Calcineurin Blockade Prevents Cardiac Mitogen-activated Protein Kinase Activation and Hypertrophy in Renovascular Hypertension*

Chronic stimulation of the renin-angiotensin system induces an elevation of blood pressure and the development of cardiac hypertrophy via the actions of its effector, angiotensin II. In cardiomyocytes, mitogen-activated protein kinases as well as protein kinase C isoforms have been shown to be important in the transduction of trophic signals. The Ca2+/calmodulin-dependent phosphatase calcineurin has also been suggested to play a role in cardiac growth. In the present report, we investigate possible cross-talks between calcineurin, protein kinase C, and mitogen-activated protein kinase pathways in controlling angiotensin II-induced hypertrophy. Angiotensin II-stimulated cardiomyocytes and mice with angiotensin II-dependent renovascular hypertension were treated with the calcineurin inhibitor cyclosporin A. Calcineurin, protein kinase C, and mitogen-activated protein kinase activations were determined. We show that cyclosporin A blocks angiotensin II-induced mitogen-activated protein kinase activation in cultured primary cardiomyocytes and in the heart of hypertensive mice. Cyclosporin A also inhibits specific protein kinase C isoforms. In vivo, cyclosporin A prevents the development of cardiac hypertrophy, and this effect appears to be independent of hemodynamic changes. These data suggest cross-talks between the calcineurin pathway, the protein kinase C, and the mitogen-activated protein kinase signaling cascades in transducing angiotensin II-mediated stimuli in cardiomyocytes and could provide the basis for an integrated model of cardiac hypertrophy.

genesis of hypertension, since drugs that inhibit Ang II production or its binding to specific receptors, reduce blood pressure in hypertensive patients. Normalization of blood pressure is followed by a regression of left ventricular hypertrophy that is usually associated with chronic hypertension (2). In addition to its hemodynamic effects, Ang II also contributes directly to the hypertrophic response via its growth factor properties on cardiac myocytes and fibroblasts (3).
Different intracellular signaling pathways have been implicated in transducing Ang II-induced hypertrophic stimuli. Signaling through the Ang II G q protein-coupled AT 1 receptors leads to the activation of phospholipase C, which results in the production of diacylglycerol and inositol triphosphate. In turn, diacylglycerol activates protein kinase C (PKC), and inositol triphosphate induces a raise in intracellular calcium (Ca 2ϩ ) (4). Ang II receptors also stimulate mitogen-activated protein kinases (MAPK) (5). At least three families of MAPK have been described, namely the extracellular-signal regulated kinases (ERK or p42/44), the c-Jun N-terminal kinases (JNK) also known as the stress-activated kinases, and the p38 MAPK (6). Activation of MAPK is the result of phosphorylation cascades involving the small GTP-binding proteins Ras and Rac. Moreover, PKC has also been suggested to participate in the activation of MAPK (7)(8)(9). Recent studies identified the pathway involving the calmodulin-dependent phosphatase calcineurin (CaN) as important in the development of cardiac hypertrophy (10). CaN dephosphorylates a family of transcription factors known as NF-ATs (nuclear factor of activated T cells) (11). In particular, NF-AT3, which is expressed in cardiomyocytes (12), binds upon dephosphorylation to the transcription factor GATA4 to up-regulate cardiac-specific gene expression (13). The immunosuppressive drug cyclosporin A (CsA) inhibits CaN activity through binding to the inhibitory protein cyclophilin (14). Therefore, CsA was shown to reduce cardiac hypertrophy induced in transgenic mice by a constitutive form of CaN (10). Recent studies using more physiological models have also reported a role for CaN in the development of cardiac hypertrophy (15,16). Interestingly, CsA inhibits Ang II-stimulated gene expression in rat primary cardiomyocytes in vitro (10). Similarly, CaN is suggested to be involved in the cardiac expression of the immediate response genes in the hypertrophied heart (16). However, the exact role of CaN in mediating Ang II-dependent cardiac growth in vivo remains to be established (17)(18)(19)(20)(21).
The two kidney-one clip (2K1C) model of renovascular hypertension greatly contributes to our knowledge of cardiovascular diseases. In this model, one renal artery is constricted to reduce renal perfusion, and the other kidney remains untouched. In response to low renal arterial pressure, plasma renin concentrations and then AngII levels are rapidly increased, resulting in chronic elevation of blood pressure and subsequent cardiac hypertrophy.
In the present study, we used mouse primary cardiomyocytes and a murine model of renovascular hypertension (22) to investigate whether CaN could play a role in cardiac hypertrophy. Results demonstrate that CsA inhibits Ang II-induced MAPK activation in primary cardiomyocytes as well as in cardiac tissue possibly via a PKC-dependent mechanism. In addition, CaN blockade also prevents the development of cardiac hypertrophy during renovascular hypertension. Taken together, these results suggest cross-talks between the CaN and the MAPK pathways for Ang II-stimulated cardiac hypertrophy

MATERIALS AND METHODS
Mice and CsA Treatment-Eight-week-old, male C57BL/6 mice (Iffa Credo, L'Arbresle, France) were injected intraperitoneally with either saline or 25 mg of CsA (Sandimmun®, Novartis, Switzerland)/kg of body weight twice daily starting on the day of clipping. Mice were maintained on sterile tap water and regular rodent diet ad libitum.
Cell Culture-Neonatal C57BL/6 mouse ventricles were separated from atria. Cardiomyocytes and nonmyocyte cells were purified as described (23) and plated at a density of 0.1 ϫ 10 6 cells/cm 2 on gelatincoated wells and noncoated wells, respectively. After 24 h in culture, cells were switched to no serum medium, and Ang II was added 24 h later. Nonmyocyte cells were passaged once before the addition of no serum medium. Cells were pretreated with either 250 ng/ml CsA or 150 ng/ml FK506 for 20 min prior to the addition of 100 nM Ang II. Stimulation time was 5 min for p38 and ERK and 10 min for JNK. Cells were then washed with phosphate-buffered saline and lysed in radioimmune precipitation buffer (150 mM NaCl, 0.25% deoxycholic acid, 1% Nonidet P-40, 1 mM NaVO 3 , 1 mM NaF, 1 g/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml pepstatin A, 1 g/ml leupeptin, 1 mM EDTA, 50 mM Tris, pH 7.5). The lysates were either heated at 95°C for 5 min for SDS-PAGE or kept frozen until used. In some experiments, cells were cultured in absence of calcium. In this case, cardiomyocytes were switched to calcium-free medium and then treated for 2 h. with 100 M thapsigargin.
Surgery-Briefly, mice were anesthetized using 1.5% halothane in oxygen. The left kidney was exposed, and a clip (0.12-mm opening) was inserted on the renal artery isolated by blunt dissection as described (22). Sham procedure included the entire surgery with the exception of artery clipping. For blood pressure and heart rate measurement, the right carotid artery was exposed through cervical incision. A silicon catheter filled with 5% glucose solution containing heparin (300 IU/ml) was inserted into the vessel and tunneled subcutaneously to exit at the back of the neck. Blood pressure and heart rate were recorded in conscious mice by connecting the catheter to a pressure transducer using a computerized data acquisition system (Notocord, Paris, France).
Tissue Sampling-Heart and kidneys were frozen in liquid nitrogen and stored at Ϫ70°C until used. Heart wet weight (without the atria) was normalized to the tibial length and to body weight.
Cardiac Protein Extract-Hearts were rapidly excised and rinsed in cold phosphate-buffered saline. Atria were removed, and ventricles were homogenized on ice in 1 ml of radioimmune precipitation buffer. Homogenates were sonicated on ice for three bursts of 5 s each and centrifuged for 15 min at 4200 rpm at 4°C. Lysates were kept frozen until used or SDS-PAGE loading buffer (125 mM Tris, 4% SDS, 20% glycerol, 100 mM dithiothreitol, 0.2% bromphenol blue, pH 6.8) was added to reach a final concentration of 25%. Lysates were then heated at 95°C for 5 min and run in 10% SDS-PAGE.
MAPK Activation-Protein concentrations were determined according to Bradford. Thirty g of soluble proteins were run in 10% SDS-PAGE and transferred onto nitrocellulose membranes (ECL nitrocellulose; Amersham Pharmacia Biotech). Phosphorylated MAPK were detected by Western blot analysis using specific antibodies for the phosphorylated forms of p38, ERK, and JNK (New England Biolabs). After overnight incubation at 4°C in blotto (20 mM Tris, pH 7.6, containing 0.1% Tween 20 and 5% dry milk), first antibodies were revealed using anti-rabbit secondary antibodies conjugated to horseradish peroxidase and a chemiluminescent detection system (Amersham Pharmacia Biotech). MAPK phosphorylation was quantified by laser-scanning densitometry of specific phosphorylated bands (NIH Image).
PKC Activation-Cells were lysed in 50 l in 12.5 mM Tris, 2.5 mM EGTA, 1 mM EDTA, 100 mM NaF, 5 mM dithiothreitol, 300 M phenylmethylsulfonyl fluoride, 120 M pepstatin A, 200 M leupeptin, 200 M aprotinin, pH 7.4. Extracts were incubated for 5 min at 4°C and centrifuged at 12,000 rpm for 15 min at 4°C. Soluble fractions were heated at 95°C for 5 min in 2ϫ SDS loading buffer. Particulate fractions were washed once in lysis buffer, resuspended in 50 l of lysis buffer containing 1% Triton X-100, and heated at 95°C for 5 min in 2ϫ loading buffer. Samples were run in SDS-PAGE and transferred on nitrocellulose membrane as above. PKC isoforms were detected using isoform-specific antibodies (Biomol). The activity of the different PKC isoforms (␣, ␤, ⑀, ) in vivo was measured by immunoprecipitating from cardiac protein extracts each isoform using specific antibodies (Biomol). Activity was assayed at 30°C in the presence of 50 M ATP, 10 Ci of [␥-32 P]ATP, and myelin basic protein (Upstate Biotechnology, Inc., Lake Placid, NY) (24). The amounts of 32 P-labeled peptide were meas-ured by liquid scintillation counting. To block the different PKC isoforms, cardiomyocytes were treated with 2 M of isoform-specific PKC inhibitors (Biomol) for 24 h.
Northern Blot-Total RNA was purified with TriPure (Roche Molecular Biochemicals). Ten g of total RNA was separated on 1% formamide-agarose gel and transferred to GeneScreen nylon membranes (PerkinElmer Life Sciences). Kidney RNAs were hybridized with a radiolabeled mouse renin probe (kindly provided by Dr. K. Nakayama, University of Tsukuba, Japan), and heart RNAs were hybridized with a mouse brain natriuretic peptide (BNP) probe. The intensity of hybridization signals was quantified with an Instant Imager detector (Packard Instruments). The blots were then stripped and hybridized with a mouse glyceraldehyde-3-phosphate dehydrogenase probe for normalization.
Statistical Analysis-All values are expressed as means Ϯ S.E.

Effect of CsA on CaN and MAPK Activation in Ang II-stimulated Primary Cardiac Myocytes and Nonmyocyte Cells-To
investigate possible cross-talks between the Ang II-activated MAPK and CaN pathways, cardiac myocytes and nonmyocyte cells were stimulated with Ang II, and phosphorylation of p38, ERK, and JNK was measured. All three kinases were activated in cardiac myocytes in response to Ang II (Fig. 1, A-D). In contrast, in nonmyocyte cells, p38 was not activated (Fig. 2,  A-D). In addition, Ang II stimulation led to a marked increase in CaN activity in myocytes (Table I). This activity was completely inhibited in the presence of CsA. Blockade of the CaN pathway in cardiomyocytes using CsA resulted in the complete inhibition of the activation of all three MAPK by Ang II (Fig. 1,  A-D). To exclude nonspecific inhibition of CsA on MAPK activation, we studied the effects of FK506, a different CaN blocker (14). This drug was as effective as CsA in inhibiting Ang IIinduced MAPK activation in cardiomyocytes (Fig. 1, A-D). Similarly, ERK and JNK phosphorylation was inhibited by CsA treatment in nonmyocyte cells (Fig. 2, A-D).
Effect of CsA on PKC Activation in Ang II-stimulated Primary Cardiomyocytes-To investigate whether PKC could be affected by CsA treatment, we determined the activity of two calcium-dependent (PKC␣ and -␤) and two calcium-independent (PKC⑀ and -) isoforms of PKC in Ang II-stimulated primary cardiomyocytes (Fig. 3A). Ang II activated all four PKC isoforms as shown by the translocation of the protein from the soluble to the particulate cellular fraction (Fig. 3A). CsA differently affected the activity of the different isoforms. First, CsA completely inhibited the Ang II-induced translocation of calcium-independent PKC isoforms. Interestingly, CsA alone was able to stimulate the activation of calcium-dependent PKC isoforms. To further study the stimulation of PKC␣ and -␤ by CsA, the experiment was repeated in the absence of calcium (Fig. 3B). In this case, CsA was not able to activate PKC␣ and -␤. In accordance with previously published data, calcium removal also inhibited Ang II-stimulated activation of PKC␣ and -␤. As expected, the absence of calcium had no effect on calcium-independent PKC isoforms (PKC⑀ and -).
Effect of PKC Inhibition on Ang II-induced MAPK Activation in Primary Cardiomyocytes-To determine whether the PKC isoforms, which were inhibited by CsA, were implicated in Ang II-induced MAPK activation in cardiomyocytes, we used a series of isoform-specific PKC inhibitors. Blockade of the activity of calcium-independent PKC isoforms (PKC⑀ and -) completely abolished Ang II-stimulated MAPK activation (Fig. 4, A-D). In contrast, the inhibition of a calcium-dependent PKC isoform (␤) had no effect on MAPK stimulation by Ang II (Fig. 4, A-D).
Hypertension and Cardiac Hypertrophy in CsA-treated 2K1C Mice-To study the possible role of CaN in the development of cardiac hypertrophy induced by a chronic elevation of blood pressure, CsA was administered to sham-operated and 2K1C mice with Ang II-dependent hypertension. Four weeks after clipping, a significant increase in systemic blood pressure was observed in untreated 2K1C mice (Fig. 5A). In these animals, elevated pressure resulted in the development of cardiac hypertrophy (Fig. 5, B and C, and Table II). In contrast, CsA treatment completely blocked the increase in cardiac mass. To exclude the possibility that a detrimental effect of CsA on animal growth could represent a confounding factor in these experiments, the development of cardiac hypertrophy was followed by calculating both heart weight to tibial length as well as heart weight to body weight ratios. Both parameters showed consistent CsA-dependent inhibition of cardiac growth. In addition, body weights did not significantly vary in CsA-treated animals (body weight in sham untreated: 25.4 Ϯ 0.7 g; sham CsA-treated: 24.7 Ϯ 0.3 g; clipped untreated: 25.2 Ϯ 0.7 g; clipped CsA-treated: 24.3 Ϯ 0.7 g). Importantly, blood pressure was as high in 2K1C mice receiving CsA as that measured in untreated 2K1C mice (Fig. 5A). To further evaluate the effect of CsA on cardiac tissues, BNP expression was measured as an index of cardiac hypertrophy (Fig. 6A). Clipped animals showed a 3-fold increase in BNP expression in cardiac tissues as compared with controls. In contrast, CsA-treatment blocked stimulated BNP transcription in 2K1C mice. To verify if the type of hypertension that developed in CsA-treated 2K1C mice was still renin-dependent, renin expression in the clipped kidney was determined. Renin mRNA levels were significantly elevated in both untreated and CsA-treated 2K1C mice as compared with values observed in respective control animals (Fig. 6B).
Effect of CsA on CaN, PKC Isoform, and MAPK Activation in the Heart of 2K1C Mice-We first confirmed that CsA treatment resulted in CaN inhibition in vivo. Data in Table I demonstrated that CaN activity in the heart of 2K1C mice was indeed completely inhibited by CsA administration. MAPK activation was then followed during the development of cardiac hypertrophy in these mice. We first confirmed that p38, ERK, and JNK were activated in response to chronic hypertension  (Table II and Fig. 7). Significant activation of both ERK and JNK was observed 3 days after clipping. In contrast, p38 appeared significantly activated only after 4 weeks. MAPK phosphorylation correlated with the progression of cardiac hypertrophy. Interestingly, maximum cardiac enlargement occurred concomitantly with p38 activation. Since CsA treatment prevented the increase in cardiac mass induced by renovascular hypertension (Fig. 5 and Table II) and inhibited MAPK phosphorylation in vitro (Figs. 1 and 2), we studied whether CaN blockade interfered with MAPK activation in the heart of 2K1C mice. Stimulation of all three MAPK pathways was fully inhibited by CsA injection (Table II and Fig. 7). Moreover, the activation of calcium-independent PKC isoforms (PKC⑀ and -) was also abolished by CsA treatment (Table III). In contrast, calcium-dependent PKC isoforms (PKC␣ and -␤) were not affected. On the contrary, in accordance to what observed in vitro, these isoforms were significantly activated by CsA (Fig. 3 and Table III).

DISCUSSION
Chronic elevation of blood pressure induces the development of cardiac hypertrophy, which appears to depend in part on MAPK activation. In the present study, we show that CaN blockade inhibits MAPK activation and prevents the development of cardiac hypertrophy in renovascular hypertension, indicating points of convergence between these two pathways. The fact that a different calcineurin inhibitor, namely FK506, also blocks MAPK activation in cardiomyocytes argues against a nonspecific inhibitory effect of CsA (Fig. 1). Ang II-induced MAPK activation appears to depend on the stimulation of calcium-independent PKC isoforms ( Fig. 4 and Table III). Since the activity of these particular PKC isoforms was found to be blocked by CsA, it suggests that CsA-mediated inhibition of MAPK activation could result from an inhibitory effect of CsA on calcium-independent PKC isoforms.
Since the initial observation on the preventive effect of CsA on the development of cardiac hypertrophy induced in transgenic mice (10), results from studies using more physiological models have been reported. Some studies fail to demonstrate an inhibitory effect of CsA on hypertrophy (15,17,19). The effectiveness of the CsA treatment could depend on the duration of the hypertrophic stimulus. In mice with transverse aortic constriction, pressure overload-induced cardiac hypertrophy was also insensitive to CsA treatment (18,20,21). However, constriction of the aortic arch does not activate the renin-angiotensin system (25). In contrast, a recent study demonstrated a role for CaN in a model of cardiac hypertrophy following abdominal banding known to be partially Ang II-dependent (16). Accordingly, our study describes the inhibitory effect of CsA in an Ang II-dependent model (22). Therefore, it is possible that CaN inhibition effectively blocks Ang II-mediated cardiac hypertrophy, such as that developing in response to abdominal aortic banding or following reduced renal perfusion in renovascular hypertension, and is less effective on mechanical stretch-stimulated cardiac growth. It is noteworthy that CsA treatment does not affect renin expression, since renin mRNA in the kidney of CsA-treated 2K1C animals is as high as that observed in untreated 2K1C mice (Fig. 6B). This suggests that CsA prevention of Ang II-induced cardiac hypertrophy does not result from an inhibition of the activity of the reninangiotensin system. Although CsA has been reported to increase renin secretion and synthesis, our result is consistent with other studies in CsA-treated rats, in which intrarenal renin mRNA and plasma renin activity were elevated for 2 weeks but not different from that of controls by 4 weeks (reviewed in Ref. 26).
Since CsA was shown to induce anorexia (27), we were concerned about possible toxic effects that could result in nonspecific reduction of cardiac tissues. We think that this is unlikely to play a significant role in the present experiments for several reasons. First, heart weight normalized to tibial length or body weight gave concordant results (Fig. 5, B and C). Second, CsA  7. Western blot analysis of MAPK activation in the heart of CsA-treated 2K1C mice. Heart proteins were analyzed 4 weeks after clipping. The phosphorylated and total p38, ERK, and JNK proteins were detected using specific antibodies. treatment led to an insignificant body weight reduction of less than 5% 4 weeks after clipping. Finally, the mortality rate was not higher in CsA-treated groups than that in untreated groups (less than 5% in each experimental group). Beside its hemodynamic properties, Ang II was shown to exert growth factor effects on cardiac myocytes and fibroblasts (3). Indeed, Ang II has been shown to stimulate several pathways involved in mitogenic signals (4). Hence, various concomitant activation of different pathways occurs in the heart during Ang II-dependent renovascular hypertension. These include the phosphatase CaN pathway. Indeed, Ang II readily activated CaN in cultured cardiomyocytes and in the heart of 2K1C mice (Table I). In addition, CaN could stimulate cardiac growth in part through MAPK activation, since CaN blockade resulted simultaneously in the prevention of cardiac hypertrophy and in a complete inhibition of the MAPK pathways in the heart of 2K1C mice. In addition, CsA inhibits Ang II-induced activation of MAPK in cultured cardiomyocytes, which also suggests cross-talks between the CaN and the MAPK pathways.
PKC has been implicated in MAPK activation in cardiomyocytes (7)(8)(9). Interestingly, CsA completely inhibits Ang II-stimulated calcium-independent PKC isoforms (PKC⑀ and -), whereas it activates by itself calcium-dependent isoforms (PKC␣ and -␤). CsA-induced activation of PKC␣ and -␤ probably occurs through the induction of calcium release, since the stimulation of PKC␣ and -␤ activity is blocked in the absence of calcium. Indeed, calcineurin has been shown to modulate intracellular calcium fluxes (28). Because calcium-dependent PKC isoforms are activated in CsA-treated sham-operated mice without concomitant stimulation of the MAPK pathways and in absence of a significant increase in cardiac mass, it suggests that these particular PKC isoforms do not play a important role in the development of cardiac hypertrophy. Moreover, specific blockade of calcium-independent PKC isoforms completely inhibited Ang II-induced MAPK activation in vitro and in vivo. Therefore, these results strongly suggest a pathway linking CaN, calcium-independent PKC isoforms, and MAPK. CaN has been suggested to synergize with PKC to activate Rac (29 -31), and CaN has also been shown to potentiate Raf-dependent activation of MAPK stimulation (32). In addition, the activation of all three MAPK pathways (i.e. ERK, JNK, and p38) has been shown to be dependent in part on PKC stimulation (7)(8)(9). The results of the present study are consistent with recently published data showing that CaN blockade prevents JNK and PKC activation in aortic banded rats (33). Interestingly, in this Ang II-independent model of pressure overload, it appears that the PKC isoforms that are inhibited by CsA are different from those affected in our study.
Finally, one cannot rule out the possibility that CaN can also act through gene transcription to subsequently activate the different MAPK pathways during chronic hypertension in vivo. CsA treatment and then CaN inhibition could block the synthesis of an unidentified MAPK activator. However, this seems rather unlikely given the fact that CsA inhibits Ang II-induced MAPK activation in cultured cardiomyocytes, in which late response genes are not activated during the short activation time. CsA treatment also induces some reduction of cardiac weight in sham-operated mice. This reduction in heart weight is likely to result from complete inhibition of pathways that are already activated in the basal state. Therefore, this CsA effect could be consistent with tonic controls of normal cardiac mass by either activated CaN, PKC, or MAPK pathways.