Endothelial TRPV4 channels modulate vascular tone by Ca2+‐induced Ca2+ release at inositol 1,4,5‐trisphosphate receptors

Background and Purpose The TRPV4 ion channels are Ca2+ permeable, non‐selective cation channels that mediate large, but highly localized, Ca2+ signals in the endothelium. The mechanisms that permit highly localized Ca2+ changes to evoke cell‐wide activity are incompletely understood. Here, we tested the hypothesis that TRPV4‐mediated Ca2+ influx activates Ca2+ release from internal Ca2+ stores to generate widespread effects. Experimental Approach Ca2+ signals in large numbers (~100) of endothelial cells in intact arteries were imaged and analysed separately. Key Results Responses to the TRPV4 channel agonist GSK1016790A were heterogeneous across the endothelium. In activated cells, Ca2+ responses comprised localized Ca2+ changes leading to slow, persistent, global increases in Ca2+ followed by large propagating Ca2+ waves that moved within and between cells. To examine the mechanisms underlying each component, we developed methods to separate slow persistent Ca2+ rise from the propagating Ca2+ waves in each cell. TRPV4‐mediated Ca2+ entry was required for the slow persistent global rise and propagating Ca2+ signals. The propagating waves were inhibited by depleting internal Ca2+ stores, inhibiting PLC or blocking IP3 receptors. Ca2+ release from stores was tightly controlled by TRPV4‐mediated Ca2+ influx and ceased when influx was terminated. Furthermore, Ca2+ release from internal stores was essential for TRPV4‐mediated control of vascular tone. Conclusions and Implications Ca2+ influx via TRPV4 channels is amplified by Ca2+‐induced Ca2+ release acting at IP3 receptors to generate propagating Ca2+ waves and provide a large‐scale endothelial communication system. TRPV4‐mediated control of vascular tone requires Ca2+ release from the internal store.


| INTRODUCTION
The endothelium plays a critical role in the regulation of numerous vascular processes such as maintaining vascular tone, regulating the passage of macromolecules and oxygen to tissues, modulating immune responses, initiating angiogenesis, and controlling vascular remodelling. The control by the endothelium of each process is mediated by the generation of various signalling molecules that include NO, prostacyclin, von Willebrand factor, tissue plasminogen activator, and endothelial-derived hyperpolarizing and contracting factors (Taylor et al., 2003;Whorton, Willis, Kent, & Young, 1984).
The generation of each of these signalling molecules is controlled tightly by changes in the cytoplasmic Ca 2+ concentration in the endothelium. Central therefore to an understanding of endothelial function is an appreciation of the control of intracellular Ca 2+ concentrations.
While large in amplitude, the Ca 2+ rise evoked by activation of endothelial TRPV4 channels is reported to be highly localized and confined to within a few micrometres of the channels (Sonkusare et al., 2012;Sonkusare et al., 2014). These highly localized Ca 2+ signals evoke endothelial and smooth muscle hyperpolarization and vascular relaxation (F. Gao & Wang, 2010;Kohler et al., 2006;Ma et al., 2013;Sonkusare et al., 2012;D. X. Zhang et al., 2009). The highly localized nature of the Ca 2+ signal provides tight spatial control of the cellular effectors activated, for example, Ca 2+ -activated K + channels (Gao & Wang, 2010;Ma et al., 2013;Sonkusare et al., 2012).
However, more widespread Ca 2+ rises throughout the cytoplasm have been implicated in the regulation of endothelial structure, maintenance of the normal orientation of endothelial cells, the control and selectively of endothelial permeability, and the production of antithrombotic factors. TRPV4 channels have been proposed to play a role in each of these processes (Noren et al., 2016;Phuong et al., 2017;Thodeti et al., 2009;Thoppil et al., 2016). The question arises as to how activation of TRPV4 channels may lead to Ca 2+ -dependent events throughout the cytoplasm if the increase in Ca 2+ concentration arising from TRPV4 channel activity remains localized to within a few microns of the channel.
The second major source of Ca 2+ in endothelial cells is the internal Ca 2+ store. Ca 2+ release from the internal store may occur via ryanodine receptors (RyRs) or IP 3 receptors (IP 3 Rs). While RyRs may be expressed in endothelial cells (Moccia, Berra-Romani, & Tanzi, 2012;Mumtaz, Burdyga, Borisova, Wray, & Burdyga, 2011;Rusko, Wang, & Vanbreemen, 1995), the functional role of the channels in the endothelium is not clear given that RyR activators such as caffeine do not increase cytoplasmic Ca 2+ in endothelial cells (Borisova, Wray, Eisner, & Burdyga, 2009;Wilson et al., 2019;Wilson, Lee, & McCarron, 2016). Ca 2+ release from the internal store in endothelial cells may be mediated mainly by IP 3 Rs (Mumtaz et al., 2011; What is already known • TRPV4 channels are key Ca 2+ permeable channels that control endothelial function.
• TRPV4 channels generate large but highly localized Ca 2+ signals.

What this study adds
• Ca 2+ -induced Ca 2+ release from IP 3 -sensitive internal store amplifies local Ca 2+ signals from TRPV4 channel activity.
• Ca 2+ -induced Ca 2+ release at IP 3 receptors mediates control of vascular contractility by endothelial TRPV4 channels.
What is the clinical significance • Endothelial TRPV4 channels are involved in several disease conditions such as cancer progression and hypertension.
• Endothelial TRPV4-mediated Ca 2+ -induced Ca 2+ release at IP 3 receptors offers a new target for drug development. et al., 2019). Ca 2+ release via IP 3 R occurs in response to physical forces and circulating vasoactive substances and contributes to the control of several vascular activities such as vasorelaxation, endothelial permeability, and production of anti-thrombotic factors (Sun, Geyer, & Komarova, 2017 (Berridge, 1997;Bootman, Niggli, Berridge, & Lipp, 1997). These waves may move through all or part of the cell . In some conditions that are not fully understood, IP 3 -evoked Ca 2+ waves may also move between cells to provide a signalling system capable of coordinating the activity of many cells (Leybaert & Sanderson, 2012).
While Ca 2+ influx and release are activated separately, they are not independent: Ca 2+ influx can trigger release, and release can trigger influx (McCarron, Chalmers, Bradley, Macmillan, & Muir, 2006). In cardiomyocytes, Ca 2+ influx via voltage-dependent Ca 2+ channels may activate RyRs to evoke a large rise in Ca 2+ , inducing cell contraction. In smooth muscle cells, TRPV4 channel activity may lead to Ca 2+induced Ca 2+ release acting at RyRs (Earley, Heppner, Nelson, & Brayden, 2005). In astrocytes, Ca 2+ -induced Ca 2+ release occurs in response to Ca 2+ influx via TRPV4 channels but at IP 3 Rs rather than at RyRs. This process amplifies and propagates the Ca 2+ signal arising from TRPV4-mediated Ca 2+ influx (Dunn, Hill-Eubanks, Liedtke, & Nelson, 2013). TRPV4-mediated Ca 2+ influx may also lead to recruitment of IP 3 Rs and Ca 2+ -induced Ca 2+ release in murine-derived cultured neuronal cells (Shen et al., 2018). These observations raise the possibility of TRPV4-mediated Ca 2+ influx being able to generate a more global increase in intracellular Ca 2+ via Ca 2+ -induced Ca 2+ release in endothelial cells.
To explore this possibility, we examined the role that the internal store plays in regulating alterations in intracellular Ca 2+ evoked by activation of TRPV4 channels. Here, we report that the IP 3 -sensitive Ca 2+ store is critical for the TRPV4-evoked global Ca 2+ rise in endothelial cells in intact blood vessels. Ca 2+ influx generates Ca 2+ -induced Ca 2+ release at IP 3 Rs to evoke Ca 2+ waves in the vascular endothelium. IP 3 R-dependent Ca 2+ waves are required for the endotheliumdependent vascular smooth muscle cell relaxation in response to activation of TRPV4 channels. Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny et al., 2010) and with the recommendations made by the British Journal of Pharmacology. The Strathclyde Biological Protection Unit is a conventional unit which undertakes FELASA quarterly health monitoring. Male Sprague-Dawley rats (10-12 week old; 250-300 g), from an in-house colony, were used for the study. Sprague-Dawley rats are a widely used experimental model with a wealth of background information to aid interpretation of results. The animals were housed three per cage, and the cage type was North Kent Plastic model RC2F with nesting material "Sizzle Nest." A 12:12 light dark cycle was used with a temperature range of 19-23°C (set point 21°C) and humidity levels between 45% and 65%. Animals had free access to fresh water and SDS diet RM1 (rodent maintenance). The enrichment in the cages was aspen wood chew sticks and hanging huts.
All experiments used first-or second-order mesenteric arteries isolated from rats killed by either CO 2 or by injection with pentobarbital; no differences in the results were observed with either method.
Controls and experimental treatments were carried out in the same tissue, so blinding and randomization were not used.

| High-resolution imaging of endothelial Ca 2+ signalling
Arteries (diameter of 200-230 μm) were isolated from the mesenteric bed, cleaned, then cut and pinned flat en face (endothelial side facing up) on a Sylgard block. The endothelium was then preferentially loaded with Cal-520/AM (5 μM; with 0.02% Pluronic F-127) in physiological saline solution (PSS) at 37°C for 30 min (Wilson, Lee, & McCarron, 2016;Wilson, Saunter, Girkin, & McCarron, 2016). Following incubation, arteries were gently washed with PSS, and the Sylgard blocks were placed face down on a custom flow chamber. Arteries were then continuously perfused with PSS at 1.5 ml·min −1 using a syringe pump. Endothelial Ca 2+ activity was stimulated by swapping the PSS syringe for one containing ACh (100 nM) or GSK1016790A (GSK, 20 nM). GSK1016790A is a selective TRPV4 channel agonist and evokes Ca 2+ influx in wild type but not TRPV4 channel knockout mice (Mannaa et al., 2018;Sonkusare et al., 2012).
In experiments examining the effects of various pharmacological agents on stimulated endothelial Ca 2+ signalling, drugs were added to the perfusate and remained thereafter. Images were acquired at 10 Hz using an inverted fluorescence microscope (Eclipse TE300, Nikon, Tokyo, Japan) equipped with a 40× objective (S Fluor, Nikon, Tokyo, Japan, NA = 1.3) and an electron-multiplying charge-coupled device camera (iXon Life; Andor Technology Limited, Belfast, Northern Ireland, UK) or on an upright epi-fluorescence microscope (FN-1, Nikon, Tokyo, Japan) equipped with a 60× objective (CFI Fluor, Nikon, Tokyo, Japan, NA = 1.0). Cal520/AM was excited at 488 nm using a monochrometer (Polychome IV, TILL Photonics, Graefelfing, Germany) or an LED illumination system (PE-300Ultra, CoolLED, Andover, UK).
In some experiments, the endothelium was loaded with Cal-520 (5 μM) and a membrane permeant, caged IP 3 (5 μM), and the Ca 2+ response to local photolysis of caged IP 3 examined. In these experiments, the endothelium was imaged as above, and a xenon flash lamp (Rapp Optoelecktronic, Germany) was used to uncage IP 3 Olson, Sandison, Chalmers, & McCarron, 2012;Wilson et al., 2019).

| Extraction and analysis of endothelial Ca 2+ signals
Endothelial Ca 2+ signals were extracted automatically from fluorescence recordings using custom written Python software (RRID: SCR_001658; Lee et al., 2018;Wilson, Lee, & McCarron, 2016;Wilson, Saunter, et al., 2016). In brief, cellular regions of interest (ROIs) were first generated from average-intensity projections of each time series recording. Intensity projections were sharpened (un-sharp mask filter) and thresholded, generating ROIs that encompassed the majority of each cell's area. To facilitate comparisons in paired experiments, for example, where Ca 2+ activity was recorded before and after pharmacological inhibition, ROIs were aligned and tracked across separate image acquisitions. Only cells that remained within the field of view for all recordings were included. Cellular Ca 2+ signals were extracted by averaging fluorescence intensity within each of the ROIs, for each frame of the image stack.
Raw fluorescence signals (F) were expressed as fractional changes in fluorescence intensity (F/F 0 ) by dividing each intensity value by the average intensity of a 100-frame period exhibiting the least activity/noise (F 0 , red in Figure S1B). The F 0 period was determined automatically by calculating, in series, the derivative of the signal, the rolling (100-frame) SD of the derivative, and the rolling (100frame) summation of the rolling SD. The minimum value of the rolling summation indicates the centre of the portion of the signal with the least activity/noise. The Ca 2+ response to activation of TRPV4 channels contained two main components: a "slow" persistent Ca 2+ elevation rise in the baseline Ca 2+ levels and fast intracellular Ca 2+ waves. To determine the underlying mechanisms, we isolated the two components from each  Figure S1C-E). Peaks in each Ca 2+ signal arising from the waves were then detected automatically, using a zero-crossing detector on derivative F/F 0 traces (Lee et al., 2018;Wilson, Lee, & McCarron, 2016).
Various signal metrics (number of peaks, peak amplitudes, peak durations, 10-90% rise time, and 90-10% fall time) were extracted from the corresponding ALS-smoothed F/F 0 trace. Cells were considered to exhibit Ca 2+ wave activity if the F/F 0 component exhibited at least one event with an amplitude more than 10-fold the SD of the baseline noise.
In some experiments, Ca 2+ signals were calibrated using Ca 2þ Images were acquired at 10 or 5 Hz (consistent within each experimental protocol) using an upright fluorescence microscope equipped with a 16× objective lens (0.8 NA; Nikon, Tokyo, Japan) and large format (1,024 × 1,024; 13-μm pixels) back-illuminated electronmultiplying charge-coupled device camera (iXon 888; Andor, Belfast, UK) and stored for offline analysis. An edge-detection algorithm (Lawton et al., 2019) was used to track the width of arteries in image recordings. The algorithm extracts an intensity profile along a scanline orientated perpendicular to the longitudinal axis of the vessel.
Each intensity profile was smoothed using a 251-point, fifth-order Savitzky-Golay filter, and the first-order derivative calculated. The edges of the artery correspond to the maxima and minima of the first-order derivative, which were identified and tracked using a zero-crossing detector. The width of the artery equates to the unfolded circumference of the intact vessel and was converted to the equivalent diameter. To control for variation in the resting diameters, summary contractile response data are expressed as the percentage reduction from resting diameter,while relaxation responses are expressed as the percentage increase in diameter compared to resting diameter ( Figure S2).

| Matching arterial tone across experimental conditions
Following the removal of the endothelium, arteries were significantly more sensitive to phenylephrine. The concentration of phenylephrine was titrated to achieve a level of tone that was comparable to that achieved in the presence of a functional endothelium ( Figure S2A).
In the presence of cyclopiazonic acid (CPA), Phenylephrine-induced contractions were transient, and it was not possible to achieve a stable contraction. Therefore, assessing the effect of store depletion on TRPV4-induced dilation was not feasible. Instead, vessels were pretreated with GSK and CPA, and the magnitude of the transient contraction induced by phenylephrine was assessed. Furthermore, after depletion of Ca 2+ stores using CPA, a significantly higher concentration of phenylephrine was required to generate comparable levels of vascular tone ( Figure S2B). Artery diameter was monitored using an sCMOS camera (DCC1545M, Thorlabs, New Jersey, USA) and recorded by VasoTracker acquisition software. Arteries were contracted using phenylephrine (~500 nM) to~80% of resting diameter, which was added to the perfusion solution. All other drugs (e.g., ACh and GSK) were applied intraluminally. Summary relaxation data are expressed as the percentage increase in diameter compared to resting diameter.

| Data and statistical analysis
The data and statistical analysis in this study comply with the recom-

| Characteristics of ACh-and GSK-evoked Ca 2+ signals
To determine if activation of TRPV4 channels in endothelial cells evoked Ca 2+ release from the internal store, endothelial cells were loaded with the fluorescent Ca 2+ indicator Cal520/AM and activated with ACh (100 nM) or the TRPV4 channel agonist GSK (20 nM).
Ca 2+ activity was then visualized in the fields of~100 endothelial cells. Cellular responses were analysed individually (Figures 1   and S1).
Ca 2+ responses evoked by ACh and GSK were clearly different ( Figure 1). ACh evoked a rapid elevation in intracellular Ca 2+ which was followed by asynchronous oscillations across the field of view ( Figure 1; Mumtaz et al., 2011). Activation of TRPV4 channels resulted in heterogeneous Ca 2+ responses ( Figure 1c, Movie S1, and Figure S3) that consisted of several distinct phases (Figures 1 and 2). The response to GSK was, initially, small localized Ca 2+ spikes which led to a slowly increasing global cytoplasmic Ca 2+ concentration (Figures 2b and S4). The GSK-evoked, global Ca 2+ elevation eventually plateaued, but the time taken to plateau was significantly longer when compared to ACh (Figure 1; 55.9 ± 6.4 s for GSK; 16.7 ± 4.7 s for ACh; n = 6). Finally, large propagating Ca 2+ waves developed (Figure 2b,c and Movies S1 and S2). These waves propagated within and between cells at a constant velocity of~5 to 15 μm·s −1 (Figure   2b,c, Movie S2, and Figure S5). GSK-evoked intracellular Ca 2+ waves were significantly lower in frequency than those for ACh (0.05 ± 0.01 Hz for GSK; 0.22 ± 0.03 Hz for ACh; n = 6).
These results suggest there are at least major two components of the Ca 2+ signal arising from activation of TRPV4 channels: (a) initial local signals which lead to a slow global rise in Ca 2+ and (b) large fast propagating Ca 2+ waves.
We next sought to examine the mechanisms that give rise to the GSK-evoked Ca 2+ signals. We first confirmed the reproducibility of GSK-evoked responses (20 nM) in the same preparation so that a   Figure S1).
To determine if each component is dependent on Ca 2+ entry, we first performed experiments using a Ca 2+ -free PSS (Figure 3a,b). The removal of extracellular Ca 2+ significantly decreased the percentage of cells exhibiting "slow" global Ca 2+ elevations and "fast" propagating waves in response to GSK (Figure 3c). The velocity of residual Ca 2+ waves was similar to those occurring in the presence of external Ca 2+ (Figure 3d; 10.0 ± 1.1 μm·s −1 for control; 8.8 ± 0.7 μm·s −1 for Ca 2+free; n = 5). However, removal of external Ca 2+ significantly reduced the amplitude (Figure 3e; 0.28 ± 0.11 ΔF/F 0 for control; 0.02 ± 0.10 ΔF/F 0 for Ca 2+ -free; n = 6) and the frequency of occurrence (Figure 3f; 4.1 ± 0.6 peaks per cell for control; 1.7 ± 0.1 peaks per cell for Ca 2+ -free; n = 6) of these residual propagating Ca 2+ waves.
These results suggest that Ca 2+ influx across the plasma membrane is essential for the slow global increase in Ca 2+ and for the propagating Ca 2+ waves evoked by GSK.
Together, these results suggest that Ca 2+ influx via TRPV4 channels is required to induce both the slow global rise component and the propagating Ca 2+ waves induced by GSK.
3.3 | GSK-induced Ca 2+ waves require a replete internal Ca 2+ store, IP 3 synthesis, and IP 3 receptor activation Having confirmed a role for TRPV4-mediated Ca 2+ influx in GSKstimulated Ca 2+ signals, we next investigated the contribution of the internal Ca 2+ store using the sarcoplasmic/endoplasmic reticulum These results suggest that the internal store is required for the propagating Ca 2+ waves.
These results suggest that the GSK-evoked slow global rise in Ca 2+ and propagating Ca 2+ waves each have contributions from PLC activation.
To investigate any role of IP 3 -sensitive Ca 2+ stores in propagating To crosscheck the contribution of IP 3 Rs in the propagating Ca 2+ waves, we used caffeine-a potent inhibitor of IP 3 Rs (Ehrlich, Kaftan, Bezprozvannaya, & Bezprozvanny, 1994;Parker & Ivorra, 1991;Saleem, Tovey, Molinski, & Taylor, 2014). Caffeine, which does not evoke Ca 2+ release in the endothelial cells under study and inhibits Ca 2+ release evoked by IP 3 , also blocked GSK-evoked propagating Ca 2+ waves ( Figure S8). Interestingly, caffeine also reduced the slow global Ca 2+ rise suggesting an effect of caffeine also on TRPV4 channels.
While Ca 2+ influx via TRPV4 channels triggered large propagating Ca 2+ waves from the internal Ca 2+ store, the release and influx did not become an uncontrolled self-regenerative process but remained under the control of Ca 2+ influx. When activation of TRPV4 channels ceases, by washout of GSK, the propagating Ca 2+ waves stop ( Figure S9 and Movie S3), and Ca 2+ returned towards resting values.

| The functional effects of TRPV4 channels in controlling vascular tone
Collectively, our results suggest that endothelial TRPV4-mediated Ca 2+ influx activates Ca 2+ -induced Ca 2+ release at IP 3 Rs. To investigate any physiological contribution of this process, we examined vascular tone in mesenteric arteries (Figure 8) with and without a functionally intact endothelial layer.
Relaxations to GSK were reversed by the selective TRPV4 channel antagonist, HC067. However, the dilations to ACh were preserved after blockade of TRPV4 channels with HC067 ( Figure S10; n = 5). These results suggest that these channels do not contribute to ACh-evoked relaxations in rat mesenteric arteries (see also Hartmannsgruber et al., 2007;Kohler et al., 2006;Wilson, Lee, & McCarron, 2016).
We next investigated the contribution of the internal Ca 2+ store to TRPV4-mediated relaxation evoked by GSK. In en face preparations (225 ± 4 μm resting diameter; n = 10), ACh and GSK each relaxed phenylephrine-constricted arteries back to pre-constricted levels ( Figure 9 and Figure S2). In these same arteries, the removal of the endothelium increased the sensitivity to phenylephrine and significantly reduced relaxation to ACh (to 26 ± 7%; n = 5 vs. control with endothelium). The residual ACh-evoked response may have arisen from the incomplete removal of the endothelium. The functional removal of the endothelium also prevented GSK-evoked relaxations.
Indeed, after the removal of the endothelium, GSK caused an increase in tone (Figure 9e; n = 5). Endothelium removal did not prevent relaxation to the endothelium-independent vasodilator SNP ( Figure 9).
These results demonstrate that vascular relaxation to GSK occurs via an endothelium-dependent mechanism and not by acting directly on the underlying smooth muscle. . Individual data points are coloured (from blue, low to red, high) according to the density (i.e., occurrence) of particular values. *P < .05, significantly different as indicated; paired Student's t test (n = 5)

| Store dependence of the TRPV4 channel response
We next examined the contribution of Ca 2+ -induced Ca 2+ release at IP 3 Rs in TRVP4 mediated relaxation. To do this, we assessed the effect of GSK on vascular reactivity before and after depletion of internal Ca 2+ stores.
Depletion of the internal Ca 2+ store using CPA (6 μM) prevents stable contractions such that it was not possible to reliably assess vasodilator responses in these arteries. Instead, we examined whether pretreatment with GSK was capable of modulating PE-evoked contraction ( Figure 10). Pretreatment with GSK significantly attenuated PE-induced contractions (Figure 10a, c).
Together, these results demonstrate that depletion of internal Ca 2+ stores reduces the inhibitory effect of GSK pretreatment on phenylephrine-induced contraction.
The response to activation of TRPV4 channels was also heterogeneous among cells. Some cells were highly responsive, while other cells responded weakly or not at all (see also Aird, 2012;Huang, Chu, Chen, & Jen, 2000;Lee et al., 2018;Marie & Beny, 2002; McCarron, Lee, & Wilson, 2017;McCarron et al., 2019;. We developed methods to automatically extract and separate the slow global increases and propagating Ca 2+ waves occurring in each endothelial cell. The changes in Ca 2+ concentration triggered by activation of TRPV4 channels consisted of different phases which operated sequentially. First, rapid localized Ca 2+ changes occurred   (Running Deer et al., 1995;Willars et al., 1998) and in living cells (Hardie et al., 2004;Willars et al., 1998) (6). The entire process remains under the control of Ca 2+ influx and terminates when TRPV4 channel activity ceases. This channel activity is suppressed by PIP 2 (7). ACh, via the M3 receptor, also activates PLC to generate IP 3 and activate Ca 2+ release from the store. The characteristics of M 3and TRPV4-mediated Ca 2+ release differ significantly (see text) waves supports the conclusion that IP 3 Rs contribute to TRPV4mediated Ca 2+ responses.
RuR and HC067047 were each used to block TRPV4 channels.
RuR is an effective TRPV4 channel antagonist but also has effects unrelated to this channnel. For example, RuR is an antagonist of channels such as the mitochondrial uniporter, the RyR, voltage-dependent Ca 2+ channels, and other TRP channels. However, the effects of RuR on these other channels are unlikely to explain the present findings.
RuR is not membrane permeant and will not have access to the cytoplasm, so the effects on the mitochondrial uniporter or RyR are unlikely to be of significance in the present study. While RuR may have effects on voltage-sensitive Ca 2+ channels (Cibulsky & Sather, 1999), these channels do not appear to play a major role in endothelial  (Adkins & Taylor, 1999;McCarron, MacMillan, Bradley, Chalmers, & Muir, 2004;Oancea & Meyer, 1996). These observations highlight a system that is activated by Ca 2+ influx, but is regulated to prevent an uncontrolled positive feedback processes dominating the TRPV4-mediated increases in cytoplasmic Ca 2+ concentrations.
As well as propagating at a constant velocity within cells, TRPV4mediated Ca 2+ waves could move seamlessly, at the same velocity, across cell boundaries into neighbouring cells as intercellular Ca 2+ waves (Movie S2). These waves were abolished by IP 3 R blockers or removal of GSK. The mechanisms of intercellular wave propagation are not entirely clear, but IP 3 Rs and (since the process stops when GSK is removed) TRPV4 channel activities are both required for these intercellular waves to occur.
In other tissues, intercellular Ca 2+ waves may be a fundamental mechanism for coordinating multicellular responses (Leybaert & Sanderson, 2012). In the endothelium of small arteries of the mouse cremaster muscle, rapidly spreading intercellular Ca 2+ waves propagate at a velocity of ∼116 μm·s −1 for distances up to ∼1 mm (Tallini et al., 2007). These intercellular waves are associated with vasodilation and are hypothesized to regulate blood flow to the parenchyma by inducing upstream dilation of arterioles (Tallini et al., 2007). The intercellular waves observed in the present study had a slower velocity (~5 to 15 μm·s −1 ) than those in mouse cremaster arterioles, suggesting a different physiological control mechanism.
In previous studies, when endothelial IP 3 R-mediated signalling was eliminated by depletion of the internal Ca 2+ store, activation of TRPV4 channels induced a large but highly localized (few square micrometres) increase in Ca 2+ concentrations (Sonkusare et al., 2012). These localized Ca 2+ signals were reported to activate endothelial Ca 2+ -activated K + channels (Sonkusare et al., 2012) and lead to endothelial hyperpolarization and vasodilation (Sonkusare et al., 2012). However, activation of TRPV4 channels is known to also control the orientation of endothelial cells, regulate endothelial permeability, and modulate the production of antithrombotic factors, each of which may require a more global [Ca 2+ ] increase throughout the cytoplasm (Noren et al., 2016;Phuong et al., 2017;Thodeti et al., 2009;Thoppil et al., 2016).
In the present study, IP 3 R-mediated Ca 2+ release generated propagating Ca 2+ waves and provided a mechanism by which TRPV4 channel activity, in the presence of a functioning Ca 2+ store, may generate large rises in Ca 2+ throughout the cytoplasm. These propagating waves are critical for TRPV4-mediated control of vascular tone. When Ca 2+ release from the internal store is inhibited, TRPV4-mediated endothelial control of tone is abolished ( Figure 10). This result suggests that control of vascular tone by TRPV4 channels is mediated by Ca 2+ -induced Ca 2+ release via IP 3 Rs. It is tempting to also speculate that the more general rise in Ca 2+ throughout the cells, which occurs as a result of the propagating IP 3 R mediated Ca 2+ waves, will facilitate TRPV4-mediated control of endothelial permeability and the production of antithrombotic factors (Noren et al., 2016;Phuong et al., 2017;Thodeti et al., 2009;Thoppil et al., 2016).
ACh evoked rapid asynchronous Ca 2+ waves in neighbouring endothelial cells. ACh-evoked changes in intracellular Ca 2+ concentration are linked closely to the IP 3 -sensitive Ca 2+ store, but they do not involve TRPV4 channels in rat arteries (Hartmannsgruber et al., 2007; see also Kohler et al., 2006;Wilson, Lee, & McCarron, 2016; Figure 10). While the response to the activation by ACh and activation of TRPV4 channels each involved IP 3 , the Ca 2+ signals generated had very different characteristics. ACh-evoked Ca 2+ increases did not appear to coordinate between cells. Activation of TRPV4 channels evoked a slowly increasing baseline and large propagating Ca 2+ waves.
Endothelial TRPV4 channels are involved in several cardiovascular control mechanisms, including inhibiting vasoconstriction in small skeletal muscle arteries in response to increases in temperature (Gifford et al., 2014), regulating myogenic tone (Bagher et al., 2012), remodelling of the cytoskeleton and reorientation of the endothelial cells in response to mechanical forces (Thodeti et al., 2009), collateral vessel growth (Sayed et al., 2010;Schierling et al., 2011;Troidl et al., 2009), and arachidonic acid-induced endothelial cell migration required for angiogenesis (Fiorio Pla et al., 2012). The present results show that when the internal IP 3 sensitive store is intact, TRPV4 channel activity evokes IP 3 R signalling to generate Ca 2+ waves that propagate within and between cells, and through the Ca 2+ store, TRPV4 channels modulate vascular contractility. The results demonstrate a link between TRPV4 channel activity and Ca 2+ -induced Ca 2+ release at IP 3 R in endothelial cells and offer a new target for drug development.