PIP2 Mediated Inhibition of TREK Potassium Currents by Bradykinin in Mouse Sympathetic Neurons

Bradykinin (BK), a hormone inducing pain and inflammation, is known to inhibit potassium M-currents (IM) and to increase the excitability of the superior cervical ganglion (SCG) neurons by activating the Ca2+-calmodulin pathway. M-current is also reduced by muscarinic agonists through the depletion of membrane phosphatidylinositol 4,5-biphosphate (PIP2). Similarly, the activation of muscarinic receptors inhibits the current through two-pore domain potassium channels (K2P) of the “Tandem of pore-domains in a Weakly Inward rectifying K+ channel (TWIK)-related channels” (TREK) subfamily by reducing PIP2 in mouse SCG neurons (mSCG). The aim of this work was to test and characterize the modulation of TREK channels by bradykinin. We used the perforated-patch technique to investigate riluzole (RIL) activated currents in voltage- and current-clamp experiments. RIL is a drug used in the palliative treatment of amyotrophic lateral sclerosis and, in addition to blocking voltage-dependent sodium channels, it also selectively activates the K2P channels of the TREK subfamily. A cell-attached patch-clamp was also used to investigate TREK-2 single channel currents. We report here that BK reduces spike frequency adaptation (SFA), inhibits the riluzole-activated current (IRIL), which flows mainly through TREK-2 channels, by about 45%, and reduces the open probability of identified single TREK-2 channels in cultured mSCG cells. The effect of BK on IRIL was precluded by the bradykinin receptor (B2R) antagonist HOE-140 (d-Arg-[Hyp3, Thi5, d-Tic7, Oic8]BK) but also by diC8PIP2 which prevents PIP2 depletion when phospholipase C (PLC) is activated. On the contrary, antagonizing inositol triphosphate receptors (IP3R) using 2-aminoethoxydiphenylborane (2-APB) or inhibiting protein kinase C (PKC) with bisindolylmaleimide did not affect the inhibition of IRIL by BK. In conclusion, bradykinin inhibits TREK-2 channels through the activation of B2Rs resulting in PIP2 depletion, much like we have demonstrated for muscarinic agonists. This mechanism implies that TREK channels must be relevant for the capture of information about pain and visceral inflammation.


Introduction
Bradykinin (BK) is released during tissue damage, it is one of the main mediators of inflammation and is able to activate and sensitize the nociceptor neurons mainly through the activation of B 2 receptors (B 2 Rs) [1][2][3]. B 2 Rs are constitutively expressed in most tissues while B 1 receptors (B 1 Rs) are induced by the inflammatory process [2][3][4][5][6][7][8].
It seems coherent that the sensation of pain produced by BK must be mediated by its action on sensory neurons (nociceptors), whether somatic [2,3] or visceral [9,10]. However, the presence of bradykinin receptors (BRs) in motor neurons of the sympathetic superior cervical ganglion (SCG) was described a long time ago. When studying the effect of BK on the movement of the nictitating membrane of cats, BK was shown to be a potent stimulator of the SCG [11]. Consistently, BK increases the excitability (membrane depolarization at rest and reduction of the spike frequency adaptation (SFA)) of rat SCG (rSCG) neurons [4,[12][13][14]. Depolarization by activation of B 2 Rs has also been reported in the entire mouse ganglion [13,15,16]; however, changes in excitability have not been investigated in isolated cultured mouse SCG (mSCG) neurons. The importance of BR expression on SCG neurons is not completely understood. Studies have suggested that depolarization of presynaptic sympathetic-neuron terminals by BK induces the release of prostanoids which acting on nociceptor terminals may induce hyperalgesia. Consistently sympathectomy strongly reduces chloroform-induced hyperalgesia and its exacerbation by BK [4,17].
BK-induced excitability has been ascribed to the inhibition of the potassium M-current and such inhibition was reported to be due to B 2 R and Gαq/11 protein activation in rat and Gα11 in mouse SCG neurons [5,12,18]. The participation of phospholipase C (PLC), inositol triphosphate (IP 3 ), Ca 2+ released from IP 3 stores and activation of Ca 2+ -Calmodulin in this process was also reported in rSCG [14,19,20]. Traditionally, the regulation of the potassium M-current by muscarinic agonists has been considered the main physiological pathway to modulate the excitability of SCG neurons [21][22][23][24]. However, researchers have also reported that the same agonists do modulate background two-pore domain potassium (K2P) channels of the "Tandem of pore-domains in a Weakly Inward rectifying K + channel (TWIK)-related channels" (TREK) subfamily in these cells [25]. The two most expressed K2P channels in mSCG neurons are "TWIK-related spinal cord potassium channel" (TRESK) and TREK-2 [26,27] and it was reasonable to hypothesize that if the activation of muscarinic receptors inhibits M and TREK currents (I TREK ) in mSCG, the activation of their bradykinin counterparts could do the same. Indeed, both types of channels, but also both types of receptors, have been thoroughly studied due to their important role on the setting of the resting membrane potential (RMP) and on the modulation of membrane excitability in many cell types.
The study of TREK currents in native neurons is challenging because the activity of these channels is reduced at room temperature and atmospheric pressure [28][29][30]. We take advantage of the neuroprotective agent riluzole, which activates the three members of the TREK subfamily (TREK-1, TREK-2 and TRAAK (TWIK-related arachidonic acid-stimulated potassium channel)) [31][32][33][34]. Riluzole does not activate other K2P channels and it was demonstrated that the outward current evoked by riluzole in mSCG neurons is mainly driven through TREK-2 channels [25,26,35]. TREK-1 and TREK-2, but not TRAAK, are well known to be modulated through G-protein coupled receptors and PLC activation [36,37]. In fact, it has been proposed that basal G protein activity may have a constant down regulating effect on these channels [38,39].
We have reported before that, in mSCG neurons, the muscarinic inhibition of TREK-2 currents following the PLC pathway requires a reduction of phosphatidylinositol 4,5-biphosphate (PIP 2 ) levels [25]. In the current study we demonstrate that, much like it was reported in rSCG neurons, BK depolarizes and increases the excitability of mSCG neurons. A good amount of evidence is given to show that BK inhibits a riluzole activated TREK-like current by reducing the membrane PIP 2 levels, via the same pathway than muscarinic agonists in the same preparation. The presence of this newly-described current could explain, at least in part, the effects of BK in SCG neurons. The main achievement of this study was to discover a new target for the control of sympathetic excitability by bradykinin. TREK channels are known to regulate the resting membrane potential in a good number of cellular types [34].

Results
Bradykinin has been shown to strongly modulate neuronal excitability by reducing M-currents in several preparations [40] including rat SCG neurons but, to our knowledge, only one study [41] has indirectly related bradykinin with TREK channels. Voltage-clamp whole-cell experiments were performed at room temperature (22-24 • C) and holding the neurons at −30 mV. A blocking "cocktail" containing tetraethylammonium (TEA, 15 mM), tetrodotoxin (TTX, 0.5 µM), and CsCl (1 mM) was used to block M-currents, voltage-dependent Na + currents, and the hyperpolarization activated cationic h-current (I h ), respectively [42,43]. This blocking cocktail reduced the basal outward current found at −30 mV (I -30 ) by about 40% [25] and does not affect the riluzole-activated outward current as previously demonstrated [26].

Bradykinin Increases the Excitability of mSCG Neurons
The resting membrane potential of mSCG neurons recorded in culture was −62.6 ± 1.3 mV (n = 33), and in order to reduce variability, we manually clamped neurons at −60 mV before applying any treatment or protocol. For the same reason, only neurons firing less than 10 action potentials in response to maximal 1 s depolarizing pulses (adapting cells), were analyzed in this study [44].
In those conditions, bath application of bradykinin 250 nM depolarized mSCG neurons by 6.3 ± 0.7 mV (n = 18, p < 0.001, Figure 1(a1)) and in three of them, BK induced cell firing (not shown). As expected, application of the recently discovered activator of TREK channels BL1249 (BL, (5,6,7,8- [45][46][47] had an opposite effect on the resting membrane potential hyperpolarizing mSCG neurons by −5.7 ± 0.9 mV (n = 7, p < 0.01) and −17.6 ± 1.5 mV (n = 8, p < 0.001, Figure 1(a2)), when applied at 3 µM and 10 µM respectively. When BL (3 and 10 µM) was applied in the presence of bradykinin 250 nM, it produced a similar and significant hyperpolarization: −8.4 ± 0.7 mV (n = 10, p < 0.001) and −18.7 ± 1.5 mV (n = 8, p < 0.001), respectively (not shown). Interestingly, when applied in the presence of BL (3 and 10 µM), the depolarization produced by BK 250 nM was not statistically different between both groups (6.5 ± 1.6 and 4.5 ± 0.8 mV, p > 0.05, n = 7 and 8 respectively) and they were also not different from the control (only BK, p > 0.05). BL1249 has been shown to activate TREK-1 and TREK-2 channels but no other K2P channels [45]. mSCG neurons are well known for their strong SFA in response to injections of depolarizing current (Figure 1(b1)) in standard solutions. Bradykinin (250 nM) provoked a small but significant increment (p < 0.05, n = 18) of the number of action potentials (Figure 1(b2)) in response to depolarizing current injections from 25 to 175 pA ( Figure 1c). In the presence of BK, BL1249 reduced the firing at both 3 and 10 µM (n = 10 and 8 respectively; p < 0.05; Figure 1(b3,b4),c). The effect of BL 10 µM was so dramatic that neurons were unable to respond at all (Figure 1(b4),c). Also when BL 10 µM was applied first, mSCG neurons stopped firing at any current injection (n = 8, p < 0.05) and subsequent application of BK (in the presence of BL) did not increase the excitability (not shown). The effect of BL 3 µM on the firing was not significant but interestingly it precluded the increase of firing normally produced by BK 250 nM (n = 5; p > 0.05).
In order to investigate the effect of BK on the action potentials of mSCG neurons, we constructed phase plots calculating the derivative of voltage with respect to time and plotting it against membrane voltage. Figure 1d shows that the action potential threshold (1), maximal up-stroke velocity (2), maximal positive voltage reached (3), and hump decrease in velocity (4) were indistinguishable before and after application of BK 250 nM. In fact, the amplitude of the action potential (121.75 ± 3.01 vs. 120.72 ± 2.78 mV, n = 17, p = 0.1), the half-amplitude duration (2.37 ± 0.15 vs. 2.24 ± 0.12 ms, n = 17, p = 0.16), the threshold (−35.07 ± 1.16 vs. −35.14 ± 1.12 mV, n = 17, p = 0.88) and the latency to the first action potential in response to depolarizing current injections at 50 pA (28.53 ± 8.94 vs. 20.61 ± 1.98 ms, n = 17, p = 0.39) were all not statistically different. respectively (not shown). Interestingly, when applied in the presence of BL (3 and 10 µM), the depolarization produced by BK 250 nM was not statistically different between both groups (6.5 ± 1.6 and 4.5 ± 0.8 mV, p > 0.05, n = 7 and 8 respectively) and they were also not different from the control (only BK, p > 0.05). BL1249 has been shown to activate TREK-1 and TREK-2 channels but no other K2P channels [45].

Bradykinin Inhibits Whole-Cell and Single-Channel TREK-2 Currents
To obtain the control I RIL (200.2 ± 39.3 pA; n = 7; Figure 2a) we applied riluzole (300 µM for at least 3 min) to neurons clamped at −30 mV in the whole-cell perforated-patch mode and in the presence of the blocking cocktail. After washing out riluzole we applied BK (250 nM) during 4 min before the second application of riluzole, which evoked an I RIL significantly smaller than I RIL control (129.8 ± 32.2 pA, n = 7; Figure 2b), showing a significant reduction of 43.2 ± 6.1% (n = 7; p < 0.01; Figure 2c). Additionally, BK reduced the I -30 in 83.8 ± 24.1 pA (n = 7; Figure 2b). We have demonstrated before that repetitive application of riluzole does not desensitize I RIL [25].
Single TREK-2 channels were identified by using voltage ramps from −100 to +100 mV in SCG neurons. Cell-attached recordings showed a high conductance at negative potentials, which was clearly reduced at positive voltages, which is characteristic of TREK-2 channels (Figure 3a). Holding the patch (cell-attached) at −60 mV we applied BK (750 nM) in order to investigate whether BK can affect TREK-2 channels using an indirect pathway through second messengers ( Figure 3b). The amplitude, open dwell time (duration of channel openings) and open probability (NP o ) were measured from one minute recordings. These recordings were taken before BK application (control) and at least 3 min after BK application (BK), when the effect of the drug was stabilized. Figure 3e shows that BK induced a reduction in the NP o (from 7,72E-4 ± 1,56E-4 to 4,17E-4 ± 1,63E-4, p < 0.05, n = 11) without affecting the current amplitude (from 7.9 ± 0.3 pA to 7.9 ± 0.3 pA, p = 0.979, n = 11, Figure 3c) nor the open dwell time (from 0.88 ± 0.14 ms to 0.71 ± 0.12 ms, p = 0.290, n = 11, Figure 3d), indicating that BK reduced I RIL by reducing the open probability of the TREK-2 channels. These experiments strongly indicated that BK can inhibit TREK-2 channels and hence I RIL through the activation of a second messenger cascade.

The Inhibition of I RIL by BK Is Mediated through B 2 Receptors
The binding of BK to specific receptors, mainly B 1 R and B 2 R, provokes a signalling cascade starting with G-protein activation. In rat sympathetic neurons, G q/11 is the main G-protein implicated in BK modulation [12]. In turn, G q/11 activates the enzyme PLC which hydrolyzes PIP 2 to give IP 3 and diacylglycerol (DAG). IP 3 binds to the IP 3 endoplasmic reticulum receptors releasing Ca 2+ and therefore increasing its intracellular concentration [48]. Finally, the Ca 2+ increase, together with DAG, activates the PKC, which can phosphorylate other proteins [49]. All intermediate products of this cascade have been shown to directly modulate ion channels in different preparations.
To investigate the participation of BK receptors in the inhibition of I RIL , we used the selective B 2 R antagonist HOE-140 (d-Arg-[Hyp 3 , Thi 5 , d-Tic 7 , Oic 8 ]BK). In a first step we applied riluzole (300 µM for 3 min in the presence of cocktail + HOE) to obtain I RIL in control conditions (158.3 ± 44.9 pA; n = 4; Figure 4a). Interestingly, the application of HOE (300 nM) to the bath solution slightly increased the steady-state current at −30 mV in 12.2 ± 1.2 pA (n = 4; not shown). Bradykinin (250 nM) was then applied in the presence of cocktail + HOE, it is relevant that in the presence of the B 2 receptor antagonist, BK did not reduce the outward current at −30 mV (See panels b in Figures 5-7). Finally, a second application of riluzole (300 µM, 3 min) in cocktail + HOE + BK (Figure 4b), induced an outward I RIL showing no significant difference with the I RIL control (142.7 ± 36.8 pA; n = 4; p = 0.2804; Figure 4c). The percentage of inhibition of I RIL by BK in presence of HOE was insignificant (5.2 ± 8.3%; Figure 4d), clearly indicating that the inhibition of I RIL by BK can be entirely explained by the activation of B 2 receptors. activates the PKC, which can phosphorylate other proteins [49]. All intermediate products of this cascade have been shown to directly modulate ion channels in different preparations.
To investigate the participation of BK receptors in the inhibition of IRIL, we used the selective B2R antagonist HOE-140 (D-Arg-[Hyp 3 , Thi 5 , D-Tic 7 , Oic 8 ]BK). In a first step we applied riluzole (300 µM for 3 min in the presence of cocktail + HOE) to obtain IRIL in control conditions (158.3 ± 44.9 pA; n = 4; Figure 4a). Interestingly, the application of HOE (300 nM) to the bath solution slightly increased the steady-state current at −30 mV in 12.2 ± 1.2 pA (n = 4; not shown). Bradykinin (250 nM) was then applied in the presence of cocktail + HOE, it is relevant that in the presence of the B2 receptor antagonist, BK did not reduce the outward current at −30 mV (See panels b in Figures 5-7). Finally, a second application of riluzole (300 µM, 3 min) in cocktail + HOE + BK (Figure 4b), induced an outward IRIL showing no significant difference with the IRIL control (142.7 ± 36.8 pA; n = 4; p = 0.2804; Figure 4c). The percentage of inhibition of IRIL by BK in presence of HOE was insignificant (5.2 ± 8.3%; Figure  4d), clearly indicating that the inhibition of IRIL by BK can be entirely explained by the activation of B2 receptors.

Manipulating PIP2 Concentration Affects the Inhibition of IRIL by Bradykinin
We have recently reported that the inhibition of TREK-2 channels by muscarinic agonists in mSCG neurons greatly depends on PIP2 depletion [25]. As B2 receptors also activate Gq/11 and PLC, the reduction of PIP2 seemed a good candidate to explain the inhibition of IRIL by BK. To investigate this issue, we incubated mSCG neurons with diC8PIP2 (0.5 µM, saturating concentration) and a histone carrier, for 1 h, before starting current recordings. As expected, in this situation, the application of BK did not affect IRIL significantly (From 161.3 ± 21.4 pA to 135.5 ± 9.3 pA, p = 0.20, n = 5; Figure 5a,b). Consistently, the percentage of current reduction (10.5 ± 10.6%) was irrelevant when compared with the control (43.2 ± 6.1%; without diC8PIP2; Figure 5c  these results demonstrate that the inhibition of the riluzole activated current by bradykinin is mediated by the reduction of PIP2 but the inhibition of I-30 is not.

Neither Protein Kinase C nor Ca 2+ Are Involved in the Inhibition of IRIL by Bradykinin
In the PLC pathway, concomitant with the reduction of PIP2 there is an increase of IP3 and DAG. DAG and the Ca 2+ released by the binding of IP3 to its receptor finally activate PKC. Indeed, research has shown that the inhibition of the M-current by BK depends on the increase of Ca 2+ produced in this way [19,50]. We investigated the role of these last messengers, on the inhibition of TREK-2 currents by BK, by inhibiting PKC and antagonizing IP3 receptors.
In order to inhibit PKC we applied bisindolylmaleimide (300 nM) for 10 min. In these conditions, bradykinin inhibited IRIL by 30.4 ± 2.7%, from 194.1 ± 43.5 to 135.7 ± 32.6 pA (p = 0.023; n = 4; Figure  6a-c). This inhibition is not different (p = 0.17) from that obtained in the absence of bisindolylmaleimide (43.2 ± 6.1%; Figure 6d). The application of bisindolylmaleimide had no effect on I-30 (7.4 ± 4.3 pA; n = 4; p = 0.18; Figure 6b) but BK application reduced significantly this current (−81 ± 22.1 pA; p = 0.035; n = 4; Figure 6b). This experiment indicated that PKC is not involved in the inhibition of IRIL or I-30 by BK.  To test the other limb of the PLC pathway we antagonized the IP3R using 2-APB (100 µM) and hence hampered the increase in Ca 2+ that is normally produced by the activation of BK receptors. Bradykinin still reduced IRIL when the increase in Ca 2+ was prevented (from 158.4 ± 30.7 pA to 94.6 ± 17.5 pA; p = 0.029; n = 5; Figure 7). The inhibition (39 ± 5.4%; Figure 7d) was similar to that found in the absence of 2-APB (43.2 ± 6.1%; p = 0.79; Figure 7d). We previously performed control experiments comparing riluzole activation in the presence and absence of 2-APB and differences were not found

Discussion
In addition to being a potent vasodilator and a pain and inflammation activator [1], bradykinin depolarizes the resting membrane potential and reduces the spike frequency adaptation in rSCG neurons [12,13]. This modulation has been related to the M-current that is well known to be inhibited by BK [12,50,52]. On the other hand, the potassium channels of the TREK subfamily have a similar role to the M channels in some excitable cells, contributing to the control of the RMP, the firing and therefore to the control of excitability [26,29]. In the present work we demonstrated, for the first time, that bradykinin also modulates TREK currents in mSCG neurons by reducing PIP2. The inhibition of these channels could contribute to the increase in excitability produced by BK in sympathetic neurons.

I-30 Includes ITREK and IM
When the membrane of SCG neurons is depolarized to −30 mV, a characteristic and constant outward current emerges, this current was initially attributed to the opening of voltage-activated, non-inactivating potassium M-channels and it explains why most voltage protocols used to study the M-current start by fixing neurons at −30 mV [53]. Our group has demonstrated the expression of background potassium TREK-2 currents in mSCG neurons but although these are considered essentially voltage-independent, its macroscopic intensity-voltage curve is almost indistinguishable from that of the M-current [25,35]. The best tool we have to isolate the TREK current from the Mcurrent is to selectively activate TREK currents with riluzole in the presence of TEA, a strong blocker of the M-currents but not affecting TREK channels [26]. We have previously demonstrated that the current activated by riluzole is not affected by blockers of other channels like voltage-gated and persistent Na + , voltage-gated Ca 2+ , hyperpolarization-activated cation, transient-receptor-potential cation, voltage-gated A and DR type K + , M type K + , and Ca 2+ -activated K + (Big K + and Small K + ) channels [26]. As riluzole activates the TREK subfamily exclusively [26,31,32], and, within the TREK

Manipulating PIP 2 Concentration Affects the Inhibition of I RIL by Bradykinin
We have recently reported that the inhibition of TREK-2 channels by muscarinic agonists in mSCG neurons greatly depends on PIP 2 depletion [25]. As B 2 receptors also activate G q/11 and PLC, the reduction of PIP 2 seemed a good candidate to explain the inhibition of I RIL by BK. To investigate this issue, we incubated mSCG neurons with diC 8 PIP 2 (0.5 µM, saturating concentration) and a histone carrier, for 1 h, before starting current recordings. As expected, in this situation, the application of BK did not affect I RIL significantly (From 161.3 ± 21.4 pA to 135.5 ± 9.3 pA, p = 0.20, n = 5; Figure 5a,b). Consistently, the percentage of current reduction (10.5 ± 10.6%) was irrelevant when compared with the control (43.2 ± 6.1%; without diC 8 PIP 2 ; Figure 5c,d). Nevertheless, BK inhibited I -30 significantly (−113.4 ± 22 pA; p = 0.0067; n = 5; Figure 5b) in presence of diC 8 PIP 2 indicating distinct pathways involved in the inhibition of I RIL and I -30 (experiments carried out in cocktail). Altogether these results demonstrate that the inhibition of the riluzole activated current by bradykinin is mediated by the reduction of PIP 2 but the inhibition of I -30 is not.

Neither Protein Kinase C nor Ca 2+ Are Involved in the Inhibition of I RIL by Bradykinin
In the PLC pathway, concomitant with the reduction of PIP 2 there is an increase of IP 3 and DAG. DAG and the Ca 2+ released by the binding of IP 3 to its receptor finally activate PKC. Indeed, research has shown that the inhibition of the M-current by BK depends on the increase of Ca 2+ produced in this way [19,50]. We investigated the role of these last messengers, on the inhibition of TREK-2 currents by BK, by inhibiting PKC and antagonizing IP 3 receptors.
To test the other limb of the PLC pathway we antagonized the IP 3 R using 2-APB (100 µM) and hence hampered the increase in Ca 2+ that is normally produced by the activation of BK receptors. Bradykinin still reduced I RIL when the increase in Ca 2+ was prevented (from 158.4 ± 30.7 pA to 94.6 ± 17.5 pA; p = 0.029; n = 5; Figure 7). The inhibition (39 ± 5.4%; Figure 7d) was similar to that found in the absence of 2-APB (43.2 ± 6.1%; p = 0.79; Figure 7d). We previously performed control experiments comparing riluzole activation in the presence and absence of 2-APB and differences were not found [26]. Interestingly, the application of 2-APB produced a fast increase of I -30 (71.4 ± 19.3 pA; p = 0.020; Figure 7b), this was a transient effect that slowly decreased in −29.9 ± 9.7 pA (p = 0.037). This was in agreement with 2-APB acting as a TREK-2 activator [51]. The application of BK (250 nM) still reduced I -30 in the presence of 2-APB (−48.3 ± 1.5 pA; p = 5.18 × 10 −6 ; n = 5; Figure 7b), probably reflecting the inhibition of the TREK-2 current, previously activated by 2-APB.

Discussion
In addition to being a potent vasodilator and a pain and inflammation activator [1], bradykinin depolarizes the resting membrane potential and reduces the spike frequency adaptation in rSCG neurons [12,13]. This modulation has been related to the M-current that is well known to be inhibited by BK [12,50,52]. On the other hand, the potassium channels of the TREK subfamily have a similar role to the M channels in some excitable cells, contributing to the control of the RMP, the firing and therefore to the control of excitability [26,29]. In the present work we demonstrated, for the first time, that bradykinin also modulates TREK currents in mSCG neurons by reducing PIP 2 . The inhibition of these channels could contribute to the increase in excitability produced by BK in sympathetic neurons.

I -30 Includes I TREK and I M
When the membrane of SCG neurons is depolarized to −30 mV, a characteristic and constant outward current emerges, this current was initially attributed to the opening of voltage-activated, non-inactivating potassium M-channels and it explains why most voltage protocols used to study the M-current start by fixing neurons at −30 mV [53]. Our group has demonstrated the expression of background potassium TREK-2 currents in mSCG neurons but although these are considered essentially voltage-independent, its macroscopic intensity-voltage curve is almost indistinguishable from that of the M-current [25,35]. The best tool we have to isolate the TREK current from the M-current is to selectively activate TREK currents with riluzole in the presence of TEA, a strong blocker of the M-currents but not affecting TREK channels [26]. We have previously demonstrated that the current activated by riluzole is not affected by blockers of other channels like voltage-gated and persistent Na + , voltage-gated Ca 2+ , hyperpolarization-activated cation, transient-receptor-potential cation, voltage-gated A and DR type K + , M type K + , and Ca 2+ -activated K + (Big K + and Small K + ) channels [26]. As riluzole activates the TREK subfamily exclusively [26,31,32], and, within the TREK subfamily, TREK-2 is largely more expressed than the other TREK subfamily channels in mSCG neurons [27], we can propose that I RIL is mainly transported through TREK-2 channels in these neurons. Assuming that the outward current observed at −30 mV is composed mainly of I M and I TREK , when TEA and other blockers not affecting TREK currents are applied, we expect the remaining current to be mainly I TREK . Even so, we call this current I -30 because we suspect that the cocktail may not be capable of removing the I M completely.

I RIL is Due to the Activation of I TREK
As BK inhibits both the M and TREK currents, the question arises of whether I RIL can be contaminated with the M-current. BK applied in the cocktail reduces I -30 and inhibits I RIL , and both effects disappear when the BK receptors are antagonized, indicating that both depend on BK receptor activation. This suggests that part of the I -30 is due to I TREK as most, if not all, I M is blocked by TEA [43].
On the contrary, when we keep high levels of PIP 2 , BK is not inhibiting I RIL but still reduces I -30 , suggesting that BK inhibits I RIL by depleting PIP 2 but it is able to reduce part of the I -30 (probably I M ) through a different pathway when PIP 2 reduction is not possible. Taken together all these data indicate that I RIL is essentially a riluzole-activated TREK current and that I -30 has two components (I TREK and a residual I M when TEA is present). This hypothesis is strongly supported by the fact that I M has been reported to be inhibited by BK through the activation of Ca 2+ -calmodulin, but not by PIP 2 depletion (like reported for muscarinic agonists), in SCG neurons [12,52,54,55]. This different mechanism could be due to the fact that muscarinic receptors are far away from IP 3 R and hence Ca 2+ release from the ER becomes difficult. However, BK receptors are close enough to provoke Ca 2+ release [52,56]. Keeping a high level of PIP 2 does not impede IP 3 production and hence Ca 2+ release, when BK stimulates its receptors [19].
Moreover, as we reported before, when using the PI3/PI4-Kinase-inhibitor wortmannin to block the replenishment of PIP 2 and to maintain PIP 2 concentration very low it can be demonstrated that I RIL , and also I -30 , need high concentrations of PIP 2 to be activated. This is consistent with both M and TREK channels being inhibited by PIP 2 depletion in SCG neurons [25,57]. Our results contrast with previous data showing that BK does not increase the concentration of PIP 2 in rat SCG neurons [14,58,59]. This discrepancy could be related to the use of different species and ages for the extraction of neurons. In fact, it has also been shown that the regulation of the M current by muscarinic agonists is quite different in rat (G q ) and mouse (G 11 , G q and PTX-sensitive G-proteins) and not less important, the inhibition of the M current by BK is mediated by G α11 in mSCG but by G αq/11 in rSCG neurons [18].

I -30 versus I TREK
When the activation of IP 3 receptors was avoided with 2-APB and hence prevented the inhibition of I M , BK still inhibited I -30 , indicating that besides the M component there exists a second component that can be inhibited through a pathway different from the Ca 2+ -calmodulin one, probably the TREK current activated by 2-APB contributes to this inhibition. Consistently, the reduction of the I -30 in these conditions resulted smaller than that obtained in cocktail alone (around 50 pA) as it represents the I TREK only while the inhibition of I RIL by BK was well preserved in those conditions.

Physiological Relevance and Conclusions
The TREK subfamily has been implicated in the transduction of mechanical, chemical, and thermal stimuli [60,61], but also related to pain transduction, as it is well expressed in dorsal root ganglia (DRG) small neurons and in trigeminal ganglia neurons [62][63][64][65]. Recent studies have proposed the TREK subfamily as a possible target for pain treatment [62,66,67] and TREK-1 seems to be involved in morphine analgesia without adverse effects in mice [68]. Also aristolochic acid, used traditionally as painkiller, enhanced TREK-1 and TREK-2 currents [69]. The present work shows that TREK-2 channels are inhibited by BK, an inductor of pain and inflammation, much like as described before for M-current inhibition [12]. The activation of TREK channels by the neuroprotector riluzole hyperpolarizes nociceptive DRG neurons resulting in a decrease of the effects of BK [70]. The TREK subfamily is well expressed in the DRG neurons [28,71,72], the TREK-2 channel being the most expressed, at least in neonatal rats [73]. Additionally, the blood vessel vasodilation produced by bradykinin is strongly attenuated in TREK-1 deleted animals [41]. These and other studies strongly relate BK with TREK channels and pain.
As summarized in Figure 8, the activation of muscarinic M 1 -receptors inhibits both I M and I TREK through the depletion of PIP 2 , while the activation of B 2 -receptors uses two different inhibitory pathways, PIP 2 reduction for I TREK and Ca 2+ -calmodulin for I M . Importantly, it has been reported that TREK-1 channels have a group of cationic aminoacids in the C-terminal region that interacts electrostatically with the inner part of the membrane. A reduction in the concentration of PIP2 causes fluctuations in this electrostatic interaction resulting in the dissociation of the c-terminal domain from the membrane and a change in channel activity [74,75]. An important conclusion of this work is that the riluzole activated current and about half of the current at −30 (in the presence of Na + , K + and h-current blockers) are carried through TREK channels. Both currents and also individual TREK-2 channels are inhibited by bradykinin through PIP 2 depletion. These results also suggest that previous interpretations on the effects of BK on the nervous system may need a review. electrostatically with the inner part of the membrane. A reduction in the concentration of PIP2 causes fluctuations in this electrostatic interaction resulting in the dissociation of the c-terminal domain from the membrane and a change in channel activity [74,75]. An important conclusion of this work is that the riluzole activated current and about half of the current at −30 (in the presence of Na + , K + and hcurrent blockers) are carried through TREK channels. Both currents and also individual TREK-2 channels are inhibited by bradykinin through PIP2 depletion. These results also suggest that previous interpretations on the effects of BK on the nervous system may need a review.

Materials and Methods
All protocols performed were approved by the Spanish Research Council and the Animal Welfare Committee of the University of Vigo (Code: 07/2014; Date: 24/10/2016). They accomplished the Spanish (RD 53/2013) and European (2010/63/EU) directives for the protection of animals used for experimental purposes.

Cell Culture
Swiss-CD1 mice of both sexes and 20 to 60-days-old were terminally anaesthetized with CO2 before decapitation. SCG ganglia were then quickly dissected and placed in L-15 culture medium to

Materials and Methods
All protocols performed were approved by the Spanish Research Council and the Animal Welfare

Cell Culture
Swiss-CD1 mice of both sexes and 20 to 60-days-old were terminally anaesthetized with CO 2 before decapitation. SCG ganglia were then quickly dissected and placed in L-15 culture medium to clean and disaggregate, first enzymatically and then mechanically. Then, neurons were placed in laminin-coated 35 mm Petri dishes and kept in culture (37 • C, 5% CO 2 ) for 12 to 24 h before recording. Detailed dissection and culture procedures have been reported before [22,76].

Single-Channel Recording
Single-channel recordings were performed in the cell-attached configuration. Bath and pipette solutions were identical and contained (mM): 150 KCl, 1 MgCl 2 , 5 EGTA, and 10 HEPES, pH 7.2 was achieved with KOH. The electrode resistance for single-channel experiments ranged from 10 to 12 MΩ. Data were sampled at 20 KHz and low-pass filtered at 2 KHz using the amplifier built-in filter. Single-channel openings faster than 50 µs were discarded. The threshold detection for single-channel openings was set at 50% of the amplitude. Single-channel parameters: amplitude, open dwell time and NP o were measured using pClamp 10.0 software. The NP o was calculated according to the following equation: NP o = t o /T, where N is the number of channels, P o is the open state probability, t o was the total time that the channel was found in the open state, and T is the total observation time. The data obtained (mean ± SEM) were analyzed and plotted using Clampfit 10 and Origin 9.0 applying Paired Student's t-test and considering significant p-values < 0.05.

Drugs
All drugs were applied directly to the bath solution (10 mL/min) during the protocol, except for diC 8 PIP 2 with a histone carrier, which was previously added to the culture dishes.
TTX and HOE-140 were purchased from Tocris Bioscience (Bristol, UK), diC 8 PIP 2 and the histone carrier were purchased from Echelon Biosciences (Salt Lake City, UT, USA), and all the other chemicals were obtained from Sigma-Aldrich (Madrid, Spain). Funding: This research was funded by the Spanish government (grant references: CONSOLIDER CSD2008-00005, BFU2014-58999-P), the regional government "Xunta de Galicia" (grant numbers INBIOMED CN2012/273, INB1-131H-2) and it was partially funded by FEDER. All the funding was awarded to J Antonio Lamas.

Conflicts of Interest:
The authors declare no conflict of interest.