Differential Involvement of Calmodulin-dependent Protein Kinase II-activated AP-1 and c-Jun N-terminal Kinase-activated EGR-1 Signaling Pathways in Tumor Necrosis Factor- (cid:1) and Lipopolysaccharide-induced CD44 Expression in Human Monocytic Cells*

CD44 plays a crucial role in cell migration, inflammation, and immune responses. Alteration in the levels of CD44 expression on monocytic cells by endotoxins and immunoregulatory cytokines may modulate the migration of immune cells to inflammatory sites and the development of immune responses. Lipopolysaccharide (LPS) and the proinflammatory cytokine, tumor necrosis factor- (cid:1) (TNF- (cid:1) ), act as important regulators of CD44 expression in human monocytic cells. We previously demonstrated that the c-Jun N-terminal kinase (JNK), a mitogen-activated protein kinase (MAPK), differentially regulated LPS- but not TNF- (cid:1) -induced CD44 expression in monocytic cells. In this study, our results suggest that the calcium signaling pathway, in particular calmodulin (CaM) and CaM-dependent protein kinase II (CaMK-II), is involved in TNF- (cid:1) - but not LPS-induced CD44 expression. CD44 promoter analysis suggested the participa-tion of distinct transcription factors AP-1 and Egr-1 AP-1 oligonucleotide probe 2 h LPS stimulation (data to the AP-1-DNA complex and blocked by competition unlabeled AP-1 oligonucleotides. Pretreatment of either of the inhibitors for calcium or the JNK MAPK the formation of the AP-1 protein-DNA complex 9 Taken together, these results show the binding of AP-1 to the CD44 promoter in both LPS and TNF- (cid:1) stimulated THP-1/CD14 cells. However, TNF- (cid:1) -induced CD44 transcription may be regulated by AP-1 through the activation of CaMK-II and not via the calcineurin or JNK pathway. of transcription TNF- (cid:1) CD44 transcription and compared the signaling cascade activated by LPS in the regulation of CD44 expression in THP-1/CD14 cells as a model system. We show for the first time the involvement of two distinct and independent signaling pathways in the regulation of LPS- and TNF- (cid:1) -induced CD44 expression. TNF- (cid:1) -induced CD44 expression was found to be regulated by AP-1 through the activation of the CaM/CaMK-II pathway. In contrast, LPS-induced CD44 transcription was regulated specifically by Egr-1 through the activation of JNK MAPK.


CD44 plays a crucial role in cell migration, inflammation, and immune responses. Alteration in the levels of CD44 expression on monocytic cells by endotoxins and immunoregulatory cytokines may modulate the migration of immune cells to inflammatory sites and the development of immune responses. Lipopolysaccharide (LPS) and the proinflammatory cytokine, tumor necrosis factor-␣ (TNF-␣), act as important regulators of CD44 expression in human monocytic cells. We previously demonstrated that the c-Jun N-terminal kinase (JNK), a mitogen-activated protein kinase (MAPK), differentially regulated LPS-but not TNF-␣-induced CD44 expression in monocytic cells. In this study, our results suggest that the calcium signaling pathway, in particular calmodulin (CaM) and CaM-dependent protein kinase II (CaMK-II), is involved in TNF-␣-but not LPS-induced CD44 expression. CD44 promoter analysis suggested the participation of distinct transcription factors AP-1 and Egr-1 in TNF-␣-and LPS-induced CD44 expression, respectively.
Furthermore, TNF-␣-induced CD44 expression was regulated by AP-1 through the activation of the CaMK-II pathway, whereas LPS-induced CD44 transcription was regulated specifically by Egr-1 through JNK activation. Overall, the results suggest the involvement of two distinct and independent signaling pathways involved in the regulation of CD44 transcription that may represent potential targets for anti-inflammatory agents capable of inhibiting CD44-mediated cell migration.
CD44, an adhesion molecule, comprises a family of 85-200-kDa transmembrane glycoproteins that are widely expressed in a variety of cell types (1,2). Interaction of CD44 with its ligand, hyaluronan (HA), a component of the extracellular matrix, has been implicated in various biological and immunological phenomena such as leukocyte adhesion to high endothelial venules, leukocyte extravasation at inflammatory sites, cell activation, and hemopoiesis (2)(3)(4). The induction of CD44 expression and its binding to HA 1 is a key event in the migration of monocytic cells to sites of inflammation or tissue injury (3,4) and as a result has been suggested to play a role in the pathogenesis of inflammatory and autoimmune diseases such as arthritis, allergy, and tumor metastasis (1)(2)(3)(4)(5)(6)(7)(8)(9)(10). The multiple functions of CD44 have been attributed to its extensive molecular heterogeneity generated by alternate mRNA splicing and post-translational modifications (1,2,(11)(12)(13)(14).
Lipopolysaccharide (LPS), a component of bacterial endotoxin, contributes to the pathogenesis of sepsis and inflammation. It is established that mononuclear phagocytes play a key role in the pathogenesis of LPS-induced syndromes and TNF-␣, a proinflammatory cytokine, is involved in LPS-induced inflammatory responses (15). Recently, we and others have shown that TNF-␣ is a potent inducer of CD44 expression and a positive regulator of LPS-induced CD44 expression and CD44-HA interactions in monocytic cells (16 -18). The effect of cytokines on CD44 expression has been documented in different cell types (19 -21). For example, IL-5, a key mediator of asthma-related eosinophilic inflammation, has been shown to enhance CD44 expression in human eosinophils (22) and in murine B cells (19). In human hematopoietic progenitor cells, IL-3 and granulocyte macrophage-colony stimulating factor are cytokines that direct leukocyte development, and induce CD44-HA binding (21). In addition, IL-12 and IL-18 were found to modulate CD44-HA interactions in human T cells (23).
Alterations in the levels of CD44 expression on mononuclear phagocytes by endotoxins and immunoregulatory cytokines may have profound effects on the migration of immune cells to sites of inflammation and in the development of immune responses. Therefore, understanding the regulation of CD44 expression and characterizing the signal transduction events involved may lead to the development of strategies for the treatment of inflammation, autoimmune diseases, and cancers. During the last few years, the signaling pathways induced in monocytes following engagement of LPS with its cognate CD14-Toll receptor complex have been investigated (24). However, the signal transduction events involved in CD44 up-regulation are not well understood. There is some evidence to suggest the involvement of the calcium signaling pathway in CD44 expression on phorbol 12-myristate 13-acetate-stimulated T lymphoma cells (25), whereas phosphatidylinositol 3-kinase and protein kinase C were shown to regulate CD44 expression in neuroblastoma cells (26). Our previous results demonstrated that the generation of HA-adhesive CD44 following TNF-␣ stimulation is regulated by p38 mitogen-activated protein kinase (MAPK) in human monocytic cells (16). Furthermore, the Egr-1 transcription factor has been implicated in the transcription of CD44 in murine B cells (27) and in the IL-1stimulated human endothelial cell line ECV304 (28). However, CD44 transcription following IL-1␤ stimulation was found to be regulated by AP-1 in rat aortic smooth muscle cells (29). In addition, a novel 22-bp regulatory element found in the upstream regulatory region of the CD44 gene was implicated in epidermal growth factor-stimulated murine fibroblasts (30).
We have previously demonstrated that the c-Jun N-terminal kinase (JNK) MAPK pathway regulated LPS, but not TNF-␣induced CD44 expression in human monocytic cells (17). To further define the signaling pathways that distinguish regulation of LPS-and TNF-␣-induced CD44 expression, we examined the role of calcium signaling in the up-regulation of CD44 expression in human promonocytic THP-1 cells as a model system. Our results suggest a distinct role for the calcium pathway, in particular, calmodulin (CaM)/CaM-dependent protein kinase II (CaMK-II) activation in the regulation of TNF-␣, but not LPS-induced CD44 expression. To further understand the involvement of CaM/CaMK-II and JNK in the regulation of CD44 transcription, we analyzed the CD44 promoter to identify the potential transcription factors involved. These results suggest for the first time that TNF-␣ and LPS induce CD44 expression through two distinct and independent signaling cascades without any evidence of cross-talk between the two pathways. TNF-␣ regulates CD44 expression specifically by AP-1 through CaM/CaMK-II activation. In contrast, LPS regulates CD44 transcription selectively by Egr-1 through JNK activation.
Flow Cytometry Analysis-CD44 expression was determined by flow cytometry as described earlier (17,20). Briefly, cells were stained with fluorescein isothiocyanate-labeled anti-CD44 mAbs and isotype (IgG2b)-matched control antibodies (BD Biosciences, Mountain View, CA). The gates were set in accordance with gates obtained with the isotype-matched control antibodies and mean fluorescence intensity was determined. Data were acquired on a BD FACScan flow cytometer (BD Biosciences) and analyzed using the WinMDI version 2.8 software package (provided by J. Trotter, Scripps Institute, San Diego, CA).
Ca 2ϩ Influx-Cells were washed with Ca 2ϩ -free phosphate-buffered saline for 5 min at room temperature and resuspended in buffer A (RPMI 1640 containing 20 mM HEPES, pH 7). The cells were washed again and resuspended in buffer A containing 1 mM Fluo3/AM (Molecular Probes, Eugene, OR) in 1 mM Me 2 SO and 3.75% Pluronic F-127 solution (Sigma) followed by incubation in the dark for 45 min at 37°C. The reaction was stopped by adding an equal volume of buffer B (buffer A containing 5% fetal bovine serum, pH 7.4) followed by incubation for 15 min at 37°C. Cells were washed and resuspended in buffer B at a final concentration of 0.5 ϫ 10 6 cells/ml followed by analysis for Ca 2ϩ levels by the FACScan flow cytometer (BD Biosciences) equipped with CellQuest software, version 3.2.1fl. Cell samples were maintained at 37°C during data acquisition. Intracellular Ca 2ϩ levels at baseline and following treatment with various inhibitors were measured.
Construction of Luciferase Reporter Vectors-A series of human CD44 promoter fragments (Ϫ1109 to ϩ53; GenBank TM accession number AH003670) were amplified from genomic DNA by PCR as described earlier (31,36). The primers with restriction sites used to amplify the promoter fragments are shown in Table I. The amplification consisted of denaturation at 95°C for 2 min, 30 cycles of denaturation at 95°C for 1 min, annealing at 60°C for 1 min, and extension at 72°C for 1 min, and final elongation at 72°C for 10 min. The amplified fragments were subcloned into the PCRII-TOPO vector, and the sequences were confirmed. The correct insertions were subcloned into the NheI and HindIII polylinker site of pGL3B (Promega, Madison, WI), and sequences were confirmed again. All DNA sequencing was performed by the Biotechnology Research Center (University of Ottawa). The Egr-1 and AP-1 mutant constructs were generated either by PCR using specific primers (Table I)   Ϫ 575/ϩ53 628 5Ј-AAGGCTAGCAAAGGCTGAACCCAATG-3Ј Ϫ 505/ϩ53 508 5Ј-AAGGCTAGCCCCCGATTATTTACAGC-3Ј Ϫ 334/ϩ53 387 5Ј-AAGGCTAGCTCTTAAAACCTCTGCGG-3Ј Ϫ 264/ϩ53 317 5Ј-AAGGCTAGCGCTTGGGTGTGTCCTTC-3Ј Ϫ 224/ϩ53 277 5Ј-AAGGCTAGCCACTGTTTTCAACCTCG-3Ј Ϫ 181/ϩ53 234 Antisense primer 5Ј-AACAAGCTTCTCAGCGGCACGAGGCAG-3Ј Transient Transfection and Measurement of Luciferase Activity-Cells were transfected with various CD44 promoter constructs by using FuGENE 6 transfection reagent (Roche Diagnostics) as described previously (31,36). Briefly, 1 g of the test plasmid and 0.5 g of pSV-␤galactosidase vectors (Promega) were incubated for 45 min at room temperature with 4.5 l of FuGENE 6 reagent in 100 l of serum-free Isocove's modified Dulbecco's medium to allow formation of DNA-liposome complexes. These complexes were added to the cell suspension (2 ϫ 10 6 cells/ml) in a 6-well plate (Falcon). Cells were cultured for 15 h prior to stimulation with LPS or TNF-␣ followed by measurement of luciferase and ␤-galactosidase activity using luciferase and ␤-galactosidase assay kits (Promega) in a Bio Orbit 1250 Luminometer (Fisher, Pittsburgh, PA) and spectrophotometer, respectively.
Measurement of CaMK-II Activity-The CaMK-II assay was performed using a CaMK-II kit (Upstate Biotechnology Inc., Mississauga, Ontario, Canada) as per the manufacturer's instructions. Cells were pretreated with inhibitors for 2 h followed by stimulation of cells with either LPS or TNF-␣ for 10 min. Cell pellets were lysed with lysis buffer (50 mM HEPES (pH 7.5), 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl 2 , 100 mM NaF, 100 mM sodium orthovanadate, 1 mM EGTA (pH 7.7), 10 g/ml each of leupeptin, aprotinin, pepstatin A, and phenylmethylsulfonyl fluoride) followed by centrifugation for 20 min at 14,000 ϫ g at 4°C. CAMK-II activity was assayed by using a peptide substrate (KKALRRQETVDAL) specific for CaMK-II. Total proteins (200 g) were mixed with 10 l of CaMK substrate, 0.4 M each of peptide inhibitors for protein kinase A and protein kinase C, and 100 Ci of MgCl 2 -[␥-32 P]ATP in ADB II buffer (20 mM MOPS, pH 7.2, 2.5 mM ␤-glycerol phosphate, 1 mM sodium orthovanadate, 1 mM dithiothreitol, and 1 mM CaCl 2 ). The reaction was incubated at 30°C for 10 min. The phosphorylated substrate was separated from the residual [␥-32 P]ATP using P81 phosphocellulose paper. The papers were washed twice in 0.75% H 3 PO 4 and once in acetone for 2 min, and were placed in 24-well Wallac plates (Turko, Finland) in scintillation fluid. Radioactivity was measured by scintillation counting using a Microbeta counter (Wallac, Turku, Finland). Blanks to correct for nonspecific binding of [␥-32 P]ATP and its breakdown products to the phosphocellulose paper and controls for phosphorylation of endogenous proteins in the sample were performed. CaMK-II activity was measured as counts/min/g of protein and calculated as pM as per the manufacturer's instructions.
Electrophoretic Mobility Shift Assays (EMSA)-EMSA was performed as described earlier (31,36). Briefly, cells were stimulated either with LPS or TNF-␣ in the presence or absence of inhibitors added 2 h prior to stimulation. Cell pellets were lysed for 10 min at 4°C with buffer A (10 mM HEPES, 10 mM KCl, 1.5 mM MgCl 2, 0.5 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride (pH 7.9)) containing 0.1% Nonidet P-40. The lysates were centrifuged at 20,000 ϫ g for 10 min at 4°C. The pellets containing the nuclei were suspended in buffer B (20 mM HEPES, 420 mM NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA and 25% glycerol) at 4°C for 15 min. Both buffers A and B contained the proteolytic inhibitors at 0.5 mM each of dithiothreitol, phenylmethylsulfonyl fluoride, and spermidine, 0.15 mM spermine, and 5 g/ml each of aprotonin, leupeptin, and pepstatin A. The nuclear extracts (5 g) were mixed with 32 P-labeled AP-1 or Egr-1 oligonucleotide probes for 20 min and the resulting complexes were separated on a 5% nondenaturing gel. To illustrate specificity of bound complexes, parallel EMSA reactions were incubated with 50 -200-fold excess of unlabeled specific and nonspecific probes for 20 min prior to the addition of labeled probe. Supershift experiments were also performed using specific mouse anti-AP-1 or anti-Egr-1 antibodies (0.5-1.0 g each) (Santa Cruz). The Egr-1 and AP-1 oligonucleotide sequences were as follows: Egr-1, 5Ј-G-CACGGGGCGGGGGCAGAGGGGCC-3Ј and AP-1, 5Ј-GCGGGCTGCT-TAGTCACAGCCCCC-3Ј.

Differential Involvement of JNK MAPK and the Calcium Signaling Pathways in LPS-and TNF-␣-induced CD44
Expression, Respectively-In this study, we confirmed our earlier observations and demonstrated the involvement of JNK in LPS-, but not in TNF-␣-induced CD44 expression in human monocytes, THP-1 (Ref. 17 and data not shown), and THP-1/CD14 cells (Fig. 1). Both LPS and TNF-␣ induced JNK phosphorylation in THP-1/CD14 cells and this phosphorylation was inhibited by the JNK-specific inhibitor SP600125 in a dose-dependent manner (Ref. 17 and data not shown). Furthermore, treatment with SP600125 inhibited LPS-induced CD44 expression in a dose-dependent manner, whereas TNF-␣-induced CD44 expression remained unaffected even at the highest dose of SP600125 (Fig. 1).
To determine the molecular mechanisms involved, we investigated the role of the calcium signaling pathway in the induction of both LPS-and TNF-␣-induced CD44 expression. LPSinduced CD44 expression in primary human monocytes is complex and is regulated by the interaction of LPS with its CD14-Toll receptor complex and endogenously produced cytokines TNF-␣ and IL-10 following LPS stimulation (17,18,24,37). The signaling pathways involved in LPS-induced CD44 expression in primary monocytes could not be investigated because of the inherent endogenous IL-10 production following LPS stimulation, and the ability of IL-10 to enhance CD44 expression in monocytes (17). Therefore, we investigated the role of the calcium signaling pathway in the regulation of LPSand TNF-␣-induced CD44 expression in IL-10-refractory THP-1/CD14 cells as a model system. We first determined whether LPS and TNF-␣ activated calcium signaling by examining calcium influx by flow cytometry using Fluo-3 as a binding dye. Both LPS and TNF-␣ induced calcium flux after 12 min of stimulation which was inhibited by EGTA, a calcium chelator, to basal levels (Fig. 2). The Ca 2ϩ ionophore A23187 was used as a positive control. To determine the effect of EGTA on CD44 expression, cells were treated with EGTA (2.5-10 mM) for 2 h prior to stimulation with either LPS or TNF-␣ for 24 h. Interestingly, EGTA inhibited TNF-␣-induced CD44 expression in a dose-dependent manner without any effect on LPS-induced CD44 expression (Fig. 3).
Elevations in cytoplasmic Ca 2ϩ concentrations occur following stimulation by diverse stimuli that activate voltage or ligand-gated Ca 2ϩ channels in the surface membrane or following release of Ca 2ϩ present in intracellular stores, mainly in the ER (38,39). To determine whether calcium release from the ER regulates CD44 expression, we used the IP 3 receptor inhibitor, 2-APB, which inhibits the release of calcium from the ER by blocking IP 3 receptor-gated Ca 2ϩ channels (33). Similar to the results obtained with EGTA, 2-APB inhibited TNF-␣-but not LPS-induced CD44 expression in a dose-dependent manner (Fig. 3). We also investigated the role of receptor-mediated entry of extracellular Ca 2ϩ following LPS or TNF-␣ stimulation by employing SKF-96365, a specific inhibitor for receptor-mediated Ca 2ϩ entry (32). Cells were pretreated with SKF-96365 for 2 h prior to stimulation with either LPS or TNF-␣. Interestingly, SKF-96365 inhibited TNF-␣-but not LPS-induced CD44 expression in a dose-dependent manner (Fig. 3). These results suggest that receptor-mediated Ca 2ϩ entry as well as blocking of Ca 2ϩ release from ER may be involved in the regulation of TNF-␣-but not LPS-induced CD44 expression.
CaM and CaMK-II Selectively Regulate TNF-␣-induced CD44 Expression-CaM, a major Ca 2ϩ receptor is present in both cytoplasmic and nuclear compartments. The complex of Ca 2ϩ /CaM regulates several downstream targets including protein kinases and protein phosphatases (40). To understand the role of CaM, we employed its specific inhibitor, W-7 (41). W-7 inhibited TNF-␣-induced CD44 expression in a dose-dependent manner without any effect on LPS-induced CD44 expression (Fig. 4B). One major family of Ca 2ϩ /CaM effectors is the CaMKs, which includes a multifunctional CaMK-II that phosphorylates a large number of signaling proteins. To gain further insight into the role of Ca 2ϩ /CaM, we examined the involvement of CaMK-II by employing the CaMK-II-specific inhibitor, KN-93 (40,41). Cells were pretreated with KN-93 for 2 h followed by LPS or TNF-␣ stimulation. KN-93 inhibited TNF-␣-but not LPS-induced CD44 expression in a dose-dependent manner (Fig. 4B). To determine the biological activities of W-7 and KN-93, cells were treated with W-7 or KN-93 for 2 h followed by stimulation with either LPS or TNF-␣ for 0 -30 min. Both LPS and TNF-␣ enhanced CaMK-II activity by 2-3fold and was inhibited by W-7 and KN-93 in a dose-dependent manner (Fig. 4A).
To confirm the involvement of CaMK-II in TNF-␣-induced CD44 expression, cells were transfected with a dominant negative CaMK-II plasmid or a control vector. CD44 expression was significantly inhibited in cells transfected with the dominant negative CaMK-II plasmid following TNF-␣, but not LPS, stimulation compared with the cells transfected with the control vector (Fig. 4C). In addition, LPS-and TNF-␣-induced CaMK-II activity was inhibited by transfecting cells with dominant negative CaMK-II plasmid. Following transfection with the dominant negative CaMK-II plasmid, CaMK-II activity was observed as 3 Ϯ 1.2 pM following stimulation with either LPS or TNF-␣ compared with the activity of 7 Ϯ 1 pM in LPS-and 6 Ϯ 1.5 pM in TNF-␣-stimulated cells transfected with the control vector.
Calcineurin is also activated by the binding of calcium to CaM, which dissociates the two proteins and allows the catalytic site of calcineurin to become accessible (41). To determine the role of calcineurin, cells were treated with cyclosporine-A or FK-506, the calcineurin inhibitors, prior to stimulation with either LPS or TNF-␣. Neither cyclosporine-A nor FK-506 inhibited CD44 expression in either LPS-or TNF-␣-stimulated cells (Fig. 5). Overall, the results suggest that TNF-␣-induced CD44 expression may be regulated selectively by CaMK-II through the activation of CaM. It may be noted that none of these inhibitors induced apoptosis at the concentrations used as determined by propidium iodide staining (data not shown). scription factors involved in CD44 regulation in monocytic cells, and particularly in response to either LPS or TNF-␣ stimulation are not known. To identify the transcription factors activated by the CaM/CaMK-II and the JNK pathways involved in the regulation of TNF-␣ and LPS-induced CD44 transcription in monocytic cells, respectively, the human CD44 promoter fragment encompassing nucleotide residues from Ϫ1109 to ϩ53 bp relative to the ϩ1 transcription site was cloned, amplified, and subcloned into the NheI and HindIII polylinker site of the luciferase reporter plasmid, GL3B (pCD44Pr-GL3B). THP-1/CD14 cells were transiently transfected with pCD44Pr-GL3B. After 15 h, cells were stimulated with either LPS or TNF-␣ following which relative luciferase activity was assessed. The luciferase activity was detected by 4 h and peaked at 8 h following stimulation with either LPS or TNF-␣ (Fig. 6A). The luciferase activity detected in unstimulated cells following 8 h of culture was 10 -15-fold higher compared with the cells transfected with the control plasmid perhaps because of the constitutive expression of CD44 in these cells. The luciferase activity increased by 2.5-3-fold following LPS or TNF-␣ stimulation compared with unstimulated cells and by 40 -50-fold compared with cells transfected with the control plasmid. The cells transfected with pGL3B alone did not show any increase in luciferase activity following stimulation with either LPS or TNF-␣ (Fig. 6A).
To determine the DNA sequences required for CD44 transcription, a series of promoter fragments (from 5Ј Ϫ1109 to 3Ј ϩ53 bp) were generated by successive deletions starting from the 5Ј-end. The CD44 promoter fragments were amplified, inserted into pGL3B, and sequenced. The exact size of the amplified product and the location of various transcription factor binding sites identified within the CD44 promoter are depicted in Fig. 6B. Examination of the DNA sequences within the CD44 promoter region containing various deletions revealed that deletion of sequences from Ϫ1109 to Ϫ334 bp had no effect on LPS-and TNF-␣-induced luciferase activity compared with the cells transfected with the full-length pCD44Pr-GL3B. However, deletion of sequences from Ϫ334 to Ϫ224 bp abrogated the LPS-and TNF-␣-induced luciferase activity, suggesting the involvement of the Ϫ334 to Ϫ224 sequence in LPS-and TNF-␣-induced CD44 transcription (Fig. 6B). The luciferase activity of cells transfected with the CD44 promoter construct containing Ϫ224 to ϩ53 bp following either LPS or TNF-␣ stimulation was comparable with the activity observed in unstimulated cells (Fig. 6B).
MAT inspector analysis of the promoter sequence between Ϫ334 to Ϫ224 revealed the existence of Egr-1 (Ϫ293 to Ϫ301 bp) and AP-1 (Ϫ238 to Ϫ243 bp) binding sites (Fig. 7A). To delineate the role of Egr-1 and AP-1, a CD44 promoter deletion construct from Ϫ264 to ϩ53 bp (pCD44Pr(Ϫ264)) containing the AP-1 but not the Egr-1-binding site was generated. LPS stimulation of cells transfected with pCD44Pr(Ϫ264) revealed an abrogation of luciferase activity compared with the cells transfected with pCD44Pr(Ϫ334) containing both Egr-1 and AP-1 binding sites, suggesting a role for Egr-1 in LPS-induced CD44 transcription (Fig. 7B, left panel). To confirm the role of Egr-1, we generated an Egr-1 mutant plasmid, pCD44Pr(Ϫ334-Egr1m). Cells transfected with pCD44Pr(Ϫ334-Egr1m) showed a significant reduction in luciferase activity following LPS stimulation (Fig. 7B, left panel) compared with stimulation with TNF-␣ (Fig. 7B, right panel) and to that of the cells transfected with pCD44Pr(Ϫ334) containing both Egr-1 and AP-1 sequences (Fig. 7B, left panel). Significantly, transfection of cells with pCD44Pr(Ϫ264) containing the AP-1 site did not

FIG. 6. LPS and TNF-␣ stimulation induces luciferase activity in THP-1/ CD14 cells transfected with a CD44
promoter/luciferase reporter gene construct. A, cells (2.0 ϫ 10 6 ) were transiently cotransfected with 1 g of either full-length CD44 promoter construct (pCD44Pr.GL3B) or pGL3B control vector and 0.5 g of ␤-galactosidase plasmid. After 15 h, the transfected cells were stimulated with either LPS (1 g/ml) or TNF-␣ (10 ng/ml) for various times followed by the measurement of luciferase and ␤-galactosidase activities in the cell lysates. Luciferase activity was normalized with ␤-galactosidase activity to get relative luciferase units. The results shown are mean Ϯ S.D. of three independent experiments performed in triplicate. B, transcriptional activities of the CD44 promoter deletion constructs in LPS-and TNF-␣-stimulated THP-1/CD14 cells. Cells (2.0 ϫ 10 6 /ml) were transiently cotransfected with various CD44 promoter deletion constructs or control pGL3B vector (1 g) and ␤-galactosidase plasmid (0.5 g). After 15 h, the transfected cells were stimulated with LPS (1 g/ml) or TNF-␣ (10 ng/ml) for another 8 h followed by measurement of luciferase and ␤-galactosidase activities. Following normalization, the luciferase activity was calculated as relative luciferase units as described above. The results shown are mean Ϯ S.D. of three independent experiments performed in triplicate.

FIG. 7. The effect of mutating the Egr-1 and AP-1 binding sites on CD44 promoter activity in LPS-and TNF-␣-stimulated THP-1/CD14 cells.
A, nucleotide sequence of the CD44 promoter from 5Ј Ϫ334 to Ϫ224 bp showing Egr-1 and AP-1 sequences and the superscript indicates mutated nucleotides. B, cells (2.0 ϫ 10 6 ) were transiently cotransfected with 1 g of either CD44 promoter deletion constructs containing mutant Egr-1 and AP-1 binding sequences or pGL3B and 0.5 g of ␤-galactosidase control plasmid. After 15 h, transfected cells were stimulated with either LPS (1 g/ml) or TNF-␣ (10 ng/ml) for 8 h followed by measurement of luciferase and ␤-galactosidase activities. Luciferase activity was normalized for ␤-galactosidase activity to calculate relative light units. The results shown are the mean Ϯ S.D. of three independent experiments performed in triplicate. C, the effect of mutating either Egr-1 or AP-1 alone, or both Egr-1 and AP-1 binding sites on CD44 promoter activity in LPS and TNF-␣ stimulated THP-1/CD14 cells. Cells (2.0 ϫ 10 6 ) were transiently cotransfected with 1 g of CD44 promoter deletion constructs containing wild type Egr-1 and AP-1 sequences pCD44Pr(Ϫ334), mutation at the Egr-1 site alone p(Ϫ334-Egr1m), AP-1 alone, p(Ϫ334-AP1m), both Egr-1 and AP-1 sites p(Ϫ334-Egr-1m, AP1m), or control pGL3B and ␤-galactosidase plasmid (0.5 g). Following stimulation with either LPS or TNF-␣ for 8 h, luciferase activity was calculated as relative light units as described above. The results shown are the mean Ϯ S.D. of three independent experiments performed in triplicate.

CD44 Regulation by LPS-and TNF-␣ in Monocytic Cells 26831
show reduction of luciferase activity upon TNF-␣ stimulation (Fig. 7B, right panel), suggesting that the AP-1-binding sequence may play a key role in TNF-␣-induced CD44 transcription. To confirm the role of AP-1, we introduced mutations in the AP-1-binding sequence (pCD44Pr (Ϫ264-AP1m)). Cells transfected with pCD44Pr(Ϫ264-AP1m) showed significant reduction in luciferase activity following stimulation with TNF-␣ as compared with cells transfected with pCD44Pr(Ϫ264) containing the wild type AP-1 sequence (Fig. 7B, right panel). These results suggest that AP-1 and Egr-1 binding sequences may play key roles in the regulation of TNF-␣-and LPSinduced CD44 transcription, respectively. To further confirm the role of Egr-1 and AP-1 in LPS-and TNF-␣-induced CD44 transcription, respectively, we generated three additional constructs from pCD44Pr(Ϫ334)-GL3B by site-directed mutagenesis. These constructs contained mutations in either Egr-1 alone (p(Ϫ334-Egr1m)), AP-1 alone (p(Ϫ334-AP1m)), or in both Egr-1 and AP-1 (p(Ϫ334-Egr1m, AP1m)). Transfection of cells with p(Ϫ334-Egr1m) containing the mutant Egr-1 sequence alone resulted in significant reduction of luciferase activity following LPS but not TNF-␣ stimulation compared with the cells transfected with pCD44Pr(Ϫ334) containing wild type Egr-1 and AP-1 sequences. Significantly, transfection of cells with p(Ϫ334-AP1m) containing the mutant AP-1 sequence alone resulted in significant reduction of luciferase activity following TNF-␣ but not LPS stimulation compared with the cells transfected with pCD44Pr(Ϫ334). As expected, transfection of cells with p(Ϫ334-Egr1m, AP1m) containing both Egr-1 and AP-1 mutant sequences resulted in abrogation of luciferase activity in both LPS-and TNF-␣-stimulated cells (Fig. 7C). These results suggest a critical role for Egr-1 and AP-1 in LPS-and TNF-␣induced CD44 transcription, respectively.
TNF-␣-induced CD44 Expression Is Selectively Regulated by AP-1 through the Activation of CaMK-II-In view of the above results, it was of interest to determine whether TNF-␣-induced CD44 expression is regulated by AP-1 through the activation of CaMK-II. To address this question, cells were transfected with pCD44Pr(Ϫ334-Egr1m) containing m-Egr-1 and wild type-AP-1 sequences. The transfected cells were treated for 2 h with the inhibitors specific for JNK or Ca 2ϩ signaling pathways prior to stimulation with TNF-␣ for 8 h. As before, following TNF-␣ stimulation, a significant 2-3-fold increase in luciferase activity was observed compared with the unstimulated cells. Prior treatment of transfected cells with EGTA, SKF, APB, W-7, and KN-93 decreased the TNF-␣-induced luciferase activity in a dose-dependent manner (Fig. 8). The luciferase activity observed with the highest concentration of the inhibitors was equivalent to the basal activity observed with unstimulated cells. Furthermore, inhibitors for calcineurin (FK506 and cyclosporine-A) and JNK (SP600125) did not affect TNF-␣-induced luciferase activity at any concentration. These results suggested that TNF-␣-induced CD44 transcription may be regulated by AP-1 through the activation of CaMK-II and not the calcineurin pathway (Fig. 8).
To further determine that TNF-␣-induced CD44 expression is regulated by AP-1 via CaMK-II, the effect of inhibitors for calcium and JNK on the binding of TNF-␣-induced AP-1 to its binding site on the CD44 promoter was investigated by EMSA using the 32 P-labeled oligonucleotide promoter sequence containing the AP-1 binding site as a probe. Maximum binding of AP-1 to the probe occurred by 120 min following TNF-␣ stimulation (Fig. 9A). The identity of a band corresponding to the AP-1 protein-DNA complex was established by competition with cold AP-1 oligonucleotides resulting in abrogation of this band, whereas nonspecific oligonucleotides had no effect on the intensity of this band. Furthermore, treatment of nuclear extracts with anti-c-Fos and c-Jun antibodies resulted in the abrogation of the AP-1-specific band. To determine the effect of the calcium pathway inhibitors on the binding of AP-1 to its binding site in the CD44 promoter, cells were treated for 2 h with various inhibitors prior to stimulation with TNF-␣ for 2 h. The Ca 2ϩ signaling inhibitors EGTA, SKF, APB, W-7, and KN-93 inhibited AP-1 binding to its probe. As before, the calcineurin and JNK inhibitors (FK506, cyclosporine-A, and SP-600125) did not affect the binding of AP-1 (Fig. 9A).
To rule out the possibility that the LPS-activated calcium pathway does not regulate AP-1 and eventual CD44 transcription, similar gel shift experiments were performed following LPS stimulation. The maximum binding of AP-1 to the 32 P-labeled FIG. 8. Calcium pathway inhibitors regulate TNF-␣-mediated CD44 promoter activity in THP-1/CD14 cells. Cells (2.0 ϫ 10 6 ) were transiently cotransfected with 1 g of CD44 promoter construct containing mutant Egr-1 binding site (pCD44Pr(Ϫ334-Egr1m)-GL3B) and 0.5 g of ␤-galactosidase plasmid and cultured for 15 h. The transfected cells were stimulated with various concentrations of inhibitors 2 h prior to stimulation with TNF-␣ (10 ng/ml) for 8 h followed by measurement of luciferase activity as described above. The results shown are the mean Ϯ S.D. of two experiments performed in triplicate.
AP-1 oligonucleotide probe occurred at 2 h following LPS stimulation (data not shown). A major band corresponding to the AP-1-DNA complex was observed and was blocked by competition with specific unlabeled AP-1 oligonucleotides. Pretreatment of cells with either of the inhibitors for calcium or the JNK MAPK pathway did not inhibit the formation of the AP-1 protein-DNA complex (Fig. 9B). Taken together, these results show the binding of AP-1 to the CD44 promoter in both LPS and TNF-␣stimulated THP-1/CD14 cells. However, TNF-␣-induced CD44 transcription may be regulated by AP-1 through the activation of CaMK-II and not via the calcineurin or JNK pathway.
Distinct Regulation of LPS-induced CD44 Expression by Egr-1 through JNK Activation-To determine whether LPSinduced CD44 expression is regulated by Egr-1 through JNK activation, cells transfected with pCD44Pr(Ϫ334-AP1m) containing wild type Egr-1, mAP-1 sequences were treated for 2 h with the JNK and Ca 2ϩ signaling pathway inhibitors prior to stimulation with LPS for 8 h. As before, following LPS stimulation, a significant 2-3-fold increase in luciferase activity was observed compared with the unstimulated cells. LPS-stimulated luciferase activity was inhibited by SP600125 in a dosedependent manner. The luciferase activity observed with the highest concentration of SP600125 was similar to the basal activity observed in unstimulated cells. However, Ca 2ϩ signal-ing inhibitors (EGTA, SKF, APB, W-7, and KN-93) did not affect LPS-induced luciferase activity (Fig. 10). These results suggested that LPS-induced CD44 transcription may be regulated by Egr-1 through the activation of JNK MAPK.
To confirm that LPS-induced CD44 expression is regulated by Egr-1 through JNK activation, the effect of calcium and JNK MAPK inhibitors on the binding of LPS-induced Egr-1 to its binding site in CD44 promoter was investigated by EMSA by using the 32 P-labeled oligonucleotide probe containing the Egr-1 sequence of the CD44 promoter. LPS induced maximum binding of Egr-1 at 60 min post-stimulation (Ref. 17 and data not shown). We observed three Egr-1 containing bands as their intensity was abrogated following specific competition with unlabeled oligonucleotides. In contrast, nonspecific oligonucleotides did not affect the intensity of any of the Egr-1 bands. Furthermore, treatment of nuclear extracts with anti-Egr-1 antibodies exhibited abrogation of the top two Egr-1 specific bands. To determine the role of JNK and calcium pathways, cells were treated for 2 h with various inhibitors prior to stimulation with LPS for 2 h. SP-600125 significantly reduced binding of LPS-induced Egr-1 to the labeled probe (Fig. 11A), whereas Ca 2ϩ signaling inhibitors including EGTA, SKF, APB, W-7, FK-506, cyclosporine-A, and KN-93 did not affect the binding, suggesting that LPS-induced CD44 transcription may FIG. 9. TNF-␣ stimulation activates the AP-1 transcription factor that is regulated by the inhibitors of calcium signaling pathway. THP-1/CD14 cells (1 ϫ 10 6 /ml) were stimulated with TNF-␣ (10 ng/ml) (A) or LPS (1 g/ml) (B) for various times ranging from 30 to 240 min followed by collection of nuclear extracts. Nuclear extracts (5 g) were probed with 32 P-labeled oligonucleotides corresponding to the AP-1-binding sequence of the CD44 promoter. To determine the specificity of AP-1 binding, the nuclear extracts were probed in the presence of different concentrations of unlabeled specific (Sp oligo) or nonspecific oligonucleotides (NS oligo). The nuclear extracts were also treated with anti-c-Jun or anti-c-Fos antibodies to identify AP-1 specific bands by supershift EMSA. To determine the effect of inhibitors of calcium and JNK MAPKs on TNF-␣and LPS-induced activation of AP-1, cells were treated with different concentrations of either SP600125 or calcium inhibitors (EGTA, W-7, SKF, APB, KN93, and FK506) for 2 h prior to stimulation with either TNF-␣ or LPS. The complexes were subjected to electrophoresis followed by autoradiography. Arrows indicate the AP-1 bands. The experiment shown is representative of three independent experiments performed. be regulated by Egr-1 through JNK activation.
To rule out the possibility of the calcium pathway regulating TNF-␣-induced Egr-1 activation, similar gel shift experiments were performed. The results show that as for LPS, three Egr-1 containing bands were observed with maximum binding occurring at 60 min following TNF-␣ stimulation (Ref. 17 and data not shown). The intensity of these bands was blocked specifically by competition with cold Egr-1 oligonucleotides. Furthermore, anti-Egr-1 antibodies abrogated the intensity of top two TNF-␣-induced Egr-1 bands. Prior treatment of cells with either of the inhibitors for JNK or the calcium pathway did not inhibit the formation of Egr-1 bands (Fig. 11B). Taken together, the results show the binding of Egr-1 to the CD44 promoter following stimulation with both LPS and TNF-␣, however, Egr-1 may play a key role in LPS-induced CD44 transcription in THP-1/CD14 cells. DISCUSSION CD44 may be induced as a result of the signals generated following association of LPS with the LPS-binding protein, and consequential binding of the LPS-LPS-binding protein complex with the CD14-Toll-like receptor-4 complex expressed on cells of the monocytic lineage (37). We and others have previously demonstrated that LPS-induced CD44 expression in monocytic cells is mediated, at least in part, by the endogenously produced TNF-␣ (17,18). Both LPS and TNF-␣ act as positive regulators of CD44 induction and its binding with HA in monocytic cells (16 -18), which has been suggested to play a vital role in their migration, and development of immune and inflammatory responses (3,4,(7)(8)(9). However, the signaling pathways underlying CD44 induction in monocytic cells especially following stimulation with TNF-␣ are not understood. We previously demonstrated the distinct involvement of JNK MAPK in LPSinduced CD44 expression in monocytic cells (17). In this study, we investigated the role of calcium signaling pathways and the activation of downstream transcription factors in TNF-␣-induced CD44 transcription and compared the signaling cascade activated by LPS in the regulation of CD44 expression in THP-1/CD14 cells as a model system. We show for the first time the involvement of two distinct and independent signaling pathways in the regulation of LPS-and TNF-␣-induced CD44 expression. TNF-␣-induced CD44 expression was found to be regulated by AP-1 through the activation of the CaM/CaMK-II pathway. In contrast, LPS-induced CD44 transcription was regulated specifically by Egr-1 through the activation of JNK MAPK.
Ca 2ϩ is an important intracellular messenger in many biological processes (38). Influx of calcium ions through ligand and voltage-gated calcium channels in the plasma membrane together with Ca 2ϩ release from ER stores results in complex calcium signaling cascades (38,39). Several mechanisms may control Ca 2ϩ entry in response to external stimuli including membrane depolarization, activation of intracellular messengers, and depletion of intracellular calcium storage (38,39). The release of Ca 2ϩ from internal stores (ER) is controlled by Ca 2ϩ itself or by an expanding group of messengers. For example, the IP 3 , produced in response to a signal from the membrane lipid phosphatidylinositol, triggers Ca 2ϩ release from the ER after binding to the IP 3 receptor (38,39). By employing a number of inhibitors specific for the Ca 2ϩ pathway, our results suggested that intracellular release as well as extracellular influx of Ca 2ϩ into the cytosol may be involved in the regulation of TNF-␣-induced CD44 expression.
CaM, a key signaling protein responsible for integrating the Ca 2ϩ signal to transcription factors, is known to regulate cell cycle and related cytoskeletal functions and ion channel activity (40). Following binding to Ca 2ϩ , CaM undergoes a conformational change that renders it active and able to recognize and bind target proteins with high affinity (40). Among the possible downstream targets of CaM are calcineurin and CaMK-II (40,42,43). As with other kinases, CaMK-II undergoes autophosphorylation on a threonine residue contained in a phosphopeptide common to its ␣ and ␤ subunits and converts it into a Ca 2ϩ /CaM independent enzyme (40,42). The results obtained by employing specific inhibitors suggested that CaMK-II may act as a key link in TNF-␣-induced CaM activation and CD44 expression.
In this study, we also investigated the downstream signaling events responsible for CD44 transcription. The transcription factors involved in the regulation of CD44 expression in monocytic cells have not been well defined. Egr-1 was shown to be involved in phorbol 12-myristate 13-acetate-induced CD44 expression in murine B cells and IL-1␣-induced CD44 expression in human endothelial cells (27,28). The Egr-1 gene is a prototypic member of a gene family encoding transcription factors that share a conserved zinc finger DNA-binding motif. This gene is induced rapidly and transiently in response to B cell receptor cross-linking or treatment with phorbol 12-myristate 13-acetate (44). Recently, AP-1 was shown to regulate CD44 transcription in murine vascular smooth muscle cells in response to stimulation with IL-1␤ (29). AP-1, a key regulator of the expression of a number of cytokines (45,46), is a het-erodimeric transcription factor comprised of members of the jun (c-Jun, JunB, and JunD) and fos (c-Fos, Fra-1, Dra-2, FosB, and FosB2) proto-oncogene families (45,47). Members of the fos and jun families dimerize via their leucine zipper domain with a variety of transcription factors including cAMP-response element-binding protein /ATF, Maf, NFB, NFAT, and GR (45, 48 -50). Various Fos and Jun proteins interact with the promoters of cytokine genes either individually as AP-1 dimers, or in cooperation with other transcription factors such as NFAT, cAMP-response element-binding protein/ATF (45, 48 -50). The results of this study show the differential involvement of Egr-1 and AP-1 in response to two distinct signaling cascades LPS and TNF-␣, respectively, in monocytic cells. This was demonstrated by analyzing the luciferase activity in cells transfected with CD44 promoter deletion constructs exhibiting mutations in the binding sites for either Egr-1 and/or AP-1.
CD44 has a relatively high constitutive expression in unstimulated THP-1 cells. It may be noted that the unstimulated cells transfected with CD44 promoter constructs (Ϫ181 to ϩ1 bp) containing mutant AP-1 and Egr-1 binding sequences exhibited a significant level of luciferase activity suggesting that the constitutive basal activity may not be mediated by either Egr-1 or AP-1 alone or through their cooperative effects. The   FIG. 11. LPS-induced Egr-1 is regulated by SP600125. THP-1/CD14 cells (1 ϫ 10 6 /ml) were stimulated with LPS (1 g/ml) (A) or TNF-␣ (10 ng/ml) (B) for 60 min followed by collection of nuclear extracts. Nuclear extracts (5 g) were probed with 32 P-labeled oligonucleotides containing the Egr-1 binding sequence of CD44 promoter. To determine the specificity of Egr-1 binding, nuclear extracts were probed in the presence of different concentrations of unlabeled specific (Sp oligo) or nonspecific oligonucleotides (NS oligo). The nuclear extracts were also treated with anti-Egr-1 antibodies to identify the Egr-1 bands by supershift EMSA. To determine the effect of inhibitors of calcium and JNK MAPKs on TNF-␣and LPS-induced activation of Egr-1, cells were treated with different concentrations of either SP600125 or calcium inhibitors (EGTA, W-7, SKF, APB, KN93, and FK506) for 2 h prior to stimulation with either TNF-␣ or LPS. The complexes were subjected to electrophoresis followed by autoradiography. Arrows indicate the Egr-1 bands. The experiment shown is representative of three independent experiments performed. transcription factors responsible for basal CD44 expression/ luciferase activity are not known. Analysis of Ϫ181 to ϩ1 bp in the CD44 promoter sequence may lead to identification of potential transcription factors responsible for the maintenance of basal levels of CD44 expression. MatInspector analysis of the Ϫ181 to ϩ1 bp sequence reveals a number of sequences at positions between Ϫ166 to Ϫ154 that are similar to the core promoter motifs such as the TATA box required for basal transcription (51,52), in addition to those of NF-1, Elk-1, NF-B, and E2F. It is possible that these transcription factors either alone or in concert with other transcription factors may regulate basal constitutive expression of CD44.
The findings of a distinct involvement of AP-1 and the upstream CaM/CaMK-II signaling molecules in TNF-␣-induced CD44 expression raised the question of how CaMK-II may serve to induce CD44 transcription. Similarly, it was interesting to determine whether JNK activation by LPS resulted in CD44 expression through Egr-1. Our results demonstrate that LPS-induced CD44 transcription is regulated specifically by Egr-1 through JNK activation, whereas TNF-␣-induced CD44 expression was independently regulated by AP-1 through CaM/ CaMK-II activation without any evidence of cross-talk between the two pathways. Although TNF-␣ induced the activation of Egr-1, it did not regulate CD44 transcription. Furthermore, the JNK inhibitor, SP600125, did not affect the binding of TNF-␣induced Egr-1 to its binding site in the CD44 promoter. Similarly, LPS induced the activation of AP-1 and upstream calcium signaling proteins including CaM and CaMK-II, activation of these proteins, however, did not affect CD44 transcription. Furthermore, the binding of LPS-induced AP-1 to its binding site on the CD44 promoter was not affected by inhibitors of the CaM/CaMK-II pathway. The molecular mechanism by which LPS-but not TNF-␣-induced JNK selectively activates Egr-1 resulting in CD44 transcription or how TNF-␣induced CaM/CaMK-II selectively activates AP-1 leading to CD44 transcription is not understood.
In addition to the signal transduction pathways, the molecular mechanisms controlling gene expression involve chromatin remodeling and DNA methylation (53). Egr-1 is known to associate with corepressor proteins such as NAB1 and NAB2 that can repress transactivation of Egr-1 target genes (54). Additionally, hypermethylation of the CD44 promoter has been shown to cause silencing of the CD44 gene (55). It is likely that the inability of LPS and TNF-␣ to induce CD44 expression through AP-1 or Egr-1, respectively, may be because of the induction of Egr-1 corepressors, or alterations in the methylation status of their binding sites in the CD44 promoter. It is also likely that LPS and TNF-␣ may activate certain histone deacetylases and chromatin remodeling complexes that may work together to silence AP-1 and Egr-1 binding sites, respectively, of the CD44 promoter. Understanding the role of transcription factor corepressors, the status of CD44 promoter methylation, and chromatin remodeling may provide the potential mechanisms controlling LPS/TNF-␣-mediated CD44 transcription.
In summary, regulation of CD44 expression in TNF-␣-and LPS-stimulated monocytic cells involves a complex sequence of intracellular signaling events. In this study, we show for the first time a role for JNK and Egr-1 in LPS-but not in TNF-␣mediated signaling resulting in CD44 induction. Moreover, TNF-␣-induced CD44 expression involved the selective activation of the AP-1 transcription factor and the calcium signaling complex comprised of CaM and CaMK-II. The JNK MAPK and CaM/CaMK-II, therefore, may represent important and novel targets for anti-inflammatory agents capable of inhibiting CD44 expression.