Voltage-gated K⁺ currents in freshly isolated myocytes of the pregnant human myometrium

Abstract

1. Voltage-gated K⁺ currents in human myometrium are not well characterized, and were therefore investigated, using the whole-cell patch clamp technique, in freshly isolated myometrial smooth muscle cells from pregnant women at term.

2. Three types of voltage-gated K⁺ currents were identified. I_K1 was a 4-aminopyridine-insensitive current with a negative half-inactivation (V_0.5 = -61 to -67 mV) and negative activation characteristics (threshold between -60 and -40 mV) and slow kinetics. I_K2 was a 4-aminopyridine-sensitive current (half-maximal block at ≈1 mM) with relatively positive half-inactivation (V_0.5 = -30 mV) and activation characteristics (threshold between -40 and -30 mV) and faster kinetics. I_K,A was a 4-aminopyridine-sensitive current with a negative inactivation and very fast inactivation kinetics.

3. Both I_K1 and I_K2 were sensitive to high concentrations of tetraethylammonium (half-maximal block at ≈3 mM) and low concentrations of clofilium (half-maximal block by 3-10 μM).

4. I_K1 and I_K2 were unevenly distributed between myometrial cells, most cells possessing either I_K1 (30 cells) or I_K2 (24 cells) as the predominant current.

5. The characteristics of these currents suggest a possible function in the control of membrane potentials and smooth muscle quiescence in the pregnant human myometrium.

The muscle of the uterus, the myometrium, is a spontaneously active smooth muscle which remains largely quiescent throughout most of pregnancy, and then at term provides powerful rhythmic contractions in order to expel the fetus. These contractions are intimately associated with fluctuations in membrane potential (Parkington, 1985; Kawarabayashi et al. 1986; Buhimschi et al. 1997). The involvement of K⁺ channels in the quiescence of the uterus during pregnancy and the initiation of contractions in labour, as well as the ionic nature of membrane potential changes in the myometrium, is not fully understood (Parkington & Coleman, 1988). Early studies in rat and guinea-pig myometrium suggested that K⁺ conductances might contribute to the resting membrane potential and slow wave depolarizations (Bulbring et al. 1968; Kuriyama & Suzuki, 1976), and to termination of action potentials (Jmari et al. 1986). The nature of these currents has not yet been defined other than that they are to some extent 4-aminopyridine (4-AP) and tetraethylammonium (TEA) sensitive (Jmari et al. 1986). In the human myometrium, calcium-activated K⁺ currents have been characterized in some detail (Anwer et al. 1993; Khan et al. 1993, 1997; Perez et al. 1993) and it has been suggested that these currents are important for terminating action potentials and preventing long tetanic contractions (Khan et al. 1997).

Voltage-gated K⁺ currents in the human myometrium, however, have only been investigated in any detail in cultured cells from non-pregnant women, where two 4-AP-sensitive currents were described. The decay rates of these currents suggested that they might represent A-type and delayed rectifier-type K⁺ currents (Erulkar et al. 1993). In other spontaneously active smooth muscles, including the dog and guinea-pig colon, at least three distinct types of voltage-gated K⁺ current have been described and potentially important functions for these currents in regulating membrane potential changes and hence cell excitability were suggested (Thornbury et al. 1992; Vogalis et al. 1993; Carl, 1995). A fuller understanding of the role of voltage-gated K⁺ currents in myometrial function may also lead to new approaches to the management of pre-term labour, as has already been suggested for the calcium-activated and ATP-sensitive K⁺ currents (Khan et al. 1998). In the present study, therefore, we have investigated the electrophysiological and pharmacological characteristics of voltage-gated K⁺ currents in freshly isolated myometrial smooth muscle cells from pregnant women.

METHODS

Cell isolation

Myometrial biopsies were taken from the middle of the upper edge of the lower segment incision in women undergoing routine elective Caesarean section at term (with the patients’ informed consent and the approval of the St Thomas’ Hospital ethics committee) and placed immediately into cold physiological salt solution (PSS). Small pieces of myometrium (2-3 mm³) were incubated for 45 min at 37°C in low-Ca²⁺ PSS containing collagenase (a 1:1 mixture of Sigma Types I and XI, total 2.5 mg ml⁻¹). After the enzymatic digestion, tissue pieces were washed in enzyme-free low-Ca²⁺ PSS and dispersed into single cells by trituration using a wide bore glass pipette. The cells were stored in low Ca²⁺ PSS at 4°C and used within 5-6 h of isolation.

Solutions and drugs

The bath solution (PSS) contained (mM): NaCl, 130; KCl, 5.0; MgCl₂, 1.2; CaCl₂, 1.5; Hepes, 10; glucose, 10. The pH was adjusted to 7.4 with NaOH. Low-Ca²⁺ PSS contained 15 μM Ca²⁺. The pipette solution contained (mM): KCl, 110; MgCl₂, 2.5; MgATP, 1.0; Hepes, 10; EGTA, 10; and the pH was adjusted with KOH. Paxilline, tetraethylammonium and 4-AP were all from Sigma UK, and were dissolved in distilled water. The pH of the 4-AP stock solution (100 mM) was adjusted to 7.4 using HCl. Clofilium was obtained from Research Biochemicals International and was dissolved in dimethyl sulphoxide (DMSO). At the highest final concentration of 0.1 %, DMSO had no significant effect on K_V currents. All other chemicals were obtained from BDH.

Data analysis and statistics

The whole-cell membrane currents were recorded at room temperature using the standard patch clamp technique, with pCLAMP 6 software and Axopatch 200B amplifier (Axon Instruments). Pipette resistance was between 3 and 5 MΩ. Cell capacitance (average ∼100 pF) was not routinely compensated. Calculated average series resistance was ∼10 MΩ which gave on average a maximum voltage error for K_V currents (average amplitude ∼1 nA) of ∼10 mV. Leak subtraction was routinely performed after recording using an algorithm in pCLAMP 6 designed for this purpose. Data analysis, statistics and curve fitting were performed with Microsoft Exel and SigmaPlot 4.01 (Jandel Scientific, San Rafael, CA, USA) software on an Elonex MTX-6233/II computer. Statistical significance was considered to exist when P < 0.05 using Student's unpaired two-tailed t test (unless otherwise stated).

RESULTS

Components of the whole-cell current

Components of the K⁺ current were initially characterized in seven cells using the pharmacological agents we have previously found to be effective in other smooth muscles (Smirnov & Aaronson, 1992; Smirnov et al. 1994). Depolarizations for 300 ms up to +60 mV in 20 mV increments from a holding potential of -80 mV were applied, firstly in normal PSS, then in 10 mM TEA, and then in the presence of both 10 mM TEA and 5 mM 4-AP.

Figure 1 shows whole-cell currents in a representative cell over a range of potentials in PSS (A), TEA (B), and both TEA and 4-AP (C). The large amplitude outward current in PSS was considerably reduced in the presence of 10 mM TEA. The further addition of 5 mM 4-AP caused additional but incomplete block in all seven cells and revealed a large transient inward current at -40, -20, 0 and +20 mV. This inward current was characteristic of the Ca²⁺ current previously described in the myometrium (Inoue et al. 1990; Young et al. 1993; Sawdy et al. 1998), and was in some cells observed even in the absence of inhibitor.

Figure 1. Whole-cell currents in a representative cell.

Currents elicited by 300 ms depolarizations to -80, -60, -40, -20, 0, +20, +40 and +60 mV from a holding potential of -80 mV using normal Ca²⁺-containing PSS (A), PSS containing 10 mM TEA (B) and PSS containing 10 mM TEA and 5 mM 4-AP (C) as the bath solution. The currents shown in D represent the 4-AP-sensitive component and are derived by subtracting the currents in C from the currents in B. Only the first 100 ms are shown in D to emphasize the transient component. Also note that the capacitance artefact is absent from D because of the subtraction.

The 4-AP-sensitive current (derived by subtracting currents present in both TEA and 4-AP from currents present in TEA only) is shown in Fig. 1D. This current had two components, a transient rapidly activating (threshold negative of -40 mV) and rapidly inactivating component and a sustained slowly inactivating component. The availability of the fast transient component of the 4-AP-sensitive current was measured in two cells, using 10 s conditioning potentials and a test potential of +60 mV. The resulting half-inactivation potentials (V_0.5) and slope factors (k) were -69.5 and 5 mV, and -70 and 6 mV for the two cells, respectively (data not illustrated). These characteristics are similar to those described for A-type K⁺ currents (I_K,A) as reported previously in myometrium (Erulkar et al. 1993, 1994).

Approximately one-quarter of the sustained outward current was not blocked by the combination of 10 mM TEA and 5 mM 4-AP (Fig. 1C). Our preliminary data showed that this K⁺ current inactivated over a relatively negative range of membrane potentials (Smirnov et al. 1995). In all subsequent experiments (unless otherwise stated) we modified the bath solution in order to isolate and further characterize the properties of this voltage-gated current. Firstly, Ca²⁺ was excluded and 0.5 mM CdCl₂ added to the bath solution to eliminate the Ca²⁺ current and diminish the Ca²⁺-activated current (I_K(Ca)). Also, Cd²⁺ is known to block I_K,A in heart (Stengl et al. 1998) and ureter smooth muscle (Imaizumi et al. 1990) and in subsequent experiments under these conditions, no further A-like currents were observed. In addition, 1 mM TEA (or in some experiments 1 μM of the selective blocker of I_K(Ca), paxilline (Sanchez & McManus, 1996), was continuously present in the bath solution to block the remaining I_K(Ca). The effects of TEA on these currents and the rationale for the use of 1 mM TEA are described in a later section.

Under these conditions, the addition of 5 mM 4-AP was found to have markedly different actions from one cell to another. In some cells, 4-AP caused considerable inhibition of the current, while in others there was very little inhibition or even a slight increase in the current when 4-AP was applied. This suggested the presence of at least two K⁺ currents, which could be distinguished by their 4-AP sensitivity, and which varied in their relative proportions between cells. Subsequent experiments supported this possibility.

Availability of the voltage-gated current

Using Ca²⁺-free PSS containing 0.5 mM CdCl₂ and 1 mM TEA in the bath solution, availability of the voltage-gated K⁺ current was investigated with the following experimental protocol. The membrane potential was held at -80 mV and 10 s conditioning pre-pulses to between -100 and +10 mV were applied, followed by a 200 ms test pulse to +60 mV (at 0.03 Hz). This protocol was then repeated in the presence of 5 mM 4-AP.

Figure 2A shows the currents elicited by the +60 mV test pulse following conditioning voltages to between -100 mV and +10 mV, in one cell. The current began to inactivate after conditioning potentials positive to -90 mV, and showed little further inactivation positive to -40 mV. The addition of 5 mM 4-AP did not on average affect the overall current amplitude following the -100 mV conditioning potential, but appeared to increase the component of the current that inactivated following the conditioning pulse to -40 mV, without significantly affecting the half-inactivation value (Fig. 2B). By contrast, in another morphologically similar cell, the same voltage protocol revealed a current that only began to inactivate when conditioning potentials positive to -50 mV were applied, and which was almost completely inactivated at 0 mV (Fig. 2C). This current was almost completely blocked by 5 mM 4-AP (not shown, but see Fig. 4).

Figure 2. Inactivation of voltage-gated currents.

A-C, examples of the 4-AP-insensitive current recorded at +60 mV following 10 s conditioning pre-pulses to between -100 and +10 mV (at 10 mV increments) in the absence (A) and presence (B) of 5 mM 4-AP. C shows the inactivation of the 4-AP-sensitive current for potentials between -90 and +10 mV in a different cell under conditions identical to those in A. Currents following the pre-pulses to -80, -40 and 0 mV are indicated by arrows. The dotted lines represent zero current. D, mean ±s.e.m. end of pulse measurements of the currents represented in A-C, normalized to the current elicited following the -100 mV conditioning pre-pulse. Points show the 4-AP-insensitive current in the absence (•, n = 6) and presence of 5 mM 4-AP (^, n = 6) and the 4-AP-sensitive current in the absence of 4-AP (▪, n = 5).

Figure 4. Decay and 4-AP sensitivity of I_K1 and I_K2.

Depolarizations for 10 s to between -50 and +10 mV from a holding potential of -80 mV for I_K1 (A and B) and I_K2 (C and D). A, I_K1 in the absence of 4-AP. B, I_K1 in the presence of 5 mM 4-AP in the same cell as in A. C, I_K2 in the absence of 4-AP. D, I_K2 in the presence of 5 mM 4-AP in the same cell as in C. The dotted lines represent zero current.

A similar correlation between the voltage range of current availability and 4-AP sensitivity appeared to exist in most cells. This correlation was further investigated using several approaches. Firstly, cells were divided into two groups on the basis of a strong (> 70 % block) or weak (little or no inhibition, or an increase, in current amplitude) block of the current by 5 mM 4-AP. The availability of the current was assessed in each cell by measuring current amplitude at the end of a 200 ms test pulse to +60 mV, which followed a conditioning potential to between -100 and +10 mV. Figure 2D illustrates the mean availability of the current in the 4-AP-sensitive (▪) and insensitive (•) groups of cells. The open circles represent the availability of the 4-AP-insensitive cells in the presence of 5-mM 4-AP. All data points have been normalized to the current amplitude observed after the conditioning step to -100 mV. In the absence and presence of 4-AP, the amplitudes of the test pulses after the conditioning step to -100 mV were 1095 ± 337 and 1269 ± 429 pA, respectively (non-significant, n = 6).

The data illustrated in Fig. 2D were fitted to the Boltzmann function in order to derive the parameters V_0.5 (half-inactivation potential), k (slope factor) and A (fraction of the current not inactivated). In the absence of 4-AP, V_0.5 of the current in the 4-AP-insensitive cells was -65 ± 3 mV, the non-inactivating component (A) was 36 ± 2 % and k was 8 ± 2 mV (n = 6). In the presence of 5 mM 4-AP, V_0.5 in these cells was -68 ± 3 mV, the non-inactivating component was 37 ± 4 % and k was 9 ± 4 mV (n = 6). Conversely, the V_0.5 of the current in the group of cells sensitive to 4-AP was -30 ± 1 mV. The non-inactivating component of the current in this group of cells was 15 ± 4 % and k was 5 ± 1 mV (n = 6). The differences in the half-inactivation potentials and size of the non-inactivating component between the two groups of cells were highly significant (P < 0.001).

We shall hereafter refer to the low threshold current which was not inhibited or slightly increased by 5 mM 4-AP, and which reached its maximal extent of inactivation at voltages negative to -30 mV (• in Fig. 2D), as I_K1. The high threshold current, which was inhibited by 4-AP and inactivated between -50 and 0 mV (▪ in Fig. 2D) will be termed I_K2.

The small degree of overlap which existed between the potential ranges over which I_K1 and I_K2 inactivated suggested that it would be possible to investigate the properties of each current by utilizing different holding potentials. The following three-step inactivation protocol based on the distinctive inactivation properties of these currents was therefore applied. Current was recorded at the end of a 200 ms test pulse to +60 mV firstly after holding at -80 mV, then after a 10 s pre-pulse to -40 mV in order to inactivate I_K1, and then after a 10 s pre-pulse to 0 mV to inactivate I_K2. The amplitude of I_K1 (in cells where I_K1 was the predominant current) was therefore estimated as the current remaining after the -40 mV pre-pulse subtracted from the current present after the -80 mV holding potential, and the amplitude of I_K2 (in cells where I_K2 was the predominant current) was estimated from the current remaining after the pre-pulse to 0 mV subtracted from the current remaining after the pre-pulse to -40 mV.

In the 14 cells where the three-step inactivation protocol was carried out in the absence and presence of 5 mM 4-AP, the cells were clearly divided into two groups (Fig. 3A). In one group (I_K1), current was inactivated by pre-pulses to -40 mV and either unaffected or enhanced by 5 mM 4-AP (n = 7, ▪). No further inactivation was observed in these cells when the conditioning potential was set at 0 mV (□). In the other group (I_K2), current was almost completely inhibited by 5 mM 4-AP, and was only slightly inactivated at -40 mV (•, n = 7). The current in this group of cells was then almost completely inactivated by prepulses to 0 mV (^).

Figure 3. Differential distribution of I_K1 and I_K2.

A, relationship between inactivation and sensitivity to 5 mM 4-AP for I_K1 (n = 7) and I_K2 (n = 7) currents. The percentage of current at +60 mV inactivated after a 10 s conditioning pre-pulse to either -40 mV (I_K1, ▪; I_K2, •) or 0 mV (I_K1, □; I_K2, ^) is plotted against the percentage change in peak current amplitude caused by the application of 5 mM 4-AP during a step to 0 mV from a holding potential of -80 mV. B, frequency distribution of inactivation of the current at +60 mV by a 10 s conditioning pulse to -40 mV in 42 cells, expressed as a fraction of the total inactivating component using the degree of inactivation by a pre-pulse to 0 mV as maximum. Data are binned at intervals of 0.1 (equivalent to 10 % of total inactivation). Two peaks are visible, one at 0.2 (corresponding to I_K2) and one at 0.9 (corresponding to I_K1).

The three-step inactivation protocol described above was carried out in 42 cells. Figure 3B depicts the extent to which the conditioning potential of -40 mV inactivated the outward current, relative to the inactivation recorded at 0 mV in each cell. These results show that two distinct groups of cells could be distinguished on the basis of the degree to which the conditioning potential-sensitive component was inactivated at -40 mV. The peak on the left represents cells in which the outward current showed relatively little inactivation at -40 mV (i.e. cells where I_K2 was dominant), while the peak on the right represents cells in which the outward current was markedly inactivated at -40 mV (cells where I_K1 was dominant). Note that the four bars on the extreme right represent cells in which the current was somewhat larger after the conditioning step to 0 mV than it was after the conditioning step to -40 mV. This effect may have been due to some activation of the minor residual component of I_K(Ca) during the conditioning step to 0 mV.

The membrane capacitance of the cells comprising each peak in Fig. 3B was calculated from the area under the capacitative artefacts as an indication of the cell surface area, and therefore cell size (Smirnov et al. 1994). The cell capacitance was 99.9 ± 9.6 pF in the 24 cells in which I_K1 predominated, and 110.6 ± 12.6 pF in the 18 cells in which I_K2 was dominant; these values were not significantly different. To our knowledge, all biopsies were taken from the same region of the myometrium from women at term who were not in active labour, and under light microscopy there were no obvious morphological differences between the two groups of cells. There was no preferential expression of either I_K1 or I_K2 in cells from any one biopsy, each isolation often producing some cells with only I_K1 and some with only I_K2.

Activation and decay

Figure 4 illustrates the effect of 5 mM 4-AP on the outward current recorded over 10 s pulses to between -50 and +10 mV from a holding potential of -80 mV. Figure 4A and B shows currents observed between -50 and +10 mV in a cell where I_K1 predominated, in the absence (A) and presence (B) of 5 mM 4-AP. Figure 4C and D shows current in a cell where I_K2 predominated, in the absence (C) and presence (D) of 5 mM 4-AP. I_K1 current appeared to be slightly enhanced by 5 mM 4-AP in the example shown here, but in a group of five cells the difference did not reach statistical significance (7 ± 10 % enhancement at 10 s, n = 5). I_K2 on the other hand, was blocked significantly by 5 mM 4-AP (63 ± 11 %, n = 6, P < 0.05).

Figure 4 also illustrates two other differences between I_K1 and I_K2. Firstly, I_K1 current was apparent at -40 to -50 mV, whereas for I_K2 current visibly activated between -40 and -30 mV. Secondly, I_K1 decayed significantly more slowly than I_K2 (66.2 ± 4.7 % decrease in amplitude over 10 s for I_K1vs. 82.7 ± 5.6 % over 10 s for I_K2, P < 0.05). To confirm this, 10 s currents were fitted with exponential curves and time constants (τ) calculated. I_K1 decay was best fitted to a double exponential with a constant whereas I_K2 decay was well fitted to a single exponential with a constant component. For I_K1 at +10 mV: τ₁ was 1016 ± 316 ms and was 22 ± 4 % of total; τ₂ was 5213 ± 1068 ms and 46 ± 5 % of total, and the constant component was 32 ± 3 % of total (n = 6). For I_K2 at +10 mV: τ was 2552 ± 233 ms and was 89 ± 3 % of total, and the constant component was 11 ± 3 % (n = 6). τ₂ was the largest component of I_K1 and was significantly larger than the τ value of I_K2 (P < 0.01). The constant component was also significantly greater for I_K1 than for I_K2 (P < 0.01). Maintaining cells with either I_K1 or I_K2 present at a holding potential of 0 mV for several minutes resulted in no net outward current, suggesting that the constant component of the 10 s currents was in fact a very slowly inactivating component.

The kinetics of activation for I_K1 and I_K2 were determined by fitting current traces to a first order power function with time constant (τ). Both types of current were found to fit better to this function than to exponential or higher order power functions. Figure 5 shows the fits for representative I_K1 and I_K2 currents at potentials between -40 and +10 mV and the mean time constants calculated from those fits in six and five cells, respectively, plotted against membrane potentials. As is shown in Fig. 5, in addition to inactivating more slowly, I_K1 activated significantly more slowly than I_K2 at all potentials.

Figure 5. Activation kinetics of I_K1 and I_K2.

A and B, currents to between -40 and +10 mV in cells where I_K1 (A) and I_K2 (B) predominate with fitted curves superimposed (see text for fitting procedure). C, a plot of time constants (tau) against membrane potential (mean ±s.e.m.) (•, I_K1; ▪, I_K2). Asterisks represent significant differences at P < 0.05. The dotted lines represent zero current.

Block of I_K1and I_K2 by TEA

The sensitivity of I_K1 and I_K2 to TEA was determined by investigating the effect of increasing concentrations of this agent (1, 3, 10 and 30 mM) in cells where either I_K1 or I_K2 predominated. As described above, I_K1 was further isolated by defining it as the difference current which was present at a holding potential of -80, but not -40 mV, while I_K2 was defined as the difference current which was present at a holding potential of -40, but not 0 mV. In addition, to determine the extent to which I_K(Ca) was blocked by 1 mM TEA, a TEA concentration-response curve was constructed when the membrane potential was held for several minutes at 0 mV in order to inactivate all voltage-gated K⁺ currents. Figure 6A shows the percentage inhibition of I_K1, I_K2 and I_K(Ca) by TEA. I_K1 and I_K2 demonstrated partial block by TEA. At 1 mM, both I_K1 and I_K2 were significantly inhibited (by 24 ± 8 and 29 ± 5 %, respectively, n = 5-6), in contrast to I_K(Ca) which was inhibited to a much greater extent (75 ± 3 %, n = 4, P < 0.01 compared with both I_K1 and I_K2).

Figure 6. Effect of TEA on I_K1, I_K2 and I_K(Ca).

A, dose-response curve for TEA, plotting the mean ±s.e.m. percentage inhibition of I_K1 (□, n = 5-6), I_K2 (▪, n = 5-6) and I_K(Ca) (•, n = 4) against TEA concentration. All three currents were significantly inhibited by all concentrations of TEA (P < 0.05). B, I-V relationship for I_K(Ca) in 4 paired cells in the absence (•) and presence (^) of 1 mM TEA, plotting the mean ±s.e.m. of current (normalized to the current at +60 mV in the absence of TEA) against membrane potential. I_K1 and I_K2 were recorded using the three-step inactivation protocol (see text). I_K(Ca) was recorded in normal PSS (1.5 mM Ca²⁺) from a holding potential of 0 mV and a test potential of +60 mV.

The I-V relationships for I_K(Ca) in the absence and presence of 1 mM TEA (in 1.5 mM Ca²⁺ PSS) are described in Fig. 6. The I-V curves, as well as the lack of outward current at the holding potential of 0 mV, suggested that I_K(Ca) activated at potentials positive to 0 mV. There was a small but insignificant additional inhibition of this current when in addition to the application of 1 mM TEA, Ca²⁺ was removed and 0.5 mM CdCl₂ was present (74 ± 5 % in PSS, n = 4, vs. 80 ± 6 % in Ca²⁺-free PSS with 0.5 mM CdCl₂, n = 4).

These experiments suggested that 1 mM TEA caused a small but significant inhibition of I_K1. To determine whether this partial block by TEA was affecting the voltage-dependent characteristics of I_K1, availability experiments with pre-pulses between -100 and +10 mV were repeated in Ca²⁺-free PSS containing 1 μM paxilline, an alternative blocker of I_K(Ca), as well as 5 mM 4-AP and 0.5 mM CdCl₂. Availability experiments were also carried out in the presence of normal-[Ca²⁺] PSS containing 10 mM TEA and 5 mM 4-AP. Paxilline (1 μM) was found to cause inhibition of I_K(Ca) comparable to that caused by 10 mM TEA (using a 0 mV holding potential; data not shown). The inactivation of I_K1 under both sets of conditions was similar to that described in Fig. 2. Half-inactivation potentials and the proportion of non-inactivating current under all three sets of conditions are summarized in Table 1.

Table 1. Effect of bath solution on I_K1 inactivation characteristics.

Bath solution	1 μM paxilline, 0.5 mM CdCl2, 5 mM 4-AP	n	1 mM TEA, 0.5 mM CdCl2	n	1 mM TEA, 0.5 mM CdCl2,5 mM 4-AP	n	10 mM TEA, 1.5 mM Ca2+,5 mM 4-AP	n
V0.5 (mV)	−61 ± 3	5	−65 ± 3	6	−66 ± 2	6	−67 ± 3	5
k (mV)	8 ± 0.4	5	8 ± 2	6	9 ± 4	6	8 ± 2	5
Non-inactivating component (%)	43 ± 3	5	36 ± 2	6	36 ± 1	6	30 ± 5	5

Half-inactivation potentials (V_0.5) and percentage of I_K1 at +60 mV which does not inactivate, calculated by fitting steady-state inactivation data to Boltzmann distribution. There are no significant differences in any of the three variables between the four sets of conditions.

The threshold of activation of I_K1 when paxilline was used to block I_K(Ca) was also similar to that seen with 1 mM TEA. To investigate this further, the I-V relationship of I_K1 was constructed using depolarizations to between -80 and +60 mV from a holding potential of -80 mV, in a Ca²⁺-free bath solution containing 1 μM paxilline, 0.5 mM CdCl₂ and 5 mM 4-AP. Figure 7 shows an example trace (A), the mean normalized I-V relationship (B) and the normalized current activation relationship calculated from this data (C, n = 5). The activation relationship was calculated from the equation y =I/(V_m − E_r) where I is current amplitude, V_m is the membrane potential and E_r is the reversal potential (calculated as -83 mV). These data were then fitted to the Boltzmann equation and the mean half-activation potential and slope factor in five cells were calculated as +1.2 ± 4.4 mV and -16 ± 1 mV, respectively.

Figure 7. Current-voltage relationship of I_K1.

A, currents elicited by 300 ms depolarizations to between -80 and +60 mV from a holding potential of -80 mV, using 1 μM paxilline, 0.5 mM CdCl₂ and 5 mM 4-AP in a Ca²⁺-free bath solution. B, plot of current measured at the end of the 300 ms pulse against membrane potential, normalized to the current at +60 mV (▪). C, activation relationship of I_K1 plotting current, converted by the equation y =I/(V_m − E_r) (see text) and normalized to the +60 mV values, against membrane potential (•).

The blockade of I_K1 by 1 mM TEA was also measured in the presence of 1 μM paxilline, defining I_K1 as the difference current present at a holding potential of -80 but not -40 mV in cells where this current predominated. Under these conditions, 1 mM TEA inhibited the current by 27 ± 1 % (n = 5). This value was similar to that recorded in the absence of paxilline.

Effects of 4-AP and clofilium on I_K1and I_K2

The effects of increasing concentrations of 4-AP (0, 1, 2, 5, 10 and 20 mM) on I_K1 and I_K2 were determined using the three-step inactivation protocol described above (pre-pulses to -80, -40 and 0 mV), and are shown in Fig. 8A and B, respectively. In Fig. 8C and D, the relative inhibition (or enhancement) by each concentration of 4-AP on I_K1 and on I_K2 components, respectively, are plotted. I_K1 was significantly enhanced by 10 and 20 mM 4-AP (P < 0.05, n = 5), whereas I_K2 was significantly inhibited by all concentrations of 4-AP tested (P < 0.001 for each concentration, n = 5).

Figure 8. Effect of 4-AP on I_K1 and I_K2.

A and B, plots of 4-AP concentration against measurements of currents elicited by +60 mV test pulses following 10 s conditioning pre-pulses to -80 mV (▪, I_K1; □, I_K2), -40 mV (•, I_K1; ^, I_K2) and 0 mV (▾, I_K1; ▿, I_K2). Points are the mean ±s.e.m. of currents normalized to the current after the -80 mV pre-pulse in the absence of 4-AP. A, 5 cells demonstrating I_K1 as the predominant current; B, 5 cells demonstrating I_K2 as the predominant current. C and D, 4-AP dose-response graphs for I_K1 (C) and I_K2 (D) plotting the mean ±s.e.m. percentage inhibition of current against 4-AP concentration. I_K1 and I_K2 were defined using the three-step inactivation protocol (see text). Asterisks show significant enhancement of I_K1 (*P < 0.05) and inhibition of I_K2 (**P < 0.001).

The effects of several concentrations of clofilium (a blocker of voltage-gated K⁺ currents) on I_K1 in a representative cell are shown in Fig. 9A. The current was elicited by a 10 s pulse to 0 mV from a holding potential of -80 mV in Ca²⁺-free PSS containing 0.5 mM CdCl₂ and 1 mM TEA. The predominance of I_K1 was confirmed in each cell with the three-step inactivation protocol. Decay of I_K1 was enhanced in a concentration-dependent manner by clofilium such that at 100 μM, the current was nearly abolished after 10 s. Percentage block of this current at the peak and end of the pulse are presented in Fig. 9B (n = 5-8). In cells where I_K2 was the predominant current (also defined using the three-step inactivation protocol), I_K2 was also profoundly inhibited by 100 μM clofilium (90 ± 5 % at 200-300 ms, 96 ± 3 % at 10 s, n = 4; Fig. 9C).

Figure 9. Effect of clofilium on I_K1 and I_K2.

A, currents elicited by 10 s pulses to 0 mV from a holding potential of -80 mV, in a cell where I_K1 predominated in the absence and presence of 3, 10, 30 and 100 μM clofilium. The dotted lines represent zero current. B, dose-response curve plotting the mean ±s.e.m. percentage inhibition of I_K1 measured at 200-300 ms (•) and 10 s (^) against clofilium concentration (n = 5-8). C, effect of 100 μM clofilium on I_K2 under conditions identical to those for I_K1 in A.

DISCUSSION

Our results suggest that there are at least four K⁺ currents in isolated pregnant human myometrial cells, three voltage-gated currents (I_K1, I_K2 and I_K,A) and a Ca²⁺-activated current (I_K(Ca)). In the present report, we have focused on characterizing I_K1 and I_K2, with an emphasis on the former current, which has unusual properties.

Isolation and characterization of I_K1

The isolation of I_K1 required the elimination of the other currents present in these cells. I_K(Ca) was routinely minimized firstly by excluding Ca²⁺ from, and adding 0.5 mM Cd²⁺ to, the bath solution. This had the additional benefits of blocking the voltage-gated Ca²⁺ current, and presumably I_K,A also (Imaizumi et al. 1990), since no transient 4-AP-sensitive current was present under these conditions. Secondly, either 1 mM TEA or 1 μM of the selective antagonist paxilline (Sanchez & McManus, 1996) was utilized to block I_K(Ca). Both combinations of conditions reduced the outward current measured at +60 mV from a holding potential of 0 mV, by about 80 %. The inactivation of the voltage-gated currents at this holding potential, combined with the concentration dependence of the block of this current by TEA (Fig. 6), suggests strongly that this current was almost entirely I_K(Ca).

Once I_K(Ca) and I_K,A were blocked, isolation of I_K1 then required the removal of I_K2. This was accomplished using several approaches. Firstly, I_K1 was always studied in cells where it was the predominant current, as confirmed using the three-step inactivation protocol described above. In some cells (e.g. results shown in Figs 5 and 9), I_K2 contributed such a small component of current that it was ignored. In other cells the I_K2 component, was blocked using 5 mM 4-AP (e.g. results shown in Fig. 7 and Table 1). Finally, in the remaining experiments (results shown in Figs 6 and 8), I_K1 was defined as the difference between the currents recorded at +60 mV after the cell was subjected to conditioning potentials at -40 and -80 mV. Similarly, I_K2 was always investigated in cells in which it predominated. In cases where this predominance was not extreme, it was defined as the difference between the currents recorded at +60 mV after the cell was subjected to conditioning potentials at -40 and 0 mV (Figs 6 and 8).

I_K1 has not, to our knowledge, been previously described in the human myometrium. This current activated and inactivated more slowly than I_K2, had a negative apparent threshold of activation (between -60 and -50 mV) and inactivation (V_0.5 between -61 and -67 mV), and, unlike most cloned delayed rectifier-type K⁺ currents (Hart et al. 1993; Grissmer et al. 1994; Schmalz et al. 1998), was not blocked by 4-AP. It was, however, inhibited by TEA, with an EC₅₀ close to 3 mM.

Currents with similar characteristics to I_K1 have been described in colonic smooth muscle of guinea-pig (Vogalis et al. 1993) and dog (Thornbury et al. 1992; Carl, 1995). In their study, Vogalis et al. (1993) also used TEA (5 mM) to block the Ca²⁺-activated current, but used 2 s conditioning pre-pulses in availability experiments to derive a V_0.5 of -55 mV, compared with our more negative value using 10 s pre-pulses. Assuming the two currents are related, this discrepancy might be due to the very slow inactivation rate of this current such that a true steady state would not have been reached after 2 s, hence shifting the inactivation curve to the right. The 4-AP-insensitive current described in the canine colon (Thornbury et al. 1992; Carl, 1995) also had a V_0.5 of close to -65 mV and was TEA sensitive (EC₅₀ 7.7 mM) but inactivated much more rapidly than I_K1. In the present study the availability of I_K1 was not significantly influenced by partial block with either 1 or 10 mM TEA.

The ‘non-inactivating component’ of I_K1 (30-43 % of total) also mirrored that of the currents described by Vogalis et al. (1993) and Carl (1995), where the 4-AP-resistant currents inactivated by only 50 and 20 %, respectively. If all of these do in fact represent current through the channel responsible for I_K1, the differences in size of the non-inactivating component may result, at least in part, from the different conditioning pulse durations of 2, 20 and 10 s used by Volgalis et al. (1993), Carl (1995), and in the present study, respectively. The absence of a substantial outward current resistant to 1 mM TEA in cells held for several minutes at 0 mV supports the suggestion that the true non-inactivating component of I_K1 was very small or non-existent. If the ‘non-inactivating component’ of I_K1 actually represents a distinct current it would have to be another voltage-gated current and the only known candidate for this is the minK/K_VLQT1 channel which is expressed in rat uterus, activates and inactivates extremely slowly and is clofilium sensitive. However, it is also partially 4-AP sensitive and less TEA sensitive than I_K1 (Boyle et al. 1987; Attali et al. 1992).

The method of block of I_K1 (and I_K2) by the class III antiarrhythmic drug clofilium (EC₅₀ between 10 and 30 μM and enhancement of current decay) resembles that described for the cloned human delayed rectifier channels K_V1.3 and K_V1.5, with approximate EC₅₀ values of 50 and 60 μM, respectively (Attali et al. 1992; Malayev et al. 1994). I_K1 does not, however, resemble either K_V1.3 or K_V1.5 in any of its other properties. The only other channel cloned to date with similar electrophysiological and pharmacological properties to I_K1 is K_V2.1, which has a slower activation than the K_V1 family, inactivates slowly, is moderately TEA sensitive and in some cell types is resistant to block by 4-AP (Kirsch & Drew, 1993; Patel et al. 1997).

The other voltage-gated K⁺ current, I_K2, was 4-AP-sensitive (approximate EC₅₀ of 1 mM), displaying an inactivation V_0.5 of -30 mV, a threshold of activation between -30 and -20 mV and a rate of inactivation much faster than I_K1. The availability of this current was similar to the sustained component of the 4-AP-sensitive current described in human non-pregnant myometrium (Erulkar et al. 1993). Currents similar to I_K1 and I_K2 have recently been described in pregnant and non-pregnant rat myometrium (Wang et al. 1998), using an experimental approach similar to ours. These two currents (C₁ and C₂) demonstrated low and high thresholds of activation and inactivation similar to I_K1 and I_K2, respectively. Both C₁ and C₂ were moderately TEA sensitive but only C₂ was 4-AP sensitive.

Other K⁺ currents described in this study

I_K,A is a transient, rapidly activating and rapidly inactivating 4-AP-sensitive current with relatively negative thresholds of activation and inactivation. This current has been previously described in primary culture of pregnant rat myometrium and non-pregnant human myometrium (Erulkar et al. 1993, 1994) as well as in other rhythmic tissues, including guinea-pig colonic smooth muscle (Vogalis et al. 1993). There is some evidence in guinea-pig colonic smooth muscle to suggest that the A-type current may act to offset action potential generation by opening before or simultaneously with voltage-gated Ca²⁺ channels, implying a function for this current in the timing and rhythmicity of contractions (Vogalis et al. 1993). A detailed characterization of this current was not carried out in the present study.

I_K(Ca) activated only at potentials positive to 0 mV under the conditions used in this study but comprised a substantial component of the whole-cell current at more positive potentials. It was inhibited by TEA with an approximate EC₅₀ of 0.3 mM. I_K(Ca) has been characterized previously in the human myometrium (Anwer et al. 1993; Perez et al. 1993; Khan et al. 1993, 1997). Since the amplitude and activation threshold of this current is influenced by Ca²⁺ influx through Ca²⁺ channels, it is believed to play a role in the termination of action potentials in the contracting myometrium (Parkington & Coleman, 1988), and may contribute to setting the resting membrane potential (Anwer et al. 1993).

Possible physiological relevance of IK₁ and IK₂

The negative threshold of activation for I_K1 (apparent current threshold between -60 and -50 mV) and its very slow inactivation suggest that this current may contribute to the resting membrane potential in the late pregnant myometrium, which has been estimated in myometrial strips from non-labouring pregnant women at between -45 and -65 mV (Kumar & Barnes, 1961; Nakajima, 1971; Nakao et al. 1997) and in isolated cells at -49 mV (Pressman et al. 1988) and -50 mV (Inoue et al. 1990). Furthermore, I_K1 decayed slowly enough so that it may be of importance to the regulation of slow wave duration in the myometrium. As far as we are aware, currents with characteristics similar to I_K1 have not been reported in vascular smooth muscle, which, with the exception of certain veins, does not demonstrate slow waves synchronized with phasic contractions. A potential role for I_K1 or I_K2 in slow wave generation or resting membrane potential is of relevance to myometrial contractility. It has been shown that 4-AP depolarizes strips of pregnant rat myometrium (Wilde & Marshall, 1988). However, until a specific blocker of I_K1 (or any 4-AP-insensitive K_V current such as K_V2.1) is found, similar experiments cannot be performed to specifically determine the role of I_K1 in membrane potential control and contraction.

There is evidence from animal studies that the resting membrane potential changes during pregnancy. For example in the late pregnant rat it is around -55 mV but as negative as -68 mV in early pregnancy (Kuriyama & Suzuki, 1976). A shift in the relative expression of different delayed rectifier currents, possibly in response to changing oestrogen and progesterone levels (Toro et al. 1990; Erulkar et al. 1994), may account for these changes. The relative expression of the voltage-gated K⁺ currents described in this study varied from one cell to the next. Most cells demonstrated either I_K1 or I_K2 as the predominant voltage-gated current, although cells with either I_K1 or I_K2 were often obtained from the same myometrial specimen. By contrast, in the study by Wang et al. (1998) in rat myometrium, all cells from both pregnant and non-pregnant tissue appeared to possess both C₁ and C₂ components.

In summary, we have demonstrated in pregnant human myometrium an unusual 4-AP-resistant voltage-gated K⁺ current (I_K1) that has characteristics similar to those expressed in other rhythmic smooth muscles. Cloning of the channel and development of drugs which selectively modulate its activity may lead to a new approach to the selective control of myometrial activity.