2.1 Isolation and cultivation of ECFCs

Blood samples (40 ml) were obtained from patients who were out of ongoing cytoreductive therapy at the moment of blood sampling for ECFC isolation and from healthy donors. The ECFC samples used in the present investigation belong to the stocks recently described and characterized by Moccia et al. (2017). To isolate ECFCs, MNCs were separated from peripheral blood by density gradient centrifugation on lymphocyte separation medium for 30 min at 400g and washed twice in endothelial basal medium-2 (EBM-2) with 2% fetal bovine serum (FBS). A median of 36 × 10⁶ MNCs (range, 18–66) were plated on collagen-coated culture dishes (BD Biosciences) in the presence of the endothelial cell growth medium EGM-2MV BulletKit (Lonza) containing EBM-2, 5% FBS, recombinant human (rh)-epidermal growth factor, rhVEGF, recombinant human fibroblast growth factor-basic, recombinant human insulin-like growth factor-1, ascorbic acid, and heparin, and maintained at 37°C in 5% CO₂ and humidified atmosphere. Discard of nonadherent cells was performed after 2 days; thereafter, the medium was changed three times a week. The outgrowth of endothelial colonies from adherent MNCs was characterized by the formation of a cluster of cobblestone-appearing cells, resembling endothelial cells. That ECFCs-derived colonies belonged to endothelial lineage was confirmed as described by Moccia et al. (2017) and Sanchez-Hernandez et al. (2010). In more detail, ECFCs-derived colonies were stained with anti-CD31, anti-CD105, anti-CD144, anti-CD146, anti-vWF, anti-CD45, and anti-CD14 monoclonal antibodies and by assessment of capillary-like network formation in an in vitro Matrigel assay. For our experiments, we have mainly used endothelial cells obtained from early passage ECFCs (P1–P3, which roughly encompasses a 15–18-day period) with the purpose to avoid (or maximally reduce) any potential bias due to possible cell differentiation. As already shown by Lodola et al. (2012), the immunophenotype of both normal and tumor ECFCs does not change at different passages in culture. We also tested whether functional differences occurred when early (P2) and late (P6) passage ECFCs were used by testing the in vitro capacity of capillary network formation in a Matrigel assay and found no differences between early and late passage ECFC-derived cells (data not shown).

2.2 Solutions

Physiological salt solution (PSS) had the following composition (in millimolar): 150 NaCl, 6 KCl, 1.5 CaCl₂, 1 MgCl₂, 10 glucose, and 10 HEPES. In Ca²⁺-free solution (0Ca²⁺), Ca²⁺ was substituted with 2 mM NaCl, and 0.5 mM egtazic acid was added. Solutions were titrated to pH 7.4 with NaOH. The osmolality of PSS as measured with an osmometer (Wescor 5500; Logan, UT) was 338 mmol/kg.

2.3 [Ca²⁺]_i measurements

ECFCs were loaded with 4 µM Fura-2 acetoxymethyl ester (Fura-2/AM; 1 mM stock in dimethyl sulfoxide) in PSS for 1 hr at room temperature. After washing in PSS, the coverslip was fixed to the bottom of a Petri dish and the cells observed by an upright epifluorescence Axiolab microscope (Carl Zeiss, Oberkochen, Germany), usually equipped with a Zeiss ×40 Achroplan objective (water immersion, 2.0 mm working distance, and 0.9 numerical aperture). ECFCs were excited alternately at 340 and 380 nm, and the emitted light was detected at 510 nm. A first neutral density filter (1 or 0.3 optical density) reduced the overall intensity of the excitation light and a second neutral density filter (optical density = 0.3) was coupled to the 380-nm filter to approach the intensity of the 340 nm light. A round diaphragm was used to increase the contrast. The excitation filters were mounted on a filter wheel (Lambda 10; Sutter Instrument, Novato, CA). Custom software, working in the Linux environment, was used to drive the camera (Extended-ISIS Camera; Photonic Science, Millham, UK) and the filter wheel, and to measure and plot online the fluorescence from 10 to 100 rectangular “regions of interest” (ROI). Each ROI was identified by a number. Since cell borders were not clearly identifiable, an ROI may not include the whole cell or may include part of an adjacent cell. Adjacent ROIs never superimposed. [Ca²⁺]_i was monitored by measuring, for each ROI, the ratio of the mean fluorescence emitted at 510 nm when exciting alternatively at 340 and 380 nm (shortly termed "ratio"). An increase in [Ca²⁺]_i causes an increase in the ratio (Sanchez-Hernandez et al., 2010). Ratio measurements were performed and plotted online every 3 s. The experiments were performed at room temperature (22°C). In the figures, each Ca²⁺ tracing refers to an individual cell and is representative of at least three independent experiments.

2.4 Preparation of NAADP-containing liposomes

NAADP-containing liposomes were prepared from lecithin by a thin film hydration method, as recently shown by Di Nezza et al. (2017) and Faris et al. (2019). A thin film was formed by dissolving the lecithin in chloroform/methanol solution (2:1 vol/vol) in a round bottom flask and following removal of the solvent under vacuum condition at room temperature, which ensured complete removal of the solvents. The film was then hydrated with phosphate-buffered saline (PBS) buffer (10 mM, pH 7.4) to make 20 ml of lipid coarse dispersion. Liposomes were prepared by adding cholesterol in an 89:20 lecithin:cholesterol molar ratio, codissolved in chloroform, and then dried. The dried film from a flask was suspended in 4 ml of rehydration solution. The resulting liposomal dispersion was sonicated 66 for 3 min (ultrasound homogenizer; BioLogics) and extruded 21 times with 100-nm filter. Finally, the mixture was dialyzed in PBS bulk for 24 hr with three bulk changes. Properties of liposomes were modulated by varying the rehydration solution composition. Liposomes were prepared from PBS solution containing only 8.70 µg of NAADP. NAADP was diluted at 1:20, as shown by Di Nezza et al. (2017).

2.5 Membrane preparation and immunoblotting

Cells were diluted in PBS and centrifuged 1,000g for 10 min. The cell pellets were resuspended with radioimmunoprecipitation assay buffer (150 mM NaCl, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecylsulphate, 50 mM Tris-HCl, and pH 8.0) containing 0.1 mg/ml phenylmethylsulfonyl fluoride, protease inhibitor cocktails (P8340; Sigma-Aldrich, St. Louis, MO) and phosphatase inhibitor (1 mM sodium orthovanadate). The homogenates were solubilized in Laemmli buffer (Laemmli, 1970) and 30 µg proteins were separated on precast gel electrophoresis (4–20% Mini-PROTEAN TGX Stain-Free Gels; Bio-Rad) and transferred to the Hybond-P polyvinylidene difluoride membrane (GE Healthcare, Italy) by electroelution. After 1 hr blocking with Tris-buffered saline (TBS) containing 3% bovine serum albumin (BSA) and 0.1% Tween (blocking solution), the membranes were incubated overnight at 4°C with the following antibody: affinity-purified rabbit anti-TPC1 (SAB2104213; Sigma-Aldrich), diluted 1:500 in 3% BSA in Tween-TBS. The membranes were washed and incubated for 1 hr with peroxidase-conjugated goat anti-rabbit IgG (AP132P; Chemicon), diluted 1:100,000 in blocking solution. The bands were detected with ECL™ Select western blot analysis detection system (GE Healthcare Europe GmbH, Italy). Prestained molecular weight markers (ab116028; Abcam) were used to estimate the molecular weight of the bands. Blots were stripped as shown in and reprobed with RabMAb anti-β-actin antibody (AB-81599; Immunological Sciences) as housekeeping. The antibody was diluted 1:2,000 in blocking solution.

2.6 Protein content

Protein contents of all the samples were determined by the Bradford's (1976) method using BSA as standard.

2.7 Gene silencing

siRNA targeting TPC1 (siTPC1) was purchased by Sigma-Aldrich MISSION esiRNA (human TPCA1, EHU016301). Scrambled siRNA was used as a negative control. Briefly, once the monolayer cells had reached 50% confluency, the medium was removed and the cells were added with Opti-MEM I reduced serum medium without antibiotics (Life Technologies). siRNAs (100 nM final concentration) were diluted in Opti-MEM I reduced serum medium and mixed with Lipofectamine™ RNAiMAX Transfection Reagent (Invitrogen) prediluted in Opti-MEM, according to the manufacturer's instructions. After 20 min incubation at room temperature, the mixes were added to the cells and incubated at 37°C for 5 hr. Transfection mixes were then completely removed and the fresh culture media was added. The effectiveness of silencing was determined by immunoblotting (see Figure S1) and silenced cells were used 48 hr after transfection. Before and after the transfection procedure, ECFCs were always maintained in the presence of the EGM-2MV BulletKit endothelial cell medium. The Trypan blue exclusion assay confirmed that the genetic silencing of TPC1 did not affect ECFC viability (Figure S2).

2.8 Proliferation assay

ECFCs were detached by trypsinization at 48 hr after transfection with scrambled siRNA and siTPC1 and then resuspended in the EBM-2 medium containing 2% FCS and 10 ng/ml VEGF, as shown by Dragoni et al. (2011, 2015). Cultures were incubated at 37°C, 5% CO_2, and cell growth assessed every day until confluence was reached in the control dishes. Cells were then recovered by trypsinization and their number assessed by counting in a hemocytometer. Three different sets of experiments were performed.

2.9 Statistics

All the data have been collected from ECFCs deriving from at least three coverslips from three distinct donors. The amplitude of intracellular Ca²⁺ release in response to each agonist (NAADP or ATP) or drug (glycyl-l-phenylalanine 2-naphthylamide [GPN], nigericin, and cyclopiazonic acid [CPA]) was measured as the difference between the ratio at the peak of intracellular Ca²⁺ mobilization and the mean ratio of 1 min baseline before the peak. For VEGF stimulation, the percentage of oscillating the cells was evaluated. Pooled data are given as mean ± standard error (SE) and statistical significance (p < .05) was evaluated by Student's t test for unpaired observations or one-way analysis of variance followed by the post hoc Dunnett's test as appropriate. Data relative to Ca²⁺ signals are presented as mean ± SE, while the number of cells analyzed is indicated in the corresponding bar histograms.

2.10 Chemicals

Fura-2/AM was obtained from Molecular Probes (Molecular Probes Europe BV, Leiden, The Netherlands). All the chemicals were of analytical grade and obtained from Sigma-Aldrich.

3.1 A functional EL Ca²⁺ pool is present in circulating ECFCs

To assess whether a function EL Ca²⁺ store is present and able to increase the [Ca²⁺]_i, ECFCs were loaded with the Ca²⁺-sensitive fluorochrome, Fura-2/AM, as shown by Di Nezza et al. (2017). Thereafter, ECFCs were challenged with the dipeptide GPN, which is cleaved by the lysosomal enzyme cathepsin C, thereby rupturing lysosomal membrane and liberating lysosomal Ca²⁺ into the cytosol (Faris et al., 2018). GPN has been widely exploited to analyze EL Ca²⁺ signaling in multiple cell types (Faris et al., 2019; Kilpatrick et al., 2013; McGuinness, Bardo, & Emptage, 2007; Moccia, Billington, & Santella, 2006; Pandey et al., 2009; Ronco et al., 2015), including vascular endothelial cells (Srinivas, Ong, Goon, Goon, & Bonanno, 2002; Zuccolo, Laforenza et al., 2019) and circulating ECFCs (Zuccolo, Dragoni et al., 2016). As expected, GPN (200 μM) caused a rapid and transient elevation in [Ca²⁺]_i in the absence of extracellular Ca²⁺ (0Ca²⁺; Figure 1a), which reflects Ca²⁺ mobilization from the endogenous lysosomal pool (Kilpatrick et al., 2013; Pandey et al., 2009). The subsequent addition of Ca²⁺ did not cause any detectable increase in [Ca²⁺]_i (Figure 1a), which suggests that acidic Ca²⁺ stores are not coupled to plasmalemmal Ca²⁺ entry pathways. Likewise, the H⁺/K⁺ antiporter, nigericin (50 μM), which dissipates the lysosomal proton gradient responsible for Ca²⁺ refilling (Faris et al., 2019, 2018), caused a transient elevation in [Ca²⁺]_i under 0Ca²⁺ conditions that was not coupled to extracellular Ca²⁺ entry upon restoration of external Ca²⁺ concentration (Figure 1b). Finally, we probed the effect of bafilomycin A1 (100 nM), a selective inhibitor of the V-type H⁺ ATPase that drives EL acidification (Faris et al., 2018; Morgan et al., 2011). Bafilomycin A1 did not cause any detectable change in [Ca²⁺]_i in circulating ECFCs (Figure 1c). As discussed by Faris et al. (2018) and Morgan et al. (2011), bafilomycin A1 depletes lysosomal Ca²⁺ only in the presence of adequate Ca²⁺ leakage. It is, therefore, conceivable that resting lysosomal Ca²⁺ permeability is low in ECFCs. The statistical analysis of these recordings is presented in Figure 1d. Overall, these observations confirm that a functional lysosomal Ca²⁺ store is present and able to increase the [Ca²⁺]_i in ECFCs.

3.2 Lysosomal Ca²⁺ release requires an intact InsP₃-sensitive ER Ca²⁺ pool

As mentioned above, lysosomal Ca²⁺ release may require a bidirectional coupling with the ER in which InsP₃Rs play a crucial role (Atakpa et al., 2018; Kilpatrick et al., 2013; Morgan, 2016; Ronco et al., 2015). To explore the role of InsP₃-dependent ER Ca²⁺ release in acidic Ca²⁺ signaling in ECFCs, we took advantage of CPA, an established inhibitor of sarco–ER Ca²⁺-ATPase activity, which impairs Ca²⁺ sequestration into the ER lumen and causes the depletion of ER Ca²⁺ pool. Pretreating ECFCs with CPA (10 μM) under 0Ca²⁺ conditions are sufficient to deplete the InsP₃-sensitive ER Ca²⁺ pool (Sanchez-Hernandez et al., 2010). This protocol significantly (p < .05) reduced the amplitude of GPN-induced endogenous Ca²⁺ release (Figures 2a,b and 2d), while it did not affect the percentage of responding cells (Figure 2c). Likewise, pretreating ECFCs with GPN (200 μM) under 0Ca²⁺ conditions significantly (p < .05) reduced the amplitude of the endogenous Ca²⁺ response to CPA (Figures 2a,b and 2d), although it did not affect the percentage of responding cells (Figure 2d). GPN was maintained throughout the recording as lysosomes readily reform upon GPN washout from the bath (Pandey et al., 2009). The same results were obtained when we compared the Ca²⁺ signals induced under 0Ca²⁺ conditions by nigericin (50 μM) and CPA (10 μM; Figure 3a) and then reversing the temporal order of agonist application (Figure 3b). The previous depletion of the ER Ca²⁺ pool with CPA did not affect the percentage of ECFCs responding to nigericin (Figure 3c) but caused a significant (p < .05) reduction in the amplitude of the endogenous Ca²⁺ mobilization (Figure 3d). Vice versa, previous depletion of the lysosomal Ca²⁺ store with nigericin did not affect the percentage of ECFCs responding to CPA (Figure 3c), but significantly (p < .05) reduced the amplitude of endogenous Ca²⁺ release (Figure 3d). Overall, these findings clearly indicate that a replete ER Ca²⁺ pool is necessary for the lysosomal Ca²⁺ store to generate a full increase in [Ca²⁺]_i in ECFCs.

InsP₃Rs may play an intermediary role in the Ca²⁺-dependent cross-talk between the lysosomal and ER Ca²⁺ stores (Atakpa et al., 2018; Churchill & Galione, 2001; Garrity et al., 2016; Kilpatrick et al., 2013). Therefore, we then measured the Ca²⁺ response to either GPN (200 μM) or nigericin (50 μM) in the presence of 2-aminoethoxydiphenyl borate (2-APB; 50 μM, 30 min), which selectively blocks InsP₃Rs in the absence of extracellular Ca²⁺ (Dragoni et al., 2011). This treatment significantly (p < .05) reduced both GPN- and nigericin-evoked increase in [Ca²⁺]_i under 0Ca²⁺ conditions (Figure 4), thereby confirming that functional InsP₃Rs are required to sustain intracellular Ca²⁺ signaling upon the liberation of the lysosomal Ca²⁺ store also in ECFCs.

3.3 TPC1 mediates NAADP-induced lysosomal Ca²⁺ release

A recent investigation revealed that liposomal delivery of NAADP (at 1:20 dilution) caused an increase in [Ca²⁺]_i that was prevented by NED-19 (Di Nezza et al., 2017), a selective TPC antagonist (Patel, 2015). In agreement with these findings, this NAADP-containing liposomal preparation induced a Ca²⁺ signal which decayed to the baseline under 0Ca²⁺ conditions, while restitution of extracellular Ca²⁺ did not evoke any detectable change in intracellular Ca²⁺ levels (Figure 5a). Therefore, NAADP does not elicit Ca²⁺ entry in ECFCs. The Ca²⁺ response to NAADP-containing liposomes displayed different kinetics, ranging from long lasting (≈1,500 s; Figure 5a) to transient (≈200 s; Figure 5b), which are likely to reflect different rates of NAADP release from the liposomes. To confirm that TPC1 mediates the intracellular Ca²⁺ response to NAADP, TPC1 was knocked down by using a specific siRNA (siTPC1), as recently shown in metastatic colorectal cancer cells (Faris et al., 2019). The efficacy of gene silencing on the expression of TPC1 protein has been displayed in Figure S1. The genetic suppression of TPC1 completely suppressed the Ca²⁺ response to NAADP, while ECFCs transfected with a scrambled construct were still responsive to this second messenger (Figure 5b,c). Resting Fura-2 fluorescence was not affected by genetic silencing of TPC1 (control: 0.120 ± 0.007 arbitrary units [a.u.], n = 301 vs. siTPC1: 0.109 ± 0.026 a.u., n = 102) Subsequently, we found that NAADP-evoked endogenous Ca²⁺ mobilization was impaired by preincubating the cells with either GPN (200 μM, 30 min) or nigericin (50 μM, 30 min; Figure 6a). Disruption of acidic Ca²⁺ stores significantly (p < .05) reduced the percentage of responding cells (Figure 6b) and the amplitude of the Ca²⁺ response to NAADP (Figure 6c). Taken together, these findings convincingly demonstrated that TPC1 mediates NAADP-induced lysosomal Ca²⁺ release in ECFCs.

3.4 InsP₃-dependent ER Ca²⁺ release is necessary for NAADP-induced lysosomal Ca²⁺ mobilization in ECFCs

In the next set of experiments, the role of InsP₃-dependent ER Ca²⁺ release in NAADP signaling was explored by exploiting the strategy introduced by Kilpatrick et al. (2013); Kilpatrick, Yates, Grimm, Schapira, and Patel, (2016). We found that pharmacological blockade of InsP₃Rs with 2-APB (50 μM, 30 min) significantly (p < .05) reduced the Ca²⁺ response to NAADP-containing liposomes (1:20), as illustrated in Figure 7a and summarized in Figure 3c. Likewise, depletion of the ER Ca²⁺ pool with CPA (10 μM, 30 min) in the absence of extracellular Ca²⁺ (0Ca²⁺) significantly (p < .05) decreased NAADP-induced intracellular Ca²⁺ release by diminishing the percentage of responding cells (Figure 7b,c). Overall, these observations endorse the notion that functional InsP₃Rs are required for the lysosomal Ca²⁺ store to generate a full increase in [Ca²⁺]_i in response to NAADP.

3.5 TPC1 triggers VEGF-induced proangiogenic Ca²⁺ oscillations, but not ATP-induced Ca²⁺ release

To assess the physiological role of NAADP-induced Ca²⁺ release in circulating ECFCs, we focused on the Ca²⁺ response to two distinct physiological autacoids, that is, ATP (100 μM) and VEGF (10 ng/ml), which trigger intracellular Ca²⁺ signals by, respectively, activating G-protein-coupled receptors (GPCRs) and tyrosine kinase receptors (Dragoni et al., 2011; Sanchez-Hernandez et al., 2010). ATP has long been used to assess the activation of GPCRs in circulating ECFCs in healthy and pathological ECFCs (Lodola et al., 2012; Sanchez-Hernandez et al., 2010; Zuccolo, Bottino et al., 2016). Depletion of the lysosomal Ca²⁺ pool with either GPN (200 μM, 30 min) or nigericin (50 μM, 30 min) induced a modest, but statistically significant, reduction in ATP-induced intracellular Ca²⁺ release (Figure 8a,b). However, the Ca²⁺ response to ATP was not affected by the pharmacological blockade of TPC1 with NED-19 (100 μM, 30 min; Figure 8c,d). Likewise, genetic silencing of NAADP-evoked lysosomal Ca²⁺ mobilization with the siTPC1 did not affect ATP-induced endogenous Ca²⁺ release (Figure 8d,e). VEGF, in turn, elicits long-lasting oscillations in [Ca²⁺]_i that stimulate ECFCs to undergo proliferation and tube formation (Dragoni et al., 2011, 2015; Lodola et al., 2017). Unlike ATP, pretreatment with either GPN (200 μM, 30 min) or nigericin (50 μM, 30 min) abolished the oscillatory Ca²⁺ response to VEGF in circulating ECFCs (Figures 9a–c and 9e). Moreover, VEGF-induced intracellular Ca²⁺ oscillations were fully prevented by pharmacological and genetic blockade of TPC1 with, respectively, NED-19 (100 μM, 30 min) and the siTPC1 (Figure 9d,e). Therefore, we examined the effect of TPC1 inhibition on VEGF-induced ECFC proliferation in the presence of EBM-2 and 5% FBS, as described by Dragoni et al. (2011, 2015) and Lodola et al. (2017). As described in Figure 9f, both NED-19 (100 μM, 30 min) and the siTPC1 prevented ECFCs from reaching the confluence after 5 days in culture in the presence of 10 ng/ml VEGF. Overall, these experiments demonstrate that NAADP-induced lysosomal Ca²⁺ release is selectively recruited by VEGF to trigger proangiogenic Ca²⁺ oscillations in ECFCs, whereas it does not support the Ca²⁺ response to ATP.

4.1 The acidic Ca²⁺ store is functionally coupled to the ER via InsP₃Rs in ECFCs

The acidic Ca²⁺ stores, including acidocalcisomes, lysosomes, lysosome-related organelles, secretory vesicles, and vacuoles (Morgan et al., 2011; Patel & Docampo, 2010), are emerging as an additional endogenous Ca²⁺ reservoir that contributes to Ca²⁺ signals in response to extracellular stimulation by establishing a bidirectional Ca²⁺-dependent cross-talk with the ER (Faris et al., 2018; Morgan, 2016; Morgan et al., 2011; Penny, Kilpatrick, Eden, & Patel, 2015). For instance, pharmacological disruption of the lysosomal Ca²⁺ store with GPN evokes long-lasting intracellular Ca²⁺ oscillations in human fibroblasts that are curtailed by emptying the ER Ca²⁺ store or inhibiting InsP₃Rs (Kilpatrick et al., 2013). Likewise, GPN- and nigericin-induced intracellular Ca²⁺ release in metastatic colorectal carcinoma cells were attenuated by preventing InsP₃-dependent ER Ca²⁺ release (Faris et al., 2019). Herein, we demonstrated for the first time in an endothelial cell type that the endogenous Ca²⁺ response to either GPN or nigericin was curbed by depleting the ER Ca²⁺ pool with CPA and inhibiting InsP₃Rs with 2-APB. This finding endorses the emerging view that acidic and neutral Ca²⁺ stores are intimately connected at the MCSs between lysosomal vesicles and ER cisternae that have been described in multiple cell types (Penny et al., 2015). This issue remains to be explored in endothelial cells, although mitochondria-associated ER membranes were recently reported in HUVECs (Yu et al., 2019). The Ca²⁺-dependent cross-talk between lysosomes and ER is largely interpreted within the framework of the “trigger hypothesis” originally put forward by Churchill and Galione (2001). These authors proposed that the Ca²⁺ response to multiple extracellular stimuli is triggered by local Ca²⁺ release from the acidic stores and then propagated into a regenerative Ca²⁺ wave by recruiting ER-embedded InsP₃Rs and RyRs through the CICR process (Churchill & Galione, 2001; Galione, 2015). Subsequently, it was found that lysosomal Ca²⁺ refilling was impaired by preventing ER-dependent Ca²⁺ release through InsP₃Rs (Garrity et al., 2016; Ronco et al., 2015) and that InsP₃Rs may concentrate at ER-lysosomes contact sites, thereby delivering Ca²⁺ directly into the lysosomal lumen (Atakpa et al., 2018). It turns out that when lysosomal Ca²⁺ uptake is prevented, the increase in [Ca²⁺]_i observed in response to ER store depletion is larger as compared to control cells (Lopez-Sanjurjo, Tovey, Prole, & Taylor, 2013). The finding that GPN and nigericin remarkably reduced ER Ca²⁺ release in response to CPA, as well to VEGF (see below), strongly suggests that acidic Ca²⁺ stores are more prone to trigger, rather than tempering, ER-derived Ca²⁺ signals in circulating ECFCs (Patel, 2019). Conversely, depletion of lysosomal Ca²⁺ did not result in extracellular Ca²⁺ entry, as also reported in human fibroblasts (Kilpatrick et al., 2013), mouse pancreatic acinar cells (Menteyne, Burdakov, Charpentier, Petersen, & Cancela, 2006), and metastatic colorectal cancer cells (Faris et al., 2019), while GPN-evoked Ca²⁺ entry has been reported in bovine corneal endothelial cells (Srinivas et al., 2002). The ER, which is seemingly recruited by lysosomal-derived Ca²⁺ in ECFCs, is a heterogeneous organelle that can be subdivided into distinct functional compartments (Berridge, 2002). For instance, a recent investigation revealed that InsP₃Rs are closely coupled to SOCE only when they are located in close proximity of the PM (Thillaiappan, Chavda, Tovey, Prole, & Taylor, 2017). It is, therefore, conceivable that in ECFCs, lysosomal Ca²⁺ release recruits a pool of InsP₃Rs which is located deeper within the cell and is, therefore, not coupled to SOCE (Taylor & Machaca, 2018).

4.2 NAADP-induced lysosomal Ca²⁺ release is mediated by TPC1 and requires InsP₃-dependent ER Ca²⁺ mobilization

NAADP has recently been shown to play a crucial role in endothelial Ca²⁺ signaling (Moccia, Berra-Romani et al., 2018). For instance, NAADP mediates acetylcholine-induced Ca²⁺ signaling and NO release in human aortic endothelial cells (Brailoiu et al., 2010) and human brain microvascular endothelial cells (Zuccolo et al., 2017). In addition, NAADP triggers Ca²⁺-dependent NO production also in human brain microvascular endothelial cells challenged with glutamate (Negri et al., 2019), histamine (Berra-Romani, Faris, Pellavio et al., 2019), and arachidonic acid (Berra-Romani, Faris, Negri et al., 2019). Likewise, NAADP drives glutamate-induced intracellular Ca²⁺ oscillations and NO production in mouse brain microvascular endothelial cells (Zuccolo, Kheder et al., 2019) as well as histamine-dependent Ca²⁺ release and vWF liberation in EA.hy926 cells (Esposito et al., 2011). Finally, a recent series of investigations revealed that NAADP promotes VEGF-induced Ca²⁺ release, proliferation, and tube formation in HUVECs and promoted angiogenesis in vivo (Favia et al., 2014, 2016; Moccia, Negri, Shekha et al., 2019). Distinct TPC isoforms were identified in mature endothelial cells depending on both species and vascular district under examination. TPC1 was the main isoform detected in mouse brain microvascular endothelial cells (Zuccolo, Kheder et al., 2019), while TPC2 was the most expressed isoform in HUVECs (Favia et al., 2014). Conversely, transcripts encoding for both TPC1 and TPC2 were found in human brain microvascular endothelial cells (Zuccolo, Laforenza et al., 2019). Preliminary observations showed that TPC1 messenger RNA was the most abundant isoform present in circulating ECFCs (Zuccolo, Dragoni et al., 2016), a finding that was confirmed by immunoblotting in the present investigation. The following pieces of evidence indicate that TPC1 mediates NAADP-evoked Ca²⁺ release from lysosomes and is then amplified into a regenerative Ca²⁺ wave by InsP₃Rs. First, NAADP-evoked intracellular Ca²⁺ mobilization was abolished by depletion of the lysosomal Ca²⁺ pool with either GPN or nigericin. Second, pharmacological (with NED-19) and genetic (through a specific siTPC1) blockade of TPC1 prevented the endogenous Ca²⁺ response to NAADP. Third, depletion of the ER Ca²⁺ reservoir with CPA and pharmacological inhibition of InsP₃Rs with 2-APB dramatically reduced NAADP-induced intracellular Ca²⁺ mobilization. Fourth, RyRs are absent in ECFCs (Sanchez-Hernandez et al., 2010; Zuccolo, Dragoni et al., 2016). As discussed above, ER Ca²⁺ mobilization is clearly tempered by lysosomal disruption. Therefore, it is reasonable to conclude that InsP₃Rs are necessary to amplify NAADP-induced local Ca²⁺ release in ECFCs, although we cannot rule out the possibility that a discrete pool of lysosomes is refilled by closely apposed InsP₃Rs in a silent manner. Early work disclosed that NAADP gates an inwardly-rectifying Ca²⁺-selective current which displayed biophysical properties similar to the Ca²⁺ release-activated Ca²⁺ current in starfish oocytes (Moccia, Billington et al., 2006; Moccia, Nusco, Lim, Kyozuka, & Santella, 2006). However, NAADP did not evoke extracellular Ca²⁺ entry in circulating ECFCs, thereby confirming that the acidic Ca²⁺ store recruited by this second messenger is uncoupled from SOCE. Collectively, these data demonstrate for the first time that EL Ca²⁺ release is functionally coupled to ER-embedded InsP₃Rs also in the endothelial lineage and suggest that the Ca²⁺-dependent interplay between TPC1 and InsP₃Rs may shape NAADP-induced endothelial Ca²⁺ signals.

4.3 NAADP mediates VEGF-induced intracellular Ca²⁺ oscillations and proliferation, but not ATP-induced Ca²⁺ release

NAADP-induced endothelial Ca²⁺ release may be selectively recruited by specific extracellular autacoids. For instance, while NAADP supported the Ca²⁺ response to acetylcholine in human aortic endothelial cells, ATP-induced Ca²⁺ signals were unaffected by bafilomycin A1 and NED-19 (Brailoiu et al., 2010). Likewise, pharmacological blockade of TPCs with NED-19 blocked histamine-, but not thrombin-induced Ca²⁺ release and vWF release in Ea.hy926 cells (Esposito et al., 2011). Therefore, we sought to assess the role of NAADP in the Ca²⁺ response to ATP and VEGF, which evoke intracellular Ca²⁺ signals in ECFCs by, respectively, activating purinergic P2Y receptors and VEGF receptor-2 (Dragoni et al., 2011; Sanchez-Hernandez et al., 2010). ATP-evoked intracellular Ca²⁺ mobilization was weakly affected by depleting the lysosomal Ca²⁺ pool with either GPN or nigericin, but it was unaffected by pharmacological (with NED-19) and genetic (with the selective siTPC1) blockade of TPC1. These findings demonstrate that TPC1 does not contribute to ATP-induced increase in [Ca²⁺]_i in circulating ECFCs, while the modest inhibition exerted by GPN and nigericin reflects the mobilization of the ER Ca²⁺ pool. As already mentioned above, this finding also shows that lysosomes do not buffer Ca²⁺ delivered into the cytosol through InsP₃Rs. We then focused on VEGF, which stimulates circulating ECFCs to proliferate and assembly into bidimensional tubular networks in Matrigel by inducing repetitive oscillations in [Ca²⁺]_i (Dragoni et al., 2011, 2015; Lodola et al., 2017). As reported in HUVECs for TPC2 (Favia et al., 2014), the Ca²⁺ response to VEGF in ECFCs was fully prevented by pharmacological and genetic deletion of TPC1. In addition, VEGF failed to induce intracellular Ca²⁺ oscillations upon depletion of the lysosomal Ca²⁺ store with nigericin and GPN. These findings strongly suggest that NAADP triggers repetitive Ca²⁺ spikes by stimulating TPC1 to mediate lysosomal Ca²⁺ release. Previous work revealed that VEGF-induced intracellular Ca²⁺ oscillations are shaped by rhythmical Ca²⁺ release through InsP₃Rs and are sustained over time by SOCE (Dragoni et al., 2011; Lodola et al., 2017). On the basis of evidence provided in the present investigation, the oscillatory Ca²⁺ activity could be initiated by NAADP via TPC1 and then transformed into global Ca²⁺ waves by InsP₃Rs, while SOCE prevents the ER (and possibly the EL) Ca²⁺ pool from depleting. The functional interaction between NAADP, InsP₃, and SOCE is necessary to promote in vitro angiogenesis as the pharmacological and genetic blockade of TPC1 prevents VEGF-induced ECFC proliferation. These findings extend to circulating ECFCs the requirement of NAADP signaling for VEGF to trigger a proangiogenic signal. It has recently proposed that genetic manipulation of the Ca²⁺ toolkit could represent an alternative strategy to boost the reparative potential of autologous ECFCs and improve the regenerative outcome of cell based therapy of ischemic diseases (Moccia, Berra-Romani et al., 2018; Moccia, Lucariello et al., 2018). Future work could challenge this hypothesis by assessing whether TPC1 overexpression in ECFCs increases their angiogenic activity in preclinical models of hindlimb ischemia or acute myocardial infarction. Interestingly, preliminary evidence demonstrated that liposomal delivery of NAADP was per se able to deliver a Ca²⁺-dependent mitogenic signal to circulating ECFCs (Di Nezza et al., 2017).

4.4 Limitations of the investigation

The present investigation demonstrated that NAADP activates TPC1 to induce intracellular Ca²⁺ release and initiated VEGF-induced proangiogenic Ca²⁺ oscillations in circulating ECFCs. A number of caveats, however, will certainly require future investigation.

First, single-channel analysis on lipid bilayers revealed that TPC1 displays the following permeability sequence: H⁺ >> K⁺ > Na⁺ ≥ Ca²⁺ (Pitt, Lam, Rietdorf, Galione, & Sitsapesan, 2014; Rybalchenko et al., 2012). Thus, its Ca²⁺ permeability is limited as compared to TPC2 (Pitt et al., 2010). Nevertheless, as recently reviewed (Galione, 2019), even a small TPC1-mediated Ca²⁺ flux could be sufficient to generate local Ca²⁺ microdomains that are then amplified by the Ca²⁺-dependent recruitment of adjoining InsP₃Rs. Indeed, assuming that TPC1-mediated Na⁺ currents drive EL Ca²⁺ release is unlike because of the unfavorable effect of the following lysosomal membrane depolarization on Ca²⁺ egress into the cytoplasm (Galione, 2019; Morgan & Galione, 2014). Patch-clamping enlarged lysosomes is required to confirm that TPC1 conducts intraluminal Ca²⁺ also in circulating ECFCs.

Second, NAADP-induced increase in Fura-2 fluorescence was fully abolished despite the 60% reduction in TPC1 expression achieved by the specific siTPC1 employed in the present investigation. Three mechanisms could be invoked to explain this seemingly contradictory result. First, the remaining TPC1 protein (≈40%) is not expressed in EL membranes. Second, our epifluorescence imaging system does not detect subcellular increases in intracellular Ca²⁺ concentration. Thus, we cannot rule out that some TPC1-mediated Ca²⁺ release events still occurred but were missed by our charge-coupled device camera. This hypothesis would be further consistent with the notion that, although lysosomes contain ≈500 µM free Ca²⁺, they occupy ≈3% of the total endothelial cell volume and are sparse throughout the cytoplasm (Lloyd-Evans & Waller-Evans, 2019). Thus, reducing TPC1 expression by 60% could lead to discrete subcellular Ca²⁺ signals that are detectable only by confocal or 2-photon microscopy (Boittin, Galione, & Evans, 2002; Kinnear et al., 2008), especially if the remaining TPC1 are uncoupled from the ER. For instance, NAADP was found to induce discrete Ca²⁺ signals that did not significantly increase Fura-2 fluorescence until they let to ER Ca²⁺ mobilization in rat pulmonary artery vascular smooth muscle cells (Kinnear et al., 2008). Third, TPC1 is sensitive to cytosolic Ca²⁺ (Lagostena, Festa, Pusch, & Carpaneto, 2017; Pitt et al., 2014). It has been postulated that upon NAADP-induced EL Ca²⁺ release, the local elevation in cytosolic Ca²⁺ could prolong TPC1 activation (Pitt et al., 2014). Thus, even a 60% reduction in TPC1 expression could reduce the Ca²⁺ response to NAADP to such an extent to ablate the Ca²⁺-dependent activation of the remaining TPC1 proteins (Pitt et al., 2014), which would, therefore, become unresponsive to NAADP stimulation. This hypothesis is further supported by the earlier report that NAADP-induced EL Ca²⁺ release in sea urchin eggs was ablated by chelating cytosolic Ca²⁺ with BAPTA (Morgan et al., 2013).

Finally, it is still unclear how NAADP gates TPCs. It has been suggested that NAADP acts through the interposition of a NAADP-binding protein, which could enable this versatile second messenger to activate either TPCs or RyR1 (Guse, 2012). Photoaffinity labeling studies revealed that a doublet of 22/23 kDa accessory proteins, rather than TPCs, bind to NAADP in naive mammalian cells (Lin-Moshier et al., 2012). It is, however, still unclear whether these NAADP-binding proteins are also present in circulating ECFCs.