Phospholamban ablation rescues SR Ca²⁺ handling but exacerbates cardiac dysfunction in CaMKIIδ_C transgenic mice

Abstract

Rationale

We previously showed that transgenic mice expressing Ca²⁺/calmodulin-dependent protein kinase II δ_C (CaMKII-TG) develop dilated cardiomyopathy associated with increased ryanodine receptors (RyR2) phosphorylation, enhanced sarcoplasmic reticulum (SR) Ca²⁺ leak and lowering of SR Ca²⁺ load. We hypothesized that phospholamban (PLN) ablation would restore SR Ca²⁺ load and prevent the decreased ventricular contractility, dilation and mortality seen in CaMKII-TG.

Objective

Our objective was to generate CaMKII-TG mice lacking PLN, determine if the maladaptive effects of cardiac CaMKIIδ_C expression were corrected and establish the mechanistic basis for these changes.

Methods and Results

CaMKII-TG were crossed with PLN knockout (PLN-KO) mice to generate KO/TG mice. Myocytes from wild type (WT), CaMKII-TG, PLN-KO and KO/TG were compared. The decreased SR Ca²⁺ load and twitch Ca²⁺ transients seen in CaMKII-TG were normalized in KO/TG. Surprisingly the heart failure phenotype was exacerbated as indicated by increased left ventricular dilation, decreased ventricular function, increased apoptosis and greater mortality. In KO/TG myocytes SR Ca²⁺ sparks and leak were significantly increased, presumably due to the combined effects of restored SR Ca²⁺ load and RyR2 phosphorylation. Mitochondrial Ca²⁺ loading was increased in cardiomyocytes from KO/TG vs. WT or CaMKII-TG mice and this was dependent upon elevated SR Ca²⁺ sparks. Cardiomyocytes from KO/TG showed poor viability, improved by inhibiting SR Ca²⁺ release and mitochondrial Ca²⁺ loading.

Conclusions

Normalizing cardiomyocyte SR Ca²⁺ loading in the face of elevated CaMKII and RyR2 phosphorylation leads to enhanced SR Ca²⁺ leak and mitochondrial Ca²⁺ elevation, associated with exacerbated cell death, heart failure and mortality.

Introduction

Ca²⁺ is a critical second messenger in cardiac function. The Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) is regulated by and involved in control of Ca²⁺ cycling in the myocardium. Several years ago we generated transgenic mice expressing CaMKIIδ, the predominant cardiac CaMKII isoform. Mice expressing the CaMKIIδ_C splice variant in the myocardium (CaMKII-TG) develop heart failure (HF) associated with ventricular dilation, marked decreases in fractional shortening, and mortality.¹ We established that there was increased phosphorylation of the cardiac ryanodine receptor (RyR2) at the CaMKII site, associated with increased sarcoplasmic reticulum (SR) Ca²⁺ spark frequency and increased diastolic SR Ca²⁺ leak. More recently we generated CaMKIIδ knockout mice and further demonstrated that HF induced by pressure overload was dependent on CaMKIIδ-mediated RyR2 phosphorylation and increased Ca²⁺ sparks.²

The increased diastolic Ca²⁺ leak seen in the CaMKII-TG resulted in profound SR Ca²⁺ depletion and reduced Ca²⁺ transients and contractions. Accordingly, we hypothesized that diminished SR Ca²⁺ load could be responsible for the decreased contractile function and concomitant morbidity that characterized these mice. Cardiac SR Ca²⁺ uptake is regulated through the interaction of phospholamban (PLN) with the SR Ca²⁺ ATPase (SERCA2a).³ PLN knockout mice (PLN-KO) exhibit enhanced myocardial contractility,⁴ which can be explained by the increased SERCA2a pump activity and higher SR Ca²⁺ load in cardiomyocytes isolated from these mice.⁵ PLN ablation was shown to rescue both the functional and structural cardiomyopathy seen in mice in which the muscle LIM protein (MLP) is genetically deleted.⁶ Rescue of this genetic model of dilated cardiomyopathy and HF led to the proposal that ablation or inhibition of PLN would have therapeutic value in treatment of HF of various etiologies. Subsequently PLN-KO was shown to rescue cardiac hypertrophy and ventricular dysfunction in calsequestrin- and β₁-adrenergic receptor-overexpressing mice.^7,8 In light of the cellular changes observed in the CaMKII-TG mice, in particular the prominent SR Ca²⁺ depletion, we postulated that cardiac dysfunction would be improved if SR Ca²⁺ was normalized by PLN ablation.

Experiments reported here examined the effects of crossing CaMKII-TG¹ with PLN-KO mice.⁴ Our studies demonstrated the desired and predicted effect of PLN ablation on SR Ca²⁺, e.g. complete rescue of the SR Ca²⁺ depletion seen in the CaMKII-TG. Additionally, contractile function of isolated myocytes (as reflected by Ca²⁺ transients) was improved. Unexpectedly, however, rather than improving the in vivo HF phenotype, the PLN-KO/CaMKII-TG (KO/TG) crosses showed exaggerated HF, with a more rapid onset of lethality and further decreases in contractile function.

The basis for the exaggerated HF phenotype, observed in the absence of changes in other Ca²⁺ handling proteins, is examined in the studies presented here. The data demonstrate that in KO/TG myocytes there are greater increases in SR Ca²⁺ sparks and leak, reflective of the now combined effects of increased SR Ca²⁺ load and RyR2 phosphorylation. Cardiomyocyte apoptosis is also increased in KO/TG. Using isolated cardiomyocytes, we observed increased mitochondrial Ca²⁺ in KO/TG and determined that the combined effects of increased SR Ca²⁺ release and mitochondrial Ca²⁺ loading contribute to diminished cardiomyocyte viability. We suggest that interventions that increase SR Ca²⁺ in the face of enhanced diastolic Ca²⁺ leak predispose to cardiomyocyte cell death and likely other Ca²⁺ mediated toxicities that compromise survival.

Methods

Mice

CaMKII-TG mice¹ were crossed with PLN-KO mice⁴ and heterozygous PLN offspring carrying the CaMKII transgene inbred with the ones without the transgene. Four genotypes of mice including wild type (WT), PLN-KO, CaMKII-TG, and PLN ablation with CaMKIIδ_C overexpression (KO/TG) were used for experiments. All mice used in the present study were of mixed gender (more male than female) at 8 weeks of age, unless otherwise noted. All procedures were performed in accordance with Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee.

Ca²⁺ Measurements of Adult Mouse Ventricular Myocytes

Mouse ventricular myocytes were isolated and loaded with indo-1-AM to assess SR Ca²⁺ content and twitch [Ca²⁺]_i transients, and with fluo-4-AM for SR Ca²⁺ spark measurements as described previously.⁹ Ca²⁺ sparks were measured in intact myocytes and were evaluated as described (IDL 5.3 computer software).⁹ Ca²⁺ spark amplitudes were normalized (F/F₀) to fluorescence baseline (F₀). Duration of the Ca²⁺ sparks was taken as the full duration above half maximum (FDHM) and width was the full spatial size above half maximum (FWHM). A crude index of diastolic SR Ca²⁺ leak was calculated as Ca²⁺ spark frequency (CaSpF) × FDHM × FWHM × amplitude. Some myocytes were pre-incubated for 30 min with a cell permeable myristoylated CaMKII inhibitor, autocamtide-2 related inhibitory peptide II (AIP, Calbiochem # 189485), before measurements.

Mitochondrial Ca²⁺ Measurements of Adult Mouse Ventricular Myocytes

Mitochondrial Ca²⁺ was measured indirectly by examining the FCCP-induced increase in [Ca²⁺]_i in myocytes loaded with 10 µM fura-2-AM (Molecular Probes; 30 min plus 10 min de-esterification). Fura-2 was excited at 340 ± 10 nm and 380 ± 10 nm, and emitted fluorescence (535 ± 20 nm) was recorded at 100 Hz. Background-subtracted ratio fluorescence (R = F₃₄₀/F₃₈₀) was converted to [Ca²⁺]_i using [Ca²⁺]_i = K_dβ(R-R_min)/(R_max –R). R_min, R_max, K_d and β were determined experimentally. To assess baseline SR Ca²⁺ content myocytes were first paced at 0.5 Hz to steady-state, and 10 mM caffeine was rapidly applied. After caffeine washout, myocytes were paced at 0.5 Hz to re-attain the same steady state. They were then exposed to 0Na⁺/0Ca²⁺/0K⁺ solution to block the Na⁺/Ca²⁺ exchanger (NCX) and Na⁺/K⁺ pump and treated with 10 µM thapsigargin for ~45 seconds to block the SERCA pump.¹⁰ Under these conditions, when 1 µM FCCP is applied to abolish the mitochondrial membrane potential, mitochondrial Ca²⁺ is released and the rise in [Ca²⁺]_i is indicative of mitochondrial Ca²⁺ content. Oligomycin (1 µM) was included with FCCP to block rapid ATP consumption. SR Ca²⁺ load assessed by application of 10 mM caffeine at the end of the protocol showed that the SR was still well Ca²⁺-loaded (and subsequent relaxation and [Ca²⁺]_i decline) confirm that ATP was not dissipated by the protocol. The SR cannot be reloaded after this depletion, confirming the thapsigargin-dependent SR Ca²⁺-ATPase inhibition.¹⁰

Transthoracic Echocardiography

Transthoracic echocardiography was performed in mice using an Agilent Technologies Sonos 5500 system with a 15 MHz transducer as described.¹

Western Blotting

Western blot analysis was carried out in cardiac homogenates as described previously.¹ The antibodies used for immunoblotting were as followings: rabbit anti-calsequestrin (CsQ), rabbit anti-SERCA2, mouse anti-RyR, mouse anti-phospho-CaMKII (Affinity Bioreagents), rabbit Ser2809 phospho-RyR2 antibody (Badrilla, UK), rabbit Ser2815 phospho-RyR2 antibody (a gift from A.R. Marks, Columbia University), and anti-NCX (monoclonal R3F1, a gift from K.D. Philipson, UCLA).

TUNEL Staining

Transverse sections o f mouse hearts were labeled with fluorescein-terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labeling (TUNEL) using a kit from Roche (Lewes, UK) according to the manufacturer’s instructions. Slides were mounted with Vectashield mounting media containing DAPI to stain for the nuclei. Wheat germ agglutinin (WGA) was used to stain cell membranes and cardiomyocytes identified by their centrally located nuclei. Labeled nuclei were counted to determine the apoptotic index (number of labeled nuclei/10⁵ total myocyte nuclei).

Cell Viability Study in Adult Mouse Ventricular Myocytes

Ventricular myocytes were isolated as described previously.¹¹ After isolation, cells were plated for 1 hour on laminin-coated dishes in MEM-HBSS (Hanks’ Balanced Salt Solutions) medium containing 5% serum. Cells were then washed, and serum-free medium added along with vehicle, 2 µM KN-93, 100 nM ryanodine, 5 µM Ru-360 or 5 µM cyclosporine A. Using a grid marking system, 8 fields were chosen per dish and the number of rod shaped living cells was counted at various times over a 12 hour time course. At the 12 hour time point, living cells were quantitated by trypan blue exclusion.

Statistical Analysis

All data are reported as mean ± SEM. Statistical significance of differences between groups was determined using one-way ANOVA with Tukey post-hoc test. P value <0.05 was considered statistically significant.

Results

SR Ca²⁺ Load and Ca²⁺ Transients in Isolated Cardiomyocytes

To determine whether PLN ablation normalized the depressed SR Ca²⁺ load observed in the CaMKIIδ_C TG mice, SR Ca²⁺ content was assessed by quantifying caffeine-induced Ca²⁺ transients after 1 Hz steady state stimulation (Fig. 1A). Consistent with our previous reports, myocytes isolated from PLN-KO showed increased SR Ca²⁺ load⁵ while those from CaMKII-TG mice exhibited significantly depressed SR Ca²⁺ load⁹ (Fig. 1A–B). SR Ca²⁺ load was restored to levels equivalent to those of WT mice in myocytes from the KO/TG crosses (Fig. 1A–B). NCX function, assessed as the rate constant of [Ca²⁺]_i decline (k_Ca) in the presence of caffeine, was significantly faster in CaMKII-TG vs. WT as shown previously,⁹ and PLN ablation did not normalize this difference in NCX function (Fig. 1C). Twitch Ca²⁺ transient amplitude (Δ[Ca²⁺]_i at 1 Hz) was also restored in KO/TG vs. CaMKII-TG myocytes (Fig. 2). In fact, Ca²⁺ transients in KO/TG were greater than those in WT, likely reflecting the combination of restored SR Ca²⁺ content along with CaMKII-dependent enhancement of fractional SR Ca²⁺ release.^9,12

Figure 1. PLN ablation in CaMKIIδc TG mice repletes SR Ca²⁺ load in isolated cardiomyocytes.

A, Representative caffeine-induced Ca²⁺ transient traces. B, Average Δ [Ca²⁺]_i for caffeine-induced Ca²⁺ transients (SR Ca²⁺ load). *P<0.05 vs. WT. # P<0.05 vs. CaMKII-TG. C, Average rate constant k_Ca for [Ca²⁺]_i decline during caffeine exposure (which indicates NCX activity). *P<0.05 vs. WT. **P<0.01 vs. WT. The number of cells studied: WT n=6, PLN-KO n=21, CaMKII-TG n=9, and KO/TG n=9.

Figure 2. PLN ablation in CaMKIIδc TG mice normalizes twitch Ca²⁺ transients in isolated cardiomyocytes.

A, Representative twitch [Ca²⁺]_i transient traces. B, Average Δ [Ca²⁺]_i for twitches-induced Ca²⁺ transients. *P<0.05 vs. WT. # P<0.05 vs. CaMKII-TG. The number of cells studied: WT n=6, PLN-KO n=21, CaMKII-TG n=9, and KO/TG n=9.

In Vivo Global Cardiac Function

To determine whether the restoration of myocyte Ca²⁺ handling was associated with a diminished hypertrophic/HF phenotype and improved hemodynamic function, we measured the heart/body weight (HW/BW) ratio and assessed in vivo cardiac diameters and function by echocardiography. As reported previously, CaMKII-TG mice showed significant cardiac enlargement at 8 weeks of age as assessed by HW/BW ratio;¹ surprisingly the increase in HW/BW was even greater in KO/TG mice (Fig. 3A) as was cell size (not shown). Left ventricular dilation and cardiac dysfunction were also worsened in KO/TG mice relative to CaMKII-TG, as revealed by significantly greater left ventricular end diastolic dimension (LVEDD), diminished factional shortening (FS), and posterior wall thinning at 6–8 weeks (Fig. 3B–C and Online Figure I). Differences between KO/TG and CaMKII-TG mice were most significant at earlier stages (4–5 weeks), becoming less divergent at 12–16 weeks (Fig. 3B–C and Online Figure I).

Figure 3. PLN ablation in CaMKIIδc TG mice exaggerates the phenotype of dilated cardiomyopathy.

A, Heart weight/body weight (HW/BW) ratio at 8 wks. B and C, Averaged echocardiographic parameters at different ages. Data are presented for B, left ventricular end-diastolic diameter (LVEDD), and for C, calculated percent fractional shortening (FS). n=4 to 6 mice per group at each age; *P<0.05 vs. WT. # P<0.05 vs. CaMKII-TG. D, Survival curves for CaMKII-TG and PLN-KO/CaMKII-TG mice. PLN ablation accentuates premature death in CaMKIIδ_C TG mice. Numbers of mice were as follows: CaMKII-TG, n=16; KO/TG, n=22. P <0.01 KO/TG vs. CaMKII-TG.

Survival Analyses

As described previously,¹ CaMKIIδ_C TG mice showed premature death. KO/TG mice demonstrated worsened mortality compared to CaMKII-TG mice with 70% of the KO/TG mice dying by 24 weeks of age (Fig. 3D).

Expression and Phosphorylation of Ca²⁺ Regulatory Proteins

Western blot analyses were performed to assess the expression levels of calsequestrin (CsQ), SERCA2a and NCX in ventricular homogenates from CaMKII-TG and KO/TG mice. As previously reported,⁹ in CaMKII-TG vs. WT we found reduced SERCA2a and elevated NCX expression, but unaltered CsQ expression (data not shown). There were no significant differences in the expression of these major Ca²⁺ regulatory proteins in KO/TG versus CaMKII-TG mice (Fig. 4A). Phosphorylation of the RyR2, assessed using antibodies to the CaMKII phosphorylation sites (Ser2815 and Ser2809), was increased significantly, but not differentially, in CaMKII-TG and KO/TG mice (Fig. 4B). These data indicate that the exaggerated phenotype observed in KO/TG mice is not secondary to changes in key Ca²⁺ handling proteins or greater RyR2 phosphorylation.

Figure 4. Quantitative immunoblotting of major Ca²⁺ handling proteins in mouse ventricular homogenates.

A, Expression levels of calsequestrin (CsQ), SR Ca²⁺ ATPase (SERCA) and Na⁺/Ca²⁺ exchanger (NCX) did not change significantly between the CaMKII-TG and the PLN-KO/CaMKII-TG. B, Phosphorylation of ryanodine receptors (RyR2) at Ser2809 and Ser2815 (normalized to total RyR2) was increased significantly but not differentially in CaMKIIδc TG and PLN-KO/CaMKII-TG mice. n=4 mice per group; *P<0.05 vs. WT.

SR Ca²⁺ Sparks and Ca²⁺ Leak in Isolated Cardiomyocytes

We previously demonstrated increased Ca²⁺ spark frequency in CaMKII-TG vs. WT myocytes.⁹ To determine whether the enhanced SR Ca²⁺ content seen with PLN ablation was associated with a further increase in SR Ca²⁺ sparks and leak, Ca²⁺ spark frequency (CaSpF), duration (FDHM), spatial width (FWHM) and amplitude were assessed. CaSpF increased dramatically in KO/TG compared to CaMKII-TG. This was entirely blocked by pretreatment with a CaMKII inhibitory peptide, AIP (Fig. 5A) indicating that it resulted from ongoing CaMKII activity. The overall diastolic SR Ca²⁺ leak index (CaSpF × FDHM × FWHM × amplitude) was also significantly increased in KO/TG mice (7.5-fold relative to WT and 2.5-fold relative to CaMKII-TG) (Fig. 5B). The augmented SR Ca²⁺ sparks and leak in KO/TG crosses can be explained in terms of the combined effects of normalized SR Ca²⁺ load in the face of enhanced RyR2 phosphorylation. Notably enhanced SR Ca²⁺ loading, Ca²⁺ spark frequency and Ca²⁺ transients are observed in KO/TG myocytes despite lower diastolic [Ca²⁺]_i levels measured under comparable conditions (Online Figure II).

Figure 5. Enhanced SR Ca²⁺ spark frequency and Ca²⁺ leak in intact cardiomyocytes from PLN-KO/CaMKII-TG mice.

A, SR Ca spark frequency (CaSpF). *P<0.05 vs. WT. # P<0.05 vs. CaMKII-TG. +P<0.05 vs. KO/TG. B, SR Ca²⁺ leak. *P<0.05 vs. WT. # P<0.05 vs. CaMKII-TG. +P<0.05 vs. KO/TG.

Apoptosis in Heart Sections

Cardiomyocyte dropout could explain why there is reduced cardiac function, ventricular thinning and worsened HF despite enhanced Ca²⁺ transients in KO/TG myocytes. TUNEL staining was performed to assess apoptosis in paraffin sections prepared from WT, PLN-KO, CaMKII-TG, and KO/TG mice. DAPI staining and WGA were used to identify cardiomyocytes (see Methods). TUNEL labeled cardiomyocyte nuclei were counted to determine the apoptotic index (number of labeled nuclei/10⁵ total myocyte nuclei). As shown in Fig. 6, the number of TUNEL positive cardiomyocytes was significantly increased (~2.6 fold) in KO/TG mouse hearts compared with WT or CaMKII-TG mice.

Figure 6. Apoptosis occurs more frequently in PLN-KO/CaMKII-TG cardiomyocytes than in CaMKIIδc TG.

A, Representative TUNEL staining images in CaMKII-TG and PLN-KO/CaMKII-TG mouse heart sections. B, Average TUNEL positive myocytes. n=4 mice per group; *P<0.05 vs. WT. Data were normalized to WT; the average number from all experiments was ~75 TUNEL positive nuclei/10⁵ myocytes in the KO/TG group.

Cell Viability Studies

The assessment of TUNEL staining captures a single time point, but suggests that the viability of cardiomyocytes is compromised in the KO/TG line. To further examine cell viability, we isolated and monitored cells by morphology and trypan blue exclusion at various times (Fig. 7). To test whether CaMKII or SR Ca²⁺ release were important in regulating cell viability, cells were pretreated with 2 µM KN-93 (to inhibit CaMKII) or 100 nM ryanodine (to deplete SR Ca²⁺ and prevent Ca²⁺ sparks). WT myocytes remained viable over a 12 hour period and survival was not appreciably altered by KN-93 or ryanodine (Fig. 7A). In contrast, the percent of live myocytes from both CaMKII-TG and KO/TG mice decreased dramatically over 12 hours (to less than 30%). Inhibition of CaMKII with KN-93 or of SR Ca²⁺ leak with ryanodine decreased the rate of myocyte death in both CaMKII-TG and KO/TG mice (Fig. 7B–C), suggesting that the CaMKII-mediated increase in SR Ca²⁺ release contributes to diminished myocyte viability.

Figure 7. Inhibition of SR Ca²⁺ leak by KN-93 or ryanodine decreases the rate of death in isolated cardiomyocytes from PLN-KO/CaMKII-TG mice.

Isolated cardiomyocytes were plated and incubated in the presence and absence of inhibitors, and cell viability assessed at various times using morphology and at 12 hours by trypan blue exclusion. A, WT myocytes showed only slight decreases in viability over 12 hours after plating. B–C, The percent of living myocytes from both CaMKII-TG and PLN-KO/CaMKII-TG mice decreased dramatically over 12 hours. Inhibition of CaMKII with KN-93 or of SR Ca²⁺ leak with ryanodine decreased the rate of myocyte death in both CaMKII-TG and PLN-KO/CaMKII-TG mice.

Mitochondrial Ca²⁺ in Isolated Cardiomyocytes

The dramatic increase in diastolic Ca²⁺ sparks could increase mitochondrial Ca²⁺ loading and thereby contribute to decreased cardiomyocyte viability. To explore this possibility we assessed mitochondrial Ca²⁺ content in isolated adult myocytes from WT, CaMKII-TG or KO/TG mice as detailed in Methods (Fig. 8A). The protocol was designed to allow mitochondrial Ca²⁺ loading to occur during regular pulses (see initial cytosolic [Ca²⁺]_i transients) and a 40 s rest period where Ca²⁺ sparks could occur. The release of mitochondrial Ca²⁺ was then induced using FCCP, under conditions where the released Ca²⁺ is trapped in the cytosol (SERCA, NCX and Na⁺/K⁺-ATPase inhibited) and quantitated by the rise in [Ca²⁺]_i. Fig. 8A demonstrates that the FCCP-induced rise in [Ca²⁺]_i in KO/TG (black trace) is substantially larger than that in WT (gray trace), reflecting a greater mitochondrial Ca²⁺ load. Mitochondrial Ca²⁺ load was more than 4 fold higher in KO/TG vs. WT or CaMKII-TG myocytes (Fig. 8B).

Figure 8. Involvement of mitochondrial Ca²⁺ in myocyte death in PLN-KO/CaMKII-TG mice.

A, Mitochondrial Ca²⁺ was measured as shown in the protocol. NT: no treatment; Thaps: thapsigargin. B, Mitochondrial Ca²⁺ was increased in cardiomyocytes isolated from PLN-KO/CaMKII-TG mice. *P<0.05 vs. WT. Mitochondrial Ca²⁺ in PLN-KO/CaMKII-TG myocytes returned to control levels with reduced external Ca²⁺ (0.3 mM) and CaMKII inhibitor AIP. C, Reduced external Ca²⁺ and CaMKII inhibition with AIP returned SR Ca²⁺ content and Ca²⁺ transients to near control myocyte levels. D, Inhibition of the mitochondrial uniporter by Ru-360 or inhibition of the mitochondrial permeability transition pore by cyclosporine A decreased the rate of death in isolated cardiomyocytes from PLN-KO/CaMKII-TG mice at 12 hours. *P<0.05 vs. WT.

To more directly test the relationship between Ca²⁺ sparks and mitochondrial Ca²⁺ loading in KO/TG myocytes, external Ca²⁺ was reduced (to 0.3 mM) and CaMKII inhibited (with AIP). This restored SR Ca²⁺ content and Ca²⁺ transients (Fig. 8C), as well as Ca²⁺ spark frequency (not shown) in KO/TG myocytes to near WT levels. Strikingly, mitochondrial Ca²⁺ was not increased in KO/TG myocytes under these conditions (Fig. 8B).

The role of mitochondrial Ca²⁺ overloading in myocyte viability was examined by treating cells with an inhibitor of the mitochondrial uniporter (Ru-360) or the mitochondrial permeability transition (PT)-pore (cyclosporine A). Both agents significantly enhanced survival of KO/TG cardiomyocytes (Fig. 8D), implicating mitochondrial Ca²⁺ overloading and the PT-pore in the diminished viability of these vs. WT cells. On the other hand, neither agent increased the survival of CaMKII-TG myocytes (Fig. 8D), suggesting that mitochondrial Ca²⁺ overloading is less critical to the viability of the CaMKII-TG cells.

Discussion

PLN as a target for HF treatment

The SR Ca²⁺ content of cardiomyocytes is a key determinant of SR Ca²⁺ release and resultant Ca²⁺ transients.¹³ SR Ca²⁺ uptake is decreased in many animal models of HF and human failing heart, and this has been suggested to cause contractile dysfunction by decreasing the stored Ca²⁺ available for release. Ca²⁺ uptake into the SR is mediated by the SR Ca²⁺ ATPase (SERCA2a) and regulated by PLN, an endogenous inhibitor of SERCA2a whose activity is regulated by phosphorylation. Ablation or inhibition of PLN has been proposed as a viable therapeutic strategy for treating HF by restoring SR Ca²⁺ load;^14,15 this notion is supported by a number of reports, including the rescue of cardiomyopathy and HF in MLP KO⁶ and calsequestrin or β₁ transgenic mice;^7,8 and the suppression of HF progression in cardiomyopathic hamsters¹⁶ and in rats following myocardial infarction.¹⁷ In sharp contrast, in the present study as in several other genetic hypertrophy and HF models,^18,19 PLN ablation clearly restored myocyte Ca²⁺ stores, Ca²⁺ transients and contraction, but failed to rescue in vivo cardiac function or the HF phenotype. Thus, although depressed SR Ca²⁺ uptake is an important feature of HF (and a target to correct cellular Ca²⁺ dysregulation), restoration of SR Ca²⁺ handling does not appear to rescue all sequelae associated with HF development. The dissociation between improved cardiomyocyte SR Ca²⁺ handling and the in vivo loss of ventricular function and survival in the KO/TG mice provided an opportunity to explore, at the cellular level, why restoring SR Ca²⁺ content fails to improve, and indeed may worsen, global cardiac function.

Ca²⁺, mitochondria and cell death

SR Ca²⁺ sparks, elementary events of diastolic SR Ca²⁺ release, have been reported to be enhanced by both CaMKII overexpression and PLN ablation.^9,20 The CaMKII effect is primarily due to an activating effect on RyR2,¹² while PLN-KO enhances Ca²⁺ sparks indirectly by increasing SR Ca²⁺ content, further enhancing RyR2 opening due to a regulatory effect of intra-SR [Ca²⁺].^12,21 Here we demonstrate that SR Ca²⁺ sparks are massively increased in the KO/TG myocytes due to the combination of CaMKII mediated RyR2 phosphorylation and elevated SR Ca²⁺ load.

Numerous studies have demonstrated that mitochondrial Ca²⁺ overload results in opening of the permeability transition (PT)-pore and that this leads to either apoptotic or necrotic cell death in various cell types including cardiomyocytes.^22–27 Cardiomyocyte loss by apoptosis has been recognized as a major factor contributing to HF development.^28,29 Necrosis also contributes to ventricular remodeling induced by ischemic damage or in HF.^29–31 Enhancement of cardiomyocyte Ca²⁺ influx, as observed in transgenic mice expressing the sarcolemmal L-type Ca²⁺ channel, is one mechanism for eliciting Ca²⁺ overload, mitochondrial PT-pore opening and necrotic myocyte death.³² Mitochondria take up cytosolic Ca²⁺ via a Ca²⁺-uniporter and a close relationship between SR and mitochondria, in which Ca²⁺ released from the SR is efficiently transmitted to the mitochondria in cardiomyocytes, has been reported.³³ We suggest that the dramatic increase in Ca²⁺ sparks in the KO/TG myocyte leads, either directly or indirectly, to increased mitochondrial Ca²⁺ loading, PT-pore opening, and subsequent cardiomyocyte loss. This mechanism is supported by data demonstrating that mitochondrial Ca²⁺ loading and cell death are diminished following inhibition of Ca²⁺ sparks or mitochondrial Ca²⁺ uptake. Loss of functional cardiomyocytes could explain why the KO/TG hearts perform more poorly than the CaMKII-TG heart, despite improved cardiomyocyte Ca²⁺ transients.

Sympathetic tone, CaMKII and HF

In HF there is typically increased sympathetic tone and PKA activation,³⁴ as well as increased expression and activation of CaMKII,^34–37 and enhanced RyR2 phosphorylation and SR Ca²⁺ leak.^34,38 Under these conditions, where phosphorylated RyR2 may already enhance diastolic SR Ca²⁺ release, increased adrenergic signaling or therapeutic attempts to replete SR Ca²⁺ stores could exacerbate increased SR Ca²⁺ leak (as with PLN ablation). As suggested above the increased leak could trigger mitochondrial Ca²⁺ overloading with concomitant cardiomyocyte death.^27,32 This may also help to explain data demonstrating that CaMKII activation contributes to cardiomyocyte cell death induced by β-adrenergic stimulation and to isoproterenol-induced apoptosis or cardiomyopathy in vivo.^39–41

Increased SR Ca²⁺ leak induced by concomitant increases in SR Ca²⁺ content and RyR2 phosphorylation, could also predispose to arrhythmias, a possible basis for the observed increase in mortality in the KO/TG mice. While arrhythmias may not occur in all forms of HF in which SERCA function is restored,⁴² arrhythmias are seen in the CaMKII-TG.⁴³ The notion that arrhythmias would be enhanced by the increased Ca²⁺ leak^44,45 is further supported by our recent demonstration that isoproterenol treatment is highly arrhythmogenic in CaMKII-TG myocytes in vitro and in vivo.⁴⁶

In conclusion, enhanced SR Ca²⁺ uptake via PLN ablation can improve myocyte Ca²⁺ transients in transgenic mice that develop CaMKII-mediated heart failure. Unfortunately, the enhanced SR Ca²⁺ content exacerbates the already high level of diastolic Ca²⁺ spark activity, which may be more arrhythmogenic and further worsen overall heart function (via mitochondrial Ca²⁺ loading and cell death). The data are consistent with the working hypothesis that in the face of phosphorylated activated RyR2 channels, repletion of Ca²⁺ stores through PLN ablation (or during sympathetic activation) can exacerbate SR Ca²⁺ leak and thereby increase mitochondrial Ca²⁺ mediated cell death or activate other Ca²⁺ dependent processes that contribute to cardiac dysfunction.

Non-standard Abbreviations and Acronyms

AIP

autocamtide-2 related inhibitory peptide II

CaMKII

Ca²⁺/calmodulin-dependent protein kinase II

CaSpF

Ca²⁺ spark frequency

CsQ

calsequestrin

FS

factional shortening

HF

heart failure

HW/BW

heart/body weight ratio

KO

knockout

LVEDD

left ventricular end diastolic dimension

NCX

Na⁺/Ca²⁺ exchanger

PLN

phospholamban

PT-pore

permeability transition pore

RyR

ryanodine receptor

SERCA

sarcoplasmic reticulum Ca²⁺ ATPase

SR

sarcoplasmic reticulum

TG

transgenic

TUNEL

transferase-mediated biotinylated UTP nick end labeling

WGA

wheat germ agglutinin

WT

wild type