|Home | About | Journals | Submit | Contact Us | Français|
Increased arterial pressure, angiotensin II (Ang II) and cytokines each result in feedback inhibition of renin gene expression. Since Ang II and cytokines can stimulate reactive oxygen species (ROS) production, we tested the hypothesis that oxidative stress may be a mediator of this inhibition. Treatment of renin-expressing As4.1 cells with the potent cytokine TNFα caused an increase in the steady state levels of cellular ROS, which was reversed by the antioxidant N-acetylcysteine. Exogenous H2O2 caused a dose- and time-dependent decrease in the level of endogenous renin mRNA, and decreased the transcriptional activity of a 4.1kb renin promoter fused to luciferase which was maximal when the renin enhancer was present. The effect of H2O2 appeared to be specific to renin as there was no change in expression of β-actin or cyclophillin mRNA, nor transcriptional activity of the SV40 promoter. The TNFα-induced decrease in renin mRNA was partially reversed by either N-acetylcysteine or panepoxydone, an NFκB inhibitor. Interestingly, H2O2 did not induce NFκB in As4.1 cells, and panepoxydone had no effect the down-regulation of renin mRNA by H2O2. The transcriptional activity of a cAMP response element (CRE)-luciferase construct was decreased by both TNFα and H2O2. These data suggest that cellular ROS can negatively regulate renin gene expression via an NFκB-independent mechanism involving the renin enhancer and inhibiting CRE-mediated transcription. Our data further suggest that TNFα decreases renin expression through both NFκB-dependent and NFκB-independent mechanisms, the latter involving the production of ROS.
Renin is the rate limiting step in the catalytic processing of angiotensinogen to angiotensin-I which is further hydrolyzed to angiotensin II (Ang II) and Ang(1-7). Renin is tightly regulated at the transcriptional, post-transcriptional, and translational levels, and its secretion is controlled by a multitude of physiological cues. The cues are derived from both systemic and local signals, including the sympathetic nervous system, circulating and tissue angiotensin peptides, endocrine factors, nitric oxide and cytokines. One hallmark of renin regulation in vivo is negative feedback which serves to tightly regulate renin expression and release in response to Ang II, NaCl at the macula densa, and renal perfusion pressure.1 The mediators of negative feedback have been the subject of extensive investigation, but the mechanisms remain incompletely understood.
Ang II can stimulate cytokine production in a variety of cell types, and cytokines are among the inhibitory signals regulating renin transcription.2-5 The cytokine TNFα, was reported to be required to mediate the drinking and pressor responses to Ang II, and is a strong negative regulator of renin expression.3, 6-8 Like Ang II, cytokines can stimulate the production of ROS and cause oxidative stress.9 Interestingly, decreasing oxidative stress in the SHR decreases blood pressure and increases plasma renin activity.10 Based on these observations, we considered the hypothesis that TNFα may cause production of ROS, and that ROS may negatively regulate renin expression. We employed As4.1 cells, an in vitro model of juxtaglomerular cells which express endogenously renin mRNA to test this hypothesis.11 A previous study reported that production of cellular ROS can be induced in As4.1 cells.12
We show that TNFα stimulates ROS production and its inhibitory effects on renin expression can be partially reversed by an antioxidant. We further show that H2O2 decreases renin expression and renin promoter activity through an NFκB-independent, but a cAMP response element (CRE)-dependent mechanism.
As4.1 cells are available through the American Type Culture Collection (CRL2193). Cells were plated 24 hours before transfection in DMEM containing 10% FBS and penicillin-streptomycin (Gibco). Prior to transfection, cells were switched to 1% FBS, then transiently transfected with a master mix containing luciferase reporter vector using FuGENE-6 (Roche). After 5-hours, cells were split by trypsin and plated into 6-well plates. Increasing doses of H2O2 were added to the media 24 hours after transfection. The cells were harvested after 48hrs, lysed, and luciferase activity was determined using the Dual-Luciferase Reporter (DLR) kit (Promega). RSV-LUC and pRLSV40 (Promega) were used as positive and internal controls, respectively. Luciferase activity was normalized to renilla and to total cellular protein and then calculated as a percentage of RSV promoter activity. Luciferase activity assays in each experiment were performed in duplicate, and the average of the 2 readings represented 1-data point. As4.1 cells were treated with 0.15 mM xanthine and 0.2 U xanthine oxidase (Sigma) in medium (1% FBS). Details on transcriptional blockade is provided in the Supplemental Methods (please see http://hyper.ahajournals.org).
The luciferase (LUC) reporter vectors m2.6, mE2.6, 4.1, 4.1-μHRE, were described previously.13, 14 Site-directed mutagenesis was performed using the QuikChange Site-Directed Mutagenesis Kit (Stratagene), and were confirmed by DNA sequencing.14 Additional information is in the Supplemental Methods. TA-luc and CRE-Luc reporter vectors were from Panomics. The vector Ad.NFκBLuc was described previously.15 Infection of Ad.NFκBLuc was performed in DMEM media containing 1% fetal bovine serum at 200 and 500 multiplicity of infection (MOI) in As4.1 and MCF-7 cells, respectively. 24 h post-infection, cells were stimulated with 1 ng/ml IL-1β, 10 ng/ml TNFα, 200μM and 1mM H2O2. Cells were lysed for measurement of luciferase activity 6 hours post-treatment.
Total RNA was extracted using the Qiagen RNeasy kit. T3 RNA polymerase was used to prepare antisense RNA for RNase protection probes. The protected fragments are 326 nucleotides for mouse REN mRNA, 105 bases for cyclophilin, 249 bp for β-actin, and 80 bp for 18S. RNase protection was performed using the RPAIII kit (Ambion Inc, Austin, TX). Protected fragments were quantified with a phosphorimager (GE Healthcare).
Dihydroethidium (DHE) and 2′, 7′- dichlorofluorescein (DCFH-DA) was used to measure the levels of cellular ROS. DHE-fluorescence is used to measure superoxide, while DCFH-fluorescence is used to measure H2O2. As4.1 cells were washed with Hanks buffer salt solution (HBSS) and labeled with 10 μM DHE (Invitrogen) for 45 min and DCFH-DA (Invitrogen, Molecular Probes) for 15 min at 37°C in the dark. Cells were trypsinized with ice cold phenol-free trypsin/EDTA. Trypsin was inactivated with medium containing 10% FBS. Cells were collected and resuspended in HBSS. Samples were analyzed using flow cytometry employing a 488 nm excitation laser. DHE was detected by a 585 nm band pass emission filter and DCFH was detected using a 530 nm band pass emission filter as previously described.16 Mean fluorescence intensity (MFI) was analyzed using FlowJo software (Tree Star). Samples were corrected for autofluorescence using unlabeled cells using DMSO.
Microarray analysis was described in the Supplemental Methods. The complete dataset is available at the Gene Expression Omnibus (GEO) at NCBI (series accession: GSE14243).
Data are presented as means ± SEM. Group comparison of the data is accomplished by one-way ANOVA with Bonferroni correction for multiple testing or by Student's t-test as appropriate using SigmaStat (SPSS Scientific).
As previously reported, TNFα markedly attenuates the level of endogenous renin mRNA and transcriptional activity of the renin promoter in As4.1 cells (Figure S1, please see http://hyper.ahajournals.org).3 We next measured the steady state levels of cellular ROS in response to TNFα in As4.1 cells (Figure 1). Antimycin A (10 μM), an inhibitor of mitochondrial electron transport augmented ROS (superoxide) levels in As4.1 cells. TNFα resulted in a time-dependent increase in both DHE and DCFH fluorescence. The specificity of the DCFH-assay for detection of cellular ROS was confirmed by the decrease in fluorescence caused by both H2O2 and TNFα in the presence of the antioxidant N-acetylcysteine (NAC).
As H2O2 is a diffusable ROS, and therefore a potential paracrine factor in the regulation of renin expression, we focused on H2O2. There was no decrease in cell viability (measured by trypan blue exclusion) when As4.1 cells were treated for 24 hours with 200 μM H2O2. Treatment of As4.1 cells with H2O2 caused a dose-dependent (Figure 2) and time-dependent (Figure S2) decrease in the levels of endogenous renin mRNA. Notably there was no change in the level of cyclophilin or β-actin mRNA indicating that the response was specific to renin. H2O2 generated by xanthine and xanthine oxidase decreased renin mRNA levels by nearly 70% after 24 hours (Figure S3). H2O2 and actinomycin D (AMD) had similar effects on renin mRNA levels suggesting the effect may be transcriptional (Figure S4A). The rate of renin mRNA decay for the first 9 hours after the addition of AMD was similar in H2O2-treated and untreated samples (Figure S4B). However, there was a steeper decline in renin mRNA between 9 and 24 hours in H2O2 pretreated cells suggesting a potential post-transcriptional component.
We determined if the renin mRNA response to H2O2 could be attributed to a decrease in renin promoter activity. As4.1 cells were transfected transiently with seven constructs carrying different lengths of the mouse renin promoter, with or without the enhancer, or containing mutations in critical transcription factor binding sites (Figure 3). Constructs carrying the renin enhancer were much more active transcriptionally than constructs lacking the enhancer (compare 2.6LUC to mE2.6LUC and 4.1LUC). Mutation of the CRE, hormone response element (HRE), or the HoxD10 binding site markedly attenuated promoter activity (Figure 3A). H2O2 caused a greater reduction in transcriptional activity of the renin promoter in constructs carrying the enhancer (reduced to 18-21% of baseline), than in constructs lacking the enhancer (reduced to 43-47% of baseline) or carrying mutations in critical elements in the enhancer or promoter (reduced to 33-48% of baseline)(Figure 3B). There was no effect of H2O2 on the activity of the renilla luciferase control nor the SV40 promoter/enhancer (pGL2C) validating the selectivity of the response.
The anti-oxidant NAC by itself had no effects on renin mRNA levels (Figure 4). However, NAC blunted the attenuation of renin mRNA by TNFα suggesting that a portion of the TNFα response is attributable to ROS. Since TNFα can inhibit renin expression through NFκB, and H2O2 was reported to induce NFκB, we determined if H2O2 acts through NFκB in As4.1 cells.3, 7 NFκB activity was assessed using an adenovirus containing a luciferase reporter driven by four repeats of the NFκB consensus site (Figure 5). TNFα and H2O2 each induced NFκB transcriptional activity in MCF7 cells, thus replicating previous results and validating the assay.17 TNFα induced NFκB transcriptional activity in As4.1 cells. Unexpectedly, H2O2 (200 μM or 1 mM) failed to induce NFκB-mediated transcription in As4.1 cells suggesting the H2O2 portion of the TNFα response may be NFκB-independent.
To gain additional evidence of specificity, we examined the global gene expression response to H2O2 by microarray analysis of As4.1 cells. Of 22,000 genes interrogated by the microarray, about 11,500 were expressed in As4.1 cells. There was no change in expression of any of the housekeeping genes, and few other genes displayed a change in expression (43 and 86 genes were decreased or increased, respectively, 2-fold or more in response to 200 μM H2O2,). Like TNFα, IL-1β stimulated NFκB transcriptional activity in both MCF7 and As4.1 cells (Figure 5). We next queried genes whose expression was induced strongly by IL-1β, but not by H2O2. The expression of 34 genes satisfied these criteria; and 25 of them are known or proposed NFκB target genes (Tables S1 and S2). This supports the conclusion that unlike IL-1β and TNFα, H2O2 does not induce NFκB activity in As4.1 cells.
To functionally validate the independence of H2O2 from NFκB, we treated As4.1 cells with the NFκB inhibitor panepoxydone. Whereas panepoxydone reversed partially the effect of TNFα on renin mRNA, it had no effect on H2O2-mediated down-regulation of renin mRNA (Figure 6). A second independent NFκB inhibitor (6-amino-4-(4-phenoxyphenylethylamino)quinazoline) also failed to reverse H2O2-mediated downregulation of renin mRNA (data not shown). TNFα has been reported to block renin promoter activity by interfering with the interaction of CREB with the renin enhancer CRE, and indeed TNFα blunts CRE-dependent transcriptional activity of a construct containing 3-copies of a CRE (Figure 7A). Similarly, H2O2 blunted both basal and forskolin-induced CRE-dependent transcriptional activity (Figure 7B).
The major findings of our study are: 1) TNFα induces the production of ROS, 2) H2O2 can negatively regulate endogenous renin mRNA and renin promoter activity through a mechanism requiring transcription factor binding sites in the renin enhancer and promoter, and 3) the negative influence of H2O2 on renin expression occurs independent of NFκB activation, but may act by modulating the activity of CREB. Our data further suggest that TNFα inhibits renin expression through both NFκB-dependent and NFκB-independent mechanisms, with the latter involving oxidative stress.
It is known that cytokines are potent modulators of renin expression (reviewed in 18). Systemic inflammation caused by lipopolysaccharide results in decreased renal renin expression with concomitant increases in several cytokines; and adenoviral over-expression of oncostatin M suppresses renal renin mRNA.4 Several cytokines blunt the activity of the renin promoter in As4.1 cells via a mechanism involving the renin enhancer.19 The renin gene enhancer consists of a close clustering of evolutionarily conserved transcription factor binding sites which strongly stimulates transcriptional activity of the renin promoter in vitro, and is required for the activity of the renin promoter in vivo.20-23 Like other cytokines, TNFα inhibits endogenous renin expression and renin promoter activity in As4.1 cells and blunts cAMP-mediated induction of renin expression in isolated native juxtaglomerular cells.8 The physiological importance of TNFα as a regulator of renin expression is evidenced by the observation that renal renin expression is significantly increased in TNFα-deficient mice.8 Todorov et al has reported that the mechanism of TNFα-mediated inhibition of renin expression involves an NFκB-dependent decrease in the binding of CREB to the renin enhancer, and a decrease in NFκB p65 transcriptional activity at the CRE, which the authors proposed is a non-canonical NFκB binding site.3, 7 Our data suggests there is also an NFκB-independent component to the TNFα response which acts through production of ROS and oxidative stress. This is consistent with our data showing a partial reversal of TNFα-mediated inhibition of renin expression after either panepoxydone (NFκB-dependent) or NAC (NFκB-independent), and the production of ROS by TNFα. Further support for an NFκB-independent effect comes from our data showing that H2O2 decreases renin expression (and renin promoter activity), but does not activate NFκB activity in As4.1 cells. This is not unique to As4.1 cells, since H2O2 does not activate NFκB, or the expression of the NFκB target gene ICAM-1 in endothelial cells.24
Our data also suggest that H2O2 may interfere with CREB/CRE mediated transcription in As4.1 cells. CREB activity has been reported to be increased by ROS, and there is evidence that the DNA binding activity of CREB can be modulated by the redox state of cysteine residues in its DNA binding domain.25 H2O2 activates CREB activity in some cells type but not in others. In HEK293 cells, H2O2 increases phosphorylation of CREB at sites other than the canonical Ser-133, causing a loss of transcriptional activity and decreased binding of CREB with CREB-binding protein (CBP).26 Similarly, in cultured neurons, oxidant stress-induced increases in lipid peroxidation caused an increase in phosphorylated CREB, but a concomitant decrease in CREB-dependent activity of the BDNF promoter.27 H2O2-induced Ser-133 phosphorylation of CREB in T cells was correlated with decreased transcriptional activity in response to activation of the T cell receptor.28 Therefore, oxidative stress-mediated decreases in CREB transcriptional activity are not unique to renin-expressing As4.1 cells.
A functional CRE sequence is located 5' of the closely linked E-box and HRE in the renin gene enhancer, the mutation of which abolishes enhancer activity (Figure 8).20 Forskolin increases the association of acetylated histone H4 with chromatin at the renin enhancer CRE.29 Recent studies show that coactivators of CREB are required for the maintenance of renin cell identity and renin expression.30 Whereas the CRE is required for enhancer activity, mutation of the CRE markedly attenuates the negative effect of cytokines on renin enhancer/promoter function suggesting that this essential transcription factor-binding site is also important for cytokine-mediated inhibition.19 Vitamin D3 has also been shown to decrease renin expression, and targeted expression of vitamin D receptor (VDR) in juxtaglomerular cells in vivo decreases renin mRNA.31, 32 It was reported that ligand-occupied VDR interacts with CREB thus blocking its ability to bind to the renin enhancer CRE.33 Our data suggests that ROS may interfere with CRE-mediated transcription in As4.1 cells. Consequently, all these data suggest that CREB may be a convergence point for physiological signals that regulate renin synthesis (Figure 8).
Increased Ang II and arterial pressure can each decrease renal renin expression and renin release in animal models. Although seemingly paradoxical, decreased plasma renin activity is not uncommon in essential (non-renovascular) hypertension, presumably because mechanisms remain intact that cause feedback inhibition of renin synthesis and release.34 Inflammatory cytokines and Ang II can promote ROS formation; and inflammation is a likely contributor to end organ damage in hypertension. Secretion and expression of TNFα in peripheral blood monocytes in response to inflammation is higher in hypertensive than in normotensive subjects.35 Supporting the hypothesis that ROS may be one of the mediators of feedback inhibition of renin expression are data from the SHR model of hypertension. In that model, a tempol-induced decrease in oxidative stress reduced blood pressure but increased plasma renin activity.10 Although it was not directly tested, we propose this may occur in response to relief of negative feedback inhibition on renin. Our study provides a potential molecular link between oxidative stress, caused by cytokines and vasopressor substances such as Ang II in hypertension, with the regulation of renin gene expression. However, proving this will require additional in vivo studies where the effects of oxidative stress and confounding variables such as arterial pressure can be carefully controlled.
We thank Dr. John Engelhardt at the University of Iowa for the gift of MCF-7 cells and the NFκB-Luc responder adenovirus.
Funding: Dr. Sigmund: NIH grants HL48058, HL61446, and HL84207, and the Roy J. Carver Trust Dr. Goswami: NIH grant CA111365.
Disclosures: Hana Itani - none Xuebo Liu - none Ehab H. Sarsour - none Prabhat C. Goswami - NIH Grants Ella Born - none Henry L. Keen - none Curt D. Sigmund - NIH Grants, Roy J. Carver Trust