Epithelial-to-mesenchymal transformation (EMT) of tubular cells into a myofibroblastic phenotype is an important mediator of renal scarring in chronic nephropathy. This study examines the role of the renin-angiotensin system (RAS) in this process. NRK-52E cells were exposed to angiotensin (ANG) II and ANG 1–7 in the presence or absence of inhibitors and agonists of RAS signaling. EMT was assessed at 3 days by expression of α-smooth muscle actin (α-SMA) and E-cadherin and the induction of a myofibroblastic phenotype. Expression of fibrogenic growth factors and matrix proteins was assessed by RT-PCR and immunofluorescence microscopy. To confirm findings in vivo, rats were also infused with ANG 1–7 (24 μg·kg−1·h−1) or saline via an osmotic minipump for 10 days, and renal fibrogenesis was then assessed. Treatment of NRK-52E cells with ANG II induced characteristic changes of EMT. Selective blockade of the AT1 receptor or the AT2 receptor failed to inhibit ANG II-induced EMT. However, blockade of the ANG 1–7 receptor, Mas-1, was able to prevent ANG II-dependent EMT. To confirm these findings, both ANG 1–7 and the selective Mas receptor agonist, AVE-0991, were able to induce NRK-52E cells in a dose-dependent manner. Exposing cells to recombinant ACE2 was also able to induce EMT. In addition, an infusion of ANG 1–7 induced the tubular expression of α-SMA and the expression of matrix proteins in the kidney. ANG II is a potent stimulus for EMT, but not through conventional pathways. This study points to the possible limitations of conventional RAS blockade, which not only fails to antagonize this pathway, but also may enhance it via augmenting the synthesis of ANG 1–7.
- renin-angiotensin system
- epithelial-to-mesenchymal transition
- proximal tubule
most chronic nephropathies share pathogenic mechanisms that contribute to disease progression, independent of the original cause or disease (5). In particular, progressive interstitial fibrosis is thought to represent a “final common pathway” that ultimately leads to irreversible loss of renal function (17, 20, 22). One important contributor to the interstitial burden of prosclerotic cells is the transition of tubuloepithelial cells into myofibroblasts [tubular epithelial-to-mesenchymal transition (EMT)] (3, 20). Another “common” mediator of renal fibrosis is activation of the intrarenal renin-angiotensin system (RAS). Although an important compensatory mechanism to maintain renal function in response to nephron loss, ultimately, activation of the intrarenal RAS is thought to contribute to progressive renal damage through systemic hypertension, increased intraglomerular capillary pressure, renal cell proliferation and hypertrophy, oxidative stress, and increased synthesis and accumulation of glomerular and interstitial matrix proteins, such as collagen and fibronectin, via upregulation of transforming growth factor (TGF)-β1 and other fibrogenic cytokines (9, 24, 34). Historically, the initiator of these changes has been thought to be increased levels of angiotensin II (ANG II), based on findings that RAS blockade is partly effective in chronic nephropathy (1, 13, 29, 38, 40). Moreover, ANG II increases the production of extracellular matrix proteins in tubular (35) and mesangial cells (16) in vitro and results in glomerular and tubulointerstitial injury when delivered as a continuous infusion (15, 16, 32, 36). However, the intrarenal RAS is more complicated than simply its regulation of ANG II. A number of other products of the RAS also have important proinflammatory and profibrotic actions that may partly mediate some of the reported effects of exposure to ANG II. This study examines, in detail, the role of the RAS in the in vitro induction of tubular EMT.
MATERIALS AND METHODS
A well-characterized, normal rat kidney cell line (NRK-52E) was obtained from the American Tissue Culture Collection (Rockville, MD). NRK-52E cells are believed to be epithelial cells of a proximal tubular origin on the basis of patterns of collagen secretion, C-type natriuretic peptide secretion, and the presence of epidermal growth factor receptors (10). Cells were maintained in DMEM containing 4.5 g/l glucose (Invitrogen, Carlsbad, CA) with 10% fetal calf serum at 37°C in a 5% CO2 atmosphere and passaged twice a week.
Cells were cultured in the presence of ANG II (1 nM-1 μM; Auspep, Parkville, Victoria, Australia), ANG 1–7 (1 nM-10 μM; Auspep), angiotensin-converting enzyme 2 (ACE2; 50 ng/ml, gift from Dr. M. Yarski), or the Mas receptor agonist AVE-0991 (10 μM; Aventis Pharma). Untreated cells were used as respective negative controls, as previously described (4).
NRK-52E cells were also pretreated 30 min before the addition of various agonists with selective inhibitors of RAS and TGF-β1 signaling. These included the type 1 and type 2 angiotensin (AT1 and AT2) receptor antagonists, Valsartan (1 μM; Novartis, Basal, Switzerland), PD123319 (1 μM; Sigma, St. Louis, MO), or both (1 μM each), the Mas receptor antagonist A779 (1 μM; Bachem Americas, Torrance, CA), and the TGF-β type I receptor (TβR-1) inhibitor SB431542 (2 μM; Sigma). The response to ANG 1–7 was also examined in cells transfected with 1 nM Silencer siRNA to connective tissue growth factor (CTGF-55; Ambion Biosystems) or negative control Silencer siRNA (Silencer) performed in accordance with the Lipofectamine 2000 siRNA transfection protocol. An 80% reduction in CTGF mRNA was achieved in NRK-52E cells transfected with CTGF-55 siRNA.
In vivo studies.
To examine the in vivo effects of ANG 1–7 on renal fibrogenesis, male Sprague-Dawley rats (200–250 g body wt) were implanted with an osmotic minipump (model no. 2002, Alzet, Cupertino, CA) through which was infused ANG 1–7 (24 μg·kg−1·h−1 sc) or 0.9% saline. Rats were housed in a 12:12-h light-dark cycle, with ad libitum food containing 0.4–0.6% NaCl (Norco) and water. After 10 days, rats were killed and kidneys were removed. Renal expression of fibrosis markers (collagen I, collagen IV, fibronectin, and FSP-1) was measured by real-time RT-PCR. Tubular expression of α-smooth muscle actin (SMA) was determined by immunostaining and quantified (as detailed below). All experiments were performed in accordance with the National Health and Medical Research Council of Australia guidelines for animal experimentation and were approved by the Animal Ethics Committee, Austin Health.
Gene expression of markers and mediators of tubular EMT and matrix production including β-catenin, CTGF, collagen I, collagen III, collagen IV, E-cadherin, fibronectin, FSP-1, α-SMA, Snail1, Snail2, TGF-β1, twist, and vimentin was analyzed by real-time RT-PCR, performed as previously described (20). Primers and TaqMan probes for the genes of interest (Supplementary Table 1; the online version of this article contains supplemental data) and the endogenous reference 18S rRNA were constructed with the help of Primer Express (ABI Prism 7700, Perkin-Elmer, Foster City, CA). Results were expressed relative to control (untreated) cells, which were arbitrarily assigned a value of 1.
NRK-52E cells were grown on coverslips and washed twice with PBS before being fixed in ice-cold acetone for 20 min at −20°C. The cells were rehydrated in PBS with two 10-min washes, the membranes were permeabilized with 0.1% Triton X-100 for 10 min, and blocked in 0.5% BSA/PBS or 10% normal rabbit serum for 30 min before incubating with primary antibody, α-SMA (1:100; Clone 1A4, DAKO, Cupertino, CA), or E-cadherin (1:100; Transduction Laboratories, Lexington, KY) for 1 h at room temperature. PBS in the absence of any primary antibody was used as a negative control. After three washes in PBS, the cells were incubated with the fluorescent secondary antibody ALEXA-488 (rabbit anti-mouse Ab, 1:200; Molecular Probes, Invitrogen) for 30 min at room temperature. For α-SMA staining, cells were counterstained with propidium iodide (red) to demonstrate the nuclei. Images were captured on a Zeiss 510 Meta laser-scanning confocal microscope (LSM; Zeiss, Oberkochen, Germany) using LSM 510 software (version 3.2 SP2; Zeiss). Relative expression was quantitated using Image J 1.40G software (National Institutes of Health, Bethesda, MD).
Western blot analysis.
Whole cell lysates that contained 10 to 50 g of protein were subjected to 10 to 12% SDS-PAGE and transferred onto polyvinylidene difluoride membranes by semidry transfer (Semi Dry Transfer Cell; Bio-Rad, Hercules, CA). After transfer, all incubations were conducted on a rocking platform at room temperature. The membrane was blocked in 5% skim milk/TBST overnight and then incubated for 1 h with α-SMA (1:2,000; Dako) or E-cadherin (1:2,500; Transduction Laboratories). The membrane was washed with TBST and then incubated with a peroxidase-conjugated goat anti-mouse secondary antibody (EnVision; Dako) for 1 h. Immunoreactivity was detected using an enhanced chemiluminescence kit (Amersham Pharmacia Biotech, Buckinghamshire, UK) and exposure to a Gel Doc XR System (Bio-Rad). Bands were quantitated using Quantity One Software (Bio-Rad).
Immunohistochemical staining for α-SMA (1:200; Dako) was performed according to standard procedures using an avidin-biotin-based system. Briefly, 4-μM sections from formalin-fixed kidneys were brought to distilled water and then transferred to PBS. Endogenous peroxidases were blocked in 0.3% hydrogen peroxide in PBS and washed in PBS before being blocked in 10% normal rabbit serum to prevent nonspecific binding. Sections were then incubated in the primary antibody overnight at 4°C, washed in PBS, and incubated in a rat anti-mouse biotinylated secondary antibody (Dako) for 30 min. After PBS washes, the addition of streptavidin-conjugated horseradish peroxidase (Vector Laboratories), and further washing in PBS, the signal was developed using diaminobenzidine (DAB; Sigma). Finally, the samples were counterstained in hematoxylin and eosin. All sections were analyzed for staining using light microscopy (Olympus BX-50; Olympus Optical, Tokyo, Japan) and digitized using a high-resolution camera (Fujix HC-2000; Fujifilm, Tokyo, Japan). Digitized images were then captured and evaluated using an image analysis system (Imaging Research, St. Catherines, ON, Canada) coupled to an IBM NT computer. Semiquantitative assessment of tubular α-SMA was performed by determination of the density of brown (DAB) staining in cortical proximal tubular cross-sections (viewed on a ×20 objective). A total of 100 tubular cross sections per rat kidney (n = 5 rats/group) were analyzed.
Values are means ± SE unless otherwise specified. Statview (Brainpower, Calabasas, CA) was used to analyze data by unpaired Student's t-test or by ANOVA and compared using Fisher's paired least significant difference post hoc test. P values <0.05 were considered significant.
ANG II-induced EMT in NRK-52E cells.
Treatment of NRK-52E cells with ANG II for 3 days resulted in a transition in cellular morphology from an epithelial to a mesenchymal phenotype, including proliferation, elongation, front-to-back polarity, separation from neighboring cells, and the loss of the typical “cobblestone” morphology of epithelial cells in culture (Fig. 1). This was associated with altered expression of established biomarkers for fibrotic (type 2) EMT (Table 1) (17). In particular, the expression of α-SMA (Figs. 1, 2A, and see Fig. 4) was increased at a gene and protein level. The protein expression of E-cadherin was also reduced following treatment with ANG II (Fig. 1C), especially at the membrane junctions (Fig. 1).
Signaling of ANG II-induced EMT.
To further explore the pathways involved in the induction of EMT by ANG II (1 nM), NRK-52E cells were pretreated with various antagonists of RAS signaling. Before commencing experiments, the expression of AT1, AT2, and Mas-1 receptors was confirmed in the cell line (data not shown). Pretreatment of ANG II-treated cells with the AT1 antagonist valsartan (1 μM) or the AT2 antagonist PD123319 (1 μM) failed to attenuate the phenotypic changes of EMT or fibrogenesis induced by ANG II treatment (Figs. 2 and 3). Indeed, valsartan modestly increased the expression of collagen IV and fibronectin in response to ANG II (Fig. 3). By contrast, pretreatment with the Mas antagonist A779 (1 μM) prevented EMT-associated induction of α-SMA expression by ANG II (Fig. 4). Blockade of the Mas receptor also prevented the ANG II-induced increase in gene expression of the mesenchymal markers, α-SMA and vimentin (Fig. 2), as well as collagen IV and fibronectin mRNA (Fig. 3).
To further explore the Mas-dependent actions of ANG II in our model of EMT, NRK-52E cells were treated with ANG 1–7 (1 nM-10 μM) or the selective Mas agonist AVE-0991 (10 μM). In both cases, this resulted in transition to a mesenchymal phenotype, including morphological and phenotypic changes of EMT, similar to those seen in cells treated with ANG II (Fig. 1). In addition, the phenotypic changes of EMT and fibrogenesis induced by ANG 1–7 were also blocked, in each case, by the Mas receptor antagonist A779 (Figs. 2, 3, and 4).
In renal tubular cells, the main source of ANG 1–7 is the carboxypeptidase ACE2. To further explore the role of ANG 1–7 in ANG II-dependent EMT, NRK-52E cells were treated with recombinant ACE2 (50 ng/ml). As with ANG 1–7, ACE2 also induced changes consistent with EMT (Fig. 1).
Downstream signaling of EMT.
ANG II (1 nM) and ANG 1–7 (10 μM) treatment of NRK-52E cells was also associated with increased gene expression of TGF-β1, a known mediator of EMT in tubular cells (4). In both cases, this increase was prevented by the Mas receptor antagonist A779 (Fig. 5A). The phenotypic changes of EMT and fibrogenesis induced by ANG II and ANG 1–7 were also partly prevented by pretreatment with the TβR-1 inhibitor SB431542 (2 μM; Figs. 2A, 3, and 4). Notably, the expression of vimentin was not changed by blockade of the TGF receptor in our cells, possibly due to TGF receptor-independent (Mas-dependent) activation of vimentin expression by angiotensin peptides.
We previously showed that CTGF is a key downstream mediator of TGF-dependent EMT (4). In the present experiments, ANG 1–7 induced the gene expression of CTGF in a TGF-β-dependent manner (Fig. 5B), as these increases were attenuated in the presence of the TβR-1 inhibitor SB431542. In addition, transfection of cells with siRNA to CTGF (CTGF-55) that resulted in an 80% reduction in CTGF gene expression also prevented the EMT-associated morphological and phenotypic changes seen following ANG 1–7 treatment, including normalizing the expression of E-cadherin (Fig. 6B).
Paradoxical effects of RAS blockade.
Treatment with angiotensin receptor blockers increases levels of ANG 1–7, via induction of renin through negative feedback and reduced receptor-dependent endocytosis (31). As we showed that ANG 1–7 induces EMT in our model, we further explored the effects of the angiotensin receptor blockers, in our NRK-52E cells, in the absence of supplemental ANG II. In these studies, the combination of AT1 and AT2 receptor blockade actually induced morphological and phenotypic changes of EMT (Fig. 7).
In vivo correlates.
To explore profibrotic effects of ANG 1–7 in vivo, rats were infused with ANG 1–7 for 10 days. Following this exposure, we were able to demonstrate an upregulation of genes associated with matrix accumulation, including collagen I, collagen IV, and fibronectin (Fig. 8). In addition, expression of the marker of EMT, FSP-1, was also significantly increased. Paralleling these changes in gene expression, the tubular expression of the α-SMA was also increased, as detected and quantified by immunohistochemistry (Fig. 8).
Activation of the intrarenal RAS is widely thought to be a key mediator of renal fibrogenesis and progressive kidney disease. While the determinant of these changes has historically been thought to be ANG II (16, 35), these studies are potentially confounded by its rapid metabolism into ANG 1–7, which in our studies appears as the central player in RAS-mediated EMT. Indeed, in our model, the effects of 1 nM ANG II on tubular EMT were exclusively ANG 1–7/mas-1 dependent. In addition, ANG 1–7 was able to induce responses, both in vitro and in vivo, that have been classically associated with ANG II signaling in tubular epithelial cells. These findings are consistent with earlier studies in which overexpression of the c-mas oncogene increased the proliferative response to ANG II in tubular cells, independent to activation of the angiotensin receptor (Am J Physiol Renal Fluid Electrolyte Physiol 263: F931–F938, 1992). Together, these findings point to the need to reassess the RAS as a purely linear pathway, conventionally portrayed to demonstrate the efficacy of RAS blockade. In fact, recent data from the RAAS and DIRECT study (7, 19) suggest that the renoprotection afforded by these agents is partial at best and has the potential for paradoxical actions.
ANG 1–7 is a heptapeptide produced from ANG II by carboxypeptidase activity (predominantly ACE2 in the kidney and prolylendopeptidase in the plasma) and from ANG I by neutral endopeptidase and other brush-border endoproteases. Rather than being an inactive peptide, as initially suggested (12), ANG 1–7 is a potent vasodilator and natriuretic (2, 14, 26) that acts predominantly via binding to the G protein-coupled Mas receptor (27). ANG 1–7 has also been shown to stimulate mesangial cell growth and production of TGF-β (28). We now show that ANG 1–7 is also able to transform NRK-52E (proximal tubular epithelial lineage) cells into profibrotic cells with a mesenchymal phenotype. This is consistent with the role of the Mas proto-oncogene that was first detected by its ability to transform NIH-3T3 cells (37).
In our model system, exposure to ANG II was able to induce EMT via upregulation of TGF-β1 expression, through activation of the TGF-β1 receptor. This is entirely consistent with the previous observation that ANG II upregulates TGF-β1 and fibrogenesis in vitro (8, 23, 33) and in vivo (1, 25). However, we propose that these findings may be partly determined by increased levels of ANG 1–7 and Mas activation in these models. Consistent with this hypothesis, Shao and colleagues (28) reported that chronic administration of ANG 1–7 increased renal expression of TGF-β1, as well as renal injury in streptozotocin diabetes. In addition, a recent study in cultured human mesangial cells reported that ANG 1–7 (100 nM) treatment increased both TGF-β1 and matrix protein synthesis (39). Furthermore, blockade of these cells with the selective Mas receptor antagonist A779 attenuated the ANG 1–7-induced increase in TGF-β1 and matrix protein expression (39).
Our findings potentially contradict those in primary proximal tubule cells in which ANG 1–7 did not affect TGF-β1 expression or fibrogenesis (30). Indeed, these studies propose that ANG 1–7 may play a renoprotective role, as preincubation of tubular cells with ANG 1–7 resulted in a partial reduction in ANG II-induced TGF-β1 protein levels (30). The reasons for these differences remain unclear. However, it is likely that the balance of the RAS (including cellular activation status, the expression of receptors, enzymatic components, and the doses of reagents used) may contribute to different findings in different settings.
The biological activities of ANG II have been traditionally thought to be mediated by binding to the high-affinity AT1 or the AT2. Certainly, AT1 receptor agonists are able to partially attenuate renal interstitial fibrosis in experimental and human nephropathy (18, 21). However, these renoprotective actions may be separate from effects on EMT. In the only previous study to explore this issue, (high-dose) ANG II-induced vimentin expression in HK-2 cells was partly attenuated by AT1 receptor blocker (6). By contrast, our studies using 1 nM ANG II and analysis after 3 days show blockade of one or both receptors did not prevent ANG-dependent EMT. Moreover, combined blockade of AT receptors in healthy cells was able to induce some components of EMT. We hypothesize that AT receptor blockade increases the formation of ANG 1–7 (through feedback induction of renin and reduced endocytosis) (31) and subsequent activation of the Mas receptor. The potential profibrotic actions of ANG 1–7 demonstrated in our study would suggest that this escape may reduce the efficacy of these agents, or antagonize the benefits achieved in the clinical setting. Recent primary prevention studies in which AT1 receptor blockade increased albuminuria in diabetic individuals may further support this assertion (7, 19).
Although we looked at ANG 1–7 as the downstream mediator of exposure to increased levels of ANG II, other products of ANG II metabolism may also have important biological effects in vitro and in vivo. Recent reports suggest that both ANG III and ANG IV have a range of biological actions. However, neither binds the Mas receptor nor can they be produced from ANG 1–7 (11), suggesting that they are not mediators of EMT in our model system. Nonetheless, their roles in other ANG II-dependent phenomena remain to be determined.
In summary, these findings demonstrate that exposure to ANG II induces EMT via its degradation to ANG 1–7 and activation of the ANG 1–7 receptor, Mas-1. These findings have important implications for the design of renoprotective strategies that block activation of the RAS and the role of ANG 1–7 in progressive renal disease.
No conflicts of interest, financial or otherwise, are declared by the author(s).
- Copyright © 2010 the American Physiological Society