Characterization of L-type calcium channel activity in atrioventricular nodal myocytes from rats with streptozotocin-induced Diabetes mellitus

Ethical approval

All experiments were performed in accordance with local Ethical Committee approval, United Arab Emirates University, Al Ain.

Induction of DM

Male Wistar rats received a single intraperitoneal injection of STZ (Sigma, St. Louis, MO) at 60 mg/kg to induce diabetes. STZ was prepared in citrate buffer (sodium citrate/citric acid, 10 mmol/L, pH 4.5) immediately before administration. The induction of a diabetic state was confirmed prior to experimentation by an elevated blood glucose concentration of 440.9 ± 12.3 mg/dL compared to control levels of 97.7 ± 6.2 mg/dL, measured using a glucometer. All animals were maintained for a minimum of 12 weeks on the same diet and water. Age-matched male rats (302.6 ± 9.6 g) were used as control animals. STZ rats had a significantly smaller body mass, 239.2 ± 6.3 g.

Atrioventricular myocyte isolation

Control (n = 14) and STZ-treated diabetic rats (n = 14) were killed by cervical dislocation and decapitated. The heart was rapidly removed from the chest cavity, and placed in ice-cold isolation solution prior to aortic cannulation, and subsequent retrograde Langendorff perfusion (6 mL/min/g heart tissue). The hearts from control animals had a mean weight of 1.17 ± 0.04 g and those from STZ animals weighed 0.96 ± 0.02 g. In the first instance, the hearts were perfused for 2 min with isolation solution containing the following constituents (in mmol/L): NaCl (130), Hepes (10), glucose (5), taurine (20), creatine (5), KCl (4.5), NaH₂PO₄ (1), Na-pyruvate (5), MgCl₂ (5), and CaCl₂ (0.75 μmol/L) (pH adjusted to 7.25 with NaOH, and bubbled with 100% O₂). The hearts were then perfused with calcium-’free’ isolation solution, containing 8 μmol/L Na-EGTA, before switching to isolation solution containing CaCl₂ (150 μmol/L), collagenase type 2 (1 mg/mL, Worthington), protease type X1V (0.1 mg/mL, Sigma), and bovine serum albumin (1%, Sigma) for 6–10 min. Hearts were then removed from the cannula, the AVN region was identified and excised, and AVN myocytes were isolated by enzymatic and mechanical dispersion (Hancox et al. 1993; Yuill et al. 2010). After isolation, cells were stored in a high [K⁺], low [Ca²⁺] storage solution at 4^°C for use on the same day. This contained (mmol/L) l-glutamate (110), KCl (30), Na-pyruvate (5); Taurine (20), creatine (5), succinic acid (5), Na₂ATP (2), β-hydroxybutyrate (5), glucose (9); MgCl₂ (5), EGTA (0.05), HEPES (10), phosphocreatine (1), K₂HPO_4, (5), and had a pH of 7.2.

Drugs and chemicals

Aristar grade chemicals (BDH) dissolved in deionized water (18.4 MΩ) were used for myocyte isolation and pipette solutions, while Analar grade chemicals (BDH) were utilized for external solutions. All drugs were supplied by Sigma-Aldrich.

Data acquisition and analysis

Voltage clamp protocols were generated using PClamp 8 (Molecular Devices). Data were digitized at 2 kHz and recorded online. Data were stored on computer, and analyzed using Clampfit (Molecular Devices), Excel (Microsoft, Redmond, WA) and Origin (Microcal, Northampton, MA) software. Numerical values in the text are shown as mean ± SEM and were compared statistically using the Student's two-way t-test in Excel (Microsoft) and Sigmaplot (Systat, San Jose, CA) software. P values of <0.05 were considered as statistically significant.

The voltage-dependent activation of I_CaL was fitted using the Boltzmann equation:

(1)

where V_m = test potential, V_0.5 is the membrane potential exhibiting half-maximal activation, and k is the slope factor.

The steady-state, voltage-dependent inactivation of I_CaL was described by the equation:

(2)

where V_m = the prepulse potential, V_0.5 = the membrane potential at which inactivation was half-maximal and k describes the slope of the inactivation curve.

The time dependence of current inactivation was described by the following equation:

(3)

where A_f and A_s are the current components described by the time constants τ_f and τ_s.

The time course of recovery from inactivation of I_CaL was described by the equation:

(4)

where y describes the fraction of recovery of I_CaL from inactivation at time t; A and B describe the data fitted by each exponential, and τ₁ and τ₂ are the time constants for each exponential.

Cellular capacitance

The AVN isolation produces a heterogeneous population of cells (Meijler and Janse 1988), so to ensure the identity of the cells included in the study were nodal in origin, we only included cells that exhibited regular spontaneous electrical activity. The cells displayed the characteristic AVN morphology described in previous studies on rabbit, guinea pig, and rat AVN myocytes (Hancox et al. 1993; Yuill and Hancox 2002; Yuill et al. 2010). Cells from the two groups of animals appeared to be similar in size and shape, which is consistent with our previous study in rat AVN that explored the effects of short-term diabetes on AVN function (Yuill et al. 2010). To assess whether there were any differences in cell size between the two populations we measured the cell capacitance for all cells studied. Although there was some degree of variation in cell size within both groups of cells, there was no significant difference (P > 0.4) between mean capacitance in the control group (29.5 ± 1.6 pF; n = 35) and the STZ (30.9 ± 1.4 pF; n = 29) (see Fig. 1B).

Diabetes mellitus reduces peak L-type calcium current

Using selective recording conditions to fully isolate I_CaL, we examined the effects of long-term DM on I_CaL channel characteristics. Inward currents were elicited by sequential step depolarizations, from a holding potential of −40 mV, to more positive test potentials, for a duration of 500 ms. Application of nifedipine (1 μmol/L), a selective L-type calcium channel blocker, abolished the inward current at all test potentials. After prolonged exposure to the diabetic state, I_CaL was clearly and significantly reduced in cells from STZ-treated rats compared to control.

Figure 2A shows representative families of currents recorded in control (Fig. 2Aa) and STZ cells (Fig 2Ab). The currents recorded in the cell from the STZ rat were substantially smaller than those seen in the age-matched control. We constructed current–voltage (I–V) relationships by plotting the peak current elicited at each test potential normalized to cell capacitance (current density), against membrane potential, and the mean values (n ≥ 12) are shown in Figure 2B. Peak I_CaL occurred at 0 mV and this was reduced by 43% in STZ cells compared to control cells (P < 0.01). Figure 2B shows the mean I–V relationships in control and STZ cells, where a significantly smaller I_CaL was observed across a range of test potentials, with significance (*P ≤ 0.05) observed at all potentials between −10 and +20 mV. Steady state activation curves using equation 1, were plotted to examine whether there was any significant effect on activation of the current. The activation curves from the pooled data are shown in Figure 2C. In control cells, the half-activation potential was −10.9 ± 1.8 mV, with a slope factor, k, of 11.1 ± 0.9 mV. This was not significantly different from the STZ cells, where the half-maximal activation was −9.4 ± 1.4 mV, and the slope factor was 10.3 ± 1.2 mV (P < 0.8, n ≥ 12). Thus, the data show that in rats with STZ-induced DM, peak I_CaL density was significantly reduced. However, there was no significant effect on the voltage dependence of activation.

L-type calcium current voltage-dependent inactivation is modified by diabetes mellitus

The steady-state inactivation parameters for I_CaL were then determined by delivering a series of prepulses (1000 ms duration) applied at a frequency of 0.1 Hz, incrementing (10 mV steps) from −90 mV to +20 mV from a holding potential of −80 mV. The conditioning pulses were followed by a test pulse to 0 mV for 500 ms. Families of selected test currents recorded in cells from control and STZ-treated animals are shown in Figures 3A and B. Mean data (Fig. 3C; n ≥ 10) were plotted as peak current measured during the test potential for each prepulse potential, and were fitted to the Boltzmann equation (2). Figure 3C shows a significant (P < 0.01) hyperpolarizing shift in the half-inactivation potential in cells from STZ-treated rats (V_0.5 = −39.6 ± 2.1 mV, k = 8.3 ± 0.6 mV), compared to age-matched controls (V_0.5, −32.3 ± 1.4 mV, k = 7.9 ± 0.8 mV).

Diabetes mellitus alters the time dependence of L-type calcium current inactivation

Examination of the traces illustrated in Figures 2 and 3 suggest differences in the kinetic properties displayed by the families of I_CaL in control compared to STZ cells. To explore this further, the time dependence of inactivation of I_CaL was compared in the two groups. The kinetics of I_CaL have been demonstrated previously to be described by a biexponential function, where a fast component of the current (A_f) can be described by the time constant (T_f), and a slow current component (A_s) can be described by a time constant (T_s) (see equation 3). Figure 4 shows representative currents elicited by a step depolarization from −40 mV to 0 mV in a control cell (Fig. 4A) and a cell from an STZ animal (Fig. 4B). The current decay was fitted with a biexponential function (illustrated by the solid line) to determine the inactivation time constants across a range (−20 to +30 mV) of test potentials. Figures 4C and D show the results of this analysis pooled from eight cells in control animals and 10 from STZ rats. The values for T_f were 5–10-fold smaller than the values for T_s. Both exhibited voltage dependence and became increasingly smaller at more positive test potentials in both age-matched control cells and STZ cells. The amount of current described by each function did not differ significantly over the range of test potentials. However, significant differences in the time constants of inactivation were observed between cells from control and STZ-treated animals, with the latter exhibiting a significant reduction in the slow decay constant, T_s, at test potentials between −20 and 0 mV (P < 0.05). These data demonstrate that in addition to a reduction in peak amplitude of I_CaL, AVN cells from animals with DM exhibit a much more rapid decay overall.

Diabetes mellitus slows the restitution properties of L-type calcium current

The prior experiments established that DM is associated with changes in the time-dependent and voltage-dependent properties of I_CaL inactivation. To complete the analysis, we then studied the effects of DM on recovery from inactivation. Restitution of I_CaL follows a biexponential time course and consists of an initial fast phase (T₁) succeeded by a slower phase (T₂). We measured restitution using a double-pulse protocol, where currents were elicited by an initial step depolarization (P1) from −40 to 0 mV, and subsequent step depolarizations (Pn) elicited with increasing intervals from P1 in variable incremental intervals. Representative current traces for control and an STZ cell are shown in Figures 5A and B, respectively. The recovery from inactivation was calculated from the current amplitude for each incrementing interval, as a fraction of the initial peak current (Pn/P1), and was then plotted against pulse number. The pooled data are shown in Figure 5C. Analysis shows that there was a significant delay in the recovery of inactivation in STZ cells compared to control (P < 0.05). When the mean data were fitted with a biexponential function (Equation 4), the time constants in control cells (T₁ = 10.4 ± 0.9 ms, T₂ = 44.3 ± 1.8 ms) were found to be significantly (P < 0.05) faster than those for DM-affected cells in STZ rats (T₁ = 20.1 ± 1.1 ms, T₂ = 149.2 ± 2.4 ms P < 0.05).

L-type calcium current in rat atrioventricular node cells

Much of what is currently known about the properties of AVN cells is derived from rabbit tissue, and this has been comprehensively reviewed previously (Hancox et al. 2003). Rat AVN has been shown to exhibit the same structural characteristics as rabbit AVN (Stephenson et al. 2012). While there is limited information regarding AVN cellular characteristics available from other animal models, such as guinea pig (Yuill and Hancox 2002) and mouse (Mangoni et al. 2006; Nikmaram et al. 2008; Marger et al. 2011), our previous study in rat (Yuill et al. 2010), reported that the action potential profile and zero current potential, were comparable to prior findings in rabbit AVN (Hancox et al. 1993; Munk et al. 1996; Hancox and Mitcheson 1997; Cheng et al. 2009). However, to date, there has been no detailed characterization, of the ion channel characteristics of rat AVN.

The generation of normal cardiac rhythm and propagation through the AVN, is dependent on the presence of voltage-gated L-type calcium channels, which are heteromeric complexes comprising four subunits (α1, β, α2/δ), and the α1 subunit exists in multiple isoforms. Typically, the two isoforms α1C (Cav1.2) and α1D (Cav 1.3) are expressed in the heart, with Cav1.3 being expressed predominantly in the SAN and AVN regions (Mangoni et al. 2003; Matthes et al. 2004). Cav1.3 has subtle differences in its kinetic profile, and is activated at more negative membrane potentials than Cav1.2, allowing it to contribute to diastolic depolarization (Takimoto et al. 1997).

Here, we have shown the peak I_CaL to be at 0 mV, and half-maximal activation in control cells to be −10.9 mV, which is comparable to previous findings in rabbit (Cheng et al. 2009). However, previous studies in rabbit AVN have shown activation parameters to be ~10 mV more depolarized (Hancox et al. 1993; Cheng et al. 2009). Mouse AVN has been demonstrated to express greater levels of Cav1.3 than Cav1.2, and I_CaL exhibited peak current around −10 mV, with a more hyperpolarized activation profile (Marger et al. 2011).

However, the voltage dependence of I_CaL inactivation parameters that we observed in rat AVN myocytes, were very similar to previous studies in rabbit (Matthes et al. 2004), V_0.5 = −25 mV; V_0.5 = −29.4 mV). We also found that the recovery from inactivation exhibited a biexponential time course, but was significantly faster than previous studies in rabbit (Workman et al. 1999; Cheng et al. 2009). The time-course of current decay exhibited a fast and slow component that were voltage dependent, and these were also similar values to those observed in rabbit AVN (Hancox and Levi 1994), but were significantly faster than AVN myocytes with abnormal morphology (Nakayama et al. 1984). Previously, nonselective recording conditions revealed that L-type Ca²⁺ current was reduced and there appeared not to be a shift in the voltage dependence of activation (Yuill et al. 2010) findings that are consistent with this study where L-type Ca²⁺ current was recorded under more selective conditions. There are clearly subtle differences in the I_CaL channel characteristics reported in AVN myocytes of different species. This could reflect the variation in recording parameters within different laboratories and the heterogeneity of the AVN region used within each study. However, these changes in profile may well be genuine physiological differences and reflect species differences in expression levels of Cav1.3 and Cav1.2 isoforms in AVN myocytes (Zhang et al. 2011). While there are differences between studies, overall, the properties of I_CaL activation, inactivation, restitution, and the time dependence of inactivation are largely comparable between rat and rabbit, suggesting that the rat model of AVN has utility for investigation of cellular electrical function of the region.

Modulation of L-type calcium current by diabetes mellitus

Changes in action potential profile are generated by alterations in sarcolemmal ionic fluxes, and previous studies have indicated that induction of diabetes can modulate a variety of cardiac ion channels, creating a significant impact on cardiac function. Available studies of the effects DM on L-type calcium channel activity in the literature, are largely focused on rat ventricle, and demonstrate a diversity of different findings. This variation may derive from differences in experimental procedures, and implementation and duration of the diabetic state. Some groups suggest that induction of DM has no significant effect on I_CaL amplitude (Choi et al. 2002; Ding et al. 2006) or voltage-dependent activation, inactivation and time-dependent current properties (Arikawa et al. 2002). However, these studies were conducted at room temperature, and this has been shown to profoundly influence L-type calcium channel properties, particularly with respect to diabetes (Pabbathi et al. 2004), and as such inferences drawn from data that was obtained under nonphysiological recording conditions may be questionable. Other studies have also observed no difference in I_CaL amplitude after the induction of DM (Lacombe et al. 2007), and while these experiments were performed at 35°C, they were conducted simultaneously with intracellular calcium measurements, and thus the chelation of intracellular calcium may have influenced the electrophysiological findings by altering calcium-dependent inactivation parameters. Conversely, others have demonstrated an augmentation of I_CaL amplitude in diabetic ventricular myocytes compared to control cells (Jourdon and Feuvray 1993). These experiments were also performed at room temperature, and the ventricular myocytes exhibited vastly depolarized resting membrane potentials (−66 mV), indicating the likelihood of cellular damage resulting from the isolation procedure, and perhaps negating the findings in this study.

There is more general consensus that diabetes inhibits I_CaL amplitude, both in STZ-induced diabetes (Wang et al. 1995; Chattou et al. 1999; Bracken et al. 2006), which is in keeping with our current findings in STZ rat AVN. Previous studies in an insulin-deficient diabetic mice model (Akita mouse) have shown that the peak amplitude of L-type calcium current is reduced compared to control (Lu et al. 2007). Studies of type II diabetes in mouse (db/db; [Pereira et al. 2006; ] and rat (Zucker diabetic fatty; (Howarth et al. 2011), have also demonstrated a significant decrease in I_CaL amplitude.

While these studies agree that I_CaL amplitude is reduced, the effects of diabetes on I_CaL gating characteristics are varied. We saw no change in the activation parameters, which contrasts with findings of an observed rightward shift in the activation variable (Lu et al. 2007). Conversely, the same authors also observed a depolarizing shift in the inactivation of the channel, while in our study, the voltage dependence of inactivation was shifted in the hyperpolarizing direction, suggesting that at physiological voltages, there would be less steady-state current available compared to control. The leftward shift of inactivation, faster inactivation time course, and slower restitution all point to stabilization of the inactivated channel state in DM. The combination of reduced L-type Ca²⁺ current amplitude and altered inactivation/current restitution might be anticipated both to slow AV nodal conduction and delay AV nodal action potential restitution. One of the most striking findings in this study was the difference in the time dependence of current decay between the control and STZ cells. We observed a significant reduction in the fast and slow time constants of current inactivation compared to control. While previous studies have observed changes in the kinetics of I_CaL decay following induction of diabetes, the time course of current decay has been shown to increase (Chattou et al. 1999; Bracken et al. 2006; Lu et al. 2007). It is likely that this difference in channel kinetics may relate to different levels of Cav1.2 and Cav1.3 channel expression between the AVN region and the ventricle, which may be further modified by diabetes.

Furthermore, previous studies have identified that intracellular calcium dynamics are greatly altered in diabetic ventricular myocytes, with a reduction in calcium transient amplitude, a prolonged time course of the calcium transient, depressed calcium loading of the sarcoplasmic reticulum, leading to an increase in basal calcium concentrations (Yaras et al. 2005; Pereira et al. 2006). We postulate that in these conditions, an increase in calcium concentration could facilitate calcium-dependent calcium current inactivation. It is possible therefore that the decrease in the time constants of current inactivation that we observed in the STZ myocytes, may have resulted from modulation of the calcium-dependent component of current inactivation. This is of particular importance, as intracellular calcium homeostasis in nodal myocytes (SAN and AVN) has been shown to contribute to spontaneous electrical activity and automaticity, in concert with sarcolemmal ionic fluxes (Ridley et al. 2008). Any changes in calcium dynamics induced by diabetes could therefore influence action potential generation, fundamentally altering the physiology of the cardiac conduction system. There is, however, little information available in the literature concerning measurement of global and local calcium in diabetic SAN and AVN myocytes, which will be of key importance for future understanding of the effects of diabetes on cardiac function. Our results demonstrate that I_CaL activity is downregulated in AVN myocytes from STZ rats, which is in keeping with previous studies in rat ventricle. While we have not specifically investigated the mechanisms that underlie these findings, we speculate that there are changes in both channel trafficking and modulation of channel activity by regulatory proteins, based on previous work (Lu et al. 2005), which demonstrated that depression of I_CaL activity coincided with a reduction in PI3 kinase signaling. Further studies indicated that in type I diabetic Akita mice, the reduction in I_CaL amplitude compared to controls, proceeded from impaired channel trafficking to the cell membrane mediated via the PI3 kinase signaling pathway (Lu et al. 2007). However, changes in the channel activation and inactivation kinetics were independent of the PI3 kinase signaling pathway, suggesting that multiple mechanisms are involved in regulation of I_CaL by diabetes. Given the changes in both channel amplitude and gating kinetics in our present study, we speculate that similar processes may be responsible for altered I_CaL in rat AVN myocytes from STZ rats.

Clinical significance and conclusions

Atrioventricular nodal dysfunction can play a vital role in a variety of cardiac arrhythmias. I_CaL is responsible for the action potential upstroke in AVN myocytes, and is therefore fundamental to conduction through the region. Depression of I_CaL activity via reduced trafficking to the membrane, or altered channel kinetics, will decrease AVN conduction velocity, and can therefore induce ECG changes, including first to third degree AV block (Watanabe 1981; Watanabe et al. 2012). There is also evidence to suggest that AV block can initiate the development of QT prolongation and of Torsade de pointes (Tsuji et al. 2002), which can result in sudden cardiac death. The reduction in I_CaL seen in this study in AVN myocytes following induction of diabetes is consistent with abnormal AVN function and could contribute to arrhythmogenic cardiac activity. Arrhythmias are a major cause of mortality in patients with diabetes (Nakou et al. 2012), and it is therefore imperative that the underlying cellular pathophysiology is investigated in order to elucidate the causes and potential therapeutic strategies, of cardiac electrical dysfunction in diabetes.