Dilated Cardiomyopathy with Increased SR Ca2+ Loading Preceded by a Hypercontractile State and Diastolic Failure in the α1CTG Mouse

Mice over-expressing the α1−subunit (pore) of the L-type Ca2+ channel (α1CTG) by 4months (mo) of age exhibit an enlarged heart, hypertrophied myocytes, increased Ca2+ current and Ca2+ transient amplitude, but a normal SR Ca2+ load. With advancing age (8–11 mo), some mice demonstrate advanced hypertrophy but are not in congestive heart failure (NFTG), while others evolve to frank dilated congestive heart failure (FTG). We demonstrate that older NFTG myocytes exhibit a hypercontractile state over a wide range of stimulation frequencies, but maintain a normal SR Ca2+ load compared to age matched non-transgenic (NTG) myocytes. However, at high stimulation rates (2–4 Hz) signs of diastolic contractile failure appear in NFTG cells. The evolution of frank congestive failure in FTG is accompanied by a further increase in heart mass and myocyte size, and phospholamban and ryanodine receptor protein levels and phosphorylation become reduced. In FTG, the SR Ca2+ load increases and Ca2+ release following excitation, increases further. An enhanced NCX function in FTG, as reflected by an accelerated relaxation of the caffeine-induced Ca2+ transient, is insufficient to maintain a normal diastolic Ca2+ during high rates of stimulation. Although a high SR Ca2+ release following excitation is maintained, the hypercontractile state is not maintained at high rates of stimulation, and signs of both systolic and diastolic contractile failure appear. Thus, the dilated cardiomyopathy that evolves in this mouse model exhibits signs of both systolic and diastolic failure, but not a deficient SR Ca2+ loading or release, as occurs in some other cardiomyopathic models.


Introduction
Much evidence has accumulated since the first cytosolic Ca 2+ transient measurements in cardiac muscle from heart failure (HF) patients [1] to support a role for alterations in myocyte ''Ca 2+ handling'' in the pathophysiology of HF. Although multiple alterations in various aspects of myocyte excitation-contraction coupling have been observed in HF, the central mechanism of the reduced Ca 2+ transient amplitude has been emphasized and has been attributed to a reduction of sarcoplasmic reticulum (SR) Ca 2+ content [1][2][3][4][5][6][7].
Longitudinal assessment of cardiac structure and function in a novel transgenic mouse that over expresses the a 1C -subunit of the L-type Ca 2+ channel (a 1C TG) indicates that like droves of other ''boutique mice'' (Table S1), this mouse, develops cardiac hypertrophy, during which adaptive Ca 2+ regulatory mechanisms are mobilized. The remarkable orchestration among cardiomyocyte Ca 2+ regulatory proteins in this model at 4months (mo) is instructive because it provides clues with respect to coordinated, adaptive remodeling of Ca 2+ regulation to maintain a normal SR Ca 2+ load at 4 mo of age [4]. Specifically, during the hypertrophic, pre-HF stage, while an L-type current of a larger amplitude triggers larger Ca 2+ release from the SR to produce a whole cell Ca 2+ transient of increased amplitude, neither the SR Ca 2+ load, as assessed by caffeine induced Ca 2+ release, nor diastolic cytosolic Ca 2+ levels, nor Ca 2+ spark characteristics, are altered [4]. An overexpression of NCX protein, which enhances Ca 2+ efflux to balance the enhanced Ca 2+ influx via the overexpressed L-type Ca 2+ channel prevents excess cytosolic calcium loading [4].
Between 8-11 mo of age, however, the adapted, hypertrophic heart of the a 1C TG mouse maladapts into a dilated, lethal cardiomyopathy [4]. The hypothesis of the present study is that adaptations in Ca 2+ regulation observed at a younger age (4 mo) in NTG cells wane with the evolution of more advanced hypertrophy and HF that accompany advancing age. Specifically, we hypothesized that these adaptations become compromised in the non-failing advanced hypertrophic stage (NFTG), and ''fail'' outright as the end-stage dilated congestive failure (FTG) evolves.

Heart and Myocyte Size
Heart weight in FTG was over two-fold greater than in NTG ( Figure 1A). Body weight did not differ between NTG and NFTG but body weight increased by 15% in FTG, likely as a result of fluid accumulation ( Figure 1B). The heart weight/body weight increased by 43% in NFTG vs NTG, by 27% in FTG vs. NFTG, and by 82% in FTG vs NTG ( Figure 1C). The average myocyte size, estimated from cell capacitance, reflected the relative heart mass. It increased by 25% in NFTG vs NTG, and in FTG further increased by 29%. Thus, myocyte size in failing FTG averaged 61% more than that in NTG ( Figure 1D).

Ca 2+ Current and Electrical Stimulation-Induced Ca 2+ Transients
Peak I Ca density in NFTG and FTG was increased compared to NTG, despite a significant increase in cell capacitance (293.5615.02 pF, 330.93613.53 pF vs 254.6167.76 pF, respectively). No statistically significant difference was observed between NFTG and FTG ( Figure 2A). Inactivation rate constants of I CaL are illustrated in Figure 2B. Tau fast did not differ among groups but Tau slow was prolonged in both TG groups compared to NTG. In rodent myocytes, the rested state Ca 2+ transient in Figure 3 is usually the maximal that can be achieved, due to a maximum SR Ca 2+ load and release achieved during rest. Upon continued stimulation from rest a typical negative staircase in Ca 2+ transient typically occurs. This reflects, in part, a net reduction in cell and SR Ca 2+ loading.
The amplitude of the Indo-1 fluorescence transient (peak-rest or peak-diastolic level) elicited by electrical stimulation declined sharply between the first stimulation following rest and steady state stimulation at 0.5 Hz. There was little further decline thereafter as the stimulation rate was increased ( Figure 3B). Following rest, and at all frequencies of stimulation, the Indo-1 transient amplitude in FTG was greater than in NTG. In NFTG group, data were available only at rest and 0.5 Hz. The Indo 1 fluorescent transient at these two frequencies was intermediate between that of FTG and NTG.
The maximum rate of rise (V MAX ) of the Indo fluorescent transient ( Figure 3C) reports the amplitude of the SR Ca 2+ release flux with fidelity [1,8,9]. Differences among the three groups in V MAX upon stimulation from rest were similar to those of the amplitude of the fluorescence transient. This suggests that the decline in the Ca 2+ transient amplitude from the post-rest state reflects a decrease in excitation induced SR Ca 2+ release. A greater Ca 2+ transient V MAX across the range of stimulation rates in FTG indicates that the magnitude of excitation-induced SR release in FTG is greater than in NTG.
Continued stimulation from rest accelerated the decay of the Ca 2+ transient ( Figure 3D). The T 50 of relaxation was prolonged in FTG vs. NTG across the range of stimulation; the difference is more marked in the resting state beat than that of 0.5 Hz. The T 50 measured over the limited range of frequencies in NFTG did not differ from NTG or FTG.
At rest there was no difference in the resting Indo-1 fluorescence ratio among the three groups of cells ( Figure 3E). Upon regular electrical stimulation from rest the diastolic Indo fluorescence at rates up to 4 Hz in NTG did not exceed that at rest. In contrast, in FTG the diastolic fluorescence increases as the stimulation rate increases, exceeding that at rest by 10% at 4 Hz. Note that the difference in the resting fluorescence ratio prior to stimulation and diastolic fluorescence ratio at 4 Hz was about 0.10 ( Figure 3E). The amplitude of fluorescence transient elicited by excitation at 4 Hz was about 0.22 ( Figure 3B). Thus, even if only 50% of the change from the resting Indo fluorescence reports a true change in diastolic cytosolic Ca 2+ , the increase in diastolic Ca 2+ in FTG with stimulation is substantial. During excitation at 4 Hz, this increase averaged 20-25% of the Ca 2+ transient amplitude.

Caffeine-Induced Ca 2+ Transients
We assessed the SR Ca 2+ load by a rapid, brief application of caffeine to the cell. In a prior study, no difference was noted, in the amplitude of caffeine-induced Indo fluorescent transient between NTG and a 1C TG at 4 mo and 9 mo. Likewise, in the present study, there was no difference in the amplitude of the caffeineinduced transient between NFTG and NTG ( Figure 4). But the caffeine-induced transient in FTG was increased compared to NFTG or NTG groups ( Figure 4). Note that the amplitude of the caffeine transient at 4 Hz did not vary from that at rest, in contrast to the amplitude of the transient induced by electrical stimulation, which is much larger at rest than that at 4 Hz ( Figure 4A). Thus, by rough calculations, the SR ''fractional release'' in response to electrical stimulation, i.e., the ratio of the electrical stimulation and caffeine Ca 2+ transients' amplitudes was reduced at 4 Hz compared to rest. The extent of this reduction was comparable in NTG and FTG: at rest it averaged 0.85 and 0.83 respectively, and at 4 Hz was 0.39 and 0.47, respectively. Figure 4A, B also shows that the decay rate of the caffeine transient is accelerated in FTG. This is more clearly illustrated by normalization of the data during the decay to the peak amplitude in each cell ( Figure 5A and B). On an average, the T 50 of relaxation of the caffeine-induced Indo transient was accelerated by 50% in FTG vs. NTG ( Figure 5C). In NFTG, the T 50 did not differ from that of NTG following rest. During continuous pacing, however, the T 50 in NFTG became less than NTG. The decay of the caffeine transient following 4 Hz stimulation was slower than at rest in both NTG and FTG. But FTG T 50 still remained substantially accelerated compared to NTG. NFTG showed no increase or difference from FTG ( Figure 5C). The ability of NFTG to maintain the same T 50 of the caffeine transient following 4 Hz as at rest suggests that NCX reserve may be greater in NFTG than in FTG or NTG cells.
A comparison of Figure 5 and Figure 3D reveals that the group differences between FTG and NTG in the T 50 of the caffeine induced transient decay and the T 50 of electrical stimulation induced transient are reversed: the electrical stimulation induced transient T 50 , which largely reflects SR Ca 2+ pumping, was more rapid in NTG than FTG. The caffeine induced transient T 50 , reflecting mostly NCX Ca 2+ extrusion, was more rapid in FTG than NTG. However, a relative failure of SERCA2a and NCX to dissipate cytosolic Ca 2+ at high rates of electrical stimulation in FTG likely explains the failure to regulate the diastolic Indo fluorescence ( Figure 3E).

Contraction Characteristics
Changes in Ca 2+ release into and removal from the cytosol, as observed in FTG, may result in contractile impairment. We first determined the extent to which Indo-1 loading affects myocyte contractile properties in both FTG and NTG cells. The contractile characteristics, i.e., twitch amplitude, shortening velocity, and relaxation time were severely distorted by Indo-1 in both groups of cells; the buffering effect on twitch amplitude was greater in FTG than in NTG myocytes (Table S2). To assess artifact-free contraction characteristics, we performed additional studies in non-Indo-1 loaded cells to analyze Ca 2+ buffering.
In response to the initial excitation following rest the absolute contraction amplitude was greater in the NFTG than in FTG or NTG ( Figure 6A-C). As stimulation from rest continues, the normal response is for contraction amplitude to decrease, as it does in NTG. During stimulation at low rates (0.5-1 Hz), the relative decline in NTG was greater than that in FTG or NFTG: the absolute contraction amplitude in NFTG and FTG exceeds that in NTG by 50%. This is consistent with a hypercontractile stage in the evolution of cardiomyopathy in other models, including humans [9]. Greater contraction amplitude in TG cells suggests a relative inability to appropriately reduce the cell Ca 2+ load in this context [1,10,11]. Contraction amplitude remained higher in NFTG cells than in NTG cells during steady state stimulation at 2-4 Hz, but in FTG it dropped to that of NTG. Note that failure of FTG to maintain augmented contraction amplitude vs. NTG, or one equivalent to NFTG at the higher stimulation rates, occurred in the context of a larger SR Ca 2+ release flux and Ca 2+ transient amplitude in FTG vs. NTG (Figure 4).
The absolute T 50 of cell relengthening (relaxation), like that of the Indo fluorescence transient (Figure 3), was prolonged in FTG in the initial excitation following rest. But T 50 of relaxation converged toward other groups with increasing stimulation rates ( Figure 6D, E). The absolute cell length at rest remained constant among the 3 groups ( Figure 6F). Both transgenic groups showed an inability to maintain a normal diastolic length as stimulation frequency increased ( Figure 6G). In FTG, this could result from an inability to maintain a normal diastolic Ca 2+ , reflected by failure to maintain a normal diastolic Indo-1 fluorescence ratio ( Figure 3E). Thus, in myocytes from the NFTG hearts, electrical stimulation at high rates uncovered a diastolic contractile failure: in the myocytes from the grossly dilated hearts of congested FTG mice, high rates  of stimulation provoked both systolic contractile failure and severe diastolic contractile failure. A ''relative contractile, i.e., failure'' also occurred within FTG at high stimulation rates, as the reduction in contraction amplitude was not accompanied by a reduction in Ca 2+ release following excitation ( Figure 4B vs. 6B). This could, in part, be due to a failure to maintain diastolic Ca 2+ levels, but cannot be attributed to a failure to release Ca 2+ into the cytosol following excitation.

Expression of Ca 2+ cycling proteins
Quantitative Western Blot techniques comparing NFTG and NTG (8-11 mo) showed: a 30% decrease in SERCA2a levels; 50% increase in NCX; 120% increase in PLN p (PLN pentamer) and 73% and 54% decrease in its phosphorylation levels of PS16 and PT17, respectively normalized to total PLN p ; by at 48% decrease in RyR2 levels but 115% hyperphosphorylation of RyRs at Ser2089 (Figure 7).

Discussion
There are several novel results of the present study. In NFTG at 8-11 mo of age, heart mass and average myocyte sizes are increased by about 25% relative to age-matched NTG as is the case in 4 mo a 1C TG vs. NTG cells [4]. During the evolution from NFTG to the FTG state, both heart mass and myocyte size are increased markedly. At low rates of stimulation several other features of 4 mo a 1C TG myocytes, [4] including an increased [Ca 2+ ] i transient elicited by electrical stimulation and a normal SR Ca 2+ load, are retained in 9 mo NFTG. Cardiomyocytes from both the NFTG and FTG myocytes at 8-11 mo showed a marked increase in I Ca,L amplitude compared to NTG. In FTG, in contrast to NFTG, both, the Ca 2+ transient amplitude following excitation and the SR Ca 2+ load are increased. While systolic Ca 2+ release in FTG is higher than in NFTG, and continues to be elevated during stimulation over a wide range of rates, mechanisms that regulate Ca 2+ removal from the cytosol are further impaired, and diastolic [Ca 2+ ] i levels increase excessively during stimulation at high frequencies, causing a more marked impairment of diastolic myocyte length than observed in NFTG (see below). Thus, although NCX function, as reflected in a reduced T 50 of the caffeine transient relaxation, is enhanced in FTG, this is not sufficient to maintain a normal diastolic Ca 2+ during high rates of stimulation. Contractile characteristics of a 1C TG myocytes have not been previously measured. Over a wide range of stimulation rates, the NFTG exhibit a hypercontractile state accompanied by diastolic failure due to an inability to maintain normal diastolic length. But in FTG, despite increased Ca 2+ transient amplitude, the contraction amplitude across a broad range of stimulation rates FTG is not greater than in NFTG. This indicates that a relative systolic contractile failure has occurred with evolution from the NFTG to FTG state. At high rates of stimulation, the contraction amplitude of FTG declines below that of NFTG to the level in NTG without a decline in SR Ca 2+ release and is accompanied by marked diastolic abnormalities. Thus NFTG and FTG differ in that the latter exhibit frank contractile systolic failure and more severe Ca 2+ and contractile diastolic failure than do the former.
Diastolic abnormalities accompany hypercontractile function in other animal experimental models [1]. The hypertrophic SHR myocytes studied at 23uC present a markedly similar profile to NFTG described herein [1,4,12]. In contrast, myocytes from some other models of cardiac hypertrophy exhibit both reduced Ca 2+ transient and contraction amplitudes. A common clinical correlate of NFTG may be hypertrophic non-dilated cardiomyopathy [13].
A hypercontractile state is attributable, at least in part, to an increased cytosolic [Ca 2+ ] i transient following excitation and may involve myofilaments, Ca 2+ energetic mechanisms, matrix properties or interactions of all.
It has recently been suggested that, like NFTG myocytes in the present study, an abnormal frequency-dependent response of human heart failure is a pathophysiologic signature of the cardiomyopathic state [1,2]. Both the hypercontractile state of NFTG myocytes, like the failing human heart, may reflect an altered functional balance between Ca 2+ influx, reuptake and extrusion from the cell [11] and can occur in the absence of systolic failure [14]. In failing ventricular myocardium, the positive FFR disappears and becomes flat or negative [1,2]. Multiple alterations in Ca 2+ handling determine negative staircase in a cellular heart failure model [15]. Interestingly, in the MLP knockout mouse with dilated cardiomyopathy (enhanced basal SERCA activity) increased PLN phosphorylation is believed to limit FFR [16]. In human heart failure reduced expression of SERCA has been demonstrated [1,2] however in animal studies, there is less consistency [1]. PLN expression levels are, however insufficient to evaluate the regulation of SERCA, as the level of inhibition of SERCA by PLN depends on the degree of phosphorylation. A decrease in the level of PLN phosphorylation at Ser16 (mainly affected by PKA activity) has been reported in some papers in heart failure, as against decreased levels of Thr17 phosphorylation (affected by the decreased CaMKII activity) [2].
Our results show that the maintenance of a normal myocyte SR Ca 2+ load afforded to older NFTG mice fails as the dilated, congested cardiomyopathic state evolves. In marked contrast to  most other cardiomyopathies, the excitation-induced Ca 2+ transient amplitude of FTG is increased. Also in contrast to most other experimental cardiomyopathies, (calcineurin overexpressing mice being an exception [17] in which the SR Ca 2+ load decreases) (Table S3), the SR Ca 2+ load of FTG increases, and, following excitation, generates a Ca 2+ transient of increased amplitude. Since neither I Ca,L amplitude nor inactivation under VC did differ between FTG and NTG, the source of the increased Ca 2+ that likely augments the SR Ca 2+ load in FTG compared to NFTG cannot be ascertained with certainty, particularly in the absence of action potential data in these two groups. Potential other sources of ''extra'' Ca 2+ in FTG mice may be increased influx via NCX [1,3,4,10] or efflux from mitochondria [18].
It is perplexing that the SR Ca 2+ load increases during the evolution from NFTG to FTG, since I Ca,L density does not increase, and Ca 2+ efflux via NCX is augmented, it might be expected that mechanisms to reduce Ca 2+ pumping into the SR might prevail: in many other HF models, SERCA2a becomes reduced, PLN increases, and its phosphorylation decreases [1,7]. The increased SR Ca 2+ load in FTG is apparently linked to an increased Ca 2+ availability for SR pumping, as evidenced by an increase in diastolic Indo-1 fluorescence during pacing. The latter may also reflect an increase in the net cell Ca 2+ load between the NFTG and FTG states.
Alterations in RyR [1] might also contribute to an increased RyR Ca 2+ leak. Unlike some other heart failure models, RyR2 and RyR2-P decrease, in the FTG (c.f., Table S3 Online Supplement), another conundrum. Some reports suggest that PKA phosphorylation of RyR2 has little functional relevance for diastolic Ca 2+ release if SR Ca 2+ levels remain constant [1,2]. Marks and coworkers in many studies and perspectives, indict RyR2 and Ca 2+ leak as causal of arrhythmias and sensitivities to heart failure [2,19,20]. There is a great deal of controversy surrounding the significance of both the increased Ca 2+ leak theory, as well as the decreased SR Ca 2+ content concept, as some studies have not been confirmed by others [1,2,21]. Our results, in fact, demonstrate that frank heart failure [10,11] can occur in the context of reduced total RyR2, phosphorylated RyR2 at Serine 2809 and increased SR Ca 2+ loading [22,23]. This underscores the complexity of ''heart failure'', and the caution that should be applied to any specific molecular conclusion at this point.
Some might argue that the FTG in which SR Ca 2+ remains elevated does not represent a ''good'' model of heart failure. But what, then, is a ''good model'' of heart failure? A ''good model'' of heart failure should represent what happens in the human condition. The presently in vogue dictum is that a reduced SR Ca 2+ load in end stage human HF is central to contractile dysfunction of human cardiomyocytes [10]. However, not all studies in human HF share this view [24]. In fresh, end stage failing hearts (a large number) obtained at transplant, using two different procedures, SR Ca 2+ accumulation demonstrated slow uptake kinetics and a complete lack of Ca 2+ release compared to non-failing samples, in which Ca 2+ release was vigorous [25,26]. These results, repeated on any human hearts are consistent with the present findings in the FTG myocytes. Further, a window of enhanced contractility, depending upon the stimulation rate, has been observed in human HF [1]. However, even in HF models that exhibit a reduced SR Ca 2+ load, muscles or myocytes under study conditions that have been employed exhibit an increase in the diastolic Ca 2+ at higher pacing rates i.e. similar patterns observed in the present study in FTG.
A recent study on failing human and mouse frank heart failure [27] has discovered a heretofore unrecognized gain of function change in the L-type Ca 2+ channel. Channel availability and single channel activity were markedly altered, the culprit being an up regulation of the b 2 -subunit of the L-VDCC complex. Further, the ensemble single channel activities of the terminally failing mouse heart and human heart were identical. Thus, quantitative measurements of single L-type channel function are vital when considering mechanisms involved in frank heart failure in all models of HF, regardless of species.
Other possible mechanisms in the conundrum of the end stage HF phenotype include heart rate responses [10], diminished contractile performance [28], increased diastolic calcium [29]. Abnormal myofilament mechanisms [30] may explain, in part, the deficient myocardial function of the intact a 1C TG, heart [4] but the mechanistic link between the two is not firm, because isolated cells, e.g., those analyzed in the present study, have no preload or afterload, two factors that, not only regulate contraction, but also regulate myofilament Ca 2+ binding and release characteristics. It is noteworthy, though, that regardless of the SR Ca 2+ load, contractile failure in cells or muscles from end stage, dilated heart failure hearts is exacerbated with increasing pacing rates in virtually all studies employing different pacing rates [10,11,14]. Myofilament abnormalities other than Ca 2+ binding or energetic fatigue, especially the activity of the b 2 -subunit of the L-VDCC, should continue to be recognized as mechanisms that underlie HF [10]. PKCa dephosphorylation of PLN was paralleled by diminished peak calcium transients.
Heart remodeling (dilatation), itself, leads to increased afterload, fibrosis, myofilament disarray and increased cell death [10], and each of these have been implicated in contractile dysfunction characteristic of FTG mice [4]. We observed that the ratio of TUNEL-positive cardiomyocytes to the total cardiomyocytes was higher in NFTG and FTG without signs of HF compared to littermate NTG (data not shown).
Of possible relevance to the a 1C -TG model, overexpression of the b 2a subunit of the L-VDCC leads to enhanced I Ca,L resulting in Ca 2+ dependent apoptosis in adult myocytes [18,31] consistent with recent findings [27]. Studies of the transition from hypertrophy to HF that occurs in SHR during advanced age [32] have also shown that increased cardiomyocyte apoptosis occurs [33].
To summarize, end stage cardiac failure phenotype is remarkably similar to that in other types of heart failure in which myocytes have a reduced SR Ca 2+ load and reduced Ca 2+ amplitude, at least under conditions that have been employed in these studies (Table S1). Similar heart failure phenotypes are associated with both an increased or decreased SR Ca 2+ load as well as SR Ca 2+ release following excitation. This suggests that there is a requirement for optimal Ca 2+ transient amplitude to prevent HF, or that the mode of measurements, type of failure and how quickly measurements are made are of significance. Heterogeneity of Ca 2+ release within and among cells in intact myocardium impairs both systolic and diastolic function and provokes arrhythmias [34]. Recent studies have indeed begun to address this issue in experimental heart failure models [2,4,7,35,36].

Materials and Methods
FBV/n in-house bred 8-11 mo old a 1C TG mice and their wildtype (NTG) littermates of the same age were used in all of the experiments. The animal protocols used in these studies were approved by the Animal Care Committees of each institution involved. At 9 mo of age some a 1C TG mice begin to manifest a dilated cardiomyopathy with markedly impaired contractile function in vivo [4]. Upon sacrifice we observed a 1C TG -mice at this age exhibited two phenotypes: the absence of gross cardiac dilatation and congestion, evidenced by the absence of pericardial and pleural effusions and atrial thrombi (n = 4), and gross cardiac dilatation with both pericardial and pleural effusions and left atrial thrombi (n = 4). We designated these two groups of mice at sacrifice as non-failing a 1C TG (NFTG) and failing a 1C TG (FTG), respectively, and structured our data analysis to compare some aspects of Ca 2+ regulation and contraction characteristics of myocytes isolated from these hearts.
We measured Ca 2+ transients in Indo-1-AM loaded myocytes at frequencies of stimulation ranging from 0.5-4 Hz. The same stimulation protocol was used to measure myocyte mechanics in both Indo-1-loaded and unloaded cells. The SR Ca 2+ load was determined measuring caffeine induced Ca 2+ transients in Indo-1-AM loaded cells. Inward Ca 2+ current through the L-type channel was measured under step-voltage protocols. We carried out quantitative and traditional Western blotting for the Ca 2+ handling proteins. FTG and NFTG subgroup comparisons were stratified prospectively and investigators were blinded when doing their experiments. All methods have been previously published and are detailed in the Methods S1.

Statistics
Results are reported as mean 6SEM. Differences among NTG and NFTG and FTG cells were compared by ANOVA. A value of p,0.05 was considered significant.