Results

Identification of KV7 channels within MA ECs
Initial investigation sought to identify KV7.4 and KV7.5 channel transcript and protein within MA ECs, as these subtypes are implicated in vascular reactivity (Barrese et al , 2018). Transcript levels for EC and VSMC markers were determined in cell lysates from MA with and without EC removal by mechanical abrasion (see Methods). In cell suspensions from EC-denuded vessels there was a reduction in EC markers (Pecam/Vwbf ) and an increase in VSMC markers (Acta2/Myh11 ) compared to lysates from unabraded MA (Figure 1). This was associated with a 1.8- and 2.1-fold decrease in relative transcript abundance of Kcnq4 and Kcnq5, (Figure 1) showing that a considerable component of these transcripts were derived from EC cells. This data suggests expression of Kcnq4 andKcnq5 within EC marker expressing cells (Figure 1).
KV7.4 was KV7.5 was detected in isolated ECs by immunodetection (Figure 2A,C). KV7.4 had a punctate distribution in isolated ECs (Figure 2A) whereas KV7.5 label appeared to be predominantly cytoplasmic with some diffuse label around the nucleus (Figure 2C). Similar to previous reports (Zhong et al. , 2010; Oliveras et al. , 2014; Mills et al. , 2015; Morales-Cano et al. , 2015, Barrese et al., 2018) KV7.4 and 7.5 were identified in isolated MA VSMCs (Figure 2B,D). Notably, KV7.4 and KV7.5 were also detected in both ECs and VSMCs inen face whole-mount tissue (Figure 2.E-H). KV7.4 was also detected in EC by immunoelectron microscopy (supplemental figure 1). Further, both KV7.4 and 7.5 were expressed at a proportion of IEL hole sites at an apparently higher level than the associated EC membrane label (Figure 3).
Removal of ECs modulates KV7.2-5 activator efficacy
A comprehensive pharmacological analysis was undertaken to determine if KV7 channels have a functional role in MA ECs. Effects of KV7 channel modulators were examined in MAs where the endothelium was intact or had been removed. S-1 and ML213 are structurally dissimilar activators of KV7.2-7.5 that interact with the same pharmacophore centered around a tryptophan in the S5 domain (Bentzen et al., 2006; Lyubov I. Brueggemann, Haick, Cribbs, & Byron, 2014; T. A. Jepps et al., 2014; Schenzer, 2005). ML277 is a potent activator of KV7.1 (Yu, 2013) with a 100-fold increase in selectivity for KV7.1 compared to KV7.2-5 (Yu, 2013). Consistent with previous findings (Chadha et al., 2012; T. A. Jepps et al., 2014), S-1- and ML213-mediated vasorelaxation was ablated by pre-incubation with 10 µmol-L-1 pan-KV7 channel inhibitor linopirdine (Schnee & Brown, 1998); Figure 4.A,B). Relaxations produced by 10 to 300 nmol-L-1 ML277 were also prevented by incubation with linopirdine (Figure 4C). However, relaxations produced by concentrations >1 µmol-L-1 ML277 were not attenuated by linopirdine and are therefore not mediated by Kv7.1 activation.
EC removal for the following experiments was confirmed by ablation of vasorelaxation in response to 10 µmol-L-1 CCh (Figure 5A). Removing the endothelium by mechanical abrasion has no impact on the peak contraction produced by 300 nmol-L-1 U46619 (Figure 4B), but significantly attenuated the potency of S-1 mediated vasorelaxation increasing EC50 from 2±0.2 µmol-L-1 to 3±0.7 µmol-L-1 (Figure 5C,D). The potency of ML213 was also impaired by endothelial removal (EC(+) EC50 = 1±0.2 µmol-L-1 vs EC(-) EC50 = 3±0.2 µmol-L-1; Figure 5E). The linopirdine-sensitive relaxation produced by ML277 was not affected by endothelial removal (Figure 5F). The accumulated findings reveal that ECs express KV7.4 and KV7.5 and the presence of the endothelium enhances the sensitivity of two structurally different KV7.2-7.5 activators.
IKCa/SKCa inhibitors have no impact on KV7 activator mediated relaxation
Having identified that the endothelium modulates responses to Kv7 activators, experiments were performed to identify the mechanism/s involved. Endothelial IKCa and SKCachannels contribute to relaxation responses in rat MA (Crane, Gallagher, Dora, & Garland, 2003). Thus, it is feasible that KV7 channels interact with other key endothelial potassium channels; particularly in microdomains (Sandow et al., 2009). However, in agreement with previous reports (T. A. Jepps, Olesen, Greenwood, & Dalsgaard, 2016), pre-incubation with a combination of IKCa inhibitor TRAM-34 (1 µmol-L-1; (Wulff et al., 2000) and SKCa inhibitor apamin (100 nmol-L-1; (Spoerri, Jentsch, & Glees, 1975) had no effect on KV7 activator mediated vasorelaxation (Figure 6A-C).
EC Kir channels modulate KV7.2-5 activator sensitivity
Within MAs of a EC-KIR2.1-/- murine model, endothelial KCNJ2-encoded KIR2.1 channels have been identified as ‘signal boosters’ that enhance EC-derived relaxation (Sonkusare et al., 2016). Comparatively, the current literature regarding rat mesenteric KIR2 channels is limited; with KIR2.1 expression being demonstrated in EC of whole mount rat MA (Dora, Gallagher, McNeish, & Garland, 2008), where inwardly rectifying Ba2+ sensitive channels are apparently restricted to the endothelial layer (Crane, Walker, Dora, & Garland, 2003), and KIR channels contribute to acetylcholine-mediated responses (Goto et al., 2004). We propose that like mice, rat mesenteric ECs express functional KIR2 channels that propagate EC signals in a similar process. Therefore, we performed a series of studies investigating the effect of two well characterized KIR2 blockers, BaCl2(Hagiwara, Miyazaki, Moody, & Patlak, 1978) and ML133 (Wang et al., 2011; Wu et al., 2010) on KV7 activator-mediated vasorelaxation.
In arteries with a functional endothelium, KIR2 blockers, BaCl (100 µmol-L-1) and ML133 (20 µmol-L-1), significantly impaired vasorelaxation by S-1 (Figure 7A, EC50= DMSO, 1.89±0.2 µmol-L-1 / BaCl, 2.3±0.31 µmol-L-1; Figure 7B, EC50= DMSO, 0.52±0.12 µmol-L-1 / ML133, 3.1±1.5 µmol-L-1, and ML213, Figure 7C,D, EC50= DMSO, 0.9±0.3 µmol/L-1 / BaCl, 2.2±0.5 µmol/L-1 / ML133, 2.5±0.25 µmol-L-1) when compared to DMSO solvent control (Figure 7A-D). No attenuation of the response to ML277 was observed consistent with EC removal data. In arteries where the endothelium had been removed neither ML133 nor BaCl had any effect on ML213 mediated vasorelaxation (Figure 7G,H). These data show that the endothelium-dependent increase in potency to the KV7 activators involves endothelial KIR channels, but not IKCa or SKCa channels.
KV7 channels contribute to CCh evoked vasorelaxation
The expression of functional KV7 channels within ECs begs the question - Do they contribute to EC-derived responses? Acetylcholine produces endothelium-dependent relaxations through NO-, EDH- and prostanoid-dependent mechanisms in rat MA (Parsons, Hill, Waldron, Plane, & Garland, 1994; Peredo, Feleder, & Adler-Graschinsky, 1997; Shimokawa et al., 1996).
A distinct rightward shift in the sensitivity to vasorelaxation in response to CCh, a synthetic acetylcholine analogue, was produced by 100 µmol-L-1 eNOS inhibitor L-NAME when compared to DMSO solvent control (EC50 DMSO = 0.59±0.1 µmol-L-1; L-NAME = 0.94±0.1 µmol-L-1; Figure 8A). A combination IKCa and SKCainhibitors, 1 µmol-L-1 TRAM34 and 100 nmol-L-1 apamin, respectively produced greater attenuation (EC50 TRAM34/apamin=1.5±0.7 µmol-L-1; Figure 8A). These results show that both NO- and EDH-dependent signaling contributes to CCh-mediated vasodilation, though the main contributor to endothelial-dependent vasodilation in 2nd order MA appears to be EDH, conferring with previously published data (Shimokawa et al., 1996). Pre-incubating vessels with 10 µmol-L-1 pan-KV7 channel inhibitor linopirdine significantly attenuated CCh-mediated vasorelaxation when compared to DMSO control (EC50 DMSO = 0.2±0.08 µmol-L-1; linopirdine = 0.7±0.3 µmol-L-1; Figure 8B). In contrast, pre-incubating vessels with either the KV7.1 specific inhibitor HMR-1556 (10 µmol-L-1; Figure 8C) or a combination of non-specific KV channel inhibitors TEA (1 mmol-L-1; (Choi, Aldricht, Yellen, & Hughes, 1991) and 4-AP (1 mmol-L-1 ; (Kurata & Fedida, 2006) had no significant effect on CCh-evoked vasorelaxation. These results imply a specific contribution of KV7.4 and KV7.5 channels to CCh-evoked relaxation in 2nd order rat MA.
Subsequent experiments explored whether linopirdine affected CCh-induced relaxations in the presence of L-NAME compared to responses in apamin/TRAM34. CCh-evoked relaxations were significantly attenuated in vessels pre-incubated in TRAM34/apamin and linopirdine compared to vessels only pre-incubated in TRAM34/apamin alone (EC50DMSO = 0.24±0.05 µmol-L-1; TRAM34/Apamin = 0.27±0.03 µmol-L-1; TRAM34/Apamin + linopirdine = 0.61±0.2 µmol-L-1; Figure 8E). In contrast, linopirdine failed to attenuate CCh relaxations in arteries pre-incubated with L-NAME (Figure 8F). These data suggest that KV7.4 and KV7.5 channels contribute to the NO pathway of CCh-mediated vasorelaxation.
KV7 channels contribute to NO-mediated vasorelaxation in a vascular bed specific manner (Jennifer B. Stott et al., 2015). The present data demonstrate that pre-incubation with the pan-KV7 channel blocker linopirdine (10 µmol-L-1) has no effect on vasorelaxation produced by the NO-donor s-nitroprusside (SNP; Figure 8G). However, in contrast with previous reports (T. A. Jepps et al., 2016), pre-incubation with L-NAME (100 µmol-L-1) significantly attenuated KV7.2-5 activator mediated vasorelaxation (Figure 8H). These data suggest that the NO axis of CCh-mediated vasorelaxation is linopirdine sensitive, upstream of NO signaling.

Discussion

The present study identified Kcnq4 and Kcnq5 transcripts within EC marker expressing cells as well as respective KV7.4 and KV7.5 protein in isolated and whole mount rat MA ECs/VSMCs, consistent with previous studies in the endothelium (Chen, Li, Hiett, & Obukhov, 2016) and vascular smooth muscle (e.g Zhong et al. , 2010; Oliveras et al. , 2014; Mills et al. , 2015, summarised in Barrese et al ., 2018). Functionally, the present study demonstrates that the relaxation produced by two structurally different KV7.2-5 activators, but not a KV7.1 activator, were modulated by the endothelium. These relaxations were also sensitive to KIR2 inhibition and suggest a novel functional interaction between KV7 channels and endothelial KIR2.x channels. Furthermore, we have demonstrated that KV7.4/7.5 channels contribute to the NO-mediated axis of CCh-evoked endothelium-dependent relaxations. Thus, KV7 channels are expressed within ECs, contribute to endothelium-derived responses and when pharmacologically upregulated, are functionally coupled to other EC potassium channels.

KV7 channel expression and function within ECs

Within smooth muscle, when active, KV7 channels hyperpolarize the membrane potential decreasing voltage-gated calcium channel (VGCC) open probability, therefore decreasing the influx of extracellular calcium, with this decrease resulting in relaxation of the cell. Within rodent models, the pharmacopeia of KV7 channel modulators has revealed KV7.4 and KV7.5 channels; 1. contribute to resting vascular tone via regulation of the resting membrane potential; 2. are upregulated during cGMP and cAMP/EPAC/PKA-mediated vasodilation, and; 3. are supressed via PKC-mediated vasoconstriction. Conjunctly, the use of molecular techniques has revealed that KV7.4/KV7.5 is the most predominant heterotetramer expressed within specific rodent arteries. Comparatively, no functional role for KV7.1 in arteries has been identified (see (Byron & Brueggemann, 2018; Barrese et al , 2018), for review). A caveat of these observations is a lack of differentiation between VSMCs and ECs. However, KV7 channels were recently identified in pig coronary artery ECs (Chen et al., 2016) where they modulate bradykinin-evoked endothelium-derived vasorelaxation (Chen et al., 2016). The data presented here demonstrate KV7 channel expression in rat mesenteric endothelium and support their functional role.
In most arteries, the endothelium and smooth muscle are electrochemically linked via MEGJs formed from connexin proteins within heterocellular communicating microdomains (see (Sandow et al., 2012) for review). Via these connections, it has been purported that current injection into ECs passes into VSMCs (Sandow, Tare, Coleman, Hill, & Parkington, 2002). As KV7 channels are expressed within rat mesenteric ECs, the present data suggest a potential role for this process in KV7 activator-mediated vasorelaxation. This conjecture is supported by a significant attenuation of vasorelaxation of pre-constricted arterial tone by two structurally different KV7.2-5 activators in the absence of ECs. In contrast, EC removal had no impact on KV7.1 activator-mediated vasorelaxation, suggesting that only KV7.4 and KV7.5 channels are functionally expressed within MA ECs. However, Chen et al, (2016) suggest that endothelial removal had no impact on ML213-mediated relaxation in pig coronary artery segments. Consequently, these collective data suggest that this phenomenon is both animal and artery bed specific, potentially being dependent on the molecular architecture of the vessel, such as MEGJ and related microdomain properties.

A novel functional interaction with endothelial KIRchannels

KIR2.x channels have been characterized in a variety of rat vascular beds including cerebral and coronary arteries (Smith et al., 2008), where their selective inhibition by Ba2+revealed KIR channel amplification of a K+ channel activator conductance (Smith et al., 2008). However, there is a degree of conflict regarding the role of KIR2.x channels within rat MAs. As above, a Ba2+ sensitive current and KIR2.x expression has been demonstrated in rat MA ECs (Crane, Walker, et al., 2003; Smith et al., 2008). Furthermore, within primary MA KIR2.x channels were purported to contribute to KCa mediated-hyperpolarization during ACh-derived EC-dependent responses (Goto et al., 2004). In contrast, Smith et al. , (2008) demonstrate that within 3rd order MA ACh-mediated K+ conductance-dependent vasorelaxation was insensitive to Ba2+, indicating that KIR2.x channels do not augment K+conductance in these vessels during receptor-mediated vasodilation.
In the present study, significant attenuation of both S-1 and ML213 KV7.2-5 activator-mediated vasorelaxation was found after pre-incubation in two structurally different KIR2.x blockers ML133 (Wu et al., 2010) and Ba2+ (Hagiwara et al., 1978). ML133 has been identified via high-throughput and mutagenesis investigation as a novel inhibitor of KIR2.1 channels with an IC50 of 1.8 μmol-L-1 at pH 7.4, with little to no selectivity against the other members of the KIR2.x family (Wang et al., 2011; Wu et al., 2010) where it exerts its activity via D172 and I176 within the M2 region of KIR2.1 (Wang et al., 2011). Presently, ML133 is the most selective inhibitor of the KIR2 family. In parallel to the endothelium denudation experiments, no effect was seen on the KV7.1 activator ML277-mediated vasorelaxation; implying a specific interplay with KV7.4 and KV7.5 channels. Furthermore, KIR2 blockers had no effect on KV7.2-7.5 activator-mediated relaxation in EC-denuded arteries, supporting the notion that functional KIR2 channels (in the rat MA bed) are restricted to the endothelium (Crane, Walker, et al., 2003), and the agents do not inhibit KV7 channels per se . The present study suggests that pharmacological activation of endothelial KV7.4/7.5 channels also activates an endothelial KIR2 channel increased K+ conductance, which in turn accounts for a significant degree of the EC augmentation of KV7.2-5 activator-mediated vasorelaxation (Figure 11). However, based on the findings described by Goto et al,. (2004) and Smith et al,. (2008), the collective data suggest that this phenomenon is dependent on the branch order of MA. Furthermore, it remains unclear if this occurs during receptor mediated signaling, or that is only present during pharmacological activation of endothelial KV7 channels.
A primary concern for identification of novel functional interactions between ion channels using pharmacological tools is off-target effects. However, KV7 activator-mediated vasorelaxation in vessels pre-incubated in 10 µmol-L-1 linopirdine was abolished. If S-1 or ML213 were activating other channels, i.e. e.g. KIR channels, a degree of vasorelaxation would still be observed in the presence of linopirdine. The present findings therefore suggest that both S-1 and ML213 work exclusively via KV7 channels and that a novel functional coupling of KV7.4/7.5 and KIR2 occurs in rat MA ECs.

The contribution of KV7 channels to CCh evoked relaxations within rat MA

KV7 channels contribute to cGMP mediated vasorelaxation in a vascular bed specific manner (Stott et al., 2015). Within rat aortic, renal and pulmonary arteries, overexpression models, and native VSMCs KV7.4/KV7.5 have been identified as downstream targets of cGMP signaling during both isometric tension recording and whole cell current recording (Mondéjar-Parreño et al., 2019; Jennifer B. Stott et al., 2015). Conversely, KV7 channels do not contribute to NO donor SNP-mediated vasorelaxation within rat renal arteries (Jennifer B. Stott et al., 2015).
In light of the significant impact of KV7 inhibition on CCh-mediated vasorelaxation during the suppression of EDH, the present study suggests that KV7 channels contribute to CCh-evoked NO-cGMP mediated relaxation within rat MA. Though, similar to renal arteries, KV7 channels do not represent downstream targets of NO signaling in rat MA as KV7 channel inhibition does not impair SNP-mediated relaxations. However, eNOS inhibition does impair relaxation to a KV7 activator, implying that KV7 channels are involved in the production or release of NO in response to CCh. Though this appears to be a vascular bed specific phenomenon, as L-NAME significantly impairs ML213 relaxations in pig coronary artery (Chen et al., 2016), but not rat penile artery (T. A. Jepps et al., 2016).

Conclusions

In conclusion, the present data identify a novel functional interaction between mesenteric endothelial KV7.4/7.5 and KIR2 channels and supports the proposition that endothelial KV7 channels contribute to endogenous endothelial-derived responses. These findings highlight the complex nature of the vascular response to KV7 channel upregulation and emphasize the importance of KV7 channels to vascular signaling cascades. The present data are consistent with KV7 channels representing a novel therapeutic target in endothelial dysfunction.