Electric field stimulation unmasks a subtle role for T-type calcium channels in regulating lymphatic contraction

We previously identified two isoforms of T-type, voltage-gated calcium (Cav3) channels (Cav3.1, Cav3.2) that are functionally expressed in murine lymphatic muscle cells; however, contractile tests of lymphatic vessels from single and double Cav3 knock-out (DKO) mice, exhibited nearly identical parameters of spontaneous twitch contractions as wild-type (WT) vessels, suggesting that Cav3 channels play no significant role. Here, we considered the possibility that the contribution of Cav3 channels might be too subtle to detect in standard contraction analyses. We compared the sensitivity of lymphatic vessels from WT and Cav3 DKO mice to the L-type calcium channel (Cav1.2) inhibitor nifedipine and found that the latter vessels were significantly more sensitive to inhibition, suggesting that the contribution of Cav3 channels might normally be masked by Cav1.2 channel activity. We hypothesized that shifting the resting membrane potential (Vm) of lymphatic muscle to a more negative voltage might enhance the contribution of Cav3 channels. Because even slight hyperpolarization is known to completely silence spontaneous contractions, we devised a method to evoke nerve-independent, twitch contractions from mouse lymphatic vessels using single, short pulses of electric field stimulation (EFS). TTX was present throughout to block the potential contributions of voltage-gated Na+ channels in perivascular nerves and lymphatic muscle. In WT vessels, EFS evoked single contractions that were comparable in amplitude and degree of entrainment to those occurring spontaneously. When Cav1.2 channels were blocked or deleted, only small residual EFS-evoked contractions (~ 5% of normal amplitude) were present. These residual, EFS-evoked contractions were enhanced (to 10–15%) by the KATP channel activator pinacidil (PIN) but were absent in Cav3 DKO vessels. Our results point to a subtle contribution of Cav3 channels to lymphatic contractions that can be unmasked in the absence of Cav1.2 channel activity and when the resting Vm is more hyperpolarized than normal.


Introduction
Collecting lymphatic vessels generate spontaneous, twitch-like contractions that propel lymph centrally, accounting for of peripheral lymph ow during quiet standing 1,2 .These contractions are triggered by action potentials (APs) in lymphatic muscle cells (LMCs), whereby a single AP evokes a transient contraction that is entrained for the length of one or more lymphangions 3,4 .Although the ionic conductances underlying the AP in LMCs have not been completely resolved, inward current during the AP spike is carried by voltage-gated calcium channels (VGCCs) with a contribution of voltage-gated sodium channels (VGSCs) in some species [5][6][7][8][9][10][11][12] .T-type VGCCs are also expressed in mesenteric lymphatic vessels from rat 13 and sheep 7 and proposed to regulate the frequency of the ionic pacemaker driving spontaneous contractions 10,13 .
In a recent study of peripheral collecting lymphatics, we con rmed that L-type VGCCs (Ca v 1.2, encoded by Cacna1c and hereafter referred to as Ca v 1.2) and T-type VGCCs (Ca v 3.1 and Ca v 3.2, encoded by Cacna1g and Cacna1h, respectively, and hereafter referred to as Ca v 3.1 and Ca v 3.2, respectively, or collectively as Ca v 3) are expressed in LMCs of both rats and mice.We demonstrated through patch clamp protocols that products of Ca v 1.2 and Ca v 3 transcription form functional calcium channels gated by depolarization in rat and mouse LMCs 14 .However, contractile tests of lymphatic vessels from Ca v 3.1 −/− mice, and Ca v 3.2 −/− mice, and even Ca v 3.1 −/− ;Ca v 3.2 −/− double knock-out (DKO) mice, exhibited nearly identical parameters of spontaneous twitch contractions as wild-type (WT) control vessels (i.e., frequency, amplitude, ejection fraction or fractional pump ow), over a wide pressure range.Thus, although functional Ca v 3 channels are expressed in murine lymphatic smooth muscle, we concluded that they do not play a detectable role in determining these parameters of spontaneous lymphatic contractions in mice.In contrast, smooth-muscle speci c deletion of Ca v 1.2 abolished all spontaneous contractions, con rming that Ca v 1.2 channels are critical for their initiation and generation.While these ndings may apply only to the mouse and not to other species, commonly-used concentrations of the T-channel inhibitors Ni 2+ and mibefradil produced consistent inhibition of spontaneous contractions in lymphatic vessels from Ca v 3.1 −/− ;Ca v 3.2 −/− (Ca v 3 DKO) mice, suggesting that these compounds inhibit lymphatic smooth muscle primarily through their actions on Ca v 1.2 channels; they, therefore, have limited use in detecting the speci c contributions of Ca v 3 channels.Until the development of truly Ca v 3 selective inhibitors, genetic deletion strategies are necessitated to provide de nitive answers about the contributions of Ca v 3 channels to lymphatic contractile function.
In the present study, we developed additional tests to determine if Ca v 3 channels play a more subtle role in lymphatic function than revealed by standard analyses of spontaneous lymphatic contractions.First, we hypothesized that Ca v 3 channels, while insu cient to initiate lymphatic action potentials in the absence of Ca v 1.2, may participate in the calcium entry required for full-amplitude contractions and normal pacemaking, and that progressive and selective inhibition of Ca v 1.2 channels by a dihydropyridine antagonist might uncover a role for Ca 2+ entry through Ca v 3 channels.That idea was tested by comparing the contractile responses of WT vessels to increasing concentrations of nifedipine (NIF), in which voltage-dependent Ca 2+ entry into LMCs could be mediated by both Ca v 1.2 and Ca v 3 channels, with the responses of Ca v 3 DKO vessels, in which voltage-gated Ca 2+ entry into LMCs could only be mediated by Ca v 1.2 channels.Second, we hypothesized that shifting the LMC resting membrane potential (Vm) to a more negative voltage might unmask a contribution of Ca v 3 channels to contractions.
Because even slight hyperpolarization can eliminate spontaneous contractions 15 , we optimized methods to evoke nerve-independent, twitch contractions from mouse lymphatic vessels using an external eld of depolarizing electrical current.In WT vessels, single, short pulses of electric eld stimulation (EFS) evoked single, large-amplitude contractions that were entrained along the length of the cannulated lymphatic vessel.This activity could be blocked by NIF in WT vessels and was absent in Ca v 1.2 KO vessels.
However, in the presence of TTX to block VGSCs and NIF to block Ca v 1.2 channels, small residual EFSevoked contractions were present in WT vessels; these were enhanced by the K ATP channel activator pinacidil (PIN) but absent in Ca v 3 DKO vessels, and thus presumably mediated by Ca v 3 channels.The responses of WT lymphatic vessels were then compared to vessels from Ca v 3 DKO mice and smoothmuscle speci c Ca v 1.2 KO mice in the presence of TTX, PIN and/or NIF.Collectively, the results of these protocols point to a subtle contribution of Ca v 3 channels to lymphatic contraction amplitude and pacemaking frequency, that can be unmasked when resting Vm is more hyperpolarized than normal.

Results
We hypothesized that, if Ca 2+ in ux through both Ca v 1.2 and Ca v 3 channels is required for full-amplitude contractions and/or to maintain a normal pacemaking frequency, progressive inhibition of Ca v 1.2 might uncover a role for Ca v 3. Thus, we predicted that lymphatic vessels from Ca v 3 DKO mice would be more sensitive than WT vessels to inhibition by NIF.To test this idea, popliteal lymphatics were isolated, cannulated and allowed to establish a regular spontaneous contraction pattern at a xed intraluminal pressure.NIF was added to the bath in cumulative concentrations from 1 nM to 30 mM, while recording spontaneous contractions for 2 min at each concentration.After the experiment, AMP and FREQ were determined from the diameter recording and FPF was calculated as described in Methods.The data were then t to the Hill equation (when possible) to determine the IC 50 values from the concentration-response relationships.
The results of the NIF protocol are illustrated in Fig. 1.Panels A and B show representative recordings of spontaneous contractions in popliteal lymphatic vessels from WT and Ca v 3 DKO mice in response to progressively higher concentrations of NIF.NIF began to inhibit contraction amplitude (AMP) at about 10 nM in both vessels but completely inhibited AMP and frequency (FREQ) of the Ca v 3 DKO vessel at 100 nM, whereas the WT vessel required a concentration of 300 nM for complete inhibition.This same pattern is evident in the summary responses shown in Fig. 1C, where there is a slight left-shift in the AMP- [NIF]   curve for the Ca v 3 DKO vessels, with statistically signi cant differences between the normalized AMP of WT vs. Ca v 3 DKO vessels at 3x10 -8 and 1x10 -7 M NIF.Likewise, the FREQ-[NIF] curve for Ca v 3 DKO vessels was left-shifted from the WT curve (Fig. 1E) by ~1/2 log order, as was the FPF-[NIF] curve (Fig. 1D), suggesting enhanced sensitivity of Ca v 3 DKO vessels to NIF.We also computed normalized FREQ (normalized to the initial average frequency of each vessel before NIF application) because, for concentration-response curves, this parameter is often a more sensitive indicator of a drug effect due to vessel-to-vessel variations in the basal FREQ that sometimes occur.When FREQ was expressed as the change from control, there was ~1-1.5xlog left-shift in the normalized FREQ-[NIF] curve for Ca v 3 DKO vessels, compared to the curve for WT vessels (Fig. 1F).We repeated this protocol using popliteal lymphatics from Ca v 3.1 -/-and Ca v 3.2 -/-(single KO) mice, as shown in Suppl.Fig. 1-2.The differences between WT and Ca v 3.1 -/-vessels were more subtle than between WT and Ca v 3 DKO vessels but showed a similar trend.The FREQ-[NIF] curve for Ca v 3.1 -/-vessels was left-shifted by ~1/2 log order (Suppl.Fig. 1E) compared to WT vessels and the normalized FREQ-[NIF] curve was left-shifted by ~1 log order (Suppl.Fig. 1F) compared to WT vessels.Ca v 3.2 -/-vessels also showed the same trend but with smaller left-shifts in AMP, FREQ, and normalized FREQ (Suppl.Fig. 2E-F).The IC 50 values for all protocols are listed in Table 1.These results are consistent with the hypothesis that Ca v 3 isoforms may have contributed to the increased sensitivity of Ca v 3 DKO vessels to NIF, with possibly a greater contribution from Ca v 3.1 than Ca v 3.2 channels.
The results shown in Fig. 1 coupled with our previous ndings 14 raise the possibility that the contributions of Ca v 3 channels to lymphatic contractions are too subtle to detect in standard tests of spontaneous contractions of mouse vessels.We hypothesized that a more de nitive role for Ca v 3 channels might be uncovered if the resting Vm prior to AP initiation was more hyperpolarized than normal, bringing Ca v 3 channels more into their optimal voltage activation window.To test this, however, would require not only hyperpolarizing the membrane, but 1) inhibiting Ca v 1.2 channels (which would otherwise predominate), 2) blocking any possible contribution from VGSC channels [in which TTXsensitive Na v 1 isoforms are predominant 6 ], which might also be enhanced at hyperpolarized potentials, and 3) evoking contractions independent of the intrinsic pacemaker, which would likely be inhibited at hyperpolarized resting potentials.We devised the protocol depicted in Fig. 2 to test our hypothesis.The theoretical window currents for Na v 1, Ca v 3 and Ca v 1.2 channels, relative to the resting Vm of mouse lymphatic smooth muscle, are illustrated in Fig. 2A.At the LMC resting Vm, Ca v 1.2 channels are predicted to be well within the range of their optimal window current, but Ca v 3 channels are predicted to be barely within their range.Fig. 2B depicts how this situation would change after inhibition of Na v 1 channels 6 by TTX and inhibition of Ca v 1.2 channels by NIF, as the latter is known to depolarize LMCs by ~10 mV 12 .
The subsequent addition of PIN to activate K ATP channels would shift Vm to a hyperpolarized value, as we have shown recently 15 , where a greater fraction of Ca v 3 current potentially would be available to be activated.Although the membrane likely would be too hyperpolarized to allow spontaneous activation of an AP by the intrinsic pacemaker potential, contractions potentially could be evoked by an external stimulus.An example recording of Vm in mouse LMCs at rest when spontaneous APs are ring, after the application of 1 mM NIF, and after subsequent addition of PIN, is shown in Fig. 2C.Resting Vm was -40 mV, depolarized to -33 mV after NIF and hyperpolarized to -40 mV after addition of 300 nM PIN, then to ~-50 mV after 1 mM and 3 mM PIN.The results of several such experiments are summarized in Fig. 2D and are consistent with the NIF-and PIN-induced shifts in Vm predicted in Fig. 2B.We were unable to directly determine the amount of LMC depolarization produced by EFS because the high voltage could have damaged the head-stage circuitry of the ampli er during Vm measurement.The amount of PINinduced hyperpolarization was quite variable between vessels and could be transient [see Fig. 2C and 15 ].
In addition, the PIN effect on a particular vessel might be su cient to hyperpolarize Vm out of the range for EFS-mediated depolarization.For these reasons, we tested a 10-fold range of PIN (0.3 to 3 mM) on each vessel, expecting that at least one of the concentrations would produce a degree of hyperpolarization that was su cient to recruit Ca v 3 channels and yet still be overcome by a subsequent EFS pulse.
We then implemented the protocol illustrated in Fig. 3A.Single, short pulses of EFS (0.1-0.2 mS, 90V) were used to elicit single contractions from WT popliteal lymphatics.The duration of the EFS pulse was set at < 0.3 mS because twitch contractions were often slow to recover when stimulus durations exceeded 1 mS and sometimes exhibited prolonged diastolic relaxation times and increased tone for seconds or minutes (note the contractions evoked by 1 and 5 mS pulses in Suppl.Fig. 3).Depending on the baseline contraction FREQ, pressure was lowered to 1 or 2 cmH 2 O to reduce the rate of spontaneous contractions, allowing for a su ciently long diastolic pause in the contraction cycle, during which we could evoke an extra contraction.Thus, EFS pulses were typically delivered within a few seconds after completion of a spontaneous contraction, and we designated EFS-induced contractions as those occurring within 50 ms following an EFS pulse.The amplitudes and durations of the evoked contractions were nearly identical to those of spontaneous contractions (Fig. 3B, withtiming of the EFS pulses shown) and the entrainment of each EFS-evoked contraction wave was similar to that of a spontaneous contraction, as measured from off-line analysis of the spatio-temporal (ST) maps (Suppl.Fig. 4).The ability of EFS to evoke entrained twitch contractions is in general agreement with the ndings of McHale et al. 16 in bovine mesenteric lymphatics, except for the speci c values of the stimulus parameters, which are expected to vary depending on a number of factors, including the species, vessel size, chamber design, electrode diameter and placement.Contractions evoked by single EFS pulses (0.1-0.3 mS, 90V) were not inhibited in the presence of TTX (Fig. 3B).Curiously, the application of TTX (1 mM) in itself had no effect on spontaneous contraction AMP or FREQ in ~50% of vessels (Fig. 3B) but caused a transient cessation of spontaneous contractions in the other ~50% of vessels; however, those vessels recovered after 1-4 min and resumed a normal FREQ and AMP in the continued presence of TTX.
Representative recordings are shown for each combination of vessel genotype and inhibitor in Fig. 4. All traces were recorded in the presence of TTX (1 mM).The set of traces at the top (Fig. 4A-C) for a WT vessel show that spontaneous contractions were blocked by NIF (1 mM) and that EFS pulses initiated only small contractions (<5 mm in AMP; Fig. 4B).Although these contractions were much weaker than those elicited in the absence of NIF, they conducted over most of the vessel (Suppl.Fig. 4B).The record in Fig 4B also shows one spontaneous contraction that occurred in the presence of NIF; events of this kind exceeding 3 mm in AMP were extremely rare, were associated with much lower conduction speeds (Suppl.Fig. 4B) and did not consistently conduct over long distances.The addition of PIN (3 mM) in the continued presence of NIF resulted in larger contraction amplitudes in response to identical EFS pulses (Fig. 4C).Spontaneous contractions over 3 mm in AMP were not observed in vessels from Ca v 1.2 smKO mice (Fig. 4D).In these vessels EFS pulses initiated only very small or negligible contractions that nevertheless were enhanced in AMP by PIN (Fig. 4E).In contrast, vessels from Ca v 3 DKO mice showed spontaneous contractions with normal AMP and EFS pulses evoked additional contractions of equivalent AMP (Fig. 4F).However, in the presence of NIF (1 mM) EFS pulses failed to evoke any residual contractions in these vessels (Fig. 4G) even after the addition of PIN (Fig. 4H).
The data for the various genotypes and pharmacological treatments are summarized in Fig. 5. TTX was present in all protocols.The amplitude of spontaneous contractions was averaged over a 2-min period.The amplitude of EFS pulses was the average for the contractions evoked by three EFS pulses, excluding any cases in which the pulse was delivered too close (< 50 ms) to a spontaneous contraction to be certain which was the initiating event.There were no signi cant differences in the average AMP of spontaneous contractions vs the average AMP of EFS-evoked contractions (Fig. 5A).However, both were signi cantly different than the average AMP of EFS-evoked contractions in the presence of NIF alone or NIF plus PIN (either 300 nM, 1 mM or 3 mM, or the largest AMP of any the three PIN concentrations for each vessel).A Wilcoxon matched pairs signed rank test was used to compare the difference in the AMP of EFS-evoked contractions in NIF alone vs NIF + the largest AMP of the three PIN concentrations.This difference was highly signi cant, indicating that PIN signi cantly increased the AMP of EFS-evoked contractions when Ca v 1.2 channels were blocked.The same analysis is presented in Fig. 5B for vessels from Ca v 1.2 smKO mice.In contrast to WT vessels, Ca v 1.2-de cient vessels had extremely small spontaneous contraction amplitudes (≤ 3 mm in all but one case) and EFS pulses likewise evoked contractions with amplitudes < 1 mm.The amplitudes of EFS-evoked contractions were enhanced by all concentrations of PIN, with the difference between EFS alone and EFS + the most effective PIN concentration (in this case always 3 mM PIN) being highly signi cant.Thus, the results shown in Fig. 5A and 5B are in agreement in showing that the amplitudes of the residual contractions evoked by EFS when Ca v 1.2 channels are deleted or blocked are enhanced when the membrane is hyperpolarized by PIN.
Finally, the same analysis is shown in Fig. 5C for vessels from Ca v 3 DKO mice.As in vessels from WT mice, there was not a signi cant difference in the amplitudes of spontaneous vs. EFS-evoked contractions.However, only very small contractions (<3 mm on average) could be evoked by EFS in the presence of NIF and these were not enhanced by any concentration of PIN.Nearly identical results for each of the three genotypes were produced when normalized AMP, rather than raw AMP, was used for the analysis (Suppl.Fig. 5).Collectively, these results suggest that Ca v 3 channels are mediating the PINinduced enhancement of the residual EFS-evoked contractions when Ca v 1.2 channels are deleted or blocked.

Discussion
In this study we asked: If functional Ca v 3 channels are expressed in lymphatic muscle, why do they not contribute a detectable component to the AP in lymphatic muscle 14 or make a signi cant contribution in the frequency or strength of spontaneous lymphatic contractions?We rst examined this issue by comparing concentration-response curves for WT and Ca v 3 DKO vessels to the Ca v 1.2 dihydropyridine antagonist NIF, reasoning that lymphatic vessels from Ca v 3 DKO mice would be more sensitive than WT vessels to inhibition by NIF because WT vessels have both Ca v 1.2 and Ca v 3 channels as Ca 2+ in ux sources whereas Ca v 3 DKO vessels have only Ca v 1.2 channels.Vessels from Ca v 3-de cient mice were indeed more sensitive to NIF (Fig. 1., Suppl.Figures 1-2), and the effect was more substantial for FREQ (~ 10-fold more sensitive than WT) than for AMP (~ 3-fold).Even though the NIF concentrations (30-100 nM) associated with leftward shifts in the AMP and FREQ of Ca v 3 DKO vessels (Table 1) were well below those causing substantial off-target effects on Ca v 3 channels (≥ 3 µM, 14 ), we could not completely rule out the possibility of off-target effects.Nevertheless, the results of that protocol were consistent with Ca v 3 channels contributing subtly to both the AMP and FREQ of spontaneous lymphatic contractions.We then devised a second set of experiments to test for a subtle role of Ca v 3 channels, reasoning that they might be mostly inactivated under the standard conditions used previously 14 to assess spontaneous contractions.EFS was used to initiate contractions after Ca v 1.2 channels had been inactivated, either by nifedipine application or by genetic deletion of Ca v 1.2 from lymphatic smooth muscle.All vessels were treated with TTX to eliminate any possible contribution of Na V channels whose activity could drive calcium in ux through the sodium-calcium exchanger (NCX) in reverse mode.Under both conditions EFS produced small, residual contractions, 2-4 µm in AMP, compared to a normal contraction AMP of ~ 40 µm.These contractions were enhanced (to 5-10 µm, equivalent to 10-15% of the AMP of a typical, spontaneous twitch contraction) after hyperpolarizing the membrane with the K ATP channel activator pinacidil prior to the EFS pulse.Importantly, this enhancement was absent in vessels from Ca v 3 DKO mice (Fig. 5), con rming that the residual EFS-evoked contractions were mediated by Ca v 3 channels.We conclude that Ca v 3 channels make < 5% contribution to the spontaneous contraction AMP and/or FREQ of mouse lymphatic vessels under normal conditions, but that this may be enhanced to 10-15% under conditions when the resting membrane potential is slightly hyperpolarized.Methodological limitations.Separating the contributions of Ca v 1.2 and Ca v 3 channels to Ca 2+ in ux has proven di cult in many different cell types, including LMCs.Studies of rat lymphatic vessels suggested a selective role for Ca v 3 channels in controlling LMC pacemaking 13 , based on the effects of inhibition with mibefradil and Ni 2+ .We previously showed that two Ca v 3 isoforms are expressed in mouse and rat LMCs and used patch clamp protocols to con rm the presence of functional channels.However, standard contraction tests revealed no signi cant differences between WT and Ca v 3 DKO vessels in either the FREQ vs. pressure or AMP vs. pressure relationships.The typical activation threshold for Ca v 3 channels is -20 to -30 mV more negative than that for Ca v 1.2 channels, and window currents for Ca v 3 channels are similarly left-shifted [17][18][19][20] .These values are estimates from arterial SM because no comparable measurements have been made in lymphatic SM.At the resting Vm that we measure in mouse LMCs (~-35 mV), it is likely that Ca v 3 channels are almost completely inactivated, unless a more left-shifted splice variant of Ca v 3 is expressed, as demonstrated for Ca v 1.2 21 .However, the resting Vm is slightly more negative in rat and human mesenteric LMCs [-40 and − 45 mV, respectively, 15,22 ], potentially enabling more basal activity of Ca v 3 channels in those species.
EFS was used in these experiments to override the normal LMC pacemaking mechanism so that contractions could be induced without the involvement of Ca v 1.2 channels.Although our results suggest that the residual contractions evoked by EFS are mediated by Ca v 3 channels, EFS could also have increased Ca 2+ in ux through other smooth muscle cation channels, e.g., TRPC6, TRPM4, and/or Ca v 2 or Ca v 1.3 channels, all of which are resistant to NIF.Although Ca v 1.3 channels in smooth muscle 23,24 are voltage-gated (but less sensitive to dihydropyridine antagonists than Ca v 1.2 25,26 ), TRPC6 and TRMP4 channels are relatively insensitive to membrane potential 27 and currents through those TRP channels would not be predicted to be signi cantly enhanced by PIN-induced hyperpolarization.Both TRPC6 and TRPM4 are expressed in mouse LMCs (our unpublished observations), but there is no evidence for the expression of Ca v 1.3 in lymphatic muscle, nor have we detected message for Ca v 1.3 channels in RT-PCR assays of puri ed mouse LMCs or in scRNA seq assays (our unpublished observations).Importantly, the possible contributions of TRPC6, TRMP4, and other channels to the residual EFS-evoked contractions should have been the same in Ca v 3 DKO and WT vessels and are therefore not consistent with the absence of those contractions in Ca v 3 DKO vessels (Fig. 5B).Another possible explanation for the residual EFS-evoked contractions is that 1 µM NIF may not have completely inhibited Ca v 1.2 channels, such that hyperpolarization prior to the EFS pulse then recruited Ca v 1.2 current rather than Ca v 3 current.Higher concentrations of NIF could have possibly blocked Ca v 3 channels 28,29 and it was for this reason that we also tested vessels from Ca v 1.2 smKO vessels.Our nding that EFS-evoked contractions in vessels de cient in Ca v 1.2 (Fig. 5) were of the nearly identical AMP as those in WT vessels + NIF argues against this possibility.Additionally, PIN treatment of Ca v 3 DKO vessels would also have recruited whatever fraction of Ca v 1.2 channels were not inhibited by NIF (presumably to the same degree as in WT vessels) and yet PIN did not potentiate evoked contractions under the same conditions in Ca v 3 DKO vessels.Physiological Relevance.Our results suggest that slight hyperpolarization of mouse LMCs can recruit additional Ca 2+ in ux through Ca v 3 channels.One implication is that rat and human LMCs, for which resting Vm levels are slightly more hyperpolarized than mouse LMCs, may normally have a larger (but probably still < 15%) contribution of Ca v 3 channels to the AMP and/or FREQ of spontaneous contractions.This conclusion is consistent with observations of Lee et al. 13 , despite the uncertainties of the off-target effects in that study of Ni 2+ and mibefradil on Ca v 1.2 channels.Although mouse Ca v 3 channels normally contribute < 5% to the contraction amplitude of mouse LMCs, if mouse LMCs were chronically hyperpolarized, e.g., by an endogenous or exogenous vasoactive agent, rapid depolarization to threshold would be predicted to recruit Ca v 3 channels to participate in a subsequent AP, and potentially enhance contraction AMP and/or FREQ.This hypothesis remains to be tested.
An incidental nding from our study is that Ca v 1.2 appears to not only mediate the upstroke of the AP in mouse LMCs (with likely contributions from Na V in rat and human LMCs), but to also modulate the pacemaker.The data in Fig. 1E-F, Suppl.Figures 1E-F, 2E-F show ~ 50% rise in FREQ that occurs in response to partial inhibition of Ca v 1.2 by low concentrations of NIF, suggesting that Ca 2+ entry through Ca v 1.2 channels normally retards the pacemaker.As multiple ion channels with interrelated activities comprise the currents that initiate and contribute to the LMC action potential, there are several potential mechanisms by which the sub-maximal NIF concentrations could drive increased frequency.Of note, 1µM NIF results in a signi cant depolarization and presumably sub-maximal concentrations might also depolarize the cell toward the threshold potential.Additionally, the activation of Ca v 1.2 channels with the agonist BayK8644 dramatically lengthens the duration of the AP plateau phase 30 , whereas inhibition of Ca v 1.2 and reduced calcium in ux during the AP would be expected to accomplish the opposite, as there would be reduced activation of Ano1 and potentially of NCX.A reduction in the plateau period would shorten the overall electrical cycle and thus a higher FREQ could be achieved.Another possibility is that while cytosolic calcium is typically considered to drive depolarization 31 , differential spatial coupling of calcium store release channels to Ano1 and Ca v 1.2 channels vs. hyperpolarizing channels such as BK 32,33 could provide a condition in which Ca 2+ entry through Ca v 1.2 channels normally retards the pacemaker.
Clinical Relevance.The relevance of Ca v 3 channels to lymphatic function in human medicine relates to their possible therapeutic targeting to reverse lymphatic collector dysfunction in chronic lymphedema.Olszewski's observations of patients with impaired lymphatic smooth muscle contraction strength and lower contraction frequency, or even complete loss of spontaneous contractions in various stages of secondary lymphedema [34][35][36] , point to a problem involving disruption of the pacemaking mechanism that potentially could be corrected pharmacologically.However, eventual therapeutic targeting of ionic dysfunction in human lymphatic muscle will require additional insights into the speci c types of ion channels involved in pacemaking, the speci c isoforms of those channels expressed in humans (which may be different than in rodents), and the development of selective inhibitors to block those channels.Whether Ca v 3 channels are expressed in human lymphatic muscle and are critical to some aspect of lymphatic function remains unknown at the present time.

Methods
Animal procedures.All procedures were 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 accordance with ARRIVE guidelines.
Animals.C57BL/6J wild-type (WT) mice were purchased from Jackson Laboratory (JAX, Bar Harbor, ME, USA).Ca v 3.1 -/-(Cacna1g null) mice on the C57BL/6J background, originally generated by Hee-Sup Shin (Korea Institute of Science and Technology; 37 , were a gift from Jeffrey Molkentin (University of Cincinnati), and rederived at MMRRHC, Columbia, MO, in the C57Bl/6 background.Ca v 3.2 -/-mice, originally generated by Chen et al. 38 , were obtained from JAX (B6;129-Cacna1h,tm1Kcam./J; #013770), bred into the C57Bl/6 background for at least 8 generations.Ca v 3.2 -/-and Ca v 3.1 -/-mice were bred to generate Ca v 3.1 /-;Ca v 3.2 -/-double KO mice on the C57Bl/6 background.Myh11-CreER T2 mice (B6.FVB-Tg(Myh11-cre/ERT2)1Soff/J), obtained from Dr. Stefan Offermanns, were bred with Ca v 1.2 f/f mice (Cacna1c tm3Hfm /J; #024714), which were purchased from JAX, to generate Myh11-CreER T2 ;Ca v 1.2 l/l mice (hereafter referred to as Ca v 1.2 smKO mice).All genotypes were veri ed by PCR.Mice from the latter strain were injected with tamoxifen (10mg/ml, 100ml i.p.) for 5 days and allowed to recover for 2 weeks before being used for experiments.Mice were provided ad libitum access to food and water and housed under normal light and dark cycles in cages of up to ve mice.Mice of either sex (except for Ca v 1.2 smKO mice) were studied at 5-10 weeks of age (18-25 g).Lymphatic vessel isolation.Mice were anesthetized with pentobarbital sodium (60mg kg − 1 , i.p.).An incision was made on the dorsal-medial side of either leg from the ankle to the groin to access the popliteal lymphatics.An excised lymphatic vessel was pinned on a Sylgard platform (Sylgard® 184, Dow Corning, Midland, MI, USA) in Krebs' buffer supplemented with 0.5% albumin, and isolated by dissection from the surrounding connective tissue and fat.After surgery, the animal was euthanized.
Pressure myography.An excised lymphatic vessel containing at least one valve was transferred to a 3 mL chamber where it was cannulated onto two micropipettes and pressurized.The bath was exchanged at a rate of 0.5 ml/min with Krebs buffer and equilibrated for 30-60 minutes at 37 o C with pressure set to 3 cmH 2 O, as previously described 14 .The pipettes contained 0.5% albumin-supplemented Krebs buffer.
Vessels used for further experimentation (except those from Ca v 1.2 smKO mice) developed robust, spontaneous contractions, with contractions that were entrained over the entire vessel length and amplitudes exceeding 30% at pressure = 3 cmH 2 O. Inner diameter at a representative region was measured continuously from video images using digital edge-detection 39 .Pressures and diameter were digitized using a National Instruments A-D system (Austin, TX) under the control of a LabVIEW program as described previously 40 .
Sharp electrode recordings of Vm.In separate experiments, Vm was recorded in the smooth muscle cell layer of pressurized WT mouse lymphatic vessels to verify the extent of PIN-induced hyperpolarization after L-type VGCC inhibition.To permit stable recordings of Vm in contracting vessels, wortmannin (1-3 mM, 20-30 min) was used to inhibit myosin light chain kinase and blunt vessel movement; the concentration and exposure time were adjusted to preserve minimal contractions (< 5 microns) that con rmed preservation of viability.The smooth muscle layer was impaled with an intracellular microelectrode (300-350 MW) lled with 1M KCl, and Vm was recorded using a NPI SEC-05x ampli er (ALA instruments, Farmingdale, NY) as previously described 31 .The ampli er output was digitized at sampled at 1 KHz using a D-A interface (National Instruments).After a successful impalement, Vm was allowed to stabilize for 15-30 seconds.The most negative value during the AP was approximately − 35 mV.After recording multiple contraction cycles, 1 mM NIF was added to the bath solution to inhibit L-type Ca 2+ channels.In some cases the impalement was lost due to the mixing procedure and, when that happened, attempts were made to impale the same cell or an adjacent cell and continue the protocol.Subsequently, PIN was added in cumulative concentrations (0.3, 1, 3 mM) while recording Vm.Once the recording was completed, the electrode was retracted from the cell and the recorded values were corrected for the offset potential.Electric eld stimulation.EFS was achieved using two 0.5 mm platinum wires (Warner Instruments, #64-1942), separated by 2.5 mm within the 3 mL bath chamber.The wires were positioned 2 mm above the bottom of the observation chamber and insulated except for the terminal 4 mm.The cannulated vessel was positioned 1 mm from the chamber bottom, equidistant between the two wires.A Grass S48 stimulator provided the depolarizing current.Initial tests showed that single twitch contractions, of amplitude comparable to those of spontaneous contractions, could be elicited with short duration (< 1 mS), single pulses of 80-90V.90V pulses were routinely used to ensure consistent responses.The synch output of the stimulator was ampli ed and digitized using an A-D interface (National Instr., Austin TX) to document pulse delivery in register with the diameter recording.For EFS protocols, pressure was usually set to either 1 or 2 cmH 2 O, depending on the spontaneous contraction rate, to provide a contraction pattern with a su ciently long diastolic period to allow for single EFS pulses to be delivered in lymphatic diastole.
Contraction wave analysis.To quantify the degree of entrainment of EFS-evoked contraction waves, bright eld videos of spontaneous contractions were acquired at video rates ranging from 30 to 50 fps.Recorded videos were then stored for o ine processing, analysis, and quanti cation of the conduction speed.Videos of contractions were processed frame by frame to generate two-dimensional spatiotemporal maps (STMs) representing the measurement of the outside diameter (encoded in 8-bit grayscale) over time (horizontal axis) at every position along the vessel (vertical axis), as described previously 3 .All video processing and analyses were performed using a set of custom-written Python programs.Conduction speed was determined for each wave by the slope of the corresponding band on the ST map (by linear t of the points de ning the leading edge) and the speeds were averaged for all the contractions in a given video.
Experimental Protocols.After a vessel established a consistent pattern of spontaneous contractions, one of two protocols was conducted.
The rst protocol assessed the concentration-dependent inhibition by NIF on spontaneous contractions.
After equilibration and establishment of a consistent pattern of spontaneous contractions at constant pressure, bath perfusion was stopped and NIF was added in cumulative concentrations (1 nM to 10 µM) to the bath.Pressure was set at either 1 or 2 cmH 2 O, depending on the spontaneous contraction rate of a given vessel.Contraction responses were recorded for 2-3 min before the next concentration was applied and the protocol was completed within 20 min, a time period found previously not to produce signi cant effects on contraction FREQ or AMP due to bath evaporation.
For the second protocol, single voltage pulses (typically 0.1-0.3mS, 90 V) were applied during the diastolic phase of the contraction cycle, with the pulses delivered 30-60 sec apart and timed to produce minimal disruption to the spontaneous contraction pattern; this was repeated 3 times.With pressure maintained at 3 cmH 2 O, the bath perfusion was stopped and TTX (1 µM) applied.After assessing the effect of TTX on the contraction pattern for 3-4 min, three identical stimulus pulses were again delivered (30-60 sec apart).For WT vessels, NIF (1 µM) was subsequently added to the bath and after 4 min the stimulus pulses were repeated.In a similar set of tests, vessels from Ca v 1.2 smKO mice were used in lieu of NIF treatment.In both cases the K ATP channel activator, pinacidil (PIN), was then added to the bath in increasing concentrations (0.3, 1, 3 µM) to hyperpolarize LMCs, allowing 2-3 min equilibration at each concentration before delivering stimulus pulses.Each time a drug was added to the bath the light path was temporarily blocked to create a vertical blanking artifact on the diameter trace.Tests using the same protocol were conducted on vessels from Ca v 3.1 -/-;Ca v 3.2 -/-mice.In each case the total protocol was completed in less than 20 min.

2 Fractional
Pump Flow (FPF)= EF•FREQ v 3.1 -/-;Ca v 3.2 -/-popliteal lymphatics are more sensitive to inhibition by NIF than WT lymphatics.A) Response of a WT popliteal lymphatic vessel to increasing concentrations of NIF (applied cumulatively).Each contraction is a downward de ection (individual contractions cannot be resolved with this compressed time scale).Vertical lines are intentional artifacts created by blanking the light path to mark when a new concentration was added, followed by ~10 secs of mixing.Pressure was held constant at 2 cmH 2 O.The cumulative DMSO concentration was < 0.4% and without effect alone.B) Response of a Ca v 3.1 -/-;Ca v 3.2 -/-popliteal lymphatic to the same NIF protocol.Contractions in the Ca v 3.1 -/-;Ca v 3.2 -/- vessel are completely inhibited at 100 nM NIF whereas the WT vessel requires at least 300 nM NIF to block contractions.C) Summary data for normalized AMP (normalized to the average AMP during the control period) as a function of NIF concentration.The curve for the Ca v 3.1 -/-;Ca v 3.2 -/-vessels is shifted to the left by ~1/2 log order, with two concentrations being signi cantly different.Summary data for FPF (D) and Frequency (E) as a function of NIF concentration.One concentration was signi cantly different for each parameter.F) Summary data for Normalized FREQ as a function of NIF concentration (normalized to the average FREQ during the control period).Two concentrations were signi cantly different and the curve for the Ca v 3.1 -/-;Ca v 3.2 -/-vessels was shifted to the left by ~1 log order.Statistical tests were two-way repeated measures ANOVAs with Tukey's multiple comparison post-hoc tests (*, p<0.05).WT: N=5; n=9.Ca v 3 DKO: N=8; n=15.

Figure 3 Protocol
Figure 3

Figure 4 Representative
Figure 4 Data analysis.Data were collected and analyzed using LabVIEW (National Instruments, Austin TX), Excel (Microsoft, Redmond, WA) and Prism 8 (Graphpad, La Jolla, CA, USA).Original recordings were plotted in IGOR (Wavemetrics, Oswego, OR).IC 50 values were determined in Prism or IGOR.The four standard tests in Prism for normality (Anderson-Darling, D'Agostino & Pearson, Shapiro-Wilk, Kolmogorov-Smirnov) were used to evaluate each data set and revealed that at least half of the data sets were not normally distributed.Subsequently, one-way ANOVAs with Krusal-Wallis post-hoc tests were performed to compare the amplitude of spontaneous and EFS-induced contractions across pharmacological treatments for each genotype, and Wilcoxon matched pairs signed rank tests were used to compare pairs of data sets within each genotype.The speci c tests used for each protocol are indicated in the gure legends.The data are expressed as mean ± standard error of the mean.P values < 0.05 were considered statistically signi cant, but other signi cance levels are marked when appropriate.N refers to the number of animals and n refers to the number of vessels or cells included per group.