Introduction

An increased risk of heart diseases, such as cardiomyopathy and heart failure, is a recognised complication of diabetes. Moreover, insulin exerts a positive inotropic effect (an increase in the strength of muscular contraction) on the normal whole heart, which is independent of its actions on glucose metabolism [1, 2, 3]. Understanding the physiological mechanisms by which insulin modulates cardiac contraction is therefore important for the development of treatment strategies for heart patients with and without diabetes. Since it is widely accepted that cardiac contraction depends on Ca2+ flux into cardiac myocytes, the inotropic effect of insulin may involve a modulation of transsarcolemmal Ca2+ movement [2]. The precise mechanism underlying the effect of insulin on Ca2+ fluxes during excitation–contraction is not known, but it could involve modulation of the activity of L-type Ca2+ channels. In studies conducted at hypothermic temperatures, insulin was found to stimulate cardiac L-type Ca2+ channel current (ICa,L) [4, 5]. The present study was undertaken to determine the effects of insulin on ICa,L at 37°C and on contraction in isolated guinea pig ventricular myocytes.

Materials and methods

Myocyte isolation and application of insulin

Male guinea pigs (400–600 g) were humanely killed using a UK Home Office-approved ‘Schedule 1’ procedure. Ventricular myocytes were isolated using an enzymatic dispersal method as previously described [6], then placed in a recording chamber and superfused with Tyrode’s solution containing 140 mmol/l NaCl, 5 mmol/l HEPES, 10 mmol/l glucose, 4 mmol/l KCl, 2.5 mmol/l CaCl2 and 1 mmol/l MgCl2 (pH adjusted to 7.45 using NaOH). Aristar grade chemicals were used (BDH, Poole, Dorset, UK). Insulin (bovine pancreas; Sigma-Aldrich, Gillingham, Dorset, UK) was added to this solution from a 35-mmol/l stock solution made up in water, adjusted to pH 2 with HCl. Prior to recording, cells were pre-incubated for 10 min in normal Tyrode’s solution alone (Control), or in insulin-containing Tyrode’s solution ([Insulin], with insulin also present throughout recording). An insulin concentration of 1 µmol/l was used in all experiments presented in the ‘Results and discussion’, except when concentration-dependent modulation of ICa,L was studied (Fig. 1e), for which a range of concentrations between 100 nmol/l and 1 µmol/l were tested.

Voltage clamp technique

Whole-cell patch clamp experiments were performed using an Axopatch 200A amplifier (Axon Instruments, Union City, Calif., USA). The methods for preparing patch pipettes and for acquiring and processing data have already been described [7].

L-type Ca2+ current (ICa,L) measurement

A Cs+-based internal dialysis solution was used for selective ICa,L recording. Its composition was as follows: 113 mmol/l CsCl (to block outward K+ currents); 5 mmol/l K2ATP; 0.4 mmol/l MgCl2; 2 mmol/l K2EGTA (to block Ca2+-activated currents); 5 mmol/l glucose; 10 mmol/l HEPES. The solution was titrated to pH 7.2 using CsOH. After the whole-cell configuration had been attained, K+-free Tyrode’s solution was applied to inhibit inwardly rectifying K+ current. As described previously, ICa,L was elicited by a two-step voltage protocol [7]. A prepulse from −80 mV to −40 mV, to activate and then inactivate the fast Na+ current, was followed by a test pulse to +10 mV to elicit ICa,L (records in Fig. 1 show only the current during the +10 mV test pulse). Voltage-dependent activation and inactivation plots for ICa,L were obtained using the protocols shown schematically in Fig. 1g, reconstructing mean activation and inactivation relations as described previously [8].

Measuring unloaded cell shortening of isolated ventricular myocytes at 37°C

For this part of the study we used a video-based edge detection system (Crescent Electronics, Sandy, Utah, USA). For experiments on undialysed cells, myocytes were field stimulated at 0.2 Hz. Where cell shortening and ionic currents were measured concurrently, patch pipettes were filled with a ‘physiological’ K+-based pipette solution of 113 mmol/l KCl, 10 mmol/l NaCl, 5 mmol/l K2ATP, 0.4 mmol/l MgCl2, 5 mmol/l glucose and 10 mmol/l HEPES. The solution was titrated to pH 7.2 with KOH. Lidocaine (200 µmol/l) was present in the external solution to block any residual Na+ current.

Data analysis

Data are expressed as means ± SEM. Numbers of cells from which particular observations were obtained are given in parentheses in the ‘Results and discussion’ or in the figure legends. Statistical comparisons were made using the Student’s t test or analysis of variance where appropriate. Statistical significance was accepted as p being less than 0.05.

Results and discussion

Figures 1a and b show representative ICa,L records and mean data in the absence and presence of 1 µmol/l insulin at room temperature (18–21°C). These data confirm previous reports [4, 5], showing an increase in ICa,L magnitude with insulin (t test; p<0.05). Figures 1c and d show corresponding data at 37°C. Surprisingly, and in marked contrast with the room temperature data, at 37°C insulin decreased the magnitude of ICa,L (p<0.05). This difference in insulin action between room temperature and 37°C was also observed when ICa,L was normalised to cell size (membrane capacitance). The inhibitory effect of insulin at 37°C was concentration dependent (Fig. 1e). The significant inhibitory effect (p<0.05) at a relatively low insulin concentration (100 nmol/l) suggests that this finding has physiological and pathological relevance. Half-maximal activation voltage and slope factor parameters derived from Boltzmann fits to current–voltage plots [7] were −11.25±2.29 mV and 7.55±1.01 mV respectively in controls (n=5), and −11.10±1.36 mV (p>0.1 vs control) and 4.56±0.32 mV (p<0.05 vs control) respectively in 1 µmol/l insulin (n=5). For voltage-dependent inactivation, half-maximal inactivation voltage and slope factor values were −30.60±1.58 mV and 7.10±1.13 mV respectively in controls (n=5), and −27.53±2.27 mV and 9.03±1.05 mV (both p>0.05 vs control) respectively in 1 µmol/l insulin (n=6). These mean values were used to produce the activation and inactivation plots in Fig. 1f [7], where the relations for control and insulin tests are closely overlaid. Thus, in contrast to previous data from vascular smooth muscle cells [8], the decrease in cardiac ICa,L with insulin did not correlate with shifts in voltage-dependent ICa,L kinetics.

Fig. 1a–f
figure 1

Effect of insulin on ICa,L. Data in bar charts are means ± SEM. *p<0.05 vs control. (a) Representative traces of peak ICa,L measured at room temperature. (b) Stimulation of ICa,L at +10 mV by insulin at room temperature (n=7 for each group). (c) Representative traces of peak ICa,L measured at 37°C. (d) Inhibition of ICa,L at 10 mV by insulin at 37°C (control, n=22; insulin, n=10). (e) Concentration dependence of insulin effect on ICa,L at 37°C (control, n=22; 100 nmol/l insulin, n=5; 500 nmol/l insulin, n=7; 1000 nmol/l insulin, n=10). (f) Steady-state voltage-dependent activation (left) and inactivation (right) curves without (■) or with (□) insulin. Half-maximal voltage and slope factor values for activation and for inactivation were derived from experiments with protocols shown in insets and were then used to simulate activation and inactivation variables at 2 mV intervals

Full size image

It is well known that ICa,L provides the main source of sarcolemmal Ca2+ entry for the process of calcium-induced calcium release during cardiac excitation–contraction coupling [9]. Intriguingly, a previous study has shown no positive inotropic effect of insulin on cultured normal ventricular myocytes at 30°C [10]. Figure 2a shows that cell shortening of field-stimulated, undialysed myocytes at 37°C was greater in the presence of insulin (1 µmol/l). Mean control shortening was 19.14±1.55% resting cell length (n=8); shortening with insulin was 27.47±0.97% resting cell length (n=8). The positive inotropic effect observed under these conditions clearly indicates a direct inotropic effect of insulin on cardiac myocytes, independent of any effects of insulin on other cardiovascular tissues (e.g. the microvasculature). Further experiments were performed on patch-clamped cells to enable joint measurement of cell shortening and ICa,L. Figure 2b shows representative recordings under control and insulin conditions, whilst Figs. 2c and d show mean data for cell shortening and ICa,L amplitude respectively. In absolute terms, cell shortening from patch-clamped cells was less than that observed from field-stimulated cells, possibly as a result of equilibration of intracellular contents with the pipette dialysate in these experiments. Nevertheless, cell shortening in dialysed cells was increased by insulin, whilst ICa,L amplitude was decreased (Fig. 2d).

Fig. 2a–d
figure 2

Effect of insulin on heart cell shortening (contraction) at 37°C. Data in bar charts are means ± SEM. *p<0.05 vs control. (a) The effect of 1 µmol/l insulin on shortening (downward deflections) of externally stimulated myocytes. (b) Combined measurements of cell shortening (upper traces) and ionic current (middle traces) under voltage clamp (lower traces) from a control myocyte and a myocyte incubated in insulin. (c) Data regarding stimulation of cell shortening in dialysed cells by insulin (n=6 for each group). (d) Data for ICa,L amplitude at +10 mV in control myocytes and in myocytes incubated in insulin (same cells as in c; n=6 for each group)

Full size image

Three main conclusions can be drawn from this study. Firstly, cardiac effects of insulin observed at ambient temperature cannot be extrapolated to the physiological situation at 37°C. This conclusion may also be relevant to the interpretation and design of studies of insulin action on other tissues and organs. Secondly, the signalling pathways downstream of the insulin receptor and/or the interaction between insulin and its receptor in cardiac myocytes are sufficiently temperature sensitive for the net effects of insulin receptor activation on whole-cell ICa,L to be opposite at room temperature from those at 37°C. The stimulatory effect of insulin on ICa,L at room temperature reported previously [4, 5] was attributed to a stimulation of cyclic adenosine monophosphate-dependent protein kinase. Clearly, the pathways responsible for the inhibitory effect of insulin on ICa,L at 37°C must now be elucidated. Thirdly, the positive inotropic action of insulin, reported in the intact normal heart, is the result of direct action of insulin on myocytes, which cannot be explained by a stimulation of Ca2+ entry via ICa,L. Rather, the overall ‘gain’ of the excitation–contraction coupling process was increased. Further investigations are warranted in order to determine which steps of the process are modulated by insulin to account for the observed inotropic effect at physiological temperature. These will probably need to combine the techniques used in the present study with measurements of intracellular Ca2+ to determine: (i) the effects of insulin on intracellular Ca2+ handling (particularly by the intracellular stores in the sarcoplasmic reticulum) and on myofilament Ca2+ sensitivity; and (ii) whether and how the ICa,L and contraction of diabetic and normal hearts differ in their response to insulin. Collectively, the present data and the proposed future work will provide information valuable to the rational development of strategies to protect and treat the myocardium in diabetic and non-diabetic patients.