Action of niflumic acid on evoked and spontaneous calcium‐activated chloride and potassium currents in smooth muscle cells from rabbit portal vein

1 The action of niflumic acid was studied on spontaneous and evoked calcium‐activated chloride (ICl(Ca)) and potassium (IK(Ca)) currents in rabbit isolated portal vein cells. 2 With the nystatin perforated patch technique in potassium‐containing solutions at a holding potential of – 77 mV (the potassium equilibrium potential), niflumic acid produced a concentration‐dependent inhibition of spontaneous transient inward current (STIC, calcium‐activated chloride current) amplitude. The concentration to reduce the STIC amplitude by 50% (IC50) was 3.6 × 10−6 m. 3 At – 77 mV holding potential, niflumic acid converted the STIC decay from a single exponential to 2 exponential components. In niflumic acid the fast component of decay was faster, and the slow component was slower than the control decay time constant. Increasing the concentration of niflumic acid enhanced the decay rate of the fast component and reduced the decay rate of the slow component. 4 The effect of niflumic acid on STIC amplitude was voltage‐dependent and at – 50 and + 50 mV the IC50 values were 2.3 × 10−6 m and 1.1 × 10−6 m respectively (cf. 3.6 × 10−6 m at −77 mV). 5 In K‐free solutions at potentials of – 50 mV and + 50 mV, niflumic acid did not induce a dual exponential STIC decay but just increased the decay time constant at both potentials in a concentration‐dependent manner. 6 Niflumic acid, in concentrations up to 5 × 10−5 m, had no effect on spontaneous calcium‐activated potassium currents. 7 Niflumic acid inhibited noradrenaline‐ and caffeine‐evoked ICl(Ca) with an IC50 of 6.6 × 10−6 m, i.e. was less potent against evoked currents compared to spontaneous currents. In contrast niflumic acid (2 × 10−6 m–5 × 10−5 m) increased noradrenaline‐ and caffeine‐induced IK(Ca). 8 The results are discussed with respect to the mechanism of block of ICl(Ca) by niflumic acid and its suitability as a pharmacological tool for assessing the role of ICl(Ca) in physiological mechanisms.


Introduction
Experiments with patch pipette techniques have revealed that noradrenaline acts on ac-adrenoceptors to stimulate simultaneously a calcium-activated chloride current (IC(Ca)), calciumactivated potassium current (IK(c.)) and a calcium-permeable cation current (Icat) in the rabbit portal vein (Byrne & Large, 1988;Wang & Large, 1991) and in the rabbit ear artery (Amrdee et al., 1990). Since it has been shown that an anion and a cation conductance increase is responsible for the noradrenaline-induced depolarization recorded with microelectrodes in isolated cells of the rabbit portal vein (Amedee & Large, 1989), it is possible that ICk,,Ca) and It,, may have important roles in producing depolarization and contraction in vascular smooth muscle. Also it is relevant that Icj(ca) has now been observed in several types of smooth muscle and can be activated by various pharmacological agents (see Introduction of Hogg et al., 1994). Therefore it would be interesting to assess the contribution of IQ(ca) and Icat to agonist-induced depolarization in smooth muscle. In order to do this it would be necessary to have selective blocking drugs to dissect out the roles of these conductance mechanisms.
Recently we have embarked on a series of experiments to investigate the characteristics of established chloride channel antagonists in blocking Icl(cI) in vascular smooth muscle cells.
It has been demonstrated that 4-acetamido-4'-isothiocyanatostilbene-2,2'-disulphonic acid (SITS), 4,4'-diisothiocyanatostilbene-2,2'-disulphonic acid (DIDS) and anthracene-9- ' Author for correspondence. carboxylic acid (A-9-C) inhibit evoked Icc) in rat portal vein (Baron et al., 1991). Also it was shown in rabbit portal vein that these compounds were more potent against spontaneous transient inward currents (STICs, chloride currents activated by spontaneous release of calcium from caffeinesensitive intracellular stores) than against Ic~ca) elicited by noradrenaline and caffeine (Hogg et al., 1994). It was found that all 3 compounds had low potency as the concentration to inhibit IcQ(c,,) by 50% was greater than 10-4 M i.e. the potency against Ia(ca) is less than the action of DIDS and SITS against their well-established effects on Cl--HCO3exchange in smooth muscle (e.g. see Aickin & Brading, 1983). Consequently it seems unlikely that these channel blockers would be of use as pharmacological tools for assessing the physiological role of Ia(c.). A more promising candidate might be niflumic acid, a non-steroidal anti-inflammatory agent, which at a concentration of 10 M produced marked attenuation of AC,(ca) in rat portal vein (Pacaud et al., 1989).
The relatively high potency of niflumic acid has been confirmed in canine and guinea-pig tracheal cells (Janssen & Sims, 1992) and in rabbit oesophageal smooth muscle (Akbarali & Giles, 1993).
In the present work we have undertaken a quantitative study of the action of niflumic acid against spontaneous and evoked Ic(ip) in the rabbit portal vein. Evidence will be presented to substantiate the relatively high potency of niflumic acid and experiments will be described which indicate that niflumic acid inhibits I'(ca), at least partly, by an open channel blocking mechanism. '." Macmillan Press Ltd, 1994 Methods Experiments were carried out on freshly dispersed smooth muscle cells from the rabbit portal vein. Rabbits (2-2.5 kg) of either sex were killed by an overdose of i.v. sodium pentobarbitone and single cells were obtained by enzymatic dissociation with collagenase and papain as described previously (Hogg et al., 1993). The cells were stored in a physiological salt solution with 0.75mM Ca2" at 4C and were used on the same day. Whole-cell currents were measured with the perforated patch method with a patch clamp amplifier (List EPC7; List-Electronic; Darmstadt, Germany) at room temperature. In order to obtain a perforated patch nystatin (75-200ligml-') was contained in the patch pipette solution. The external salt solution contained (mM): NaCl 126, KC1 6, MgCl2 1.2, CaC12 1.5, HEPES 10 and glucose 11 and the pH was adjusted to 7.2 with NaOH. In potassium-free conditions, 6mM KCI was omitted and in some experiments 5 mM tetraethylammonium chloride (TEA-Cl) was added to the bathing solution. The pipette solution contained (mM): KCI 126, MgCl2 1.2, HEPES 10, glucose 11 and EGTA 0.2. In potassium-free conditions 126 mM KCl was replaced by an equimolar amount of NaCl. In some experiments the effect of niflumic acid on voltage-gated divalent cation currents was studied. For these studies 10 mM BaCl2 was added to the bathing solution instead of calcium so that Ba2' was the charge carrier. Also in these latter studies the pipette solution contained 126 mM CsCI and 10mM EGTA. In experiments where noradrenaline was used, 10-6 M propranolol was included in the bathing solution to remove any P-adrenoceptor-mediated response. The recordings were made in a static bathing solution and the external solution was changed with a push-pull arrangement with syringes. Some records were illustrated by playback from videotape onto a Gould brush recorder. Analysis of the time course of spontaneous transient currents was carried out on a personal computer. Signals were filtered at 400 Hz prior to digitisation and currents were sampled at 800 Hz using a CED 1401 laboratory interface and captured on the hard disk of the computer. Capture and averaging of signals were performed with SIGAVG signal-averaging programme and curve fitting was done with a voltage clamp analysis programme (both Cambridge Electronic Design, Science Park, Cambridge). Exponential fits were obtained using a least squares fitting routine and the fitting procedure was weighted towards large current values. From each cell 10-20 spontaneous chloride or potassium currents were averaged to obtain amplitude and time course values. The Langmuir adsorbtion isotherm curve in Figure lb was drawn using an IC5o value (concentration of niflumic acid to inhibit STIC amplitude to 50% of the control value) obtained by linear regression of the experimental data points. In the text, n values refer to the number of cells used to obtain the mean determinations. The values given in the text are means ± s.e.mean and statistical significance was estimated with either Student's t test or paired t test. Chemicals used were bovine serum albumin (fatty acid free), caffeine, dithiothreitol, noradrenaline bitartrate, nystatin, papain (type IV), TEA-Cl (all Sigma, Poole, Dorset); niflumic acid (Aldrich, Gillingham, Dorset); collagenase (CLS2 247 umg', Worthington, Reading, Berkshire).

Effect of niflumic acid on STICs
In the first series of experiments we investigated the action of niflumic acid on STICs in potassium-containing solutions. Figure la shows records from a cell which was held under voltage clamp at a command potential of -77 mV (the potassium equilibrium potential, EK) and niflumic acid was added to the bathing solution in a cumulative manner. It can be seen quite clearly that there was progressive diminution of STIC amplitude and in the presence of 10-4 M niflumic acid the STICs were blocked. The onset of inhibition was rapid and was apparent within about 10 s of adding niflumic acid to the external solution. Also the antagonism was readily reversible and usually complete washout occurred within 1-3 min after removing niflumic acid from the cells ( Figure  la). The concentration-effect of niflumic acid on STIC amplitude is illustrated graphically in Figure lb. The continuous curve is drawn according to the Langmuir isotherm and the estimated concentration to inhibit STIC amplitude to 50% of the control value (ICw) at -77 mV with potassium containing solutions was 3.6 x 10-6M. The potency of niflumic acid against STIC amplitude at any given membrane potential did not differ in potassium-free conditions which were used in some experiments to investigate the voltage-dependence of the action of niflumic acid.
It was apparent in the majority of cells at -77 mV that the reduction in STIC amplitude by niflumic acid was accompanied by a marked alteration of the STIC time course. Figure 2a illustrates averaged STICs in the absence and presence of various concentrations of niflumic acid taken from the same cell as shown in Figure 1. In control conditions the STIC decay time course can be described by a single exponential with a time constant (Tco) of 77 ms (control curve in Figure 2b). In the presence of niflumic acid the STIC decay time course became more complex and appeared to consist of two distinct phases ( Figure 2a). The semilogarithmic plots show that the initial decay was faster than the control while the slow component was slower than the control decay ( Figure 2b). In Figure  STOCs are insensitive to niflumic acid (see Figure 6).
Evidence has been put forward to suggest that the STIC decay represents closure of the calcium-activated chloride channels (Hogg et al., 1993) and the alteration of the STIC decay by niflumic acid could therefore be explained by blockade of open ion channels. The usual scheme (see Colquoun & Sheridan, 1981) to describe open channel block is: Assuming the normal opening rate (P') is faster than the closure rate (a) of calcium-activated chloride channels then STIC decay in control conditions is determined by a( = /T. see Hogg et al., 1993). In the presence of an open channel blocking agent (with association and dissociation rate constants of respectively k+B and k-B) it is expected that the STIC decay should consist of two components as was seen experimentally. Moreover it is predicted with certain assumptions (see Colquhoun & Sheridan, 1981 b then it is expected that a plot of (l/Tf + 1/T-I/TO) against niflumic acid concentration would be linear. The results from the cell in Figure 2 are plotted according to equation (2) in Figure 3. It can be seen that a linear relationship fits the data points well and the intercept (kB) is 25 sand the slope (k+B) was 3.8 x i05 M-1 s'l. In 5 out of 6 cells in potassiumcontaining solutions at -77 mV, niflumic acid produced similar results to those illustrated in Figures 2 and 3 and the mean k+B was 5.8 ± 1.1 x 105 M1 s ' and k.B was 26 + 5.1 s'l. In the sixth cell the fast component was not present and niflumic acid appeared only to increase the T value (i.e. corresponding to T). It seems unlikely that the fast component was not seen in this cell because of the voltagedependent action of niflumic acid and the fast component was not observed at more depolarized potentials (e.g. see Figure 5 and see discussion). Nevertheless, overall the data are consistent with open channel block by niflumic acid.
Voltage-dependent effect of niflumic acid Previously we had demonstrated that A-9-C reduced STICs in a voltage-dependent manner whereas the inhibitory action of DIDS and SITS was not influenced by membrane potential (Hogg et al., 1994). We have carried out similar experiments with niflumic acid in potassium-free bathing and pipette solutions to eliminate STOCs which become prominent at potentials positive to -50 mV. Figure 4a shows the effect of 10-6 M niflumic acid at -50 mV and + 50 mV in the same cell. At -50 mV this concentration of the blocker reduced the STIC amplitude by about 10% whereas at + 50 mV the attenuation was more marked (about 30%). The concentration-effect curves for several cells are shown in Figure 4b at -50 and + 50 mV and the curve at -77 mV from Figure lb is also added for comparison. The calculated ICm values at -50 and + 50 mV were respectively 2.3 x 10-6 M and 1.1 X 10-6 M (3.6 x 10-6 M at -77 mV). Therefore it is concluded that the potency of niflumic acid was increased by depolarization by about two fold between -50 and + 5O mV. We also investigated the effect of niflumic acid on the STIC T at various membrane potentials and these experiments were also carried out in potassium-free conditions to remove any interference from STOCs. Figure 5a shows averaged STICs in the absence (control) and in the presence of 5 x 10-6 M niflumic acid at holding potentials of -50 mV and + 50 mV. It was apparent that the fast component of decay was no longer observed in the presence of niflumic acid at these potentials (to be discussed later) and that only the slow component was present (cf. Figure 2a). At -50 and + 50 mV, niflumic acid greatly prolonged the STIC decay in addition to reducing the STIC amplitude (i.e. qualitatively similar to the effect of niflumic acid on the slow component of decay at -77 mV, Figure 2). The decays of the averaged currents in Figure

Effect of niflumic acid on evoked calcium-activated chloride and potassium currents
It has been demonstrated previously that DIDS, SITS and A-9-C were less potent in inhibiting evoked chloride currents compared to STICs (Hogg et al., 1994). Consequently we investigated the effect of niflumic acid on noradrenalineand caffeine-evoked IcQ(ca) in potassium-free conditions at a holding potential of -50 mV. Figure 7 shows the effect of two concentrations of niflumic acid on noradrenaline-induced IC1(C.) (Figure 7a) and caffeine-evoked I'(ca) (Figure 7b) and it can be seen that niflumic acid produces a concentration-A A NA 5 min wash 5 min wash a A NA acid the STICs decayed exponentially and the control time constants at -50 and + 50 mV were respectively 96 ms and 213 ms. In the presence of 5 x 10-6 M niflumic acid the ? values were 263 ms (-50 mV) and 485 ms (+ 50 mV). The effect of two concentrations of niflumic acid at -50 mV and + 50 mV are shown in Table 1 and there are several conclusions. First, the increase in the STIC T value was concentration-dependent as 5 x 10-6 M niflumic acid produced a greater effect on T than 10-6 M niflumic acid. Secondly, although the absolute T values are larger at + 50 mV than -50 mV (see Hogg et al., 1993), the ratios of the T values (drug:control) are no greater at + 50 mV than at -50. Therefore the increase in T does not appear to be voltage-dependent.

Effect of niflumic acid on STOCs
In order to ensure that the effect of niflumic acid was not mediated by an action on the intracellular calcium store which is the primary source of calcium for triggering STICs we investigated the action of niflumic acid on spontaneous transient outward currents (STOCs). These are spontaneous calcium-activated potassium currents which are triggered by the same calcium store responsible for STICs (Wang et al., 1992). Figure 6a shows a continuous record of STOCs in a cell held at 0 mV (i.e. close to Ecr) before and after the addition of 5 x 10-M niflumic acid to the bathing solution. From this trace and averaged STOCs illustrated in Figure 6b it is apparent that niflumic acid did not affect STOCs. With 5 x 10-6 M, 10-5 M and 5 x 10-5 M, (n = 3 at each concentration) niflumic acid did not alter the amplitude, time to peak, half-decay time and frequency of STOCs. It can be concluded that niflumic acid does not modify the intracellular  dependent inhibition of I'cA) which is rapidly reversible. It is interesting that in Figure 7b niflumic acid appears to prolong the duration of the induced current which persisted after washout of the drug. There is no obvious explanation for this observation. Niflumic acid produced similar effects on both noradrenalineand caffeine-evoked currents and consequently the data were pooled. Figure 8 illustrates the concentration-effect relationship of niflumic acid on evoked Aa(c) at -50 mV obtained from several cells and the'estimated ICo was 6.6 x 10-6 M. Also included on this graph are the data on STICs at -50 mV (appropriate curve from Figure 4b) and it can be seen quite clearly that the potency of niflumic acid against STICs (IC5 = 2.3 x 10-6 M) is greater than against evoked chloride currents.
We also studied the effect of niflumic acid on IK(Ca) stimulated by noradrenaline in potassium-containing solutions at OmV (which is close to the chloride equilibrium potential, -2 mV). A typical experiment is shown in Figure  7c where it can be seen that 5 x 10-5 M niflumic acid potentiates the noradrenaline-evoked IK(ca) and this enhancement is sustained in the continued presence of niflumic acid. The potentiating effect of niflumic acid was concentration-dependent as in the presence of 2 x 10-6 M and 5 x 1i-0M niflumic acid the noradrenaline-evoked IK(c.) was increased respectively 1.7 ± 0.4 fold (n = 6) and 2.5 ± 0.3 (n = 8) fold. Therefore niflumic acid increases the amplitude of evoked IK(Ca) without affecting spontaneous calcium-activated potassium currents.
Effect of niflumic acid on voltage-activated divalent cation currents The relative high potency (see Discussion) of niflumic acid against Ia(Ca) suggests that this agent might be used to appraise the role of calcium-activated chloride currents in producing membrane depolarization and contraction evoked by noradrenaline (or other excitants) in smooth muscle. Consequently it seemed worthwhile to see if niflumic acid inhibited voltage-gated calcium currents which might be expected to be the essential link between depolarization which is generated by Icj(ca) and muscle contraction. In these experiments we used pipettes filled with 126 mM CsCl and 10 mM EGTA (see Methods) and the external solution contained 10 mM BaCl2 (no added calcium) so that Ba2+ was the charge carrier. amplitude of inward currents evoked by command steps duration of (1500 ms) to 0 mV was not altered by 2 x 10-6M -5 x 10-5 M niflumic acid. For example, in one series of experiments the amplitude of the control barium current was 58 ± 12 pA and in the presence of 2 X 10-6 M niflumic acid the current 59 ± 9 pA (n = 4). In another series of experiments in 5 x 10-5 M niflumic acid the inward current was 79 13 pA compared to a control value of 73 ± 12 pA (n= 3). Therefore it can be concluded that niflumic acid in concentrations up to 5 x 10-5 M does not inhibit the influx of divalent cations through voltage-gated calcium channels.

Discussion
This paper demonstrates that niflumic acid is a relatively potent blocker of calcium-activated chloride currents in rabbit portal vein and inhibits Ia(C^) in the micromolar range. It will be suggested later that niflumic acid inhibits Ia(c.) by blocking open chloride channels and therefore comparison of quantitative data in different preparations might give an indication whether the calcium-activated chloride conductance differs from tissue to tissue. Dose-response curves with niflumic acid have not been constructed in other smooth muscle tissues but it has been reported that in the presence of 10-4 M niflumic acid a tail current, presumed to be chloride current activated by the influx of calcium, was greatly reduced but a small component still persisted in the presence of 10-4 M niflumic acid in rabbit oesophageal smooth muscle (Akbarali & Giles, 1993) whereas this concentration of niflumic acid blocked STICs in the present work. This suggests that there might be a difference in Ia(ca) in portal vein and oesophagus or that another conductance might contribute to the tail current in rabbit oesophageal. smooth muscle cells. The ICo of niflumic acid against Ic(Ca) in Xenopus oocytes was estimated to be 17 #iM (White & Aylwin, 1990) which is almost an order less potent than against STICs at -50 mV (IC5o of 2.2 x 10-6 M) in the present study. This difference is sufficiently large to suggest that there might be a difference between ICc,1) in rabbit portal vein and oocytes. It should be noted that niflumic acid is not specific for IC(Ca) because this, agent inhibits cyclic AMP-activated chloride currents in amphibian retinal pigment epithelial cells with an IC5o of about 2.5 x 10-1 M (Hughes & Segawa, 1993). Also niflumic acid is a potent inhibitor (ICm of 6.3 x 10-7 M) of anion exchange in human red cells where it has been suggested that niflumic acid interacts with the band 3 protein (Cousin & Motais, 1979). It is interesting that many compounds that block chloride channels also inhibit anion transport in red blood cells (also see Cousin & Motais, 1982a,b). A striking result was that niflumic acid inhibited STICs more potently than evoked currents. Since niflumic acid increased the noradrenaline-evoked IK(Ca) without altering STOCs (spontaneous calcium-activated potassium currents) it is possible that this compound increases the evoked release of calcium from the intracellular store without altering the spontaneous release of calcium. It was observed previously that DIDS, SITS and A-9-C increased the evoked IvCac) (Hogg et, al., 1994) and it may be a general property of chloride channel blocking agents to enhance the amount of calcium released from the intracellular store.
There is some evidence to suggest that niflumic acid inhibits Ica(c.) by blocking open chloride channels. First, the action of niflumic acid is voltage-dependent which suggests the blocking site is within the membrane electrical field. The rapid onset and reversibility indicates the site with which niflumic acid binds is readily accessible from the external solution and therefore taking these two points together it is tempting to speculate that the ion channel is the site of action. Secondly, at -77 mV in the presence of niflumic acid the STIC decay consisted of two exponentials rather than one and the time constants of both exponentials were depen-dent on concentration of niflumic acid in a manner consiste with open channel block according to scheme (1). Moreover the data represented in Figure 3 give quantitative support for an open channel blocking mechanism over a large concentration range. However, it is necessary to explain why in the presence of niflumic acid the STIC decay was described by two exponentials at -77 mV but only a single component was observed at more positive potentials. It should be emphasized that the fast component at -77 mV was not a STOC (see Results for arguments) but was a chloride current. The STIC decay at -50 mV and + 50 mV (e.g. see Figure 5a) resemble in some respects the slow component recorded at -77 mV (e.g. Figure 2a). At -50 and + 50 mV in the presence of niflumic acid the STIC decay was exponential and had a similar absolute value, for any given concentration of niflumic acid, to ;s at -77 mV. Therefore it can be concluded that the fast component is not observed at potentials more positive than -77 mV. A common observation is that STOCs and STICs often occur as biphasic events which suggests that the same calcium pulse activates both potassium and chloride currents (see Hogg et al., 1993). Therefore since the STOC time course is not altered by niflumic acid it is unlikely that the slow decay of STICs in the presence of the blocking agents can be attributed to a prolonged time course of the calcium signal that triggers the STIC. From scheme (1) two components of STIC decay would be recorded only if chloride channel opening was rapid compared to the blocking rate by niflumic acid. We have no information on the rate of opening of calcium-activated chloride channels in smooth muscle but the STIC rise time is 40-60 ms which is likely to be slow relative to chloride channel opening. There are no data on the factors that determine the STIC rise time although the rate of rise of calcium at the internal surface of the cell membrane which contains calcium-activated chloride channels remains a distinct possibility. Nevertheless since the STIC rise time is relatively slow it is probable that significant block of the chloride channels will occur during the STIC rise time. If the rate of association (k+B) of niflumic acid with the chloride channel was rapid compared to the STIC rise time it might not be expected to record all (or any) of the rapid phase of block and only the component associated with unblocking of the channel might be observed. In the context of the present work it might be necessary to postulate that the fast component of block is resolved at -77 mV but not at -50 mV or + 50 mV because the rate of association of niflumic acid with the channel is slow compared to the STIC rise time at -77 mV but not at more positive potentials. For this to occur it would be necessary for k+B to be voltagedependent and at more depolarized potentials block of channels might occur during the STIC rise time and only the unblocking slow component is observed. The experimental data confirm that membrane depolarization increases the potency of niflumic acid which indicates that as the inside of membrane is made more positive the rate of association of the negatively changed molecules of niflumic acid with binding sites inside the open chloride channel is increased. Previously it has been demonstrated at the frog motor endplate in the presence of lignocaine derivatives, which block channels opened by acetylcholine, the decay of the endplate current (e.p.c.) is prolonged but remains mono-exponential at -30 mV whereas at more negative potentials the e.p.c. decay is obviously bi-exponential (Beam, 1976). These data are qualitatively similar to the results of the present experiments.
In conclusion, it seems likely that at least part of the blocking action of niflumic acid on IC(Ca,) can be attributed to block of open chloride channels. The equilibrium constant of niflumic acid for the open chloride channel (KB) can be estimated from keB/k+B which is calculated to be 4.5 x 10-5 M at -77 mV. This is about an order greater than the IC5o (3.6 x 10-6 M) calculated from inhibition of STIC amplitude at -77 mV which suggests that niflumic acid might have an additional action (other than blocking open channels) to inhibit Ic(ca). Alternatively the value of KB might be an overestimate resulting from the imperfections of using STICs as a model as outlined in the previous paragraph. Finally, it is worth commenting on the potential usefulness of niflumic acid as a pharmacological tool. It is evident that niflumic acid does not inhibit either the a,-adrenoceptor recognition site (as the noradrenaline-induced IK(ca) was not blocked) or the voltage-dependent calcium channel. Also in two cells, noradrenaline-evoked I,,, (unpublished observation) was not reduced and thus it seems that niflumic acid might be useful in defining the role of Icj(ca) in producing noradrenaline-evoked depolarization and contraction. However it seems that niflumic acid appears to increase the amount of calcium released from the intracellular store in response to stimulation with noradrenaline. If this increased release of calcium can activate contractile proteins niflumic acid will tend to increase contractility of smooth muscle to excitants such as noradrenaline in addition to blocking calcium-activated chloride channels which are also stimulated by noradrenaline. Therefore the effect of niflumic acid to inhibit ICkCa) in smooth muscle and subsequent contraction might be overcome, at least partly, by the ability to increase the release of calcium from the sarcoplasmic reticulum and hence the contractility of vascular smooth muscle. In these circumstances it might be difficult to assess the role of ICpa,) in smooth muscle contraction with niflumic acid. This work was supported by The Wellcome Trust.