T-type, but not L-type, voltage-gated calcium channels are dispensable for lymphatic pacemaking and spontaneous contractions

The spontaneous contractions of collecting lymphatic vessels provide an essential propulsive force to return lymph centrally. These contractions are driven by an intrinsic electrical pacemaker, working through an unknown underlying ionic mechanism that becomes compromised in some forms of lymphedema. In previous studies, T-type voltage-gated Ca2+ channels (VGCCs) were implicated in this pacemaking mechanism, based on the effects of the reputedly selective T-type VGCC inhibitors mibefradil and Ni2+. Our goal was to test this idea in a more definitive way using genetic knock out mice. First, we demonstrated through both PCR and immunostaining that mouse lymphatic muscle cells expressed Cav3.1 and Cav3.2 and produced functional T-type VGCC currents when patch clamped. We then employed genetic deletion strategies to selectively test the roles of each T-type VGCC isoform in the regulation of lymphatic pacemaking. Surprisingly, global deletion of either, or both, isoform(s) was without significant effect on either the frequency, amplitude, or fractional pump flow of lymphatic collectors from two different regions of the mouse, studied ex vivo. Further, both WT and Cav3.1−/−; 3.2−/− double knock-out lymphatic vessels responded similarly to mibefradil and Ni2+, which substantially reduced contraction amplitudes and slightly increased frequencies at almost all pressures in both strains: a pattern consistent with inhibition of L-type rather than T-type VGCCs. Neither T-type VGCC isoform was required for ACh-induced inhibition of contraction, a mechanism by which those channels in smooth muscle are thought to be targets of endothelium-derived nitric oxide. Sharp intracellular electrode measurements in lymphatic smooth muscle revealed only subtle, but not significant, differences in the resting membrane potential and action potential characteristics between vessels from wild-type and Cav3.1−/−; 3.2−/− double knock-out mice. In contrast, smooth-muscle specific deletion of the L-type VGCC, Cav1.2, completely abolished all lymphatic spontaneous contractions. Collectively our results suggest that, although T-type VGCCs are expressed in mouse lymphatic smooth muscle, they do not play a significant role in modulating the frequency of the ionic pacemaker or the amplitude of spontaneous contractions. We conclude that the effects of mibefradil and Ni2+ in other lymphatic preparations are largely or completely explained by off-target effects on L-type VGCCs, which are essential for controlling both the frequency and strength of spontaneous contractions.

pressure myography. A segment of 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 to 3 cmH 2 O. The bath solution containing Krebs buffer was exchanged at a rate of 0.5 mL/min and equilibrated for 30-60 minutes, at 37 °C, as previously described 52 . The inner diameter was tracked continuously from video images using a digital edge-detection method 54 . Vessels were included in the data analysis only if 1) spontaneous contractions and tone developed within the first 60 minutes; 2) normalized contraction amplitude exceeded 30% at 2 cmH 2 O intraluminal pressure (except for vessels from SMMHC-CreER T2 ; Cav1.2 fl/fl mice).
Each vessel was tested for contractile function using pressure steps between 0.5-8 cmH 2 O (for IALs) or 0.5-10 cmH 2 O (for popliteal lymphatics, PLs). After initial equilibration at P = 3 cmH 2 O, pressure was lowered to 2, 1, then 0.5 cmH 2 O, and then raised in steps from 0.5 to 10 cmH 2 O, while recording diameter for 2-3 min at each pressure. Contraction amplitude was typically optimal at P = 1 or 2 cmH 2 O while frequency increased with each pressure step from 0.5 to 5 cmH 2 O and reached a plateau at ~8 cmH 2 O. Tests using ACh, Bay K8644 or nifedipine were conducted at the pressure that produced the largest contraction amplitude (usually 2-3 cmH 2 O). To test a potential role for T-type VGCCs in lymphatic contractions, the inhibitors Ni 2+ (15 µM-2 mM), nifedipine (1 nM-10 µM), mibefradil (1 nM-100 µM) or TTA-A2 (100 nM-100 µM) were added to the bath solution. Acetylcholine (ACh, 1-300 nM) was applied to test its ability to inhibit contractions through NO production 55 . The L-type VGCC activator Bay K8644 (1 µM) was added to test residual Ca v 1.2 activity in vessels from tamoxifen-induced SMMHC-CreER T2 ; Ca v 1.2 fl/fl mice. Inhibitors were applied in increasing concentrations to generate cumulative concentration-response curves. Contraction responses were recorded for 3 min before the next concentration was applied. At higher concentrations of some pharmacological inhibitors, the changes in inner diameter were <5 µm and were not entrained over the length of the vessel-i.e. they resembled the irregular vasomotion often observed in arteries; these changes were not counted as bona fide propulsive contractions 52 . At the end of every pressure myograph experiment the vessels were superfused for 30 min with Ca 2+ -free Krebs buffer solution containing 3 mM EGTA to obtain the passive diameter at each pressure. calculation of contractile parameters. From internal diameter measurements, end diastolic diameter (EDD) and end systolic diameter (ESD) were determined for each contraction cycle, after which the following contraction parameters were calculated:  (4) where MaxD represents the maximum passive diameter obtained under Ca 2+ -free conditions at the same pressures at the end of the experiment. FREQ (min −1 ) was also calculated. For protocols using pharmacological inhibitors, the amplitude and frequency were sometimes normalized to the control values recorded prior to the first application of inhibitor as follows: Sharp electrode recording of Vm. Mouse IALs lymphatic vessels were isolated and pressurized as described above. Once the vessel had stabilized at a pressure of 3 cmH 2 O, wortmannin was added (1-3 μM) and allowed to incubate for 30-60 minutes to inhibit myosin light chain kinase 56 and blunt vessel movement associated with contractions to <5 μM, as assessed by edge-tracking of the vessel wall. Impalements into lymphatic smooth muscle were made using intracellular microelectrodes (300-350 MΩ) filled with 1 M KCl and membrane  ACT TGT GGG-3′  414   exon31  NM_009781  F: 5′-GAG CAT AAT TGA TGT CAT TC-3′  136  R: 5′-CTT CAC CAG GCG CAT GAC-3′   exon32  NM_001255997  F: 5′-TGT TGA TAT AGC AAT CAC C-3′  130  R: 5′-CTT CAC CAG GCG CAT GAC-3′   Table 1. Primers used in end-point PCR. F: forward primer -sense, R: reverse primer. potential (Vm) was sampled at 1 KHz using a NPI SEC-05x amplifier (ALA instruments, Farmingdale, NY). Once impaled, Vm was allowed to settle for 15-30 seconds until the membrane potential stabilized. Vm was corrected for the offset value recorded upon withdrawal of the microelectrode from the cell. Acquired data files were then processed and action potentials (APs) were measured and characterized using an in-house Python algorithm described previously 52 . Mean total integrated signal per AP and percent-integrated signals for each phase of an AP were computed for each trace consisting of three or more APs per vessel per mouse.
Single lymphatic smooth muscle cell isolation. After  25 °C) at 37 °C for 10 min, and then went through two digestion steps: (1) with 1 mg/mL dithioerythritol (Sigma, D8161-5G) and 1 mg/mL (mouse) papain (Sigma, P4762-1G) added to the low-Ca 2+ -digestion solution; (2) with 0.5 mg/mL collagenase H (Sigma, C8051-1G), 0.7 mg/mL collagenase F (Sigma, C7926-1G) and 1 mg/mL soybean trypsin inhibitor II (Sigma, T9128-1G) added to the low-Ca 2+ -digestion solution. After digestion, the remaining vessel fragments were rinsed with ice-cold digestion solution and triturated with a fire-polished Pasteur pipette to release single cells. Isolated muscle cells were stored in ice-cold solution to stop enzymatic digestion and used within 4-6 hr 21 . The resulting LMC cell homogenates were used either for Ca 2+ current recording with the patch clamp technique or collection of single LMCs for end-point PCR. For LMC collection, the cell suspension was placed in Krebs containing 0.1 mM CaCl 2 in a 0.5 mL chamber at room temperature and a 30 µm-pipette was used to collect single, spindle-shaped cells.
patch clamp recordings. LMCs isolated from the digestion of lymphatic vessels were placed in a recording chamber under an inverted microscope and perfused with physiological solution containing 1.8 mM Ca 2+ . Whole-cell patch clamp recording of current through VGCCs at room temperature was performed using an EPC9 amplifier (HEKA, Bellmore, NY, USA) controlled by Pulse software (HEKA) as described previously 57 ). Recording electrodes (resistance = 3-5 MΩ) were back-filled with pipette solution containing Cs + to block K + current. Pipette movement was controlled with a hydraulic manipulator (MO-102, Narishige, Tokyo, Japan). After the gigaseal was formed and stabilized in physiological solution, all subsequent recordings were performed in bath solution containing 10 or 20 mM BaCl 2 to enhance current through Ca v 1.2/Ca v 3×. Cell capacitance, an indication of cell membrane surface area, ranged from 9-25 pF. The recorded current value (pA) was normalized to cell capacitance (pF) for each cell to obtain current density (pA/pF). Holding potential was −70 mV. Two voltage clamp protocols were subsequently performed: (1) a voltage ramp, from −100 to +80 mV at 1 mV/ms, to rapidly obtain current-voltage relations; (2) voltage steps from a holding potential of −70 mV with a prepulse to −90 mV for 200 ms, followed by steps to a range of voltages from −80 to +40 mV (in 10 mV-increments, 2 s each), and a final step back to the holding potential; this protocol provided a measurement of current kinetics in response to different voltage steps 11 . Solutions and chemicals. Lymphatic vessels were isolated and cannulated in 0.5% albumin-supplemented Krebs buffer; the cannulating pipettes contained the same solution. Pressure myograph experiments were conducted in Krebs buffer without albumin. Krebs buffer contained (in mM): NaCl, 146.9; KCl, 4.7; CaCl 2 , 2; MgSO 4 , 1.2; NaH 2 PO4.H 2 O, 1.2; NaHCO 3 , 3; Na-HEPES, 1.5; D-glucose, 5 (pH 7.4, 37 °C). For patch clamp recordings, the physiological Ca 2+ bath solution contained (in mM): 1.8 CaCl 2 , 110 NaCl, 1 CsCl, 1.2 MgCl 2 , 10 HEPES, 10 D-glucose (pH adjusted to 7.4 with NaOH). Ba 2+ bath solution contained (in mM): 10 or 20 BaCl 2 , 110 NaCl, 1 CsCl, 1.2 MgCl 2 , 10 HEPES, 10 D-glucose (pH adjusted to 7.4 with NaOH). Patch pipettes contained (in mM): 135 CsCl, 10 HEPES, 10 EGTA, 5 GTP·Mg, (pH adjusted to 7.2 with CsOH). All chemicals were obtained from Sigma (St. Louis, MO, USA), except for BSA (US Biochemicals; Cleveland, OH, USA), MgSO 4 , HEPES (Fisher Scientific; Pittsburgh, PA, USA), and TTA-A2 (Alomone, Israel). TTA-A2, nifedipine and Bay K8644 were dissolved in DMSO. Ni 2+ , mibefradil and acetylcholine were dissolved in Krebs buffer (pre-warmed for Ni 2+ ) without BSA. DMSO by itself, at its maximal concentration (0.033%) had no effect on contractions. end-point pcR. End-point PCR experiments detected message for VGCCs in mouse whole PLs and IALs and in LMCs freshly isolated from either of the two types of vessels. Vessels used for PCR were dissected at 4 °C. Cleaned lymphatic vessels or collected LMCs were directly transferred to RNAlater solution (Qiagen, Hilden, Germany) and stored at −80 °C. RNA was pooled from 100-300 LMCs collected by micropipette aspiration of spindle-shaped cells digested from 6-16 lymphatic vessels (from three or more mice). RNA was extracted using the Arcturus PicoPure RNA Isolation Kit (Life Technology, Carlsbad, CA, USA). cDNA then was synthesized using the Superscript III first Strand synthesis system (Invitrogen, Carlsbad, CA, USA). PCR was subsequently performed using GoTaq Flexi Reverse Transcriptase (Promega, Madison, WI, USA). Nested PCR was performed for weak signals such as Ca v 3.1, Ca v 3.2 and Ca v 3.3. Nested PCR used a second internal set of primers to amplify the PCR product of the first external set of primers. The amplified PCR products were loaded onto 1.5-2% Agarose gel (Invitrogen) + Gel Red (1:10000, Biotum, Fremont, CA, USA) and the gel was run at 70-80 V for 60-90 min.
immunostaining. IALs were isolated and transiently pressurized to ~8 cmH 2 O and stretched axially to remove slack. The vessels were fixed in 2% paraformaldehyde (v/v) (19943-1 L, Affymetrix, Cleveland, OH, USA) for 30 minutes while pressurized to 3 cmH 2 O. The vessels were then washed in Phosphate Buffered Saline (PBS, 21600-010, Thermo Fisher Scientific, Waltham, MA, USA), and permeabilized with 0.3% Triton-X (v/v) (32737-1000, Acros Organics, Springfield Town, NJ, USA) for 30 minutes and blocked with 5% donkey serum (v/v) for 1 hr at room temperature. The vessels were incubated in α-actin primary antibody (Sigma, A2547) (1:200) and one of the primary antibodies for VGCCs (Alomone, Jerusalem, Israel): Ca v 3.1 (#ACC-021), Ca v 3.2 (#ACC-025) (1:100), Ca v 1.2 (#ACC-003) diluted in PBS containing 5% donkey serum (017-000-121, Jackson ImmunoResearch www.nature.com/scientificreports www.nature.com/scientificreports/ Laboratories, West Grove, PA, USA), and 0.1% Triton-X (AC2156825000, Thermo Fisher Scientific). The vessels were washed with PBS at least 3 times for 20 min and incubated with an appropriate secondary antibody diluted 1:200 in PBS containing 5% donkey serum and 0.1% Triton-X for 1 h at room temperature. The secondary antibodies used were goat anti-mouse antibody Alexa Fluor 647 (A21241, Invitrogen, Carlsbad, CA, USA) and donkey anti-rabbit Alexa Fluor 448 (A21206, Invitrogen). The vessels were subsequently washed again with PBS at least 3 times for 20 min. Subsequently, ProLong ® Gold Antifade Reagent with DAPI (P36931, Thermo Fisher Scientific) was used in some cases to stain nuclei and to mount the vessels onto imaging slides. The mounted vessels were covered with a #1 glass coverslip, allowed to cure overnight and sealed the following day with nail polish. Ca v 1.2 staining served as a positive control and α-actin co-staining was used to outline the structure and location of LMCs. Negative controls included omission of Ca 2+ channel primary antibody or primary antibody pre-absorption using the corresponding antigens (provided by Alomone). The vessels were recannulated and repressurized before collection of fluorescence images. Fluorescence emission was imaged using an Olympus IX-81 microscope equipped with an Andor/Yokogawa spinning disk confocal head and Hamamatsu Flash 4.0 camera (Hamamatsu Photonics, Shizuoka, Japan), with illumination from diode lasers (Andor, Belfast, Northern Ireland) using Metamorph (Nashville, TN, USA). Images were processed with NIH ImageJ. Data analysis. All data were analyzed using Prism 5 (Graphpad, La Jolla, CA, USA). Two-way ANOVAs with Bonferroni's post hoc tests were performed to compare contractile function parameters between transgenic and WT vessels at different concentrations during concentration-response curve protocols. One-way ANOVAs were performed to compare pharmacological treatments with control parameters in the same vessels. P values less than 0.05 were considered statistically significant. The data are expressed as mean ± standard error of the mean.

Results
Effects of mibefradil and Ni 2+ on rat mesenteric lymphatics. The goal of this study was to test the hypothesis that T-type VGCCs play an important role in lymphatic vessel pacemaking. Prior to performing experiments on mouse lymphatic vessels, we sought to confirm the expression of Ca v 3.1 and Ca v 3.2 in rat mesenteric LMC and the effects of the T-type VGCC inhibitors mibefradil and Ni 2+ on rat mesenteric lymphatics as reported by Lee et al. 35 . To assess contractile parameters of lymphatic collectors under controlled conditions (fixed transmural pressure, absence of luminal flow), single-valve segments of rat mesenteric lymphatics were studied ex vivo using established pressure myograph methods 55 . All vessels used for further experimentation developed robust spontaneous contractions when pressurized to 3 cmH 2 O at 37 °C. With pressure maintained at that level, mibefradil was added to the bath in cumulative concentrations while assessing its effects on contraction amplitude and frequency for 2 min at each concentration. A representative recording of a spontaneously contracting rat mesenteric lymphatic during mibefradil application is shown in Fig. 1A. In this recording, Mibefradil slowed the contraction frequency, starting at concentrations below 1 nM and reaching a maximum effect at ~20 nM, but FREQ partially recovered at concentrations of 50 and 100 nM. Contraction amplitude was remarkably constant over the concentration range 1-100 nM, but spontaneous contractions stopped completely at 200 nM. The summary data in Fig. 1B,C reveals the same pattern of contractile regulation for 8 rat vessels, with a gradual reduction in FREQ occurring at all concentrations but only being significantly different from control at concentrations >10 nM. In contrast, there was a trend for AMP to increase slightly up to mibefradil concentrations of 100 nM, above which it fell precipitously. All vessels stopped contracting at the higher concentrations of mibefradil and the large error bars for the points at concentrations between 50-200 nM reflect the fact that some vessels stopped contracting at slightly different concentrations than others. The IC 50 of mibefradil for AMP was 372 nM and the IC 50 for FREQ was 56 nM. The lower IC 50 for FREQ is consistent with the results of Lee et al. 35 , who concluded that mibefradil selectively inhibited FREQ, but not AMP, of rat mesenteric lymphatics, although those authors tested only a single concentration (100 μM) of mibefradil.
Complete concentration-response curves for Ni 2+ were also performed on rat mesenteric lymphatics. With pressure maintained at 3 cmH 2 O, Ni 2+ was added to the bath in increasing concentrations while recording its effects on contraction AMP and FREQ for 2 min at each concentration. A representative recording of a rat mesenteric lymphatic responding to Ni 2+ is shown in Fig. 1D. FREQ began to slow when Ni 2+ reached a concentration of 15 μM and slowed further at concentrations of 20 and 25 μM, then increased somewhat at 50 and 100 μM before slowing further at 150 µM and stopping completely at 200 μM. In contrast, AMP declined progressively throughout the entire concentration range of Ni 2+ , beginning at ~15 μM. The summary data for 8 vessels in Fig. 1E,F showed a similar response pattern: a progressive reduction in AMP with increasing concentrations of Ni 2+ , becoming significantly different from control above concentrations of 50 μM. FREQ exhibited a biphasic concentration-response relationship to Ni 2+ , with significant inhibition at concentrations between 10-40 μM, a rebound at concentrations between 50-105 μM followed by further inhibition and complete cessation of contractions at concentrations >200 μM. The IC 50 for Ni 2+ on AMP was 227 μM and, although the FREQ data could not be fit to the Hill equation, the interpolated IC 50 for Ni 2+ on FREQ was 287 μM. These results are somewhat, but not completely, consistent with those of Lee et al. 35 . The inhibition of FREQ by low concentrations of Ni 2+ could be explained by inhibition of Ca v 3.2 channels 23,33 ; however, the complete inhibition of contraction by both drugs at intermediate-to-high concentrations suggests off-target effects on L-type VGCCs, which would be consistent with previous reports 14, 58 . What is not clear is the extent to which either or both, Ni 2+ and/or mibefradil, might have inhibited L-type VGCCs at lower concentrations. These uncertainties point to the need for using a different method to inhibit T-type VGCCs, which is the reason we turned to genetic approaches in the mouse.
VGcc isoform expression in mouse lymphatics. Before testing the possible role of T-type VGCCs in the spontaneous contractions of mouse lymphatics, we examined which Ca v 3 isoforms were expressed in mouse lymphatic vessels. We tested two types of collecting lymphatic vessels from different regions of the mouse, based (2020) 10:70 | https://doi.org/10.1038/s41598-019-56953-3 www.nature.com/scientificreports www.nature.com/scientificreports/ on the following reasoning. Popliteal afferent lymphatics (PLs) previously have been demonstrated by our laboratory 55 and others 59 to exhibit robust spontaneous contractions ex vivo comparable in magnitude to those from other species 60 ; further, these contractions are modulated by pressure in the same way, and over approximately the same range, as those of collecting lymphatic vessels from other species, in particular rat mesentery 61 . In contrast, the IAL is an efferent vessel that demonstrates strong spontaneous contractions 62 but is larger, easier to clean and cannulate, and more amenable to electrophysiology studies. PL and IAL vessels were excised from WT (C57BL/6J) mice and thoroughly cleaned of fat and connective tissue. RNA was extracted and end-point PCR performed on single (whole) vessels, 2-3 mm in length. Message for Ca v 1.2 was detected in both types of vessels. Message for Ca v 3.1 and Ca v 3.2 was detected in PLs, and message for all three Ca v 3 isoforms was detected in IALs (Suppl. Fig. S1). However, Ca v 3.3 was not detected by immunostaining (Suppl. Fig. S11). Because analysis of mRNA obtained from whole lymphatic vessels includes mRNA from multiple cell types, including LMCs, LECs, dendritic cells 63,64 , macrophages, mast cells 65,66 , lymphocytes, and possibly neuronal axons and terminals, we could not be confident which VGCC isoforms are actually expressed by LMCs per se. Therefore, we repeated the procedure with the additional step of enzymatically digesting the vessels with papain and collagenase followed by vigorous trituration in a fire-polished Pasteur pipette to release single cells. Spindle-shaped cells, presumably representing LMCs, were then collected into a 20-μm tip micropipette using suction from a micrometer-controlled glass syringe. The collected cells (~100-300 per animal) were deposited into a clean glass chamber, pooled, RNA was extracted, and PCR performed with the same primer sets used above. LMCs from both vessel types expressed message for Ca v 3.1, Ca v 3.2 and Ca v 1.2 (Fig. 2). The absence of Ca v 3.3 in both purified LMC preparations suggests that it was contributed by other cell types in the initial whole-vessel PCR analyses. The PL sample continued to show a band for eNOS, suggesting some contamination from LECs; however, the IAL sample did not. Because previous studies suggest that T-type VGCCs might be expressed in endothelium 67 , we checked for Ca v 3 expression specifically in LECs by performing additional PCR analyses on LEC tubes isolated from PLs. Vessels were dissected and cleaned but then subjected briefly to collagenase (alone) treatment and gentle trituration using a micropipette with a tip diameter just larger than that of the vessel. This procedure has been used consistently to isolate pure EC tubes from arterioles 68 and we adapted it recently for lymphatic vessels 69 ). Analysis of LEC tubes obtained from PLs showed a prominent band for eNOS, no band for SMA and no message for Ca v 3.1 or Ca v 3.2 (Suppl. Fig. S1D). The LEC tube sample failed to show message for Ca v 1.2 (not shown).
Expression levels of Ca v 3 isoforms were too low and sample sizes too small to make Western blotting feasible. Therefore, to test for protein expression, we performed immunostaining for Ca v 3.1 and Ca v 3.2 using IALs from WT mice. Both Ca v 3 isoforms showed prominent signals in the green fluorescence channel for the secondary antibody that appeared to substantially colocalize with the signal for SMA in the red channel (Fig. 3), albeit with some non-specific staining. For negative controls, WT vessels were stained using the same primary and secondary antibodies and image acquisition/analysis settings after preabsorption with the respective antigenizing peptides (Suppl. Figs. S2 and S3); only non-specific background signal, which did not colocalize with SMA, was detected in each type of vessel. As additional controls, we stained IALs from Ca v 3. tests for functional ca v 3 channels in mouse LMCs. To test for the expression of functional Ca v 3 channels, we enzymatically isolated LMCs from IALs or PLs and performed whole-cell patch clamp experiments using solutions and conditions designed to maximize VGCC current. 20 mM Ba 2+ or Ca 2+ was used as the charge carrier in the bath solution; outward K + currents, that otherwise would have masked smaller currents through VGCCs, were blocked by dialyzing the cells with 135 mM Cs + solution. After enzyme treatment and trituration, most LMCs were easily distinguished as spindle shaped cells, with slightly rounded ends (Fig. 4A), but only a minority of such cells produced recordings with measurable (>20 pA per cell) inward divalent cation current. We confirmed the spindle-shape morphology of LMCs by using the same isolation procedure on lymphatic vessels from SMMHC GFP mice and verifying the typical spindle shape of a GFP + cell. However, isolation of cells from Prox1 GFP vessels also yielded LECs that were spindle-shaped, perhaps explaining the low success rate in finding cells with measurable inward divalent cation current and accounting for residual eNOS message in spindle-shaped cells collected from popliteal lymphatics via micropipette aspiration for PCR ( Fig. 2A).
The first patch clamp protocol was designed to evoke the opening of VGCCs using a voltage ramp from −100 to +80 mV from a holding potential of −70 mV (Fig. 4B). A hyperpolarizing step to −100 mV was expected to reset T-type VGCCs, if present, so they could contribute to inward current during the depolarizing ramp. As shown in the example in Fig. 4C, the (current-voltage) I-V curve produced by this protocol in a LMC isolated from a mouse IAL shows inward current activating at ~−20 mV and peaking at ~+20 mV. Such curves are expected to be right-shifted by 20-30 mV due to both the use of Ba 2+ as the charge carrier and the use of a concentration that is higher than physiological levels (2 mM 72 ). Attempted recordings in 2 mM Ca 2+ bath solution typically produced inward currents that were <5 pA; for this reason, we were unable to determine the I-V curve in physiological Ca 2+ solution, which would have been useful for assessing the window currents for T-(and L-) type VGCCs. The application of nifedipine during the recording in Fig. 4C reduced peak inward current from −80 pA to −20 pA and shifted the peak from +22 mV to +5 mV; this shift is consistent with unmasking of a T-type VGCCC current 21,70 . In this cell, the remaining (nifedipine-insensitive) current was blocked ~60% by the addition of 1 mM Ni 2+ (lower Ni 2+ concentrations produced less block), leaving some residual current for which the peak shifted right-ward to a more positive voltage; the residual current may reflect the presence of another Ca v isoform (although it is unlikely to be Ca v 1.3, see RT-PCR in Suppl. Fig. S1A-C) or other Ca 2+ -permeable cation channel(s). The I-V curve for 5 LMCs from IALs is summarized in Fig. 4D and shows the same pattern as described for the recording in panel C. Recordings from LMCs isolated from mouse popliteal lymphatics and from rat mesenteric lymphatics also showed the same type of inward currents with similar I-V relationships and nifedipine and Ni 2+ sensitivities, but with smaller whole-cell currents (Suppl. Fig. S4A,B).
Because some inactivation of VGCCs is likely to occur during a voltage ramp, we also used voltage step protocols, as depicted in Fig. 4E, stepping first to −90 mV to relieve inactivation, followed by depolarizing steps from −80 to +40 in 10 mV increments in successive sweeps. Fig. 4Fa shows a typical recording, with each of the resultant current traces overlaid. Comparison of the control traces (in the absence of inhibitor) to those with 1 μM nifedipine (Fig. 4Fb) revealed a transient, rapidly-inactivating inward current that was nifedipine-insensitive (arrow); in this cell the transient current was completely blocked by 1 mM Ni 2+ (Fig. 4Fc). The individual current traces for the same cell are aligned vertically in Fig. 4G to reveal that the transient current begins to activate at −20 mV, is the only inward current activated at −10 and 0 mV, and then is masked by a longer-lasting, slowly-inactivating inward current at more positive step potentials. These transient currents were not blocked by TTX (not shown). Collectively, the kinetics and nifedipine-resistance of this current are consistent with those of Ca v 3 channels, in agreement with a preliminary report by Hollywood et al. 73 using sheep LMCs.

Effects of Ni 2+ and TTA-A2 on mouse lymphatic contractions. Having established that functional
T-channels are present in mouse lymphatic smooth muscle, we then tested the effects of Ni 2+ on the spontaneous contractions of PLs and IALs. A representative recording of spontaneous contractions in a PL during the application of increasing concentrations of Ni 2+ is shown in Fig. 5A. Ni 2+ produced a concentration-dependent reduction in AMP at all concentrations, but a slight increase in FREQ at concentrations up to 250 μM and then a reduction in FREQ at 500 μM before inhibiting contractions completely at 1 mM. Summary data for Ni 2+ on PL contractions showed the same patterns (Fig. 5B,C). Interestingly, the IC 50 for Ni 2+ on AMP was 97 μM but on FREQ was 452 μM, which is opposite to the predictions and conclusions of Beckett et al. 8 and Lee et al. 35 . A representative concentration-response curve for an IAL is shown in Fig. 5D. Increasing concentrations of Ni 2+ led to progressive reductions in AMP and increases in FREQ until at 1 mM, spontaneous contractions ceased completely for ~45 sec, followed by partial recovery. Summary data for Ni 2+ on 8 IALs showed the same pattern: a progressive reduction in AMP with a concomitant increase in FREQ until contractions stopped at a concentration of 2 mM (Fig. 5E,F). We also tested a recently developed and reputedly more selective inhibitor of Ca v 3 channels, TTA-A2, which has been reported to be effective on Ca v 3 channels at concentrations of 1 µM 74 . However, TTA-A2 had only a modest effect on both PLs and IALs, inhibiting AMP by ~20% at a concentration of 100 μM and having no significant effect of FREQ at any concentration (Suppl. Fig. S5). IC 50 values for Ni 2+ , mibefradil and TTA-A2 on the various lymphatic preparations are listed in Table 2. In summary, the inconsistent effects of mibefradil and Ni 2+ on AMP and FREQ of rat mesenteric lymphatics and of Ni 2+ and TTA-A2 on mouse PLs and IALs highlight the importance of using more selective methods, e.g. genetic approaches, to assess the role(s) T-type VGCCs in lymphatic contraction.  Fig. 6C-E. This summary analysis confirmed a lack of consistent, significant differences in AMP, FREQ or calculated FPF between WT and Ca v 3.2 −/− PLs. A similar analysis is shown for IALs from WT and Ca v 3.2 −/− mice in Fig. 7. Again, there were no significant differences between vessels from the two genotypes, except that AMP was higher in Ca v 3.2 −/− IALs at a single pressure, 0.5 cmH 2 O (Fig. 7C). Additional contractile data for each of the two genotypes (EDD, EF, tone) are shown in Suppl. Fig. S6 and also reveal no significant differences for either vessel type compared to their respective control vessels. Because the penetrance of some transgenic traits can be strain-specific, we obtained an additional strain of Ca v 3.2 −/− mice on a mixed background (mostly C129SVE) from the MMRRC (Columbia, MO) and tested the responses of six PLs to pressure. The representative sample traces and summary data show that the contraction patterns were nearly identical to those of WT PLs and Ca v 3.2 −/− PLs on the C57BL/6J background (Suppl Fig. S7A-C).
Because Ca v 3.2 activation has been shown to be linked to calcium spark activity, smooth muscle hyperpolarization and negative feedback inhibition of tone in cerebral and mesenteric arteries 23,33 , we tested whether the known effects of vasodilation and inhibition of lymphatic pumping produced by ACh would be altered in Ca v 3.2 −/− lymphatics. PLs were isolated, cannulated and pressurized to 3 cmH 2 O. After spontaneous contractions developed, ACh was added to the bath in cumulative concentrations. ACh caused a concentration-dependent reduction in both AMP and FREQ of WT PLs (Suppl. Fig. S8A), typically beginning at 1-3 nM, which is consistent with previously published data 55 . Nearly identical inhibitory responses were observed for PLs from Ca v 3.2 −/− mice (Suppl. Fig. S8B). Summary data are shown in Suppl. Fig. S8C-F, where FREQ is expressed both in absolute values and normalized to the respective control values in the absence of ACh. No significant differences were evident between the two data sets nor for the additional Ca v 3.2 KO strain (Suppl. Fig. 7D-F). The trend for FREQ and normalized FREQ to increase at the highest concentration (Suppl Fig. 8B) reflects a stimulatory action of ACh 55 on the smooth muscle layer of some vessels (even higher concentrations would have stimulated all vessels). Collectively, these results strongly suggest that Ca v 3.2 does not play a detectable role in the regulation of either the FREQ or AMP of spontaneous contractions by pressure, or for ACh-mediated inhibition of lymphatic contractions in mouse PLs and IALs.  www.nature.com/scientificreports www.nature.com/scientificreports/ There were no significant differences in AMP, FREQ or FPF at any pressure between WT and Ca v 3.1 −/− vessels. These data are inconsistent with a critical contribution of Ca v 3.1 channels to the FREQ or AMP of spontaneous contractions. We also tested ACh responses of PLs from Ca v 3.1 −/− mice because that isoform in smooth muscle has been implicated as a target of endothelium-derived NO 21,75 ; the results are shown in Suppl. Fig. S9. There were no significant differences between the two groups in the effectiveness of ACh to inhibit spontaneous contractions, although there was a trend for the AMP of Ca v 3.1 −/− vessels to be more resistant to ACh (Suppl. Fig. S9C). Collectively, the results in Fig. 8 and Suppl Fig. 9 contradict the hypothesis that Ca v 3.1 channels are important In some knock-out mouse models, deletion of one protein isoform is known to be compensated by upregulation of another isoform of that protein [76][77][78][79] . Compensation becomes more likely in global knock-out animals because it can occur gradually during development. We considered the possibility that deletion of Ca v 3.2 could be compensated by upregulation of Ca v 3.1 and vice-versa. Therefore, we generated mice deficient in both Ca v 3 isoforms by breeding Ca v 3.2 −/− mice with Ca v 3.1 −/− mice to obtain offspring deficient in both isoforms. PLs and IALs from Ca v 3.1 −/− ; Ca v 3.2 −/− double KO mice were then studied at pressures from 0.5 to 10 cmH 2 O and the results are summarized in Fig. 9. Again, there were no significant differences in AMP, FREQ or FPF between Ca v 3.1 −/− ; Ca v 3.2 −/− double KO vessels compared to WT controls for either PLs or IALs. We then tested the ACh responses of PLs from Ca v 3.1 −/− ; Ca v 3.2 −/− mice and found no significant or consistent differences between the two groups at any ACh concentration (Suppl Fig. S10). Collectively, these results contradict the hypothesis that either Ca v 3.1 or Ca v 3.2 channels are critical regulators of spontaneous lymphatic contractions or ACh-mediated contraction inhibition.  Table 2. Summary of IC 50 s for pharmacological inhibitors of T-type VGCCs on lymphatic contraction amplitude and frequency. nd = not-determined.  35 . Therefore, we performed protocols similar to theirs, in which the contractile parameters of mouse PLs were measured at each pressure between 0.5 and 10 cmH 2 O before and after application of Ni 2+ (100 μM) or mibefradil (100 nM)-the same concentrations used in their study. The prediction based on their results is that both inhibitors would reduce frequency at most or all pressures but have no effect on amplitude at any pressure. The Ni 2+ data are shown in Fig. 10 and the mibefradil data are shown in Fig. 11. Contrary to the prediction, our results show that in WT lymphatics both Ni 2+ and mibefradil inhibited contraction amplitude at all pressures, whereas Ni 2+ increased frequency at all pressures and mibefradil increased frequency at most pressures (Figs. 10A,B and  11A,B, respectively). The reduction in amplitude with an increase or no change in frequency points to a primary (or exclusive) effect of both inhibitors on L-type channels in mouse lymphatics.
We also repeated these protocols on PLs from Ca v 3.1 −/− ; Ca v 3.2 −/− double KO mice, reasoning that, if T-type channels were important and the inhibitors were specific for those channels, we would see inhibition of pacemaking frequency in WT vessels but not in Ca v 3.1 −/− ; Ca v 3.2 −/− vessels. However, as shown in panels C-D of Figs. 10 and 11, the effects of Ni 2+ and mibefradil were very similar in both WT and Ca v 3.1 −/− ; Ca v 3.2 −/− PLs; moreover, the most consistent inhibitory effect was on amplitude not frequency. Because the effects of both inhibitors were nearly identical in both genotypes this again suggests primary actions on L-type channels (vessels from both genotypes have L-type channels, but only WT vessels have T-type channels).
Contractile responses of lymphatics from smooth muscle-specific Ca v 1.2 KO mice. Next, we hypothesized that deletion of L-type (Ca v 1.2) channels in the lymphatic smooth muscle layer would unmask a role for Ca v 3 channels such that Ca v 3-mediated spontaneous contractions, perhaps weaker ones, would become evident. We reasoned that such events might not be observed in previous studies using nifedipine or other DHPs, even at low concentrations, if those inhibitors had partial inhibitory actions on Ca v 3 channels. Since global deletion of Ca v 1.2 is lethal 14 , a conditional tissue-specific knockout strategy was needed. Previously, deletion of Ca v 1.2 in smooth muscle, using an inducible SM22α-Cre, resulted in mice with lower blood pressure and impaired myogenic constriction in arteries and arterioles 14 . In addition those mice began to die at ~15 days post www.nature.com/scientificreports www.nature.com/scientificreports/ induction due to a severe loss of GI motility and intestinal impaction 14 . However, those studies did not examine any aspect of lymphatic function. We first checked if the smooth muscle isoform of Ca v 1.2 was the predominant, or only, isoform expressed by mouse lymphatic vessels. Ca v 1.2 was expressed in lymphatic smooth muscle based on both PCR (Fig. 2, Fig. S1C) and immunostaining (Suppl. Fig. S11). As shown in Suppl. Fig. S12, Ca v 1.2a was highly expressed in mouse heart and Ca v 1.2b was expressed in mouse thoracic aorta and IALs; Ca v 1.2b was also expressed in brain and heart, presumably due to contamination by blood vessels (panel A). Preliminary analyses failed to reveal possibly unique splice variants of Cav1.2b in lymphatic vessels. The common splice variant exon 9 was found in all samples (panel B), exon 8a and 8b were found in one of the two IAL samples (panel C) and exons 31 and 32 were found in thoracic aorta but in neither of the lymphatic vessel samples (panel D).
We generated our own smooth muscle-specific, inducible Ca v 1.2 (Ca v 1.2b) KO mouse by crossing SMMHCCreER T2 mice to Ca v 1.2(Ca v 1.2b) flox mice. The tamoxifen induction protocol (daily i.p. injection for 5 days) was optimized first using SMMHCCreER T2 ; Rosa26 mTmG reporter mice and revealed >99% recombination in LMCs (not shown). Induction of SMMHCCreER T2 ; Ca v 1.2 fl/fl mice produced animals that began to die between days 12-15 from weight loss and impaction of the intestinal tract. PLs from these mice were therefore studied ex vivo 10-12 days after induction. PLs never exhibited any spontaneous contractions larger than 5 μM, even over observation periods of 3-4 hours. Changing pressure over the range 0.5 and 10 cmH 2 O also failed to evoke robust, spontaneous contractions. Example recordings are shown in Fig. 12A,B, where the control vessel (panel A) is from a Ca v 1.2 f/f mouse (without Cre) that was injected with tamoxifen for 5 days. The control vessel exhibited robust spontaneous contractions at all pressure levels in contrast to the complete lack of contractions in the SMMHCCreER T2 ; Ca v 1.2 fl/fl vessel (panel B). Examination of diameter tracings from the latter vessels at higher resolution showed small, high-frequency fluctuations (panel C) resembling the vasomotion often observed in WT arterioles [80][81][82] . However, these fluctuations were also observed in PLs from control (both WT and Ca v 1.2 f/f ) mice and in vessels from Ca v 3.1 −/− , Ca v 3.2 −/− , and Ca v 3.1 −/− ; Ca v 3.2 −/− mice (not shown) during the diastolic period of the lymphatic contraction cycle. Fluctuations were more obvious in the smooth muscle-specific Ca v 1.2 KO mice due to absence of superimposed strong contractions. Calcium imaging of the smooth layer suggested that the small fluctuations corresponded to calcium waves in LMCs (not shown), similar to what has been reported for arteriolar smooth muscle 83,84 . The localized, non-propagating nature of these small diameter fluctuations also distinguished them from the large-amplitude, propagating contractions of WT  Fig. 12D-G, where AMP, FREQ and FPF are zero at all pressures, in contrast to the corresponding parameters from 7 PLs from Ca v 1.2 fl/ fl mice. Contractions for this (and all other) analyses of FREQ were defined as propagating events larger than 5 μM in amplitude. Interestingly, tone was ~2.5-fold higher in these mice, which is opposite to that reported for arteries from SM22aCreER T2 ; Ca v 1.2 fl/fl mice 14 . The Ca v 1.2 antagonist nifedipine did not alter the level of tone (1 μM, tested at 3 cmH 2 O), nor did the Ca v 1.2 agonist BayK8644 (1 μM). Neither did Bay K8644 reverse the lack of contractions in smooth muscle-specific Ca v 1.2 KO mice. In contrast, contractions of (control) Ca v 1.2 fl/fl vessels from mice treated with tamoxifen were slightly enhanced in amplitude by BayK8644 and completely inhibited by nifedipine (Suppl. Fig. S13). and be evident in Ca v 3-deficient vessels as a reduction in the peak or upstroke velocity of the AP. To test this hypothesis sharp electrode recordings of membrane potential (Vm) were made in the LMC layer of pressurized IALs from WT and Ca v 3.1 −/− ; Ca v 3.2 −/− double KO mice; wortmannin treatment minimized the contraction amplitude so that the electrode would not be dislodged 10 . We previously reported that APs in LMCs from WT mice appear to be initiated around −30 mV by a gradual diastolic depolarization, peak around −5 mV, followed by a plateau lasting 1-2 secs at ~−10 mV 52 . As shown by the example recordings in Fig. 13A,B, the APs from LMCs of WT and Ca v 3.1 −/− ; Ca v 3.2 −/− vessels were remarkably similar in appearance. Because the peak of the AP and the upstroke velocity are sensitive to even partial inhibition of whatever voltage-gated cation channel carries current (Na v , Ca v 3, Ca v 1, etc.), we compared those parameters, and others, of the AP in lymphatic smooth muscle from WT and Ca v 3.1 −/− ; Ca v 3.2 −/− vessels (Fig. 13C-J). No parameters of the AP or resting Vm were significantly different between WT and Ca v 3.1 −/− ; Ca v 3.2 −/− vessels. These results suggest that Ca v 3 channels do not significantly affect the upstroke velocity or the peak of the AP (i.e. characteristics of the spike), which is what might have been expected. Diastolic depolarization rate tended to be lower in Ca v 3.1 −/− ; Ca v 3.2 −/− than WT vessels, but that change did not result in a significant change in the pacemaking frequency.

Discussion
The aim of this study was to test the hypothesis that T-type VGCCs play an important role in modulating the frequency of lymphatic vessel pacemaking. We confirmed the expression of two Ca v 3 channel isoforms (Ca v 3.1, Ca v 3.2) by RT-PCR in rat mesenteric lymphatic smooth muscle. We confirmed previous findings that the FREQ, but not AMP, of spontaneous contractions of rat mesenteric lymphatics was sensitive to inhibition by mibefradil, although we were unable to confirm a similar, selective inhibitory effect of Ni 2+ on the FREQ of rat mesenteric lymphatics. In the mouse, we demonstrated the expression of Ca v 3.1 and Ca v 3.2 in popliteal and IAL lymphatic smooth muscle by RT-PCR and immunofluorescence and verified the presence of functional VGCCs by patch clamp protocols, in which the kinetics and nifedipine resistance of whole-cell Ca 2+ currents were consistent with those of Ca v 3 channels. In both popliteal and IALs of the mouse, Ni 2+ had equivalent inhibitory effects on both the FREQ and AMP of spontaneous contractions. The equivalent IC 50 values for Ni 2+ , coupled with the likelihood that L-type channels were also inhibited at those concentrations, led us to explore genetic deletion methods to more selectively test the role of Ca v 3 channels in lymphatic smooth muscle. We tested two previously-documented global knock-out models, Ca v 3.1 −/− and Ca v 3.2 −/− mice, and also generated the first double KO Ca v 3.1 −/− ; Ca v 3.2 −/− mouse, assessing the contractile properties of both popliteal and IAL vessels in each mouse model. We also generated and tested a smooth-muscle specific KO of the L-type calcium channel, Ca v 1.2. Spontaneous contractions were completely absent from lymphatic vessels from Ca v 1.2-deficient mice, and they could not be evoked by pressure changes. This finding indicates that Ca v 1.2 channels are obligatory for spontaneous twitch contractions. In contrast, none of the contractile parameters tested, including FREQ and AMP, were significantly different in any of the three Ca v 3 KO mouse strains compared to those of WT vessels. This finding held up for both popliteal and IAL vessels at all pressures tested (from 0.5 to 10 cmH 2 O), with the only exceptions being an accentuated AMP at a few selective pressures in some Ca v 3-deficient vessels. Comparisons of the effects For this analysis we did not count contractions less than 5 μm in amplitude at pressures ≤ 3 cmH 2 O because they were non-propulsive and usually not highly entrained. Counting these contractions would have led to even higher determinations of frequency at those pressures.
www.nature.com/scientificreports www.nature.com/scientificreports/ of channel deletion with the effects of Ni 2+ and mibefradil suggest that previous conclusions concerning a role for Ca v 3 channels in LMCs using so-called selective T-type channel inhibitors likely are explained by off-target effects on L-type calcium channels. This conclusion is further strengthened by the results of protocols in which we compared the effects of reputedly selective concentrations of mibefradil and Ni 2+ on WT and Ca v 3.1 −/− ; Ca v 3.2 −/ lymphatic vessels. Ideally, if mibefradil and Ni 2+ selectively inhibited T-type channels, neither compound would have an effect on Ca v 3.1 −/− ; Ca v 3.2 −/− vessels (which lack T-type channels), but they would decrease pacemaking frequency in WT vessels without any change in amplitude. However in both genotypes both compounds produced a profound impairment of contraction amplitude and a slight increase in pacemaking frequency at almost all pressures. This result further reinforces our conclusion that these compounds primarily or exclusively inhibit L-type VGCCs in mouse lymphatic vessels. Finally, we measured membrane potentials and action potentials in the smooth muscle layer of mouse IAL vessels and found slight, but not significant, differences in the resting Vm, the duration of the AP plateau and the rate of diastolic depolarization and upstroke velocity in Ca v 3 DKO vessels compared to WT vessels. We conclude that these differences in the characteristics of the AP are too subtle to result in significant differences in the amplitude or frequency of spontaneous lymphatic contractions.
Our work is the first report of a lymphatic phenotype in smooth-muscle-specific Ca v 1.2 KO mice. We engineered this mouse to avoid possible and subtle off-target effects of nifedipine on Ca v 3 channels, which are known to occur with recombinant Ca v 3 channels 85 . Lymphatic vessels from smooth muscle-specific Ca v 1.2 KO mice have no spontaneous, propagated contractions whatsoever and contractions could not be stimulated by multiple interventions, including input and/or output pressure elevation. Thus, in the absence of Ca v 1.2 channels, Ca v 3 channels are insufficient to evoke APs and propulsive contractions in mouse LMCs, which is further confirmation of a dominant role for Ca v 1.2 channels. Interestingly, lymphatic tone in smooth muscle-specific Ca v 1.2 KO vessels was elevated 3-fold over that of control vessels, which is opposite to that observed in arteries of SM22aCreER T2 ; Ca v 1.2 f/f mice 14 ; the subsequent elimination of that tone in Ca 2+ free Krebs solution implies that lymphatic www.nature.com/scientificreports www.nature.com/scientificreports/ smooth muscle tone is strongly regulated by VGCC-independent calcium-influx mechanisms, perhaps through Rho kinase 86 , and perhaps even being attenuated by VGCC-dependent Ca 2+ influx.
If functional Ca v 3 channels are expressed in mouse LMCs, why does their absence neither affect the AP in LMCs in an obvious way nor alter the characteristics of the spontaneous twitch contraction? Based on numerous studies of native and recombinant Ca v 3 channels, the typical activation threshold for Ca v 3 channels is −70 mV 17 , with a current window between −65 and −45 mV 87,88 ; in comparison, the current window for L-type VGCCs is typically between −30 and 0 mV 89 . At the resting Vm that we measured in mouse LMCs (~−35 mV) ( Fig. 13 and ref. 52 ), it is likely that Ca v 3 channels are completely inactivated. Because Vm values in rat and human mesenteric LMCs are only slightly more negative (−40 and −45 mV, respectively; ref. 90 and unpublished observations in pressurized human mesenteric lymphatics), a similar degree of inactivation is likely to prevail in lymphatic vessels of those species (unless a more right-shifted splice variant of Ca v 3.1 is expressed 91 ). It is possible, however, if LMCs were chronically hyperpolarized, e.g. by a vasoactive agent, that rapid depolarization to threshold would recruit Ca v 3 channels to participate in a subsequent AP. This hypothesis remains to be tested. There was substantial cell to cell variation in the AP spike height, but these recordings are reasonably representative. (C-J) Summary of comparisons of resting Vm (C) and AP characteristics (D-J) as analyzed in the same manner described in ref. 52 . There were no significant differences in any of the parameters between the two mouse strains, although there were trends for a few parameters (e.g. AP threshold, diastolic depolarization rate, upstroke exponential depolarization time constant) to be lower in Ca v 3.1 −/− ; Ca v 3.2 −/− vessels.
In mouse LMCs, Ca v 3 channels do not appear to play any detectable role in mediating the inhibition of spontaneous lymphatic contractions by ACh: we could find no significant differences in the concentration-dependent inhibition by ACh in Ca v 3-deficient vessels compared to WT vessels. A possible role for Ca v 3 channels in arterial endothelium was suggested by Figueroa et al. 67 who hypothesized that an electrically-induced signal could be conducted along the arterial endothelium and sequentially activate both Ca v 3.2 and Na v channels to produce Ca 2+ influx that would subsequently regulate NO production. Gilbert et al. 92 found evidence that Ca v 3.1 channels in mouse pulmonary endothelium contribute to ACh-induced [Ca 2+ ] increases that drive NO production. This idea, however, is highly controversial 93,94 in large part due to lack of solid electrophysiological evidence for the expression of functional Ca v 3 or Na v (or Ca v 1) channels in intact endothelium or native endothelial cells. In confirmation of the skeptical view we were unable to detect either Ca v 3.1 or Ca v 3.2 (or Ca v 1.2) message in pure LEC tubes. Neither could we find a consistent impairment in ACh-induced inhibition between Ca v 3.1 −/− , Ca v 3.2 −/− or Ca v 3.1 −/− ; Ca v 3.2 −/− vessels and WT vessels (Suppl. Figs. S7-S10). Previously we showed that ACh-induced inhibition of mouse popliteal lymphatic contractions was solely dependent on NO 55 . If Ca v 3 channels were an important target of LEC-derived NO, as shown for some blood vessels 21,75 , we would have expected the ACh concentration-response relationships for AMP or FREQ, or both, to be right-shifted compared to WT vessels. Although there was a trend for this to occur in Ca v 3.1 −/− vessels, the differences were not significant. In cerebral artery smooth muscle cells Ca v 3.2 channels colocalize with RYR2 to microdomains and their blockade also disrupts RYR-mediated spark production and subsequent BK channel activation that normally counters cerebral arterial constriction 23,33 . Our data do not directly address this issue; however, a role for Ca 2+ sparks in lymphatic smooth muscle, or even their occurrence, has not been demonstrated 7 . Moreover, ryanodine is reportedly without any effect on guinea pig lymphatic smooth muscle 10 although its actions in mouse lymphatic smooth muscle are currently being investigated with genetic tools.
Despite the analysis above, it would be premature to conclude that Ca v 3 channels have no important role whatsoever in lymphatic function. The observations that two different Ca v 3 isoforms are expressed and can be activated by depolarization (under voltage clamp conditions) suggest that they should likely participate in some aspect of lymphatic function. We can conclude from our data only that Ca v 3 channels play no overt role in normal regulation of the amplitude or frequency of spontaneous twitch contractions in lymphatic vessels of the mouse nor are they required for ACh-mediated inhibition of those contractions. However, we cannot rule out the possibility that Ca v 3 channels might play a more substantial role in spontaneous contractions of lymphatic vessels from other species, or that they may be critical for other aspects of lymphatic function even in the mouse, including responses to agonists other than ACh, to nerve stimulation, for remodeling, or mediating immune cell interactions. These possibilities, and others, await to be explored in future studies.
The question of whether T-type calcium channels play a role in lymphatic function is relevant to human medicine because of the possibility for therapeutic targeting of dysfunctional lymphatic vessels in chronic lymphedema. Olszewski's observations of patients with impaired lymphatic smooth muscle contraction strength and/or pacemaking in chronic lymphedema [46][47][48] point to a problem involving an ionic mechanism that potentially might be corrected pharmacologically. However, eventual therapeutic targeting of ionic dysfunction in human LMCs will require knowledge of the specific VGCC isoforms expressed in humans as well as the development of selective inhibitors to block those channels. Whether Ca v 3 channels are expressed and are critical to some aspect of lymphatic function in humans remains unknown.

Data availability
The datasets analyzed during the current study are available from the corresponding author on reasonable request.