CaMKII‐dependent ryanodine receptor phosphorylation mediates sepsis‐induced cardiomyocyte apoptosis

Abstract Sepsis is associated with cardiac dysfunction, which is at least in part due to cardiomyocyte apoptosis. However, the underlying mechanisms are far from being understood. Using the colon ascendens stent peritonitis mouse model of sepsis (CASP), we examined the subcellular mechanisms that mediate sepsis‐induced apoptosis. Wild‐type (WT) CASP mice hearts showed an increase in apoptosis respect to WT‐Sham. CASP transgenic mice expressing a CaMKII inhibitory peptide (AC3‐I) were protected against sepsis‐induced apoptosis. Dantrolene, used to reduce ryanodine receptor (RyR) diastolic sarcoplasmic reticulum (SR) Ca2+ release, prevented apoptosis in WT‐CASP. To examine whether CaMKII‐dependent RyR2 phosphorylation mediates diastolic Ca2+ release and apoptosis in sepsis, we evaluated apoptosis in mutant mice hearts that have the CaMKII phosphorylation site of RyR2 (Serine 2814) mutated to Alanine (S2814A). S2814A CASP mice did not show increased apoptosis. Consistent with RyR2 phosphorylation‐dependent enhancement in diastolic SR Ca2+ release leading to mitochondrial Ca2+ overload, mitochondrial Ca2+ retention capacity was reduced in mitochondria isolated from WT‐CASP compared to Sham and this reduction was absent in mitochondria from CASP S2814A or dantrolene‐treated mice. We conclude that in sepsis, CaMKII‐dependent RyR2 phosphorylation results in diastolic Ca2+ release from SR which leads to mitochondrial Ca2+ overload and apoptosis.

of intracellular Ca 2+ and more recently, its activity has also been shown to be modulated by ROS-dependent oxidation. 7,8 CaMKII is known to play an important physiological role in the regulation of cardiac excitation-contraction coupling. 9 However, under pathological conditions CaMKII has been shown be detrimental. [10][11][12] Indeed, we have recently shown that in sepsis, CaMKII is activated by oxidation and once active it phosphorylates the cardiac ryanodine receptor (RyR2) leading to inappropriate diastolic Ca 2+ release from the sarcoplasmic reticulum (SR), known as SR Ca 2+ leak. This SR Ca 2+ leak reduces SR Ca 2+ content and thus systolic Ca 2+ release, mediating at least in part the contractile dysfunction associated with sepsis. 13 In addition, CaMKII has been shown to be a common intermediate of diverse death stimuli that induce apoptosis of cardiac cells. 14 Consistent with these results, we have previously demonstrated that CaMKII activation is a critical step in the signalling cascade that leads to apoptosis in ischaemia/reperfusion injury, ouabain toxicity, sustained Angiotensin II stimulation and rapid pacing. 8,[15][16][17] Moreover, we and others have provided evidence demonstrating that CaMKII, through the phosphorylation of the RyR2, enhances SR Ca 2+ leak which is then taken up by the mitochondria, leading to mitochondrial Ca 2+ overload which triggers the apoptotic cascade. [17][18][19][20][21][22] The above-mentioned events could also hold true for sepsis. Thus, we hypothesized that CaMKII-dependent SR-Ca 2+ leak mediates apoptosis in sepsis.
The results presented herein, using a colon ascendant stent peritonitis (CASP) model of sepsis that closely mimics the clinical course of diffuse peritonitis with early and steadily increasing systemic infection and inflammation, 23 enhanced apoptosis 24 and contractile dysfunction 13 provide mechanistic insight on sepsis-induced apoptosis. We show that in sepsis, cardiac cell death involves CaMKII activation, which through the phosphorylation of RyR2 enhances SR-Ca 2+ leak leading to mitochondrial Ca 2+ overload and apoptosis. Importantly, we further show that the RyR2 plays a critical role in sepsis-induced cell death and that agents that reduce RyR2 open probability (RyR2 stabilizers) are effective in preventing apoptosis, suggesting a potential therapeutic advantage of these compounds to reduce apoptosis and contractile dysfunction associated with sepsis. or mutant mice where the CaMKII-dependent phosphorylation site on the RyR2 (site Serine 2814) is mutated to Alanine (S2814A) were used. All genetically modified mice were generated on a C57BL/6 background.

| Mouse model of sepsis
Colon ascendens stent peritonitis (CASP) surgery was performed to induce sepsis as previously described. 13 In brief, a small stent was inserted into the ascending colon of mice leading to the continuous leakage of intestinal bacteria into the peritoneal cavity resulting in peritonitis and systemic bacteraemia. Sham surgery was carried out using the identical surgical procedure, but without stent implantation. For more detail see online Supporting Information.

| Cardiomyocyte isolation
Myocytes were isolated by enzymatic digestion. 13 Details are provided under Supporting Information.

| Isolation of mouse heart mitochondria
Isolated hearts were washed and homogenized in ice-cold isolation solution (IS) consisting of 75 mmol/L sucrose, 225 mmol/L mannitol and 0.01 mmol/L EGTA neutralized with Trizma buffer at pH 7.4. After tissue pieces were settled, the entire supernatant was discarded and fresh IS was added and the mixture was transferred to a hand homogenizer. Proteinase (0.08 mg, bacterial, type XXIV, Sigma,) was added just before starting the homogenization procedure. The homogenate was carefully transferred after each step to a polycarbonate centrifuge tube. After 5 minutes of 750 × g of centrifugation to discard unbroken tissue and debris, the supernatant was centrifuged at 8000 × g for 10 minutes to sediment the mitochondria. The mitochondrial pellet was washed twice with IS and the last one with suspension solution (IS without EGTA, SS) at 8000 × g for 5 minutes each. The residue was washed and re-suspended in SS. The mitochondrial protein concentration was evaluated by the Bradford method 25 using bovine serum albumin as standard.

| Apoptosis assessment
Mitochondrial Cytochrome C levels and the ratio between proand anti-apoptotic proteins Bax and Bcl-2, respectively, were used as an index of apoptosis (see online Supporting Information for details). Apoptosis was also determined by TUNEL assay (In Situ Cell Death Detection Kit, TMR red, Roche, Mannheim, Germany).
TUNEL-positive cells were imaged under a fluorescence microscope (100 × magnification) and counted in 10 random fields from each experimental situation. The results were expressed as percentage of TUNEL-positive cells related to total number of cells. DAPI (1 µg/mL, 4', 6-Diamidino-2-phenylindole dihydrochloride, Sigma, St. Louis, MO) was used for nuclear staining.

| Immunoblotting
Homogenates, cytosolic fractions and SR membranes were prepared from the pulverized ventricular tissue. Proteins were electrophoresed and transferred to polyvinylidene fluoride membranes. 26 Blots were probed with antibodies raised against, anti-Bcl2 and anti-Bax. Immunoreactivity was visualized by a peroxidase-based chemiluminescence detection kit (Immobilon Western Millipore) using a Chemidoc Imaging System (Bio-Rad, Hercules, CA). Details are provided in the Supporting Information.

| Mitochondrial Membrane Potential
Mitochondrial membrane potential changes were evaluated by Isolated mitochondria loaded with RH-123 were excited at 503 nm, and fluorescence emission was detected at 527 nm. 27 During the measurements, the reaction medium containing mitochondria (0.1 mg/mL) was continuously stirred. Mitochondrial membrane potential (ΔΨm) was calculated following the instructions previously detailed using the Nernst-Guggenheim Equation 28 According to the authors, RH-123 uptake is in proportion to ΔΨm; therefore, the rate of fluorescence quenching is a function of ΔΨm, as well as the steady-state level of fluorescence decrease.

| Mitochondrial calcium retention capacity assay
The ability of mitochondria to retain exogenous Ca 2+ before the irreversible opening of the mitochondrial permeability transition pore (mPTP) occurs was monitored by following the changes in the fluorescence of the Ca 2+ sensitive indicator, Calcium Green-5N. 29 The reaction begins with the addition of successive pulses of 5 µmol/L Ca 2+ .
After each addition, the fluorescence in the medium increases until the mitochondria start to take up Ca 2+ and then it decreases. When mitochondria are sufficiently loaded with Ca 2+ , the opening of the mPTP occurs and the release of the mentioned ion in the medium increases again the fluorescence. 30 For this assay, 0.3 mg/mL of isolated mito- Calcium retention capacity (CRC) was defined here as the amount of Ca 2+ required to trigger a massive Ca 2+ release by isolated cardiac mitochondria. 29 CRC is used as an indicator of the resistance of the mPTP to opening after matrix Ca 2+ accumulation and expressed as nmol CaCl 2 per mg of mitochondrial proteins.

| Statistical analysis
The normal distribution of the data was corroborated with the Kolmogorov-Smirnov test; unpaired Student's t test, Mann-Whitney test and one-way ANOVA were applied accordingly using GraphPad Prism 5.0 (GraphPad Software, USA). The data are presented as means ± SEM. The level of significance was set at P < .05.

| Apoptosis is enhanced in a polymicrobial model of sepsis
Apoptosis was assessed by four independent methods, in heart homogenates, cardiomyocytes and cardiac mitochondria isolated from colon ascendens stent peritonitis operated mice (CASP) and from Sham-operated controls. Heart homogenates, cardiomyocytes and cardiac mitochondria were prepared 24 hours after surgery when contractile dysfunction is already installed. 13 Figure 1 depicts typical blots and overall results showing that the ratio between pro-apoptotic (Bax) and anti-apoptotic (Bcl2) proteins is increased in CASP homogenates relative to Sham, indicative of enhanced apoptosis in septic hearts. Similarly, mitochondrial cytochrome C content and mitochondrial membrane potential (ΔΨm) were significantly reduced in CASP vs Sham hearts showing that mitochondrial pathway-induced apoptosis was enhanced in septic hearts. In addition, there was a significant increase in TUNEL-positive nuclei in myocytes isolated from CASP compared to myocytes isolated from Sham mice.

| CaMKII mediates apoptosis associated with sepsis
We recently reported that CaMKII mediates at least part of the contractile dysfunction associated with sepsis, and several laboratories including our own have provided clear evidence that CaMKII is involved in triggering apoptosis in several disease models. [15][16][17]20,31 To test whether CaMKII is also involved in sepsis-induced apoptosis, we assessed apoptosis in heart homogenates and isolated mitochondria from CASP-and Sham-operated transgenic mice, expressing either a CaMKII inhibitory peptide, AC3-I, or a scramble non-inhibitory peptide, AC3-C.    Ca 2+ , mitochondrial CRC is exceeded and stored mitochondrial Ca 2+ is released into the extramitochondrial space through the opening of the mPTP. As observed in the typical traces shown in Figure 5A, mitochondria isolated from CASP-operated C57 mice resisted the addition of significantly fewer 5 µmol/L pulses of Ca 2+ before the opening of the mPTP, suggesting that these mitochondria have higher resting intramitochondrial Ca 2+ than Sham. In addition, Figure 5B

| D ISCUSS I ON
Our previous study demonstrates that CaMKII-dependent SR-Ca 2+ leak reduces the cytosolic Ca 2+ transient amplitude, contributing to the reduced contractility associated with sepsis. 13 In addition, several reports have proposed myocyte cell death by apoptosis as a critical process that contributes to contractile dysfunction of the septic heart. [37][38][39] Interestingly, sustained activation of CaMKII has been shown to be pro-apoptotic. 10 Indeed, experimental evidence from our laboratory demonstrates that CaMKII activation is a primary event in the signalling cascade that leads to apoptosis in ischaemia/ reperfusion injury, ouabain toxicity, rapid pacing and under sustained Angiotensin II stimulation. 8,[15][16][17] These results suggest that CaMKII, in addition to mediating a reduction in myocyte Ca 2+ transient amplitude in sepsis, could favour the apoptotic process. However, the subcellular mechanisms underlying CaMKII-dependent apoptosis in sepsis have yet to be determined. Thus, in the present study, using a colon ascendens stent peritonitis (CASP) mouse model that closely mimics the clinical situation of abdominal sepsis, we investigated the mechanisms involved in CaMKII-dependent apoptosis.
Consistent with previous reports, we confirmed that the CASP model of sepsis promotes cardiac apoptosis, as evidenced by an enhanced ratio between pro-and anti-apoptotic proteins, Bax and Bcl2, respectively (Bax/Bcl2), a decrease in mitochondrial cytochrome C levels, a significant decrease in mitochondrial membrane potential and by an increase in TUNEL staining. These results showing altered mitochondrial homeostasis further suggest that cardiac myocyte cell death, in sepsis, is mediated at least in part, by the mitochondrial pathway of apoptosis.

| CaMKII mediates sepsis-induced cell death
Oxidative stress has been shown to be enhanced in sepsis, and experimental evidence demonstrates that antioxidants can prevent F I G U R E 4 Reducing RyR2 open probability prevents sepsis-inducedapoptosis. Typical blots and overall results of the effect of dantrolene pretreatment on sepsis-induced apoptosis. WT-CASP + dantrolene mice hearts showed non-significant (NS) changes in Bax/Bcl2, mitochondrial cytochrome C expression and ∆ψm compared to Sham + dantrolene-operated WT mice (n = 4 per group). All data are expressed as means ± SEM * P < .05 (unpaired t test) vs Sham)

| Mechanism underlying CaMKII-dependent apoptosis in sepsis
Recent experimental evidence from our laboratory, in the setting of ischaemia and reperfusion injury, suggests that CaMKII may promote apoptosis by altering the coupling between SR Ca 2+ release and mitochondrial Ca 2+ uptake, resulting in mitochondrial Ca 2+ overload and opening of the mPTP. 20 Indeed, the interplay between SR and mitochondria under different stimuli has been known for many years to be pivotal in triggering apoptotic signals. 18,19 Supporting the critical role played by the SR/mitochondrial interaction in CaMKII-dependent apoptosis, Zhang et al concluded that enhanced activity of CaMKII results in RyR2 phosphorylation leading to enhanced SR Ca 2+ leak and mitochondrial Ca 2+ elevation, associated with exacerbated cell death in transgenic mice lacking phospholamban and overexpressing CaMKIIαδc. 33 Thus, we hypothesized that sepsis-induced cell death could be, at least in part, due to CaMKII-dependent post-translational modification of the RyR2 resulting in SR Ca 2+ leak which would be taken up by the mitochondria provoking mitochondrial Ca 2+ overload. To test this hypothesis, we used myocytes isolated from transgenic mice with RyR2 site Serine 2814 mutated to Alanine (S2814A) and therefore not phosphorylatable by CaMKII. We observed that S2814A CASP mice were protected from sepsis-induced apoptosis ( Figure 3). Moreover, CASP-operated S2814A mice had a higher survival rate than CASP-operated WT mice (see results in Supporting Information). We have previously shown that CASP S2814A mice have a similar degree of CaMKII activation than CASP wild-type mice. Moreover, site Thr17 of phospholamban, a specific target of CaMKII, also has a similar level of phosphorylation F I G U R E 5 Role of mitochondrial Ca 2+ in sepsis. A, Typical tracings of changes in the fluorescence of Ca 2+ indicator Calcium Green-5N. Arrows indicate Ca 2+ addition to the mitochondrial suspension. Increasing extramitochondrial Ca 2+ results in a rapid increase in the fluorescent signal followed by a decline in the fluorescence intensity of the sensor resulting from mitochondrial Ca 2+ uptake. WT-Sham mitochondria showed enhanced tolerance to extramitochondrial Ca 2+ addition compared to WT-CASP mitochondria. The measurement of extramitochondrial Ca 2+ shows that WT-Sham cardiac mitochondria are able to accumulate significantly more Ca 2+ than WT-CASP mitochondria (CRC) (n = 4). Mann-Whitney test results are expressed as mean ± SEM. B: Typical traces and overall results of changes of Calcium Green fluorescence after Ca 2+ addition in samples of mitochondria from S2814A mutant CASP mice CASP mice compared to Sham. Mitochondria from CASP mice showed preserved CRC compared to Sham (n = 4). Unpaired t test, results are expressed as mean ± SEM  Interestingly, a recent report has shown that dantrolene preserves cardiac mitochondria and prevents contractile dysfunction associated with sepsis. 34 However, if this compound can also prevent sepsis-induced apoptosis has not been evaluated. Consistent with this possibility, we recently reported that rapid pacing-induced apoptosis, which is also CaMKII dependent, can be prevented by reducing RyR2 open probability with the carvedilol-non-β-blocking analog, VK-II-86. 20 Thus, we hypothesized that reducing RyR2 open probability with dantrolene could be effective preventing sepsis-induced cell death. We show herein that dantrolene treatment is able to prevent sepsis-induced cell death ( Figure 4) and that this treatment also enhances survival rate of CASP-operated WT mice (see results in online Supporting Information). These results highlight the possibility of using RyR2 stabilizers as a therapeutic strategy to reduce the adverse remodelling associated sepsis.
To assess whether CaMKII-dependent increase is SR Ca 2+ leak results in mitochondrial Ca 2+ overload, we measured mitochondrial CRC in mitochondria isolated from Sham and CASP mice. As observed in Figure 5A, CRC was significantly reduced in wild-type CASP mitochondria, suggesting that these mitochondria were Ca 2+ overloaded and therefore tolerated less Ca 2+ loading until mPTP opening compared to Sham. Consistent with mitochondrial Ca 2+ overload resulting from CaMKII-dependent RyR2 phosphorylation and enhanced Ca 2+ leak from the SR, the decrease in CRC observed in CASP mitochondria was absent in CASP mitochondria isolated from S2814A mice ( Figure 5B) and from mice treated with dantrolene ( Figure S1). Taken together, and similar to the conclusion of Zhang et al 33 Figure S2). Based on these observations, it is likely that the protective effect of dantrolene is mainly due to its effect on RyR2. Another potential limitation of our study is that we did not examine the mechanism underlying the enhanced survival of Taken together, these findings propose CaMKII and RyR2 as new therapeutic targets whose inhibition or stabilization, respectively, could serve to ameliorate the cardiac symptoms of sepsis.

ACK N OWLED G EM ENTS
The technical support of Omar Castillo is gratefully acknowledged. We also wish to thank Dr Xander Wehrens from University of Houston, Texas, and Dr Mark Anderson from Johns Hopkins University, Baltimore, Maryland, for their generous gift of Ser2814A and AC3 mice breeding pairs, respectively.

CO N FLI C T O F I NTE R E S T S
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.