ERG K+ channels mediate a major component of action potential repolarization in lymphatic muscle

Smooth muscle cells in the walls of collecting lymphatic vessels fire spontaneous action potentials (APs), which conduct rapidly over the muscle layer to initiate contractions that propel lymph. Several ion channels have been implicated in the currents underlying the AP spike and the preceding diastolic depolarization, but the molecular identities of K+ channels involved in AP repolarization are unknown. Based on previous studies of other rhythmically active smooth muscles, we hypothesized that ether-a-go-go related gene (ERG) K+ channels (Kv11) play an important role in repolarization of the AP in lymphatic muscle. Message for one or more ERG channel isoforms was detected by RT-PCR analysis of lymphatic vessels from mice, rats and humans. Membrane potential recordings in smooth muscle cells of rat and human lymphatics revealed that nanomolar concentrations of ERG-1 inhibitors (E-4031 and BeKm-1) prolonged the duration of the AP plateau (normally ~ 1 s in duration) and induced multiple spikes, whereas ERG-1 activators (ICA-105574 and RPR-260243) shortened the plateau and could completely inhibit spontaneous APs. At relatively high inhibitor concentrations, the AP plateau duration lasted as long as 24 s. ERG activators reversed the effects of ERG inhibitors and vice-versa. In pressure myograph studies, ERG channel inhibition prolonged the diastolic repolarization phase of the contraction cycle and reduced the frequency of spontaneous contractions. This is the first evidence for a specific K+ channel contributing to the AP in lymphatic muscle. Our results imply that lymphatic contractile dysfunction may occur in long QT type II patients with mutations that result in ERG channel loss-of-function or impaired trafficking of the channel to the cell membrane.

ERG inhibitors delay LMC repolarization and prolong lymphatic diastole.Next, we investigated whether APs in LMCs were sensitive to the widely used ERG1 channel inhibitor, E-4031 27,31,33 .Of the three species-rat, mouse and human-mouse LMCs were the most difficult from which to make successful Vm measurements.Difficulties with obtaining successful recordings from human lymphatics limited experimentation in that species to a few Vm recordings (see Suppl.Methods).Thus, the majority Vm measurements were recorded from LMCs of pressurized rat mesenteric lymphatic collectors.A representative Vm recording from the LMC layer of a rat mesenteric lymphatic vessel is shown in Fig. 2.After impalement, there was a rapid drop in Vm to ~ − 46 mV, followed by a short period of stabilization.Insert (a) shows the shape of a typical AP in a rat LMC, with a resting Vm ~ -42 mV, gradual depolarization in diastole to a threshold for AP initiation of ~ − 40 mV, a single spike to ~ 0 mV, followed by a plateau at ~ -15 mV lasting 1.0 s, ending with rapid repolarization and after-hyperpolarization to ~ -50 mV.Subsequent application (b) of the ERG1 inhibitor E-4031 (10 nM final bath concentration) led to a slight widening of the AP plateau (to 1.4 s). 100 nM E-4031 caused a slight depolarization (~ 2 mV) and further widening of the plateau (to 3.4 s) along with a blunted second spike (c).300 nM E-4031 greatly extended the plateau duration (to 7.0 s), which contained multiple blunted spikes (d).These effects were even more exaggerated in some APs (e).Higher concentrations of E-4031 sometimes caused a sustained depolarization that abolished APs altogether (not shown).
Summary data for the effects of E-4031 on APs of rat mesenteric LMCs are shown in Fig. 3. Increasing concentrations of E-4031 caused progressive widening of the AP plateau, measured as the length of time from the spike to the point during repolarization when Vm crossed the original resting Vm; the IC 50 was 240 nM and concentrations above 100 nM produced plateau durations that were statistically increased from control (Fig. 3A).
Likewise, E-4031 led to an increase in the number of spikes per AP (Fig. 3B), with IC 50 = 250 nM; concentrations above 300 nM were statistically different from control.When high concentrations of E-4031 were applied to naïve LMCs, larger secondary spikes were often observed (Suppl.Fig. 1).No significant changes in AP plateau duration were observed during a separate series of experiments testing for time and/or vehicle effects (Suppl.Fig. 2).Increasing concentrations of E-4031 produced progressive depolarization (Fig. 3C), reaching ~ 6 mV at 1 μM E-4031, but these changes were not statistically significant.Increasing concentrations of E-4031 caused reductions in the frequency of AP firing (Fig. 3D).The magnitude of the after-hyperpolarization was also reduced by E-4031 in some cells (see Fig. 2 for an example), but this effect was not statistically significant.
We also tested the effects of the peptide ERG channel inhibitor, BeKm-1, which works through a different mechanism of action than E-4031.While E-4031 works an open channel blocker, possibly binding to the S6 domain 37 , BeKm-1 is an ERG-specific toxin that binds near the pore to stabilize the close state 38 .BeKm-1 also increased the duration of the AP plateau (Fig. 3E) and increased the number of spikes per AP (Fig. 3F), albeit at higher concentrations than E-4031 (IC 50 = 0.92 and 1.3 μM, respectively).BeKm-1 did not produce significant effects on resting Vm (Fig. 3G) or the magnitude of the after-hyperpolarization (not shown).Increasing concentrations of BeKm-1 caused reductions in the frequency of AP firing (Fig. 3H), although the differences were not statistically significant.
In parallel protocols using pressurized rat mesenteric lymphatic vessels in which diameter, rather than Vm, was measured (in the absence of wortmannin), E-4031 significantly altered several components of diastole.An example is shown in Fig. 4A.At this compressed time scale, the most obvious effect of E-4031 was to decrease the contraction frequency by increasing the time between contractions, but analyses of single, twitch contractions at higher time resolution (Fig. 4B) revealed substantial increases in the durations of the twitch contractions, primarily due to increases in the time required for relaxation.Systole was also prolonged, contributing to the increase in contraction duration and leading to an increase in contraction amplitude.In some cases, there were two components to the twitch contraction (arrowheads in Fig. 4A and expanded traces in Fig. 4B; see also the diameter trace in Suppl.Fig. 1A), events which were extremely rare under normal conditions.Summary data in Fig. 4C-F showed that ERG1 inhibition significantly increased the time to half maximal relaxation (Fig. 4C) and the area under the diameter-time curve (Fig. 4D); EC 50 = 150 and 200 nM, respectively, as estimated from curve fits in IGOR.E-4031 slightly increased the contraction amplitude (Fig. 4E) and significantly decreased contraction frequency (Fig. 4F).The reduction in frequency occurred both as a result of an increased time to www.nature.com/scientificreports/half maximal relaxation (the early phase of diastole) and an increase in the time between twitch contractions (i.e., the later phase of diastole).
ERG activators shorten the duration of the AP plateau phase.Next, we tested the effects of the ERG channel activator, ICA-105574 27 , which attenuates ERG inactivation by shifting its voltage dependence to more positive potentials 39 .Increasing concentrations of ICA-105574 resulted in progressive narrowing of the AP plateau.An example recording is shown in Fig. 5 in which the ultimate effect of 3 μM ICA-105574 was to decrease the width of the AP plateau to ~ 30% of control prior to complete cessation of spontaneous APs.In some cells, ICA-105574 also appeared to increase the magnitude of the after-hyperpolarization (see example in Fig. 5Bd), but analysis of the summary data indicated that this effect was not statistically significant.The effects of ICA-105574 and another ERG channel activator RPR-260243, which works by slowing the kinetics of ERG channel deactivation 40 , are summarized in Fig. 6.Both activators produced a decrease in the duration of the AP plateau (Fig. 6A,E; EC 50 = ~ 800 nM and 2 μM, respectively).No significant changes in AP plateau duration were observed during a separate series of experiments testing for time and/or vehicle effects (Suppl.Fig. 3).Neither activator produced a statistically significant increase in the resting Vm (Fig. 6C,G), or frequency (Fig. 6D,H), although in both cases there was a trend for frequency to decrease with increasing activator concentration.

ERG activators and inhibitors have counteracting effects on the lymphatic AP.
We also tested whether ERG channel activators and inhibitors would antagonize each other in their effects on the lymphatic muscle AP.Representative recordings of this protocol are shown in Figs.7 and 8.As before, the application of E-4031 led to an increase in the duration of the AP plateau (Fig. 7A), which became particularly evident at 100 nM E-4031 (Fig. 7Bc).Subsequent addition of ICA-105574 (3 μM, but not 1 μM) in the continued presence of E-4031, reversed the effects of the inhibitor, shortening the duration of the plateau to the extent that it was even less than control (Fig. 7Be; 0.73 s for control, 0.45 s in 100 nM E-4031 + 3 μM ICA-105574).Summary data for AP plateau duration and normalized AP plateau duration are given in Fig. 7C,D.An example of the reverse protocol is shown in Fig. 8. Here, ICA-105574 was applied in increasing concentrations, again leading to progressive narrowing of the AP plateau at 1, 3 and 6 μM (Fig. 8Bc-d), with plateau width at 6 μM being reduced to 44% of control.Subsequent application of E-4031, in the continued presence of ICA-105574, led to an increase in the frequency of spontaneous APs and a plateau duration that was 27% greater than control.Summary data for AP plateau duration and normalized AP plateau duration are given in Fig. 8C,D.
ERG inhibitors prolong AP duration and diastole in human lymphatic muscle.Lastly, we obtained two recordings of Vm in wire myograph-mounted human lymphatic vessels in response to multiple concentrations of E-4031 that showed the same pattern of response as in rat mesenteric LMCs, i.e., prolongation of the AP plateau with the appearance of multiple spikes.One of those recordings is shown in Suppl.Fig. 4,  where the AP plateau duration increased from 1.7 s (control) to 5.7 s, with multiple spikes, after application of 1.2 μM E-4031.In addition, the effects of E-4031 on the diameter of a pressurized human mesenteric vessel were similar to those observed in rat mesenteric vessels, as confirmed in the example shown in Suppl.Fig. 5. Increasing concentrations of E-4031 prolonged the duration of the contraction phase and were associated with an increase in the number of "double" contractions (arrowheads, Suppl.Fig. 5A), in which second contractions were often initiated before the vessel had relaxed to the previous end diastolic diameter (Suppl.Fig. 5B).Increasing concentrations of E-4031 led to progressive increases in the half time to relaxation (Suppl.Fig. 5C), increases in area under the diameter-time curve (Suppl.Fig. 5D), increases in contraction amplitude (Suppl.Fig. 5E), and decreases in frequency (Suppl.Fig. 5F).The similarities between the responses to E-4031 of human and rat vessels in both Vm (Fig. 2, Suppl.Fig. 4) and diameter recordings (Fig. 3, Suppl.Fig. 5) suggest that hERG channels in human LMCs play similar roles in AP repolarization as in rodent LMCs.

Discussion
Our results support the conclusion that ERG1 (Kv11.1)channels mediate a major component of AP repolarization in lymphatic muscle.This is the first report of a specific K + channel contributing to AP repolarization in this cell type.ERG1 channel message is expressed in mouse, rat and human lymphatic vessels, and FACS sorting of murine lymphatic vessels expressing a GFP reporter in the smooth muscle layer show that ERG1 is present in LMCs.Vm measurements in rat and human vessels reveal that ERG1 inhibitors delay AP repolarization and ERG1 activators accelerate repolarization, with IC 50 and EC 50 values in the nanomolar range.In pressurized mesenteric lymphatics from rats and humans, the ERG1 inhibitor E-4031 has clear effects on the relaxation phase of spontaneous contractions, lengthening the time required for relaxation and decreasing the frequency of spontaneous contractions.We conclude that ERG channels play a critical role in the repolarization of the AP generated by LMCs.Loss of ERG channel activity results in prolonged AP duration and a lower frequency of active lymphatic pumping, which are predicted to lead to a net impairment in lymph transport.
The role of ERG1 channels in controlling AP repolarization and contraction frequency in LMCs is consistent with their documented contribution to the rhythmicity of other types of smooth muscle.The spontaneous APs of rat gall bladder smooth muscle have nearly identical characteristics to the APs in rat LMCs shown in Figs. 2, 5, 7 and 8: a resting Vm between − 40 and − 45 mV, a single AP spike to + 5 mV, an AP plateau lasting ~ 1 s and spontaneous firing that is initiated by gradual diastolic depolarization 34 .In gall bladder, E-4031 had very similar effects (at somewhat higher concentrations) to what we observed in rat LMCs: prolongation of the AP plateau, induction of multiple spikes, delayed repolarization, increased contraction amplitude and decreased frequency 34 .ERG channels also contribute to the spontaneous oscillations in Vm and/or contractility in bladder 31 , esophagus 32 , jejunum 41,42 , portal vein 33,43 , epididymis 44 and myometrium 28,29 .In mouse and human myometrium, ERG1 activity decreases in late pregnancy, due to upregulation of inhibitory accessory KCNE channel β-subunit(s) 28,29 , which reduce ERG activity by increasing the rate of channel deactivation 45 .Unlike some other types of smooth muscle [32][33][34]43 , ERG1 inhibitors had no significant effect on the resting Vm of LMCs, suggesting that ERG1 channels are not strongly activated until an AP spike is initiated. Hoever, we did observe E-4031-mediated depolarization in some LMCs (Fig. 2) and more consistently so if naïve cells were exposed to high, rather than gradually increasing, concentrations of the inhibitor (Suppl.Fig. 1).In gall bladder, a clear depolarizing effect of E-4031 was recorded when APs were first blocked by diltiazem 34 , but we did not test that protocol.
It is interesting that the application of both ERG1 activators and inhibitors led to a decrease in the frequency of LMC APs and lymphatic vessel contractions.Explaining the effect of the inhibitors on frequency is straightforward: lengthening the AP plateau duration from 1 to 24 s (in the case of 300 nM E-4031 in Fig. 2Be) must necessarily result in a lower AP and contraction frequency.Although ERG1 activators accelerated repolarization and narrowed the AP plateau width, higher concentrations of ICA-105574 lowered both AP and contraction frequency by reducing overall LMC excitability, which is evident in the traces shown in Figs. 5 and 7, where diastolic depolarization was slowed or reversed, such that Vm underwent multiple oscillations before reaching threshold.This reduced overall excitability would account for the trend in Fig. 6 for frequency to be progressively reduced as the concentrations of the ERG-1 activators increased.
Although the inhibitors and activators used in our protocols are reported to be selective for Kv11 channels at low concentrations 37,38 , we are unable to completely rule out contributions from other Kv channel family members-Kv1 in particular-to the observed effects of these agents on AP repolarization in LMCs.TEA and 4-AP were found in previous lymphatic studies to increase the duration of the AP plateau in rat LMCs, but these  compounds produced substantial depolarization and repetitive AP firing compared to E-4031 22,23,46 .The differences in the pattern of responses to TEA and 4-AP (Suppl.Fig. 6) vs. ~ 10,000-fold lower concentrations of E-4031 (Fig. 4A) suggest that E-4031 does not exert its effects on AP plateau duration in LMCs by inhibiting other Kv isoforms.A way to resolve this issue would be to perform voltage-clamp studies of single, isolated LMCs.ERG1 current is most definitively distinguished in voltage-clamp protocols of K + tail currents by an initial 'hook' that results from partial inactivation during depolarization and which is removed (reversed) with a fast time course by hyperpolarization 27,28,33,47 .This rapid voltage-dependent inactivation and recovery from inactivation presumably makes the ERG channel well suited to determine the duration of the plateau phase of an AP 27 .Although such protocols were beyond the scope of the present study, future experiments utilizing voltage-clamp protocols on single, enzymatically isolated LMCs could be used to identify the ERG1 current signature, determine the sensitivity of that current to ERG1 inhibitors/activators, and to determine the fraction of whole-cell K + current that is sensitive to concentrations of ERG1 inhibitors/activators.Additional electrophysiology studies in transgenic mice, e.g., Vm recording and/or patch clamp protocols in mouse LMCs, could test the impact of conditional ERG1 deletion on the AP shape and the frequency and amplitude of spontaneous contractions.
ERG-1 currents can contribute to pacemaking if they are active at the resting Vm and their slow deactivation after an AP promotes diastolic depolarization.This property has been demonstrated in SA nodal cells [48][49][50][51][52] , but it could also facilitate pacemaking potentials in other cell types, including LMCs.Again, the potential role of ERG1 channels in lymphatic pacemaking would best be studied using voltage or current clamp protocols in LMCs and in mice with conditional knock out of ERG1 channels.The contribution of ERG channels to LMC pacemaking is predicted to depend on the resting Vm and thus may vary between lymphatics of different body regions and between species, as Vm appears to be more negative in human than rat or mouse lymphatic vessels 10,13,35,36,53 .
ERG-1 channels mediate the rapid component of the delayed rectifier current controlling AP repolarization in cardiac myocytes and ERG1 mutations account for about one fourth of all hereditary arrhythmias in humans, including class II long QT syndrome.Because ERG1 channels exert similar control over the AP plateau duration in LMCs, is likely that the same ERG1 mutations also affect lymphatic contractile function, as well as the function of other tissues/organs with rhythmically active smooth muscle.Our data show that the loss of ERG channel activity produces a substantial prolongation of the AP plateau duration in LMCs, resulting in a reduction in the normal frequency of spontaneous lymphatic contractions.These effects are predicted to impair both active lymphatic pumping and lymph transport, which is a question that could be explored in clinical studies.

Methods
Animal procedures.All procedures were carried out in accordance with relevant guidelines and regulations as approved by the animal care committee at the University of Missouri and complied with the standards stated in the "Guide for the Care and Use of Laboratory Animals" (National Institutes of Health, revised 2011).The study is reported in accord with ARRIVE guidelines.
Human tissues.Protocols on human lymphatics were conducted as approved by the Human Research Protection Office at Washington University (protocol #201111038 to GJR) and conformed to the principles of the Declaration of Helsinki and with relevant guidelines and regulations.Informed consent was obtained from all subjects and/or their legal guardian(s).Discarded samples of mesenteric fat and sometimes gut wall were collected from de-identified consenting patients undergoing intestinal surgical resections for bowel inflammation, obstruction, cancer or polyps at Washington University Barnes-Jewish Hospital.Staff affiliated with the Washington University Digestive Diseases Research Core Center obtained the informed consent, provided tissue collection oversight, and de-identification services as per protocol.
Mice and rats.Male Sprague-Dawley rats, 180-200 g, were obtained from Envigo (Indianapolis, IN).Male or female C57Bl/6J mice were obtained from JAX (JAX: 000664) and used at 2-3 months of age.For FACS analyses, Myh11-CreER T2 mice (B6.FVB-Tg(Myh11-cre/ERT2)1Soff/J) obtained from Stefan Offermanns (Max-Planck Institute, Bad Neuheim) were crossed to ROSA26mT/mG reporter mice (JAX: 007676).The male offspring were induced with i.p. injections of 100 mg tamoxifen (10 mg/ml in safflower oil) on five consecutive days.For genotyping, genomic DNA was extracted from tail clips using the HotSHOT method.Genotypes were determined by PCR with 2 × PCR Super Master Polymerase Mix (Catalog # B46019, Bimake, Houston, TX) according to the provider's instructions.Mice were provided ad libitum access to food and water and housed under normal light and dark cycles in cages of up to five mice.
Mouse and rat vessel isolation and cannulation.Mice and rats were anesthetized with ketamine/xylazine (100/10 mg/kg, i.p.) and placed face down on a heated pad.Inguinal-axillary lymphatic vessels in the mouse were exposed through a superficial incision in the side, removed and transferred to a dissection chamber filled with Krebs-albumin solution for further dissection.Rat mesenteric lymphatics were isolated by opening the abdomen and removing the entire small intestine, which was then pinned out for dissection of individual collecting vessels.Regardless of the vessel source, a short vessel segment containing at most 1 valve was then transferred to a 3-ml myograph chamber containing Krebs-albumin solution and cannulated at each end with a glass micropipette (60-80 μm OD), pressurized to 3 cmH 2 O, and further cleaned to remove excess fat and connective tissue.The chamber, with associated micropipettes, pipette holders and micromanipulators, was transferred to the stage of an inverted microscope.The micropipettes were connected to either a moveable reservoir for manual control of pressure or a microfluidic controller (Elveflow Instr., Paris) for computer control of pressure.Custom LabVIEW programs (National Instruments; Austin, TX) acquired analog data from low pressure sensors connected to the controller, simultaneously with vessel inner diameter, as detected from video images acquired

Figure 4 .Figure 5 .
Figure 4. Effects of E-4031 on contractions of rat pressurized mesenteric lymphatics.(A) E-4031 increased the amplitude and decreased the frequency of spontaneous contractions.Vertical lines (to Diam = 0) reflect transient blanking of the light path to mark addition and mixing of each concentration E-4031.(B) E-4031 prolonged the diastolic relaxation phase of the contraction cycle and higher concentrations were often associated with double contractions (also marked arrowheads in A), which were extremely rare under normal conditions in rat mesenteric lymphatics.(C-F) Summary of E-4031 effects on time to 50% relaxation (C), area under the diameter vs. time curve (D), contraction amplitude (E) and frequency (F).*Significantly different from control value using ANOVA with Dunnett's post-hoc tests, p < 0.05.Error bars are ± SD.N = 4, n = 9.

Figure 6 .
Figure 6.Summary of effects of the ERG channel activators ICA-105574 and RPR-260243 on Vm of rat pressurized mesenteric lymphatics.(A,B) ICA-105574 decreases the duration of the AP plateau phase, and the effect is equally significant if the duration is normalized to the initial control value (B).ICA-105574 does not significantly alter the resting Vm (C), or frequency (D), although there is a trend for frequency to decrease.RPR-260243, another ERG channel activator that works through a different mechanism of action, produces a similar effect on the plateau duration (E,F) but requires higher concentrations to reach significance.RPR does not significantly affect the resting Vm (G) or frequency (H), although there is a trend for frequency to decrease.*Significantly different from control value using a one-way ANOVA with Dunnett's post-hoc tests, p < 0.05.Error bars are ± SD.N = 5, n = 9.

Figure 7 .
Figure 7. ICA-105574 reverses the effects of E-4031.(A) After impalement and Vm stabilization, bath application of E-4031 leads to progressive widening of the AP plateau such that at time point (c) there are multiple secondary "spikes".The subsequent addition of 1 μM ICA-105574 in the continued presence of E-4031 leads to the narrowing of some (arrowheads) but not all APs, whereas 3 mM ICA-105574 consistently reduces plateau duration (e) to values lower than control (a).(B) Expanded time scales show how the duration of the plateau phase of the AP is altered first by the addition of E-4031 and then subsequent addition of ICA-105574.Individual APs marked (a-e) correspond to the times marked (a-e) in (A).Calibration bar: 10 mV•2 s −1 for all traces except (c) where it is 10 mV•5.8 s −1 .(C,D) Summary data for 4 vessels from 2 animals, using either the raw (C) or normalized (D, to control) data for AP plateau duration.The concentration of E-4031 was 100 nM; the concentration of ICA-105574 was 3 μM.Error bars are ± SD.Significance determined using a one-way ANOVA with Tukey's multiple comparison tests.*Indicates p < 0.05; **indicates p < 0.01; ***indicates p < 0.001; ns = not significant at p < 0.05.

Figure 8 .
Figure 8. E-4031 reverses the effects of ICA-105574.(A) After impalement and Vm stabilization, bath application of increasing concentrations of ICA-105574 leads to progressive narrowing of the AP plateau with eventual slowing of frequency.The subsequent application of E-4031 (3 μM) increases the frequency and results in widening of the AP plateau to a value slightly wider than control.(B) Expanded time scales show how the plateau phase of the AP changes first in response to ICA-105574 and then to ICA-105574 + 3 μM E-4031.Individual APs marked (a-e) correspond to the times marked (a-e) in (A).Note: for the "control" AP, an early recording in 100 nM ICA-105574 was selected because Vm had not completely stabilized before the first application of ICA-105574.Recording is representative of 3 similar experiments.Summary data for 10 vessels from 7 animals, using either the raw (C) or normalized (D, to control) data for AP plateau duration.Error bars are ± SD.Significance determined using a one-way ANOVA with Tukey's multiple comparison tests.**Indicates p < 0.01; ***indicates p < 0.001; ****indicates p < 0.0001.