CD45-induced Tumor Necrosis Factor α Production in Monocytes Is Phosphatidylinositol 3-Kinase-dependent and Nuclear Factor-κB-independent*

The pro-inflammatory cytokine tumor necrosis factor (TNF)-α plays a pivotal role in the pathogenesis of rheumatoid arthritis. The mechanisms involved in regulating monocyte/macrophage TNFα production are not yet fully understood but are thought to involve both soluble factors and cell/cell contact with other cell types. Ligation of certain cell surface receptors, namely CD45, CD44, and CD58, can induce the production of TNFα in monocytes. In this paper, we investigate further the signaling pathways utilized by cell surface receptors (specifically CD45) to induce monocyte TNFα and compare the common/unique pathways involved with that of lipopolysaccharide. The results indicate that monocyte TNFα induced upon CD45 ligation or lipopolysaccharide stimulation is differentially modulated by phosphatidylinositol 3-kinase and nuclear factor-κB but similarly regulated by p38 mitogen-activated protein kinase. These results demonstrate that both common and unique signaling pathways are utilized by different stimuli for the induction of TNFα. These observations may have a major bearing on approaches to inhibiting TNFα production in disease where the cytokine has a pathogenic role.

CD45 is a membrane-anchored protein-tyrosine phosphatases found exclusively on all nucleated hemapoietic cells (24,25). The role of CD45 in T cells has been the subject of much investigation and has been shown to play an important costimulatory role in intracellular signal transduction in T lymphocytes (26 -31). While ligation of CD45 on monocytes has been shown to induce synthesis of cytokines, including TNF␣, IL-1␤, and macrophage-colony stimulating factor (M-CSF) (18,19), the signaling mechanisms involved and the functional relevance of CD45 on monocyte/macrophages remain unclear.
We have investigated the signaling pathways utilized upon CD45 ligation on monocytes leading to TNF␣ production and compared this with the conventional stimulus, LPS. We demonstrate that CD45 ligation (but not LPS) activates the phosphatidylinositol 3-kinase (PI3K) pathway and that inhibitors of PI3K activation block CD45-but not LPS-induced TNF␣ synthesis. The differences in signaling also extended to nuclear factor-B (NF-B), which, unlike LPS, was not required by CD45-induced TNF␣ synthesis. In contrast, CD45, like LPS, activated p38 MAPK.
Antibodies-Rabbit antisera to p38 MAPK was provided by Prof. J Saklatvala (Kennedy Institute of Rheumatology, London, UK) (32) and the antibody to the p85␣ subunit of PI3K was kindly provided by Dr. D. Cantrell (ICRF, London, UK). The antibody to p70 S6K was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies to phosphorylated protein kinase B (pPKB) and PKB were obtained from New England Biolabs (Hitchin, Herts, UK). Mouse IgG2a mAb HB196 (4B2 anti-CD45) and mouse isotype control IgG2a mAb OKT8 (anti-CD8) and OX12 were obtained as hybridomas from ATCC, and antibodies were subsequently purified using a protein-G Sepharose column (Millipore, * 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.
Monocyte Purification-Human peripheral blood monocytes were isolated from single donor platelet pheresis residues purchased from the North London Blood Transfusion service (Colindale, UK) as described previously (16). Briefly, mononuclear cells were isolated by Ficoll/ Hypaque centrifugation (specific density 1.077 g/ml; Nycomed Pharma A.S., Oslo, Norway), prior to cell separation in a Beckman JE6 elutriator. Monocyte purity was assessed by flow cytometry using fluorochrome-conjugated anti-CD45 and anti-CD14 mAb (Becton Dickinson, Oxford, UK) and routinely consisted of Ͼ85% CD45-or CD14-expressing cells, respectively.
Monocyte Culture-Monocytes were cultured in complete medium at 4 ϫ 10 6 cells/ml in flat-bottomed 96-well culture plates (Nunc Life Technologies Ltd., Paisley, Scotland). At the start of the culture period, cells were either left unstimulated or were cultured with the following reagents as indicated in the text: 10 ng/ml LPS, 10 g/ml immobilized anti-CD45 mAb or immobilized isotype-matched controls OX12 (IgG2a) and OKT8 (IgG2a). In some experiments monocytes were pre-treated for 15 min with wortmannin or LY294002, or for 1 h with SB203580 or rapamycin at the indicated concentrations prior to stimulation. After 18 h in culture at 37°C with 5% CO 2 , supernatants (200 l/well, 3 wells/condition) were harvested and stored at Ϫ20°C until used. All experiments were performed at least three times, and the figures show representative examples of these experiments.
Measurement of TNF␣ by Sandwich ELISA-Reagents for the TNF␣ ELISA were provided by Dr W. Buurman (Rijks Universiteit Limbury, Maastricht, The Netherlands). The ELISA was performed as described previously (33) using immobilized anti-TNF␣ mAb 61E71 and developed using a rabbit polyclonal anti-TNF␣ antibody. The rabbit polyclonal antibody was detected using a peroxidase-conjugated goat antirabbit IgG (HϩL) (Jackson Immunoresearch Laboratories, Westgrove, PA) followed by an appropriate substrate. The range of the assay was 1.6 -5000 pg/ml. Results are expressed as the mean concentration of cytokine Ϯ S.D. per condition above the minimum sensitivity of the ELISA.
Western Blotting-Western blotting for phosphorylated forms of p38 MAPK and PKB was performed according to the antibody manufacturer's instructions (New England Biolabs, Hitchin, Herts, UK).
PI3K Assay-Beads containing immunoprecipitates were washed three times in PI3K lysis buffer, once in PBS, twice in 500 mM lithium chloride, once in water and once in PI3K assay buffer ( protein kinase A inhibitor (Santa Cruz Biotechnology Inc.), and 5 Ci of [␥-32 P]ATP. The reactions were performed at 37°C for 30 min, and the products were separated by gel electrophoresis in the presence of urea. Products were visualized by autoradiography using Hyperfilm MP (Amersham Pharmacia Biotech).
Adenoviral Infection-Adenoviral infection was performed using an adenovirus encoding porcine IB␣ under the control of the cytomegalovirus promoter and a nuclear localization sequence (AdvIB␣) (36) or control adenovirus containing no insert (Adv0). Adenoviral infection of monocytes was performed as described previously (37). Briefly, freshly elutriated monocytes were cultured at 1 ϫ 10 6 /ml for 2-3 days with 100 ng/ml M-CSF. Following culture, M-CSF-treated monocytes were washed once with PBS to remove non-adherent cells and the remaining adherent monocytes were incubated with 10 ml of cell dissociation solution (Sigma, UK) for 30 -45 min to dissociate from the plastic. M-CSF-treated monocytes were resuspended to 2 ϫ 10 6 cell/ml prior to stimulation with either 10 g/ml immobilized anti-CD45 or 10 ng/ml LPS as indicated in the text for 18 h. Supernatants were harvested and assayed for TNF␣ production. Fig. 1a illustrates TNF␣ synthesis following CD45 ligation on monocytes stimulated by immobilized anti-CD45 antibody in a concentrationdependent manner. There was also synergy between CD45 ligation and stimulation with LPS (10 ng/ml), as TNF␣ production was enhanced 4 -6-fold (Fig. 1b) over that observed with LPS alone. In all experiments, immobilized isotype control antibodies did not induce TNF␣ production over that of cells alone.

CD45 Induces Monocyte TNF␣ in Peripheral Blood Monocytes in a Concentration-dependent Manner-
Inhibition of PI3K Differentially Modulates Anti-CD45-and LPS-induced Monocyte TNF␣ Production-The signaling pathways involved in monocyte TNF␣ following CD45 ligation are unknown. In contrast, signaling pathways involved in LPSinduced TNF␣ production have received much attention. We have investigated the signaling pathways utilized upon CD45 ligation and compared these to LPS. Initial investigations focused on PI3K, which is reported to be activated in monocytes upon LPS stimulation (38). Monocyte TNF␣ induced by anti-CD45 antibody (10 g/ml) was inhibited in a dose-dependent manner by the PI3K inhibitor wortmannin (Fig. 2a) with an IC 50 of ϳ0.07 nM. In contrast wortmannin was found to synergize with LPS (10 ng/ml) to enhance TNF␣ production (Fig. 2b).
To determine if the effects seen with wortmannin were due to inhibition of PI3K and not another signaling pathway, we studied the effects of another, structurally unrelated PI3K inhibitor, LY294002. LY294002, like wortmannin, was shown to inhibit anti-CD45 antibody (10 g/ml)-induced monocyte TNF␣ production (IC 50 ϳ0.07 M), while having little effect on LPS (10 ng/ml)-induced monocyte TNF␣ (Fig. 2, c and d).
CD45 Induces PI3K Activity in Peripheral Blood Monocytes-Due to the observed effects of the PI3K inhibitors, wortmannin and LY294002, we investigated PI3K activity. Engagement of CD45 on monocytes induced a transient increase in lipid kinase activity, maximal at 20 min and associated with immunoprecipitates of the anti-p85␣ subunit of PI3K (Fig. 3). Treatment of these monocytes with wortmannin prior to stimulation with anti-CD45 antibody totally inhibited kinase activity (Fig. 3). In contrast, only a weak activation of PI3K was observed following LPS stimulation and none at all in control immunoprecipitates of isotype-matched monoclonal antibodies. Similar experiments with LY294002 were not possible, because unlike wortmannin, this compound does not covalently bind to the enzyme and thus is removed during the assay procedure.
Ligation of CD45 Phosphorylates and Activates Downstream Effectors, PKB and p70 S6K-Recent studies suggest that PI3K-mediated events are transduced via protein kinase B (PKB) (39). Ligation of CD45 resulted in phosphorylation of PKB with similar kinetics to that seen for activation of PI3K, and found to be maximal at 20 min. Fig. 4 illustrates PKB phosphorylation in monocytes following CD45 ligation, which was inhibited by pre-incubation with wortmannin or LY294002. In contrast, LPS induced only a weak phosphorylation of PKB, similar to that seen with the isotype control antibody. We next investigated the involvement of another known downstream effector of PI3K, p70 S6K (40). Ligation of CD45 on monocytes also resulted in activation of p70 S6K (Fig.  5), which was maximal at 30 min and was inhibited by pretreatment with rapamycin. Interestingly, however, the inclu- sion of rapamycin did not inhibit anti-CD45-induced monocyte TNF␣ production (results not shown).
Anti-CD45-induced Monocyte TNF␣ Production Is NF-Bindependent-After demonstrating that TNF␣ production was differentially modulated by PI3K, we investigated the involvement of other factors known to regulate TNF␣ gene expression. We focused upon the transcription factor NF-B, the activation of which has previously been shown to be important in TNF␣ production following LPS stimulation (37). Furthermore, it has recently been reported that NF-B is activated by PI3K. Fig. 6a illustrates NF-B binding activity following 30 min stimulation with LPS (0.1-10 ng/ml). Virtually maximal activation was observed with 1 ng/ml LPS, whereas in contrast anti-CD45 antibody (10 g/ml) resulted in only a weak activation of NF-B. It is unlikely that the difference in activation of NF-B between LPS and CD45 ligation was simply due to a weaker stimulation provided by anti-CD45, because similar amounts of TNF␣ (750 pg/ml) were induced with anti-CD45 (10 g/ml) and LPS (1 ng/ml) (Fig. 6b). These differences between LPS and CD45 ligation were further supported by the observation (Fig.  6c) that TNF␣ synthesis in LPS-but not anti-CD45-stimulated monocytes was inhibited by Ͼ80% when monocytes were infected with an adenoviral vector expressing the inhibitor of NF-B (AdvIB␣).
p38 MAPK Pathway-We have demonstrated that TNF␣ production in monocytes is differentially modulated by both PI3K and NF-B. Numerous studies have demonstrated the importance of MAPKs, in particular the p38 MAPK, in LPS-induced TNF␣ production (9,41). Therefore, we have investigated whether p38 MAPK is also involved in CD45-induced TNF␣ production, using an inhibitor of p38 MAPK, SB203580. SB203580 was found to inhibit both anti-CD45-and LPS-induced monocyte TNF␣ production, IC 50 values ϳ0.005 and 0.006 M, respectively (Fig. 7, a and b). It has previously been demonstrated that LPS can activate p38 MAPK, with maximal stimulation seen at 10 min, followed by rapid loss of activation. 2 CD45 ligation also induced activation of p38 MAPK (maximal at 10 min) displaying similar kinetics to LPS (Fig. 8). Similarly, we have demonstrated that ligation of monocyte CD45 results in activation of p44/p42 MAPK with similar kinetics to LPS (results not shown).

DISCUSSION
In this paper we investigated the signaling pathway(s) involved in monocyte TNF␣ production following ligation of the cell surface receptor CD45 or LPS. Our results reveal the unexpected finding that CD45 ligation results in TNF␣ production that is dependent upon the activation of PI3 kinase but independent of the transcription factor NF-B. In contrast, LPS-induced TNF␣ production was dependent upon NF-B activation as previously reported (37) while PI3K-independent. These observations indicate that, while NF-B has previously been shown to be important in TNF␣ production, it is not always necessary/required.
The importance of the cell surface receptor, CD45 in the activation of T and B cell antigen receptor-mediated signaling pathways and subsequent cellular responses has been well documented. Engagement of CD45 is known to regulate Src tyrosine kinases (p59 fyn , p56 lck , and p70 zap ) phosphorylation (42,43), phospholipase C␥1 regulation (44), inositol phosphate production (45), diacylglycerol production, PKC activation, and calcium mobilization (46). Ligation of CD45 has previously been shown to induce production of cytokines in monocytes (18,19); however, the signaling pathways utilized upon CD45 ligation in monocytes have received little attention.
Ligation of monocyte CD45 results in activation of PI3K and the known downstream effectors PKB and p70 S6K. We have shown the anti-CD45-induced monocyte TNF␣ production is inhibited by the PI3K inhibitors, wortmannin and LY294002. However the inhibitor of p70 S6K activation, rapamycin, did not inhibit anti-CD45-induced TNF␣ production. These findings suggest that TNF␣ production is p70 S6K-independent and other, as yet unidentified, downstream components of PI3K pathway are involved.
In contrast, wortmannin but not LY294002 enhanced LPSinduced monocyte TNF␣ production, suggesting that the effects observed with wortmannin are not specific to PI3K activation. Wortmannin has other targets including PLA 2 (47), and we have shown that the PLA 2 inhibitor, AKTA, also enhances LPS-induced TNF␣ production in monocytes, 3 suggesting that the effect of wortmannin on LPS-induced TNF␣ production may be due to PLA 2 inhibition. How PLA 2 negatively regulates TNF␣ production is unclear, but this enzyme is required for synthesis of PGE 2 , an inhibitor of TNF␣ production (48). Wortmannin is known to stimulate the stress-activated protein kinase pathway (49), and this may also have a positive effect on TNF␣ production. Furthermore, we observed only a weak increase in PI3K and p70 S6K activity following LPS stimulation, suggesting that neither of these pathways play a major role in LPS-mediated events in monocytes. These findings contradict with those performed by Herrera et al. (38), in which LPS was demonstrated to induce PI3K activity in monocytes, using similar methods to those described here. The reason for these apparently contradicting findings remain unclear. These studies have focused upon class 1 A PI3Ks, specifically those involving the p85␣ subunit and the involvement of other PI3K subclasses including those regulated by G-proteins and those which are wortmannin-insensitive have not been investigated. FIG. 8. CD45 ligation and LPS stimulation of monocytes activates p38 MAPK. 5 ϫ 10 6 monocytes were cultured with LPS (10 ng/ml), immobilized anti-CD45 (10 g/ml), or isotype control (IC) antibodies (10 g/ml) for given times. Postnuclear lysates were incubated with a suspension of protein G and anti-p38. p38 MAPK activity was assessed via [␥-32 P]ATP incorporation into ATF-2. Phosphorylated products were visualized by autoradiography using Hyperfilm. duced TNF␣ production in monocytes is p70 S6K-independent. Furthermore, while ligation of CD45 induces phosphorylation of PKB, the involvement of PKB in monocyte TNF␣ production at this stage cannot be verified due to the lack of specific PKB inhibitors. These findings indicate that there must be a bifurcation of the signaling pathways downstream of PI3K that regulate TNF␣ production. Several signaling molecules have been shown to directly and/or indirectly regulate PI3K, leading to the activation of transcription factors, e.g. atypical PKC and PKC. (52). Unfortunately, inhibitors of PKC were found to be toxic to monocytes and as such the involvement of PKC in anti-CD45-induced TNF␣ production has not been assessed. Other potential downstream effectors include Rac, Rab5 (53,59), Bruton's tyrosine kinase (55,56), and JNK/stress-activated protein kinase (57,58). The involvement of these molecules in PI3K-dependent TNF␣ production still remains to be determined.
Several studies have suggested that LPS-induced TNF␣ production in monocytes/macrophages is NF-B-dependent. Protease inhibitors, gliotoxin, and free radical scavengers have all been used to block NF-B activity; however, the lack of specificity of these reagents remains a constant problem. More recently, the over expression of IB␣ following adenoviral infection (AdvIB␣) has been demonstrated to inhibit LPS-induced TNF␣ production in monocytes (37). Curiously, we demonstrated that ligation of CD45 induced IB␣ degradation (results not shown) but only a weak NF-B binding activity; the reasons for this remain unclear, although it suggests further complexity of the NF-B system. Overexpression of AdvIB␣ did not inhibit anti-CD45-induced TNF␣ production. These findings indicate that other, as yet unidentified, transcription factors are involved in anti-CD45-induced TNF␣ production monocytes.
In T cells, induction of TNF␣ gene expression is regulated by the nuclear factor of activated T cells (NFAT), not NF-B (59,60,61). NFAT binds to the 3 element of the TNF␣ gene (located Ϫ97 and Ϫ88 nucleotides relative to the TNF␣ start site), in association with ATF-2 and c-Jun proteins, which bind to the cyclic AMP response element site (62). NFAT DNA binding activity in activated T cells is prevented by the immunosuppressive drugs cyclosporin A (CsA), and FK506 (63,64,61). CsA and FK506 form complexes with their intracellular receptors (immunophilins), and inhibit the activity of calcineurin (protein phosphatase 2B), a ubiquitous calcium-and calmodulin-dependent phosphatase (reviewed in Ref. 65). Induction of TNF␣ mRNA gene transcription in T cells can be blocked by CsA and FK506 (62), and expression of calcineurin is sufficient to activate a reporter gene whose transcription is driven by the TNF␣ promoter (60). The involvement of NFAT in monocyte TNF␣ production remains to be confirmed. However, CsA and FK506 failed to inhibit anti-CD45-induced TNF␣ production in monocytes (results not shown), but this does not discount the involvement of CsA-insensitive NFAT in the regulation of monocyte TNF␣ production. These findings suggest that NFAT, like NF-B, is not required for anti-CD45-induced TNF␣ production.
LPS has previously been shown to activate the three major mammalian MAPK pathways, p42/44 (extracellular signal-regulated kinases 1/2), p38, and p54 MAPK (stress-activated protein kinase), in monocyte/macrophages (9,10,11). However, the relationship between the activation of these signaling molecules cytokine expression remains to be clarified. p38 MAPK is the only kinase that has been shown to play a pivotal role in the production of TNF␣ (66). Previous studies have suggested that the post-transcriptional regulation of TNF␣ is mediated through adenosine-uridine (AU)-rich elements present within the 3Ј-untranslated region of the TNF␣ mRNA (67). Deletion of this region leads to the constitutive synthesis of TNF␣ in cell lines (68) and transgenic animals (69). TNF␣ reporter gene constructs that do not contain the 3Ј-AU-rich element regions lose their sensitivity to inhibition by the p38 inhibitor, SB203580, and it has been suggested that the p38 MAPK cascade is mediating the release of translational repression of TNF␣ (66). The pyridinyl imidazole compound, SB203580, has been used to determine the involvement of p38 MAPK in the regulation of numerous pro-inflammatory cytokines including IL-1, IL-6, and TNF␣ (9). Recently, SB203580 has been shown to inhibit TNF␣ protein and mRNA induced by LPS, suggesting that TNF␣ is being inhibited at the pre-translational level (70,71). We have demonstrated that monocyte TNF␣ production is regulated by distinct transcriptional mechanisms. Furthermore, we have demonstrated that both LPS-and anti-CD45induced TNF␣ production is regulated by p38 MAPK suggesting that both stimuli utilize similar translational mechanisms to regulate TNF␣ production. We observed that ligation of CD45 resulted in activation of the MAPKs p38 and p42/p44 (results not shown) with similar kinetics to that observed with LPS. Furthermore, inhibitors of p38 MAPK (SB203580) and p42/44 MAPK (PD98059) (results not shown) were shown to block both anti-CD45 and LPS-induced TNF␣ production. At higher concentrations SB203580 is known to inhibit the activity of JNK2 and JNK3 (72); however, the IC 50 values observed for SB203580 inhibition of anti-CD45-and LPS-induced monocyte TNF␣ synthesis are consistent with its effects on p38 MAPK and not JNK, although the nonspecific actions of this drugs cannot be disregarded. These findings indicate that TNF␣ production is regulated by distinct transcriptional signaling mechanisms, while the translational mechanisms appear to be identical.
In summary, this study demonstrates that TNF␣ production in monocytes is regulated by multiple signaling pathways. The initiating signals for TNF␣ production in inflammatory disorders such as rheumatoid arthritis are unknown. However, these findings suggest that engagement of specific cell surface receptors may be important in regulating TNF␣ production via distinct signaling pathways and investigation of these mechanisms in both physiological and pathological systems is currently being investigated. 3