Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

KCNQ potassium channels occur in the heart, nervous tissue, and many epithelia (Busch and Suessbrich, 1997) and have important physiological functions. KCNQ1 (K_VLQT₁) can form heteropolymeric channels with an accessory subunit, KCNE1 (minK) to form I_Ks channels, responsible for the slow component of the delayed rectifier potassium current that contributes to the repolarization of the cardiac action potential (Sanguinetti et al., 1996). We know that the sensitivity of KCNQ1 channels to both activators and inhibitors, such as mefanamic acid and 293B, respectively, is enhanced by complex formation with minK (Busch et al., 1997). A recent study can be interpreted to mean that chromanols, such as 293B, bind to the subunit KCNQ1, with the minK unit acting as an allosteric activator (Lerche et al., 2000). The cognitive enhancer XE991 increases stimulus evoked transmitter release in the central nervous system (Zaczek et al. 1998), probably by blockade of the M-channel (Aiken et al., 1996), a heteromultimer of KCNQ2 and KCNQ3 channel subunits (Wang et al., 1998a). XE991 also blocks KCNQ1 channels expressed in oocytes, but its efficacy is reduced when the channel is expressed with KCNE1 (Wang et al., 2000). Presumably, if the activity of XE991 had been enhanced when KCNQ1 channels were expressed with KCNE1, the potential to cause long QT syndrome would have increased. Thus, the selectivity of the cognitive enhancer is dependent on the type of subunits colocalized with particular K⁺ channels.

Activation or inhibition of K⁺ channels in epithelia, particularly those secreting chloride, has potentially useful therapeutic applications in diseases as diverse as cystic fibrosis, where chloride secretion is deficient, and secretory diarrhea, where it is in excess. We have shown previously that the chromanol 293B, although very active on intestinal epithelia, is virtually inactive in airway epithelia, whereas for clofilium, the situation is reversed (MacVinish et al., 1998; Cuthbert et al., 1999b). In this report, we examine the effect of XE991 on K⁺ channels in chloride secretory epithelia from the intestine and the airways of mice. Our objective was to examine if the differences seen with 293B also applied to XE991 and, if so, whether this could be correlated with differences in the type of K⁺ channels present in the two epithelia. Again we found that the effects of XE991 are very different in these two types of epithelia. Sequences from various murine KCNQ-type K⁺ channels and minK-like peptides were amplified by PCR to allow correlation of function with channel type. Although the colonic tissue showed the presence of both KCNQ1 and KCNE3 mRNA, the airway epithelium also showed, uniquely, the presence of mRNA for KCNQ2.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

All the SCC experiments were performed on the isolated colonic mucosae or nasal epithelium of balbC mice. Mice were killed by CO₂ narcosis and the colons were placed in cold Krebs Henseleit solution (KHS) of the composition: 117 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl₂, 1.2 mM MgCl₂, 1.2 mM KH₂PO₄, 25 mM NaHCO₃, and 11.2 mM glucose, pH 7.6 when bubbled with 95% O₂/5% CO₂ at 37°C. The muscle layers were dissected away and small pieces of mucosae were mounted in Ussing chambers with a window area of 20 mm². Up to four pieces of mucosae were used from a single mouse. Tissues were bathed on both sides with 20 ml of KHS at 37°C that was continually circulated with the gas lift using 95% O₂/5% CO₂. The tissues were voltage-clamped at 0 mV to measure the short circuit current (SCC) using a WPI Dual Voltage Clamp (Stevenage, Herts, UK) connected via an ADI Instruments Maclab 8e (Hastings, Sussex, UK) to a computer. The arrangement of voltage sensing and current passing electrodes was as described elsewhere (Cuthbert et al., 1999a) and current compensation for the current × time drop between the voltage electrodes was used. When colonic epithelia from IsK knockout mice were used, comparisons were made with tissues from their wild-type littermates. Nasal epithelia were mounted in Ussing chambers and voltage clamped as for the colonic epithelia and bathed in KHS solution as for the colon. However, the window area in the chambers was only 1.8 mm². The technique for the removal of the nasal epithelium from the nasal cavity was exactly as described elsewhere (MacVinish et al., 1997)

Detection of K⁺ Channel Messenger RNA. Heart, lung, nasal epithelium, colon and jejunal tissues were dissected from balbC mice, and quick frozen in isopentane. Total RNA was isolated from each sample using the TRIzol extraction method (Invitrogen, Paisley, UK) (Chomczynski and Sacchi, 1986). For nasal epithelium tissues from ten mice were pooled; all other tissues were sampled from single animals. RNA samples were run on ethidium bromide agarose gels and gave the appropriate 28S to 18S bands, confirming the lack of degradation. Samples were double DNase digested before reverse transcription (Superscript II; Invitrogen, Paisley, UK) and the resulting cDNA used for PCR. Sense primers 5' and antisense primers 3' were designed for each K⁺ channel or minK-like protein using Primer 3 software (http://www.genome.wi.mit.edu/genome_software/other/primer3.html) using the GenBank accession numbers given in Table 1. After reverse transcription using 1 µg total RNA, specific PCR products were obtained, but not in the absence of reverse transcription. The nature of the PCR products was confirmed by sequencing the amplified products on an ABI 310 sequencer (Applied Biosystems, Warrington, UK).

Expression in Xenopus laevis Oocytes. Full-length clones of hKCNQ1 and murine KCNE3 were obtained. The former was a kind gift from Dr. J. Vandenberg, and the latter was prepared by the usual methods. Standard PCR procedures were used to construct a Kozak consensus sequence GCCACC, immediately upstream of the ATG initiation codon in cDNA clones for KCNE3. These constructs were then subcloned into the HindIII-BamHI site of the pBG7.2 vector, which provides the 5'- and 3'-untranslated regions of the X. laevis -globin gene. In vitro transcription was achieved using capped cRNA made from NdeI-linearized cDNA using the T7 polymerase mMESSAGE mMACHINE Kit (Ambion, Austin, TX) according to the manufacturer's instructions.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of XE991 on Colonic Epithelia. The effects of XE991 on the chloride secretory responses to two agonists, forskolin and EBIO, were investigated by short circuit current procedures. In all experiments, 100 µM amiloride was present in the apical bathing solution to eliminate electrogenic sodium absorption. Addition of 30 µM XE991 to colonic epithelia without prior stimulation caused a significant reduction in the basal SCC. In 11 preparations, the basal current was 54.5 ± 6.9 µA cm², which was reduced to 29.2 ± 4.1 µA cm² (P < 0.005) in 10 min after addition of XE991, representing a 46.4% reduction in basal SCC. Because the basal current is known to result in part from the secretion of chloride ions (Cuthbert et al., 1999a) XE991 seems to inhibit this activity in the absence of a secretory stimulus.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Effects of XE991 on the responses to forskolin in colonic epithelia. SCC records from a pair of colonic epithelial preparations. a, the tissue was exposed to 10 µM forskolin, and after the current had stabilized, 30 µM XE991 was added. b, the tissue was exposed to 30 µM XE991 for 20 min before exposure to forskolin. Note that 1 mM furosemide produces an increase in SCC after exposure to XE991. c, a partial concentration response curve to forskolin is shown in the presence and absence of XE991 (30 µM). The inhibition of 10 µM forskolin responses by XE991 is shown in d, giving an EC50 value of 10 µM. In c and d, the number of observations is given in parentheses.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Contrasts effects of XE991 on the responses to EBIO in the presence and absence of ChTX. a, XE991 (30 µM) was added basolaterally to a colonic epithelium after the response to 600 µM EBIO had reached steady state. b, the same procedure was used except the tissue had been preincubated with 50 nM ChTX for 20 min before EBIO was added.

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Contrasts the inhibitory effects of XE991 on the colonic epithelial responses to forskolin and to EBIO. The inhibitory effects of 30 µM XE991 on the responses to forskolin are shown in a and b. a, XE991 was added to the basolateral side after the response to 10 µM forskolin had reached a steady state value. b, tissues were exposed to 30 µM XE991 for 20 min before forskolin responses were assessed and the results compared with those in the absence of XE991. Significant inhibitions of the responses to forskolin were found in a and b; mean inhibition corresponded to 79.2% in a (P < 0.0002) and 70.6% in b (P < 0.0002). In c and d, the secretory response was evoked with 600 µM EBIO. Addition of 30 µM XE991 in c had no significant effect on the current, reducing the response by only 3.9%. In d, 50 nM ChTX was added to the basolateral side for 20 min before EBIO was added. Addition of 30 µM XE991 after EBIO caused a significant reduction of the response of 81.8% (P < 0.04).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Inhibition by ChTX together with XE991 on the responses to EBIO. The effect of 600 µM EBIO on SCC and its inhibition by 1 mM furosemide is shown in a. b, a paired preparation was subjected to the same protocol after incubation for 20 min with 50 nM ChTX and 30 µM XE991 applied basolaterally. Note the reversal of the effect of furosemide in b. Quantitative data from four pairs of colonic epithelial preparations, treated as in a and b, are presented in c and d.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Effect of XE991 on epithelial K+ currents. Colonic epithelia were subjected to an apical to basolateral potassium ion gradient. Essentially, this consisted of replacing NaCl in the apical solution with potassium gluconate and in the basolateral solution with sodium gluconate (see Cuthbert et al., 1999b). Nystatin (180 µg/ml) was used to permeabilize the apical membranes, resulting in an inward K+ current. Inhibition by 30 µM XE991 in the presence and absence of 600 µM EBIO is shown in a and b, respectively. The insets show the data from several experiments, the reduction in current being significant in both instances (paired t test). a, XE991 caused a reduction in current of 87.1 ± 10.3% and of 26.2 ± 6.8% in b; these values were significantly different (P < 0.002)

Effects of XE991 on Nasal Epithelia. In a previous report, we showed that the effects of K⁺ channel blockers on chloride secretion depended on the particular epithelium investigated (MacVinish et al., 1998; Cuthbert et al., 1999). For example, the chromanol 293B was effective in the colon but had only a minor effect on murine nasal epithelium, indicating that different K⁺ channels were involved in the two tissues. The same proved to be true for XE991, which caused less than 10% inhibition of the forskolin-activated chloride current in murine nasal epithelium. The ineffectiveness of XE991 in nasal epithelium was not different when used alone or in the presence of ChTX (Fig. 6, a to c). By contrast, clofilium, a class III antiarrhythmic agent, was effective in blocking the chloride secretory effect, either alone or in the presence of ChTX plus XE991.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Relative lack of effect of XE991 on nasal epithelium. a, addition of 100 µM XE991 to the basolateral side of murine nasal epithelium, previously stimulated with 10 µM forskolin had a small but significant effect on the SCC using a paired t test. The reduction in the current had a mean value of 11.5%. b, record showing the effect of 100 µM XE991 and 1 mM furosemide on SCC after stimulation with 10 µM forskolin. c, as in b, showing that 50 nM ChTX does not increase the inhibitory effect of XE991. Clofilium (100 µM) rapidly reversed the effect of forskolin.

Effects of XE991 on Jejunal Epithelium. To test whether the profile described for murine colonic epithelium was unique, a limited study was made with murine jejunal epithelium. Phloridzin (100 µM) was added to the apical bathing solution to inhibit electrogenic sodium/glucose transport. The tracing illustrated in Fig. 7 shows that, like the colon, the forskolin-sensitive current is sensitive to XE991, whereas the residual current is insensitive to clofilium but sensitive to furosemide. Consequently, the jejunum is similar to the colon and unlike the nasal epithelium.

View larger version (11K): [in this window] [in a new window]   Fig. 7.   Sensitivity of jejunal epithelium to XE991. Murine jejunal epithelium was treated with amiloride and phloridzin (both 100 µM) on the apical side to block electrogenic sodium absorption and electrogenic sodium-glucose transport. Forskolin (10 µM) increased SCC, and this was sensitive to 100 µM XE991, causing 85% inhibition. The residual current was insensitive to 100 µM clofilium, but sensitive to 1 mM furosemide.

Which K Channels Are Present in Murine Colonic and Nasal Epithelium? To discover whether the difference in response to XE991 in nasal and colonic epithelia can be ascribed to different K⁺ channels, RNA was extracted from various tissues and reverse transcribed. The resulting cDNAs were then used for PCR. Pairs of primers were selected for the  and  subunits of the KCNQ group of K⁺ channels and for KCNN4, an intermediate-conductance, calcium-activated potassium channel. No primers for KCNQ3 and KCNE2 were made because the murine sequences for these are unknown. The primers were designed to give 200-bp products from samples extracted from heart, lung, nasal epithelium, colon and jejunum. A typical gel, showing the presence of 200-bp products using primers for KCNQ1, KCNQ2, KCNQ4, KCNQ5, KCNE1, KCNE3, KCNE4, and KCNN4, is shown in Fig. 8.

View larger version (60K): [in this window] [in a new window]   Fig. 8.   PCR products obtained using primers specific for murine K+-channels. The expected size of the products was 200 bp in all instances. Ladders (1 kilo-bp) are shown between each group. H, L, N, C, and J correspond to cDNAs made from murine heart, lung, nasal epithelium, colon, and jejunum.

Observations with IsK Knockout Mice. Further evidence for the involvement or otherwise of KCNE1 subunits (minK) in the chloride secretory response was obtained by experiments with tissues from IsK knockout mice comparing these with wild-type controls. Inhibition of responses to forskolin by clofilium and 293B were measured in nasal and colonic epithelia, respectively, whereas XE991 was examined in only the mutant colonic tissue. No differences in the inhibitory responses were found between tissues without KCNE1 and those possessing the subunit, suggesting that the KCNE1 subunit is not involved in the inhibition of chloride secretion by K⁺ channel blockers (Table 2).

Expression in X. laevis Oocytes. Indirect evidence above suggests that KCNQ1/KCNE3 multimers is the target for XE991. To test this directly, KCNQ1 was expressed in X. laevis oocytes, either alone or together with KCNE3, and examined by two-electrode voltage clamping. When KCNQ1 was expressed alone, a voltage-dependent current-voltage relationship was found, whereas in the presence of KCNE3, the relationship was linear (Fig. 9c). XE991 was more effective in inhibiting the clamp current, measured at 6 s after stepping the voltage from 80 mV to +40 mV, when both cRNAs were expressed than when KCNQ1 was expressed alone (Fig. 9d). The EC₅₀ value was approximately 20 µM in the presence of both KCNQ1 and KCNE3, but was >100 µM in the presence of KCNQ1 alone. Water- or KCNE3-injected oocytes showed virtually no current responses to changes in voltage.

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Data obtained with X. laevis oocytes injected, 3 to 4 days previously, with cRNA for KCNQ1, KCNE3, or both. a and b, families of current curves for voltage-clamped oocytes in response to stepwise changes in membrane potential from 80 mV to +40 mV, in steps of 20 mV. a, an oocyte was injected with cRNAs for both KCNQ1 and KCNE3, whereas in b, the oocyte was injected with the cRNA for KCNQ1 only. c, data for five oocytes injected with both messages and 12 oocytes injected with only the message for KCNQ1. Mean values and S.E. are shown. At all voltages, the current values for oocytes injected with both cRNAs was significantly greater than the currents obtained with oocytes injected with only KCNQ1 (by at least P < 0.01). The mean current generated in water injected oocytes voltage clamped at +40 mV was 0.18 µA (n = 3) and for oocytes injected with the cRNA for KCNE3 only and clamped at +40 mV was 0.16 µA (n = 3). d, the percentage inhibition of the voltage clamp current at 6 s and +40 mV by XE991, added externally to the bath. Mean values and S.E. are given for five oocytes injected with cRNAs for KCNQ1 and KCNE3 and for seven oocytes injected with cRNA for KCNQ1 only. The values at XE991 concentrations of 30 µM and 100 µM were significantly different (P < 0.012 and P < 0.019, respectively).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have shown that XE991 can reduce electrogenic chloride secretion in the murine colon when it is added basolaterally. This result supports the general thesis that hyperpolarization increases chloride efflux through the apical membrane by increasing the electrical gradient. Conversely, blockade of basolateral K⁺-channels, by causing depolarization, will reduce chloride secretion. Because XE991 reduces the basal SCC without prior stimulation of SCC, this suggests that some K⁺ channels are constitutively open, as has been argued by others (Schroeder et al., 2000).

Both forskolin and EBIO have multiple actions on chloride secretory epithelia. Both activate apically located CFTR chloride channels, either by increasing cAMP (Cuthbert et al., 1999b) or by direct action (Devor et al.,1996a,b), and both activate basolateral K⁺ channels, again either via cAMP or by a mixture of direct and indirect actions (Devor et al.,1996a,b; Cuthbert et al., 1999b; Syme et al., 2000). However, as shown here for the colon, XE991 inhibits the response to forskolin without a significant effect on the response to EBIO. This suggests that the target for XE991 is the cAMP-sensitive K⁺ channel, which has been shown to be a KCNQ1-KCNE3 heteropolymeric channel (Schroeder et al.,2000). Messenger RNAs for both the channel alpha subunit [KCNQ1(KvLQT1)] and the associated minK-related peptide (KCNE3) are present in the murine colon (Fig. 7). Furthermore, it is presumably these channels that are constitutively open in the absence of agonists.

EBIO activates intermediate conductance Ca²⁺-sensitive K⁺ channels by increasing number × open probability (Syme et al., 2000). XE991 had no significant effect on the responses to EBIO unless the tissues were pretreated with ChTX, after which XE991 was effective. The presence of mRNA for the intermediate conductance Ca²⁺-sensitive K⁺ channel (KCNN4) was shown to be present in the epithelial tissues used in this study. Because, it is argued, the cAMP-sensitive K⁺-channels are constitutively open, then in the presence of EBIO, both the Ca²⁺- and the cAMP-sensitive K⁺ channels will be activated and hyperpolarization is probably supramaximal; the rate of chloride secretion is rate-limited by the chloride permeability of the apical membrane. Under these conditions, blockade of either the cAMP- or Ca²⁺-sensitive K⁺ channels will have little effect and blockade of both types of channel is necessary to see inhibition. Most of the data presented with the IsK knockout mice tissues were obtained at a much earlier time, before the work on XE991 was commenced. At that time, the existence of multiple minK-related peptides was only postulated (Attali et al., 1993) and the failure to see any differences in responses in mice with and without minK was puzzling. Recently, we carried a few further studies using IsK knockout tissues in which XE991 was used instead of 293B. Again, no differences were found, confirming the earlier findings with 293B. Two conclusions from these findings might be as follows: first, the KCNQ1/KCNE1 heteropolymer is not the target for either XE991 or 293B; second, XE991 and 293B probably have the same target. Because the mRNA for KCNE1, the original minK, was found predominantly in the heart (Fig. 7), it is unlikely that this minK peptide is involved in functional responses in the colonic or nasal epithelium. However, it does not follow that XE991 and 293B do not interact with the heteropolymeric channel KCNQ1/KCNE1; indeed, both drugs have been shown to be active on cardiac tissues or oocytes expressing both components (Bosch et al., 1998; Wang et al., 2000). It is clear that the target for 293B in the colon is the KCNQ1/KCNE3 K⁺ channel (Schroeder et al., 2000; Kunzelmann et al., 2001), and because XE991 and 293B are equally effective in wild-type and IsK knockout mice, obviously KCNE1 is not essential for the activity of either agent. However, when KCNQ1 is expressed with KCNE1 in oocytes, the sensitivity to XE991 is reduced (Wang et al., 2000), whereas it is enhanced to 293B (Busch et al., 1997). Assuming KCNE3 also reduces the affinity of XE991 for the colonic K⁺ channel, then the EC₅₀ value of 10 µM found in this study is impressive. It is possible that, in the colon, there is some KCNQ1 uncomplexed with KCNE3 and that XE991 interacts with this entity, whereas 293B reacts with the KCNQ1/KCNE3 complex.

The reversal of the responses to furosemide after XE991 is to be expected. The chloride secretory response in the colon is accompanied by modest K⁺ secretion (Cuthbert et al., 1999a). When the former is inhibited by XE991, subsequent blockade of the Na-K-2Cl cotransporter will inhibit the latter, giving an increase in SCC.

PCR analysis of cDNAs derived from a variety of murine tissues, including epithelia, confirm that KCNN4, a Ca²⁺-dependent K⁺ channel, is present in both nasal and colonic epithelia. We and others have shown that EBIO activates this channel (Cuthbert et al., 1999b; Syme et al., 2000), although its effects on nasal epithelium are small and somewhat transient compared with the effects on the colon. In airway epithelia from other species (i.e., man), the effects of EBIO are more pronounced (Devor et al., 2000). The reasons for these differences are unknown, and we have no data on the expression levels of the Ca²⁺-sensitive K⁺ channel protein in the nasal epithelium of mice. The nasal epithelium of the mouse was unique in another way: it was the only tissue containing the mRNA for KCNQ2. Whether KCNQ2 can form heteropolymers with KCNE3, which are insensitive to XE991, is unknown and can only be determined by expression studies in cells lacking these entities. It is known that KCNQ2 is rather insensitive to clofilium (Yang et al., 1998), whereas it is sensitive to XE991. Thus the pharmacological profile of clofilium and XE991 in the nasal epithelium remains obscure. Possibly other -subunits, still undiscovered, will provide the answer. KCNQ2 can form modulated K⁺ channels with KCNE2 subunits (Tinel et al., 2000) but the murine homolog of KCNE2 has yet to be discovered. 293B, which seems to share with XE991 the same target in colonic epithelia, is similarly unable to inhibit the forskolin-generated current in nasal epithelium (MacVinish et al., 1998). XE991 is thought to act on KCNQ2/KCNQ3 multimers in the central nervous system (Wang et al., 1998a) when acting as a cognitive enhancer and will block KCNQ1 channels when expressed in oocytes (Wang et al., 2000), its activity being reduced when expressed with KCNE1. A drastic reduction in sensitivity to XE991 through a similar mechanism is unlikely in murine nasal epithelium, because little or no mRNA for KCNE1 is detectable. Although the concentration of XE991 was raised to 100 µM, inhibition of the forskolin responses in nasal epithelium was only 10%. We have no evidence that this is caused by inhibition of uncomplexed KCNQ1. Forskolin-stimulated current in nasal epithelium was blocked by clofilium. The long QT syndrome is associated with mutations in the KCNQ1 gene that forms channels with the KCNE1 -subunit. When expressed together in oocytes, currents were inhibited by clofilium (Yang et al., 1997), yet this agent is ineffective in the colon (MacVinish et al., 1998). Because the KCNQ1/min K family of channels can have multiple stoichiometries (Wang et el., 1998b) the differences between nasal and colonic epithelia and the actions of inhibitors such as XE991, clofilium, and 293B may be very subtle. The mRNA for KCNQ5 was found only in the lung and is therefore irrelevant to this discussion. KCNQ4, generally associated with sensory pathways in neuronal tissue (Kharkovets et al., 2000), may be expressed in both the nasal and colonic epithelium. Whether KCNQ4, either alone or complexed with a minK-like protein, may be a target for clofilium or XE991 requires expression studies.

From the limited study of the mouse jejunum, its profile with respect to the sensitivity to XE991 seems to be identical to that of the colon. mRNA for KCNQ1 and KCNE3 are present in this tissue and this together with the ability of XE991 to inhibit the chloride secretory current generated by forskolin supports this view.

Finally, to show that KCNQ1/KCNE3 multimers are the target for XE991, at least in colonic epithelium, we expressed both these proteins in X. laevis oocytes. We obtained current voltage relationships for KCNQ1 alone and combined with KCNE3 that were similar to those reported by others (Schroeder et al., 2000; Kunzelmann et al., 2001). Importantly, although the EC₅₀ value for XE991 was 20 µM for oocytes expressing both KCNQ1 and KCNE3, the value was more than 100 µM for oocytes expressing KCNQ1 alone. This contrasts with the finding that expression of KCNQ1 with KCNE1 reduces the sensitivity to XE991 (Wang et al., 2000). The EC₅₀ value for the KCNQ1/KCNE3 multimers expressed in oocytes (20 µM) is near the value found for the inhibition of the forskolin induced SCC by XE991. Thus, the results in oocytes support the deductions we have made for the murine colon.

In summary, XE991 is able to inhibit the chloride secretory current generated by activation of CFTR chloride channels and cAMP-sensitive K⁺ channels by targeting the latter, which are made up of KCNQ1/KCNE3 multimers, while having no effect on Ca²⁺-sensitive K⁺ channels. The low sensitivity of airway epithelial K⁺ channels to XE991 may result from subtle, as-yet-undiscovered differences in the assembly of  and  subunits of the channel.