Angiotensin II–Induced Oxidative Stress Resets the Ca²⁺ Dependence of Ca²⁺–Calmodulin Protein Kinase II and Promotes a Death Pathway Conserved Across Different Species

Experimental evidence indicates that a critical factor in the transition from compensated to noncompensated cardiac hypertrophy is myocyte cell loss by apoptosis.¹ The circulating levels of Ang II are increased in heart failure and may constitute one of the major causes of cell death in this transition.² The apoptotic effects of Ang II have been reported to be mediated by different signaling molecules, eg, Ca²⁺, protein kinase (PK)C-δ, p38 mitogen-activated protein kinase (p38MAPK), or reactive oxygen species (ROS).^2–5 Recent reports have shown that activation of the multifunctional Ca²⁺/calmodulin–dependent protein kinase (CaMK)II is a common intermediate of diverse death stimuli–induced apoptosis in cardiac cells.⁶ Supporting these results, we have demonstrated that CaMKII activation is a critical event in the signaling cascade that leads to apoptosis and necrosis occurring in reperfusion injury.⁷ More importantly, we and others have shown that Ang II–induced apoptosis is caused by an increase in CaMKII activity mediated by ROS.^8,9 Recent experimental evidence further indicated that ROS-induced oxidation of methionine residues sustains CaMKII activity in the absence of Ca²⁺/calmodulin (Ca²⁺/CaM). However, this action requires previous binding of Ca²⁺/CaM to expose the autoinhibitory domain of CaMKII for oxidation.⁸ Although these experiments suggest the concept that ROS can reset the Ca²⁺ dependence of CaMKII activation, this possibility has never been tested. Moreover, other findings have pointed to the role of phosphatase inhibition in ROS-induced CaMKII activation.¹⁰

In spite of these important achievements in the understanding of the proapoptotic molecules activated by Ang II, the signaling cascade whereby the peptide produces apoptosis remains uncertain. For instance, it is not clear whether the proapoptotic molecules mentioned above are interconnected in one single pathway initiated by Ang II or whether they constitute part of different and alternative cascades by which Ang II exerts its apoptotic action. To examine this issue, we investigated the signaling cascade by which Ang II produces cell death in cultured cardiomyocytes from rat and cat, 2 species in which Ang II produces opposite effects on cardiac contractility.^11,12 The results presented herein, using pharmacological tools and genetic manipulation, show that species with opposite inotropic response to acute Ang II administration share a common apoptotic pathway triggered by the sustained stimulation with the peptide, involving ROS, CaMKII, and p38MAPK. Although these molecules have been previously implicated in the apoptotic effect of Ang II, the present findings demonstrate that they constitute a previously undefined cascade of concatenated events that lead to apoptosis. More importantly, the results indicate that in the presence of ROS, CaMKII may be activated in nominally Ca²⁺-free conditions. These results reveal a new mechanism by which ROS radically reset the Ca²⁺ dependence of CaMKII to extremely low values.

Methods

Adult cats, Wistar rats, or transgenic mice expressing either a CaMKII inhibitory peptide (AC3-I) or a scrambled control peptide (AC3-C)⁸ were used in accordance with the Guide for the Care and Use of Laboratory Animals (NIH Publication No.8-23, revised 1996).

An expanded Methods section is available in the Online Data Supplement at http://circres.ahajournals.org.

Results

ROS- and CaMKII-Dependent Ang II—Induced Apoptosis

Figure 1A shows that 1 μmol/L Ang II significantly reduced the number of viable rat and cat myocytes to 38.9±2.6% and 43.7±3.0%, respectively, after 24 hours of culture. This was prevented in cat myocytes by the Ang II type 1 receptor blocker losartan but not by the AT2 receptor antagonist PD123.319 (41.9±2.9%, P<0.05 versus control). Figure 1B and 1C shows that Ang II increased caspase-3 activity and the number of TUNEL-positive nuclei in cat myocytes. Overall, TUNEL-positive nuclei increased from 3.9±0.2% to 10.1±0.1% (P<0.05, control versus Ang II), confirming the apoptotic effect of sustained Ang II administration.²

• Download figure
• Download PowerPoint

Figure 1D depicts typical experiments and overall results showing that Ang II increased ROS production in rat cells. This effect was prevented by 10 μmol/L of the flavoprotein inhibitor diphenyleneiodonium (DPI). Moreover, cotreatment of either rat or cat myocytes with Ang II plus DPI or the free radical scavenger MPG (2-mercaptopropionylglycine) (1 mmol/L) prevented cell death (Figure 1E).

Figure 2A depicts immunofluorescence images from cat cultured myocytes, showing that phosphorylated (P)-CaMKII was increased by Ang II. This increase was prevented by 1 μmol/L of the CaMKII inhibitor KN-93 but not by the inactive analog KN-92. Similarly, KN-93 prevented the increase in P-CaMKII and P-Thr17 of phospholamban, a CaMKII substrate, when evaluated by Western blots (Figure 2B). KN-93, when given alone, failed to significantly affect basal CaMKII activity, as assessed by P-Thr17 (82.2±16.5% of control, n=7). Figure 2C shows that CaMKII inhibition did not prevent the increase in ROS production induced by Ang II in rat myocytes, indicating that Ang II–induced ROS formation is upstream of CaMKII. Figure 2D depicts overall results showing that CaMKII inhibition, either by KN-93 or by a more specific CaMKII inhibitor, AIP (autocamtide 2-related inhibitory peptide) (2.5 μmol/L) also prevented Ang II–induced cell death in both, cat and rat myocytes. Moreover, transgenic myocytes expressing the CaMKII inhibitory peptide (AC3-I) were protected from the deleterious effect of Ang II when compared with myocytes expressing the scrambled control peptide (AC3-C). The images on the right of Figure 2D depict representative experiments showing that CaMKII inhibition prevents the increase in caspase-3 activation and TUNEL-positive cells. Overall, TUNEL-positive nuclei decreased from 10.1±0.1% to 5.73±0.5% and caspase-3 activity from 57.9±17.9% to 27.5±13.9% (Ang II versus Ang II+KN-93).

• Download figure
• Download PowerPoint

Taken together, these results indicate that ROS and CaMKII activation are necessary steps of the pathway by which Ang II induces apoptosis, both in cat and rat myocardium.

Ang II–Induced ROS-Dependent Activation of CaMKII Occurs Even in a Nominally Ca²⁺_i-Free Medium

Results show that prolonged exposure of rat myocytes to Ang II is associated with the activation of CaMKII. However, the source of Ca²⁺ for CaMKII activation is not straightforward in cultured unpaced myocytes. To assess this issue, we evaluated diastolic Ca²⁺ and Ca²⁺_i transient amplitude (Ca_iT) in electrically field stimulated (0.5 Hz) rat cells, after 24 hours of culture. Ang II did not modify neither diastolic Ca²⁺ nor peak Ca_iT in rat myocytes (diastolic Indo-1 fluorescence, 0.72±0.05 and 0.70±0.05; peak fluorescence, 1.34±0.04 and 1.35±0.05, control versus Ang II). Similarly, sustained administration of Ang II failed to increase diastolic Ca²⁺ or Ca_iT amplitude in cat myocytes, a species in which acute administration of Ang II effectively increases Ca²⁺_i (Figure 3A and 3B).¹¹ Furthermore, estimation of sarcoplasmic reticulum Ca²⁺ content by caffeine pulses showed that there was no significant differences between control and Ang II treated cells. In contrast, myocytes stimulated for 24 hours with 1 μmol/L isoproterenol (an intervention known to promote Ca²⁺-dependent apoptosis¹³) showed a significant increase in the amplitude of the Ca_iT and sarcoplasmic reticulum Ca²⁺ content (Figure 3C). Diastolic Ca²⁺ was also measured every 10 minutes during the first hour of incubation with Ang II and every hour in the following 4 hours, to exclude the existence of a possible transitory increase in Ca²⁺_i at the beginning of the incubation period, undetectable after 24 hours of Ang II stimulation but still responsible for Ang II–induced CaMKII activation. Again, no increase in Ca²⁺_i was detected in Ang II–treated cells with respect to controls in 4 experiments in which at least 20 cells were observed at each time point (Figure 4A). These experiments indicate that there was no significant increase in Ca²⁺_i in unpaced myocytes with sustained Ang II stimulation, suggesting that Ang II–induced CaMKII activation could occur in the absence of any increase in Ca²⁺_i. Because Ang II increases inositol triphosphate (IP₃), it is conceivable that the peptide could produce a compartmentalized increase in Ca²⁺ through the IP₃ receptors, undetectable when measuring bulk cytosolic Ca²⁺. To explore this possibility, rat myocytes were cultured in the presence and absence of 25 μmol/L 2-APB (2-aminoethoxydiphenyl borate), an IP₃ receptor inhibitor.¹⁴ 2-APB did not prevent Ang II induced cell death: cell viability was 38.9±2.6% (Ang II) versus 36.0±3.2% (2-APB+Ang II) (n=4). These results make unlikely the possibility of a restricted increase in Ca²⁺ produced by IP₃ being the source for CaMKII activation. The effect of Ang II on cell viability and CaMKII activity was also assessed in the presence of 1 μmol/L of the Ca²⁺ chelator BAPTA-AM. BAPTA-AM was added to the culture when the cells were plated. After 1 hour, the medium was changed for a fresh medium containing 1 μmol/L BAPTA-AM and Ang II±KN-93. Control experiments showed that this concentration of BAPTA-AM prevented cell death and the increase in Ca²⁺ levels induced by exposing the cells to the Ca²⁺ ionophore A23128 (Figure 4B). Furthermore, BAPTA abolished twitch contraction and the associated Ca_iT and decreased Indo-1 ratio to values lower than the initial resting levels in electrically stimulated myocytes (results not shown). These experiments indicate that BAPTA-AM was properly loaded into the cell and that Ca²⁺ levels reached subdiastolic values in the presence of BAPTA-AM. Figure 4C shows that Ang II induced a decrease in cell viability and an increase in CaMKII activity and that these effects could be prevented by KN-93. Because Ang II increases ROS (Figure 1D), we assessed whether ROS could promote CaMKII activation and apoptosis in the presence of BAPTA-AM by measuring P-CaMKII and caspase-3 activity by fluorescence microscopy in myocytes stimulated with 200 μmol/L H₂O₂. Both P-CaMKII and caspase-3 activity were increased by H₂O₂ and again these increases were prevented by cotreatment with KN-93 (Figure 4D). These results indicate that Ang II–induced ROS production may activate CaMKII even at subdiastolic Ca²⁺ levels.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

To further explore whether the effect of ROS on CaMKII activation was attributable to a direct action of ROS on CaM, we performed experiments in isolated myocytes and in cardiac homogenates in vitro with the CaM inhibitor W-7. Figure 5A and 5B show that in cultured myocytes pretreated with BAPTA-AM, W-7 failed to affect H₂O₂-induced cell death and CaMKII activity (P-CaMKII and P-Thr17), whereas KN-93 prevented both effects. Of note, control experiments showed that when given alone, H₂O₂ produced an increase in CaMKII activity (assessed by P-Thr17 site of phospholamban) of 224±51% (n=3) that was associated with the decrease in cell viability observed in Figure 5A. BAPTA-AM, KN-93, and W-7 failed to significantly affect basal CaMKII activity when given alone. Similar results to those obtained in myocytes were obtained in vitro where H₂O₂ increased P-CaMKII in a nominally Ca²⁺-free medium (0 Ca²⁺-EGTA), both in the absence or in the presence of W-7 and no exogenous addition of CaM (Figure 5C). These results were obtained in homogenates from nonbeating hearts perfused with low Ca²⁺, nifedipine and EDTA, to favor Ca²⁺_i depletion, and in the presence of 1 to 10 mmol/L of EGTA in the in vitro phosphorylation assay. Because CaM is a protein with many methionine residues susceptible to oxidation,¹⁵ we cannot exclude the possibility that the lack of effect of W-7 could be attributable to the fact that oxidized CaM becomes resistant to the inhibitory effects of W-7. However, in in vitro experiments, W-7 was effective in diminishing by 60.8±9.9% the increase in P-CaMKII produced by Ca²⁺ addition in the presence of H₂O₂. These results would support the view that W-7 is able to inhibit CaM, even in the presence of H₂O₂ and would indicate that ROS can either directly activate CaMKII or shift the Ca²⁺/CaM dependence of CaMKII to extremely low Ca²⁺ and CaM values.

• Download figure
• Download PowerPoint

p38MAPK Activation Is Required for Ang II–Induced Cell Death

p38MAPK mediates apoptosis in several models of cardiac stress.^16,17 To evaluate the role of p38MAPK in Ang II–induced apoptosis, 2 different approaches were used: (1) rat and cat myocytes were cultured with Ang II±10 μmol/L of the p38MAPK inhibitor SB202190 (SB) (Figure 6A) and (2) gene transfer was performed to overexpress p38MAPK (Adv.p38) in rat myocytes (Figure 6B). In these experiments, β-galactosidase (Adv.βgal) was overexpressed to serve as control. Figure 6A shows that SB protected cat cardiomyocytes from the deleterious effect of Ang II. Similar results were obtained in rat cells (data not shown). Figure 6B shows that after 24 hours, cells infected with the adenovirus robustly expressed the reporter gene green fluorescent protein (GFP), indicating that the gene of interest was also overexpressed. Results show that although the viability of the p38MAPK-overexpressing cells was not different from that in β-galactosidase–overexpressing cells under control conditions, cells infected with Adv.p38 and incubated with Ang II showed reduced viability compared to β-galactosidase cells. These experiments reveal that p38MAPK is involved in Ang II–induced cell death. Figure 6C further shows that the increase in the activity of p38MAPK induced by Ang II was completely prevented by KN-93. However, SB failed to affect Ang II–induced CaMKII activity (Ang II: 194.1±26.8% of control; Ang II+SB: 180.9±21.5% of control). These findings indicate that activation of p38MAPK is downstream of CaMKII in the Ang II–induced apoptotic signaling cascade.

• Download figure
• Download PowerPoint

Discussion

Ang II has been shown to be implicated in many cellular physiological and pathological processes. In the near term, Ang II has been shown to modulate contractility and excitation–contraction coupling. Recent experiments demonstrated that the positive and negative inotropic effects of the peptide occur through different mechanisms. Whereas the positive inotropic effect was associated with an increase in intracellular Ca²⁺,^11,14 negative inotropism occurs in the rat, because of a p38MAPK-mediated decrease in myofilament Ca²⁺ responsiveness.¹² In the long term, Ang II has been associated with cardiomyocyte apoptosis. Since the pioneering studies of Anversa, this apoptotic action was associated with an increase in Ca²⁺_i.² From then, different molecules have been suggested to mediate the apoptotic effects of Ang II.^4,5,18 However, all of these previous reports failed to explore the sequence of events triggered by angiotensin II type 1 receptor activation that culminate in apoptotic cell death. The present results describe for the first time a sequence of concatenated events by which Ang II produces apoptosis. More importantly, the results indicate that activation of CaMKII may occur at subdiastolic Ca²⁺_i levels. Finally, it is shown that this pathway is common to species in which Ang II has opposite inotropic effects. Thus, the interspecies differences observed in the contractile effect of Ang II, appear to be absent in its apoptotic action, suggesting that a similar apoptotic cascade induced by Ang II may be operative across a wide range of species, including the human.

Ang II–Induced Apoptotic Cell Death Involves ROS-Dependent Activation of CaMKII

The present results demonstrate that Ang II induces ROS production in association with an increase in CaMKII activity. The inhibition of ROS and the prevention of CaMKII activation were both able to inhibit Ang II–induced cell death. Moreover, KN-93 diminished CaMKII phosphorylation and myocyte apoptosis without affecting ROS production, indicating that ROS are upstream of CaMKII (Figures 1 and 2). These results fully confirm the recently discovered ROS-induced CaMKII activation in the apoptotic cascade of Ang II.^8,9,19

CaMKII Activity Can Be Increased by ROS in Nominally Ca²⁺_i-Free Conditions and After Inhibition of CaM

Recent experiments indicate that sustained activation of CaMKII by ROS occurs by a modification of CaMKII at M281/282 residues. CaMKII activation by ROS appears to require previous “opening” of the kinase by Ca²⁺/CaM to allow access to the autoinhibitory domain for oxidation.⁸ Because our experiments were performed in quiescent myocytes, the source of Ca²⁺ required to initially activate the kinase was not obvious. Indeed, when Ca²⁺_i was measured in both species after 24 hours of Ang II stimulation, no increases in either diastolic and systolic Ca²⁺ or sarcoplasmic reticulum Ca²⁺ content was detected in stimulated myocytes (Figure 3). These results prompted us to explore in depth the putative source of activating Ca²⁺, considering different hypotheses. (1) The increase in Ca²⁺ occurred within the first hours during the culture period. This possibility is supported by experiments describing an increase in diastolic Ca²⁺ after incubation with Ang II.² (2) There is a compartmentalized increase in Ca²⁺, not detectable by our epifluorometric measurements of global cytosolic Ca²⁺ but nevertheless able to activate CaMKII. A release of Ca²⁺ by IP₃ receptors is a putative mechanism by which Ang II may produce a restricted increase in Ca²⁺_i.²⁰ (3) ROS may alter the affinity of CaM for CaMKII. It is known that CaM weakly associates with the CaM binding domain of the kinase even before binding to Ca²⁺²¹ and several cellular processes are regulated by Ca²⁺-free CaM.²² (4) Ang II does not affect basal Ca²⁺ levels but, through the increase in ROS, produces an inhibition of phosphatases that would allow the phosphorylation of CaMKII at resting Ca²⁺. This possibility was suggested in lymphocytes.¹⁰ (5) There might be conditions in which ROS may activate CaMKII directly without the requirement of Ca²⁺/CaM. (6) ROS may radically reset the CaMKII Ca²⁺ dependence to subphysiological values.

The first 2 hypotheses are not supported by the present results because: (1) we did not detect any significant increase in diastolic Ca²⁺ levels during the first 4 hours of incubation with Ang II and (2) inhibition of IP₃ receptors failed to prevent Ang II–induced cell death. More importantly, incubation of myocytes with BAPTA-AM or of cardiac homogenates with 0 Ca²⁺-EGTA was unable to inhibit both cell death and CaMKII activation induced by Ang II and H₂O₂, which makes an undetected compartmentalized increase in Ca²⁺ an unlikely candidate. The finding that the inhibition of CaM with W-7 in the presence of BAPTA-AM did not affect H₂O₂-induced CaMKII phosphorylation (Figure 5) would exclude the third possibility, ie, a direct CaM regulation by ROS. Indeed, recent experiments showed that oxidation of CaM diminishes rather than induces CaM-dependent CaMKII activation.²³ The fourth possibility, ie, an increase in CaMKII activity caused by a ROS-induced phosphatase inhibition able to increase the phosphorylated state of the kinase, is hard to reconcile with the fact that autophosphorylation of the kinase requires previous activation.²⁴ An inhibition of phosphatases would only be able to increase CaMKII activity, measured as P-CaMKII, if the kinase was previously activated to trigger its autophosphorylation (see further discussion below). According to this, one might conclude, at first sight, that Ang II induces a Ca²⁺/CaM–independent ROS-dependent activation of CaMKII in isolated myocytes. There are at least 3 main reasons, however, that strongly conspire against this possibility. First, experiments by Anderson and colleagues clearly showed that an increase in Ca²⁺_i is a prerequisite to produce Ang II–induced CaMKII activation.⁸ Second, although we used different strategies to decrease Ca²⁺_i, we are aware of the fact that a complete Ca²⁺_i depletion is virtually impossible in living systems.²⁵ Third, we were able to prevent Ang II–dependent CaMKII activation by KN-93, known to inhibit CaMKII by competing with Ca²⁺/CaM. Thus, our results strongly suggest that the presence of ROS resets the dependence of CaMKII to Ca²⁺ to extremely low, subdiastolic Ca²⁺_i levels. In this scenario, it is conceivable that an inhibition of phosphatases by ROS could also contribute to sustained CaMKII activation.¹⁰ In line with this idea, recent experiments indicate the need of phosphatase inhibition for the progression of CaMKII autophosphorylation.²⁶ Of note, buffering Ca²⁺_i (EGTA or BAPTA-AM) (Figures 4 and 5) may decrease the activity of the Ca²⁺-dependent phosphatase PP2B, which, by increasing inhibitor-1 phosphorylation, would inhibit PP1, one of the main phosphatases negatively involved in CaMKII phosphorylation.²⁷ This mechanism may be responsible for part of the basal phosphorylation levels of CaMKII or Thr17 site of phospholamban observed under these conditions. A contribution of phosphatases to the enhanced phosphorylation of these proteins over basal values, under the same Ca²⁺ buffering conditions or in the presence of phosphatase inhibitors (Figure 5C), is more difficult to conceive. However, based on the present results their contribution cannot be completely excluded.

Apoptotic Role of p38MAPK

Our results using SB and Ad.p38 demonstrate that p38MAPK is a crucial step in the apoptotic cascade triggered by sustained Ang II stimulation. Because KN-93 significantly decreased p38MAPK activity, whereas SB has no effect on Ang II–induced CaMKII activity, p38MAPK would be located downstream CaMKII activation (Figure 6). The potential for CaMKII to activate p38MAPK is not unique to this study. Several reports have shown that p38MAPK can be regulated by CaMKII through the activation of different intermediate molecules, including, apoptosis signal-regulating kinase 1,²⁸ Janus kinase 2,²⁹ and protein kinase R.³⁰ Although the signaling sequence downstream of p38MAPK was not studied in the present experiments, previous reports indicate that CaMKII activates p38MAPK,²⁹ which, in turn, activates mitochondrial death pathway by producing an unbalance between the expression of pro- and antiapoptotic proteins, Bax and Bcl-2.³¹ These results would suggest a downstream location of p38MAPK in the cascade of events initiated by Ang II that leads to apoptosis.

In summary, the present results describe for the first time the sequence of events whereby Ang II mediates apoptosis in cardiac cells. In this cascade, Ang II–induced ROS production appears as the initial trigger for the subsequent activation of CaMKII and the onset of the apoptotic process. The results presented herein further demonstrate that in the presence of ROS, activation of CaMKII may occur without the need of the conventional increase in Ca²⁺ and only requires extremely low Ca²⁺/CaM levels, revealing a novel and previously unrecognized mode of CaMKII activation in which ROS reset the Ca²⁺ dependence of CaMKII to subphysiological concentrations. Finally, this study shows that the Ang II–induced signaling pathway that results in cardiomyocyte apoptosis is conserved across different species and is operative independently of the inotropic responses elicited by the peptide. It is tempting to speculate that similar apoptotic signaling could occur and even be exacerbated in the failing heart, where Ang II levels, as well as CaMKII and p38MAPK activities, are increased.

Footnotes