Effect of the microtubule polymerizing agent taxol on contraction, Ca²⁺ transient and L-type Ca²⁺ current in rat ventricular myocytes

Abstract

1. Microtubules form part of the cytoskeleton. Their role in adult ventricular myocytes is not well understood although microtubule proliferation has previously been linked with reduced contractile function.

2. We investigated the effect of the anti-tumour drug taxol, a known microtubule polymerizing agent, on Ca²⁺ handling in adult rat ventricular myocytes.

3. Treatment of cells with taxol caused proliferation of microtubules.

4. In taxol-treated cells there was a reduction in the amplitude of contraction, no significant effect on the amplitude of L-type Ca²⁺ current, but a significant reduction in the amplitude of the Ca²⁺ transient.

5. Caffeine was used to release Ca²⁺ from the sarcoplasmic reticulum (SR). There was a significant reduction in the ratio of electrically stimulated : caffeine-induced Ca²⁺ transients in taxol-treated cells. This observation is consistent with the hypothesis that taxol reduces fractional SR Ca²⁺ release.

6. We suggest that the negative inotropic effect of taxol may, at least in part, be the result of reduced release of Ca²⁺ from the SR. Microtubules may be important regulators of Ca²⁺ handling in the heart.

The cytoskeleton of the ventricular myocyte includes the interconnections of the contractile system, the membrane subcortical lattice of spectrin, filamin and ankyrin, microtubules and transmembrane attachments to extracellular components (Ganote & Armstrong, 1993). In isolated myocytes the microtubules are abundant around the nucleus and in the extramyofibrillar space. They are associated with the myofibrils in a helical arrangement and form a network that runs transversely at the level of the I band and axially between the myofibrils (Goldstein & Entman, 1979; Rappaport & Samuel, 1988). Microtubules have been consistently found in close association with the mitochondria and sarcoplasmic reticulum (SR), as well as with myofilament bundles (Goldstein & Entman, 1979).

In adult ventricular myocytes the role of the microtubules is not well understood. They may be implicated in cellular morphogenesis and provide a framework to which membranous organelles such as the nucleus are attached (Goldstein & Entman, 1979). Cytoskeletal disruption has been implicated in the response to ischaemia (Ganote & Armstrong, 1993). In cardiac cells from newborn rats, disruption of the microtubules has been reported to stimulate the rate of spontaneous contraction (Klein, 1983; Lampidis et al. 1992). Taxol is a known microtubule polymerizing agent (Schiff et al. 1979; Tsutsui et al. 1994; Johnson et al. 1996) and has been shown to slow inactivation kinetics of the L-type Ca²⁺ current (I_Ca,L). Conversely, the reverse effect has been observed with the microtubule disrupter colchicine, although neither agent affected I_Ca,L amplitude (Johnson & Byerly, 1994; Galli & Defelice, 1994).

Proliferation of the microtubules has been demonstrated in experimental pressure-overloaded hypertrophy. This response was associated with reduced contractile function of single myocytes. The negative inotropic effect was reversed by exposure to colchicine and mimicked by exposure of normal cells to taxol (Tsutsui et al. 1993, 1994; Ishibashi et al. 1996; Tagawa et al. 1996, 1997). It was suggested that increased microtubule content imposed additional load on the sarcomeres and that this was responsible for the negative inotropic response (Tsutsui et al. 1993, 1994; Ishibashi et al. 1996; Tagawa et al. 1996, 1997). However, it is possible that microtubule proliferation may influence other cellular mechanisms that modulate contraction. Therefore, in this study we have investigated the effects of taxol on contraction and Ca²⁺ handling mechanisms in ventricular myocytes isolated from rat heart.

METHODS

Cell isolation

Rats (Wistar, 250-300 g) were killed by cervical dislocation and hearts removed. Single ventricular myocytes were isolated by a combination of enzymatic and mechanical dispersion, according to previously described techniques (Frampton et al. 1991).

Immunofluorescence confocal microscopy

The microtubules were visualized using immunofluorescence techniques. Myocytes were washed in phosphate-buffered saline (PBS, for composition see below) for 10 min, centrifuged (400 r.p.m., for 40 s), then resuspended in a low Ca²⁺, low Mg²⁺ buffer (LCMB, for composition see below) for 10 min. Myocytes were permeabilized for 2 min with 1 % Triton X-100 in LCMB, washed twice in the same buffer and fixed for 30 min with 3·7 % formaldehyde. To minimize non-specific binding of antibodies, myocytes were incubated with 10 % (v/v) donkey serum in PBS containing 0·1 M glycine for 30 min. Myocytes were then incubated overnight (16 h) at 4°C with a 1 : 50 dilution of a monoclonal antibody to β-tubulin (N-357; Amersham International). Myocytes were then washed several times with PBS and incubated for 1½ h at room temperature (24°C) with a 1 : 50 dilution of fluorescein (FITC)-conjugated affinipure donkey anti-mouse IgG (715-095-150; Jackson, West Grove, PA, USA). After several washes with PBS, myocytes were mounted with 50 % Vectashield mounting medium (H-1000; Vector Laboratories, Inc., Burlingame, CA, USA). Sections of approximately 2 μm thickness were acquired by confocal laser microscopy (Leica True Confocal Scanner SP). An argon laser provided excitation light at 488 nm and emissions were collected between 500 and 600 nm.

On-line quantitative analysis software (TCS NT version 1.6.551; Leica) was used to measure the intensity of fluorescence emissions from the cell and hence provide a comparative estimate of microtubule density. The fluorescence intensity was normalized to scan surface area and measurements were made at the level of the nucleus. Comparisons were made only between cells that had undergone an identical antibody labelling and imaging process.

Measurement of contraction

Myocytes were electrically stimulated by 5-10 ms current pulses delivered by external platinum electrodes at 1 Hz. Cell shortening was used as our index of contraction and was monitored with a video edge detector (Crescent Electronics, Sandy, UT, USA).

Measurement of intracellular calcium

To measure intracellular calcium transients (Ca²⁺ transients), myocytes were loaded with the fluorescent indicator fura-2 AM (Molecular Probes, Eugene, OR, USA); 6·25 μl of a 1·0 mM stock solution of fura-2 AM (dissolved in dimethylsulphoxide, DMSO) was added to 2·5 ml of cells to give a final fura-2 concentration of 2·5 μM. Myocytes were shaken gently for 10 min at 24°C (room temperature). After loading, myocytes were centrifuged, washed with normal Tyrode solution (see below) to remove extracellular fura-2 and then left for 30 min to ensure complete hydrolysis of the intracellular ester.

To measure intracellular calcium ([Ca²⁺]_i), myocytes were alternately illuminated by 340 nm and 380 nm light using a rotating filter wheel which changed the excitation filter every 2 ms. The resultant fluorescent emission at 510 nm was recorded by a photomultiplier tube and the ratio of the emitted fluorescence at the two excitation wavelengths (340/380 ratio) was calculated to provide an index of [Ca²⁺]_i.

Direct effects of taxol on fura-2 in vitro were tested for in an intracellular simulating solution, containing 1 μM Ca²⁺, in the presence and absence of 10 μM taxol, using a Perkin-Elmer (MPF-44) fluorescence spectrophotometer (Frampton et al. 1991; Harrison et al. 1992). Excitation light was varied between 200 and 500 nm and fura-2 fluorescence at 510 nm was measured. The excitation spectrum of fura-2 was not affected by taxol.

The in vivo relationship between fura-2 and Ca²⁺ in untreated and taxol-treated cells was investigated using a modification of the method described by Frampton et al. (1991). In brief, fura-loaded cells which had been previously incubated with 10 μM taxol or vehicle solution for 2 h at 37°C, were resuspended in glucose-free Tyrode solution containing 5 μM carbonyl cyanide m-chlorophenyl hydrazone and 5 μM rotenone (pH 8·0). The [Ca²⁺] of this solution was varied to allow measurement of the maximum fluorescence ratio (R_max), the minimum fluorescence ratio (R_min), and the K_d of fura-2. R_max was determined in the presence of a solution containing 3 mM Ca²⁺, R_min was determined in the presence of 0 Ca²⁺ and 10 mM EGTA. A range of solutions containing free [Ca²⁺] (10⁻⁸ to 10⁻⁵ M) were used to measure K_d, which was calculated as described by Harrison & Bers (1987). After 5 min incubation in the appropriate solution, 50 μM ionomycin was added to the cell suspension and emission at 510 nm following excitation at 340 and 380 nm was recorded using the spectrophotometer described above. There was no significant difference (P > 0·05) in R_max (vehicle, 4·14 ± 0·99 ratio units; taxol, 4·80 ± 0·66 ratio units), R_min (vehicle, 0·71 ± 0·04 ratio units; taxol, 0·77 ± 0·01 ratio units) or K_d (vehicle, 300 ± 113 nM; taxol, 289 ± 142 nM) values recorded from cells incubated in the presence (n= 5) or absence (n= 5) of taxol.

Whole-cell patch clamp recording

Glass pipettes were pulled from non-heparinized haematocrit tubes to a resistance of 2-5 MΩ. The pipette solution for the whole cell patch clamp studies was (mM): CsCl, 130; K₂ATP, 8; glucose, 5; NaCl, 10; MgCl₂, 8; BAPTA, 5; Hepes, 10; and pH adjusted to 7·2 with 1 M CsOH. A voltage command pulse was applied to the pipette using an Axoclamp-2B amplifier (Axon Instruments, Inc.) controlled by a CED interface and software (Cambridge Electronic Design Ltd).

The protocol used to construct activation curves for L-type Ca²⁺ current (I_Ca,L) was as follows: from a holding potential of -40 mV, the membrane potential was stepped to potentials between -50 and +60 mV, with 10 mV increments, for 200 ms. Prior to each test pulse a train of four conditioning pulses (-40 to 0 mV, 200 ms) was applied at 1 Hz. The amplitude of I_Ca,L was measured as the difference between the peak inward current at the start of the test pulse and the steady-state current at the end of the 200 ms pulse (London & Krueger, 1994; Vornanen et al. 1994). I_Ca,L was normalized to cell capacitance (Satoh et al. 1996). Voltage-dependent inactivation curves were obtained by applying 200 ms pre-pulses using the activation procedure (range -50 to +60 mV). After stepping back to -40 mV for 10 ms, a depolarizing pulse to 0 mV was applied for 200 ms to assess the availability of I_Ca,L at each of the pre-pulse potentials.

Solutions

Normal Tyrode solution contained (mM): NaCl, 113; Hepes, 5; Na₂HPO₄, 1; MgSO₄.7H₂O, 1; KCl, 5; CaCl₂, 1; glucose, 10; sodium acetate, 20; insulin, 5 units l⁻¹; pH adjusted to 7·4. Taxol (T-7402; Sigma) and colchicine (C-9754; Sigma) were stored at -20°C as 10⁻² M stock solutions in methanol.

Unlike colchicine, taxol is poorly soluble in aqueous solution. During the preparation of experimental solutions the solubility of taxol in Tyrode solution was improved by warming the solution to 37°C and a 10 min period of sonication. The final concentration of methanol in the test solutions was 0·1 %. A control solution was used in the experiments containing 0·1 % methanol (vehicle).

PBS solution contained (mM): NaCl, 138; KCl, 2·7; KH₂PO₄, 2; Na₂HPO₄ (P-3813; Sigma), 8. LCMB solution contained (mM): Mes, 100; MgSO₄, 1; EGTA, 2; EDTA, 0·1; pH was adjusted to 6·75.

Statistics

Results are expressed as means ±s.e.m. Unless stated otherwise statistical analysis was performed by Student's independent sample t test. P values < 0·05 were considered significant.

RESULTS

Microtubule organization in taxol-treated myocytes

Immunofluorescence techniques were used to confirm microtubule polymerizing effects of taxol under the conditions of our experiments and to investigate the distribution of microtubules within rat ventricular myocytes. Myocytes were incubated for 4 h, at room temperature, in either 10 μM taxol or vehicle solution. Confocal images, approximately 2 μm in thickness, were recorded at the level of the nucleus. Figure 1A shows microtubules in a vehicle-treated cell and Fig. 1B shows microtubules in a taxol-treated cell. The intensity of fluorescence, normalized to cell surface area at the level of each scan, was significantly greater in cells treated with taxol than in untreated cells (P < 0·01) (Fig. 1C).

Figure 1. Immunofluorescence labelling of β-tubulin in rat ventricular myocytes.

Immunofluorescence confocal micrographs showing organization and distribution of microtubules in ventricular myocytes incubated in either vehicle (A) or 10 μM taxol (B) solutions for 4 h at room temperature (24 °C). Scale bars represent 10 μm. Immunolabelling protocol is described in Methods. The intensity of FITC emissions collected between 500 and 600 nm, normalized to section surface area, imaged at the level of the nucleus, are shown in C. Values are means ±s.e.m. and numbers above bars indicate number of cells. Normalized intensity was significantly higher (P < 0·01) in taxol-treated cells compared with vehicle cells.

Effects of taxol on contraction

Ventricular myocytes were incubated in either 10 μM taxol or vehicle solution for 4 h at room temperature. In agreement with the findings of Tsutsui et al. (1994) in feline myocytes, the effects of taxol on contraction after 2 h incubation were qualitatively the same but quantitatively less than after 4 h incubation (results not included for brevity). Myocytes were superfused with either taxol or vehicle solutions and contraction was recorded during the application of electrical stimulation at 1 Hz. Figure 2A shows typical recordings of contraction and Fig. 2B-D shows contraction amplitude measured as a percentage of resting cell length, time from stimulation to peak contraction (t_pk) and time from peak contraction to half-relaxation (t_½). Also shown for comparison are values of contraction, t_pk and t_½ in myocytes which had been stored in Tyrode solution at room temperature, soon after isolation (within 2 h) (Early control) and 8-10 h after isolation (Late control). Storage over this period did not have a significant (P > 0·05) effect on the kinetics of contraction.

Figure 2. Effect of taxol on the contraction of rat ventricular myocytes.

Ventricular myocytes were incubated in either 10 μM taxol or vehicle solution for 4 h at room temperature. Contraction was then measured in electrically stimulated (1 Hz) cells, superfused with either taxol or vehicle solutions at room temperature. A, single contraction traces comparing contraction in the presence and absence of taxol. B, contraction expressed as a percentage of resting cell length. C, time from stimulation to peak contraction (t_pk). D, time from peak contraction to half-resting level (t_½). Data are also given for untreated cells, taken within 2 h of isolation (Early control) and 8 h after isolation (Late control). Values are means ±s.e.m. and numbers above bars indicate number of cells. *P < 0·05, **P < 0·01 statistical comparisons between vehicle and taxol shown.

Resting cell length was not significantly different (P > 0·05) in taxol-treated myocytes (114·6 ± 3·5 μm, n= 14) compared with vehicle-treated myocytes (111·1 ± 3·7 μm n= 10). There was a significant (P < 0·01) reduction in the amplitude of contraction in cells treated with taxol (10·0 ± 1·0 %, n= 18) compared with cells treated with vehicle (14·4 ± 0·99 %, n= 10). The t_½ was significantly (P < 0·05) longer in cells treated with taxol (336·4 ± 11·9 ms, n= 16) compared with cells treated with vehicle (296·1 ± 9·9 ms, n= 10).

Having established that under our experimental conditions taxol was polymerizing microtubules (Fig. 1) and decreasing contraction (Fig. 2), we investigated the possibility that the negative inotropic effect of taxol was due to modification of Ca²⁺ handling in the cell.

Effects of taxol on the Ca²⁺ transient

Ventricular myocytes were incubated in either 10 μM taxol or vehicle solution for 4 h at room temperature. Myocytes were then superfused with either taxol or vehicle solutions and Ca²⁺ transients were recorded during the application of electrical stimulation at 1 Hz. Figure 3A shows typical recordings of Ca²⁺ transients and Fig. 3B-D shows the amplitudes of the Ca²⁺ transient (peak ratio minus resting ratio), time from stimulation to peak Ca²⁺ transient (t_pk) and time from peak Ca²⁺ transient to half-decay (t_½) of Ca²⁺ transients. Also shown for comparison are values of Ca²⁺ transient amplitudes, t_pk and t_½ in myocytes which had been stored in Tyrode solution at room temperature and studied within 2 h of isolation and 8-10 h after isolation. Storage over this period did not have a significant (P > 0·05) effect on the Ca²⁺ transient.

Figure 3. Effect of taxol on the Ca²⁺ transient.

Ventricular myocytes were incubated in either 10 μM taxol or vehicle solution for 4 h at room temperature. [Ca²⁺]_i transients were measured in electrically stimulated (1 Hz) cells superfused with either taxol or vehicle solutions at room temperature. A, original traces showing the effects of taxol on the Ca²⁺ transient. B, amplitude of Ca²⁺ transient (peak ratio - resting ratio). C, time from stimulation to the peak Ca²⁺ transient (t_pk). D, time from the peak of the Ca²⁺ transient to half-resting level (t_½). Data also given for untreated cells, taken within 2 h of isolation (Early control) and 8 h after isolation (Late control). Values are means ±s.e.m. and numbers above bars indicate number of cells. **P < 0·01.

Resting fura-2 ratios in taxol-treated myocytes were significantly (P < 0·01) smaller (0·19 ± 0·01, n= 22) than in vehicle-treated cells (0·27 ± 0·01, n= 20). There was a significant (P < 0·01) reduction in the amplitude of the Ca²⁺ transient in myocytes treated with taxol (0·07 ± 0·01, n= 22) compared with myocytes treated with vehicle (0·17 ± 0·01, n= 20). Thus it is possible that the negative inotropic response to taxol is mediated via a fall in the Ca²⁺ transient. The mean t_½ in myocytes treated with taxol was greater than in myocytes treated with vehicle but this difference was not significant (P > 0·05).

Comparison of the reduction in the amplitude of contraction and the Ca²⁺ transient by taxol (Fig. 2B and 3B) suggested that the effect on Ca²⁺ transients might be greater. We tested this possibility further by measuring contraction and Ca²⁺ transients simultaneously in a population of cells. When contraction and Ca²⁺ transients were measured in the same cell, there was a significant fall in both these parameters in taxol-treated cells (n= 17) compared with vehicle-treated cells (n= 18), but the fall in the Ca²⁺ transient was not significantly (P > 0·05) greater than the fall in contraction.

During the course of the experiments it was found that incubation of cells with taxol for 2 h at 37°C and subsequent superfusion with Tyrode solution in the experimental chamber, gave equivalent effects on contraction as a 4 h incubation at room temperature followed by superfusion with taxol (results not shown for brevity). Ca²⁺ transients were also measured following incubation with taxol for 4 h at room temperature (Fig. 3) and after 2 h at 37°C in the absence of superfusing taxol (Fig. 4). The magnitude and time course of the transients in response to the taxol treatments were not significantly different from each other (P > 0·05). As the magnitude of the reduction in the transients was similar (61 %, Fig. 3; 45 %, Fig. 4), we performed the remaining experiments using the shorter, warmer, incubation period. This approach allowed us to exclude the effects of external taxol on observed responses.

Figure 4. Effect of colchicine on Ca²⁺ transients from taxol-treated cells.

Ventricular myocytes were incubated in either 10 μM taxol or vehicle solution for 2 h at 37 °C. Some taxol-treated cells were then treated with 10 μM colchicine for a further 2 h, whilst the rest were left in vehicle solution. Ca²⁺ transients were measured in electrically stimulated (1 Hz) cells superfused with normal Tyrode solution at room temperature. A, original traces showing the effects of taxol or taxol followed by colchicine (Tax-Colch) on the Ca²⁺ transients. B, amplitude of Ca²⁺ transient. C, time to peak (t_pk) of the Ca²⁺ transient. D, time from the peak of the Ca²⁺ transient to half-resting level (t_½). Values are means ±s.e.m. and numbers above bars indicate number of cells. **P < 0·01.

Effects of colchicine on Ca²⁺ transients in taxol-treated myocytes

We next attempted to test whether the effects of taxol could be reversed by the microtubule depolymerizing agent colchicine. Colchicine has previously been shown to depolymerize microtubules in feline cardiac myocytes (Tsutsui et al. 1993, 1994; see Introduction). In untreated cells, colchicine does not modify the contraction of feline (Tsutsui et al. 1993, 1994) or rat myocytes (Brette et al. 1996).

Ventricular myocytes were incubated in either 10 μM taxol or vehicle solution for 2 h at 37°C. Some cells were then treated with 10 μm colchicine for a further 2 h, whilst the remaining vehicle and taxol-treated cells were incubated for a further 2 h in vehicle solution. Myocytes were then superfused with Tyrode solution and Ca²⁺ transients were recorded during the application of electrical stimulation at 1 Hz. Figure 4A shows typical recordings of Ca²⁺ transients and Fig. 4B-D shows the amplitude of the Ca²⁺ transient, t_pk and t_½ of Ca²⁺ transients, respectively. There was a significant (P < 0·01) reduction in the amplitude of the Ca²⁺ transient in myocytes treated with taxol (0·06 ± 0·004, n= 15) or taxol and colchicine (0·06 ± 0·003, n= 15) compared to vehicle (0·11 ± 0·01, n= 14). However, there was no significant difference between the amplitude of the Ca²⁺ transients in myocytes treated with taxol, or taxol and colchicine. There was a significant (P < 0·01) increase in the t_½ in myocytes treated with either taxol (281·1 ± 12·2, n= 15) or taxol and colchicine (276·9 ± 8·3, n= 15) compared to vehicle (240·9 ± 5·2, n= 14). Thus, colchicine was not able to reverse the effects of taxol (see Discussion).

Microtubule organization in taxol and taxol-colchicine-treated myocytes

Immunofluorescence techniques were used to visualize the effects of colchicine on myocytes which had been incubated with 10 μM taxol for 2 h at 37°C. Figure 5A shows a cell treated with taxol and Fig. 5B shows a cell following taxol and colchicine treatment.

Figure 5. Effect of colchicine on immunofluorescence labelling of β-tubulin in rat ventricular myocytes treated with taxol.

Immunofluorescence confocal micrographs showing organization and distribution of microtubules in ventricular myocytes incubated for 2 h in either 10 μM taxol (A) or 10 μM taxol followed by 10 μM colchicine (B) for a further 2 h at 37 °C. Scale bars represent 10 μm. The immunolabelling protocol is described in Methods. The intensity of FITC emissions collected between 500 and 600 nm, normalized to surface area, from sections imaged at the level of the nucleus are shown in C. Normalized intensity was not significantly different (P > 0·05) in cells treated with taxol and colchicine compared with taxol alone.

The intensity of microtubule labelling, at the level of the nucleus, was slightly less in taxol-treated cells exposed to colchicine when compared with cells exposed to taxol only, but the difference was not significant (P > 0·05). Thus it would appear that under the conditions of our experiments, exposure of cells to colchicine was unable to reverse microtubule polymerization induced by taxol (see Discussion).

Effects of taxol on L-type Ca²⁺ current

One possible way in which Ca²⁺ transients might be reduced is by a reduction in Ca²⁺ influx via I_Ca,L. To test for this possibility, ventricular myocytes were incubated in either 10 μM taxol or vehicle solution for 2 h at 37°C. During the experiment myocytes were superfused with Tyrode solution at room temperature. Myocytes were held at -40 mV to inactivate fast Na⁺ current. The pipette solution contained Cs⁺ to block K⁺ currents and BAPTA to increase the buffering of intracellular Ca²⁺ and thus abolish [Ca²⁺]_i-activated currents such as inward Na⁺-Ca²⁺ exchange. To standardize membrane conditions (such as voltage-dependent channel availability) a train of four 200 ms conditioning pulses were applied to 0 mV, at 1 Hz, prior to each test pulse.

To study the activation kinetics of I_Ca,L the membrane was depolarized from -40 mV to +60 mV in 10 mV steps each of 200 ms duration (Fig. 6A). Figure 6B is a typical recording of I_Ca,L, during a step to 0 mV in myocytes which had been pre-treated with either taxol or vehicle. The two current records are superimposed. The mean current-voltage relationship of I_Ca,L is shown in Fig. 6C. The amplitude of I_Ca,L during a depolarizing step to 0 mV was 2·86 ± 0·43 pA pF⁻¹ (n= 11) in taxol-treated cells and 3·38 ± 0·63 pA pF⁻¹ (n= 12) in vehicle-treated cells. The time to peak of I_Ca,L in taxol-treated cells was 7·95 ± 0·58 ms (n= 11) and in vehicle-treated cells was 8·83 ± 0·62 ms (n= 12). The time constant of inactivation of the currents was significantly (P < 0·05) longer (37·3 ± 2·13 ms; n= 11) in taxol-treated cells compared with vehicle-treated cells (30·8 ± 1·20 ms; n= 12).

Figure 6. Effect of taxol on I_Ca,L.

Ventricular myocytes were incubated in either 10 μM taxol or vehicle solution for 2 h at 37 °C then perfused with Tyrode solution. A, experimental protocol. Prior to each test pulse a train of four pulses (each to 0 mV for 200 ms) was applied to re-establish steady-state I_Ca,L. B, records showing the effect of 10 μM taxol on I_Ca,L, elicited by a 200 ms depolarization from -40 to 0 mV. Current records of a cell treated with taxol and a cell treated with vehicle solution are superimposed. Current-voltage relationships for activation and inactivation of I_Ca,L are shown in C and D, respectively. Values are means ±s.e.m.

To study the voltage-dependent inactivation of I_Ca,L, the membrane potential was returned to -40 mV for 10 ms and then stepped to 0 mV for 200 ms after each of the activation steps described above (Fig. 6A). Currents during the second step to 0 mV were normalized to maximum I_Ca,L and plotted against the membrane potential during the first step. The inactivation curves are shown in Fig. 6D. Taxol had no significant effect on voltage-dependent inactivation in the range -50 to +40 mV.

Although mean I_Ca,L was reduced by taxol (Fig. 6C), this effect was not statistically significant, therefore we must conclude that reduced Ca²⁺ influx via I_Ca,L is not responsible for the negative inotropic effect of taxol.

Effects of taxol on SR Ca²⁺ release

Experiments to investigate the effect of taxol on SR Ca²⁺ release were performed in ventricular myocytes incubated with 10 μM taxol or vehicle solution for 2 h at 37°C. During the experiment, myocytes were superfused with Tyrode solution at room temperature. Ca²⁺ transients were recorded in vehicle (Fig. 7A, upper panel) or taxol-treated cells (Fig. 7A, lower panel) during electrical stimulation at 1 Hz. Stimulation was then switched off and after a delay of 10 s, a rapid switch was made from Tyrode solution to 5 mM caffeine for a period of 10 s. The caffeine-induced release of Ca²⁺ from the SR was recorded.

Figure 7. Effect of taxol on SR Ca²⁺ release.

Ventricular myocytes were incubated in either 10 μM taxol or vehicle solution for 2 h at 37 °C and superfused with Tyrode solution. Initially cells were electrically stimulated at 1 Hz. Electrical stimulation was then turned off for 10 s and a switch to 10 mM caffeine was applied for 10 s. A, chart recordings showing the effects of caffeine on SR Ca²⁺ release in vehicle-treated (upper panel) and taxol-treated (lower panel) cells. B, amplitudes of Ca²⁺ transient (fura-2 ratio) during electrical stimulation or exposure to caffeine in taxol- or vehicle-treated cells. C, amplitude of Ca²⁺ transient during electrical stimulation expressed as a percentage of the Ca²⁺ transient during application of caffeine in vehicle- or taxol-treated cells. Values are means ±s.e.m.**P < 0·01 (n= 18).

Ca²⁺ release during electrical stimulation and application of caffeine is shown in Fig. 7B. The amplitude of the Ca²⁺ transient during electrical stimulation in taxol-treated cells was significantly (P < 0·01) lower (0·09 ± 0·01, n= 18) than that in vehicle-treated cells (0·14 ± 0·01, n= 18). The amplitude of the caffeine-evoked Ca²⁺ transient was similar (P > 0·05) in taxol-treated cells (0·19 ± 0·01, n= 18) and vehicle-treated cells (0·20 ± 0·01, n= 18). The time taken for Ca²⁺ to fall from peak caffeine-induced levels to half-resting level was decreased from 1103 ± 65 ms (n= 18) in vehicle-treated cells to 908 ± 53 ms (n= 18) in taxol-treated cells (P < 0·05). The amplitude of the electrically evoked Ca²⁺ transient expressed as a percentage of the caffeine-evoked Ca²⁺ transient has been referred to as fractional Ca²⁺ release, and is thought to be the percentage of Ca²⁺ stored within the SR that is released on electrical stimulation (Bassani et al. 1995). The percentage of the caffeine-induced Ca²⁺ transient, evoked by electrical stimulation was significantly (P < 0·01) lower in taxol-treated cells (51·1 ± 2·7 %, n= 18) compared with vehicle-treated cells (69·5 ± 2·9 %, n= 18). Our observations are consistent with the hypothesis that the reduction in the electrically evoked Ca²⁺ transient amplitude in the presence of taxol is the result of decreased release of Ca²⁺ from the SR and not decreased SR Ca²⁺ loading (see Discussion).

DISCUSSION

Previous studies have shown that proliferation of the microtubules, for example during experimentally induced cardiac hypertrophy, is associated with a reduction in contractile function of the heart and that depolymerization of the microtubules can normalize function (Tsutsui et al. 1993, 1994; Ishibashi et al. 1996). It has been suggested that the microtubules, when present in excess, impede sarcomere motion (Tsutsui et al. 1994; Tagawa et al. 1997) and that this may be responsible for the reduced contractile function. The anti-tumour agent taxol acts as a promoter of microtubule assembly (Schiff et al. 1979) and this property has been used to induce microtubule assembly in cardiac cells (Lampidis et al. 1992; Tsutsui et al. 1994) and reduce contractile function (Tsutsui et al. 1994; Ishibashi et al. 1996). In this study we investigated the effects of taxol on contraction and Ca²⁺ handling in ventricular myocytes isolated from the rat heart.

Effect of taxol on microtubules and the amplitude of contraction

Confocal images of taxol-treated cells showed an increase in the density of microtubules. Pre-treatment of myocytes with taxol caused a reduction in the amplitude of contraction. These findings are consistent with the findings of other workers (e.g. Tsutsui et al. 1993, 1994).

Effect of taxol on the Ca²⁺ transient

In order to account for possible errors that may result from comparison of Ca²⁺ transients between cells, only cells taken from the same hearts and loaded with the same batches of fura-2 were compared with each other. For this reason several sets of data for vehicle- and taxol-treated cells are presented. Incubation of myocytes with taxol caused significant reductions in the amplitude of the Ca²⁺ transient. As the amplitude of the Ca²⁺ transient is a major modulator of cardiac muscle contraction (Cannell et al. 1987; Cleemann & Morad, 1991) it seems likely that this is responsible, at least in part, for the negative inotropic effect of taxol.

The Ca²⁺ transient and the resting fura-2 ratio were significantly lower in taxol-treated cells compared with vehicle cells. One possible explanation for this is a stimulatory effect on the sarcolemmal Na⁺-Ca²⁺ exchanger which contributes to the removal of Ca²⁺ from the cytoplasm (Bers et al. 1991). The speeding of the decay of the caffeine-induced Ca²⁺ transient in taxol-treated cells would support this interpretation. In addition, it has been suggested that microtubules may affect local Ca²⁺ buffering (Johnson & Byerly, 1994); if this were to occur, it could influence the excitation-contraction coupling process in taxol-treated cells by affecting local [Ca²⁺]_i (see following sections).

Colchicine has been used to depolymerize microtubules (Klein, 1983; Tsutsui et al. 1994; Ishibashi et al. 1996; Tagawa et al. 1997). However in the present study, colchicine was not able to reverse the effects of taxol on the Ca²⁺ transient (Fig. 4). Immunofluorescence studies showed that the density of microtubules was slightly, but not significantly, reduced following colchicine treatment of cells previously exposed to taxol (Fig. 5). Thus it is interesting to note that, under the conditions of our experiments, subsequent incubation of taxol-treated cells with colchicine is not an appropriate experimental procedure to test for the reversibility of microtubular-dependent processes. The interactions of taxol and colchicine are difficult to explain because, although distinct binding sites have been identified for these agents and these binding sites may be linked, the mechanisms of action of taxol and colchicine in isolation are currently unknown (Arnal & Wade, 1995; Han et al. 1998). The effects of taxol on microtubules have also been shown to be resistant to the depolymerizing effects of cold and millimolar concentrations of Ca²⁺ (Schiff et al. 1979).

Effect of taxol on I_Ca,L

Ca²⁺ influx via I_Ca,L plays a major role in the triggering of Ca²⁺ release from SR Ca²⁺ stores. Blockade of I_Ca,L results in a reduction in the Ca²⁺ transient and thus a fall in the amplitude of contraction (Cannell et al. 1987; Cleemann & Morad, 1991). However, taxol did not significantly reduce the amplitude of I_Ca,L. This suggests that the negative inotropic effect of taxol is not mediated through an effect on I_Ca,L. The delay in I_Ca,L inactivation in taxol-treated cells is consistent with previous reports in which chick embryonic ventricular cells were superfused for 8 min with 50 μM taxol (Galli & Defelice, 1994), and in acutely isolated hippocampal pyramidal neurons from adult rats where 20 μM taxol was dialysed into the cell (Johnson & Byerly, 1994). The latter authors suggest that microtubules have a Ca²⁺ buffering effect close to the channel pore and that this fall in local Ca²⁺ concentration attenuates Ca²⁺-dependent inactivation. A similar mechanism could explain our findings. Another explanation is that the fall in the cytosolic Ca²⁺ transient, in response to taxol, reduces Ca²⁺-dependent inactivation, but it should be noted that our electrophysiological experiments were performed in the presence of the intracellular Ca²⁺ buffer BAPTA.

Taxol and SR Ca²⁺ release

A reduction in the Ca²⁺ transient could be explained by reduced release of Ca²⁺ from the SR. This might occur either as a result of reduced SR Ca²⁺ loading or reduced fractional SR Ca²⁺ release. Our experiments with caffeine indicate that the former is not the case as caffeine-induced transients in taxol-treated cells were not smaller than in untreated cells. Assuming the buffering of the caffeine-induced and electrically evoked Ca²⁺ transients are not differentially affected by taxol, this suggests that the target for taxol action is SR Ca²⁺ release. If microtubules act as local Ca²⁺ buffers (Johnson & Byerly, 1994), buffering of Ca²⁺ in the region of the SR Ca²⁺ release sites might result in a reduced trigger for release.

In summary, it has been shown that treatment of ventricular myocytes with taxol increased cellular microtubule content and reduced the amplitude of contraction. We suggest that this negative inotropic effect is due to a fall in the Ca²⁺ transient which is mediated neither by effects on I_Ca,L nor by a reduction in SR Ca²⁺ load. Our observations are consistent with the hypothesis that taxol reduces SR fractional release of Ca²⁺.