Phospholamban ablation in hearts expressing the high affinity SERCA2b isoform normalizes global Ca2+ homeostasis but not Ca2+-dependent hypertrophic signaling

William E. Louch, Peter Vangheluwe, Virginie Bito, Luc Raeymaekers, Frank Wuytack, Karin R. Sipido


Cardiomyocytes from failing hearts exhibit reduced levels of the sarcoplasmic reticulum (SR) Ca2+-ATPase (SERCA) and/or increased activity of the endogenous SERCA inhibitor phospholamban. The resulting reduction in the Ca2+ affinity of SERCA impairs SR Ca2+ cycling in this condition. We have previously investigated the physiological impact of increasing the Ca2+ affinity of SERCA by substituting SERCA2a with the higher affinity SERCA2b pump. When phospholamban was also ablated, these double knockouts (DKO) exhibited a dramatic reduction in total SERCA levels, severe hypertrophy, and diastolic dysfunction. We presently examined the role of cardiomyocyte Ca2+ homeostasis in both functional and structural remodeling in these hearts. Despite the low SERCA levels in DKO, we observed near-normal Ca2+ homeostasis with rapid Ca2+ reuptake even at high Ca2+ loads and stimulation frequencies. Well-preserved global Ca2+ homeostasis in DKO was paradoxically associated with marked activation of the Ca2+-dependent nuclear factor of activated T-cell-calcineurin pathway known to trigger hypertrophy. No activation of the MAP kinase signaling pathway was detected. These findings suggest that local changes in Ca2+ homeostasis may play an important signaling role in DKO, perhaps due to reduced microdomain Ca2+ buffering by SERCA2b. Furthermore, alterations in global Ca2+ homeostasis can also not explain impaired in vivo diastolic function in DKO. Taken together, our results suggest that normalizing global cardiomyocyte Ca2+ homeostasis does not necessarily protect against hypertrophy and heart failure development and that excessively increasing SERCA Ca2+ affinity may be detrimental.

  • phospholamban ablation
  • cardiac hypertrophy
  • cardiomyocyte Ca2+ homeostasis

in cardiomyocytes, sarcoplasmic reticulum (SR) Ca2+ ATPase (SERCA) function is an important determinant of SR Ca2+ content and cellular contractility. SERCA activity is regulated by phospholamban (PLB), which inhibits SERCA by decreasing its apparent Ca2+ affinity. The pivotal role of PLB as an affinity modulator has been explored in a number of transgenic mouse models. Decreasing the affinity of SERCA for cytosolic Ca2+ by increasing inhibition from PLB causes decreased cardiac function and heart failure (7, 15, 34, 37). In contrast, increasing the ratio of the cardiac-specific SERCA isoform (SERCA2a) to PLB improves cardiac function (12, 20, 30, 33) and can alleviate heart failure (8, 22). However, ablation of PLB does not rescue cardiac dysfunction in all mouse models of heart failure (24, 35). As well, a mutated form of PLB that cannot inhibit SERCA has been linked to heart failure in humans (11). Thus the therapeutic potential of increasing the Ca2+ affinity of SERCA is not unambiguous (23). This issue is particularly relevant since the benefits of SERCA overexpression in heart failure patients are already being examined in an ongoing clinical trial (14).

Since increasing the Ca2+ affinity of SERCA2a has potential to improve cardiac contractility, it is quite surprising that the heart expresses SERCA2a rather than the ubiquitous Ca2+ pump SERCA2b. Indeed, SERCA2b has a higher Ca2+ affinity but lower maximal turnover rate than SERCA2a and, as could be expected from a housekeeping pump, is normally expressed at low levels also in cardiomyocytes (9, 26). To address this, we have previously investigated mice in which the SERCA2a splicing mechanism was disrupted, resulting in expression of only SERCA2b in the heart (SKO mice) (28). SKO mice exhibited mild left ventricular hypertrophy and minor cardiac dysfunction but a dramatic reduction in SERCA levels and increased PLB expression. To investigate whether alleviation of the increased PLB suppression of SERCA would improve cardiac function and prevent hypertrophy, SKO mice were crossed with PLB−/− mice (27). Although double knockout (DKO) mice exhibited a high SERCA2 Ca2+ affinity, total SERCA levels were further reduced to 30% of wild-type (WT) values and the mice developed severe hypertrophy, diastolic dysfunction, and reduced contractile reserve (27). We presently examined the role of cardiomyocyte Ca2+ handling in the functional defect, which could be related to a reduced capacity for SR Ca2+ uptake. Reduced Ca2+ removal capacity may also trigger the structural remodeling in DKO hearts through activation of Ca2+-dependent signaling pathways.


Experiments were conducted with adult mice (age 10–16 wk) of either sex in accordance with the Guide for the Care and Use of Laboratory Animals (National Institute of Health). Protocols for handling experimental animals were also approved by the ethical commission of KU Leuven. For analysis of RNA and protein expression levels, hearts were rapidly excised and frozen. Ventricular homogenates were prepared as in (27). Standard Western blotting techniques were employed (27) using polyclonal antibodies directed against calcineurin Aβ (CnAβ; Chemicon, Billerica, MA, USA) and monoclonal antibodies against the kinases of the mitogen-activated protein kinase (MAPK) pathway (total levels and phosphorylated levels): JNK, P-JNK, p39, P-p38, ERK, and P-ERK (Cell Signaling Technology, Beverly, MA). Ventricular total RNA was isolated as in Vangheluwe et al. (27), and mRNA levels of modulatory calcineurin-interacting protein 1.4 (MCIP1.4) and calcineurin A were determined by reverse transcription PCR. RNA was treated with DNase I, and cDNA was prepared from 1 μg total RNA in 20 μl reverse transcriptase buffer according to the manufacturer (Thermoscript; Invitrogen, Merelbeke, Belgium). Negative controls without reverse transcriptase were tested. The number of PCR cycles was chosen to fall in the exponential phase of amplification. Primers were as follows: MCIP1.4: forward primer, 5′-AAGGAACCTCCAGCTTGGGCT-3′, reverse primer, 5′-CCCTGGTCTCACTTTCGCTG-3′, Ta = 54°C, fragment of 159 bp, 20 cycles; and calcineurin A: forward primer, 5′-CCACAGGGATGTTGCCTAGTG-3′, reverse primer, 5′- GTCCCGTGGTTCTCAGTGGTA-3′, Ta = 55°C, fragment of 244 bp, 20 cycles.

Ventricular myocytes were enzymatically isolated as described previously (4). Cells were stored at room temperature, and experiments were performed within 6 h of isolation. Myocytes were plated in an open-perfusion chamber mounted on the stage of an inverted microscope. Cell contractions were recorded with a video-edge detector (Crescent, Salt Lake City, UT) at a 240-Hz frame rate, and data are presented as fractional shortening (%diastolic length). Ca2+ transients were recorded by whole cell photometry in cells loaded with fluo-3. Ca2+ transient recordings from field-stimulation experiments are presented normalized to baseline fluorescence (F/F0), while fluo-3 signals from voltage-clamped myocytes have been calibrated to intracellular Ca2+ concentration ([Ca2+]i). Calibration of [Ca2+]i was performed as described by Trafford et al. (25) using Fmax values measured at the end of the experiment; with this method, resting [Ca2+]i can be measured. In some cells, Fmax values were not obtained, and Ca2+ transients were instead calibrated by the approach of Cheng et al. (6) using measured resting [Ca2+]i (WT = 137.8 nmol/l; DKO = 130.8 nmol/l).

Basic characterization of cellular contractions, Ca2+ transients, and the frequency response was performed during perfusion with a modified Tyrode solution (in mmol/l: 137 NaCl, 5.4 KCl, 0.5 MgCl2 , 1.0 CaCl2, 11.8 Na-HEPES, and 10 glucose, pH 7.4). Myocytes were either field stimulated (1-Hz, 5-ms square pulses, 37°C) or studied in whole cell voltage clamp mode (30°C). The pipette solution contained the following (in mmol/l): 120 K-aspartate, 20 KCl, 10 K-HEPES, 5 MgATP, 10 NaCl, and 0.05 K5-fluo-3, pH 7.2. The voltage-clamp protocol in these experiments consisted of 25-ms voltage steps from −70 to +20 mV, given at a range of frequencies. The effects of 10 μM forskolin (Sigma-Aldrich, Bornem, Belgium) treatment were examined in some experiments.

Membrane currents were sampled at 4 kHz and filtered at 2 kHz. For measurement of L-type Ca2+ current (ICaL), K+ currents were blocked by replacing K+ with Cs+. The external solution contained the following (in mmol/l): 130 NaCl, 10 CsCl, 0.5 MgCl2, 1.0 CaCl2, 11 Na-HEPES, and 10 glucose, pH 7.4, 22°C. The pipette solution contained the following (in mmol/l): 120 Cs-aspartate, 10 TEACl, 10 Cs-HEPES, 0.5 MgCl2, 5 MgATP, and K5-fluo-3, pH 7.2. The voltage-clamp protocol consisted of a 250-ms prepulse from −70 mV to −45 mV, followed by a series of 250-ms depolarizing voltage steps ranging between −45 mV and +65 mV. Cellular capacitance was measured during a voltage step from −70 mV to −80 mV.

To measure releasable SR Ca2+ content, a train of 10 conditioning pulses (1 Hz, −70 to +20 mV) was used to induce stable SR loading, after which cells were clamped at −70 mV and rapidly perfused with 10 mmol/l caffeine. SR content was estimated by measuring the magnitude of the caffeine-elicited Ca2+ transient and by integrating the caffeine-induced Na+-Ca2+- exchange (NCX) current.

Elevated Ca2+ loads were induced by blocking Ca2+ extrusion via the Na+-Ca2+ exchanger as described previously (4). The pipette solution in these experiments contained the following (in mmol/l): 120 Cs-aspartate, 10 TEACl, 10 Cs-HEPES, 0.5 MgCl2, 5 MgATP, and 0.05 K5-fluo-3, pH 7.2. After the SR was depleted with a rapid caffeine application, cells were exposed to a Na+-free solution (in mmol/l: 120 NMDGCl, 20 TEACl, 11 HEPES, 0.5 MgCl2, 5.4 CaCl2, and 10 glucose, pH 7.4, 22°C) during 0.25-Hz stimulation with 100-ms voltage steps from −70 to 0 mV.

Statistical analyses were performed using SigmaPlot and SigmaStat software, and statistical significance (P < 0.05) was determined by paired or unpaired Student t-tests. Data are displayed as mean values ± SE.


Ca2+-dependent hypertrophy signaling pathways are activated in DKO.

We (27) have previously observed modest concentric hypertrophy in SKO hearts that was exacerbated in DKO (Fig. 1A). For insight into the signaling pathways underlying hypertrophy in these mice, we here examined the calcineurin-nuclear factor of activated T-cell (NFAT) system, as this pathway is both required and sufficient for hypertrophic signaling (29). This was done by measuring the NFAT-dependent transcription activation of MCIP1.4, which, as an endogenous reporter for calcineurin activity, contains multiple NFAT binding sites in an internal promoter (32). MCIP1.4 mRNA levels (Fig. 1, B and C) were increased by approximately two- and threefold in SKO and DKO, respectively, correlating well with the degree of hypertrophy observed in these mice. Calcineurin mRNA levels were unaltered (Fig. 1, B and C), as were the ratios between phosphorylated and unphosphorylated JNK, ERK, and p38 (Supplemental Fig. S1; Supplemental Material for this article is available online at the Am J Physiol Heart Circ Physiol website), i.e., components of the MAP kinase signaling pathway. These data indicate an involvement of Ca2+-dependent (calcineurin-NFAT) but not MAP kinase signaling pathways in SKO and DKO remodeling.

Fig. 1.

Hypertrophy in single and double knockout (SKO and DKO) hearts is correlated with activation of Ca2+-dependent hypertrophic signaling pathways. Mild concentric hypertrophy was observed in SKO hearts, and was exacerbated in DKO, as was shown before in (25) (A). The degree of hypertrophy was correlated with modulatory calcineurin interacting protein 1.4 (MCIP1.4) transcription (B and C), an endogenous reporter for calcineurin-nuclear factor of activated T-cell activation. Calcineurin mRNA levels were unaltered in SKO and DKO. CnAβ, calcineurin Aβ. nhearts = 4, 3, and 5 in wild-type (WT), SKO, and DKO, respectively; *P < 0.05 vs. WT.

Myocyte contraction and relaxation are improved in DKO.

Because the extent of hypertrophy in SKO and DKO correlates with calcineurin activation, we explored WT, SKO, and DKO myocytes for altered Ca2+ signaling. Representative recordings of myocyte contractions and Ca2+ transients are shown in Fig. 2, A and B, respectively. Mean systolic cell shortening (Fig. 2C) was significantly larger in DKO cells than in SKO. Relaxation tended to be slower in SKO than in WT, as indicated by larger tau values, but was significantly faster in DKO (Fig. 2D). Ca2+ transients were of similar amplitude in WT, SKO, and DKO (Fig. 2E), and Ca2+ decayed more rapidly in DKO than either WT or SKO (Fig. 2F). Thus cytosolic Ca2+ transient amplitude and kinetics did not predict the degree of activation of Ca2+-dependent signaling pathways in the three groups (Fig. 1). Altered Ca2+ homeostasis could also not explain diastolic dysfunction observed in vivo (27).

Fig. 2.

Contraction and relaxation are not impaired in isolated DKO cardiomyocytes. Representative measurements of cell shortening (A) and mean data show that contractions elicited during 1 Hz field stimulation tended to be larger in DKO cells (C), and relaxation more rapid (D; ncells = 13, 21, and 17 in WT, SKO, and DKO). Ca2+ transients (B) were of similar magnitude in WT, SKO, and DKO (E), while Ca2+ decline was most rapid in DKO (F; ncells = 18, 14, and 16). F/F0, baseline fluorescence. *P < 0.05 vs. WT; †P < 0.05 vs SKO.

The rapid decline of Ca2+ transients in DKO is surprising given the extremely low SERCA expression levels (30% of WT). We therefore examined the possibility that an increased rate of Ca2+ removal by other pathways might be a contributing factor. Patch-clamped cells were stimulated with a train of voltage steps (−70 to +20 mV) at 1 Hz, followed by rapid application of caffeine (Fig. 3A). Ca2+ declined more slowly in DKO than WT during caffeine application (Fig. 3, A and B). This was not related to a reduced Ca2+ removal by NCX (peak values of NCX current normalized to [Ca2+]: 0.13 ± 0.04 nA/μM in DKO vs. 0.14 ± 0.03 nA/μM in WT; P = NS). Comparison of these [Ca2+] measurements with simultaneously recorded NCX currents in a pilot experiment on five cells indicated no change in global Ca2+ buffering (data not shown). DKO cells instead exhibited slowed Ca2+ removal by non-NCX (and non-SERCA) mechanisms, as indicated by a protracted decay of caffeine transients in the presence of 5 mM Ni (Fig. 3C). Such slowing of Ca2+ removal could theoretically result from a reduced rate of mitochondrial Ca2+ uptake; however, there is controversy concerning the kinetics of Ca2+ removal by this pathway (18). Another possibility is reduced Ca2+ extrusion by the plasmalemmal Ca2+ ATPase (PMCA). Although we (27) have previously reported no alterations in PMCA expression in DKO, it is possible activity is nevertheless reduced. We conclude that the rapid decline of Ca2+ transients in DKO must result from markedly enhanced SR Ca2+ reuptake, presumably due to the high affinity of SERCA2b and absence of PLB.

Fig. 3.

Ca2+ extrusion is slowed in DKO. Representative recordings (A) and mean data (B; ncells = 8 and 12 in WT and DKO) show protracted decline of Ca2+ following application of 10 mM caffeine in DKO myocytes. Similar observations were made during Na+-Ca2+-exchange (NCX) inhibition (5 mM Ni) (C; ncells = 8 and 7), suggesting that non-NCX mediated Ca2+ extrusion is reduced in DKO. [Ca2+]i, intracellular Ca2+ concentration. *P < 0.05 vs. WT.

SR content and release are normal in DKO.

Ca2+ transients elicited by voltage steps at 1 Hz were of similar magnitude in DKO cells as in WT, as illustrated by representative examples (Fig. 3A) and mean data (Fig. 4A). SR content was also similar in WT and DKO, as assessed by the magnitude of the caffeine-elicited Ca2+ transient (Figs. 3A and 4B) and integrated NCX current (Fig. 4C). The trigger for SR Ca2+ release, ICaL, was of comparable magnitude and voltage dependence in WT and DKO (Fig. 4D). Thus, at low pacing rates, global Ca2+ homeostasis was not impaired in DKO and could not account for structural and functional remodeling observed in vivo.

Fig. 4.

SR Ca2+ content and L-type Ca2+ current are unaltered in DKO. In patch-clamped cells, 1 Hz voltage steps from −70 to + 20 mV elicited similar amplitude Ca2+ transients in WT and DKO (A; ncells = 8 and 11 in WT and DKO). Following the conditioning train, caffeine application induced similar magnitude Ca2+ transients (B; ncells = 8 and 12) and integrated NCX current (C; ncells = 14 and 13) in WT and DKO, indicating normal SR Ca2+ content. Peak L-type Ca2+ current elicited by voltage-clamp steps from −45 mV was also unaltered in DKO (D; ncells = 10 and 8). INCX, Na+-Ca2+-exchange current.

Ca2+ cycling is preserved at high pacing rates.

In addition to exhibiting functional deficits at baseline, DKO mice also show impaired cardiac reserve (27). An important component of cardiac contractile reserve is an ability of myocytes to enhance SR Ca2+ reuptake as the stimulation frequency and cytosolic Ca2+ load increase. We therefore examined SR Ca2+ uptake under these two conditions.

We first tested the frequency response by stimulating myocytes with 25-ms voltage steps from −70 to +20 mV at different frequencies. Additionally, each stimulus train was followed by a 1-s pause and then a test pulse to assess SR Ca2+ release with ICaL fully recovered (Fig. 5A) (3). Increasing stimulation frequency increased diastolic [Ca2+]i in WT, but, interestingly, diastolic values were not significantly elevated in DKO (Fig. 5B). At low frequencies, the rate of Ca2+ decay was faster in DKO than WT, but frequency-dependent acceleration of Ca2+ decline only occurred in WT. Thus, at 8 Hz, time to half decay was similar in the two groups (Fig. 5C). These findings indicate that at physiological stimulation frequency diastolic Ca2+ homeostasis is not impaired in DKO despite a 70% reduction in SERCA levels.

Fig. 5.

SR Ca2+ uptake and release are maintained at high stimulation frequencies in DKO. Cells were stimulated with conditioning pulses at a range of frequencies, followed by a 1-s pause and test pulse (A). Increasing conditioning pulse stimulation frequency significantly elevated diastolic [Ca2+]i (B) and reduced Ca2+ decay time (C) in WT but not DKO (B; ncells = 11 and 7). Conditioning pulse (D) and test pulse (E) magnitudes were not significantly different between WT and DKO, across the range of frequencies. At high frequencies, SR content was similar in WT and DKO, although only WT cells demonstrated a frequency-dependent increase (F; ncells = 5 and 7). *P < 0.05 vs. WT; †P < 0.05 vs. 1 Hz.

Both WT and DKO exhibited a decline in the amplitude of Ca2+ transients as frequency increased (Fig. 5D). This is likely due to ICaL inactivation (3), since the amplitude of test pulses applied after a 1-s pause (to allow ICaL recovery) was maintained in DKO and tended to increase in WT (Fig. 5E). In agreement with this observation, caffeine applications administered at the end of the stimulus train indicated increasing SR load with frequency in WT and maintained SR load in DKO (Fig. 5F). Thus DKO remain capable of loading the SR at higher frequencies, although this ability appears to be modestly impaired compared with WT. These results show that at physiological frequency, there are no alterations in the Ca2+ transient (diastolic [Ca2+]i, peak [Ca2+]i , or Ca2+ decay rate) that would be expected to activate calcineurin or result in the diastolic dysfunction observed in vivo.

Ca2+ uptake is preserved at high Ca2+ load.

We further challenged SR functional capacity in DKO by experimentally increasing SR load. Figure 6A shows representative Ca2+ transient recordings during progressive Ca2+ loading in Na+-free conditions with high extracellular [Ca2+] (5.4 mmol/l). [Ca2+]i increased more rapidly in DKO than WT, as indicated by mean measurements of the peak [Ca2+]i attained during each Ca2+ transient (Fig. 6B). Since L-type Ca2+ current is of similar magnitude in WT and DKO, the more rapid increase in [Ca2+]i observed in DKO likely results from decreased Ca2+ extrusion by non-NCX mechanisms and/or the high Ca2+ affinity of SERCA2b. Despite this large increase in [Ca2+]i in DKO, the time to half decline of the Ca2+ transient remained markedly below WT values throughout the protocol (Fig. 6B). By comparison, we have previously observed that SKO cells exposed to this protocol exhibit impaired Ca2+ reuptake as Ca2+ levels rise, presumably due to increased PLB expression (4). Thus, in the absence of phospholamban expression, the low level of SERCA2b in DKO myocytes can maintain efficient SR Ca2+ reuptake even at elevated Ca2+ load. This indicates that under these conditions the Ca2+ affinity of SERCA is more important than the maximal turnover rate.

Fig. 6.

SR Ca2+ uptake is maintained in DKO at high Ca2+ loads. Cellular Ca2+ load was elevated by increasing Ca2+ influx and inhibiting Ca2+ efflux (extracellular Ca2+ concentration = 5.4 mM and Na+ concentration = 0 mM). During 5 consecutive voltage steps (−70 to 0 mV; top), Ca2+ transients progressively increased in magnitude, as indicated by representative examples (A) and mean data (B). In DKO, transient half-decay time remained markedly reduced from WT values, even at high Ca2+ loads (C; ncells = 6 and 5 in WT and DKO); *P < 0.05 vs. WT.


In the present study, we have demonstrated that myocytes in which SERCA2a is replaced by the high affinity SERCA pump, SERCA2b, in the absence of PLB exhibit remarkably normal global Ca2+ homeostasis. Despite the low SERCA levels in DKO, we observed rapid Ca2+ reuptake even at high Ca2+ loads and stimulation frequencies. The well-preserved Ca2+ homeostasis in DKO was paradoxically associated with marked activation of Ca2+-dependent signaling pathways.

Comparison with in vivo cardiac function.

Some observations of isolated cardiomyocyte function in DKO are consistent with the in vivo phenotype of these mice. Ca2+ transients were found to be of near-normal magnitude over a wide range of stimulation frequencies in DKO. In fact, at low stimulation frequencies SR Ca2+ content, Ca2+ transient magnitude, and contraction were observed to be slightly larger in DKO than in WT, although these trends did not reach statistical significance in all cases. These data support the in vivo finding of preserved load-independent systolic function (27). However, in vivo diastolic function (assessed by dP/dtmin) and overall cardiac output are depressed in DKO (27). This observation is at odds with observations made in cells where Ca2+ reuptake is more rapid in DKO than WT at low stimulation frequencies and normal at physiological rates. Impaired diastolic function in the intact heart may instead result from an observed stiffening of the myocardium (27). Although we have not observed fibrosis in DKO hearts (27), other unknown alterations in the extracellular matrix could compromise tissue compliance. The geometric constraints of hypertrophy should also be considered, as wall thickening and reduction in chamber size during concentric hypertrophy can impair diastolic function (16).

In the normal heart, an increase in heart rate elevates SR Ca2+ content. This response is an important component of contractile reserve and believed to be largely mediated via phospholamban (36). Indeed, we (27) previously observed impaired contractile reserve in DKO mice. We presently showed that although SR content did not significantly increase with frequency in DKO, releasable Ca2+ levels were close to WT values. This observation suggests that alterations at the level of the cardiomyocyte likely only minorly contribute to impaired in vivo performance of the heart at high frequencies. Thus other limitations such as the remodeled myocardial geometry and possibly a reduced response to adrenergic stimulation may impair contractile reserve in DKO.

Data from a number of transgenic mouse models have suggested that the PLB-to-SERCA ratio is an important determinant of cardiac contractility and relaxation (21). For example, cardiac-specific overexpression of PLB reduces contraction and relaxation both in vivo and in isolated cells (15). Opposite effects have been observed in mice overexpressing SERCA (12). However, it appears that the PLB-to-SERCA ratio may only be a good indicator of contractility and relaxation in a SERCA2a background. We observed that in vivo and isolated cardiomyocyte function were only mildly affected in SKO despite a fourfold increase in PLB to SERCA2b (4, 28). Using mathematical modeling, we predicted that lower SERCA levels in these mice compensate for the presence of the high Ca2+ affinity SERCA isoform to prevent excessive SR Ca2+ uptake and abnormally low diastolic [Ca2+]i (4). When PLB is ablated, SERCA levels are further reduced, which, as indicated by our present findings, maintains Ca2+ cycling in a normal range. Thus it appears that the Ca2+ affinity of SERCA is tightly regulated within a narrow window to minimize deficiencies in Ca2+ handling. In this regard, it appears that the Ca2+ affinity of SERCA is more important than the expression level and maximal turnover rate.

Implications for hypertrophy signaling.

It is intriguing that although we observed marked activation of the Ca2+-dependent NFAT-calcineurin signaling pathway in DKO, we could not detect alterations in global Ca2+ homeostasis in cardiomyocytes. This finding suggests that local Ca2+ signaling may be importantly modified in DKO. We observed that diastolic Ca2+ levels are ∼150 nM in DKO cells at physiological frequency (Fig. 5B), and recent evidence based on Ca2+-dependent modulation of LTCC and NCX indicates that local Ca2+ levels in the dyadic cleft normally reach extremely high levels during Ca2+ release (10–15 μM) (1). Importantly, we (27) have observed that the KD of SERCA for Ca2+ in DKO is reduced to only 100 vs. 590 nM in WT. Thus, with high Ca2+ affinity and low expression levels, SERCA2b might be saturated with Ca2+ in DKO. Since global Ca2+ homeostasis remains near normal, SERCA2b saturation may be modest but sufficient to locally reduce Ca2+ buffering and activate Ca2+-dependent signaling molecules. Notably, the calcineurin-interacting protein calsarcin-2 couples calcineurin to α-actinin at the Z-line of the sarcomere, which might link calcineurin activation to local Ca2+ changes in this region (10). Therefore, while upregulation of PLB in SKO and downregulation of SERCA in DKO may be aimed at maintaining global Ca2+ homeostasis and contractility, such alterations may inappropriately alter local Ca2+ homeostasis and trigger hypertrophy signaling pathways. The role of local microdomains in Ca2+-dependent hypertrophic signaling is likewise supported by the finding that hypertrophy is triggered by overexpression of the β2a-subunit of the L-type Ca2+ channel (5) but not by overexpression of the T-type Ca2+ channel (13). This differential effect has been attributed to distinct localization of these Ca2+ channels, with L-type channels present in the T-tubules and T-type channels present exclusively in the surface sarcolemma.

An interesting comparison can be made between the DKO mouse and the recently described conditional SERCA KO mouse (2, 19). Ablation of the Serca2 gene in adult mice does not result in cardiac hypertrophy despite a reduction in SERCA levels to below 5% of normal values (2, 19). Based on our discussion above, one might have speculated that with such extremely low SERCA levels, the conditional SERCA KO mice would be even more prone to hypertrophy, which is not the case. However, two important differences between the SERCA KO and DKO mice should be considered. First, the conditional SERCA KO predominantly expresses SERCA2a (WT background), and second, Ca2+ transient amplitude is reduced in these cells. Together, this may mean that remaining SERCA2a pumps are not easily saturated with Ca2+. Furthermore, SERCA KO myocytes exhibit an impressive upregulation of Ca2+ extrusion pathways, including both NCX and non-NCX systems, which has been shown to maintain diastolic [Ca2+]i at normal values (17, 19). In DKO myocytes, on the other hand, the expression of NCX and PMCA are unaltered (27), and we even observed a somewhat impaired non-NCX mediated Ca2+ extrusion (Fig. 3). Indeed, Wu et al. (31) have suggested that PMCA4b activity controls a sub-sarcolemmal Ca2+ microdomain critical for calcineurin activation. Thus the hypertrophy signaling we observed in DKO could theoretically result from reduced extrusion from such a subsarcolemmal compartment.

Estimation of in vivo NFAT activity was accomplished by measuring MCIP1.4 expression since, with its 15 NFAT binding sites, this is a robust reporter of NFAT activity (32). Since MCIP is endogenous, this approach is arguably better suited for measuring NFAT activity than alternative methods requiring exogenous introduction of NFAT-dependent reporter genes. Commercially available kits are also not well suited for studies such as ours since they only allow calcineurin activity to be assessed at a fixed in vitro Ca2+ concentration.

While hypertrophy signaling pathways were presently examined in tissue from animals not deliberately exposed to stress protocols, it remains possible that these animals are submitted to intermittent stress in the animal facility. Thus the question could be raised as to whether baseline or fluctuating catecholamine levels affect Ca2+ signaling in DKO to an extent that hypertrophy pathways are triggered. It is certainly true that due to the absence of PLB in DKO myocytes, these cells would be unable to respond normally to β-adrenergic stimulation by increasing the affinity of SERCA for Ca2+. In this situation, stress-induced alterations such as enhanced L-type Ca2+ current could theoretically activate NFAT signaling. Our observation that DKO myocytes can efficiently handle large increases in Ca2+ load (Fig. 6) argues against this possibility. However, to be sure that a poor stress response is not involved with NFAT signaling in DKO, future work may examine whether β-blockers protect against hypertrophy development in these animals. It would also be interesting to examine NFAT activity in isolated myocytes where confounding issues of animal stress levels can be ruled out.


Our findings demonstrate that in mice expressing the high affinity SERCA2b pump instead of SERCA2a, ablation of PLB rescues global cardiomyocyte Ca2+ homeostasis but does not prevent activation of Ca2+-dependent growth signaling pathways. These observations suggest that important differences in local Ca2+ homeostasis might occur in DKO, perhaps due to reduced microdomain Ca2+ buffering by SERCA2b. Global Ca2+ homeostasis can also not explain impaired in vivo diastolic function in DKO, suggesting that this deficit instead results from hypertrophy-associated geometric constraints and/or loss of myocardial compliance.


P. Vangheluwe is a postdoctoral fellow of the Flemish Research Foundation FWO. This study was supported by a FWO Grant G. 0617.09N (to K.R. Sipido) and G.0646.08 (to F. Wuytack).


No conflicts of interest, financial or otherwise, are declared by the author(s).


Author contributions: W.E.L., P.V., L.R., F.W., and K.R.S. conception and design of research; W.E.L., P.V., and V.B. performed experiments; W.E.L., P.V., and V.B. analyzed data; W.E.L., P.V., V.B., F.W., and K.R.S. interpreted results of experiments; W.E.L. and P.V. prepared figures; W.E.L., P.V., F.W., and K.R.S. drafted manuscript; W.E.L., P.V., V.B., L.R., F.W., and K.R.S. edited and revised manuscript; W.E.L., P.V., V.B., L.R., F.W., and K.R.S. approved final version of manuscript.


We thank Dr. Niall Macquaide for helpful discussions.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
View Abstract