Comparison of the inflammatory cytokine profile in bone marrow-derived macrophages (BMDMs) from normal and Src homology domain 2-containing tyrosine phosphatase-1 (SHP-1)–deficient Motheaten (me/me) mice revealed a dramatic suppression of IL-6 transcript and protein in me/me BMDMs after LPS stimulation. Interfering with SHP-1 expression using antisense SHP-1 oligonucleotides led to a significant downregulation of IL-6 in normal BMDMs. Conversely, reconstitution of me/me BMDMs with the SHP-1 gene using adenoviral vectors restored IL-6 production. Expression of only SHP-1 Src homology region 2 domains in normal BMDMs inhibited IL-6 production, confirming that IL-6 regulation depends on SHP-1 phosphatase activity. We further demonstrated that loss of SHP-1 function affects proper phosphorylation of Erk1/2 MAPKs and, to a lesser degree, of NF-κB downstream of TLR4 in BMDMs. Inefficient phosphorylation of Erk1/2 MAPKs abrogated the activation of C/EBPβ transcription factor, which was reversed on restoration of SHP-1 function and led to a concomitant enhancement of IL-6 production. We demonstrate that IL-6 production is regulated by a complex network of signaling pathways that include SHP-1–dependent activation of Erk1/2–C/EBPβ and NF-κB, in addition to SHP-1–independent IκB pathway through the activation of protein tyrosine kinases downstream of TLR4. Taken together, these results revealed for the first time, to our knowledge, a positive and critical role of SHP-1 in IL-6 regulation and dependence of Erk1/2–C/EBPβ pathway in addition to that of IκB on SHP-1 activity required for IL-6 induction after LPS stimulation.
Interleukin-6, a B cell-stimulating factor, is capable of inducing the proliferation and differentiation of B and T cells (1–4). It is produced by monocytes/macrophages, T cells, B cells, endothelial cells, and glial cells in response to IL-1, TNF-α, platelet-derived growth factor, and microbial components such as LPS (5, 6). IL-6 is a multifunctional cytokine that regulates the immune response, hematopoiesis, bone remodeling, the acute-phase response, and inflammation (2). IL-6 has also been shown to play a key role in the development of autoimmune diseases (7). Deregulation of IL-6 production underlies the pathology of multiple myeloma, rheumatoid arthritis, Castleman’s disease, psoriasis, Alzheimer’s disease and postmenopausal osteoporosis (8). Therefore, understanding IL-6 regulation and characterizing the signal transduction events particularly after its induction by LPS may lead to the development of strategies for the treatment of autoimmune diseases and cancer.
LPS binds to TLR4 on the surface of macrophages triggering the activation of signal transduction pathways including a well-characterized MyD88/IL-1R–associated serine/threonine kinase/TNFR-associated factor 6 pathway in addition to IκB and MAPK pathways (9, 10). The MAPK pathway has been implicated in processes regulating cell growth, differentiation, apoptosis, and inflammation (11, 12). In mammals, there are three MAPKs subgroups, namely, the ERKs (Erk1/2), the JNK/stress-activated protein kinases, and the p38 MAPKs (11). Generally, Erk1/2 MAPKs respond to mitogens and growth factors that regulate cell proliferation and differentiation, whereas JNK and p38 MAPKs are activated by environmental stresses, such as UV radiation, heat shock, and proinflammatory cytokines such as IL-1 and TNF-α (11, 13, 14). LPS-activated JNK and p38 were shown to regulate the production of TNF-α, IL-1β, and IL-6 in murine macrophages (15, 16). LPS-activated NF-κB has also been shown to be critically involved in the transcriptional regulation of the IL-6, TNF-α, and IL-1 genes in macrophages (17). In LPS-stimulated macrophages, C/EBP has also been implicated in IL-6 transcription (18).
There is evidence suggesting the involvement of protein tyrosine kinases (PTKs) in LPS induction of proinflammatory cytokines including IL-6 in murine macrophages (19). However, the precise role of PTKs in the regulation of IL-6 production by LPS-stimulated macrophages remains unknown. The level of tyrosine phosphorylation is controlled by opposing activities of PTKs and protein tyrosine phosphatases (PTPs) (20). PTPs have been described as both positive and negative regulators of signaling processes (21). For example, the cytosolic Src homology region 2 (SH2) domain-containing PTP, Src homology domain 2-containing tyrosine phosphatase-1 (SHP-1), is a negative regulator of cellular responses initiated by CSF-1, IL-3, c-Kit, erythropoietin, and IFN-α/β in hematopoietic cells and IFN-γ in astrocytes (22–29). In contrast, SHP-1 positively regulates epidermal growth fac-tor and IFN-γ signaling in malignant epithelial cells (30). The critical involvement of SHP-1 in the negative regulation of normal hematopoiesis is best illustrated by severe immunodeficiency and vast expansion of myeloid cell lineages in SHP-1–deficient me/me mice (31, 32). Me/me mice manifest autoimmune disorders with an increased CD5+/CD5− B cell ratio, a phenotype that correlates with autoantibody production (33). This high expression of autoantibodies could be a consequence of deregulated production of proinflammatory cytokines such as IL-6.
In one study, SHP-1 was shown to negatively regulate IL-6, IL-10, TNF-α, and IFN-γ production in T and B cells and in sera from mev mice (34). However, a role for SHP-1 in the regulation of IL-6 production in LPS-stimulated murine macrophages is yet to be examined. In this study, using normal and me/me mice as an experimental model, we show for the first time, to our knowledge, that IL-6 production in LPS-stimulated bone marrow-derived macrophages (BMDMs) is positively regulated by two distinct SHP-1–dependent C/EBPβ and NF-κB pathways and an SHP-1–independent NF-κB pathway through the activation of PTKs.
Materials and Methods
C3HeB/FeJ (B6) Motheaten (me/me) and their phenotypically normal littermates (The Jackson Laboratory, Bar Harbor, ME) were used. The mice were of mixed sex, and their ages at the time of sacrifice were between 10 and 12 d. They were genotyped using PCR amplification of tail DNA (35). The breeding and handling of mice were approved by Animal Care and Use Committee, Health Canada (Ottawa, ON, Canada).
Cell isolation, cell culture, reagents, and Abs
36). Rabbit anti–β-actin Ab was obtained from Sigma-Aldrich. HRP-conjugated goat anti-mouse IgG or goat anti-rabbit IgG were supplied by Bio-Rad (Richmond, CA).5 M 2-ME (Sigma-Aldrich, St. Louis, MO), and 20% L929 cell-condition medium (LCM) as a source of M-CSF. The pharmacological inhibitors, SB202190 (p38 MAPK inhibitor), PD98059 (Erk1/2 MAPK inhibitor), SP600125 (JNK MAPK inhibitor), Herbimycin A (PTK inhibitor), and caffeic acid phenethyl ester (CAPE; NF-κB inhibitor) were obtained from Calbiochem (San Diego, CA). All other chemicals for Western blotting and gel shifts were obtained from Sigma-Aldrich. LPS derived from Escherichia coli
Cell stimulation and quantitation of cytokine production
The bone marrow (5 × 106 cells/ml) cells from normal and me/me mice were cultured at 37°C and allowed to adhere to the 96-well culture plate. Adherent cells were either unstimulated or stimulated with 10 ng/ml of either LPS or TNF-α for different periods. The purity of BMDMs in culture was >95% as confirmed by FACS analysis (data not shown). The levels of TNF-α, IL-1β, and IL-6 in the culture supernatants were measured by commercial ELISA kit (Biosource International, Camarillo, CA) as recommended by the manufacturers. To determine the involvement of p38, Erk1/2, JNK, NF-κB, and PTK in LPS-induced IL-6 production, we pretreated BMDMs (5 × 106 cells/ml) with specific inhibitors for 2 h. The cells were then unstimulated or stimulated with LPS (10 ng/ml) for 24 h, and culture supernatants were collected and analyzed for IL-6 production by ELISA. For supershift analyses to identify NF-κB and C/EBPβ, we used specific mouse anti–NF-κB p50 and p65 mAbs (Santa Cruz Biotechnology) and rabbit anti-C/EBPβ polyclonal Abs (New England Biolabs), respectively.
Phosphorothioate-modified SHP-1, SHP-2, or CD4 antisense oligonucleotides were obtained from University Core DNA and Protein Services (University of Calgary, Calgary, AB, Canada). The oligonucleotide sequences were as follows: SHP-1, 5′-AAA CCA CCT CAC CAT-3′; SHP-2, 5′-TCT CCG CGA TGT CAT-3′; and control, CD4, 5′-AGG GAC TCC CCG GTT CAT-3′. Normal BMDMs (1 × 106/ml) were treated with SHP-1, SHP-2, or CD4 antisense oligonucleotides (15–30 μm) for 4 h. Cells were then stimulated with LPS (10 ng/ml) for 24 h, after which culture supernatants were collected and analyzed for IL-6 production. In parallel, normal BMDMs (2 × 106/ml) were pretreated with SHP-1 antisense oligonucleotides (15–30 μM) for 4 h followed by LPS stimulation for 15 min. The BMDMs lysates were then subjected to Western immunoblot analysis using either anti–SHP-1 or anti–β-actin Abs.
Synthetic, single-stranded, HPLC-purified, phosphorothioate oligodeoxynucleotides (ODNs) specific for C/EBP, together with matching mutated ODN as a control, were obtained from GeneDetect (Bradenton, FL). They were supplied as double stranded, and the efficiency of hybridization was 98% when tested by the company. The sequences of the double-stranded ODNs were as follows (underlined letters denote phosphorothioate bonded bases, bold letters denote the core binding sequence for the transcription factor, and italic letters denote the mutated bases):
C/EBP (consensus), 5′-TGCAGATTGCGCAATCTGCA-3′ and 3′-ACGTCTAACGCGTTAGACGT-5′; and C/EBP (mutant), 5′-TGCAGAGACTAGTCTCTGCA-3′ and 3′-ACGTCTCTGATCAGAGACGT-5′.
BMDMs (1 × 106
RNA isolation and semiquantitative RT-PCR
Semiquantitative PCR was performed using 5 μl cDNA, 10X PCR buffer (500 mM KCl, 25 mM MgCl2Taq DNA polymerase (Roche Diagnostics, Mannheim, Germany) in a final volume of 50 μl. The oligonucleotide primer sequences for mouse TNF-α, IL-1β, IL-6, and β-actin (Stratagene, La Jolla, CA) are as follows: sense strand TNF-α primer, 5′-ATG AGC ACA GAA AGC ATG ATC-3′; antisense strand TNF-α primer, 5′-TAC AGG CTT GTC ACT CGA ATT-3′; sense strand IL-1β primer, 5′-CAG GAT GAG GAC ATG AGC ACC-3′; antisense strand IL-1β primer, 5′-CTC TGC AGA CTC AAA CTC CAC-3′; sense strand IL-6 primer, 5′-GAC AAA GCC AGA GTC CTT CAG AGA G-3′; antisense strand IL-6 primer, 5′-CTA GGT TTG CCG AGT AGA TCT C-3′; sense strand β-actin primer, 5′- TGT GAT GGT GGG AAT GGG TCA G-3′; antisense strand β-actin primer, 5′-TTT GAT GTC ACG CAC GAT TTC C-3′. The PCR was carried out in a PTC-100 programmable thermal controller (MJ Research, Watertown, MA). The conditions for amplification are as follows: IL-1β, TNF-α, and β-actin: denaturation at 94°C for 45 s, annealing at 60°C for 45 s, extension at 72°C for 1:30 min 25 cycles (IL-1β, β-actin) or 30 cycles (TNF-α) and a final extension at 72°C for 10 min; IL-6: denaturation at 94°C for 1 min, annealing at 55°C for 1 min, extension at 72°C for 1 min 28 cycles, and a final extension at 72°C for 7 min. The amplified products for TNF-α (276 bp), IL-1β (447 bp), IL-6 (229 bp), and β-actin (514 bp) were resolved by 1.2% PAGE. Quantitation was performed by densitometry.
Western blot analysis
Phosphorylation of p38, Erk1/2, JNK, and IκB-α was determined by Western blot analysis according to the protocol described previously (36).
Nuclear extracts for EMSAs were prepared as previously described (37) and carried out using a nonradioactive Light Shift chemiluminescent EMSA kit (Pierce, Rockford, IL) according to the manufacturer’s recommendation. The oligonucleotide sequences corresponding to the NF-κB and C/EBP binding sites in the mouse IL-6 promoter were as follows: NF-κB, 5′-CAA ATG TGG GAT TTT CCC ATG A-3′ and 5′-TCA TGG GAA AAT CCC ACA TTT G-3′; C/EBP, 5′-CTA AAC GAC GTC ACA TTG TGC AAT CTT AAT AAG GTT-3′ and 5′-T GGA AAC CTT ATT AAG ATT GCA CAA TGT GAC GTC GTT TAG-3′. The mutant oligonucleotide sequences used as cold competitors corresponding to the NF-κB or C/EBPβ binding sites were as follows: NF-κB (m), 5′-CAA ATG TGG GAT TTT aga cTG A-3′ and 5′-TCA gtc tAA AAT CCC ACA TTT G-3′; C/EBPβ (m), 5′-CTA AAC GAC GTC ACA gat atc AAT CTT AAT AAG GTT-3′ and 5′-T GGA AAC CTT ATT AAG ATT gat atc TGT GAC GTC GTT TAG-3′.
All experiments were repeated at least two to three times; however, data from one representative experiment showing similar trends are depicted in the Results. In each experiment, all samples were set up in triplicate, and the data shown represent the mean ± SE. Comparisons between two groups were made using Student t test. Comparisons between more than two groups were made using one-way ANOVA. A p value < 0.05 is regarded as statistically significant.
Deregulated production of proinflammatory cytokines in BMDM cultures of SHP-1 null me/me mice
Homozygous SHP-1 null Motheaten (me/me) mice experience development of systemic autoimmunity and alopecia as a result of severe inflammation (31). To determine the involvement of SHP-1 in inflammatory responses, we first compared the profile of proinflammatory cytokines TNF-α, IL-1β, and IL-6 in LPS-stimulated normal and me/me BMDMs. The results show 2- to 5-fold greater levels of TNF-α production in unstimulated and LPS-stimulated me/me BMDMs compared with the normal controls. However, IL-1β production by BMDMs of normal and me/me mice in either LPS- stimulated or unstimulated cultures was not found to be significantly different. In contrast, IL-6 production after LPS stimulation was markedly diminished by 95% in me/me BMDMs compared with the controls (Fig. 1A). Similar results were obtained at transcriptional levels. Semiquantitative RT-PCR analysis revealed enhanced levels of TNF-α mRNA and unaltered levels of IL-1β mRNA, whereas IL-6 mRNA in LPS-stimulated me/me BMDMs was significantly decreased compared with the controls (Fig. 1B).
We further determined the kinetics from 0–96 h of LPS-induced IL-6 production in both normal and me/me BMDMs. Our results confirmed that normal BMDMs were capable of producing >20,000 pg/ml IL-6 protein between 72 and 96 h after LPS stimulation. In contrast, IL-6 production in me/me BMDMs attained a peak of 1100 pg/ml between 24 and 48 h, after which the IL-6 production was not sustained and declined to basal levels (Fig. 2A). Furthermore, suppression of IL-6 was observed irrespective of LPS dose in me/me BMDMs (Fig. 2B). These results suggest that SHP-1 positively regulates cytoplasmic signaling pathways downstream of LPS/TLR4 that culminate in IL-6 production.
Because our results showed that me/me BMDMs produce large quantities of TNF-α on LPS stimulation, it was therefore of interest to examine whether endogenously produced TNF-α plays a role in IL-6 suppression. For this, we interfered with TNF-α signaling by treating normal and me/me BMDMs with either anti–TNF-α R1(Fig. 3A) or anti–TNF-α Abs (Fig. 3B) before LPS stimulation (38). The results show that the treatment of BMDMs with anti–TNF-α R1 (7 μg/ml) or anti–TNF-α (1 μg/ml) Abs did not affect LPS-induced IL-6 production in both normal and me/me BMDMs (Fig. 3A). As a control, we demonstrated that rTNF-α (10 ng/ml) induces only a modest level of IL-6 (∼1000 pg/ml) in normal BMDMs (Fig. 3A, 3B, left panels). We further showed that this amount of IL-6 could be effectively neutralized by either anti–TNF-α R1 or anti–TNF-α Abs, respectively, in BMDMs, thus confirming their specific effects (Fig. 3A, 3B, left panels). These results suggest that the endogenously produced TNF-α does not modulate LPS-induced IL-6 production in BMDMs.
Crucial role of SHP-1 phosphatase activity in LPS-induced IL-6 production
To understand the role of SHP-1 in LPS-induced IL-6 expression, we mimicked the me/me phenotype by treating normal BMDMs with SHP-1 antisense oligonucleotides (39). The results revealed that antisense SHP-1 oligonucleotides significantly reduced IL-6 production in a dose-dependent manner. This suppression was not observed when normal BMDMs were treated with control antisense oligonucleotides to either a closely related tyrosine phosphatase SHP-2 or unrelated antisense CD4 oligonucleotides, thus confirming the specificity of this effect (Fig. 4A). The expression of SHP-1 was markedly decreased to the basal levels in BMDMs treated with antisense SHP-1oligonucleotides in a dose-dependent manner as determined by Western blot analysis (Fig. 4B).
To further confirm the critical role of SHP-1 in LPS-induced IL-6 production, we reconstituted me/me BMDMs with SHP-1 by transduction with adenovirus vector carrying the full-length SHP-1 gene (Ad-SHP-1). Reconstitution of me/me BMDMs with SHP-1 gene led to almost 300% enhancement in LPS-induced IL-6 production at high multiplicity of infection (MOI) compared with cells transduced with Ad-GFP vector, and that was equivalent to 70% of normal LPS-stimulated BMDMs (Fig. 5A, left panel). We observed dose-dependent increase in IL-6 production when me/me BMDMs were transduced with increasing infective doses of Ad-SHP-1, from MOI 8–100 (Fig. 5A, left panel). In contrast, reconstitution of me/me BMDMs with Ad-GFP alone did not induce any IL-6, and further stimulation of these cells with LPS resulted in only basal levels of IL-6 similar to those produced by mock-infected LPS-stimulated me/me BMDMs (Fig. 5A, left panel). The expression of SHP-1 in me/me BMDMs transduced with the full-length SHP-1 gene was confirmed by Western blot analysis, and the amount of SHP-1 increased in a MOI-dependent manner, reaching almost normal levels at the highest MOI (Fig. 5A, right panel).
The role of SHP-1 in IL-6 production was further demonstrated by experiments in which normal BMDMs were transduced with adenovirus vector carrying a dominant negative SHP-1 gene lacking the catalytic activity (SHP-1 DN SH2). Normal BMDMs when transduced with increasing infective doses of this construct showed reduced IL-6 production similar to that seen in SHP-1–deficient BMDMs (Fig. 5B, left panel). In three experiments, the expression of the SHP-1 DN SH2 construct (at low MOI) resulted in reduction of IL-6 by >40%. The expression of DN SHP-1 in these cells was shown by Western blot analysis to increase in an MOI-dependent manner (Fig. 5B, right panel). These results suggested that the SHP-1 activity was mandatory for efficient LPS-induced IL-6 production.
SHP-1 regulates the activation of Erk1/2 MAPKs, which are necessary for efficient LPS-induced IL-6 production
The MAPK pathway has been shown to be activated by LPS and implicated in cytokine production (12, 37, 40). Furthermore, SHP-1 has also been implicated in the regulation of the MAPK pathway in human neutrophils, brain pericytes, and the renal cortex of diabetic mice (41, 42). Therefore, to examine the involvement of MAPKs pathway in LPS-induced IL-6 production, we compared the activation of MAPK family members in LPS-stimulated BMDMs derived from me/me and normal mice. LPS stimulation of normal BMDMs induced intense and prolonged phosphorylation of JNK and Erk1/2 MAPKs, whereas activation of these kinases in me/me BMDMs was transient. Specifically, Erk1/2 was transiently phosphorylated only at 15 and 20 min in me/me BMDMs compared with the intense phosphorylation observed from 5–40 min post-LPS stimulation in normal BMDMs. In contrast, the activation of p38 MAPKs remained the same in BMDMs from both normal and me/me mice showing slight induction at 5 min and lasting up to 40 min post-LPS stimulation (Fig. 6A).
To determine the role of each MAPK in LPS-induced IL-6 production, we used pharmacological inhibitors specific to distinct members of MAPK family: SB202190, PD98059, and SP600125 for p38, Erk1/2, and JNK MAPK, respectively. LPS activation of JNK and p38 MAPKs in either me/me or normal BMDMs did not affect IL-6 production in either cell type (Fig. 6B). However, PD98059 at 10-μM concentration inhibited IL-6 production by almost 70% in normal BMDMs (Fig. 6B, left panel). This was in sharp contrast with the effect of PD98059 on LPS induced IL-6 production in me/me BMDMs where IL-6 production was not further inhibited beyond the already reduced amount (Fig. 6B, right panel). The biological activity of SB202190, PD98059, and SP600125 was confirmed by their ability to affect the phosphorylation of p38, Erk1/2, and JNK MAPK in a dose-dependent manner (Fig. 6C). These results suggest that LPS-induced IL-6 production may be regulated by two distinct Erk-dependent and -independent pathways in normal BMDMs, whereas in me/me mice, IL-6 production may be regulated only by a SHP-1– and Erk-independent pathway.
This differential effect of PD98059 on LPS-induced IL-6 production in normal and me/me BMDMs suggested a role for SHP-1 in Erk1/2 activation. This notion was confirmed by treatment of normal BMDMs with antisense SHP-1 oligonucleotides showing a dose-dependent inhibition of LPS-induced Erk1/2 phosphorylation (Fig. 6D). Because me/me mice lack SHP-1, as expected treatment of me/me BMDMs with antisense SHP-1 oligonucleotides or with unspecific antisense CD4 oligonucleotides did not have any effect on the levels of LPS-induced Erk1/2 phosphorylation (data not shown). These results suggest that IL-6 production in response to LPS occurs, at least in part, through Erk1/2 signaling and requires SHP-1. When SHP-1 is missing, a small amount of IL-6 is still produced but in an SHP-1– and Erk-independent manner.
LPS-induced IL-6 production is regulated by NF-KB in both normal and me/me BMDMs
NF-κB is a critical transcription factor involved in the regulation of a number of cytokines including IL-6 after stimulation with LPS, TNF-α, or IL-1β (43, 44). To investigate the role of SHP-1 in IL-6 gene regulation and NF-κB activation, we examined the effect of CAPE, an NF-κB–specific inhibitor on IL-6 production. As expected, CAPE at a low concentration of 1 μM significantly inhibited LPS-induced IL-6 production by 70% in normal BMDMs and further reduced the already low amount of IL-6 produced by me/me BMDMs (Fig. 7A). To determine the effect of CAPE on the binding of NF-κB to the IL-6 promoter-derived oligonucleotides, normal BMDMs were pretreated with CAPE and the nuclear extracts were subjected to EMSA. The addition of CAPE completely abrogated the binding of NF-κB to IL-6 promoter (Fig. 7B) strongly suggesting a role for NF-κB in LPS-induced IL-6 production in normal BMDMs.
SHP-1 regulates NF-κB activation and IκB phosphorylation in LPS-stimulated BMDMs
To get further insights into any differences in the activation of NF-κB in normal and me/me BMDMs, we examined LPS-induced IκB phosphorylation. Although a much higher and sustained phosphorylation of IκB was observed in normal BMDMs, LPS-induced phosphorylation of IκB was rapid but was poorly sustained beyond 10 min in me/me BMDMs (Fig. 8A). These results were further supported by measuring the DNA binding activity of NF-κB from LPS-stimulated nuclear extracts to biotin-labeled oligonucleotides encoding NF-κB sequences derived from the murine IL-6 promoter. Stimulation of BMDMs with LPS resulted in substantially higher DNA binding activity of NF-κB in normal compared with me/me BMDMs (Fig. 8D, lane 1 versus 6), suggesting a requirement of SHP-1 activity for efficient binding of NF-κB to IL-6 promoter and full induction of the IL-6 gene.
To further assess the role of SHP-1 in LPS-induced NF-κB activation and IL-6 production, we reconstituted me/me BMDMs with wild type (WT) SHP-1 gene and examined NF-κB binding to its oligonucleotides derived from the murine IL-6 promoter. The expression of SHP-1 gene in me/me BMDMs resulted in increased NF-κB binding activity in an MOI-dependent manner, reaching levels close to that seen in normal BMDMs (Fig. 8B, left panel, lanes 3–5 versus 6). In parallel, transduction of normal BMDMs with the DN mutant of SHP-1 (DN SHP-1) reduced NF-κB–binding activity as seen in me/me BMDMs (Fig. 8B, left panel, lanes 7, 8 versus 2). As a control for specificity, transduction of me/me and normal BMDMs with empty vector (Ad-GFP) did not have any effect on NF-κB binding activity (Fig. 8B, left panel, lanes 9, 10). The specificity of NF-κB bands was ascertained by supershift analysis using either anti-p50 (NF-κB 1) or anti-p65 (Rel A) Abs (Fig. 8B, right panel, lanes 2–5). Furthermore, the shifted band disappeared when either unlabeled (200-fold) or mutant oligonucleotides were used (Fig. 8B, right panel, lanes 7, 8). Overall, the results suggest a role for SHP-1 in NF-κB activation and IL-6 production in BMDMs. Moreover, suppression of IL-6 production in LPS-stimulated me/me BMDMs may be a consequence of inefficient NF-κB activation possibly because of an SHP-1 defect upstream of NF-κB. These results suggest a critical role for SHP-1 in LPS-induced NF-κB activation and IL-6 production.
LPS-induced NF-κB activation is not regulated by the Erk1/2 MAPKs
In light of our findings that both Erk1/2 MAPKs and IκB were required for IL-6 production in LPS-stimulated normal BMDMs, we determined whether Erk1/2 contributed to IκB activation by examining the effect of PD98059 on LPS-induced IκB phosphorylation and NF-κB binding to the IL-6 promoter–derived oligonucleotides. Our results show that PD98059 did not influence the level of IκB phosphorylation (Fig. 8C) or the binding of NF-κB to its oligonucleotides derived from murine IL-6 promoter in LPS-stimulated either normal or me/me BMDMs (Fig. 8D).
PTKs are involved in the regulation of LPS-induced NF-κB and Erk1/2 activation and IL-6 production in BMDMs
In view of the earlier observations and the fact that SHP-1 causes tyrosine dephosphorylation of its targets (22–29, 45), we hypothesized that SHP-1 may activate Erk1/2 or NF-κB through an intermediary of SHP-1 target. Because PTKs are SHP-1 targets in hematopoietic cells (22–29, 46), we determined whether PTKs play a role in the residual IL-6 production in me/me BMDMs. Therefore, we examined PTKs’ role in LPS-induced IL-6 regulation by using Herbimycin A, a PTK inhibitor (47–50). The results show that even very low concentrations (1 μM) of Herbimycin A markedly inhibited LPS-induced IL-6 production in both normal and me/me BMDMs (Fig. 9A). To establish a link between the potential role of PTKs in the activation of either NF-κB or Erk1/2 MAPKs pathways, we treated normal and me/me BMDMs with Herbimycin A for 2 h before stimulation with LPS for 15 min. The results show that LPS-induced phosphorylation of IκB (Fig. 9B) and Erk1/2 MAPKs (Fig. 9C) were markedly inhibited by Herbimycin A in both normal and me/me BMDMs. These results suggest that PTK activity is critical for SHP-1–mediated NF-κB activation and Erk1/2 phosphorylation after LPS stimulation of normal BMDMs. However, the identity of the PTKs involved remains unknown.
SHP-1 regulates LPS-induced IL-6 production via C/EBP activation through PTK and Erk1/2 phosphorylation in normal BMDMs
The earlier results suggest that in normal BMDMs, LPS-induced IL-6 production is regulated by two distinct signal transduction pathways, one of which is Erk dependent, NF-κB dependent, and the second one constitutes Erk-independent activation of NF-κB. To analyze a potential interaction or cooperation between these two pathways, we searched for potential transcription factors activated by ERK1/2. Studies with transcriptional regulation of IL-6 have identified a C/EBP site immediately upstream of the NF-κB site, and cooperation between the two elements created optimal IL-6 gene activation in the presence of stimuli that induced NF-κB and C/EBP (51–53). Therefore, we first examined whether LPS can induce the activation of C/EBP in normal and me/me BMDMs. The results show that significant binding of C/EBP to the C/EBP oligonucleotides occurred between 5 and 20 min post-LPS stimulation in normal BMDMs. In contrast, me/me BMDMs showed insignificant binding of C/EBP to the C/EBP oligonucleotides (Fig. 10A). The binding of C/EBP was completely blocked by competition with cold C/EBP oligonucleotides in both types of cells (Fig. 10A). Furthermore, IL-6 promoter-derived oligonucleotides bearing mutations in the C/EBP binding site failed to bind LPS-activated C/EBP, thus demonstrating the specificity of the binding reaction (Fig. 10A).
To examine whether LPS-activated C/EBP regulates increased levels of IL-6, we used a C/EBP decoy to inhibit binding of C/EBP to its cognate DNA sequences in BMDMs. The results showed that the levels of IL-6 were significantly suppressed on treatment with WT decoy, whereas the mutant control had little or no effect in normal BMDMs (Fig. 10B). Furthermore, treatment with WT, but not with the mutant decoy, resulted in 90% decrease in binding of C/EBP to the labeled C/EBP probe (Fig. 10C).
We further show that C/EBP binding to IL-6 promoter was sensitive to Erk1/2 and PTK inhibition. PD98059 (Fig. 11A) and Herbimycin (Fig. 11B) at very low concentrations significantly inhibited the binding of C/EBP to the C/EBP probe in LPS-stimulated normal BMDMs (Fig. 11A, 11B). Herbimycin also inhibited the residual IL-6 production in me/me BMDMs (data not shown). To further determine the dependence of C/EBP activation on SHP-1 expression, we reconstituted me/me BMDMs with the WT SHP-1 gene. The reconstitution of SHP-1 expression in me/me BMDMs rescued LPS-induced activation of C/EBP in a transduction dose-dependent manner, reaching the levels close to those observed in normal BMDMs (Fig. 11C, lanes 3–5 versus 9). The activation and identity of C/EBP was confirmed by supershift analyses using specific rabbit anti-C/EBPβ polyclonal Ab (Fig. 11C, lanes 3–5, 9). The specificity of supershifted protein–DNA complexes were also assessed by competition assays with 200-fold unlabeled or mutated oligonucleotide to C/EBP binding site (Fig. 11C, lanes 7, 8). In parallel, when normal BMDMs were transduced with DN mutant of SHP-1 lacking catalytic activity, the activation of C/EBP was completely abolished (Fig. 11C, lanes 10, 11), thus confirming the role of SHP-1 in C/EBP regulation. These results clearly suggest that SHP-1 is a critical signaling molecule necessary for proper and effective regulation of C/EBP, as well as NF-κB, leading to LPS-induced IL-6 production in BMDMs.
IL-6 is induced in response to a variety of mediators of inflammation such as TNF-α, IL-1, and PGDF, and viral and bacterial infections (2, 8, 54). During the last several years, Erk1/2 and p38 MAPKs and NF-κB have been implicated in the regulation of LPS-induced IL-6 production in murine macrophages (53). In this work, we studied the role of SHP-1 in the context of LPS/TLR4-signal transduction pathways leading to IL-6 induction by using SHP-1 null me/me BMDMs. We demonstrate that SHP-1 is a critical positive regulator of IL-6 in LPS-stimulated BMDMs. Our results suggest for the first time, to our knowledge, that IL-6 production is regulated by a complex network of pathways that include SHP-1–dependent activation of Erk/EBPβ pathway and Erk-independent activation of NF-κB transcription factor through the involvement of PTK. The absence of SHP-1 not only leads to a defect in the function of Erk1/2-dependent C/EBP pathway, but affects the integrity of the NF-κB pathway unmasking the SHP-1–independent NF-κB component of IκB pathways that may contribute to the residual level of IL-6 in me/me BMDMs (Fig. 12). Because SHP-1 acts as a negative regulator of cytokine production, including TNF-α in hematopoietic cells (39, 55), our results suggest a dual role for SHP-1 activity both as a negative and as a positive regulator in LPS/TLR4 signal transduction pathways leading to cytokine gene induction.
SHP-1 is believed to act as a negative regulator of cytokine production through activation of NF-κB in T and B cells derived from mev mice. SHP-1 deficiency in mev/mev mice led to the increased production of IL-6, IL-10, TNF-α, and INF-γ in anti-CD3–stimulated T cells or LPS-stimulated B cells (34). Our results confirmed a negative regulatory role of SHP-1 showing that loss of SHP-1 results in dramatically increased levels of TNF-α production in LPS-stimulated BMDMs from me/me mice. Interestingly, our findings provide evidence for a mandatory positive role of SHP-1 in LPS-induced IL-6 production in BMDMs, thus suggesting a dual role of SHP-1 in the regulation of LPS/TLR4-induced cytokine production. The positive role of SHP-1 was demonstrated first by showing that SHP-1 deficiency in me/me mice led to a dramatic suppression of IL-6 transcript and proteins after LPS stimulation. Moreover, suppression of SHP-1 by antisense oligonucleotides or DN-SH2-SHP-1 in normal BMDMs significantly inhibited IL-6 production, thus mimicking SHP-1−/− me/me phenotype. Conversely, reconstitution of me/me BMDMs with the SHP-1 gene using adenoviral vector restored IL-6 production, suggesting that SHP-1 phosphatase activity and possibly a tyrosine dephosphorylation step is required in SHP-1–dependent, LPS-activated signaling cascade leading to IL-6 gene induction.
Our results also suggest that SHP-1 positively regulates the activation of Erk1/2 MAPKs, C/EBPβ, and NF-κB during LPS-induced IL-6 production in normal BMDMs. LPS induced IL-6 production was highly sensitive to NF-κB inhibition in both normal and me/me BMDMs, suggesting a key role for NF-κB in IL-6 regulation. The loss of SHP-1 resulted in only transient activation of NF-κB and C/EBP in LPS-stimulated me/me BMDMs. The positive role of SHP-1 in LPS-induced NF-κB and C/EBP activation was confirmed by reintroducing SHP-1 gene into me/me BMDMs, as well as by inhibiting SHP-1 activity by DN SHP-1 in normal BMDMs. These observations parallel the study by You et al. (56) showing diminished NF-κB binding activity and IκB phosphorylation in SHP-2−/− fibroblasts resulting in impaired IL-6 production after IL-1/TNF-α stimulation that could be restored by the reintroduction of SHP-2 gene into SHP-2 knockout fibroblasts.
Studies by Georganas et al. (57) identified a dominant role for NF-κB, but not C/EBPβ, in the regulation of human IL-6 expression in rheumatoid arthritis synovial fibroblasts. We defined the involvement of NF-κB in IL-6 production by demonstrating that NF-κB is activated by SHP-1–dependent and –independent mediators. We have shown that the pathway operative in SHP-1−/− me/me BMDMs producing residual levels of IL-6 in response to LPS involves, at least in part, SHP-1–independent activation of NF-κB. In me/me BMDMs, the basal production of IL-6 cannot be attributed to Erk1/2 MAPK because of its poor and transient phosphorylation in these cells. Our results suggest that the residual IL-6 produced in me/me BMDMs may be regulated by poorly activated NF-κB, possibly through PTKs. The precise mechanism by which residual levels of IL-6 are produced in LPS-activated me/me BMDMs requires further investigation.
In addition to NF-κB, we demonstrate the involvement of C/EBPβ in the regulation of LPS-induced IL-6 production in murine BMDMs. C/EBP family members (C/EBPα, C/EBPβ, C/EBPγ, C/EBPδ, and C/EBPε), bind to the same DNA consensus sequence and have a conserved region at the C terminus, containing the basic amino acid DNA binding motif followed by a leucine zipper, bZIP, needed for dimerization. The activation domains responsible for transcriptional activation and/or repression are located in the N termini. Previous studies identified the presence of the binding sites for C/EBPβ and C/EBPδ in the promoter of the IL-6 gene. It was also shown that both C/EBPβ and C/EBPδ are activated in LPS-stimulated BMDMs (58). Furthermore, Erk1/2 was shown to activate C/EBP in response to IFN-γ in RAW 264.7 macrophages (59).
Our results demonstrate a critical role played by the Erk1/2–C/EBPβ pathway in addition to that of NF-κB in efficient SHP-1–dependent induction of IL-6 in LPS-stimulated BMDMs. The C/EBP transcription factors have been shown to play a role in the regulation of cellular growth, differentiation, and immune and inflammatory processes (60). In view of our data demonstrating the existence of an Erk1/2-independent NF-κB pathway regulating the induction of the IL-6 gene in LPS-stimulated BMDMs, we examined the activation of C/EBPβ transcription factor through Erk1/2 MAPKs. We demonstrate for the first time, to our knowledge, a critical involvement of SHP-1 function in the regulation of LPS-induced IL-6 production through activation of the Erk1/2–C/EBPβ pathway. Inhibition of SHP-1 in normal BMDMs with antisense SHP-1 oligonucleotides resulted in the loss of Erk1/2 activation, whereas reintroduction of SHP-1 gene into me/me BMDMs enhanced the activation of C/EBP with concomitant induction of IL-6 production. Furthermore, LPS-induced phosphorylation of Erk1/2 MAPKs was downregulated in me/me BMDMs, which prevented efficient activation of C/EBPβ and concomitantly led to reduced levels of LPS-induced IL-6 production.
In an effort to understand how SHP-1 may affect the regulation of LPS/TLR4-activated signal transduction pathways, we examined the involvement of PTKs, which constitute SHP-1 targets and are localized in close proximity to the receptor. Because treatment of normal and me/me BMDMs with PTK-specific inhibitor Herbimycin A inhibited LPS-induced IL-6 production, the results suggest a requirement of an SHP-1–dependent dephosphorylation event in the activation of PTKs regulating Erk1/2, C/EBP, and NF-κB signaling proteins downstream of TLR4, leading to IL-6 gene induction in normal BMDMs. These observations suggest a proximal function of SHP-1 in conjunction with PTKs downstream of TLR4. LPS-induced PTKs have been shown to play a role in the induction of IL-6, IL-1β and TNF-α, as well as NF-κB activation in monocytes (61–63). The Src family PTKs, Hck, Fgr, and Lyn, are expressed in macrophages and have been implicated in LPS/TLR4 signaling pathways (64–66). Although our results implicate the involvement of PTKs in IL-6 regulation, the identity of specific PTKs involved and their activation by SHP-1 downstream of LPS/TLR4 signaling and specifically those leading to IL-6 gene transcription remain to be investigated. Currently, it is not known how SHP-1 regulates MAPK and IκB signal transduction pathways downstream of TLR4. We envisage that SHP-1 protein may form a part of TLR4 proximal signalosome, which also includes PTKs and perhaps a component of cytoskeleton. SHP-1–mediated dephosphorylation of a tyrosine residue on a currently unknown member of the PTK family may be required for the activation of this PTK, which, in turn, may lead to MAPK pathway and C/EBP activation. Similarly, a proximal signalosome complex including SHP-1–PTK may activate IκB and C/EBP, as suggested in results shown in Fig. 9A and 9B. In fact, the Src family of PTKs has recently been shown to cooperate with Erk and JNK MAPKs downstream of polymeric Ig receptor signaling in epithelial cells (67).
Taken together, in addition to a well-established role of SHP-1 as a negative regulator of signal transduction pathways in hematopoietic cells, this study shows for the first time, to our knowledge, a positive role for SHP-1 in IL-6 expression after LPS stimulation of BMDMs. We demonstrate that IL-6 gene transcription is regulated by a complex network of signaling pathways including SHP-1–dependent Erk1/2–C/EBPβ and NF-κB pathways, in addition to SHP-1–independent IκB pathway through PTK activation after LPS stimulation of BMDMs (Fig. 12). Further studies are needed to identify PTKs involved in activation of Erk, NF-κB, and C/EBPβ, the downstream signaling molecules involved in IL-6 induction in the LPS/TLR4 signaling cascade.
The authors have no financial conflicts of interest.
We thank Animal Care Division, Health Canada, for assistance with maintaining and breeding the Motheaten mice colony. We thank Dr. Aurelia Busca for critically reading the manuscript.
This work was supported by a Biotechnology Strategy research grant from Health Canada (to M.K.).
Abbreviations used in this article:
- bone marrow-derived macrophage
- caffeic acid phenethyl ester
- dominant negative
- IκB kinase
- multiplicity of infection
- protein tyrosine kinase
- protein tyrosine phosphatase
- Src homology region 2
- Src homology domain 2-containing tyrosine phosphatase-1
- wild type.
- Received November 11, 2010.
- Accepted February 24, 2011.
- Copyright © 2011 by The American Association of Immunologists, Inc.