C/EBPβ and Its Binding Element Are Required for NFκB-induced COX2 Expression Following Hypertonic Stress*

NFκB plays a critical role mediating COX2 expression in renal medullary interstitial cells (RMICs). The trans-activating ability of NFκB can be modified by another nuclear factor C/EBPβ that can physically bind to NFκB and regulate its activity. Because the COX2 promoter also contains a C/EBPβ site adjacent to the NFκB site, the present study examined whether these two transcription factors cooperate to induce COX2 expression following hypertonic stress. Hypertonicity markedly induced COX2 expression in cultured medullary interstitial cells by immunoblot analysis. The tonicity-induced COX2 expression was suppressed by mutant IκB (IκBm) that blocks NFκB activation, demonstrating that tonicity-induced COX2 expression depends on NFκB activation. However, mutation of the NFκB site in the COX2 promoter failed to abolish tonicity-induced COX2 reporter activity. IκB kinase-1 (IKK1) significantly induced COX2-luciferase activity by 2.3-fold (n = 10, p < 0.01); mutation of the NFκB site also failed to abolish IKK1-stimulated COX2 reporter activity (86 ± 3.1% of wild type, p > 0.05, n = 4). Interestingly, mutation of the C/EBPβ site of the COX2 gene significantly reduced both IKK1 and hypertonicity-induced COX2 reporter activity (p < 0.01). To further examine the potential role of C/EBPβ in tonicity-induced COX2 expression, a dominant negative C/EBPβ-p20 was transduced into RMICs. C/EBPβ-p20 markedly suppressed hypertonic (550 mOsm) induction of COX2 (immunoblot) to a similar extent as IκBm. No additional suppression was observed when both NFκB and C/EBPβ were simultaneously blocked by IκBm and C/EBPβ-p20. Interestingly, IKK-induced COX2 expression was not only blocked by IκBm, but also completely abolished by C/EBPβ-p20. Further studies demonstrated physical association of C/EBPβ to NFκB p65 by coimmunoprecipitation. Importantly, this interaction between C/EBPβ and NFκB was greatly enhanced following hypertonic stress. These studies indicate C/EBPβ is required for the transcriptional activation of COX2 by NFκB, suggesting a dominant role for the C/EBPβ pathway in regulating induction of RMIC COX2 by hypertonicity.

Cyclooxygenase (COX) 1 is a key enzyme in the conversion of arachidonic acid to prostaglandin H, which is further catalyzed to five major bioactive prostaglandins (e.g. PGE2, PGI2, PGF2␣, PGD2, and TXA2) through their distinct synthases. Two isoforms of COX have been identified, designated COX1 and COX2 (1,2). COX1 is constitutively expressed in most tissues detected and is thought to carry out housekeeping functions, such as cytoprotection of the gastric mucosa, regulation of renal blood flow, and control of platelet aggregation. In contrast, COX2 mRNA and protein are normally undetectable in most tissues, but can be rapidly induced by a variety of stimuli, including various cytokines, growth factors, oncogenes, endotoxins, and chemicals (2). Accumulating evidence suggests that COX2-mediated prostaglandins play important roles in regulating cellular homeostasis, inflammation, and tumorigenesis (2)(3)(4)(5).
The kidney is one of the few organs where constitutive COX2 expression is detected. Renal medullary interstitial cells (RMICs) are a major site of COX2 expression in the kidney (6 -8). Recent studies indicate that the hypertonic environment in renal medulla is an important factor contributing to COX2 expression (7,9). Expression of COX2 plays an important role promoting renal medullary interstitial cells to survive otherwise lethal changes in environmental tonicity (7,10), which is critical to the regulation of urinary concentrating ability. The mechanism by which renal medullary interstitial cell COX2 expression is regulated following hypertonic stress has only been partially characterized (7,9). Studies suggest that in RMICs, hypertonic stress activates nuclear factor NFB, and this is critical for induction of COX2 expression in renal medullary interstitial cells (7). NFB has also been reported to be an important signaling pathway promoting COX2 expression by such stimuli as hypoxia and tumor necrosis factor, etc. (11)(12)(13)(14)(15)(16). NFB binding sites have been identified in the promoter region of the COX2 gene (17,18), making it likely that binding of the NFB protein to the NFB cis-acting element is responsible for increased COX2 expression. However, recent studies indicate that the mechanism underlying NFB-associated COX2 expression is more complex. Interactions between NFB and other nuclear factors such as C/EBP, SP1, and PPAR have been reported (19 -21). Cross-talk among these transcriptional factors can be critical for their transcriptional activity (22)(23)(24). The present studies examined the mechanism by which NFB activates COX2 gene expression in cultured renal medullary interstitial cells.

MATERIALS AND METHODS
Cell Culture-Rabbit medullary interstitial cells were cultured as described previously (6). Briefly, female New Zealand White rabbits were anesthetized (44 mg/kg ketamine and 10 mg/kg xylazine, i.m.). The left kidney was removed, and the medulla was dissected and minced with a razor blade under sterile conditions in 5 ml of sterile RPMI 1640 plus 10% (v/v) fetal bovine serum (Hyclone, Logan, Utah). This homogenate was injected subcutaneously in the abdominal wall using a 14-gauge needle. Twenty days postsurgery, subcutaneous nodules appeared. The rabbits were re-anesthetized and sacrificed by decapitation, and the nodules removed under sterile conditions. Nodules were minced into 1-mm fragments and explanted in 75-cm 2 tissue culture plates. Cells were cultured in RPMI 1640 tissue culture medium supplemented with 10% (v/v) fetal bovine serum, and streptomycin and penicillin. Cultures were incubated at 37°C in 95% O 2 , 5% CO 2 . Tissue culture medium was changed every 48 -72 h. Mouse RMICs were prepared as reported (7). C57BL/6J mice were sacrificed, and kidneys were rapidly removed and washed in Ringer's solution. The renal medulla was excised, minced, and placed in Ringer's solution containing collagenase (1 mg/ml) at 37°C for 1 h with occasional agitation. The collagenase-treated tissue was then washed in Dulbecco's modified Eagle's medium (DMEM) three times and cultured in DMEM containing 10% fetal bovine serum. Cells were studied in their third to fourth passages. These cells exhibited characteristic abundant oil red O-positive lipid droplets, a characteristic of type I RMICs (25).
Immunoblotting-Immunoblots were performed on whole cell lysates from cultured RMICs. The protein concentration was determined using the bicinchoninic acid protein assay (Sigma). Thirty micrograms of protein extract were loaded in each lane of a 10% SDS-PAGE minigel and run at 120 V. Protein was transferred to a nitrocellulose membrane at 22 V overnight at 4°C. The membrane was washed three times with TBST (50 mM Tris, pH 7.5, 150 mM NaCl, 0.05% Tween 20) and then incubated in blocking buffer (150 mM NaCl, 50 mM Tris, 0.05% Tween 20, and 5% Carnation nonfat dry milk, pH 7.5) for 1 h at room temperature. The membrane was then incubated with an antihuman COX2 antibody (1:1,000, 160106, Cayman), anti-C/EBP␤ (1:400 sc-150, Santa Cruz Biotechnology), or anti-p-C/EBP␤ (1:500, 3084, Cell Signaling Technology) antibody in blocking buffer overnight at 4°C. Following washing (3ϫ), the membrane was incubated with a horseradish peroxidase-conjugated secondary antibody (1:20,000, Jackson Immuno-Research Laboratories) for 1 h at room temperature, followed by three 15-min washes. Antibody labeling was visualized by addition of the chemiluminescence reagent (Renaissance, PerkinElmer Life Sciences), and the membrane was exposed to Kodak XAR-5 film.
Nuclear Protein Extraction and Immunoprecipitation-Cultured cells were washed with phosphate-buffered saline and lysed on ice for 15 min in hypotonic lysis buffer (10 mM HEPES, 5 mM KCl, 1.5 mM MgCl 2 , 1 mM NaF, 1 mM Na 3 VO 3 , and 0.08% Nonidet P-40) containing proteinase inhibitor mixture (1 tablet/10 ml, Complete Mini, Roche Applied Science). The cell lysate was centrifuged at 4°C at 3,000 rpm for 5 min. The supernatant (cytoplasmic proteins) was stored at Ϫ80°C. The pellet was washed with hypotonic lysis buffer two times and centrifuged at 13,000 rpm for 5 s. The supernatant was removed, and the pellet was resuspended in 50 l of Dignin solution (20 mM HEPES, 1.5 mM MgCl 2 , 0.2 mM EDTA, 420 mM NaCl, 50 mM ␤-glycerophosphate, 1 mM NaF, 1 mM Na 3 VO 4 , 1 mM dithiothreitol, 25% glycerol, pH 7.9) for 30 min and centrifuged for 10 min at 13,000 rpm. The supernatant nuclear protein was used for immunoprecipitation. 50 g of nuclear protein extract was added to 500 l of IP buffer (Tris 20 mM, pH 7.5, NaCl 150 mM, EDTA 1 mM, EGTA, 1 mM, Triton-100 1%). The nuclear protein was precleared by adding 0.2 g of rabbit IgG and 20 l of 25% protein A-agarose, incubated at 4°C for 30 min, and centrifuged at 3,000 rpm. The supernatant was collected, and 0.4 g of anti-C/EBP␤ antibody was added and incubated at 4°C for 2 h. 20 l of 25% protein A-agarose beads were added and incubated at 4°C overnight with mixing. The beads were washed three times with IP buffer and were resuspended in 30 l of 2ϫ sample buffer. The samples were boiled for 2 min, and 20 l of precipitated proteins were added to each lane of an SDS-PAGE gel.
Ad-IBmu, Ad-IKK␣, Ad-CEBP-p20, and Ad-GFP-Adenoviral vectors, encoding a dominant negative IB and a constitutively active IB kinase 1 (IKK1) or a dominant negative C/EBP␤-p20, were used to modulate NFB and C/EBP activity, respectively, in cultured renal medullary interstitial cells. The trans-dominant inhibitor of NFB, IBmut (avian IB␣S36/40A) was provided by Dr. Timothy Blackwell (7). Ad-C/EBP␤-p20 was provided by Dr. Linda Sealy. Constitutively active IKK1 (IKK1) cDNA was kindly provided by Dr. Frank Mercurio (Signal Pharmaceutical, San Diego, CA) and subcloned into pACCMV for IKK1 adenovirus construction (7). The IKK1 was made constitutively active by Ser-Glu mutations in Ser 176 and Ser 180 residues (26). An adenovirus expressing green fluorescent protein was constructed as described (27) for a control adenovirus. For infection of RMICs, 200 l of virus (multiplicity of infection, 100) was added to each culture dish, and GFP adenovirus was used to adjust for equal loading. After a 2-h incubation, the virus was removed, and fresh Dulbecco's modified Ea-gle's medium with 10% fetal bovine serum was added. Experiments were carried out 48 -72 h after infection.
COX2 Reporter Studies-An 891-bp human COX2 luciferase reporter construct was generously provided by Dr. Lee-Ho Wang (17). A 327-bp human COX2 luciferase reporter construct, and its NFB and C/EBP␤ site mutants were provided by Dr. Hiroyasu Inoue (28). The NFB and C/EBP␤ site mutants have been shown to lack the ability to bind to NFB and C/EBP, respectively (28,29). Two NFB sites in the 891-bp COX2 reporter construct were mutated via site-directed mutagenesis using primers: CGGCGGCGGGAGAGCTCATTCCCTGCGCCC (5Ј sense), CAGGAGAGTGGCCACTACCCCCTCTGCT (3Ј sense) (30) (QuikChange II Site-Directed Mutagenesis kits, Stratagene, La Jolla, CA). The firefly luciferase COX2 reporter plasmid and a plasmid containing Renilla luciferase driven by the TK promoter (Promega) were transfected into cells using SuperFect (Qiagen). Cells were lysed 48 h after transfection for luciferase activity measurement using the Dual Luciferase assay system (Promega). COX2 luciferase activity was adjusted by Renilla luciferase activity.
Chromatin Immunoprecipitation (ChIP) Assay-The ability of NFB and C/EBP␤ to bind to endogenous COX2 promoter was examined using the ChIP assay according to the manufacturer's protocol (Upstate Technologies, Lake Placid, NY). Briefly, cultured mouse renal medullary interstitial cells (7) were exposed to isotonic or hypertonic media for indicated periods of time. Cells were then cross-linked with 1% formaldehyde for 5 min. After washing with phosphate-buffered saline, cells were lysed with SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8.1) containing proteinase inhibitors (Complete Mini, EDTA-free). The chromatin was sheared by sonication (strength, 20%; pulse, 12 s ϫ three times). The cross-linked chromatin was quantified to determine the initial amount of DNA present in the different samples. 100 ng of DNA were used as input. The remaining chromatin fractions were precleared with salmon sperm DNA/protein A-agarose for 1 h and immunoprecipitated with antibodies (NFB-p65 or CEBP␤, 200 g/ml, Santa Cruz Biotechnology) overnight at 4°C. The COX2 promoter DNA, bound to p65 and C/EBP␤, was analyzed by PCR using primers: sense, CGGAGGGTAGTTCCATGAAA; antisense, CAGGCTTTTACCCACG-CAAA. PCR was performed at 94°C for 30 s, 58°C for 30 s, and 72°C for 30 s, for 35 cycles.
To further examine whether the C/EBP␤ site in the COX2 promoter plays an important role in mediating NFB binding to the COX2 gene, wild type or mutants of COX2 promoter constructs containing 327-bp human COX2 promoter sequences, were transfected into mouse interstitial cells using SuperFect (Qiagen). 24 h after transfection, cells were exposed to hypertonic stress for 1 h. The cells were cross-linked and precipitated as described in ChIP assay. The transfected human COX2 promoter bound to NFB was detected by PCR using primers specific for the human COX2 gene. PCR primers: sense, CCCCTCTGCTC-CCAAATT; antisense, CGCTCACTGCAAGTCGTAT. The PCR was performed at 94°C for 30 s, 58°C for 30 s, and 72°C for 30 s, for 35 cycles. Genomic DNA from a human cell line HEK293 cells was used as a positive control. Genomic DNA extracted from mouse renal medullary interstitial cells without transfection was used as a negative control.

Mutation of the NFB site of the COX2 Promoter Fails to Suppress Induction of the COX2 Reporter by Hypertonic
Stress-Our previous studies demonstrate that hypertonicity activates NFB, and blocking NFB by a mutant IB dramatically suppresses hypertonic induction of COX2, suggesting that NFB mediates hypertonicity-induced COX2 expression (7). Two NFB binding sites have been identified in the human COX2 promoter (Ϫ446 to Ϫ437 and Ϫ223 to Ϫ214) (31). To examine whether hypertonicity-induced COX2 expression is mediated via binding of NFB protein to the NFB element of the COX2 gene, a COX2 luciferase transcription reporter system with mutant NFB element was used. Hypertonic stress in RMICs significantly increased COX2 reporter activity in both 891-bp COX2 luciferase reporter construct (Fig. 1)-and 327-bp COX2 reporter construct (Fig. 2)-transfected cells. Surprisingly, mutation of NFB sites in the COX2 promoter luciferase reporters failed to abolish hypertonic stress-induced COX2 reporter activity in either COX2 reporter constructs. In contrast, mutation of the C/EBP␤ binding site completely blocked hypertonic activation of COX2 reporter activity (Fig. 2).

Blocking of C/EBP␤ Suppresses Hypertonic Induction of COX2 Protein
Expression-To further examine the involvement of C/EBP␤ in COX2 expression following hypertonic stress, a dominant negative isoform of C/EBP␤, C/EBP␤-p20 (p20) was used to block C/EBP␤ activity (32,33). As shown in Fig. 3, induction of COX2 expression by hypertonic stress was suppressed by IB mutant that blocked NFB activation, consistent with our previous findings (7). These studies now find that a dominant negative C/EBP␤-p20 also dramatically reduced the ability of hypertonicity to induce COX2 expression. More importantly, combined treatment with C/EBP␤-p20 and IBm did not further reduce COX2 expression, suggesting these two factors participate in the same signaling pathway.
C/EBP␤-p20 Suppresses IKK-induced COX2 Protein Expression in RMICs-To further test the hypothesis that NFB and C/EBP␤ participate in the same signaling pathway, we examined the effect of inhibiting C/EBP␤ in NFB-induced COX2 expression. NFB was activated by adenoviral transduction with IB kinase 1 (IKK1). As expected, IKK1, which phosphorylates IB and activates NFB, dramatically induced COX2 expression. However, IKK1-induced COX2 expression was blocked not only by an inactive IBm, but also by blocking C/EBP␤ with C/EBP␤-p20 adenovirus (Fig. 4).
Mutation of the COX2 Promoter C/EBP␤ Binding Site Suppresses IKK-activated COX2 Reporter Activity-To further investigate whether C/EBP␤ is involved in the transcription mechanisms underlying NFB-induced COX2 expression in cultured RMICs, the effect of IKK on the COX2 luciferase reporter system was examined. IKK1 increased COX2 reporter activity by 3-fold (p Ͻ 0.01, Fig. 5). However unexpectedly, mutation of the NFB site failed to completely abolish IKK1-induced COX2 reporter activity. In contrast, mutation of C/EBP␤ site completely abolished IKK-induced COX2 reporter activity.
Hypertonic Stress Enhances Interaction of C/EBP␤ and p65 in Cultured Renal Medullary Interstitial Cells-To further examine whether C/EBP␤ is associated with NFB, we examined whether physical interaction between NFB and C/EBP␤ could be detected by coimmunoprecipitation. Nuclear protein extract FIG. 1. Effect of NFB site mutation on hypertonicity-induced COX2 luciferase reporter activity in cultured renal medullary interstitial cells. Cultured RMICs were co-transfected with wild-type or mutant COX2 promoter-driven firefly luciferase vector and TK-driven Renilla luciferase plasmid. Cells were exposed to isotonic (300 mOsm) or hypertonic (500 mOsm) medium. 24 h later, luciferase activities were determined as described under "Materials and Methods." **, p Ͻ 0.01 versus isotonic medium, n ϭ 6. KBM, NFB site mutation.

FIG. 4. Effect of C/EBP␤-p20 on IKK1-induced COX2 expression.
Cultured RMICs were transduced with GFP, IKK1, IKK1 plus IBm, or IKK plus C/EBP-p20 via adenoviral vectors. 24 h later, cellular proteins were extracted and immunoblotted for COX2. A, representative autoradiograph of immunoblot for COX2. B, densitometry analysis of immunoblot for COX2. **, p Ͻ 0.01; n ϭ 6 was immunoprecipitated using anti-C/EBP␤ antibody and separated with SDS-PAGE. As shown in Fig. 6, C/EBP␤ antibodyimmunoprecipitated proteins from cultured medullary interstitial cells include NFB p65 immunoreactive protein, consistent with a physical association of p65 with C/EBP␤. The interaction between p65 and C/EBP␤ appears to be specific, because no p65 was coprecipitated by PPAR␦ (data not shown), a transcription factor abundantly expressed in renal medullary interstitial cells (34). More importantly, this physical interaction was dramatically enhanced following hypertonic stress, despite the fact that hypertonic stress did not change C/EBP␤ protein expression (Fig. 6). Only C/EBP␤ but not C/EBP␣, ␦, and ␥ were detected in cultured renal medullary interstitial cells by immunoblot. Furthermore, none of these C/EBP isoforms was induced by hypertonic stress (data not shown). Hypertonicity did not change C/EBP␤ phosphorylation (Thr-235) (Fig. 6C), suggesting that phosphorylation of the Thr-235 residue is not critical for hypertonic activation of C/EBP␤.
Hypertonic Stress Increases Binding of C/EBP␤ and NFB p65 to the Endogenous COX2 Promoter-To examine whether hypertonic stress can enhance the binding of C/EBP␤ and NFB to the endogenous COX2 promoter, a chromatin precipitation assay was conducted. An expected PCR product (417 bp) was obtained. Nucleotide sequencing confirmed that the PCR product was identical to the mouse COX2 promoter from Ϫ568 to Ϫ151. As shown in Fig. 7, hypertonic stress enhanced the binding of both NFB p65 and C/EBP␤ to the COX2 promoter in a time-dependent manner, with maximal binding at 1 h following hypertonic stress. This binding of p65 and C/EBP to the COX2 promoter was specific, because transcription factor Sp1 antibody failed to pull-down the COX2 gene detected using the same PCR primers (data not shown).
C/EBP␤ Site Is Required for NFB to Bind to the COX2 Promoter-To further determine whether the C/EBP␤ site in the COX2 promoter is involved in NFB binding to the COX2 promoter, human COX2 promoter constructs with or without C/EBP␤ site mutation were transfected into cultured mouse interstitial cells. The binding ability of NFB to the COX2 promoter constructs was determined by a modified ChIP assay. Because the transfected constructs were from the human COX2 promoter and the host cells were from mouse, this allowed us to specifically amplify the transfected human COX2 promoter using PCR primers specific for human COX2, to examine the effect of mutation of transcription factor binding elements on NFB binding. An expected PCR product (241 bp) was obtained from cells transfected with the human COX2 promoter, but not cells transfected with control vector. As shown in Fig. 8, hypertonic stress increased binding of p65 to the wild-type COX2 promoter. This hypertonic stress-associated binding of p65 was not abolished in cells transfected with a NFB binding site mutant construct, but was abolished by mutation of both the NFB and C/EBP␤ binding sites. These results were consistent with functional studies using the luciferase reporter assay (Fig.   FIG. 5. IKK-associated COX2 luciferase reporter activity. Cultured RMICs were co-transfected with wild-type or mutant COX2 promoter-driven firefly luciferase vector and TK-driven Renilla luciferase plasmid. Cells were transduced with AdGFP or AdIKK. 24 h later, luciferase activities were determined as described under "Materials and Methods." WT, wild type; KBM, NFB site mutant; CEBPM, C/EBP␤ site mutant. **, p Ͻ 0.01 versus AdGFP   FIG. 6. Effect of hypertonic stress on C/EBP␤ expression (A), interaction between C/EBP␤ and NFB p65 (B), and C/EBP␤ phosphorylation (C). Renal medullary interstitial cells were cultured to confluent and exposed to hypertonic stress (550 mOsm) for indicated periods of time. A, whole cell protein extracts were separated on SDS-PAGE and blotted for C/EBP␤. B, nuclear protein extracts were immunoprecipitated by C/EBP␤ antibody. C/EBP␤ immunoprecipitated proteins were blotted for p65 as described under "Materials and Methods." C, whole cell protein extracts were blotted with anti-pC/EBP and C/EBP antibodies.

FIG. 7. Effect of hypertonicity on binding of p65 (A) and C/EBP␤ (B) to the COX2 promoter in renal medullary interstitial cells in vivo.
Cultured renal medullary interstitial cells were exposed to hypertonic medium for 0, 0.5, 1, and 3 h. Cells were fixed with formaldehyde. p65-or C/EBP␤-bound DNA was isolated via immunoprecipitation using anti-p65 or anti-C/EBP␤ antibodies. 100 ng of genomic DNA was used as input. The p65-or C/EBP␤-bound COX2 promoter DNA was detected by PCR as described under "Materials and Methods." 2), supporting a role for the C/EBP␤ site in promoting NFBmediated, hypertonicity-induced COX2 expression.

DISCUSSION
COX2 is an inducible form of cyclooxygenase, and its expression levels regulate endogenous prostaglandin synthesis. Numerous studies have indicated that COX2-derived prostaglandins play an important role in modulating organ development, cardiovascular homeostasis, and inflammatory reaction. Conversely, aberrant expression of COX2 is associated with tumorigenesis. Elucidating the mechanism by which COX2 expression is regulated will be crucial in understanding these COX2regulated physiological and/or pathophysiological processes. COX2 expression is regulated at multiple levels, including transcriptional and post-transcriptional levels. Several putative cis-acting elements have been identified in the 5Ј-upstream region flanking the COX2 gene, including AP2, STAT1, STAT3, NFB, SP1, NF-IL6 (C/EBP␤), and CRE sites (17,18). Several transcription factors, including NFB, C/EBP, CREB, AP-1, and PPAR␥, have been reported to regulate COX2 expression (28,(35)(36)(37)(38)(39)(40). However, the signal transduction pathways leading to activation of these transcription factors are extremely diverse and depend on the cell types studied. The present studies demonstrate a novel transcriptional mechanism underlying NFB regulation of COX2 expression. In medullary interstitial cells, activation of COX2 by the NFB pathway relies on an intact C/EBP␤ element, rather than the NFB element alone. These studies demonstrate positive interaction between NFB and C/EBP␤ binding sites on the COX2 gene.
The presence of mechanisms facilitating survival in the hypertonic conditions is an important characteristic of the cells residing in the renal medulla. The importance of COX activity in maintaining viability of renal medullary cells has long been recognized, based on observations that COX2-inhibiting NSAIDs may cause severe renal medullary injury including papillary necrosis (41). Recent studies show that hypertonicity induces COX2 and that this plays an important role in promoting survival of renal medullary interstitial cells residing in this otherwise lethal hypertonic environment (7,10,34,42). Our previous studies indicate that hypertonicity-induced COX2 expression in RMICs is mediated by NFB. These studies showed that water deprivation not only increased renal medullary COX2 expression, but also increased renal NFB activity (7). Blocking NFB activation using an IB mutant dramatically suppressed hypertonic induction of COX2 expression in cultured renal medullary interstitial cells (7). Although NFB activation is also reported to promote COX2 expression by other stimuli (11)(12)(13)(14)(15)(16), the promoter-based mechanisms have not been fully characterized, partially because the presence of the putative NFB site in the COX2 gene has led to the assumption that this site is the target of NFB.
The present study unexpectedly found that mutation of NFB site in the COX2 gene failed to block COX2 expression by hypertonic stress, suggesting that the NFB element in the COX2 gene promoter is not critical. In contrast, mutation of the C/EBP␤ binding site, which is located adjacent to the NFB site, abolished induction of COX2 expression by hypertonicity. The involvement of C/EBP␤ in hypertonic trans-activation of COX2 expression is also supported by studies showing increased binding of C/EBP␤ as well as NFB p65 to the endogenous COX2 promoter. The C/EBP pathway does not appear to be separate from the NFB pathway, because the additive effect of C/EBP blockade and NFB blocking was not observed. Moreover, mutation of C/EBP site not only abolished hypertonicity-induced COX2 expression, but also abolished IKK-induced COX2 expression, whereas mutation of the NFB site of the COX2 gene failed to abolish IKK-induced COX2 expression, suggesting that the NFB cis-acting site is not critical for IKK-induced COX2 expression. Rather the C/EBP site appears to be integral to the mechanism of NFB activation, leading to COX2 expression.
C/EBP belongs to the basic leucine zipper C/EBP family that is comprised of six members, C/EBP␣, ␤, ␥, ⑀, ␦, and . C/EBP␤ is closely related to C/EBP␣ and C/EBP␦, but is distantly related to C/EBP␥, C/EBP⑀, and C/EBP (43,44). Several truncated forms of C/EBP␤ have been reported (45). The low molecular weight form of C/EBP␤ (C/EBP␤-p20) has been shown to function as a dominant negative form of C/EBP (46). Other studies demonstrate that C/EBP family members are capable of interacting with members of NFB (Rel) family members (22)(23)(24). Overlapping or adjacent NFB/CEBP binding sites are located within the promoter regions of IL-6, IL-8, IL-12, angiotensinogen, serum amyloid A, and COX2 genes (24,47,48), indicating a close relationship between NFB and C/EBP in transcriptional regulation of these proteins (19). Adams et al. (49) reported that nuclear Rel/CEPB␤ heteromer is important in PGG-glucan-induced Rel-A/CEBP␤-related transcription. A p65/CEBP␦ complex, activated following lipopolysaccharide liver, is a potent activator of serum amyloid-A expression, promoting transcription from either NFB or C/EBP elements within the promoter (24). The present studies now show that the C/EBP␤ site of the COX2 promoter is more critical for activation of COX2 expression than the NFB site, because mutation of the C/EBP site significantly blocked IKK-induced COX2 reporter activity, whereas mutation of the NFB site failed to block IKK-associated COX2 expression. The in vivo DNA binding studies show that the C/EBP␤ site on the COX2 promoter plays an important role in mediating p65 binding to the COX2 promoter (Fig. 8). Based on these observations, it may be hypothesized that activated Rel protein(s) may interact with C/EBP(s) in renal medullary interstitial cells. This protein complex may be recruited to COX2 promoter DNA through interaction at the C/EBP␤ site of the COX2 gene, thereby enhancing transcription of COX2 expression. This hypothesis is further supported by coimmunoprecipitation studies demonstrating increasing physical association between Rel A (p65) and C/EBP␤ following hypertonic stress.
Although the cis-acting site for the ␤ isoform of C/EBP has been identified in the COX2 promoter, other C/EBP family members could also bind to the C/EBP␤ site and trans-activate COX2 gene expression (39). Overexpression of murine C/EBP␤ and C/EBP␦ produced a dose-dependent increase in basal and IL-1-stimulated COX2 luciferase reporter activity. C/EBP␦ FIG. 8. Effect of mutation of NFB and C/EBP␤ sites on hypertonicity-induced binding of p65 to the COX2 promoter. Cultured mouse renal medullary interstitial cells were transfected with the human COX2 327-bp promoter construct with or without mutations of NFB and C/EBP binding sites and exposed to hypertonic medium for 1 h. Cells were fixed with formaldehyde. p65-bound DNA was isolated via immunoprecipitation using anti-p65 antibody. The p65-bound-transfected COX2 promoter DNA was detected by PCR as described under "Materials and Methods." ϩ, genomic DNA from human HEK293 cells was used as a positive control; Ϫ, DNA from mouse renal medullary interstitial cells without transfection was used as a negative control. caused a greater enhancement of basal and IL-1-stimulated COX2 promoter activity than C/EBP␤, suggesting that C/EBP␦ is a stronger trans-activator. Overexpression of C/EBP␤-p20, a dominant negative C/EBP inhibitor, which retains the C-terminal DNA binding domain and the leucine zipper region but lacks the N-terminal trans-activating domain of C/EBP␤ (50), not only blocks C/EBP␤-induced COX2 expression, but can also block C/EBP␦-induced COX2 expression (51). Nevertheless, in the present study, C/EBP␣ and -␦ do not seem to be involved, because immunoblotting failed to detect C/EBP␣ and -␦ expression in cultured renal medullary interstitial cells. It has been reported that C/EBP␤ phosphorylation (Thr-235) is associated with ERK/Ras-induced activation of C/EBP␤ (52,53). However, Thr-235 phosphorylation of C/EBP␤ does not seem to be critical in mediating interaction with p65 and promoting COX2 transcription following hypertonic stress, because hypertonicity did not change C/EBP␤ phosphorylation (Fig. 6C). The mechanism by which hypertonicity enhanced interaction of C/EBP␤ and NFB remains to be explored.
In summary, the present study indicates that C/EBP␤ is required for the transcriptional activation of COX2 by NFB following hypertonic stress, suggesting a dominant role for the C/EBP␤ pathway in regulating induction of RMIC COX2 by hypertonicity.