Formamide-induced osmotic shock has been used to detubulate isolated adult rat ventricular myocytes (i.e., disrupt the surface membrane-T tubule junction). Cell volume, calculated from cell length and width, rapidly decreased and increased upon application and removal of formamide, respectively. After treatment with formamide, membrane capacitance decreased by 26.4% (from 199.4 ± 18.7 pF in control cells to 146.7 ± 6.4 pF in formamide-treated cells;n = 13,P < 0.05). However, the amplitude of the L-type Ca2+ current (I Ca) decreased by a greater extent (from 0.75 ± 0.14 to 0.18 ± 0.03 nA;n = 5,P < 0.05) so that the density ofI Ca decreased by 74.5%. Simultaneous measurements ofI Ca and Ca2+ transients (monitored using fura 2) showed that both decreased rapidly upon removal of formamide. However, the Ca2+ content of the sarcoplasmic reticulum showed little change. Cross-striations, visualized with the fluorescent dye di-8-aminonaphthylethenylpyridinium, were sparse or absent in cells that had been treated with formamide, suggesting that formamide can successfully detubulate cardiac cells and thatI Ca is concentrated in the T tubules, which therefore play an important role in excitation-contraction coupling.
- transverse tubules
- calcium current
- cell volume
- sarcoplasmic reticulum
Ca2+ -induced Ca2+ release (CICR) is the major mechanism for Ca2+ release from the sarcoplasmic reticulum (SR) of cardiac muscle (1, 8, 9). Ca2+ entering the cell via L-type Ca2+ channels (1) and, under some conditions, the Na+/Ca2+exchange mechanism (18), triggers the release of further Ca2+ from the SR. Current evidence suggests that, of these two pathways, Ca2+ entering via the L-type Ca2+ channels is the main trigger for Ca2+ release (e.g., Ref. 7 and see discussion) and that the transverse (T) tubules are a particularly important site for the coupling of Ca2+ entry across the cell membrane to Ca2+ release from the SR (28).
T tubules are invaginations of the sarcolemma that occur perpendicular to the longitudinal axis of the cell at intervals of ∼1.8 to 2.0 μm (23, 25). The percentage of the total surface membrane contained in the T tubules depends on the species (24) and the region of the heart (e.g., see Ref. 13). By analogy with skeletal muscle, it has been suggested that T tubules help the spread of excitation to the center of the cell so that activation of Ca2+ release is synchronized to produce homogenous Ca2+ release throughout the cell (3). This suggestion is supported by three lines of evidence. First, Ca2+ sparks evoked by electrical stimulation occur predominantly close to T tubules (28), suggesting an important role for the T tubules in excitation-contraction coupling. Second, atrial cells (that lack a T tubular system) exhibit Ca2+release that is initially localized near the subsarcolemmal space and subsequently spreads to the central regions of the cell (13). Finally, cultured ventricular myocytes, which lose their T tubules, exhibit biphasic Ca2+ transients that may result from spatially nonuniform SR Ca2+ release (20). However, in the latter two cases, other aspects of cell function (e.g., protein expression) are also different. Thus direct evidence for the role of T tubules is still sparse.
Previous studies designed to elucidate the role of the T tubules in skeletal muscle have used membrane-permeant solutes such as glycerol (12) and formamide (26) to produce an osmotic shock that disrupts the normal coupling between the T tubules and the surface membrane (detubulation). Such methods have not been successful in multicellular cardiac preparations. However, we have used formamide to produce osmotic shock in single isolated cardiac cells to investigate the functional role of the T tubules in excitation-contraction coupling.
Isolation of myocytes and measurement of cell length and width. Ventricular myocytes were isolated from female adult Wistar rat hearts. The animals were stunned and then killed by cervical dislocation, in accordance with the Home OfficeGuidance on the Operation of the Animals (Scientific Procedures) Act of 1986. Myocytes were isolated as described previously (10) and allowed to settle in the recording chamber, which was mounted on an inverted microscope (Diaphot; Nikon). The setup was configured for the simultaneous recording of cell length or width, fluorescence transients, and membrane currents (10, 14).
Myocytes were illuminated with red light (>610 nm) to generate an image of the cell, which was detected by a camera mounted on the side limb of the microscope and displayed on a monitor. The length or width of this image was monitored using an edge detection system (Crescent Electronics). These measurements were made on myocytes that had been loaded with the Ca2+ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA)-AM (5 μmol/l for 15 min) to avoid formamide-induced changes in intracellular Ca2+concentration ([Ca2+]i; see Fig. 4) that could also alter the dimensions of the cell.
Measurement of Ca2+ current and [Ca2+]i.
Two series of voltage-clamp experiments were carried out to determine the effects of formamide on Ca2+current (I Ca). In the first, I Caand [Ca2+]iwere monitored simultaneously in myocytes loaded with fura 2-AM (5 μmol/l for 15 min). The membrane potential was controlled using the perforated patch-clamp technique, with amphotericin B (250 or 500 μg/ml) as the pore-forming agent, as described previously (14). The pipette solution used for these recordings contained (in mmol/l) 120.0 cesium glutamate, 20.0 KCl, 10.0 NaCl, 1.0 MgCl2, 1.0 CaCl2, and 10.0 HEPES, adjusted to pH 7.3 with tetraethylammonium hydroxide (TEA-OH). In the second series of experiments, membrane capacitance andI Ca were monitored using conventional whole cell voltage-clamp in BAPTA-loaded cells. The pipette solution for these experiments contained (in mmol/l) 120.0 cesium glutamate, 20.0 KCl, 10.0 NaCl, 5.0 Mg-ATP, 1.0 BAPTA, and 10.0 HEPES, adjusted to pH 7.3 with TEA-OH. The amplitude ofI Ca was measured as the difference between peak inward current and the current remaining at the end of the 300-ms voltage-clamp pulse. Cs-based pipette solutions and 20.0 mmol/l CsCl in the external solution were used in both sets of experiments to avoid contamination ofI Ca by potassium currents.
All data were digitized by an analog-to-digital (A/D) converter (model DR-890; Neuro Data Instruments) and stored on video tape (model SR-330MS; Victor Company of Japan, Tokyo, Japan) and on the hard disk of a microcomputer via a CED 1401plus A/D interface and Sigavg or Vclamp software (Cambridge Electronic Design).
Osmotic shock treatment of ventricular myocytes with formamide and confocal imaging of T tubules. Myocytes were bathed in a solution containing (in mmol/l) 113.0 NaCl, 5.0 KCl, 1.0 MgSO4, 1.0 CaCl2, 1.0 Na2HPO4, 20.0 sodium acetate, 10.0 glucose, 10.0 HEPES, and 5.0 U/l insulin, pH adjusted to 7.4 with NaOH. To induce osmotic shock, cells were exposed to this solution plus formamide (1.5 mol/l) for 15 min, before returning the cells to control solution.
In some experiments, control (untreated) cells and cells that had undergone this osmotic shock procedure were then labeled with 5.0 μmol/l di-8-aminonaphthylethenylpyridinium (Di-8-ANNEPS) for 2 min, washed in control solution for 5 min, and then imaged using a confocal microscope (Leica), using 488 nm excitation light and detection at 514 nm.
Direct effects of formamide on fura 2 fluorescence. To determine whether formamide had direct effects on fura 2 fluorescence, solutions containing a range of Ca2+ concentrations, the pentapotassium salt of fura 2 (10.0 μmol/l), and EGTA (10.0 mmol/l) were prepared as described previously (16), and the effect of 1.5 mol/l formamide on fura 2 fluorescence was determined in vitro. Formamide had no effect on fura 2 fluorescence: pCa1/2 was 5.91 in the absence and 5.95 in the presence of formamide.
Chemicals and reagents. All solutions were made using ultrapure water supplied by a Milli-Q system (Millipore). Salts and other solution constituents were reagent grade and were purchased from Sigma unless stated otherwise. Stock solutions of BAPTA-AM (10.0 mmol/l) and fura 2-AM (1.0 mmol/l) were prepared in DMSO, and Di-8-ANNEPS (10.0 mmol/l) was dissolved in DMSO plus 20% pluronic acid (all from Molecular Probes). Amphotericin B (2.5 mg/30 μl) was also prepared in DMSO and then diluted in the pipette solution.
Statistical analysis. Data are presented as means ± SE, and statistical analysis was performed using paired or unpaired t-tests, as appropriate.
Effects of formamide on cell volume. We first determined the effects of formamide on cell volume, because detubulation is thought to occur in response to the increase in volume that occurs on washout of a membrane-permeable solute (e.g., Refs. 12and 26 and see discussion).
Figure1 A,top trace, shows that addition of 1.5 mol/l formamide to the bathing solution caused a rapid (<1 min) decrease in resting cell length, from 106.6 ± 2.0 to 96.8 ± 2.1 μm (n = 5,P < 0.001). This was followed by a further slower decrease, to 84.0 ± 2.1 μm, during the remainder of the 15-min exposure. The second trace shows that changes in cell width occurred over a similar time course, although the initial decrease, from 30.2 ± 2.0 to 28.2 ± 2.1 μm (P < 0.01,n = 5), was followed by an increase (rather than a further decrease) in the continued presence of formamide. Washout of formamide caused the opposite effects. These data were used to calculate cell volume (v), as described previously (2) wherel andw are cell length and width, respectively. Cell depth (d) was derived by assuming the cell to be an elliptical cylinder with a minor-to-major axis ratio of 1:3 (2). Cell volume calculated using the top two traces in Fig.1 A is shown in the third trace. Figure 1 A,bottom, shows that average cell volume decreased by ∼21.0% within 1 min of adding 1.5 mol/l formamide and increased by 17.8% (n = 5) after washout of formamide, again in <1 min of washout, before recovery toward control levels.
Effects of formamide on membrane staining with Di-8-ANNEPS. To investigate whether these changes in volume caused detubulation, we used confocal microscopy to visualize staining of the cell membrane with the lipophilic fluorescent indicator Di-8-ANNEPS (13, 20). The confocal images in Fig. 1 show Di-8-ANNEPS staining of a representative control myocyte (Fig.1 B), characterized by regular cross-striations at intervals of 1.82 ± 0.02 μm (n = 9) along the longitudinal axis of the cell, as described previously for staining of the T tubules (23,25). Similar staining patterns were observed in all the sections taken at 1.0-μm intervals through the depth of the cell. In contrast, there were no such cross-striations in 609 out of 700 cells examined after treatment with formamide (Fig. 1 C). These data suggest that the Di-8-ANNEPS no longer had access to the T tubules; this could be either because the T tubule membranes had been lost or because the T tubules had been uncoupled from the surface membrane. However, because labeling with Di-8-ANNEPS was carried out after the osmotic shock treatment (see methods), we cannot distinguish between these possibilities.
Effects of formamide on ICa and Ca2+ transient in voltage-clamped cells.
If detubulation occurs during the rapid swelling of the cell upon washout of formamide (26), changes inI Ca and the Ca2+ transient might also be expected to occur during this period. Figure2 A shows a portion of the cell volume record (taken from Fig. 1) on an expanded time scale and records ofI Ca and Ca2+ transients recorded simultaneously using the perforated patch-clamp technique and fura 2, respectively (Fig. 2 B), showing that cell expansion after washout of formamide is accompanied by a rapid decrease in the amplitude ofI Ca and the accompanying Ca2+ transient. Figure 2 C shows these changes on a faster time base, showing more clearly the parallel decrease in these parameters. These data are compatible with the idea that detubulation occurs during cell swelling upon washout of the solute (26).
To determine whether these changes were associated with a loss of membrane and whether the channels underlyingI Ca were uniformly distributed, we measured membrane capacitance andI Ca in control and formamide-treated myocytes loaded with BAPTA. Figure3 shows typical records ofI Ca from a control cell (A) and a cell that had been treated with formamide (B) and the associated current-voltage relationships (C;n = 5) forI Ca obtained by depolarizing the membrane potential in 4-mV increments from a holding potential of −40 mV at 0.5 Hz. The myocytes were also subjected to 10-mV hyperpolarizing pulses from a holding potential of −80 mV and the capacity currents at the beginning of the pulse integrated to determine membrane capacitance.
Membrane capacitance decreased from 199.4 ± 18.7 to 146.7 ± 6.4 pF (i.e., by 26.4%) after treatment of myocytes with formamide (n = 13,P < 0.05), compatible with loss of membrane and hence detubulation. However,I Ca decreased by 75.8% (at 0 mV), suggesting that this fraction ofI Ca is within the T tubular membrane. Because this decrease (75.8%) is greater than the apparent percentage decrease in surface area (26.4%), as indicated by cell capacitance, this suggests thatI Ca is concentrated within the T tubules. Figure3 D shows this graphically, showing that I Ca density decreased after formamide treatment. Together, these data suggest that L-type Ca2+ channel density is 8.7 times greater in the T tubules than in the surface membrane.
Effects of formamide on Ca2+ transients in field-stimulated myocytes.
Figure 4 Ashows that addition of formamide caused an initial decrease and then an increase in the amplitude of the Ca2+ transient. This was followed by a slower decrease in amplitude to near control levels during the continued presence of formamide. Diastolic Ca2+ remained elevated throughout the exposure to formamide.
Removal of formamide was followed by two patterns of change in the amplitude and time course of the Ca2+ transient. In one group of cells (∼70%), amplitude decreased from 0.52 ± 0.05 to 0.15 ± 0.01 ratio units (n = 14,P < 0.001) after 5 min of washout, with little change in the time course of the transient (Fig.4 B,middle). In 12 out of 26 cells, the transient remained unchanged, whereas in the remaining 14 cells the amplitude recovered to 0.28 ± 0.03 ratio units (P < 0.001) at 30 min of washout. In cells in which the Ca2+ transients showed recovery, the time course of the Ca2+ transients became prolonged, with the initial rapid rise of Ca2+ being followed by a secondary slower increase (Fig. 4 B,right). In the second group of cells, such biphasic Ca2+transients were observed immediately after washout of formamide.
The Ca2+ content of the SR, assessed using caffeine, showed little change after removal of formamide (Fig. 4, A andC): the amplitude of caffeine-induced Ca2+ release was 0.72 ± 0.12 ratio units in control, 0.64 ± 0.11 after 5 min (n = 9,P < 0.05), and 0.64 ± 0.10 (not significant) after 30 min washout of formamide. Thus the ratio of the amplitude of electrically stimulated Ca2+ transient to the amplitude of caffeine-induced Ca2+ transient (“fractional release”) decreased significantly, from 0.78 ± 0.06 in control cells to 0.26 ± 0.05 after 5 min washout of formamide (n = 9,P < 0.001).
The time constant of the decline of caffeine-induced Ca2+ transients was measured to obtain an indication of changes in the rate of Ca2+ extrusion across the cell membrane (Fig. 4 C). This increased by a small but statistically insignificant amount, from 1.91 ± 0.25 s in control to 2.23 ± 0.41 s after 5 min and 2.66 ± 0.53 s after 30 min of washout (n = 7).
Detubulation of single ventricular myocytes. The high osmotic pressure of the formamide solution (1,780 mosmol), compared with the control solution (286 mosmol), would be expected to cause the cell to lose water and shrink (Fig. 1 A). However, formamide is membrane permeant and therefore enters the cell down its concentration gradient, accompanied by water, so the cell subsequently expands in the continued presence of formamide (Fig.1 A). During washout, the intracellular concentration of formamide is initially high so that water enters the cell, causing the rapid expansion (Figs.1 A and2 A) that is thought to cause the T tubules to break from the surface membrane while the surface membrane reseals (26). The data shown in Figs. 1 and 2 are compatible with this idea. Formamide will then leave the cell, taking water with it, to allow the cell volume and shape to recover slowly.
Several lines of evidence suggest that formamide-induced changes in cell volume are effective in detubulating cardiac myocytes. First, the absence of cross-striations from the center of cells stained with the lipophilic dye Di-8-ANNEPS after formamide treatment (Fig.1 C) is strong evidence that T tubules were either disrupted or uncoupled from the cell surface in these cells. Second, detubulation would be expected to occur predominantly during the rapid expansion that occurs upon washout of formamide. The data in Figs. 1 and 2 show that cell expansion does occur rapidly and that this is accompanied by a decrease in the amplitude of functional parameters such asI Ca and the Ca2+ transient (Fig.2 B). Third, membrane capacitance (and presumably membrane area, if capacitance is uniformly 1.0 μF/cm2) was decreased by 26.4% after detubulation. This is close to previous estimates of the proportion of the membrane contained in the T tubules in rat ventricular myocytes (33%; see Ref. 24) and therefore also suggests that T tubules have been largely removed. Finally, the biphasic Ca2+ transients shown in Fig.4 B,right, are similar to the transients seen in cultured myocytes in which such changes were suggested to result from the detubulation that occurs in culture (Ref. 20 and see below).
An alternative that should be considered, however, is the possibility that formamide has direct effects on cardiac cells that might account for the present data. During exposure to formamide, the amplitude of the Ca2+ transient changes (Fig.4 A) and, in the cells in which it was monitored throughout exposure to formamide,I Ca initially increased and subsequently decreased. Although these changes could be due to formamide itself or secondary to the induced changes of volume (Fig. 1 A), they raise the possibility that formamide has direct effects on cardiac cells. There are a number of lines of evidence, however, that make it unlikely that such effects can account for the present results. First, the changes ofI Ca and the Ca2+ transient reported here occurred on wash off, not wash on, of formamide (Fig. 2), which is unexpected for a direct effect of formamide, and occurred with a similar time course to the change in cell volume (Fig. 2), consistent with this being important in the observed changes. Second, the characteristics ofI Ca, including its configuration and voltage dependence (Fig. 3), and its response to BAY K 8644 and isoprenaline (not shown) were unchanged after formamide treatment, as was the response to caffeine (Fig. 4), so that any direct effects of formamide would have to be very specific to account for the observed changes, and it seems unlikely that direct effects of formamide would be so specific as to produce a series of changes so compatible with detubulation (see above).
Previous attempts to use this technique in cardiac muscle have been unsuccessful (5), possibly because multicellular preparations were used in which diffusion delays could have slowed changes in cell volume and hence decreased the efficacy of the shock. It is also possible that the most commonly used solute (glycerol) crosses the cell membrane too slowly to increase intracellular concentration sufficiently to induce osmotic shock; in support of this idea, in an earlier series of experiments we were unable to achieve effective detubulation using glycerol (data not shown). Although excitation-contraction coupling remains disrupted 30 min after washoff of formamide (e.g., Fig. 4), it is unclear whether the observed changes reverse over longer periods than those monitored in the present study.
Functional consequences of detubulation on Ca2+ handling in cardiac muscle.
The percentage decrease in the amplitude ofI Ca was greater than the percentage decrease in membrane capacitance (and hence membrane area). Thus the density ofI Ca decreased, consistent with previous suggestions that L-type Ca2+ channels are concentrated in the T tubules (19). This has implications for excitation-contraction coupling in cardiac cells. Current models of this process suggest that local Ca2+ release, from a cluster of ryanodine receptors, depends on local trans-sarcolemmal Ca2+ influx, which increases local Ca2+ concentration, thus increasing the open probability of adjacent ryanodine receptors (4, 21,22, 27). Such Ca2+ release has previously been shown to be triggered by the L-typeI Ca in a stochastic manner (4, 21, 22), with a voltage dependence similar to that of the L-typeI Ca (22, 27); such release does not, however, normally appear to be triggered by Ca2+ influx via the Na+/Ca2+exchanger (22). These data suggest that the main trigger for release is normally Ca2+ influx via the L-type Ca2+ channels, rather than influx via the Na+/Ca2+exchanger. Labeling studies have shown that the L-type Ca2+ channel and the ryanodine receptor are concentrated in, and close to, the T tubules, respectively, and are closely juxtaposed (19), suggesting a possible mechanism for the potency of Ca2+influx via I Ca as a Ca2+ trigger. However some studies also show a concentration of the Na+/Ca2+exchange protein in the T tubules (6, 11); in this case, the relative lack of potency of Ca2+ influx via the exchanger as a trigger of SR Ca2+ release could be due to the position of the exchanger relative to the ryanodine receptor or to lower Ca2+ flux rate through the exchanger, since the size and rate of change of Ca2+ at the surface of the SR is known to determine the effectiveness of the Ca2+ trigger (8). However, other studies show a more even distribution of the exchange protein throughout the T tubules and surface membrane (17). The present observation that the rate of decline of caffeine-induced Ca2+ release decreases by approximately the same proportion as cell surface area after treatment with formamide is compatible with such a distribution (see below); in this case, the relative ineffectiveness of Ca2+ influx via the exchanger to trigger SR Ca2+ release might be because the Na+/Ca2+exchanger is not concentrated near, or closely juxtaposed to, the ryanodine receptors.
Because the L-type Ca2+ channel appears to be the main source of trigger Ca2+, the observation that L-type Ca2+ channels are concentrated in the T tubules can explain why electrical stimulation gives rise to local SR Ca2+ release (Ca2+ sparks) predominantly close to the T tubules (28). The present study suggests that this is unlikely to be due to better coupling between the L-type Ca2+ channels and the ryanodine receptors in the T tubules because the Ca2+ transient remaining after detubulation had a rapid rising phase, which would suggest efficiently coupled CICR at the surface membrane. If the Ca2+ influx pathways responsible for triggering this rapid Ca2+ release are located mainly on the surface membrane, then comparison of the relationship between the decrease in the amplitude ofI Ca and the accompanying decrease in the amplitude of the Ca2+ transient with previous work on intact myocytes (14) suggests that the decrease in the Ca2+ transient is that expected from the decrease in I Ca and hence that Ca2+ release is equally well coupled to Ca2+ entry in the T-tubules and the surface membrane. It appears more likely, therefore, that the concentration of L-type Ca2+ channels in the T tubules results in greater Ca2+ influx viaI Ca in the T tubules, which, in conjunction with the local concentration of ryanodine receptors, will give rise to more Ca2+ sparks in the vicinity of the T tubules. Similarly, during an action potential, the majority of Ca2+ influx viaI Ca will occur in the T tubules, at the site where the ryanodine receptors are concentrated, so that most Ca2+release would be expected to occur in the vicinity of the T tubules. Because the coupling ofI Ca to Ca2+ release appears to be similar in the surface membrane and T tubule (see above), it seems likely that the decrease in the amplitude of the Ca2+ transient after detubulation can be explained simply by the loss ofI Ca. Thus the present evidence suggests that the T tubules are important in excitation-contraction coupling because they contain a high proportion of the total I Ca, which, in turn, appears to be the main influx pathway for triggering SR Ca2+ release.
It is unlikely that changes in the Ca2+ content of the SR play a role in the observed changes, because SR Ca2+ content does not change appreciably (Fig. 4 C). The relatively small effect of detubulation on SR Ca2+ content is perhaps surprising, sinceI Ca is thought to help load the SR with Ca2+ (9,15). One possibility is that Ca2+content of the SR is maintained by Ca2+ influx pathways in the surface membrane (Ca2+ channels and Na+/Ca2+exchange) rather than the T tubules under these conditions.
The slowly rising phase of the Ca2+ transient observed after the initial rapid rise is consistent with slower regenerative CICR from the SR within the cell after the initial triggered release at the cell surface (20).
During relaxation of the twitch, [Ca2+]iis normally sequestered by the SR and extruded across the cell membrane, mainly by Na+/Ca2+exchange and the sarcolemmal Ca2+-ATPase. The observation that the percentage decrease in the rate constant of decline of the caffeine-induced Ca2+ transient after formamide treatment (14.5–28.2%) was similar to the percentage decrease in apparent surface area (26.4%; see above) suggests that the Ca2+ extrusion mechanisms (principally the Na+/Ca2+exchanger in rat myocytes) are distributed uniformly across the cell membrane rather than being concentrated in the T tubules.
In summary, the data in the present study show that osmotic shock with formamide can successfully detubulate isolated myocytes. The observation thatI Ca appears to be concentrated within the T tubules suggests these to be important sites for triggering SR Ca2+ release, possibly for synchronizing CICR throughout the cell. In contrast, Ca2+ extrusion mechanisms appear to be more uniformly distributed across the T tubules and the surface membrane.
We are grateful to the British Heart Foundation and the Uehara Memorial Foundation of Japan for their support.
Address for reprint requests and other correspondence: C. H. Orchard, School of Biomedical Sciences, Univ. of Leeds, Leeds LS2 9NQ, UK (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 1999 the American Physiological Society