- Top of page
The effects of the Cl− channel antagonists, niflumic acid (NFA), dichloro-diphenylamine 2-carboxylic acid (DCDPC) and diisothiocyanato-stilbene-2,2′-disulphonic acid (DIDS) on Ca2+-activated Cl− current (ICl(Ca)) evoked by adding fixed intracellular calcium concentrations ([Ca2+]i) to the pipette solution were studied in rabbit pulmonary artery myocytes. With 250 and 500 nm[Ca2+]i bath application of NFA (100 μm) increased inward current at negative potentials, but inhibited outward current at positive potentials. On wash out of NFA, ICl(Ca) was greatly enhanced at all potentials. When external Na+ ions were replaced by N-methyl-d-glucamine (NMDG+) NFA still enhanced ICl(Ca) at negative potentials but the increase of ICl(Ca) on wash out was blocked. When the mean reversal potential (Er) of ICl(Ca) was shifted to negative potentials by replacing external Cl− with SCN−, NFA increased inward current but blocked outward current suggesting that the effect of NFA is dependent on current flow. Inclusion of NFA in the pipette solution had no effect on ICl(Ca). Voltage jump experiments indicated that ICl(Ca) displayed characteristic outward current relaxations at +70 mV and inward current relaxations at −80 mV that were abolished by NFA. DCDPC (100 μm) produced similar effects to NFA but 1 mm DIDS produced inhibition of ICl(Ca) at both positive and negative potentials and there was no increase in current on wash out of DIDS. These results suggest that NFA and DCDPC, but not DIDS, simultaneously enhance and block ICl(Ca) by binding to an external site, probably close to the mouth of the chloride channel.
Calcium-activated chloride currents (ICl(Ca)) have been recorded in many smooth muscle types and have many distinctive properties including voltage-dependent kinetics, modulation by external anions and inhibition by a range of chemically disparate agents (see reviews by Large & Wang, 1996 and Greenwood & Large, 1999a). The majority of experiments studying the properties of ICl(Ca) in vascular smooth muscle cells have utilised the perforated patch configuration of the whole-cell recording technique where the resting intracellular Ca2+ concentration ([Ca2+]i) is normally below threshold for activation of ICl(Ca). The channel is then stimulated by an increase in [Ca2+]i elicited by several methods. ICl(Ca) can be activated by agonist-dependent or spontaneous intracellular Ca2+ release (the latter have been termed spontaneous transient inward currents or STICs, Wang et al. 1992) or as a consequence of Ca2+ influx through voltage-gated Ca2+ channels in response to depolarising pulses (Lamb et al. 1994; Greenwood & Large, 1996; Yuan, 1997).
Using these techniques we have studied the mechanism of action of several chloride channel antagonists (e.g. see review by Large & Wang, 1996). The most potent blocker of ICl(Ca) studied so far in smooth muscle is niflumic acid (NFA) which inhibits ICl(Ca) in micromolar concentrations (Pacaud et al. 1989; Hogg, et al. 1994b) and produces distinctive effects on the decay time course of STICs. Previously it had been proposed that the mono-exponential decay of STICs represents closure of the chloride channels (Hogg et al. 1993) but in the presence of NFA the STIC decay time course can be described by two exponentials (Hogg et al. 1994b). The data could be interpreted by a scheme in which NFA acts as an open channel blocking agent which remains in the channel for a short time relative to the mean open duration of the channel, i.e. NFA is a rapidly dissociating open channel blocker (Hogg et al. 1994b). Due to the potency of NFA and its action on the channel molecule NFA is frequently used to characterise ICl(Ca) in electrophysiological experiments and also to assess the functional role of ICl(Ca) in intact tissue preparations (e.g. see Greenwood & Large, 1999a).
In studies on ICl(Ca) using the above techniques the activating Ca2+ pulse is free to change. However, ICl(Ca) can also be activated persistently by a clamped [Ca2+]i imposed by dialysis from the patch pipette. This approach has been used to study ICl(Ca) in other cell types (e.g. rat parotid acinar cells, Arreola et al. 1996; cultured calf pulmonary artery endothelial cells, Nilius et al. 1997; Xenopus oocytes, Kuruma & Hartzell, 1999) and more recently in vascular smooth muscle cells (Greenwood et al. 2001). In the present study we used [Ca2+]i of 0.25–1 μm to activate ICl(Ca) as this is the concentration range which is likely to be achieved physiologically during stimulation of vascular smooth muscle. During the course of these experiments we observed a surprising and unexpected result in that in some experimental conditions NFA did not inhibit ICl(Ca) activated by a maintained concentration of Ca2+ but actually increased ICl(Ca). This is a novel mechanism of drug action as well as an interesting characteristic of the calcium-activated chloride conductance. Therefore we explored this effect in detail and we propose that NFA can both increase and decrease ICl(Ca) by binding to an extracellular site, probably close to the mouth of the chloride channel.
- Top of page
The main finding of the present work is that in rabbit pulmonary artery smooth muscle cells the chloride channel antagonist NFA has a dual effect on ICl(Ca) when activated by 250 and 500 nm[Ca2+]i imposed by dialysis from the pipette solution. At negative potentials NFA increased the amplitude of ICl(Ca) but at positive potentials NFA inhibited the current, i.e. NFA has a dual effect on ICl(Ca). This is a surprising result because in all previous reports it has been shown that NFA behaves only as a blocker of ICl(Ca) not only in smooth muscle but also in other cell types (e.g. see Large & Wang, 1996). However, using similar techniques of activating ICl(Ca) NFA increases ICl(Ca) in rabbit portal vein myocytes (I. A. Greenwood & N. Leblanc, unpublished results). The most likely explanation for the discrepancy in the results with NFA is a difference in the experimental conditions used. In previous studies where only inhibitory effects of NFA have been recorded in single cells NFA has been added to the cells when [Ca2+]i was at resting levels, probably around 50–100 nm, which is below the threshold for activation of the chloride channels (see Large & Wang, 1996). In our experiments with 50 nm[Ca2+]i there was no activation of ICl(Ca) and NFA did not activate ICl(Ca) at −50 mV in these conditions. When STICs are recorded there are sporadic transient increases of Ca2+ at the sarcolemma but for most of the time [Ca2+]i is at resting levels. In the present work NFA is added to cells in which ICl(Ca) has been persistently activated by a relatively high [Ca2+]i (e.g. 500 nm) imposed by the patch pipette solution. Therefore, it seems that the effect of NFA is altered when the chloride channels are tonically activated by constantly elevated [Ca2+] ions in single cells.
The increase in current observed at negative potentials produced by external application of NFA appeared to be due to an increase solely in Cl− conductance since the effect of NFA was similar when the impermeant cation NMDG+ replaced Na+ ions in the pipette and bathing solutions. Since NFA produced no increase in membrane conductance when [Ca2+]i was 50 nm this is good evidence that the increased current observed at −50 mV is ICl(Ca). It is worth noting that with inside-out patches there is only one conductance state activated by 500 nm Ca2+ (A. S. Piper & W. A. Large, unpublished results) suggesting that there is only one class of Ca2+-activated chloride channel in rabbit pulmonary artery myocytes. Therefore it can be concluded that in these conditions NFA can both increase and decrease ICl(Ca) depending on the membrane potential.
Voltage-dependence of the effect of NFA
The effect of NFA was potential sensitive and depended on the direction of the net current flow. It was found that NFA increased ICl(Ca) at potentials negative to Er but decreased the current at potentials positive to Er even when Er was changed by about −50 mV by substituting external NaCl with NaSCN. Therefore, when the net flux of anions was outward, NFA produced an overall increase of the conductance, but when there was net inward movement of anions, NFA produced an overall decrease of ICl(Ca) irrespective of Er.
In voltage-jump experiments on stepping the potential from negative to positive potentials in control conditions the instantaneous current was followed by an outward relaxation reflecting an increase in conductance. It has been suggested that this increase in conductance is due to a smaller channel closing rate (Hogg et al. 1993) and an increased affinity of Ca2+ ions for the intracellular binding site on the channel (Arreola et al. 1996) at positive, compared to, negative membrane potentials. On returning to negative potentials there was a much larger instantaneous current, representing the increased conductance caused by the depolarising step to positive potentials, followed by an outward relaxation which represents closing of the channels opened by the depolarising step. In the presence of NFA the relaxations were abolished indicating that NFA had altered these voltage-dependent properties of the conductance. A salient observation was that on stepping from negative potentials, where ICl(Ca) was enhanced by NFA, to positive potentials, where ICl(Ca) was decreased, there was no discernible inward relaxation corresponding to the transition from the increased to the decreased state of the conductance. Equally on returning to negative potentials it was not possible to resolve an inward relaxation representing the transition from the decreased to the increased state of the conductance. It is possible that the rate of these transitions were faster than the capabilities of our recording techniques. However we favour the possibility that NFA simultaneously increases ICl(Ca) and blocks the conductance and the net effect observed depends on the direction of ion flux.
Large increase of ICl(Ca) on wash out of NFA
Another notable feature of the effect of NFA in the present study was that on wash out of NFA there was a large increase of ICl(Ca). Measurement of Er in the presence of external NaSCN confirmed that the increase in current at −50 mV after wash out of NFA was indeed ICl(Ca) and moreover the conductance was increased at positive, as well as, negative membrane potentials. In the voltage-jump experiments the large washout current was associated with marked increases in the instantaneous currents to voltage steps but the amplitude of the relaxations at positive and negative membrane potentials were greatly reduced. Similar effects were observed with the chemically related DCDPC. These observations are consistent with the idea that the NFA-induced wash out current is due to a large increase in the probability of chloride channel opening and suggest that reversal of the inhibitory effect of NFA occurs more rapidly than the reversal of the potentiating effects of this agent.
It seems possible that the enhancement of ICl(Ca) on wash out of NFA is linked to the increase of the current produced in the presence of NFA. With the structurally related compound DCDPC which increased ICl(Ca) at −50 mV there was a large current on wash out of the agent. In contrast with DIDS, which simply inhibited ICl(Ca) at all potentials in its presence, there was no increase in ICl(Ca) when the agent was removed from the bathing solution. An argument against the above hypothesis is that when NMDG+ was used in the pipette and bathing solutions NFA increased ICl(Ca) at −50 mV but there was no increase in conductance on removal of NFA. This would imply that the NFA induced, and ‘wash out’ increase, in current amplitude are not related or that they are related but external Na+ ions are necessary for the development of the washout current. This interpretation is supported by the observation that in the experiments with NMDG+ a washout current was observed when NMDG+ was replaced by Na+ on wash out of NFA.
It is interesting that replacement of external Na+ ions with NMDG+ significantly reduced ICl(Ca) which was also observed by Arreola et al. (1996) in rat parotid acinar cells which suggests that external (and possibly internal) cations modulate ICl(Ca). After achieving the whole-cell conformation ICl(Ca) was approximately −160 pA but then declined to approximately one third of this value after a few minutes. Therefore many channels are in the inactivated state and it is possible that the equilibrium between open and inactivated states is sensitive to external cations. Some of the inactivation of ICl(Ca) appears to be due to the activation of CaMKII as KN93, a CaMKII inhibitor, was able to increase ICl(Ca) activated by 500 nm[Ca2+]i both in the present and in previous studies (Greenwood et al. 2001). However it is apparent that the majority of inactivation of ICl(Ca) recorded in the present study is not mediated by CaMKII. It is possible that another Ca2+-dependent inactivation process is occurring, or alternatively, that inactivation is an inherent property of the channel, as is the case with other ion channels e.g. voltage-dependent Ca2+ channels (Stotz & Zamponi, 2001) or voltage-gated potassium channels (Martens et al. 1999).
Mechanism of action of NFA
Previously it has been proposed that NFA is a channel blocker due to its voltage dependence and effect on the decay of STICs (see Hogg et al. 1994b for details). This implies that the binding site for the blocking effect of NFA is in the conducting pathway. However, there have been no previous reports on its ability to increase ICl(Ca) and it is evident that the potentiating effect of NFA is revealed with a sustained level of [Ca2+]i. We have not found any precedent in the literature for a channel blocker increasing a conductance but nevertheless it is worth considering putative mechanisms. It is possible that NFA and DCDPC might activate ICl(Ca) in addition to blocking the conductance. It has been shown that NFA and DCDPC evoke tetraethylammonium-sensitive Ca2+-activated potassium current (IK(Ca)) in rabbit portal vein myocytes (Greenwood & Large, 1995, 1998). This effect was not due to an action on internal Ca2+ stores and we proposed that this was an action on the channel protein (Greenwood & Large, 1995). This property is shared by other Cl− channel antagonists (Toma et al. 1996). At present we do not understand the molecular mechanism for this action but it is possible that these compounds might directly open the K+ channels. It is possible that NFA may also have a similar effect on the Ca2+-activated Cl− conductance and increase the probability of chloride channel opening by a mechanism similar to that with IK(Ca). Thus the observed data result from a net effect of NFA increasing ICl(Ca) as well as simultaneously inhibiting the conductance by a channel blocking mechanism. This model explains why at negative potentials NFA increases ICl(Ca) when low [Ca2+]i was used to activate ICl(Ca) but there was a trend towards inhibition when relatively high (1 μm) [Ca2+]i was used to evoke the current. In the former conditions the probability of channel opening will be considerably smaller and therefore the facilitatory effect of NFA on ICl(Ca) predominates. With 1 μm[Ca2+]i the probability of channel opening is much higher and therefore the blocking effect of NFA is predominant. The large increase in the amplitude of ICl(Ca) on wash out of NFA may be explained by a faster reversal of the inhibitory, compared to the facilitatory effect. This might occur, for example, if the facilitatory and blocking effects of NFA were mediated by an action on two distinct binding sites. However the present data do not show whether the dual effects of NFA are produced via one or two sites and further work is needed to solve this problem. Nevertheless it is clear that NFA must be applied extracellularly to produce both inhibitory and facilitatory effects and it is probably that the binding site(s) is close to the external mouth of the conducting pore. This is in contrast to Xenopus oocytes where NFA produced a significant block of ICl(Ca) when applied to the internal surface of inside-out patches (Qu & Hartzell, 2001).
Physiological implication of present results
NFA has often been used to probe the role of ICl(Ca) in producing smooth muscle contraction (e.g. Large & Wang, 1996; Greenwood & Large, 1999). This approach was based on the observation that in previous studies NFA was shown to be potent compared to other blockers of ICl(Ca) as well as relatively selective. However the present work shows that under conditions where [Ca2+]i is maintained at a sustained level at physiological resting membrane potentials (about −50 to −60 mV) NFA produced an increase in ICl(Ca) with [Ca2+]i at 250 and 500 nm and only a small decrease of ICl(Ca) when [Ca2+]i was 1 μm. These [Ca2+]i are readily achieved physiologically during smooth muscle contraction, for example to agonists. Therefore, if the present results observed in single cells also occur in intact preparations it is possible that there may be physiological responses involving ICl(Ca) which may not be sensitive to NFA, notably in cells where [Ca2+]i is elevated to around 500 nm before NFA is applied, i.e. the lack of effect of NFA on vasoconstrictor responses may not necessarily mean that ICl(Ca) is not involved in that response.