:
- bone marrow-derived macrophages
- LPS binding protein
- Src family kinase
- wild type.
- Received May 7, 2009.
- Accepted March 5, 2010.
- Copyright © 2010 by The American Association of Immunologists, Inc.