Simultaneous measurements of changes in sarcoplasmic reticulum and cytosolic [Ca²⁺] in rat uterine smooth muscle cells

Abstract

1. The role of the sarcoplasmic reticulum (SR) was investigated in spontaneous and agonist-induced uterine Ca²⁺ transients, by combining low- (mag-fluo-4) and high-affinity (fura-2) indicators to measure intraluminal SR ([Ca²⁺]_L) and cytosolic ([Ca²⁺]_i) calcium concentration, simultaneously, in single smooth muscle cells from pregnant rat uterus.

2. Carbachol or ATP, in the absence of extracellular Ca²⁺, decreased [Ca²⁺]_L and increased [Ca²⁺]_i. Although some replenishment (around 50 %) occurred in its absence, extracellular Ca²⁺ was required for full replenishment of the SR Ca²⁺.

3. In 4/15 cells, ATP evoked oscillations of [Ca²⁺]_i. These were accompanied by successive release and re-uptake of SR Ca²⁺. Inhibition of the SR Ca²⁺-ATPase with thapsigargin abolished the oscillations and luminal changes.

4. Spontaneous [Ca²⁺]_i transients produced no detectable changes in [Ca²⁺]_L. The larger [Ca²⁺]_i transients evoked by high-K⁺ depolarisation increased [Ca²⁺]_L. Spontaneous activity was inhibited when [Ca²⁺]_L was increased.

5. These data show that it is possible to simultaneously measure SR and cytosolic [Ca²⁺], and to investigate their response to agonist application and spontaneous activity.

Understanding the factors that control intracellular calcium in uterine myocytes is crucial to elucidating the mechanism of labour, as uterine contractions are triggered by increases in cytosolic Ca²⁺ concentration ([Ca²⁺]_i) (Szal et al. 1994; Taggart et al. 1996). The rise in [Ca²⁺]_i is due to entry through the surface membrane calcium channels and, possibly, Ca²⁺ release from the sarcoplasmic reticulum (SR) (Shmigol et al. 1998a). For example, SR calcium release mediated by inositol trisphosphate (InsP₃) has been shown in human myometrium (Luckas et al. 1999), and contributes to agonist-induced force production in smooth muscles. Calcium-induced calcium release (CICR), however, appears to play little (Shmigol et al. 1998a) or no (Taggart & Wray, 1998) role in spontaneous force production and global [Ca²⁺] transients in the uterus. It may be that, as in vascular smooth muscle, the SR or CICR is more important for relaxation than contraction. Thus in vascular smooth muscle, Ca²⁺ release from the SR (‘Ca²⁺ sparks’) activates K⁺ channels (K_Ca) (Zhou et al. 2000) and is associated with relaxation, as Ca²⁺ entry is inhibited (Nelson et al. 1995). Such Ca²⁺-activated K⁺ channels have also been reported in myometrium (Vivat et al. 1992; Khan et al. 1993).

A detailed understanding of the role played by the SR in smooth muscle is limited by the lack of direct SR Ca²⁺ measurements (ZhuGe et al. 1999). Information on the SR has relied on indirect measurements of its Ca²⁺ content and pharmacological approaches. In the present study, we report the first measurements of simultaneous, time-resolved SR and cytosolic free Ca²⁺ changes, in mammalian muscle. We have used measurements in uterine myocytes (i) to investigate the changes in SR calcium during agonist-induced and spontaneous [Ca²⁺]_i transients and (ii) to determine how the SR calcium content affects spontaneous activity.

Part of this study has been published in abstract form (Shmigol et al. 1999a).

METHODS

Experiments were performed on acutely isolated uterine smooth muscle cells. Pregnant female Wistar rats (19-21 days gestation) were killed by cervical dislocation under CO₂ anaesthesia, and cells were enzymatically isolated from the longitudinal myometrium (Shmigol et al. 1998a). The myocytes were kept in KB medium (Klockner & Isenberg, 1985) at 4 °C, and retained their normal physiological properties for up to 36 h. To assess the intraluminal SR Ca²⁺ concentration ([Ca²⁺]_L) simultaneously with [Ca²⁺]_i, we used a low-affinity calcium indicator, mag-fluo-4 (K_d= 22 μm), and a high-affinity dye, fura-2 (K_d= 0.14 μm), differentially loaded into the SR and cytosol. To obtain differential loading, the cells were incubated with 5 μm mag-fluo-4 AM at 37 °C for 40 min, washed with indicator-free solution and stored overnight at 4 °C to unload the cytosolically located mag-fluo-4. The cells were then loaded with 2-3 μm fura-2 AM at room temperature for 20 min, and then left in KB medium at room temperature for at least 40 min to allow de-esterification of the indicators. The cell suspension was put into a perfusion chamber (volume 0.2 ml) mounted on the stage of an inverted microscope (Nikon Diaphot 200). After they had settled in the chamber, cells were superfused with Krebs solution (35 °C, 5 ml min⁻¹) for around 1 h. For fluorescence measurements, the cells were illuminated at 340 and 380 nm for fura-2 excitation, and at 465 nm for mag-fluo-4 excitation using a monochromator (Cairn Optoscan, Faversham, Kent, UK). The frequency of wavelength alternation (33 Hz) was adequate to resolve rapid changes in Ca²⁺ levels. The fluorescence from both indicators was separated from the excitation light by a dichroic mirror, 470DCXR, and a long-pass interference filter, E480LP (Chroma Technology Corp., Brattleboro, VT, USA), and measured using a PMT. Verification of differential loading of mag-fluo-4 into the SR was undertaken as follows. (i) The surface membrane was permeabilised with saponin (50 μg saponin dissolved in solution containing 140 mm NaCl and 5.5 mm KCl; free Ca²⁺ in this solution was set to approximately 100 nm using 0.066 mm CaCl₂ and 0.1 mm EGTA, 35 °C, pH 7.4). No change in the mag-fluo-4 fluorescence was detected, which suggested that little or no mag-fluo-4 was retained in the cytosol after unloading. (ii) The mitochondria were discharged with carbonyl cyanide 3-chlorophenylhydrazone (CCCP). There was no decrease in the mag-fluo-4 fluorescence in the presence of CCCP, indicating that the calcium affinity of mag-fluo-4 was too low to detect changes in intramitochondrial calcium (and suggesting mitochondrial free Ca²⁺ may be low). The dual loading did not alter cell function, as judged by comparing, in single- and dual-loaded cells, Ca²⁺ transients, levels of resting [Ca²⁺] and responses to high-K⁺ solution.

As single-wavelength indicators are difficult to calibrate in terms of absolute Ca²⁺ concentrations, we used normalised mag-fluo-4 signals (F/F₀) to express changes in [Ca²⁺]_L. [Ca²⁺]_i was calculated from the fura-2 ratio based on in situ calibration using ionomycin and Calcium Calibration Buffer kit no. 2 (Molecular Probes, Eugene, OR, USA) with magnesium.

Agonists were applied from a blunt pipette positioned close to the cell. Contractile activity of cells was viewed in some experiments simultaneously with the fluorescence measurements, using an infrared video camera.

Where appropriate, the results are presented as means ±s.d.; otherwise, the traces shown represent typical results from at least four similar experiments, obtained from different cells, from at least three animals. Statistical differences were tested using ANOVA.

Solutions

Modified Krebs solution (referred to simply as Krebs solution henceforth) contained (mm): NaCl, 140; KCl, 5.4; CaCl₂, 2; MgCl₂, 1.2; glucose, 10; Hepes, 10; adjusted to pH 7.4 with NaOH. Ca²⁺-free solution contained 0.2 mm EGTA and 5 mm magnesium, with no Ca²⁺. Hanks’ solution contained (mm): NaCl, 137; KCl, 5.1; KH₂PO₄, 0.44; Na₂HPO₄, 0.26; glucose, 5.5; Hepes, 10; adjusted to pH 7.2 with NaOH. KB medium contained (mm): KCl, 40; K₂HPO₄, 10; KOH, 105; creatine, 5; sodium pyruvate, 5; taurine, 10; glucose, 11; EGTA, 0.04; Hepes, 5; adjusted to pH 7.2 with glutamic acid. ATP, carbachol and thapsigargin were used at 100, 50 and 0.2 μm, respectively. Mag-fluo-4 AM, fluo-4FF AM and fura-2 AM were from Molecular Probes. All other reagents were from Sigma (Sigma-Aldrich Co., Dorset, UK).

RESULTS

Validation and reproducibility

Simultaneous cytosolic and SR Ca²⁺ records were reproducibly obtained in the uterine myocytes. Figure 1 shows simultaneous measurements of agonist-induced changes in [Ca²⁺]_L and [Ca²⁺]_i. In these experiments, the agonists were dissolved in a Ca²⁺-free solution, while the bath was superfused with Krebs solution. Hence, trans-sarcolemmal Ca²⁺ entry should not occur during agonist application but, upon washout, extracellular Ca²⁺ will be available for SR refilling. Repetitive applications of carbachol (Fig. 1A) elicited very similar responses, as judged by the time course and magnitude, in [Ca²⁺]_i and SR [Ca²⁺]_L (4 cells from 4 animals). Reproducible transients were also obtained with ATP ( cells from 8 animals, Fig. 1B).

Figure 1. Simultaneous and reproducible measurements of changes in SR and cytosolic [Ca²⁺] in response to agonists.

Cells were superfused with Krebs solution (2 mm Ca²⁺). Agonists (in Ca²⁺-free solution) were applied as indicated by the filled bars. Repetitive applications of 50 μm carbachol (A) or 100 μm ATP (B) evoked reproducible [Ca²⁺]_i elevations (lower traces) and [Ca²⁺]_L decreases (upper traces). Left-hand axes, [Ca²⁺]_i; right-hand axes, SR [Ca²⁺]_L.

Magnesium sensitivity

To ensure that the mag-fluo-4 signals were not contaminated by changes in [Mg²⁺], we used the Mg²⁺-insensitive indictor fluo-4FF (loaded in the same way as mag-fluo-4). The two indicators detected similar changes in SR Ca²⁺ content during ATP application, indicating that mag-fluo-4 was indeed detecting [Ca²⁺]_L but not [Mg²⁺]_L (not illustrated). The SR loading with mag-fluo-4 was better than with fluo-4FF, and therefore we performed the rest of the investigation with mag-fluo-4.

Movement artefacts

Single-wavelength indicators are prone to movement artefacts with contraction. We minimised such artefacts by measuring fluorescence from the entire cell and ensuring that the illumination intensity was as uniform and constant as possible. It can be seen in Fig. 2 that upon application of ATP, the fura-2 fluorescence at 340 and 380 nm excitation wavelengths changed in opposite directions. This indicates that the increase in [Ca²⁺]_i was not greatly contaminated by any movement artefact, despite the contraction of the cell, as shown in the images at the bottom of Fig. 2. In all experiments used for analysis, opposite changes in the fura-2 excitation wavelengths occurred.

Figure 2. Changes in [Ca²⁺]_L and [Ca²⁺]_i and cell contraction in response to 100 μm ATP applied in the absence of extracellular Ca²⁺.

Traces show (from top to bottom): normalised mag-fluo-4 fluorescence, and fura-2 fluorescence at 340 and 380 nm excitation (arbitrary units, a.u.). Cell images were taken at the time points indicated by corresponding numbers in the top trace.

Thus, having established that we could directly detect SR Ca²⁺ changes and simultaneously measure [Ca²⁺]_i, we proceeded to investigate the relationship between the two, during agonist-induced changes in [Ca²⁺]_i and spontaneous activity.

Changes in the sr [ca²⁺]_L in response to agonist

In order to investigate in detail and over a prolonged period the effects of agonist on SR [Ca²⁺]_L, both the agonist and the bath perfusate contained 0 Ca²⁺. As shown in Fig. 3A, ATP caused a significant (P < 0.01) rise in [Ca²⁺]_i (from 80 ± 20 to 600 ± 70 nm, mean ±s.d.), and a decrease in SR [Ca²⁺]_L (mag-fluo-4 signal decreased to 0.78 ± 0.13 of its initial value, 11 cells from 4 animals). [Ca²⁺]_i was restored to baseline values in the continued presence of ATP and 0 Ca²⁺ solution. Interestingly, the SR [Ca²⁺]_L started to recover despite the continued presence of ATP and 0 Ca²⁺ solution. This recovery was variable, from 25 to 75 % between cells (compare Figs 1B and 3). There was no apparent effect of agonist removal on the rate of [Ca²⁺]_L recovery. When external Ca²⁺ was restored, a rapid and complete refilling of the SR occurred (Fig. 3A).

Figure 3. Partial re-accumulation of Ca²⁺ during application of agonist.

Changes in [Ca²⁺]_i (bottom traces, left-hand axes) and [Ca²⁺]_L (top traces, right-hand axes) elicited by ATP. A, ATP (100 μm)-induced changes in the absence and then presence of 0.2 μm thapsigargin. B, oscillatory response to ATP and then ATP plus thapsigargin.

Figure 3A also shows that thapsigargin, an inhibitor of the SR Ca²⁺-ATPase, caused a slow decline in SR [Ca²⁺]_L, accompanied by a small increase in [Ca²⁺]_i. ATP, in the presence of thapsigargin, caused a marked and rapid reduction in SR [Ca²⁺]_L. This decrease was significantly larger than that produced by ATP alone (0.54 ± 0.11 and 0.78 ± 0.13, respectively, 4 cells, P < 0.05). It can also be clearly seen that there was no refilling of the SR in the presence of thapsigargin.

In 4/15 cells, from four animals, ATP produced oscillating Ca²⁺ transients, as shown in Fig. 3B. In this cell, ATP evoked four distinct cytosolic Ca²⁺ transients of diminishing amplitude and time course, accompanied by successive waves of Ca²⁺ release and re-uptake by the SR. Again, the restoration of SR Ca²⁺ was incomplete until external Ca²⁺ was restored. In all oscillating cells (4 cells), thapsigargin eliminated oscillations and prevented the SR from refilling (Fig. 3B).

Changes in the sr [ca²⁺]_L during spontaneous activity

In approximately 20 % of cells (from 5 animals), spontaneous contractions and changes in [Ca²⁺]_i were seen (Fig. 4A). These were most probably due to spontaneous action potentials, as their magnitude and kinetics were similar to those recorded from multicellular preparations generating spontaneous electrical activity (compare with Parkington et al. 1999). The rapid increases in [Ca²⁺]_i were abolished in Ca²⁺-free or 10 μm nifedipine solution (not shown). The mean magnitude of the spontaneous [Ca²⁺]_i transients was significantly (P < 0.01) smaller (265 ± 32 nm, 14 measurements on 4 cells, vs. 600 ± 70 nm, 11 measurements) and briefer (7.0 ± 1.5 s compared with 20 ± 5 s), than those in response to ATP. No changes in SR [Ca²⁺]_L could be detected during these spontaneous transients.

Figure 4. SR modulates spontaneous activity.

[Ca²⁺]_i transients due to spontaneous action potentials (bottom trace, left-hand axis); note no measurable changes in [Ca²⁺]_L (top trace, right-hand axis). An increase in the frequency of spikes induced by 30 mm KCl increased [Ca²⁺]_L and suppressed spontaneous activity.

When the frequency of the action potentials was increased by applying 30 mm KCl, the rise in [Ca²⁺]_i was augmented due to the summation of individual transients, and a clear rise in SR [Ca²⁺]_L occurred (Fig. 4). When the cell was repolarised by removal of the high-K⁺ solution, the elevated [Ca²⁺]_L slowly declined. Interestingly, spontaneous activity was suppressed until SR Ca²⁺ levels returned to their initial values (Fig. 4, 4 cells).

DISCUSSION

We have combined mag-fluo-4 with fura-2 to measure [Ca²⁺]_L and [Ca²⁺]_i simultaneously. To the best of our knowledge, there are no previous reports of such measurements in mammalian smooth muscle cells. We found uptake of Ca²⁺ into the SR in the continued presence of calcium-releasing agonists, but this was incomplete in the absence of external Ca²⁺. This refilling required SR Ca²⁺-ATPase activity and the SR Ca²⁺ content decreased substantially when it was inhibited. Agonist-induced oscillations of [Ca²⁺]_i were found to be accompanied by successive Ca²⁺ release and re-uptake by the SR and hence were abolished when it was inhibited. Furthermore, our data indicate that the SR Ca²⁺ content influences spontaneous activity in the uterine myocytes. Low levels of SR Ca²⁺ were associated with spontaneous Ca²⁺ transients and high levels of SR Ca²⁺ were associated with their suppression.

Methodology and validation

There have been few previous studies in any cell type showing direct and simultaneous calcium measurements in cytosol and SR. In a paper on toad stomach cells, ZhuGe et al. (1999) showed a figure where mag-fura-2 was used to record SR Ca²⁺ and Ca²⁺ Green dextran to record cytosolic Ca²⁺. In hepatocytes, Chatton et al. (1995) used mag-fura-2 and fluo-3. Other less direct approaches include recording SR Ca²⁺ and using digital imaging to record cytosolic Ca²⁺ (Golovina & Blaustein, 1997), assessing SR content via Ca²⁺-activated membrane currents (Trafford et al. 1995), or require genetic manipulations (Monteith, 2000).

Therefore, as few previous studies have reported methods to make direct SR Ca²⁺ measurements simultaneously with [Ca²⁺]_i, we sought initially to develop and validate the appropriate methodology. Control experiments showed that both indicators performed in a reliable and reproducible manner. The suitability and reliability were ascertained as follows: (i) when the surface membrane was permeabilised no change in the mag-fluo-4 fluorescence occurred, indicating little or no mag-fluo-4 in the cytosol; (ii) large alterations of mitochondrial function with CCCP did not decrease the mag-fluo-4 signal, indicating that mitochondrial Ca²⁺ was not contributing to the SR signal, in agreement with ZhuGe et al. (1999); (iii) data with the Mg²⁺-insensitive dye fluo-4FF reported similar changes to SR Ca²⁺ as mag-fluo-4, indicating that changes in [Mg²⁺] were not influencing the SR Ca²⁺ signal, in agreement with others (Hofer & Machen, 1993; Golovina & Blaustein, 1997); and (iv) the behaviour of both indicators was consistent with their proposed locations, e.g. upon agonist application the mag-fluo-4 signal decreased and the fura-2 signal increased, and thapsigargin produced a large decrease in the former, with little alteration of the latter.

The signals remained stable for up to 40 min and gave reproducible changes in response to repeated manoeuvres e.g. application of agonist, and were free from movement artefacts. Single-wavelength indicators are difficult to calibrate in terms of absolute [Ca²⁺], hence we show the mag-fluo-4 fluorescence in arbitrary units. We appear to be within the working range of mag-fluo-4 (non-saturated) as both increases and decreases from the resting levels could be detected, although it is appreciated that caution must be used when drawing any quantitative rather than qualitative conclusions.

Effect of agonists

The rise in [Ca²⁺]_i produced by agonists in the absence of external Ca²⁺ was accompanied by a decrease in [Ca²⁺]_L. When the SR was allowed to recover between applications of agonist, this decrease in SR Ca²⁺ was of a very similar magnitude and rate each time. The agonist concentrations were such as to give maximal stimulation, yet as seen in Fig. 3 and 4 with thapsigargin, the SR still contained Ca²⁺. This lack of emptying of the SR could be due to inactivation of the IP₃ receptor, the bell-shaped relationship between Ca²⁺ and IP₃ effectiveness (Hirose et al. 1998) or Mg²⁺ activating the mag-fluo-4, or may represent a pool of Ca²⁺ that is not accessed by the agonist. Work on cultured human uterine myocytes loaded with fluo-3FF to visualise the SR also showed that agonists did not deplete the store under normal conditions, but did when thapsigargin was present (Young & Mathur, 1999).

SR refilling

The data also clearly show that the SR started taking up Ca²⁺ and refilling in the continued presence of agonist. The amount of this refilling was variable between cells. As thapsigargin inhibited this uptake, it is clear that it was brought about by the SR Ca²⁺-ATPase. There is evidence that high cytosolic and low luminal [Ca²⁺] activate the Ca²⁺-ATPase (Saiki & Ikemoto, 1997; Mogami et al. 1998). Thus, if the SR Ca²⁺ content is a balance between uptake and release, the behaviour of the SR in the presence of agonist is that of a temporal sequence of IP₃-induced Ca²⁺ release followed by desensitisation to IP₃ and low [Ca²⁺]_L with elevated [Ca²⁺]_i favouring SR uptake. In the absence of extracellular calcium, this restoration was only partial, i.e. [Ca²⁺]_L never reached pre-stimulation levels. These data indicate that in the absence of extracellular Ca²⁺ much of the Ca²⁺ released from the SR is extruded by the surface membrane calcium transporting systems such as Ca²⁺-ATPase and the Na⁺-Ca²⁺ exchanger, in agreement with previous data (Shmigol et al. 1998b, 1999b). Our data also show that the cytoplasmic Ca²⁺ signal was terminated, i.e. resting levels were restored, more rapidly than SR Ca²⁺ level recovery. This has been reported for other cell types in which it has been studied, e.g. ZhuGe et al. (1999) and Mogami et al. (1998). In some cells, ATP evoked [Ca²⁺]_i oscillations accompanied by reciprocal changes in [Ca²⁺]_L. Oscillations in [Ca²⁺]_i have been described in other smooth muscle preparations following electrical or agonist stimulation (Kasai et al. 1997; Miriel et al. 1999). Although these authors did not measure SR Ca²⁺ content, they did show that the oscillations were not abolished by nifedipine, but were by cyclopiazonic acid, another blocker of the SR Ca²⁺-ATPase (Miriel et al. 1999), or ryanodine (Prakash et al. 1997). Our experiments have directly shown that Ca²⁺ release and uptake by the SR accompany the oscillations and that SR Ca²⁺ is essential for their occurrence. Thus although the precise mechanism of [Ca²⁺]_i oscillations in smooth muscle is still unknown, it is clear from our data that the sarco-endoplasmic reticulum Ca²⁺-ATPase (SERCA) pump is essential for their generation.

Spontaneous contractions

Our most intriguing finding is that there was no appreciable change in [Ca²⁺]_L during spontaneously generated [Ca²⁺]_i transients. Pregnant rat myometrial cells have been reported to have all three ryanodine receptor isoforms (Martin et al. 1999), and thus would be expected to respond to the rise in [Ca²⁺]_i. The lack of response might be due to (i) the SR playing no role in Ca²⁺ handling during spontaneous activity or (ii) changes in [Ca²⁺]_L taking place that are far too localised or small to be resolved with our technique. We also found that an increased SR Ca²⁺ content was associated with a suppression of the spontaneous [Ca²⁺]_i transients (see Fig. 4). These data support the idea that the uterine SR free Ca²⁺ content does not change during spontaneous contraction, and that the SR may provide a negative feedback to the surface membrane when it is full and thereby regulates Ca²⁺ entry through voltage-gated Ca²⁺ channels. Thus when the SR is full either stimulated or spontaneous (sparks) Ca²⁺ releases from the SR activate K_Ca and reduce [Ca²⁺]_i and contraction, as shown in vascular smooth muscle (Nelson et al. 1995). Further experiments will be required to determine the details and physiological importance of this process and the relationship of any sparks to K_Ca activity.

Conclusions

In summary, we have developed the use of dual labelling with Ca²⁺-sensitive indicators in smooth muscle cells. We have found that the free Ca²⁺ content in the SR does not measurably change with spontaneous contractions, suggesting that it is not involved in their generation, verifying previous studies in intact tissue. Re-filling of the SR occurs in the presence of agonist and this presumably reflects the SR activity as a sink for Ca²⁺, which will be extruded via the surface membrane efflux mechanisms. Oscillations of [Ca²⁺]_i produced in response to agonist stimulation are due to Ca²⁺ release and re-uptake into the SR.