J Muscle Res Cell Motil (2006) 27:297–306 DOI 10.1007/s10974-006-9089-y
ORIGINAL PAPER
Effect of changes in action potential spike configuration, junctional sarcoplasmic reticulum micro-architecture and altered t-tubule structure in human heart failure M. B. Cannell Æ D. J. Crossman Æ C. Soeller
Received: 17 May 2006 / Accepted: 5 July 2006 / Published online: 4 August 2006 Springer Science+Business Media B.V. 2006
Abstract Using a Monte–Carlo model of L-type Ca2+ channel (DHPR) gating, we have examined the effect of changes in the early time course of the action potential as seen in human heart failure on excitation contraction coupling. The time course of DHPR Ca2+ influx was coupled into a simple model of sarcoplasmic reticulum Ca2+ release. Our model shows that the loss of the initial spike in human heart failure should reduce the synchrony of Ca2+ spark production and lead to the appearance of late Ca2+ sparks and greater nonuniformity of intracellular Ca2+. Within the junctional space of the cardiac dyad, a small increase in the mean distance of a DHPR from a RyR results in a marked decrease in the ability of the DHPR-mediated increase in local [Ca2+] concentration to activate RyRs. This suggests that the efficiency of EC coupling may be reduced if changes in micro-architecture develop and such effects have been noted in experimental models of heart failure. High resolution imaging of t-tubules in tachycardia-induced heart failure show deranged t-tubule structure. While in normal human hearts t-tubules run mainly in a radial direction, t-tubules in the heart failure samples were oriented more toward the long axis of the cell. In addition, t-tubules may become dilated and bifurcated. Our data suggest that changes in the micro-architecture of the cell and membrane structures associated with excitation–contraction coupling, combined with changes in early action potential configuration can reduce the efficiency by which Ca2+
M. B. Cannell (&) Æ D. J. Crossman Æ C. Soeller Department of Physiology, Faculty of Medicine and Health Sciences, University of Auckland, 85 Park Road, Grafton, Auckland, New Zealand e-mail:
[email protected]
influx via DHPRs can activate SR calcium release and cardiac contraction. While the underlying cause of these effects is unclear, our data suggest that geometric factors can play an important role in the pathophysilogy of the human heart in failure. Keywords Heart failure Æ t-tubule Æ Calcium Æ Action potential Æ EC coupling Æ DHPR Æ SR Æ Human
Introduction Excitation–contraction (EC) coupling hinges on Ca2+induced Ca2+ release (CICR) (Fabiato 1985) wherein a small Ca2+ influx due to L-type Ca2+ channels triggers a larger Ca2+ release from the sarcoplasmic reticulum (SR) (for review see Bers 2002). During EC coupling, Ca2+ sparks (Cheng et al. 1993; Cannell et al. 1994) are evoked by the local increase in [Ca2+] produced by L-type Ca2+ channels (or dihydropyridine receptors—DHPRs; Lew et al. 1991) in the sarcolemma. As might be expected, these evoked sparks occur at z-lines (Shacklock et al. 1995) where t-tubules enable rapid propagation of excitation into the cell interior (Cheng et al. 1994). The amount of Ca2+ released depends not only on the SR Ca2+ content but also on the ability of the DHPR to activate ryanodine receptors (RyRs) within the junctional regions of the SR. The discovery of Ca2+ sparks has shown that EC coupling must be considered a ‘‘local control’’ phenomenon where the local micro-environment and micro-architecture will be critical factors in determining the properties of macroscopic (whole cell) EC coupling (Stern 1992; Cannell et al. 1995; Lopez-Lopez et al. 1995).
123
298
In heart failure, there is an inability of the heart to meet the body’s requirements for blood flow without bringing systemic compensatory mechanisms into play (such as an increase in filling pressure). The myocyte origin of decreased contractile performance has been extensively studied and contributory factors are highly variable among animal models and human tissue samples. As a major determinant of contractility, abnormal Ca2+ regulation might be expected to play a role in heart failure. This expectation has been met in some human studies (Beuckelmann et al. 1992; Kubo et al. 2001; Piacentino et al. 2003), but the underlying changes and their cause remain unclear. In addition, in animal models, disparate findings occur. For example, in the hypertensive Dahl salt sensitive rat model, Gomez et al. (1997) found a functional mismatch between L-type Ca2+ current (ICa) density and the release of Ca2+ by the SR. A recent study on SHR also showed a reduced EC coupling efficiency, which was linked to disruption in t-tubular structure and physical delocalization of DHPRs and RyRs (Song et al. 2006). On the other hand, Shorofsky et al. (1999) found increased SR release without changes in SR load or ICa density. Ward et al. (2003) found that Ca2+ transients recorded from an intact muscle preparation (trabeculae) from SHR in heart failure appeared nearly normal, despite a large reduction in contractile strength. The latter authors also found a marked change in collagen structure, suggesting that the defect was not in the myocytes per se but in an ability of the myocytes within the trabeculae to develop external force. In connection with this point, it has been suggested that changes in expression of matrix proteins may provide a ‘‘fingerprint’’ for human failure (Tan et al. 2002). Nevertheless, it is clear from several studies that there may be large changes in Ca2+ handling in some types of human heart failure (for reviews see Houser et al. 2000; Hasenfuss and Pieske 2002; Birkeland et al. 2005; Rodriguez and Kranias 2005; Taur and Frishman 2005; Yano et al. 2006). Recent large scale gene array profiling also suggests that down regulation of Ca2+ signalling pathways may play a role (Barrans et al. 2002). However, it remains unclear whether these changes are common in dilated cardiac myopathies or whether they may be secondary to some other changes such as a deranged myocyte and tissue micro-architecture. Evidence for the possible disruption of structures associated with EC coupling in heart failure is quite limited. Although changes in the gross structure of the dilated myocardium are well established (e.g. Norton et al. 2002), with wall remodelling and myocyte hypertrophy being common findings, how the myocytes may respond to the consequent changes in mechanical
123
J Muscle Res Cell Motil (2006) 27:297–306
stress is largely unknown. Page and McCallister (1973) noted the appearance of enlarged t-tubules in a rat model although there were no changes in overall surface to volume ratio. In a dog model of tachycardiainduced heart failure (He et al. 2001) there was a ~25% reduction in t-tubule density as well as a decrease in ICa. While the reduction in ICa in this model would be expected to directly reduce the strength of EC coupling, current views of EC coupling suggest that it is not the absolute magnitude of ICa which is important but rather how that influx is linked to the activation of local RyRs. In connection with this point, Song et al. (2006) suggested that, while ICa density was not changed in their SHR model of heart failure, the reduction of co-localization of RyRs with DHPRs as well as the observed loss of t-tubules might explain the reduced Ca2+ transient they observed. Another change commonly observed in heart failure is a change in action potential configuration. In human failure, there is an alteration in the ‘‘spike and dome’’ morphology and the action potential becomes longer (Beuckelmann et al. 1993; Tomaselli et al. 1994; Piacentino et al. 2003). In the mouse, rapid repolarization associated with the spike may serve to synchronize the activation of Ca2+ sparks by allowing a synchronous increase in Ca2+ entry via DHPRs (Bridge et al. 1999). While the human action potential has a smaller spike and higher plateau, the loss of the spike may have important effects on the time course of CICR activation and Ca2+ sparks. In this paper, we will show that small changes in geometry between DHPRs and RyRs can have profound effects on the ability of a DHPR to activate adjacent RyRs in the junctional SR. In addition, we will present evidence that in human heart failure, remodelling of the t-tubular system can occur and that the change in action potential configuration in heart failure can change the synchrony and strength of EC coupling across the cell. Although the general applicability of these results is unclear, our data emphasize that EC coupling may be altered in a number of subtle ways in heart failure which can be contributory to the problem of reduced contraction strength.
Materials and methods Computer simulation A Markov model of DHPR gating (Imredy and Yue 1994; Iyer et al. 2004) was coupled into a dyad model implemented in Facsimile (Chance et al. 1977). The dyadic space model was similar to that used by Sobie
J Muscle Res Cell Motil (2006) 27:297–306 Table 1 Values of model parameters describing RyR gating and dyadic space properties. Symbols are as defined in (Hinch 2004)
299
sopen KRyR k
1 ms 8 lM 0.2
VJSR VDS gD
10–17 l 10–19 l 1.4 10–13 l/s
et al. (2002) and Hinch (2004) (for parameters see Table 1). The DHPR model was a 12-state Markov gating scheme with voltage dependent transition rates that incorporate Ca2+ dependent inactivation using a mode switching approach (mode-Ca vs. mode-normal, Fig. 1a). Numerical parameters were as given in Iyer et al. (2004). RyR gating was described by a 2-state model (Hinch 2004) derived from a 3-state model with two closed states (see Fig. 1) with Ca2+ dependent transitions between the states but without coupling between receptors (i.e. CF+=CF– = 1 in the terminology of Hinch 2004). The opening rate was described by the equation
RyR open time RyR Ca2+ binding constant RyR open probability at saturating [Ca2+] Volume of junctional SR Volume of dyadic space Diffusive rate from dyadic space into cytoplasm
cleft [Ca2+] time courses were calculated by summation of contributions from all couplons. Whole cell ICa was estimated from the simulated DHPR current density in t-tubule junction and scaled to the whole cell capacitance on the basis of t-tubule membrane parameters for rabbit ventricle (Page and Surdyk-Droske 1979). No adjustment was made for differing DHPR channel density or micro-environment between surface sarcolemma and t-tubules DHPRs in calculating this estimate. A simple spark model similar to that described in Smith et al. (1998) was used to simulate spark records arising from couplons firing in response to DHPR activation. Human t-tubule labelling
kþ ¼
½Ca2þ 4ds ksopen ð½Ca2þ 4ds þ
4 KRyR Þ
ð1Þ
and the closing rate was k– = 1/sopen, for numerical values of these parameters see Table 1. The stochastic gating of DHPRs and RyRs was simulated using a Monte–Carlo approach as described by Stern et al. (1997). Model action potentials were based on the data published by Piacentino et al. (2003) (see Fig. 3a,b) adjusted for 1 Hz heart rate. The action potentials were used to drive the DHPR gating model (see Fig. 1) by incorporating a look up table which provided the integrator the with membrane potential as a function of time [Vm(t)]. To obtain the whole cell behaviour 5,000 (independent) dyad junctions (‘‘couplons’’) were simulated and whole cell current and average
Fig. 1 Gating schemes for DHPR and RyR channels. The channel states and life times were calculated a Monte–Carlo method. Rate constants for the DHPR are as described by Iyer et al. (2004) The RyR rate constants are given in Table 1, see text
Small samples of left ventricular tissue were obtained from a male patient aged 60 with tachycardia-induced heart failure after the diseased heart was removed during cardiac transplantation. Control left ventricular heart samples were obtained from an unmatched donor heart (Male aged 68). Both patients were receiving dopamine infusion for cardiovascular support. The protocols for harvesting tissue were approved by the local ethics committees. The samples were fixed with 4% paraformaldehyde in PBS for 4 h at room temperature. The samples were embedded in agar and a vibratome was used to cut 100 lm thick free floating sections. The sections and remaining agar block were stored in 0.5% BSA, 0.1% NaAzide in PBS at 4C before staining. Sections were washed 3 · 5 min in PBS and then a 1 in 50 dilution of WGA-488 (Molecular Probes) in PBS applied overnight in the cold room on a rocker. A blank section was incubated in PBS only. Sections were then washed 3 · 30 min at room temperature with PBS before mounting on slides with Prolong gold (Molecular Probes) and allowed to set overnight at room temperature. Sections were imaged with an LSM 410 (Zeiss) using 488 nm excitation and fluorescence emission at 535 ± 25 nm. Images were taken with a 63· 1.25NA objective and sampled at 0.1 · 0.1 · 0.3 lm (x, y, z). Results To examine the sensitivity of RyR activation to distance from the DHPR (labelled s in Fig. 2a) within the
123
300
dyad we used the model described in detail in elsewhere (Cannell and Soeller 1997). As shown in Fig. 2b, even a brief DHPR opening is capable of causing a large increase in RyR activation probability (Pa). For RyRs situated opposite a DHPR, the Pa rises to ~0.8 at 0.3 ms and ~0.9 at 10 ms. As the RyR is displaced across the junctional region there is a continuous reduction in Pa, an effect that arises from the marked gradients of [Ca2+] from the open DHPR within the junction (Soeller and Cannell 1997). Even a small displacement of the RyR by 10 nm (about half the width of an RyR) reduces Pa to ~0.6 at 0.3 ms, an effect that is partly offset at later times (due to the slower dissipation of [Ca2+] gradients within the junctional space). In an intact junction with many RyRs, this effect would be offset by the open DHPR
Fig. 2 Panel a shows a two-dimensional cartoon of the junctional geometry. The gap between SR and t-tubule membranes was 15 nm and the junctional region was circular with a diameter of 400 nm. The model included surface charge and electrodiffusion as described by Soeller and Cannell (1997) Panel b shows the activation probability (Pa) for a RyR situated at varying distances (s) from the DHPR. The solid lines show the time course of evolution of Pa for a DHPR opening of 0.3 ms. The dotted lines show Pa for an indefinite DHPR open time. Note that small increases in s at near molecular scale strongly affect Pa
123
J Muscle Res Cell Motil (2006) 27:297–306
becoming closer to another RyR within the junction but this would not be the case for a DHPR located near the edge of the junction. It should be noted that these effects are not related to the RyR gating model as we are simply considering the binding of Ca2+ by the RyR which should be determined by the diffusion-limited RyR on rate, Kd for Ca2+ binding and local [Ca2+]. RyR gating should reduce the overall RyR open probability below that predicted from the Pa calculated here. Nevertheless, a reduction of Pa should directly translate to RyR open probability since the latter should be proportional to Pa. In addition, the model shows that Pa responds quickly to the increase in [Ca2+] produced by a DHPR opening, this is a consequence of both the small diffusion distances involved as well as the large local dyad [Ca2+] produced by the DHPR opening. This model prediction is consistent with the very short (
