Synergistic interactions between Ca²⁺ entries through L-type Ca²⁺ channels and Na⁺–Ca²⁺ exchanger in normal and failing rat heart

Methods

Confocal Ca²⁺ imaging and patch-clamp measurements

Single ventricular myocytes were isolated by enzymatic dissociation from hearts from male Sprague–Dawley rats (Charles River Laboratories, Raleigh, NC, USA). Adult rats (200–300 g) were killed with an overdose of sodium pentobarbital in accordance with the guidelines of the Institutional Animal Care and Use Committee. Whole-cell transmembrane ionic currents and changes in membrane potential were recorded with an Axopatch 200B amplifier (Axon Instruments, USA) and pClamp (6.03) software. The bathing solution contained (mm): 140 NaCl, 5.4 KCl, 1 CaCl₂, 0.5 MgCl₂, 10 Hepes and 5.6 glucose; pH 7.3. The pipette solution contained (mm): 70 caesium aspartate, 50 CsCl, 3 ATP(Mg²⁺), 4 NaCl, 5 Hepes, 0.05 fluo-3 potassium salt; pH 7.3. Voltage pulses of 150 ms duration were applied from a holding potential of −50 mV at 1 min intervals. Signals were sampled at 5 kHz and filtered at 2 kHz. Cytosolic [Ca²⁺] changes in patch-clamped myocytes were measured with an Olympus Laser Scanning Confocal System (LSM-GB200, Olympus Optical Co. Ltd, Tokyo, Japan), equipped with an Olympus ×60 oil, 1.4 NA objective. Fluo-3 was excited at the 488-nm line of an argon laser, with emission collected through a 515-nm long-pass filter. Fluorescence images were recorded in the line-scan mode at the rate of 2 ms per line. In experiments involving measurements of trans-sarcolemmal Ca²⁺ entries, the pipette solution contained 0.1 mm fluo-3 potassium salt. In certain experiments, the SR Ca²⁺ content was increased by applying a loading protocol consisting of a train of 50 ms pulses from −50 mV holding potential to 0 mV at 2 Hz. All measurements were performed at 23°C. KB-R7943 (2-(2-(4(4-nitrobenzyloxy) phenyl)ethyl)isothiourea mesylate) was obtained from Tocris (USA). All other reagents were from Sigma-Aldrich (USA).

Aortocaval shunt-induced HF

The aortocaval shunt (ACS) was produced according to the method described by Garcia & Diebold (1990). Briefly, after the animal was anaesthetized with pentobarbital, the inferior vena cava and abdominal aorta were exposed by laparotomy. An 18 gauge hypodermic needle was advanced into the aorta distal to the left renal artery and through the opposite wall into the inferior vena cava in order to establish a communication between the two vessels. Bulldog vascular clamps were placed across the aorta above and below the entry point, the needle was withdrawn, and the aortic puncture point was sealed with a drop of cyanoacrylate glue. The clamps were removed 30 s later. The presence of a patent shunt was confirmed visually after release of the clamps by the occurrence of swelling of the vena cava and admixture of arterial and venous blood. The laparotomy was then closed. Sham-operated animals (n= 6) serving as controls were subjected to the same surgical procedures except for creation of the shunt. Compensated hypertrophy occurs after 5–6 week of ACS, with progression to HF after 8–12 week (Garcia & Diebold, 1990; Pieruzzi et al. 1995). In vivo cardiac haemodynamic measurements and confirmation of the shunt were done using the Hewlett Packard Sonos 5500 Ultrasound system (12 MHz transducer). M-mode echocardiograms were recorded at the left ventricular (LV) papillary muscle level to measure end-diastolic and end-systolic dimension, fractional shortening, and end-diastolic posterior wall thickness. ACS rats without LV dilatation and decreased shortening after 8–12 weeks of ACS were excluded from the study.

Isoproterenol-induced cardiomyopathy

Chronic administration of isoproterenol (ISO) was performed either by daily subcutaneous injections or by means of surgically implanted osmotic minipumps for a period of 28 days. Rats were injected daily with either ISO (Sigma-Aldrich, USA; 0.3 mg kg⁻¹n= 6) or 0.9% saline (control, n= 4). An osmotic minipump was surgically implanted subcutaneously into each rat previously anaesthetized with pentobarbital. In the ISO group (n= 10), ISO was delivered at a rate of 4.2 mg kg⁻¹ day⁻¹. In the control group (n= 10), the pumps were filled with 0.9% saline. Cardiomyopathy, manifested by cardiac hypertrophy and reduced functional performance, develops within 28 days of ISO administration (Deshaies et al. 1981). LV hypertrophy and decreased shortening in ISO rats were confirmed by echocardiography as previously described.

Statistics

The results are presented as means ±s.e.m. Statistical differences were tested using ANOVA (P < 0.05).

Results

Effects of db-cAMP on voltage-dependent Ca²⁺ release in normal myocytes

It has been shown previously that β-adrenergic stimulation results in dramatic alterations in the voltage dependency of SR Ca²⁺ release in rat ventricular myocytes (Piacentino et al. 2000; Viatchenko-Karpinski & Györke, 2001). In the presence of β-adrenergic agonists, depolarization triggers SR Ca²⁺ release at membrane potentials as high as 120 mV, where normally no Ca²⁺ release occurs. This results in the widening and flattening of the bell-shaped, voltage-dependency curve of the amplitude of Ca²⁺ transients (Piacentino et al. 2000; Viatchenko-Karpinski & Györke, 2001). These effects were reported to take place at a constant SR Ca²⁺ content and, importantly, require the presence of Na⁺ in the extracellular solution. This suggested the potential involvement of NCX exchange (Piacentino et al. 2000; Viatchenko-Karpinski & Györke, 2001). In accordance with these previous findings, Ca²⁺-transient voltage dependence became widened on application of the membrane-permeable cAMP analogue, db-cAMP (Fig. 1A and B). Also consistent with our previous report, we were able to obtain these effects upon SR Ca²⁺ release at relatively low levels of PKA stimulation (300 μm db-cAMP), which resulted only in modest increases in Ca²⁺current (I_Ca) (<20%; Fig. 1B). Following the addition of a specific inhibitor of the reverse-mode NCX, KB-R7943, the voltage-dependence relationship of Ca²⁺ release assumed its normal bell shape. We note that KB-R7943 at the concentration used did not affect I_Ca. Again, these effects of PKA stimulation were not accompanied by any changes in the SR Ca²⁺ content, as documented by the observation that application of caffeine induced similar-sized Ca²⁺ transients or NCX currents before and after treatment of myocytes with db-cAMP (Fig. 1C). As shown previously, the lack of changes in SR Ca²⁺ content in quiescent myocytes exposed to cAMP may be attributable to the fact that the increase in SR Ca²⁺ uptake is paralleled by an increase in SR Ca²⁺ leak, so that no net gain in the SR Ca²⁺ content is attained (Viatchenko-Karpinski & Györke, 2001; Szentesi et al. 2004). The elevation of SR Ca²⁺ content commonly observed following β-adrenergic stimulation in myocytes undergoing periodic pacing is apparently due to the enhanced Ca²⁺ entry via LCCs (Szentesi et al. 2004). Since the SR Ca²⁺ release at high membrane potentials seemed to rely for its initiation on Ca²⁺ entries through both LCCs and the reverse-mode NCX instead of just one of the Ca²⁺-entry pathways, we refer to it as dual trigger-activated Ca²⁺ release. The experiments below were designed to explore the mechanisms of this newly identified form of release, and how it is influenced by pathological states such as HF.

LCC- and NCX-mediated Ca²⁺ entries in normal myocytes

To directly quantify trans-sarcolemmal Ca²⁺ entries mediated by LCC and NCX in control and db-cAMP-treated myocytes, we carried out measurements of cytosolic [Ca²⁺] using confocal fluorescent microscopy and fluo-3. DelPrincipe et al. (2000) demonstrated that fluorescence measured in the presence of 0.1 mm fluo-3 is a linear function of the amount of Ca²⁺ entry determined by integration of I_Ca in rat ventricular myocytes. Since fluo-3 fluorescence is proportional to the number of indicator molecules that bind Ca²⁺, it can be used to measure Ca²⁺ influx (Zhou & Bers, 2000; DelPrincipe et al. 2000). In these experiments, myocytes were treated with ryanodine and thapsigargin to disable the SR as a Ca²⁺-storage site. As shown in Fig. 2A, depolarization to membrane potentials ranging from −30 to 150 mV induced a measurable Ca²⁺ entry into the cytosol under reference conditions. The voltage dependency of the peak amplitude of the [Ca²⁺] inflow signal (Fig. 2B, squares) followed that of I_Ca and SR Ca²⁺ release in myocytes with intact SR. Application of db-cAMP resulted in dramatic increases in [Ca²⁺] entry at all membrane potentials, and especially at highly positive membrane potentials (>60 mV). In contrast, only a relatively small amount of Ca²⁺ entry was detected in the absence of db-cAMP (Fig. 2B, circles). Thus, the voltage dependency of Ca²⁺ entry paralleled the voltage dependency of SR Ca²⁺ release in cAMP-stimulated myocytes. These results support the hypothesis that the effects of cAMP on SR Ca²⁺ release could be due to changes in the properties of the Ca²⁺ trigger signal rather than the result of possible changes in the properties of the mechanism for Ca²⁺ release from the SR.

To further explore interactions between LCCs and NCX in cAMP-stimulated myocytes, we measured separately the Ca²⁺ entries mediated by these Ca²⁺-transport mechanisms. The LCC- and NCX-mediated components of the Ca²⁺ entry were separated pharmacologically by using the specific inhibitors of LCCs and reverse-mode NCX, nifedipine and KB-R7943, respectively (Fig. 3). KB-R7943 did not affect I_Ca under the conditions of our experiments (Fig. 1). Nifedipine is a selective blocker of LCCs, with no reported side-effects on NCX at concentrations of the drug used in our study (Shen et al. 2000). Thus the effects of these agents upon NCX and LCCs could be considered indeed specific. Under reference conditions, depolarization to 90 mV induced a relatively large total Ca²⁺-inflow signal (ΔF_Tot), and smaller signals in the presence of KB-R7943 or nifedipine (ΔF_LCC and ΔF_NCX, respectively; Fig. 3A, left panel). Simultaneous application of both compounds eliminated entirely the Ca²⁺-inflow signal, suggesting that the ΔF_Tot was entirely due to Ca²⁺ passing through LCCs and NCX. The averaged peak fluorescence signals for different experimental conditions are plotted in Fig. 3B (grey bars). The sum (ΔF_Tot) of the KB-R7943- and nifedipine-insensitive components, ΔF_LCC and ΔF_NCX, respectively (black bar inside grey bar), corresponded to the total Ca²⁺ entry, suggesting that these two transport systems operate independently of each other in the absence of adrenergic stimulation. Following application of cAMP, ΔF_Tot increased ∼threefold (Fig. 3A, right panel, and Fig. 3B, hatched bars). At the same time, the KB-R7943-insensitive entry increased only by ∼50%, and no significant changes were detected in the amplitude of the nifedipine-insensitive Ca²⁺ entry. Consequently, the magnitude of the total Ca²⁺ signal was threefold higher than the arithmetic sum of the NCX- and LCC-mediated components (Fig. 3B, black bar inside hatched bar). Thus, Ca²⁺ entries via LCCs and NCX can sum in a nonlinear fashion implying that a synergistic interaction between the two Ca²⁺-entry systems takes place in myocytes treated with db-cAMP.

ISO-induced model of HF

HF is characterized by impaired intracellular Ca²⁺ signalling. In order to study whether the synergistic interactions between LCCs and NCX are altered in HF, we performed experiments using two different models of this disease state: HF induced by chronic administration of ISO, and HF induced by chronic volume overload. In myocytes from ISO-treated rats, the amplitudes of Ca²⁺ transient were reduced, and decay rates slowed under reference conditions (absence of cAMP) (Fig. 4A and B; Table 1). Additionally, the SR Ca²⁺ content was reduced to ∼80% of control (Fig. 4C, Table 1). Such alterations in Ca²⁺ transients are typical of HF (Yao et al. 1998; Houser et al. 2000; Piacentino et al. 2000). Application of db-cAMP to HF myocytes did not lead to the broadened Ca²⁺-voltage dependency observed in normal myocytes (Fig. 4B, circles). Since SR Ca²⁺ release is highly sensitive to the SR Ca²⁺ content, we considered the possibility that the absence of release at positive membrane potentials was due to the reduced SR Ca²⁺ content in HF myocytes. Therefore, we increased the SR Ca²⁺ content in HF myocytes by using a loading protocol (see Methods). This protocol allowed us to increase the SR Ca²⁺ content of HF myocytes to a level similar to that in normal myocytes (Fig. 4C). However, as shown in Fig. 4B (triangles), increasing the SR Ca²⁺ content had only a minimal effect on the voltage dependence of SR Ca²⁺ release. Thus, the loss of Ca²⁺ release induced by a combination of Ca²⁺ triggers in db-cAMP stimulated HF myocytes was not simply due to changes in the SR Ca²⁺ content. To further explore the effects of HF on LCC–NCX interactions, we performed fluorometric Ca²⁺-entry measurements similar to those carried out in myocytes from normal hearts. In HF myocytes, the LCC-mediated Ca²⁺ signal remained unchanged, but the total, and especially the NCX-mediated, Ca²⁺ entries were reduced when compared with normal myocytes (Fig. 5). As in normal myocytes, stimulation by db-cAMP increased the LCC-mediated Ca²⁺ entry, but had no significant effect on the NCX-mediated Ca²⁺ signal. However, unlike in normal myocytes, the total Ca²⁺ influx did not change significantly on addition of db-cAMP, and it continued to correspond to the sum of the NCA and LCC components (black bars). Thus, nonlinear summation of Ca²⁺ entries mediated by LCC and NCX appears to be lost in the ISO model of HF. This could be associated with the reduced NCX activity in myocytes from ISO-treated rats.

ACS model

The results obtained in myocytes isolated from hearts of animals with overload-induced HF are shown in Fig. 6. Similar to the results obtained with the ISO model (Fig. 5; Table 1), the amplitude of Ca²⁺ transients was reduced and their decay rate slowed in overload-induced HF myocytes. Also, similar to the ISO model, db-cAMP failed to induce Ca²⁺ release at positive membrane potentials. The normal bell-shaped relationship between membrane potential and the amplitude of Ca²⁺ transients was unchanged, even when the SR content was increased by using a prepulse protocol. Interestingly, as indexed by the amplitude of the NCX current, NCX activity was increased in overload-induced HF myocytes, rather than decreased as was observed in the ISO-induced HF myocytes (Fig. 6C, Table 1). This suggests that the loss of synergistic interactions occurs regardless of changes in overall NCX activity in HF.

Discussion

In this study, we investigated the respective roles of LCCs and reverse-mode NCX in promoting Ca²⁺ entry to trigger SR Ca²⁺ release in myocytes from normal and diseased hearts. Our principal finding is that Ca²⁺ entries via LCCs and NCX interact in a synergistic manner resulting in a larger total Ca²⁺ inflow than the sum of entries mediated by each of these Ca²⁺ transport mechanisms individually. This synergism between the two Ca²⁺ entry pathways was dependent on stimulation of PKA, and it led to the activation of SR Ca²⁺ release at highly positive membrane potentials where no or only minimal Ca²⁺ release occurred in the absence of PKA stimulation. The nonlinear summation of Ca²⁺ triggers and the associated Ca²⁺-release component was abolished in the two experimental models of HF (ISO-induced and overload-induced). Our results provide new insights into the mechanisms involved and the modulation of cardiac EC coupling in normal and failing hearts.

Mechanism of nonlinear summation of Ca²⁺ entries via NCX and LCCs

The precise molecular mechanism of synergism between Ca²⁺ entries through LCCs and NCX is not known. Both transporters are localized in the T-tubular membrane of cardiac myocytes (Frank et al. 1992; Kieval et al. 1992; Carl et al. 1995; Sun et al. 1995). While LCCs are restricted mainly to the dyads, the precise localization of NCXs is less certain. Thus, Scriven et al. (2000), using immunofluorescence and wide-field microscopy, have concluded that NCXs, although located in T-tubules, are excluded from the dyads. However, more recently, Thomas et al. (2003), using a new antibody to NCX in combination with confocal imaging and electron microscopy, found that RyRs, LCCs and NCX are closely associated in the dyadic region. Such colocalization of LCC and NCX would make it possible for communication between these Ca²⁺ transporters. One possibility is that the crosstalk between LCCs and NCX is mediated by Ca²⁺. Haworth et al. (1991) and Matsuoka et al. (1995) have shown that NCX can be activated by Ca²⁺ entry through LCCs, presumably by binding to the cytosolic, catalytic Ca²⁺-activation site on the exchanger. According to several recent studies, the affinity of the cytosolic Ca²⁺ activation sites is in the range of 300–600 nm (Levitsky et al. 1994; Reeves & Condrescu, 2003) (although some studies reported lower values, e.g. Miura & Kimura, 1989; Weber et al. 2001). This would imply that NCX is only partially activated in resting myocytes. Ca²⁺ entry through the LCCs on depolarization could enhance Ca²⁺ binding to the NCX regulatory sites to increase the Ca²⁺ inflow via the exchanger and, hence, result in stronger Ca²⁺-release activation. Another possibility is that the synergism between LCCs and NCX involves a Ca²⁺-dependent signalling pathway, such as protein phosphorylation by CAMKII. It has been reported that NCX is positively modulated by CAMKII (Isosaki et al. 1994). Ca²⁺ entry through L-type channels could enhance NCX-mediated Ca²⁺ entry by promoting CAMKII-dependent phosphorylation of NCX, again resulting in an increased net Ca²⁺ entry into the cell. In both of these scenarios, activation of NCX by Ca²⁺ would be expected to be greater after stimulation of I_Ca by cAMP, potentially accounting for the role of PKA stimulation in the observed effects. Clearly, more studies are needed to determine the mechanisms responsible for the synergism between Ca²⁺ entry through LCCs and NCX.

Activation of SR Ca²⁺ release

While our study demonstrated that synergistic interactions between LCCs and NCX resulted in an increased total Ca²⁺ entry, it is not clear whether this increased Ca²⁺ entry by itself was solely responsible for SR Ca²⁺ release activation at positive membrane potentials. Indeed, several additional processes in the junctional SR could contribute to activation of SR Ca²⁺ release. For example, pre-elevation of Ca²⁺ at the cytosolic face of the RyR has been shown to increase RyR activation probability in response to a subsequent Ca²⁺ stimulus by displacing Mg²⁺ from the RyR activation sites (Zahradnikova et al. 2004). Recordings of single RyR channel activity in response to isolated and paired photolytically induced Ca²⁺ stimuli revealed that RyR opens only when all four of its Ca²⁺-activation binding sites are occupied by Ca²⁺, and blocking any of these sites by Mg²⁺ prevents channel opening. Thus, an initial Ca²⁺ entry to the dyadic space via the L-type channels could sensitize RyRs for a subsequent activation by NCX-mediated Ca²⁺ entry by replacing Mg²⁺ at the activation sites. PKA phosphorylation of RyR and/or other proteins of the junctional complex could also contribute to dual activation of Ca²⁺ release by LCCs and NCX by increasing the functional activity of RyRs in cAMP-treated myocytes (Marx et al. 2000; Viatchenko-Karpinski & Gyorke, 2001; Lindegger & Niggli, 2005). However, the issue of effects of phosphorylation and dephosphorylation on RyR function remains controversial (Bers, Eisner & Valdivia, 2003; Terentyev et al. 2003).

Implication for EC coupling

The dual activation of release by LCCs and NCX occurred at positive membrane potentials of 30 mV and above. Therefore, it would be expected to operate at the peak of the action potential (40–50 mV) in cardiac myocytes. In our experiments, the synergistic interactions between LCCs and NCX required at least some degree of activation of the PKA pathway. Previous studies, however, provided evidence for synergism between NCX and LCCs in activation of Ca²⁺ release, even in the absence of PKA stimulation, but usually under conditions that stimulate reverse-mode NCX, such as elevated cytosolic [Na⁺] (Litwin et al. 1998; Vila Petroff et al. 2003). Thus, although PKA stimulation may not be essential to the observed interplay between LCCs and NCX, it can clearly facilitate this process and the resultant Ca²⁺ release. It is unclear whether endogenous cytosolic cAMP levels in vivo would be sufficient for dual activation of Ca²⁺ release by LCCs and NCX to the extent observed in this study in myocytes treated with db-cAMP. It is tempting to speculate that this form of release acts in cardiac muscle as a reserve mechanism that is engaged during adrenergic stimulation. In any event, our study, together with those referenced, suggest that the trigger for SR Ca²⁺ release in cardiac muscle may be more complex than commonly envisioned, and elucidation of interactions between LCCs and NCX is needed for complete understanding of cardiac EC coupling.

Implication for HF

One of the most important results of our study is that the nonlinear summation of Ca²⁺ entries and the resulting Ca²⁺ release was absent in myocytes from hearts from two different HF models. In the case of ISO-induced HF, this result was not surprising considering the reduction in NCX activity found with this model. However, in the case of overload-induced HF, NCX activity was markedly increased. The loss of dual activation of release by LCCs and NCX that occurred despite the enhanced overall NCX Ca²⁺-transport activity observed in this model appears paradoxical. Several types of explanations can be proposed for this unexpected result. One possibility is that degradation of the T-tubular system in HF myocytes disrupts the crosstalk between NCX, LCC and RyR. HF is generally accompanied by derangement of the ultrastructural architecture of the myocyte membrane system, including reduction of the T-tubular system (Balijepalli et al. 2003). Thus, despite the increased overall number of NCXs, a potential change in distribution of this protein in the sarcolemma could disrupt the crosstalk among NCX, LCC and RyR in the dyadic region required for Ca²⁺-release initiation. Another possibility is that the cytosolic Ca²⁺ regulatory site of NCX involved in the summation phenomenon becomes altered in HF as a result of either altered interactions of the exchanger with other proteins or its post-translational modification. Finally, it is conceivable that downregulation of the β-adrenergic signalling pathway, seemingly required for nonlinear summation of triggers in control myocytes, contributes to the loss of this phenomenon in HF myocytes. It is to be noted, however, that, in our experiments, we activated PKA directly with a membrane-permeable form of cAMP. Thus alterations in the cascade would have to be downstream of cAMP. Even though we do not as yet know why dual trigger-activated Ca²⁺ release does not operate in myocytes from failing hearts, it is reasonable to hypothesize that the loss of this process could account for, or contribute to, the reduced contractile performance and diminished contractile reserve in HF.

Summary

Our results demonstrate that in cAMP-stimulated cardiomyocytes, Ca²⁺ entries via LCCs and reverse-mode NCX interact in a synergistic manner resulting in a larger total Ca²⁺ entry than the sum of entries mediated by each of these transport systems individually. This compound Ca²⁺-entry signal is at least partially responsible for the induction of SR Ca²⁺ release at highly positive membrane potentials where no Ca²⁺ release occurs in the absence of PKA activation. The exact linkage between the two Ca²⁺-entry pathways and its dependence on PKA activation remains to be determined. The demonstration of the interactive relationship between the LCCs and NCX in cAMP-stimulated myocytes provides new insight into the underlying mechanisms of the positive ionotropic effects of catecholamines in the heart. Further, the finding that this interactive mechanism is nonoperational in myocytes from animals in HF suggests at least a partial explanation for the decreased contractile force as well as the inability of increased adrenergic stimulation to produce the necessary compensatory increase in the force of contraction in the failure state.