- Email: [email protected]
PGE2 induces COX-2 expression in podocytes via the EP4 receptor through a PKA-independent mechanism
Cellular Signalling 20 (2008) 2156–2164
Contents lists available at ScienceDirect
Cellular Signalling j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c e l l s i g
PGE2 induces COX-2 expression in podocytes via the EP4 receptor through a PKA-independent mechanism ☆ Wissam H. Faour, Kaede Gomi, Christopher R.J. Kennedy ⁎ Kidney Research Centre, Division of Nephrology, Department of Medicine, the Ottawa Hospital, Ottawa, Ontario, Canada K1H 8M5 Ottawa Health Research Institute, Ottawa, Ontario, Canada K1H 8M5 Department of Cellular and Molecular Medicine, University of Ottawa, Ottawa, Ontario, Canada K1H 8M5
a r t i c l e
i n f o
Article history: Received 2 April 2008 Received in revised form 23 July 2008 Accepted 11 August 2008 Available online 15 August 2008 Keywords: Podocyte Eicosanoids p38 MAPK cAMP AMPK Cyclooxygenase
a b s t r a c t Cyclooxygenase-2 (COX-2)-dependent prostaglandin E2 (PGE2) synthesis correlates with the onset of proteinuria and increased glomerular capillary pressure (Pgc) glomerular disease models. We previously showed that an in vitro surrogate for Pgc (cyclical mechanical stretch) upregulates the expression of both COX-2 and the PGE2 responsive E-Prostanoid receptor, EP4 in cultured mouse podocytes. In the present study we further delineate the signaling pathways regulating podocyte COX-2 induction. Time course experiments carried out in conditionally-immortalized mouse podocytes revealed that PGE2 transiently increased phosphorylated p38 MAPK levels at 10 min, and induced COX-2 protein expression at 4 h. siRNA-mediated knockdown of EP4 receptor expression, unlike treatment with the EP1 receptor antagonist SC 19220, completely abrogated PGE2-induced p38 phosphorylation and COX-2 upregulation suggesting the involvement of the EP4 receptor subtype. PGE2-induced COX-2 induction was abrogated by inhibition of either p38 MAPK or AMP activated protein kinase (AMPK), and was mimicked by AICAR, a selective AMPK activator, and by the cAMP-elevating agents, forskolin (FSK) and IBMX. Surprisingly, neither PGE2 nor FSK/IBMX-dependent p38 activation and COX-2 expression were blocked by PKA inhibitors or mimicked by 8-cPT-cAMP a selective EPAC activator, but were instead abrogated by Compound C, suggesting the involvement of AMPK. These results indicate that in addition to mechanical stretch, PGE2 initiates a positive feedback loop in podocytes that drives p38 MAPK activity and COX-2 expression through a cAMP/AMPKdependent, but PKA-independent signaling cascade. This PGE2-induced signaling network activated by increased Pgc could be detrimental to podocyte health and glomerular filtration barrier integrity. © 2008 Elsevier Inc. All rights reserved.
1. Introduction Podocytes are a major component of the glomerular filtration barrier [1]. Numerous studies show that podocyte injury is a key initiating event leading to filtration defects and progressive glomerular damage [2,3]. Cyclooxygenases (COX) and their metabolites e.g., prostaglandin E2 (PGE2), are involved in both hemodynamic and inflammatory-based pathogenesis of progressive nephropathies. Increased COX-2 levels correlate significantly with the extent of renal damage and high levels of both cyclooxygenase isoforms (COX-1 and COX-2) are reported in immunological and nonimmunological nephropathies, such as systemic lupus erythematosus [4], glomerulo☆ C. R. J. Kennedy is the recipient of a New Investigator Award from the Canadian Institutes of Health Research. Dr. W. H. Faour is the recipient of the KRESCENT (Kidney Foundation of Canada–Canadian Institutes of Health Research) joint fellowship. This work was supported by the Canadian Institutes of Health Research. ⁎ Corresponding author. Ottawa Health Research Institute, Division of Nephrology, Ottawa Hospital and University of Ottawa, 451 Smyth Rd., Rm 2501, Ottawa, Ontario, Canada K1H 8M5. Tel.: +1 613 562 5800 8529; fax: +1 613 562 5487. E-mail addresses: [email protected] (W.H. Faour), [email protected] (C.R.J. Kennedy). 0898-6568/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.cellsig.2008.08.007
sclerosis [5], Heymann nephritis [6] and renal ablation [7]. Furthermore, chronic use of COX inhibitors attenuates renal injury and albuminuria in rodents with subtotal nephrectomy, a model of chronic kidney disease [8]. A role for COX-derived prostaglandins in glomerular disease is further supported by recent studies showing that cyclooxygenase overexpression predisposes to proteinuria, podocyte damage, and loss. Cheng et al.[9] showed that podocyte-specific overexpression of COX-2 enhances podocyte sensitivity to either adriamycin-induced nephropathy or subtotal nephrectomy in mice. However, the specific prostaglandin receptor that underlies such damage has not been defined. PGE2, the major renal metabolite produced by COX-2, acts through at least four seven-transmembrane receptor subtypes (EP1, EP2, EP3, and EP4) [10,11]. A role for the EP1 receptor subtype in mediating glomerular damage and proteinuria is suggested by studies of spontaneously hypertensive rats and those with streptozotocininduced diabetic nephropathy. Treatment with an EP1 receptor antagonist reduced proteinuria and indices of glomerulosclerosis in both models of nephropathy [12]. Conversely, EP1-null mice are more susceptible than their wild type littermates to renal dysfunction in a model of nephrotoxic serum nephritis [13,14]. However, in these
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
studies, proteinuria levels were similar between wild type and EP1-null mice. PGE2 signaling in podocytes is mediated by the EP1 and/or the EP4 receptor subtypes [15]. We previously showed that equibiaxial mechanical stretch, a mimic of intraglomerular pressure, induces COX2 and PGE2 EP4 receptor expression in a mouse podocyte cell line [16]. Moreover, the enhanced EP4 receptor expression resulted in greater cAMP accumulation in response to exogenous PGE2, and yielded a significant effect upon the podocyte actin cytoskeleton [16]. However, the effect of such enhanced PGE2 signaling upon COX-2 expression remains unresolved. The present study was undertaken to characterize the signaling events initiated by PGE2 and leading to COX-2 upregulation in podocytes. We found that PGE2 initiates a positive feedback loop of COX-2 upregulation driven by EP4 receptor activation, a phenomenon entirely p38 MAPK dependant as p38 inhibition completely abrogated COX-2 induction. Our data support a cAMP-dependant, yet PKAindependent model of COX-2 induction as several PKA inhibitors failed to block PGE2 or FSK/IBMX mediated COX-2 induction. Specifically, we show that the heterotrimeric kinase AMPK [17,18] acts upstream of p38 MAPK to induce COX-2 expression. Triggering of this PGE2-driven signaling network in vivo by increased Pgc could be detrimental to podocyte health and the integrity of the glomerular filtration barrier. 2. Materials and methods 2.1. Cell culture Culture of conditionally-immortalized mouse podocytes, kindly provided by Dr. K. Endlich, was carried out as previously described [19]. Experiments were also carried out in another line of conditionally-immortalized mouse podocytes, kindly provided by Dr. P. Mundel, with similar results. Briefly, cells were grown on type-I collagencoated plastic tissue culture dishes in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 0.1 mg/ml streptomycin. Podocytes were routinely propagated at 33 °C in RPMI culture media supplemented with 10 U/ml mouse recombinant γ-interferon to promote the expression of the temperaturesensitive large T-antigen. Differentiation was induced by maintaining cultures at 37 °C in media without γ-interferon for 7–14 days. Cell culture materials were purchased from Life Technologies (Burlington, ON). For experimental protocols and conditions employed in this study, differentiated cells were trypsinized and transferred to the 6 well plates, and cultured for an additional 3 days. Following overnight serum starvation in RPMI-1640 medium supplemented with 0.1% FBS, podocytes were subjected to stimulation with PGE 2 (Cayman Chemical), forskolin/3-isobutyl-1-methylxanthine (IBMX) or 5aminoimidazole-4-carboxamide-1-{beta}-D-ribofuranoside (AICAR) (Cell Signaling) for 10 min to 16 h, or 8-(4-chlorophenylthio)-2′-Omethyladenosine 3′,5′-cyclic monophosphate (8-cPT-2′-O-MecAMP, Alexis Biochemicals) for 10 min to 4 h. Control cells (designated as ‘non-stimulated’) were cultured under identical conditions but were not exposed to bioactive chemicals. MAPK inhibitors SB 202190, (Tocris; Ellisville, MO), wortmannin (Sigma), compound C (Calbiochem), H89 (BIOMOL) and KT 5720 (Calbiochem) were employed at 25 µM, 200 nM, 3 µM, 10 µM and 2 µM respectively, in RPMI-1640 + 0.1% FBS, with a 30 min incubation prior to stimulation experiments. Indomethacin (10 µM) was added 45 min prior to all experimentation to prevent endogenous PGE2 synthesis.
2157
(Amersham Pharmacia Biotech; Baie d'Urfé, QC). Membranes incubated with blocking buffer for 1 h were probed overnight at 4 °C with rabbit antibodies directed against COX-2 (1:1000 dilution, Cayman Chemical Corporation — Ann Arbor, MI), actin (1:1000 dilution, SigmaAldrich), or a rabbit antibody (New England Biolabs) which recognizes the phosphorylated (activated) form of p38 MAP kinase (1:1000 dilution of anti-active p38). After incubation with an appropriate HRPconjugated secondary antibody (1:2000 to 1:40,000) (GE Healthcare UK), blots were incubated in chemiluminescent substrate (Pierce — Rockford, IL) and exposed to blue light-sensitive film (Kodak). Densitometric analyses of scanned blots were carried out using Alpha Imager software. 2.3. EP4 siRNA transfection assays Transient transfections were performed using HiPerfect transfection reagent (Qiagen) according to the manufacturer's protocol. Briefly, EP4 siRNA or scrambled non-silencing off-target siRNA (Qiagen) was diluted to a final concentration of 5 nM in 100 µL serum free HyClone RPMI medium (Fisher Scientific) followed by the addition of the transfection reagent. The resulting lipid-RNA complexes were incubated for 10–20 min at room temperature and then added drop-wise to the podocytes (1.5–3 × 105 cells) seeded onto collagen-coated 6-well cluster dishes. Cells were transfected for 48 h prior to use in experiments. 2.4. Real-time RT-PCR of EP4 mRNA in mouse podocytes Cells grown on type-I collagen-coated 60 mm dishes at 37 °C were transfected with EP4 siRNA as described above. RNA was extracted using an RNeasy kit according to the manufacturer's instructions (Qiagen) and subjected to reverse transcription. EP4 mRNA levels were assessed by real-time RT-PCR using TaqMan RT-PCR Master Mix Reagents and an ABI Prism 7000 Sequence Detection System (Applied Biosystems). Reactions were performed using the following
2.2. Immunoblotting Differentiated podocytes grown on 6 well plates were serumstarved, subjected to the appropriate experimental conditions, and were washed twice with ice-cold PBS. Cells lysed with 1× Laemmli buffer (62.5 mM Tris–HCl, pH 6.8, 2% w/v SDS, 10% glycerol, 50 mM DTT, 0.1% w/v bromophenol blue), were electrophoresed on 10% resolving gels, and electrotransferred to nitrocellulose membranes
Fig. 1. PGE2 induces COX-2 expression in mouse podocytes. Conditionally-immortalized differentiated mouse podocytes were incubated with vehicle (ethanol) as control (C) or 1 µM of PGE2 for 2 h, 4 h, 8 h and 16 h, as indicated and lysed with Laemmli buffer. Protein extracts were resolved by SDS-PAGE and immunoblotted with a COX-2 antibody. Expression levels were normalized to β-actin protein content as assessed by densitometric analysis (n = 3). ⁎P b 0.01 vs. vehicle control.
2158
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
conditions: 50 °C for 2 min, 95 °C for 10 min, and 45 cycles of 95 °C for 15 s and 60 °C for 1 min. Primers and TaqMan probe for murine EP4: sense, 5′-ATG GTC ATC TTA CTC ATC GCC AC-3′; antisense, 5′-CTT TCA
CCA CGT TTG GCT GAT-3′; probe 6FAM-CAT CTG CTC CAT TCC GCT CGT GGT-TAMRA. Values were normalized to GAPDH mRNA levels in each sample.
Fig. 2. PGE2-induced COX-2 expression is EP4 receptor and p38 MAPK-dependent. (A) Conditionally-immortalized differentiated mouse podocytes were incubated for 4 h with either vehicle, PGE2 alone (1 µM), or in combination with SB 202190 (25 µM), wortmannin (200 nM), SC 19220 (10 μM) or with sulprostone alone (10 μM). Cell lysates were resolved by SDSPAGE and immunoblotted with a COX-2 antibody. Expression levels were normalized to β-actin protein content as assessed by densitometric analysis (n = 3). ⁎P b 0.05 vs. vehicle control; ⁎⁎P b 0.01 vs. vehicle control. (B) Podocytes were transfected for 48 h with either 5 nM scrambled siRNA or 5 nM EP4-siRNA and subsequently stimulated for 4 h with either vehicle or PGE2 (1 μM) as indicated. COX-2 protein levels were determined as in (A). (C and D) Podocytes were transfected for 48 h with either scrambled 5 nM siRNA or 5 nM EP4-siRNA. (C) EP4 mRNA levels were quantified by analyzing total cellular RNA by RT-PCR using specific TaqMan primer/probe sets and normalizing the values to GAPDH levels. (D) Cells were stimulated with PGE2 (1 μM) or FSK (10 μM) in the presence of 0.5 mM IBMX. Intracellular cAMP levels were assayed as described in the Materials and methods (n = 2).
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
2159
Fig. 3. Time and dose dependent p38 MAPK phosphorylation by PGE2. Conditionally-immortalized differentiated mouse podocytes were stimulated with 1 μM PGE2 over a range of time points as indicated (Fig. 3A) or over a range of PGE2 concentrations (10− 9M–10− 6M) for 10 min (Fig. 3B). Cells were lysed with Laemmli buffer and analyzed by western blot using a phospho-p38 antibody and normalized for total p38 protein content using a p38 antibody and analyzed by densitometry (n = 3). ⁎P b 0.01 vs. vehicle control.
2.5. cAMP assay Podocytes were grown on collagen I-coated 6-well cluster dishes at 37 °C and transfected with EP4 or scrambled siRNA. Cells were preincubated with HyClone RPMI-1640 (Fisher Scientific) containing 0.1% FBS, 5 μM indomethacin (Sigma-Aldrich Canada Ltd.), and 0.5 mM IBMX (Sigma-Aldrich) for 10 min. Cells were then stimulated for a further 10 min with either PGE2 (1 μM; Cayman Chemical Company) or forskolin (10 μM; Sigma-Aldrich). Intracellular cAMP was determined by cAMP EIA Kit purchased from Cayman Chemical.
3.2. PGE2–COX-2 induction is mediated through activation of the EP4 receptor subtype Bek and coworkers [15] showed that both EP1 and EP4 receptor subtypes, but not EP2 and EP3 subtypes, are expressed in conditionally-immortalized mouse podocytes. We previously demonstrated that mechanical stretch of these same cells increases EP4, but not EP1 subtype expression. Stretch thereby increases the podocyte's response
2.6. Statistics Data are expressed as means of duplicate determinations from individual experiments and are presented as means ± SEM where n ≥ 4 or means ± SD where n = 3 experiments. Statistical significance was accepted at P b 0.05 as determined by ANOVA followed by a Newman– Keuls multiple comparisons test or alternatively determined by a paired t-test where appropriate. 3. Results 3.1. PGE2 induces COX-2 expression in cultured mouse podocytes Podocyte COX-2 levels are increased in vivo in models of glomerular damage that involve elevated intraglomerular capillary pressure [8]. Such forces can be mimicked in vitro by the application of equibiaxial mechanical stretch which induces expression of both COX2 and the EP4 receptor subtype in cultured podocytes [16]. In the present study, we asked whether PGE2 in podocytes may act in an autocrine manner through the EP4 receptor to further enhance COX-2 expression. We therefore carried out time course experiments in which cultured mouse podocytes were incubated with 1 µM PGE2. As shown in Fig. 1, western blot analysis indicates that PGE2 upregulates COX-2 protein in a time dependent manner. PGE2 rapidly induced COX-2 protein, reaching significant levels after 2 h of stimulation and a steady state level after 4 h, declining gradually until returning to near baseline levels by 16 h.
Fig. 4. PGE2-induced COX-2 expression is mimicked by cAMP-elevating agents. Conditionally-immortalized differentiated mouse podocytes were incubated with vehicle control (C) or stimulated for 4 h with PGE2 (1 μM), FSK (10 μM) + IBMX (0.5 mM), or in combination with SB 202190 (25 μM). Cells were lysed with Laemmli buffer and protein extracts were resolved by SDS-PAGE and immunoblotted with a COX2 antibody. Expression levels were normalized to β-actin protein content and analyzed by densitometry (n = 3). ⁎P b 0.05 vs. vehicle control; ⁎⁎P b 0.01 vs. vehicle control.
2160
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
to PGE2, to yield a greater synthesis of intracellular cAMP [16]. To determine which EP subtype accounted for the observed effects of PGE2 upon COX-2 induction, mouse podocytes were incubated with PGE2 for 4 h in the presence or absence of EP agonist/antagonists or EP4 expression was knocked down by siRNA. While PGE2 upregulated COX-2 protein levels, sulprostone (1 µM), an EP1/EP3 receptor agonist, failed to induce a statistically significant increase in COX-2 expression (Fig. 2A). Likewise, SC 19220 (1 µM), a relatively selective EP1 receptor antagonist, failed to reverse PGE2 dependant COX-2 upregulation. Additionally, SC 19220 by itself, or in combination with PGE2, induced COX-2 protein expression indicating a possible non-specific effect (Fig. 2A). The involvement of the EP4 subtype was further suggested as PGE2 failed to induce COX-2 protein levels when podocytes were transfected with EP4-siRNA for 48 to 72 h prior to experimentation (Fig. 2B, C). Knockdown of the receptor was confirmed by quantitative RT-PCR of EP4 mRNA levels, as well as by failure of the EP4-siRNA transfected cells to elicit a cAMP response following PGE2 stimulation of the podocytes (Fig. 2). These findings suggest that the EP4 receptor subtype mediates PGE2-stimulated COX-2 expression in podocytes. 3.3. Time course and dose dependent activation of p38 MAPK by PGE2 Previous work in our lab showed that COX-2 induction in response to mechanical stretch is p38 MAPK-dependent. Stretch induced maximal p38 phosphorylation after 10 min in cultured mouse
podocytes [16]. We asked whether PGE2 transduces its effects on COX-2 expression in podocytes through a similar activation of p38 MAPK. As shown in Fig. 2, PGE2-stimulated COX-2 induction was p38dependant as the selective p38 inhibitor, SB 202190, blocked this effect, whereas wortmannin (a PI3K/AKT inhibitor) was without effect. We also performed both time course and dose response experiments of p38 MAPK phosphorylation in response to PGE2. As judged by western blot, stimulation of mouse podocytes with 1 µM of PGE2 resulted in a transient but statistically significant 3-fold stimulation of p38 phosphorylation after 10 min which declined rapidly thereafter (Fig. 3A). Maximal p38 phosphorylation levels were achieved with 1 µM PGE2 (Fig. 3B). These findings clearly demonstrate that PGE2 can activate p38 MAPK to enhance COX-2 expression in cultured podocytes. 3.4. PGE2 upregulates COX-2 through a cAMP-dependant but PKAindependent mechanism In order to examine the role of cAMP and cAMP-dependent protein kinase A (PKA) in PGE2-mediated p38-dependent COX-2 induction, we performed western blot experiments using the cAMP-elevating agents, forskolin (10 μM) and IBMX (0.5 mM), and the PKA-selective inhibitors, H89 (25 μM) and KT5720 (2 µM). Cultured podocytes incubated with PGE2 or FSK/IBMX for 4 h showed significant COX-2 protein biosynthesis (Fig. 4). In keeping with a role for p38 MAPK in COX-2 expression, SB 202190 abrogated both PGE2 and FSK/IBMX-
Fig. 5. PGE2-induced p38 MAPK activation and COX-2 expression requires cAMP, but is independent of PKA. Cells were treated with vehicle or stimulated for 4 h with PGE2 (1 μM), FSK (10 μM) + IBMX (0.5 mM), or in combination with either H89 (10 μM) or KT5720 (2 μM) (Fig. 5A and 5C), 10 min (Fig. 5B) or 1 h (Fig. 5D). Cell lysates were analyzed by western blot using antibodies against either COX-2, phospho-p38, or phospho-CREB-1/ATF-1. Expression levels were normalized to β-actin protein content and signal intensity quantified by densitometry (n = 3). 5A) ⁎⁎P b 0.01 vs. vehicle control or H89 alone; ⁎P b 0.05 vs. vehicle control.
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
induced COX-2 expression. However, H89 failed to block either COX-2 upregulation (Fig. 5A) or p38 MAPK phosphorylation driven by either PGE2 or FSK/IBMX (Fig. 5B). Surprisingly, incubation with H89 alone tended to increase COX-2 expression and phospho-p38 levels. However, experiments performed using another PKA-selective inhibitor (i.e., KT5720) were likewise without effect on PGE2-induced COX-2 expression (Fig. 5C). Inhibition of PKA activation by KT5720 was confirmed by experiments showing that a FSK/IMBX mediated increase of phosphorylated CREB, a known target of PKA could be abrogated by incubation with this inhibitor (Fig. 5D). These data suggest that a PKA-independent mechanism likely underlies PGE2dependent COX-2 induction. 3.5. PGE2-induced p38 phosphorylation and COX-2 protein upregulation occur independently of EPAC activation EPAC1, a cAMP-dependent guanine nucleotide exchange factor for the small GTPase RAP, is a well known mediator of PKA-independent signaling. We therefore investigated the role of EPAC1 in PGE2 mediated induction of COX-2 protein translation. Immunoblotting of lysates from podocytes treated with a specific and cell membrane
2161
permeable activator of EPAC1 and EPAC2, 8-cPT-2′-O-Me-cAMP (8cPT-cAMP) [20–22] showed that 8-cPT-cAMP had no effect on both p38-MAPK phosphorylation and COX-2 protein levels. These data suggest that EPAC does not mediate PGE2-induced COX-2 expression in podocytes (Fig. 6). 3.6. AMPK activation mediates PGE2 upregulation of COX-2 expression and p38 phosphorylation The heterotrimeric kinase, AMPK, is activated by intracellular energy deficits brought about by increased AMP-to-ATP ratio when the cell's energy state is low [17,18]. Other reports indicate that AMPK activation can induce p38 MAPK activity [23]. In order to test whether AMPK plays a role in PGE2-mediated COX-2 gene induction in mouse podocytes, we performed western blot experiments using the AMPKselective inhibitor, compound C, and the AMPK activator, 5-aminoimidazole-4-carboxamide-1-D-ribofuranoside (AICAR). Cultured podocytes incubated with either PGE2 or FSK/IBMX for 4 h exhibited significant COX-2 induction (Fig. 7A) that was abrogated by coincubation with compound C. Activation of AMPK by incubation of podocytes with AICAR significantly induced COX-2 expression that was likewise susceptible to inhibition by compound C. Finally, inhibition of AMPK by compound C abolished both PGE2 and FSK/IBMX induced p38 MAPK phosphorylation (Fig. 7B). 4. Discussion
Fig. 6. PGE2-induced p38 MAPK activation and COX-2 expression are independent of EPAC. Cells were treated with vehicle control (c) or stimulated with PGE2 (1 μM), FSK (10 μM) + IBMX (0.5 mM), or 8-cPT (EPAC activator) for 4 h or 10 min (Fig. 6) as indicated. Cell lysates were analyzed by western blot using either COX-2 antibody or phospho-p38 antibody and normalized to β-actin protein content and signal intensity quantified by densitometry.
Induction of glomerular COX-2 is reported in various glomerularbased diseases, correlating significantly with proteinuria and the extent of renal damage [4–7]. COX-2 is barely detectable in healthy podocytes, yet levels rise rapidly during disease progression. Mesangial cells strongly upregulate COX-2 mRNA and protein following stimulation with proinflammatory cytokines, such as IL-1β, by a p38dependant mechanism [24]. Similarly, increased podocyte COX-2 levels are evident in rodent models of glomerular injury, including diabetic nephropathy [25,26], subtotal renal ablation [27], and adriamycin-induced nephrotoxicity [9]. Furthermore, podocyte-specific transgenic overexpression of COX-2 renders mice more susceptible to the development of proteinuria and renal injury following puromycin [28] or adriamycin-induced nephrotoxicity [9]. Importantly, inhibition of prostaglandin synthesis abrogates proteinuria and reduces disease progression in each of these models. Together, these findings strongly suggest that the induction of COX-2 in the glomerular tuft is an important step in the etiology of damage to the filtration barrier. However, the mechanisms and signal transduction pathways that regulate podocyte COX-2 expression remain incompletely resolved. PGE2 is a major COX-2 metabolite synthesized by glomerular podocytes. The role played by PGE2 in podocyte biology is unclear. However, our data clearly indicate that this eicosanoid can significantly and rapidly induce COX-2 expression in conditionally-immortalized mouse podocytes. PGE2 can potentially exert its effects on podocytes via either EP1 and/or EP4 receptor subtypes, both of which are expressed in this cell type [15]. EP4 receptor activation yields cAMP production in order to regulate target genes in podocytes [29], while little is known regarding signaling through the EP1 receptor other than it promotes transient intracellular Ca2+ fluxes [15]. Our data favor an EP4-dependent mechanism for COX-2 gene induction since siRNAmediated knockdown of the EP4 receptor abolished the ability of PGE2 to induce COX-2 expression. Furthermore, an EP1-selective antagonist failed to inhibit PGE2-dependent COX-2 upregulation. It was interesting to note that SC-19220 promoted COX-2 induction, suggesting possible off-target effects. However, the participation of the EP1 subtype is unlikely as cAMP production by forskolin and IBMX significantly upregulated COX-2 expression while sulprostone, an EP1/ EP3 agonist was without effect. Therefore a positive feedback loop
2162
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
Fig. 7. AMPK mediates PGE2–COX-2 induction and p38 MAPK phosphorylation. Fig. 7A) Differentiated podocytes were stimulated for 4 h with either vehicle control (c), PGE2 (1 μM), FSK (10 μM) + IBMX (0.5 mM), AICAR (2 mM), or in combination with compound C (comp C) (3 μM). The cells were then processed for COX-2 immunoblotting as described earlier, normalized for β-actin protein content and analyzed by densitometry (n = 3). ⁎P b 0.01 vs. vehicle control and compound C alone; ⁎⁎P b 0.001 vs. vehicle control and compound C alone. Fig. 7B) Cells were stimulated as in 7A, but for 10 min and the cell lysates analyzed for phosphorylated p38, normalized for β-actin protein content and analyzed by densitometry (n = 3).
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
appears to operate wherein COX-2-derived PGE2 acts in an autocrine manner via the EP4 subtype to further enhance both COX-2 expression and PGE2 synthesis. How is this signaling loop achieved? Our previous studies showed that mechanical stretch of podocytes, a mimic of hydrostatic forces encountered in vivo, yields COX-2 expression in a p38 MAPKdependent manner [16]. We investigated whether PGE2 would similarly require p38 activation to increase COX-2 expression. Other studies showed that COX-2 mRNA induction can be integrated within a posttranscriptional feedback loop involving PGE2 production and p38 activation [30,31]. In this regard, PGE2, by promoting p38 activation, stabilizes COX-2 mRNA induced by IL-1β [31]. However, in those studies PGE2 was shown to induce a sustained p38 phosphorylation that enhanced COX-2 mRNA stability [31]. In contrast, our data indicate that p38 is transiently activated (10 min) by PGE2 only to rapidly decline thereafter, suggesting a transcriptional model of COX-2 gene regulation in podocytes. The classical pathway leading to p38 MAPK activation proceeds through a GTPase signaling cascade involving raf/ras and MKK3/6 [32,33]. However, our findings suggest that PGE2-induced p38 signaling involves a distinct signaling cascade that requires cAMP production. Such p38 MAPK activity likely proceeds through a PKAindependent mechanism since several PKA-selective inhibitors (i.e., H89, KT5720; Fig. 5) each failed to block PGE2-induced p38 MAPK phosphorylation and COX-2 induction. A recent study demonstrated that PGE2-induced osteoclast differentiation requires p38 MAPK activation, yet signals through PKA-dependent phosphorylation of transforming growth factor beta-activated kinase 1 (TAK1), a mitogen activated kinase kinase [34]. However, these authors demonstrated that PGE2 alone failed to induce osteoclast differentiation, and that TAK1 activity requires additional phosphorylation achieved by coincubation with a TNF family member, RANKL [34]. A similar scenario is unlikely in PGE2-stimulated podocytes as neither PGE2 nor FSK/IBMX induced immuno-detectable phosphorylated TAK1 (unpublished results). Alternatively, we considered another PKAindependent signaling route that involves the exchange protein activated by cAMP-1 (EPAC1), a cAMP-dependent guanine nucleotide exchange factor for the small GTPase RAP [20–22]. In cerebellar neurons, EPAC mediates p38 MAPK activation, independent of PKA [22]. However, our data suggest that EPAC is not implicated in PGE2dependent p38 activation and COX-2 induction in podocytes. Accordingly, the EPAC-selective activator, 8-cPT-cAMP failed to stimulate either p38 MAPK phosphorylation or COX-2 induction (Fig. 6). Taken together, our results suggest that the ability of PGE2 to increase COX-2 expression in podocytes does not rely upon the recruitment of either TAK1 or EPAC. Interestingly, a similar cAMPdependent, PKA-independent and EPAC-independent pathway may operate in mouse neonatal cardiac fibroblasts, where β2-adrenergic signaling induces p38-dependent IL-6 transcription [35]. We next tested whether another cAMP-dependent, PKA-independent signaling route could account for the effects of PGE2 on COX-2 induction. AMP activated protein kinase (AMPK) serves as an energy sensor regulated by the intracellular AMP-to-ATP ratio, and it promotes glucose uptake via the GLUT4 transporter in a variety of mammalian cell types. AMPK activation was shown to induce p38 MAPK phosphorylation in a rat liver cell line [36] and in cardiac muscle [23]. In the latter studies, AMPK was shown to recruit the scaffold protein, transforming growth factor-β-activated protein kinase 1-binding protein 1 (TAB1) which then facilitated p38 autophosphorylation. Conversely, this signaling paradigm may not be broadly applicable since activation of p38 MAPK appears to be independent of AMPK in skeletal muscle [37]. In the present study, activation of podocyte AMPK by AICAR mimicked, while inhibition of AMPK by compound C, abrogated both PGE2 and FSK/IBMX induced COX-2 expression. This AMPK-driven activation of p38 represents a novel link between podocyte dependent energy status and downstream targeted genes, such as COX-2. As podocyte COX-2
2163
expression increases in rodent models of diabetic nephropathy, our findings suggest that AMPK may likewise be induced rendering the podocyte's energy status extremely sensitive to changes in glucose levels which in turn modulate critical cellular events involving control of cellular hypertrophy, adhesion and survival which are hallmarks of diabetic podocytes. Interestingly, a recent study demonstrated that AMPK-induced p38 phosphorylation promotes actin cytoskeleton reorganization and enhances cardiac lipoprotein lipase activity in a rat model of acute diabetes [38]. The authors concluded that such changes in cardiac fatty acid metabolism could translate into increased cardiovascular damage following diabetes. Our previous studies indicated that mechanically stretched podocytes, which exhibit enhanced COX-2 and EP4 receptor expression, respond to PGE2 by dramatically reordering their actin cytoskeleton [16]. However, the impact of PGE2-driven AMPK activity in podocytes in the context of diabetic nephropathy, as well as other glomerular injury contexts in vivo awaits investigation. In summary, our data shows that the EP4 receptor subtype can trigger a signaling paradigm that promotes AMPK activation and p38dependent COX-2 induction to further enhance PGE2 synthesis within the podocyte. This combination of COX-2 expression and ensuing PGE2 production may have deleterious effects on the podocyte leading to a compromised glomerular filtration barrier. Acknowledgments We thank Drs. K. Endlich and P. Mundel for the kind gift of the murine podocyte cell lines. References [1] H. Pavenstadt, W. Kriz, M. Kretzler, Physiological Reviews 83 (1) (2003) 253. [2] S.J. Shankland, Kidney international 69 (12) (2006) 2131. [3] B.L. Wharram, M. Goyal, J.E. Wiggins, S.K. Sanden, S. Hussain, W.E. Filipiak, T.L. Saunders, R.C. Dysko, K. Kohno, L.B. Holzman, R.C. Wiggins, Journal of the American Society of Nephrology 16 (10) (2005) 2941. [4] S. Tomasoni, M. Noris, S. Zappella, E. Gotti, F. Casiraghi, S. Bonazzola, A. Benigni, G. Remuzzi, Journal of the American Society of Nephrology: JASN 9 (7) (1998) 1202. [5] W. Weichert, A. Paliege, A.P. Provoost, S. Bachmann, American Journal of Physiology. Renal Physiology 280 (4) (2001) F706. [6] T. Takano, A.V. Cybulsky, American Journal of Pathology 156 (6) (2000) 2091. [7] J.L. Wang, H.F. Cheng, M.Z. Zhang, J.A. McKanna, R.C. Harris, The American Journal of Physiology 275 (4 Pt 2) (1998) F613. [8] C.K. Fujihara, G.R. Antunes, A.L. Mattar, N. Andreoli, D.M. Malheiros, I.L. Noronha, R. Zatz, Kidney international 64 (6) (2003) 2172. [9] H. Cheng, S. Wang, Y.I. Jo, C.M. Hao, M. Zhang, X. Fan, C. Kennedy, M.D. Breyer, G.W. Moeckel, R.C. Harris, Journal of the American Society of Nephrology 18 (2) (2007) 551. [10] Y. Boie, R. Stocco, N. Sawyer, D.M. Slipetz, M.D. Ungrin, F. Neuschafer-Rube, G.P. Puschel, K.M. Metters, M. Abramovitz, European journal of pharmacology 340 (2–3) (1997) 227. [11] M. Kiriyama, F. Ushikubi, T. Kobayashi, M. Hirata, Y. Sugimoto, S. Narumiya, British journal of pharmacology 122 (2) (1997) 217. [12] T. Suganami, K. Mori, I. Tanaka, M. Mukoyama, A. Sugawara, H. Makino, S. Muro, K. Yahata, S. Ohuchida, T. Maruyama, S. Narumiya, K. Nakao, Hypertension 42 (6) (2003) 1183. [13] S. Rahal, L.I. McVeigh, Y. Zhang, Y. Guan, M.D. Breyer, C.R. Kennedy, Canadian journal of physiology and pharmacology 84 (8–9) (2006) 877. [14] C.R. Kennedy, H. Xiong, S. Rahal, J. Vanderluit, R.S. Slack, Y. Zhang, Y. Guan, M.D. Breyer, R.L. Hebert, American journal of physiology.Renal physiology 292 (2) (2007) F868. [15] M. Bek, R. Nusing, P. Kowark, A. Henger, P. Mundel, H. Pavenstadt, Journal of the American Society of Nephrology: JASN 10 (10) (1999) 2084. [16] L.C. Martineau, L.I. McVeigh, B.J. Jasmin, C.R. Kennedy, American Journal of Physiology. Renal Physiology 286 (4) (2004) F693. [17] D.G. Hardie, Nature reviews.Molecular cell biology 8 (10) (2007) 774. [18] D.G. Hardie, Science (New York, N.Y.) 315 (5819) (2007) 1671. [19] P. Mundel, J. Reiser, AeZÃiMa Borja, H. Pavenstädt, G.R. Davidson, W. Kriz, R. Zeller, Experimental Cell Research 236 (1) (1997) 248. [20] K.S. Lyle, J.H. Raaijmakers, W. Bruinsma, J.L. Bos, J. de Rooij, Cell Signal 20 (6) (2008) 1104. [21] H. Rehmann, A. Wittinghofer, J.L. Bos, Nature reviews. Molecular Cell Biology 8 (1) (2007) 63. [22] J. Ster, F. DeBock, N. Guerineau, A. Janossy, S. Barrere-Lemaire, J. Bos, J. Bockaert, L. Fagni, Proceedings of the National Academy of Sciences 104 (7) (2007) 2519. [23] J. Li, E.J. Miller, J. Ninomiya-Tsuji, R.R. Russell III, L.H. Young, Circulation research 97 (9) (2005) 872.
2164
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
[24] Z. Guan, S.Y. Buckman, B.W. Miller, L.D. Springer, A.R. Morrison, Journal of Biological Chemistry 273 (44) (1998) 28670. [25] R. Komers, J.N. Lindsley, T.T. Oyama, S. Anderson, Clinical and Experimental Pharmacology and Physiology 34 (1–2) (2007) 36. [26] R. Komers, J.N. Lindsley, T.T. Oyama, W.E. Schutzer, J.F. Reed, S.L. Mader, S. Anderson, Journal of Clinical Investigation 107 (7) (2001) 889. [27] J.L. Wang, H.F. Cheng, S. Shappell, R.C. Harris, Kidney International 57 (6) (2000) 2334. [28] Y.I. Jo, H. Cheng, S. Wang, G.W. Moeckel, R.C. Harris, Nephron. Experimental Nephrology 107 (3) (2007) e87. [29] N. Endlich, K. Endlich, Microscopy Research and Technique 57 (4) (2002) 228. [30] P. Akarasereenont, K. Techatrisak, S. Chotewuttakorn, A. Thaworn, Mediators of inflammation 8 (6) (1999) 287.
[31] W.H. Faour, Y. He, Q.W. He, M. de Ladurantaye, M. Quintero, A. Mancini, J.A. Di Battista, Journal of Biological Chemistry 276 (34) (2001) 31720. [32] E. Herlaar, Z. Brown, Molecular Medicine Today 5 (10) (1999) 439. [33] J. Kyriakis, J. Avruch, Physiological Reviews 81 (2) (2001) 807. [34] Y. Kobayashi, T. Mizoguchi, I. Take, S. Kurihara, N. Udagawa, N. Takahashi, Journal of Biological Chemistry 280 (12) (2005) 11395. [35] F. Yin, Y.Y. Wang, J.H. Du, C. Li, Z.Z. Lu, C. Han, Y.Y. Zhang, Journal of Molecular and Cellular Cardiology 40 (3) (2006) 384. [36] X. Xi, J. Han, J.Z. Zhang, Journal of Biological Chemistry 276 (44) (2001) 41029. [37] R.C. Ho, N. Fujii, L.A. Witters, M.F. Hirshman, L.J. Goodyear, Biochemical and Biophysical Research Communications 362 (2) (2007) 354. [38] M.S. Kim, G. Kewalramani, P. Puthanveetil, V. Lee, U. Kumar, D. An, A. Abrahani, B. Rodrigues, Diabetes 57 (1) (2008) 64.
Contents lists available at ScienceDirect
Cellular Signalling j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c e l l s i g
PGE2 induces COX-2 expression in podocytes via the EP4 receptor through a PKA-independent mechanism ☆ Wissam H. Faour, Kaede Gomi, Christopher R.J. Kennedy ⁎ Kidney Research Centre, Division of Nephrology, Department of Medicine, the Ottawa Hospital, Ottawa, Ontario, Canada K1H 8M5 Ottawa Health Research Institute, Ottawa, Ontario, Canada K1H 8M5 Department of Cellular and Molecular Medicine, University of Ottawa, Ottawa, Ontario, Canada K1H 8M5
a r t i c l e
i n f o
Article history: Received 2 April 2008 Received in revised form 23 July 2008 Accepted 11 August 2008 Available online 15 August 2008 Keywords: Podocyte Eicosanoids p38 MAPK cAMP AMPK Cyclooxygenase
a b s t r a c t Cyclooxygenase-2 (COX-2)-dependent prostaglandin E2 (PGE2) synthesis correlates with the onset of proteinuria and increased glomerular capillary pressure (Pgc) glomerular disease models. We previously showed that an in vitro surrogate for Pgc (cyclical mechanical stretch) upregulates the expression of both COX-2 and the PGE2 responsive E-Prostanoid receptor, EP4 in cultured mouse podocytes. In the present study we further delineate the signaling pathways regulating podocyte COX-2 induction. Time course experiments carried out in conditionally-immortalized mouse podocytes revealed that PGE2 transiently increased phosphorylated p38 MAPK levels at 10 min, and induced COX-2 protein expression at 4 h. siRNA-mediated knockdown of EP4 receptor expression, unlike treatment with the EP1 receptor antagonist SC 19220, completely abrogated PGE2-induced p38 phosphorylation and COX-2 upregulation suggesting the involvement of the EP4 receptor subtype. PGE2-induced COX-2 induction was abrogated by inhibition of either p38 MAPK or AMP activated protein kinase (AMPK), and was mimicked by AICAR, a selective AMPK activator, and by the cAMP-elevating agents, forskolin (FSK) and IBMX. Surprisingly, neither PGE2 nor FSK/IBMX-dependent p38 activation and COX-2 expression were blocked by PKA inhibitors or mimicked by 8-cPT-cAMP a selective EPAC activator, but were instead abrogated by Compound C, suggesting the involvement of AMPK. These results indicate that in addition to mechanical stretch, PGE2 initiates a positive feedback loop in podocytes that drives p38 MAPK activity and COX-2 expression through a cAMP/AMPKdependent, but PKA-independent signaling cascade. This PGE2-induced signaling network activated by increased Pgc could be detrimental to podocyte health and glomerular filtration barrier integrity. © 2008 Elsevier Inc. All rights reserved.
1. Introduction Podocytes are a major component of the glomerular filtration barrier [1]. Numerous studies show that podocyte injury is a key initiating event leading to filtration defects and progressive glomerular damage [2,3]. Cyclooxygenases (COX) and their metabolites e.g., prostaglandin E2 (PGE2), are involved in both hemodynamic and inflammatory-based pathogenesis of progressive nephropathies. Increased COX-2 levels correlate significantly with the extent of renal damage and high levels of both cyclooxygenase isoforms (COX-1 and COX-2) are reported in immunological and nonimmunological nephropathies, such as systemic lupus erythematosus [4], glomerulo☆ C. R. J. Kennedy is the recipient of a New Investigator Award from the Canadian Institutes of Health Research. Dr. W. H. Faour is the recipient of the KRESCENT (Kidney Foundation of Canada–Canadian Institutes of Health Research) joint fellowship. This work was supported by the Canadian Institutes of Health Research. ⁎ Corresponding author. Ottawa Health Research Institute, Division of Nephrology, Ottawa Hospital and University of Ottawa, 451 Smyth Rd., Rm 2501, Ottawa, Ontario, Canada K1H 8M5. Tel.: +1 613 562 5800 8529; fax: +1 613 562 5487. E-mail addresses: [email protected] (W.H. Faour), [email protected] (C.R.J. Kennedy). 0898-6568/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.cellsig.2008.08.007
sclerosis [5], Heymann nephritis [6] and renal ablation [7]. Furthermore, chronic use of COX inhibitors attenuates renal injury and albuminuria in rodents with subtotal nephrectomy, a model of chronic kidney disease [8]. A role for COX-derived prostaglandins in glomerular disease is further supported by recent studies showing that cyclooxygenase overexpression predisposes to proteinuria, podocyte damage, and loss. Cheng et al.[9] showed that podocyte-specific overexpression of COX-2 enhances podocyte sensitivity to either adriamycin-induced nephropathy or subtotal nephrectomy in mice. However, the specific prostaglandin receptor that underlies such damage has not been defined. PGE2, the major renal metabolite produced by COX-2, acts through at least four seven-transmembrane receptor subtypes (EP1, EP2, EP3, and EP4) [10,11]. A role for the EP1 receptor subtype in mediating glomerular damage and proteinuria is suggested by studies of spontaneously hypertensive rats and those with streptozotocininduced diabetic nephropathy. Treatment with an EP1 receptor antagonist reduced proteinuria and indices of glomerulosclerosis in both models of nephropathy [12]. Conversely, EP1-null mice are more susceptible than their wild type littermates to renal dysfunction in a model of nephrotoxic serum nephritis [13,14]. However, in these
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
studies, proteinuria levels were similar between wild type and EP1-null mice. PGE2 signaling in podocytes is mediated by the EP1 and/or the EP4 receptor subtypes [15]. We previously showed that equibiaxial mechanical stretch, a mimic of intraglomerular pressure, induces COX2 and PGE2 EP4 receptor expression in a mouse podocyte cell line [16]. Moreover, the enhanced EP4 receptor expression resulted in greater cAMP accumulation in response to exogenous PGE2, and yielded a significant effect upon the podocyte actin cytoskeleton [16]. However, the effect of such enhanced PGE2 signaling upon COX-2 expression remains unresolved. The present study was undertaken to characterize the signaling events initiated by PGE2 and leading to COX-2 upregulation in podocytes. We found that PGE2 initiates a positive feedback loop of COX-2 upregulation driven by EP4 receptor activation, a phenomenon entirely p38 MAPK dependant as p38 inhibition completely abrogated COX-2 induction. Our data support a cAMP-dependant, yet PKAindependent model of COX-2 induction as several PKA inhibitors failed to block PGE2 or FSK/IBMX mediated COX-2 induction. Specifically, we show that the heterotrimeric kinase AMPK [17,18] acts upstream of p38 MAPK to induce COX-2 expression. Triggering of this PGE2-driven signaling network in vivo by increased Pgc could be detrimental to podocyte health and the integrity of the glomerular filtration barrier. 2. Materials and methods 2.1. Cell culture Culture of conditionally-immortalized mouse podocytes, kindly provided by Dr. K. Endlich, was carried out as previously described [19]. Experiments were also carried out in another line of conditionally-immortalized mouse podocytes, kindly provided by Dr. P. Mundel, with similar results. Briefly, cells were grown on type-I collagencoated plastic tissue culture dishes in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 0.1 mg/ml streptomycin. Podocytes were routinely propagated at 33 °C in RPMI culture media supplemented with 10 U/ml mouse recombinant γ-interferon to promote the expression of the temperaturesensitive large T-antigen. Differentiation was induced by maintaining cultures at 37 °C in media without γ-interferon for 7–14 days. Cell culture materials were purchased from Life Technologies (Burlington, ON). For experimental protocols and conditions employed in this study, differentiated cells were trypsinized and transferred to the 6 well plates, and cultured for an additional 3 days. Following overnight serum starvation in RPMI-1640 medium supplemented with 0.1% FBS, podocytes were subjected to stimulation with PGE 2 (Cayman Chemical), forskolin/3-isobutyl-1-methylxanthine (IBMX) or 5aminoimidazole-4-carboxamide-1-{beta}-D-ribofuranoside (AICAR) (Cell Signaling) for 10 min to 16 h, or 8-(4-chlorophenylthio)-2′-Omethyladenosine 3′,5′-cyclic monophosphate (8-cPT-2′-O-MecAMP, Alexis Biochemicals) for 10 min to 4 h. Control cells (designated as ‘non-stimulated’) were cultured under identical conditions but were not exposed to bioactive chemicals. MAPK inhibitors SB 202190, (Tocris; Ellisville, MO), wortmannin (Sigma), compound C (Calbiochem), H89 (BIOMOL) and KT 5720 (Calbiochem) were employed at 25 µM, 200 nM, 3 µM, 10 µM and 2 µM respectively, in RPMI-1640 + 0.1% FBS, with a 30 min incubation prior to stimulation experiments. Indomethacin (10 µM) was added 45 min prior to all experimentation to prevent endogenous PGE2 synthesis.
2157
(Amersham Pharmacia Biotech; Baie d'Urfé, QC). Membranes incubated with blocking buffer for 1 h were probed overnight at 4 °C with rabbit antibodies directed against COX-2 (1:1000 dilution, Cayman Chemical Corporation — Ann Arbor, MI), actin (1:1000 dilution, SigmaAldrich), or a rabbit antibody (New England Biolabs) which recognizes the phosphorylated (activated) form of p38 MAP kinase (1:1000 dilution of anti-active p38). After incubation with an appropriate HRPconjugated secondary antibody (1:2000 to 1:40,000) (GE Healthcare UK), blots were incubated in chemiluminescent substrate (Pierce — Rockford, IL) and exposed to blue light-sensitive film (Kodak). Densitometric analyses of scanned blots were carried out using Alpha Imager software. 2.3. EP4 siRNA transfection assays Transient transfections were performed using HiPerfect transfection reagent (Qiagen) according to the manufacturer's protocol. Briefly, EP4 siRNA or scrambled non-silencing off-target siRNA (Qiagen) was diluted to a final concentration of 5 nM in 100 µL serum free HyClone RPMI medium (Fisher Scientific) followed by the addition of the transfection reagent. The resulting lipid-RNA complexes were incubated for 10–20 min at room temperature and then added drop-wise to the podocytes (1.5–3 × 105 cells) seeded onto collagen-coated 6-well cluster dishes. Cells were transfected for 48 h prior to use in experiments. 2.4. Real-time RT-PCR of EP4 mRNA in mouse podocytes Cells grown on type-I collagen-coated 60 mm dishes at 37 °C were transfected with EP4 siRNA as described above. RNA was extracted using an RNeasy kit according to the manufacturer's instructions (Qiagen) and subjected to reverse transcription. EP4 mRNA levels were assessed by real-time RT-PCR using TaqMan RT-PCR Master Mix Reagents and an ABI Prism 7000 Sequence Detection System (Applied Biosystems). Reactions were performed using the following
2.2. Immunoblotting Differentiated podocytes grown on 6 well plates were serumstarved, subjected to the appropriate experimental conditions, and were washed twice with ice-cold PBS. Cells lysed with 1× Laemmli buffer (62.5 mM Tris–HCl, pH 6.8, 2% w/v SDS, 10% glycerol, 50 mM DTT, 0.1% w/v bromophenol blue), were electrophoresed on 10% resolving gels, and electrotransferred to nitrocellulose membranes
Fig. 1. PGE2 induces COX-2 expression in mouse podocytes. Conditionally-immortalized differentiated mouse podocytes were incubated with vehicle (ethanol) as control (C) or 1 µM of PGE2 for 2 h, 4 h, 8 h and 16 h, as indicated and lysed with Laemmli buffer. Protein extracts were resolved by SDS-PAGE and immunoblotted with a COX-2 antibody. Expression levels were normalized to β-actin protein content as assessed by densitometric analysis (n = 3). ⁎P b 0.01 vs. vehicle control.
2158
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
conditions: 50 °C for 2 min, 95 °C for 10 min, and 45 cycles of 95 °C for 15 s and 60 °C for 1 min. Primers and TaqMan probe for murine EP4: sense, 5′-ATG GTC ATC TTA CTC ATC GCC AC-3′; antisense, 5′-CTT TCA
CCA CGT TTG GCT GAT-3′; probe 6FAM-CAT CTG CTC CAT TCC GCT CGT GGT-TAMRA. Values were normalized to GAPDH mRNA levels in each sample.
Fig. 2. PGE2-induced COX-2 expression is EP4 receptor and p38 MAPK-dependent. (A) Conditionally-immortalized differentiated mouse podocytes were incubated for 4 h with either vehicle, PGE2 alone (1 µM), or in combination with SB 202190 (25 µM), wortmannin (200 nM), SC 19220 (10 μM) or with sulprostone alone (10 μM). Cell lysates were resolved by SDSPAGE and immunoblotted with a COX-2 antibody. Expression levels were normalized to β-actin protein content as assessed by densitometric analysis (n = 3). ⁎P b 0.05 vs. vehicle control; ⁎⁎P b 0.01 vs. vehicle control. (B) Podocytes were transfected for 48 h with either 5 nM scrambled siRNA or 5 nM EP4-siRNA and subsequently stimulated for 4 h with either vehicle or PGE2 (1 μM) as indicated. COX-2 protein levels were determined as in (A). (C and D) Podocytes were transfected for 48 h with either scrambled 5 nM siRNA or 5 nM EP4-siRNA. (C) EP4 mRNA levels were quantified by analyzing total cellular RNA by RT-PCR using specific TaqMan primer/probe sets and normalizing the values to GAPDH levels. (D) Cells were stimulated with PGE2 (1 μM) or FSK (10 μM) in the presence of 0.5 mM IBMX. Intracellular cAMP levels were assayed as described in the Materials and methods (n = 2).
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
2159
Fig. 3. Time and dose dependent p38 MAPK phosphorylation by PGE2. Conditionally-immortalized differentiated mouse podocytes were stimulated with 1 μM PGE2 over a range of time points as indicated (Fig. 3A) or over a range of PGE2 concentrations (10− 9M–10− 6M) for 10 min (Fig. 3B). Cells were lysed with Laemmli buffer and analyzed by western blot using a phospho-p38 antibody and normalized for total p38 protein content using a p38 antibody and analyzed by densitometry (n = 3). ⁎P b 0.01 vs. vehicle control.
2.5. cAMP assay Podocytes were grown on collagen I-coated 6-well cluster dishes at 37 °C and transfected with EP4 or scrambled siRNA. Cells were preincubated with HyClone RPMI-1640 (Fisher Scientific) containing 0.1% FBS, 5 μM indomethacin (Sigma-Aldrich Canada Ltd.), and 0.5 mM IBMX (Sigma-Aldrich) for 10 min. Cells were then stimulated for a further 10 min with either PGE2 (1 μM; Cayman Chemical Company) or forskolin (10 μM; Sigma-Aldrich). Intracellular cAMP was determined by cAMP EIA Kit purchased from Cayman Chemical.
3.2. PGE2–COX-2 induction is mediated through activation of the EP4 receptor subtype Bek and coworkers [15] showed that both EP1 and EP4 receptor subtypes, but not EP2 and EP3 subtypes, are expressed in conditionally-immortalized mouse podocytes. We previously demonstrated that mechanical stretch of these same cells increases EP4, but not EP1 subtype expression. Stretch thereby increases the podocyte's response
2.6. Statistics Data are expressed as means of duplicate determinations from individual experiments and are presented as means ± SEM where n ≥ 4 or means ± SD where n = 3 experiments. Statistical significance was accepted at P b 0.05 as determined by ANOVA followed by a Newman– Keuls multiple comparisons test or alternatively determined by a paired t-test where appropriate. 3. Results 3.1. PGE2 induces COX-2 expression in cultured mouse podocytes Podocyte COX-2 levels are increased in vivo in models of glomerular damage that involve elevated intraglomerular capillary pressure [8]. Such forces can be mimicked in vitro by the application of equibiaxial mechanical stretch which induces expression of both COX2 and the EP4 receptor subtype in cultured podocytes [16]. In the present study, we asked whether PGE2 in podocytes may act in an autocrine manner through the EP4 receptor to further enhance COX-2 expression. We therefore carried out time course experiments in which cultured mouse podocytes were incubated with 1 µM PGE2. As shown in Fig. 1, western blot analysis indicates that PGE2 upregulates COX-2 protein in a time dependent manner. PGE2 rapidly induced COX-2 protein, reaching significant levels after 2 h of stimulation and a steady state level after 4 h, declining gradually until returning to near baseline levels by 16 h.
Fig. 4. PGE2-induced COX-2 expression is mimicked by cAMP-elevating agents. Conditionally-immortalized differentiated mouse podocytes were incubated with vehicle control (C) or stimulated for 4 h with PGE2 (1 μM), FSK (10 μM) + IBMX (0.5 mM), or in combination with SB 202190 (25 μM). Cells were lysed with Laemmli buffer and protein extracts were resolved by SDS-PAGE and immunoblotted with a COX2 antibody. Expression levels were normalized to β-actin protein content and analyzed by densitometry (n = 3). ⁎P b 0.05 vs. vehicle control; ⁎⁎P b 0.01 vs. vehicle control.
2160
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
to PGE2, to yield a greater synthesis of intracellular cAMP [16]. To determine which EP subtype accounted for the observed effects of PGE2 upon COX-2 induction, mouse podocytes were incubated with PGE2 for 4 h in the presence or absence of EP agonist/antagonists or EP4 expression was knocked down by siRNA. While PGE2 upregulated COX-2 protein levels, sulprostone (1 µM), an EP1/EP3 receptor agonist, failed to induce a statistically significant increase in COX-2 expression (Fig. 2A). Likewise, SC 19220 (1 µM), a relatively selective EP1 receptor antagonist, failed to reverse PGE2 dependant COX-2 upregulation. Additionally, SC 19220 by itself, or in combination with PGE2, induced COX-2 protein expression indicating a possible non-specific effect (Fig. 2A). The involvement of the EP4 subtype was further suggested as PGE2 failed to induce COX-2 protein levels when podocytes were transfected with EP4-siRNA for 48 to 72 h prior to experimentation (Fig. 2B, C). Knockdown of the receptor was confirmed by quantitative RT-PCR of EP4 mRNA levels, as well as by failure of the EP4-siRNA transfected cells to elicit a cAMP response following PGE2 stimulation of the podocytes (Fig. 2). These findings suggest that the EP4 receptor subtype mediates PGE2-stimulated COX-2 expression in podocytes. 3.3. Time course and dose dependent activation of p38 MAPK by PGE2 Previous work in our lab showed that COX-2 induction in response to mechanical stretch is p38 MAPK-dependent. Stretch induced maximal p38 phosphorylation after 10 min in cultured mouse
podocytes [16]. We asked whether PGE2 transduces its effects on COX-2 expression in podocytes through a similar activation of p38 MAPK. As shown in Fig. 2, PGE2-stimulated COX-2 induction was p38dependant as the selective p38 inhibitor, SB 202190, blocked this effect, whereas wortmannin (a PI3K/AKT inhibitor) was without effect. We also performed both time course and dose response experiments of p38 MAPK phosphorylation in response to PGE2. As judged by western blot, stimulation of mouse podocytes with 1 µM of PGE2 resulted in a transient but statistically significant 3-fold stimulation of p38 phosphorylation after 10 min which declined rapidly thereafter (Fig. 3A). Maximal p38 phosphorylation levels were achieved with 1 µM PGE2 (Fig. 3B). These findings clearly demonstrate that PGE2 can activate p38 MAPK to enhance COX-2 expression in cultured podocytes. 3.4. PGE2 upregulates COX-2 through a cAMP-dependant but PKAindependent mechanism In order to examine the role of cAMP and cAMP-dependent protein kinase A (PKA) in PGE2-mediated p38-dependent COX-2 induction, we performed western blot experiments using the cAMP-elevating agents, forskolin (10 μM) and IBMX (0.5 mM), and the PKA-selective inhibitors, H89 (25 μM) and KT5720 (2 µM). Cultured podocytes incubated with PGE2 or FSK/IBMX for 4 h showed significant COX-2 protein biosynthesis (Fig. 4). In keeping with a role for p38 MAPK in COX-2 expression, SB 202190 abrogated both PGE2 and FSK/IBMX-
Fig. 5. PGE2-induced p38 MAPK activation and COX-2 expression requires cAMP, but is independent of PKA. Cells were treated with vehicle or stimulated for 4 h with PGE2 (1 μM), FSK (10 μM) + IBMX (0.5 mM), or in combination with either H89 (10 μM) or KT5720 (2 μM) (Fig. 5A and 5C), 10 min (Fig. 5B) or 1 h (Fig. 5D). Cell lysates were analyzed by western blot using antibodies against either COX-2, phospho-p38, or phospho-CREB-1/ATF-1. Expression levels were normalized to β-actin protein content and signal intensity quantified by densitometry (n = 3). 5A) ⁎⁎P b 0.01 vs. vehicle control or H89 alone; ⁎P b 0.05 vs. vehicle control.
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
induced COX-2 expression. However, H89 failed to block either COX-2 upregulation (Fig. 5A) or p38 MAPK phosphorylation driven by either PGE2 or FSK/IBMX (Fig. 5B). Surprisingly, incubation with H89 alone tended to increase COX-2 expression and phospho-p38 levels. However, experiments performed using another PKA-selective inhibitor (i.e., KT5720) were likewise without effect on PGE2-induced COX-2 expression (Fig. 5C). Inhibition of PKA activation by KT5720 was confirmed by experiments showing that a FSK/IMBX mediated increase of phosphorylated CREB, a known target of PKA could be abrogated by incubation with this inhibitor (Fig. 5D). These data suggest that a PKA-independent mechanism likely underlies PGE2dependent COX-2 induction. 3.5. PGE2-induced p38 phosphorylation and COX-2 protein upregulation occur independently of EPAC activation EPAC1, a cAMP-dependent guanine nucleotide exchange factor for the small GTPase RAP, is a well known mediator of PKA-independent signaling. We therefore investigated the role of EPAC1 in PGE2 mediated induction of COX-2 protein translation. Immunoblotting of lysates from podocytes treated with a specific and cell membrane
2161
permeable activator of EPAC1 and EPAC2, 8-cPT-2′-O-Me-cAMP (8cPT-cAMP) [20–22] showed that 8-cPT-cAMP had no effect on both p38-MAPK phosphorylation and COX-2 protein levels. These data suggest that EPAC does not mediate PGE2-induced COX-2 expression in podocytes (Fig. 6). 3.6. AMPK activation mediates PGE2 upregulation of COX-2 expression and p38 phosphorylation The heterotrimeric kinase, AMPK, is activated by intracellular energy deficits brought about by increased AMP-to-ATP ratio when the cell's energy state is low [17,18]. Other reports indicate that AMPK activation can induce p38 MAPK activity [23]. In order to test whether AMPK plays a role in PGE2-mediated COX-2 gene induction in mouse podocytes, we performed western blot experiments using the AMPKselective inhibitor, compound C, and the AMPK activator, 5-aminoimidazole-4-carboxamide-1-D-ribofuranoside (AICAR). Cultured podocytes incubated with either PGE2 or FSK/IBMX for 4 h exhibited significant COX-2 induction (Fig. 7A) that was abrogated by coincubation with compound C. Activation of AMPK by incubation of podocytes with AICAR significantly induced COX-2 expression that was likewise susceptible to inhibition by compound C. Finally, inhibition of AMPK by compound C abolished both PGE2 and FSK/IBMX induced p38 MAPK phosphorylation (Fig. 7B). 4. Discussion
Fig. 6. PGE2-induced p38 MAPK activation and COX-2 expression are independent of EPAC. Cells were treated with vehicle control (c) or stimulated with PGE2 (1 μM), FSK (10 μM) + IBMX (0.5 mM), or 8-cPT (EPAC activator) for 4 h or 10 min (Fig. 6) as indicated. Cell lysates were analyzed by western blot using either COX-2 antibody or phospho-p38 antibody and normalized to β-actin protein content and signal intensity quantified by densitometry.
Induction of glomerular COX-2 is reported in various glomerularbased diseases, correlating significantly with proteinuria and the extent of renal damage [4–7]. COX-2 is barely detectable in healthy podocytes, yet levels rise rapidly during disease progression. Mesangial cells strongly upregulate COX-2 mRNA and protein following stimulation with proinflammatory cytokines, such as IL-1β, by a p38dependant mechanism [24]. Similarly, increased podocyte COX-2 levels are evident in rodent models of glomerular injury, including diabetic nephropathy [25,26], subtotal renal ablation [27], and adriamycin-induced nephrotoxicity [9]. Furthermore, podocyte-specific transgenic overexpression of COX-2 renders mice more susceptible to the development of proteinuria and renal injury following puromycin [28] or adriamycin-induced nephrotoxicity [9]. Importantly, inhibition of prostaglandin synthesis abrogates proteinuria and reduces disease progression in each of these models. Together, these findings strongly suggest that the induction of COX-2 in the glomerular tuft is an important step in the etiology of damage to the filtration barrier. However, the mechanisms and signal transduction pathways that regulate podocyte COX-2 expression remain incompletely resolved. PGE2 is a major COX-2 metabolite synthesized by glomerular podocytes. The role played by PGE2 in podocyte biology is unclear. However, our data clearly indicate that this eicosanoid can significantly and rapidly induce COX-2 expression in conditionally-immortalized mouse podocytes. PGE2 can potentially exert its effects on podocytes via either EP1 and/or EP4 receptor subtypes, both of which are expressed in this cell type [15]. EP4 receptor activation yields cAMP production in order to regulate target genes in podocytes [29], while little is known regarding signaling through the EP1 receptor other than it promotes transient intracellular Ca2+ fluxes [15]. Our data favor an EP4-dependent mechanism for COX-2 gene induction since siRNAmediated knockdown of the EP4 receptor abolished the ability of PGE2 to induce COX-2 expression. Furthermore, an EP1-selective antagonist failed to inhibit PGE2-dependent COX-2 upregulation. It was interesting to note that SC-19220 promoted COX-2 induction, suggesting possible off-target effects. However, the participation of the EP1 subtype is unlikely as cAMP production by forskolin and IBMX significantly upregulated COX-2 expression while sulprostone, an EP1/ EP3 agonist was without effect. Therefore a positive feedback loop
2162
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
Fig. 7. AMPK mediates PGE2–COX-2 induction and p38 MAPK phosphorylation. Fig. 7A) Differentiated podocytes were stimulated for 4 h with either vehicle control (c), PGE2 (1 μM), FSK (10 μM) + IBMX (0.5 mM), AICAR (2 mM), or in combination with compound C (comp C) (3 μM). The cells were then processed for COX-2 immunoblotting as described earlier, normalized for β-actin protein content and analyzed by densitometry (n = 3). ⁎P b 0.01 vs. vehicle control and compound C alone; ⁎⁎P b 0.001 vs. vehicle control and compound C alone. Fig. 7B) Cells were stimulated as in 7A, but for 10 min and the cell lysates analyzed for phosphorylated p38, normalized for β-actin protein content and analyzed by densitometry (n = 3).
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
appears to operate wherein COX-2-derived PGE2 acts in an autocrine manner via the EP4 subtype to further enhance both COX-2 expression and PGE2 synthesis. How is this signaling loop achieved? Our previous studies showed that mechanical stretch of podocytes, a mimic of hydrostatic forces encountered in vivo, yields COX-2 expression in a p38 MAPKdependent manner [16]. We investigated whether PGE2 would similarly require p38 activation to increase COX-2 expression. Other studies showed that COX-2 mRNA induction can be integrated within a posttranscriptional feedback loop involving PGE2 production and p38 activation [30,31]. In this regard, PGE2, by promoting p38 activation, stabilizes COX-2 mRNA induced by IL-1β [31]. However, in those studies PGE2 was shown to induce a sustained p38 phosphorylation that enhanced COX-2 mRNA stability [31]. In contrast, our data indicate that p38 is transiently activated (10 min) by PGE2 only to rapidly decline thereafter, suggesting a transcriptional model of COX-2 gene regulation in podocytes. The classical pathway leading to p38 MAPK activation proceeds through a GTPase signaling cascade involving raf/ras and MKK3/6 [32,33]. However, our findings suggest that PGE2-induced p38 signaling involves a distinct signaling cascade that requires cAMP production. Such p38 MAPK activity likely proceeds through a PKAindependent mechanism since several PKA-selective inhibitors (i.e., H89, KT5720; Fig. 5) each failed to block PGE2-induced p38 MAPK phosphorylation and COX-2 induction. A recent study demonstrated that PGE2-induced osteoclast differentiation requires p38 MAPK activation, yet signals through PKA-dependent phosphorylation of transforming growth factor beta-activated kinase 1 (TAK1), a mitogen activated kinase kinase [34]. However, these authors demonstrated that PGE2 alone failed to induce osteoclast differentiation, and that TAK1 activity requires additional phosphorylation achieved by coincubation with a TNF family member, RANKL [34]. A similar scenario is unlikely in PGE2-stimulated podocytes as neither PGE2 nor FSK/IBMX induced immuno-detectable phosphorylated TAK1 (unpublished results). Alternatively, we considered another PKAindependent signaling route that involves the exchange protein activated by cAMP-1 (EPAC1), a cAMP-dependent guanine nucleotide exchange factor for the small GTPase RAP [20–22]. In cerebellar neurons, EPAC mediates p38 MAPK activation, independent of PKA [22]. However, our data suggest that EPAC is not implicated in PGE2dependent p38 activation and COX-2 induction in podocytes. Accordingly, the EPAC-selective activator, 8-cPT-cAMP failed to stimulate either p38 MAPK phosphorylation or COX-2 induction (Fig. 6). Taken together, our results suggest that the ability of PGE2 to increase COX-2 expression in podocytes does not rely upon the recruitment of either TAK1 or EPAC. Interestingly, a similar cAMPdependent, PKA-independent and EPAC-independent pathway may operate in mouse neonatal cardiac fibroblasts, where β2-adrenergic signaling induces p38-dependent IL-6 transcription [35]. We next tested whether another cAMP-dependent, PKA-independent signaling route could account for the effects of PGE2 on COX-2 induction. AMP activated protein kinase (AMPK) serves as an energy sensor regulated by the intracellular AMP-to-ATP ratio, and it promotes glucose uptake via the GLUT4 transporter in a variety of mammalian cell types. AMPK activation was shown to induce p38 MAPK phosphorylation in a rat liver cell line [36] and in cardiac muscle [23]. In the latter studies, AMPK was shown to recruit the scaffold protein, transforming growth factor-β-activated protein kinase 1-binding protein 1 (TAB1) which then facilitated p38 autophosphorylation. Conversely, this signaling paradigm may not be broadly applicable since activation of p38 MAPK appears to be independent of AMPK in skeletal muscle [37]. In the present study, activation of podocyte AMPK by AICAR mimicked, while inhibition of AMPK by compound C, abrogated both PGE2 and FSK/IBMX induced COX-2 expression. This AMPK-driven activation of p38 represents a novel link between podocyte dependent energy status and downstream targeted genes, such as COX-2. As podocyte COX-2
2163
expression increases in rodent models of diabetic nephropathy, our findings suggest that AMPK may likewise be induced rendering the podocyte's energy status extremely sensitive to changes in glucose levels which in turn modulate critical cellular events involving control of cellular hypertrophy, adhesion and survival which are hallmarks of diabetic podocytes. Interestingly, a recent study demonstrated that AMPK-induced p38 phosphorylation promotes actin cytoskeleton reorganization and enhances cardiac lipoprotein lipase activity in a rat model of acute diabetes [38]. The authors concluded that such changes in cardiac fatty acid metabolism could translate into increased cardiovascular damage following diabetes. Our previous studies indicated that mechanically stretched podocytes, which exhibit enhanced COX-2 and EP4 receptor expression, respond to PGE2 by dramatically reordering their actin cytoskeleton [16]. However, the impact of PGE2-driven AMPK activity in podocytes in the context of diabetic nephropathy, as well as other glomerular injury contexts in vivo awaits investigation. In summary, our data shows that the EP4 receptor subtype can trigger a signaling paradigm that promotes AMPK activation and p38dependent COX-2 induction to further enhance PGE2 synthesis within the podocyte. This combination of COX-2 expression and ensuing PGE2 production may have deleterious effects on the podocyte leading to a compromised glomerular filtration barrier. Acknowledgments We thank Drs. K. Endlich and P. Mundel for the kind gift of the murine podocyte cell lines. References [1] H. Pavenstadt, W. Kriz, M. Kretzler, Physiological Reviews 83 (1) (2003) 253. [2] S.J. Shankland, Kidney international 69 (12) (2006) 2131. [3] B.L. Wharram, M. Goyal, J.E. Wiggins, S.K. Sanden, S. Hussain, W.E. Filipiak, T.L. Saunders, R.C. Dysko, K. Kohno, L.B. Holzman, R.C. Wiggins, Journal of the American Society of Nephrology 16 (10) (2005) 2941. [4] S. Tomasoni, M. Noris, S. Zappella, E. Gotti, F. Casiraghi, S. Bonazzola, A. Benigni, G. Remuzzi, Journal of the American Society of Nephrology: JASN 9 (7) (1998) 1202. [5] W. Weichert, A. Paliege, A.P. Provoost, S. Bachmann, American Journal of Physiology. Renal Physiology 280 (4) (2001) F706. [6] T. Takano, A.V. Cybulsky, American Journal of Pathology 156 (6) (2000) 2091. [7] J.L. Wang, H.F. Cheng, M.Z. Zhang, J.A. McKanna, R.C. Harris, The American Journal of Physiology 275 (4 Pt 2) (1998) F613. [8] C.K. Fujihara, G.R. Antunes, A.L. Mattar, N. Andreoli, D.M. Malheiros, I.L. Noronha, R. Zatz, Kidney international 64 (6) (2003) 2172. [9] H. Cheng, S. Wang, Y.I. Jo, C.M. Hao, M. Zhang, X. Fan, C. Kennedy, M.D. Breyer, G.W. Moeckel, R.C. Harris, Journal of the American Society of Nephrology 18 (2) (2007) 551. [10] Y. Boie, R. Stocco, N. Sawyer, D.M. Slipetz, M.D. Ungrin, F. Neuschafer-Rube, G.P. Puschel, K.M. Metters, M. Abramovitz, European journal of pharmacology 340 (2–3) (1997) 227. [11] M. Kiriyama, F. Ushikubi, T. Kobayashi, M. Hirata, Y. Sugimoto, S. Narumiya, British journal of pharmacology 122 (2) (1997) 217. [12] T. Suganami, K. Mori, I. Tanaka, M. Mukoyama, A. Sugawara, H. Makino, S. Muro, K. Yahata, S. Ohuchida, T. Maruyama, S. Narumiya, K. Nakao, Hypertension 42 (6) (2003) 1183. [13] S. Rahal, L.I. McVeigh, Y. Zhang, Y. Guan, M.D. Breyer, C.R. Kennedy, Canadian journal of physiology and pharmacology 84 (8–9) (2006) 877. [14] C.R. Kennedy, H. Xiong, S. Rahal, J. Vanderluit, R.S. Slack, Y. Zhang, Y. Guan, M.D. Breyer, R.L. Hebert, American journal of physiology.Renal physiology 292 (2) (2007) F868. [15] M. Bek, R. Nusing, P. Kowark, A. Henger, P. Mundel, H. Pavenstadt, Journal of the American Society of Nephrology: JASN 10 (10) (1999) 2084. [16] L.C. Martineau, L.I. McVeigh, B.J. Jasmin, C.R. Kennedy, American Journal of Physiology. Renal Physiology 286 (4) (2004) F693. [17] D.G. Hardie, Nature reviews.Molecular cell biology 8 (10) (2007) 774. [18] D.G. Hardie, Science (New York, N.Y.) 315 (5819) (2007) 1671. [19] P. Mundel, J. Reiser, AeZÃiMa Borja, H. Pavenstädt, G.R. Davidson, W. Kriz, R. Zeller, Experimental Cell Research 236 (1) (1997) 248. [20] K.S. Lyle, J.H. Raaijmakers, W. Bruinsma, J.L. Bos, J. de Rooij, Cell Signal 20 (6) (2008) 1104. [21] H. Rehmann, A. Wittinghofer, J.L. Bos, Nature reviews. Molecular Cell Biology 8 (1) (2007) 63. [22] J. Ster, F. DeBock, N. Guerineau, A. Janossy, S. Barrere-Lemaire, J. Bos, J. Bockaert, L. Fagni, Proceedings of the National Academy of Sciences 104 (7) (2007) 2519. [23] J. Li, E.J. Miller, J. Ninomiya-Tsuji, R.R. Russell III, L.H. Young, Circulation research 97 (9) (2005) 872.
2164
W.H. Faour et al. / Cellular Signalling 20 (2008) 2156–2164
[24] Z. Guan, S.Y. Buckman, B.W. Miller, L.D. Springer, A.R. Morrison, Journal of Biological Chemistry 273 (44) (1998) 28670. [25] R. Komers, J.N. Lindsley, T.T. Oyama, S. Anderson, Clinical and Experimental Pharmacology and Physiology 34 (1–2) (2007) 36. [26] R. Komers, J.N. Lindsley, T.T. Oyama, W.E. Schutzer, J.F. Reed, S.L. Mader, S. Anderson, Journal of Clinical Investigation 107 (7) (2001) 889. [27] J.L. Wang, H.F. Cheng, S. Shappell, R.C. Harris, Kidney International 57 (6) (2000) 2334. [28] Y.I. Jo, H. Cheng, S. Wang, G.W. Moeckel, R.C. Harris, Nephron. Experimental Nephrology 107 (3) (2007) e87. [29] N. Endlich, K. Endlich, Microscopy Research and Technique 57 (4) (2002) 228. [30] P. Akarasereenont, K. Techatrisak, S. Chotewuttakorn, A. Thaworn, Mediators of inflammation 8 (6) (1999) 287.
[31] W.H. Faour, Y. He, Q.W. He, M. de Ladurantaye, M. Quintero, A. Mancini, J.A. Di Battista, Journal of Biological Chemistry 276 (34) (2001) 31720. [32] E. Herlaar, Z. Brown, Molecular Medicine Today 5 (10) (1999) 439. [33] J. Kyriakis, J. Avruch, Physiological Reviews 81 (2) (2001) 807. [34] Y. Kobayashi, T. Mizoguchi, I. Take, S. Kurihara, N. Udagawa, N. Takahashi, Journal of Biological Chemistry 280 (12) (2005) 11395. [35] F. Yin, Y.Y. Wang, J.H. Du, C. Li, Z.Z. Lu, C. Han, Y.Y. Zhang, Journal of Molecular and Cellular Cardiology 40 (3) (2006) 384. [36] X. Xi, J. Han, J.Z. Zhang, Journal of Biological Chemistry 276 (44) (2001) 41029. [37] R.C. Ho, N. Fujii, L.A. Witters, M.F. Hirshman, L.J. Goodyear, Biochemical and Biophysical Research Communications 362 (2) (2007) 354. [38] M.S. Kim, G. Kewalramani, P. Puthanveetil, V. Lee, U. Kumar, D. An, A. Abrahani, B. Rodrigues, Diabetes 57 (1) (2008) 64.