Increased InsP3Rs in the junctional sarcoplasmic reticulum augment Ca2+ transients and arrhythmias associated with cardiac hypertrophy

Cardiac hypertrophy is a growth response of the heart to increased hemodynamic demand or damage. Accompanying this heart enlargement is a remodeling of Ca2+ signaling. Due to its fundamental role in controlling cardiomyocyte contraction during every heartbeat, modifications in Ca2+ fluxes significantly impact on cardiac output and facilitate the development of arrhythmias. Using cardiomyocytes from spontaneously hypertensive rats (SHRs), we demonstrate that an increase in Ca2+ release through inositol 1,4,5-trisphosphate receptors (InsP3Rs) contributes to the larger excitation contraction coupling (ECC)-mediated Ca2+ transients characteristic of hypertrophic myocytes and underlies the more potent enhancement of ECC-mediated Ca2+ transients and contraction elicited by InsP3 or endothelin-1 (ET-1). Responsible for this is an increase in InsP3R expression in the junctional sarcoplasmic reticulum. Due to their close proximity to ryanodine receptors (RyRs) in this region, enhanced Ca2+ release through InsP3Rs served to sensitize RyRs, thereby increasing diastolic Ca2+ levels, the incidence of extra-systolic Ca2+ transients, and the induction of ECC-mediated Ca2+ elevations. Unlike the increase in InsP3R expression and Ca2+ transient amplitude in the cytosol, InsP3R expression and ECC-mediated Ca2+ transients in the nucleus were not altered during hypertrophy. Elevated InsP3R2 expression was also detected in hearts from human patients with heart failure after ischemic dilated cardiomyopathy, as well as in aortic-banded hypertrophic mouse hearts. Our data establish that increased InsP3R expression is a general mechanism that underlies remodeling of Ca2+ signaling during heart disease, and in particular, in triggering ventricular arrhythmia during hypertrophy.

I n response to increased hemodynamic requirements or damage the heart undergoes a hypertrophic growth response. Hypertrophy is induced by physiological stimuli, such as exercise or pregnancy and by pathological conditions such as hypertension and ischemic heart disease. Although hypertrophy can initially be an adaptive compensatory response, chronically it may become decompensated. As a result, cardiac function is decreased and the heart exhibits an increased propensity for arrhythmias that together ultimately lead to heart failure and death (1).
Ca 2ϩ is a fundamental regulator of cardiac function causing myocyte contraction via excitation-contraction coupling (ECC) (2), and stimulating the gene transcription that underlies hypertrophy (3). Accompanying cardiac hypertrophy and failure is a remodeling of Ca 2ϩ signaling (4). Whilst enhanced Ca 2ϩ transients facilitate greater myocyte contraction during adaptive hypertrophy, Ca 2ϩ fluxes are diminished during heart failure and thereby contribute to decreased cardiac output (5). Remodeling of the Ca 2ϩ signaling proteome also underlies the increased arrhythmias associated with hypertrophy and heart failure (6).
In addition to the RyRs that mediate ECC-dependent Ca 2ϩ fluxes, cardiomyocytes also express InsP 3 Rs, albeit outnumbered by RyRs at approximately 50:1 (7). Mammals have 3 InsP 3 R isoforms (types 1-3) (8), with InsP 3 R2 being the main isoform in cardiomyocytes (9,10). Although Ca 2ϩ flux via these InsP 3 Rs is relatively small in comparison to the large Ca 2ϩ transients occurring during every heartbeat, recent data suggests that InsP 3 Rs have an important role in cardiac physiology. We, and others have shown, that Ca 2ϩ release through InsP 3 Rs contributes to the inotropic, arrhythmogenic, and hypertrophic effect of G␣ q -coupled agonists such as the vasoactive peptide ET-1 (11)(12)(13)(14)(15)(16). Whether altered InsP 3 R signaling also contributes to remodeling of Ca 2ϩ homeostasis during cardiac hypertrophy is not yet determined. An increase in InsP 3 R expression has however been reported during heart failure in humans (17). Moreover, InsP 3 -induced Ca 2ϩ release (IICR) is increased in SR microsomes prepared from hypertrophic myocytes (18).
Here, we hypothesized that enhanced Ca 2ϩ release via InsP 3 Rs contributes to remodeling of ECC-mediated Ca 2ϩ transients, and to the increased arrhythmogenic Ca 2ϩ signals observed in ventricular cardiomyocytes during compensated hypertrophy. To test these hypotheses, in a model that reflects the slow development of hypertrophy in humans, Ca 2ϩ fluxes and contractility were investigated in hypertrophic ventricular myocytes isolated form SHRs (19). We found that the increase in amplitude of ECC-mediated Ca 2ϩ transients and propensity for extra-systolic spontaneous Ca 2ϩ signals, characteristic of hypertrophic myocytes, was caused by augmented InsP 3 R signaling. This profound effect of enhanced InsP 3 R activity in hypertrophic myocytes was due to an increase in InsP 3 R expression, specifically in the junctional SR membrane in close proximity to RyRs. At this location, Ca 2ϩ release via InsP 3 Rs acted to sensitize RyRs, thereby enhancing Ca 2ϩ release during ECC and inducing spontaneous elementary Ca 2ϩ -release events and extra-systolic Ca 2ϩ transients. InsP 3 R2 expression was also increased in hypertrophic cardiomyocytes isolated from aortically banded mice and in human hearts displaying ischemic dilated cardiomyopathy. We propose that InsP 3 Rs play a fundamental role in the physiology of hypertrophic hearts contributing to remodeled cardiac function and triggering ventricular arrhythmia.

Results
SHR Cardiomyocytes Develop Hypertrophy. As previously described, at 6 months, cardiomyocytes from SHRs are hypertrophic (20). Cardiomyocyte width was increased in SHRs compared to WKY The authors declare no conflict of interest. 1 M.D.B. and H.L.R. contributed equally to this work. 2 To whom correspondence should be addressed. E-mail: llewelyn.roderick@bbsrc.ac.uk.
To reveal whether increased Ca 2ϩ release via InsP 3 Rs contributes to remodeling of Ca 2ϩ signaling and myocyte function during hypertrophy, we measured global Ca 2ϩ transients and cellular contraction under conditions where InsP 3 Rs were either inhibited or activated. As Ca 2ϩ transients are greater under basal conditions in hypertrophic SHR than in WKY myocytes, the systolic Ca 2ϩ amplitude was normalized to that before treatment. Inhibition of InsP 3 Rs with 2-APB (2 M) (11) caused a greater reduction in ECC-mediated Ca 2ϩ transient amplitude in SHR myocytes compared with WKY controls (Fig. 1A). Concurrently, 2-APB also decreased the magnitude of contraction in SHRs (Fig. 1 Aii). These data suggested that Ca 2ϩ release via InsP 3 Rs contributes to the greater basal ECC-associated Ca 2ϩ transients observed in SHR myocytes.
Direct activation of InsP 3 Rs with InsP 3 ester (11) promoted a greater increase in Ca 2ϩ transient amplitude and inotropy in SHR compared to WKY myocytes ( Fig. 1 B and C), which was abrogated in both strains by 2-APB ( Fig. 1 Bii and Cii). No difference in ECC-mediated Ca 2ϩ transient amplitude or cellular contraction was observed between myocytes isolated from 12-week-old WKY rats or SHRs (Fig. S1 A).

InsP3R2 Expression Is Increased in Hypertrophic Cardiomyocytes.
Next, we analyzed whether an increase in InsP 3 R expression underlies altered InsP 3 signaling during hypertrophy. At 6 months, InsP 3 R2 mRNA and protein levels were higher in SHR than in WKY myocytes, whereas at 12 weeks, InsP 3 R2 mRNA and protein levels were lower in SHRs compared to WKY controls ( Fig. 2 A and  B). RyR2 protein levels were not different between the 2 strains at the age of 12 weeks or 6 months (Fig. 2B).
Immunofluorescent labeling revealed that in WKY myocytes, InsP 3 R2 was predominantly expressed in the perinuclear regions with weaker staining along the SR membrane, where RyRs are localized (Fig. 2C). InsP 3 R2 was also expressed in the perinuclear regions of SHR cardiomyocytes, but compared to WKY cells its expression was significantly greater along the RyR2-stained striations outside the nuclear region. Thus, the ratio of cytosolic/nuclear InsP 3 R2 immunofluorescence was increased (Fig. 2D). No difference in RyR2 immunostaining between the 2 strains was observed (Fig. 2C). In both WKY and SHR myocytes, InsP 3 R2 co-localized with RyR2s (intensity profile along indicated line, Fig. 2 C and F) and Pearson's coefficient (Fig. 2E), indicating that like RyR2s, InsP 3 R2s are located at dyadic junctions alongside T-tubule membranes (Fig. 2C). The co-localization of these 2 channels was markedly increased in hypertrophic myocytes ( Fig. 2 E and F). We concluded that, as a result of hypertrophic remodeling, the number of InsP 3 R2s located in the junctional SR membrane is increased, thereby mediating their greater co-localization with RyR2s. InsP 3 R2 expression was also increased in hearts from mice after aortic banding (Fig. 2G) and in human patients with ischemic dilated cardiomyopathy (Fig. 2H). These data suggested that increased InsP 3 R2 expression is a general hallmark of hypertrophy.

Increased InsP3R Expression in the Junctional SR Causes a Spatially
Restricted Remodeling of ECC-mediated Ca 2؉ Transients during Hypertrophy. To establish how increased InsP 3 R expression in the junctional SR membrane impacted on ECC-mediated Ca 2ϩ signals, we performed confocal Ca 2ϩ imaging experiments. In addition to directly stimulating InsP 3 Rs with a membrane-permeant InsP 3 ester, we tested the effect of physiologically activating InsP 3 Rs with InsP 3 generated following ET-1 stimulation (11,22). In SHR myocytes, stimulation with ET-1 and InsP 3 ester increased the amplitude of nuclear and cytosolic Ca 2ϩ transients during electrical pacing ( Fig. 3 A-E). Contrastingly, in WKY myocytes, only nuclear systolic Ca 2ϩ transients were augmented ( Fig. 3 C and E). The enhancement of systolic Ca 2ϩ transients by ET-1 and InsP 3 ester was sensitive to 2-APB, further indicating that this effect was mediated by InsP 3 Rs (Fig. 3 Bii-Eii). The increase in nuclear Ca 2ϩ transient amplitude in SHRs was comparable to that observed in WKY myocytes (Fig. 3 C and E). To accommodate for variation between cells in the absolute magnitude of Ca 2ϩ changes, the ratio of nuclear to cytosolic Ca 2ϩ transient amplitude was calculated. This ratio was increased in WKY myocytes following ET-1 or InsP 3 ester stimulation whereas no change was observed in SHR myocytes ( Fig. S2 A). The difference in ratio between the 2 strains is explained by restriction of the ET-1-and InsP 3 ester-stimulated increase in Ca 2ϩ transient amplitude to the nuclear compartment in WKY myocytes, whereas, in SHR myocytes nuclear and cytosolic Ca 2ϩ transient amplitude were both increased. These data indicate that in non-hypertrophied myocytes, Ca 2ϩ release via InsP 3 Rs impacts more profoundly on nuclear Ca 2ϩ transients, whereas in hypertrophic myocytes, increased junctional InsP 3 R expression specifically augments the cytosolic Ca 2ϩ transients. There was no difference in the ratio of nuclear to cytoplasmic Ca 2ϩ transients between 12-week-old WKY and SHRs under basal conditions, or during ET-1 or InsP 3 ester stimulation (Fig. S1B).
Maximal Ca 2ϩ release from nuclear and cytosolic Ca 2ϩ stores induced by 10 mM caffeine (RyR agonist) was not significantly different between WKY and SHR myocytes (Fig. S2B), indicating that differences in Ca 2ϩ store content do not underlie the changes in ECC-associated Ca 2ϩ transients during ET-1 and InsP 3 ester stimulation.
Extra-Systolic InsP 3 -dependent Ca 2؉ -release Events Are Increased in SHR Myocytes. In atrial cardiomyocytes, which express approximately 6 fold more InsP 3 Rs than ventricular myocytes, Ca 2ϩ release via InsP 3 Rs underlies the induction of extra-systolic Ca 2ϩ transients (11,12). We therefore investigated whether the increased InsP 3 R expression and activity observed in SHR ventricular myocytes caused them to exhibit more extra-systolic Ca 2ϩ -release events. Extra-systolic events were determined as rises in Ca 2ϩ concentration that were temporally distinguished from signals induced by field stimulation and that also impacted on contraction (see arrows  Table S2). However, the number of cells that exhibited extra-systolic Ca 2ϩ transients and the frequency of events per cell were greater in SHRs than WKYs (ET-1: WKY: 26% vs. SHR: 50%; InsP 3 ester: WKY: 29% vs. SHR: 57%, Table S2 and Fig. 4B). In both strains, extra-systolic Ca 2ϩrelease events began to occur within a few minutes of InsP 3 ester or ET-1 stimulation and increased throughout the time-course of the experiment (Fig. 4B). The rate at which the frequency of the extra-systolic Ca 2ϩ transients increased following InsP 3 ester or ET-1 stimulation was greater for SHR myocytes than WKY myocytes. After 1,000 s, the incidence of extra-systolic Ca 2ϩ transients was significantly higher in SHRs than WKY cells (Fig. S3). No difference in the frequency of extra-systolic Ca 2ϩ transients was observed between the 2 strains at 12 weeks (Fig. S1 C and D). These data indicate that activation of InsP 3 Rs was responsible for the initiation of the extra-systolic Ca 2ϩ transients and provides an explanation for the increased frequency of extra-systolic Ca 2ϩ transients during hypertrophy.

Enhanced Ca 2؉ Release via InsP3Rs Increases the Rate of Rise of Systolic Ca 2؉ Transients and Elevates Diastolic [Ca 2؉ ] in SHRs.
A hypertrophy-associated increase in InsP 3 R-mediated Ca 2ϩ flux via junctional InsP 3 Rs acting to induce Ca 2ϩ release via neighboring RyRs could provide a mechanism to accentuate Ca 2ϩ signaling during ECC. To test this hypothesis, the effect of ET-1 and InsP 3 ester on the rate of rise of pacing-evoked systolic Ca 2ϩ transients was measured. During both ET-1 and InsP 3 ester stimulation, the rate of rise of the Ca 2ϩ transient was faster in hypertrophic SHR than in WKY cells (Fig. 5A). The effects of ET-1 and InsP 3 ester were abrogated by adenoviral-mediated expression of a cherry fluorescent protein-tagged InsP 3 5Ј-phosphatase, which disrupts InsP 3 signaling (5ЈP; Fig. 5B) (16). There was no difference in the rate of rise of the systolic Ca 2ϩ transient in myocytes from 12-weekold WKY and SHRs (Fig. S1E).
As  Fig. 5C), whereas no change was seen in WKY cells (Fig. 5C). 2-APB or 5ЈP expression abrogated the increase in diastolic [Ca 2ϩ ] caused by InsP 3 ester or ET-1 in SHRs (Fig. 5 C and D) without effecting diastolic [Ca 2ϩ ] under basal conditions. At 12 weeks of age, there was no difference in diastolic [Ca 2ϩ ] during stimulation of SHR myocytes with ET-1 or InsP 3 ester (Fig. S1F).

Frequency of Elementary InsP3-dependent Ca 2؉ -release Events Is
Increased during Hypertrophy. To further resolve the consequences of increased InsP 3 R expression for Ca 2ϩ signaling, elementary Ca 2ϩ -release events were analyzed. Under normal paced conditions, Ca 2ϩ events during the diastolic period were of greater amplitude in the hypertrophic SHR myocytes than in WKY cells (WKY: ⌬F/F 0 ϭ 0.26 Ϯ 0.01 vs. SHR: 0.34 Ϯ 0.04, Table S3 and Fig.  5E). Under conditions where RyRs were blocked with 1 mM tetracaine, InsP 3 ester application stimulated elementary Ca 2ϩrelease events (Fig. 5E) that occurred at a greater frequency in hypertrophic myocytes (WKY: 2.12 Ϯ 0.46 vs. SHR: 6.84 Ϯ 0.65, Table S3). These data suggest that InsP 3 R-mediated Ca 2ϩ signals contribute to the greater amplitude of diastolic Ca 2ϩ events observed in SHR myocytes and may underlie the elevated diastolic [Ca 2ϩ ] observed in SHR myocytes stimulated with InsP 3 or ET-1.
Our data suggest a model to explain the enhanced ECCmediated Ca 2ϩ signals and increased extra-systolic Ca 2ϩ -release events observed during hypertrophy (Fig. 5F). Key to this model is a hypertrophy-associated increase in InsP 3 R expression in the dyadic region. Thus, more InsP 3 Rs are in close proximity to RyRs in the SR membrane (Fig. 2 E and F). Ca 2ϩ released via these InsP 3 Rs sensitizes their adjacent RyRs, bringing them closer to threshold for activation. Under conditions of increased [InsP 3 ], elementary InsP 3 -dependent Ca 2ϩ -release events are increased in frequency and diastolic [Ca 2ϩ ] is elevated. Consequently, RyRs are triggered to generate extra-systolic Ca 2ϩ signals and to accelerate the rate of rise of pacing-evoked Ca 2ϩ transients (Fig. 5F).

Discussion
Here we demonstrate that enhanced Ca 2ϩ signaling via InsP 3 Rs located in the dyadic cleft remodels Ca 2ϩ signaling during hypertrophy.
In agreement with previous data, we found that the amplitude of ECC-mediated Ca 2ϩ transients under basal conditions was significantly greater as a result of hypertrophy in SHR myocytes (in the absence of any other stimulation) (20). Significantly, we determined that this increased amplitude of basal ECC-mediated Ca 2ϩ transients was due to augmented Ca 2ϩ release via InsP 3 Rs. These data demonstrated that InsP 3 Rs could contribute to ECC-mediated Ca 2ϩ fluxes without additional neurohormonal input, thereby modifying myocyte Ca 2ϩ signaling.
A greater role for InsP 3 Rs in regulating ECC-mediated Ca 2ϩ transients during hypertrophy was revealed following their direct activation with cell-permeant InsP 3 ester. These data showed that increased activation of InsP 3 Rs could augment the amplitude of ECC-mediated Ca 2ϩ transients mediated via RyRs even further. Consistent with previous reports (20,23,24), no increase in SR releasable Ca 2ϩ was observed in hypertrophic SHR myocytes,  thereby indicating that the enhancement of ECC-associated Ca 2ϩ flux by InsP 3 Rs was not due to an increase in store loading. InsP 3 R2 expression was elevated as a result of hypertrophy thereby providing a mechanism for increased Ca 2ϩ release via InsP 3 Rs. InsP 3 R expression in the heart has previously been reported to be modified following disease. In particular, InsP 3 R expression is increased in atrial myocytes of humans and dogs during atrial fibrillation (AF) (25,26). Furthermore, elevated InsP 3 R levels and increased InsP 3 binding was reported in the left ventricle during human heart failure (17). Consistent with these reports and our observations in rats, we found that InsP 3 R2 expression was significantly elevated in cardiac tissue from aortically-banded hypertrophic mice and from human hearts showing ischemic dilated cardiomyopathy. Due to its very low expression and insensitivity to hypertrophy in rat cardiac fibroblasts (Fig. S4), we considered that the changes in InsP 3 R2 expression detected in human and mouse cardiac tissue was due solely to InsP 3 R2 in cardiac myocytes. Our findings in rats, mice, and humans therefore suggested that increased InsP 3 R expression is a general feature of cardiac disease, raising the possibility that increased Ca 2ϩ release via InsP 3 Rs contributes to pathological changes in Ca 2ϩ signaling.
The enhanced InsP 3 R2 expression had a striking spatial aspect in that InsP 3 R expression was specifically increased in the junctional SR. Detailed analysis showed that these junctional InsP 3 Rs colocalized with RyRs, which reside primarily in the dyadic cleft. This profound remodeling in InsP 3 Rs expression and distribution had significant functional consequences. In particular, the increased number of dyadic InsP 3 Rs augmented the amplitude of the cytosolic ECC-mediated Ca 2ϩ transients and enhanced the positive inotropic effect of InsP 3 ester. Similarly, cytosolic ECC-mediated Ca 2ϩ transient amplitude and contraction were enhanced when InsP 3 Rs were engaged by InsP 3 generated following application of ET-1. Given that ET-1 is a potent pro-hypertrophic agonist, and its levels are elevated during heart failure, these findings have significant implications for cardiac function during hypertrophy (16,22,27). The activation of InsP 3 Rs in SHR myocytes by ET-1 is in agreement with data from our laboratory and elsewhere showing that stimulation of the InsP 3 signaling cascade in cardiomyocytes with ET-1 modifies Ca 2ϩ fluxes and contractility (11,13,28). The increase in nuclear Ca 2ϩ transient amplitude during ECC by ET-1 and InsP 3 ester was not altered during hypertrophy reflecting the lack of a change in InsP 3 R expression in this region. Together, these data suggested that Ca 2ϩ release via InsP 3 Rs in the dyadic region primed ECC-mediated Ca 2ϩ -induced Ca 2ϩ release via RyRs (see Fig. 5F). Specifically, Ca 2ϩ release via InsP 3 Rs could elevate diastolic [Ca 2ϩ ] closer to the threshold for activation of RyRs. Thus, we established that increased Ca 2ϩ release via InsP 3 Rs in hypertrophic myocytes can significantly contribute to remodel ECCmediated Ca 2ϩ signals.
At the most fundamental level, in the absence of RyR activity, SHR myocytes exhibited an increased frequency of elementary InsP 3 -dependent Ca 2ϩ -release events. Interestingly, the amplitudes of those events were no different between WKY and SHR myocytes. This is not surprising given that Ca 2ϩ puffs are fundamental Ca 2ϩ signals that are conserved between cells as diverse as Xenopus oocytes and HeLa human epithelial cells (29). At the molecular level, Ca 2ϩ puffs arise via the stochastic recruitment of neighboring InsP 3 Rs (a cluster) until a threshold number required for puff generation is reached (30). Thus, it is plausible that greater InsP 3 R expression in SHR myocytes simply increases the probability of recruiting this puff-generating threshold number of receptors without altering the properties of puffs. As a result, only the frequency of elementary events is increased in SHRs. The greater abundance of these elementary events may explain the elevated diastolic [Ca 2ϩ ] observed in SHR myocytes stimulated with InsP 3 ester and ET-1. These data are consistent with the requirement for InsP 3 Rs for the ET-1-stimulated increase in diastolic [Ca 2ϩ ] observed in atrial myocytes (which have Ϸ6-fold greater InsP 3 R expression than ventricular myocytes) (10,12). As elementary Ca 2ϩ -release events (Ca 2ϩ sparks and puffs) are the building blocks of higher order Ca 2ϩ transients, it was not surprising that SHR myocytes also exhibited an increased frequency of extra-systolic Ca 2ϩ transients. Similarly, stimulation of atrial myocytes with InsP 3 or InsP 3 -generating ago- nists such as ET-1, potently induced arrhythmogenic Ca 2ϩ -release events that were dependent on InsP 3 R2 expression (11)(12)(13)28).
By bringing Ca 2ϩ levels closer to the threshold for activation of RyRs, InsP 3 -mediated sensitization of RyRs also served to increase the rate of rise of ECC-mediated Ca 2ϩ transients. This may remediate the deterioration in Ca 2ϩ signaling that occurs as hypertrophy progresses to failure. In particular, extra Ca 2ϩ release via dyadic InsP 3 Rs may compensate for the decreased coupling efficiency between L-type Ca 2ϩ channels and RyRs due to a deterioration in the T-tubular network and increased width of the dyadic cleft that occurs during disease (31).
The arrhythmogenic effect of InsP 3 R activity in the ventricles may have profound consequences. Coupled with increased systemic levels of InsP 3 -generating agonists, such as ET-1 during hypertension and heart failure, it provides a possible mechanistic explanation why hypertrophic hearts are more likely to develop potentially lethal ventricular arrhythmias (32).
As InsP 3 R2 is increased during cardiac hypertrophy, yet is dispensable for the normal physiological function of the healthy heart (12), it may represent an ideal target to which pharmacological modulators could be developed to intervene in both the induction of the hypertrophic gene program and the generation of arrhythmias.

Materials and Methods
Detailed methods for myocyte isolation, adenoviral infection, photometric, and confocal measurements of [Ca 2ϩ ]i, immunoblotting, immunofluorescence, quantitative RT-PCR, and cell length measurements are provided elsewhere (14,33) and in SI Methods and Fig. S5.
Animal Models. Male SHRs and normotensive Wistar-Kyoto (WKY) rats were obtained from Harlan and were housed under control conditions with ad libitum food and water. All experiments were performed in accordance with the guidelines from the code of practice for humane killing under Schedule 1 of the Animals (Scientific Procedures) Act 1986. Constriction of the transverse thoracic aorta was performed on 3-month-old male mice as described in SI Methods. The sham procedure was identical but without aortic ligation.
Patients. Left ventricular tissue samples of human failing hearts were from individuals undergoing heart transplantation due to end-stage heart failure. All samples were obtained from male caucasians, aged 41-62. Samples from nonfailing donor hearts were provided by the U.K. Human Tissue Bank. After cardiectomy, left ventricular samples were frozen in liquid nitrogen and stored at Ϫ80°C. Detailed information about the patients can be found in SI Methods. All experiments involving human tissue samples have been approved by the Cambridgeshire Research Ethics Committee.

Recordings of Myocyte Contraction and [Ca 2؉ ]i.
All experiments, unless otherwise stated, were performed at 22°C on myocytes electrically paced with field electrodes at 0.33 Hz. This condition is referred to as the basal condition. Detailed procedures can be found in SI Methods.
Statistics. Data are expressed as mean Ϯ SEM. Statistical comparisons were carried out with Student's t test or 2-way ANOVA. Statistically significance was accepted at P Ͻ 0.05.