The human amniotic fluid stem cell secretome triggers intracellular Ca2+ oscillations, NF‐κB nuclear translocation and tube formation in human endothelial colony‐forming cells

Abstract Second trimester foetal human amniotic fluid‐derived stem cells (hAFS) have been shown to possess remarkable cardioprotective paracrine potential in different preclinical models of myocardial injury and drug‐induced cardiotoxicity. The hAFS secretome, namely the total soluble factors released by cells in their conditioned medium (hAFS‐CM), can also strongly sustain in vivo angiogenesis in a murine model of acute myocardial infarction (MI) and stimulates human endothelial colony‐forming cells (ECFCs), the only truly recognized endothelial progenitor, to form capillary‐like structures in vitro. Preliminary work demonstrated that the hypoxic hAFS secretome (hAFS‐CMHypo) triggers intracellular Ca2+ oscillations in human ECFCs, but the underlying mechanisms and the downstream Ca2+‐dependent effectors remain elusive. Herein, we found that the secretome obtained by hAFS undergoing hypoxic preconditioning induced intracellular Ca2+ oscillations by promoting extracellular Ca2+ entry through Transient Receptor Potential Vanilloid 4 (TRPV4). TRPV4‐mediated Ca2+ entry, in turn, promoted the concerted interplay between inositol‐1,4,5‐trisphosphate‐ and nicotinic acid adenine dinucleotide phosphate‐induced endogenous Ca2+ release and store‐operated Ca2+ entry (SOCE). hAFS‐CMHypo‐induced intracellular Ca2+ oscillations resulted in the nuclear translocation of the Ca2+‐sensitive transcription factor p65 NF‐κB. Finally, inhibition of either intracellular Ca2+ oscillations or NF‐κB activity prevented hAFS‐CMHypo‐induced ECFC tube formation. These data shed novel light on the molecular mechanisms whereby hAFS‐CMHypo induces angiogenesis, thus providing useful insights for future therapeutic strategies against ischaemic‐related myocardial injury.


| INTRODUC TI ON
Endothelial colony-forming cells (ECFCs) represent vasculogenic cells that are truly committed to the endothelial lineage and, throughout postnatal life, are mobilized upon ischaemic injury to restore the damaged vascular network. 1,2 ECFCs possess a high clonogenic potential, form capillary-like structures in in vitro Matrigel tubulogenesis assays, integrate into pre-existing vasculature and rescue local blood perfusion in murine models of ischaemia. 2,3 Therefore, ECFCs hold remarkable promise as the most suitable cellular substrate to induce therapeutic angiogenesis in ischaemic disorders, such as acute myocardial infarction (AMI), peripheral artery disease, and stroke. [2][3][4] Genetic manipulation and pharmacological conditioning could be exploited to improve ECFCs' angiogenic activity and/or to improve their survival/engraftment within the harsh microenvironment of the ischaemic tissue. 2,5 Mesenchymal stromal progenitors obtained from leftover samples of second trimester amniotic fluid for prenatal diagnosis have been lately described as appealing therapeutics in several preclinical models of disease.
In particular, human amniotic fluid-derived stem cells (hAFS) have been shown to express a remarkable pro-angiogenic paracrine potential. 6,7 A more recent investigation showed that the hAFS secretome collected under hypoxic conditions stimulated circulating ECFCs to assemble into bidimensional capillary-like networks in vitro through an oscillatory increase in intracellular Ca 2+ concentration ([Ca 2+ ] i ) 8,9 ; furthermore, it actively induced local angiogenesis and promoted cardiac repair in a murine model of AMI. 8,9 Therefore, the local injection of the hypoxic hAFS secretome could represent an alternative strategy to recruit circulating ECFCs towards the damaged myocardium and induce therapeutic angiogenesis. 2 The paracrine therapy of AMI would be promptly accessible to the patients and could overcome many of the drawbacks associated with cell-based therapy, including the time-consuming procedure and the requirement for huge amounts of cells. 10,11 Clinical translation of this approach would benefit from the elucidation of the signalling pathways whereby the hAFS secretome triggers the pro-angiogenic Ca 2+ response in ECFCs.

| hAFS secretome separation and concentration
The total cell secretome, as represented by the cell-conditioned medium (hAFS-CM) from either hAFS in control normoxic condition (hAFS-CM Normo ) or hypoxic preconditioning (hAFS-CM Hypo ), was collected as previously described. 8,9,20 Briefly, hAFS-CM formulations were centrifuged to remove cell debris and further concentrated using ultrafiltration membranes with a 3 kDa selective cut-off (Amicon Ultra-15; Millipore) at 4°C first at 3000 ×g for 90' and then at 3000 ×g for additional 30'. hAFS-CM protein concentration was assessed by BiCinchoninic Acid (BCA) assay (Gibco-Thermo Fisher Scientific). hAFS-CM was used for in vitro experiments as 80 mg/ mL solution to be added to the cell culture medium as from previous studies. 8 Hepes. In Ca 2+ -free solution (0Ca 2+ ), Ca 2+ was substituted with 2 mmol/L NaCl, and 0.5 mmol/L EGTA was added. Solutions were titrated to pH 7.4 with NaOH. The osmolality of PSS as measured with an osmometer (Wescor 5500) was 338 mmol/kg.

| [Ca 2+ ] i measurements
Endothelial colony-forming cells were loaded with 4 µmol/L fura-2 acetoxymethyl ester (fura-2/AM; 1 mmol/L stock in dimethyl sulfoxide) in PSS for 1 hour at room temperature. The details of the Ca 2+ recording set-up have been described in REF 9,14 and are reported in the Supplementary Information. All the experiments were performed at room temperature. All the data have been collected from ECFCs isolated from peripheral blood of at least three healthy volunteers.

| Immunofluorescence
Twenty-four hours before treatment with hAFS-CM Hypo and the specific blockers of intracellular Ca 2+ signalling, 6 × 10 4 ECFCs were plated onto 13 mm coverslips in 24-well plates. ECFCs were fixed in 4% formaldehyde in PBS for 15 minutes at room temperature, permeabilized for 7 minutes in PBS with 0.1% Triton X-100 and blocked for 30 minutes in 2% gelatin. Then, primary (incubated for 1 hour at 37°C) and secondary (incubated for 1 hour at room temperature) antibodies were applied in PBS with 2% gelatin. The primary anti-p65 (NF-κB subunit) antibody specific for immunocytochemistry (Santa Cruz Biotechnology, catalog no.  All the chemicals were of analytical grade and obtained from Sigma Chemical Co.

| Statistics
All the data have been collected from ECFCs deriving from at least three distinct donors. Pooled data are given as mean ± SE and statistical significance (P < .05) was evaluated by Student's t test or one-way ANOVA followed by the post hoc Dunnett's test as appropriate. Data relative to Ca 2+ signals are presented as mean ± SE, while the number of cells analysed is indicated in the corresponding bar histograms.

| Extracellular Ca 2+ entry triggers hAFS-CM Hypoinduced intracellular Ca 2+ oscillations in circulating ECFCs
Human amniotic fluid-derived stem cells medium conditioned under hypoxia (hAFS-CM Hypo ) ( Figure 1A), but not under normoxia However, the spiking Ca 2+ signal promptly resumed upon Ca 2+ restitution to the recording solution ( Figure 2B). Therefore, extracellular Ca 2+ entry was required to trigger hAFS-CM Hypo -induced intracellular Ca 2+ oscillations. SOCE, which can be recruited by a spatially restricted InsP 3 -induced ER Ca 2+ pulse undetectable by epifluorescence imaging, 26,27 represents the main Ca 2+ -entry pathway in circulating ECFCs. 28 The pyrazole derivative BTP-2 has been shown to specifically inhibit SOCE in ECFCs. 29,30 Pre-treating the cells with BTP-2 (10 µmol/L, 20 minutes) did not prevent the onset of the Ca 2+ response to hAFS-CM Hypo ( Figure 2C,D), but significantly (P < .05) reduced the percentage of oscillating cells ( Figure 2E) and the amplitude of the 1st Ca 2+ spike ( Figure 2F). Furthermore, BTP-2 curtailed the frequency of the Ca 2+ transients to 1-2 oscillations/h ( Figure 2G). These data demonstrate that SOCE is required to maintain the oscillations over time but is not responsible for the onset of the Ca 2+ response to hAFS-CM Hypo . Therefore, a store-independent Ca 2+ -permeable route initiates the oscillatory signal recorded in the presence of extracellular Ca 2+ , as previously shown in UCB-derived ECFCs stimulated with VEGF. 13

| hAFS-CM Hypo -induced intracellular Ca 2+ oscillations are triggered by TRP Vanilloid 4 (TRPV4) and shaped by InsP 3 receptors (InsP 3 Rs) and TPC1
Phospholipase C (PLC) plays a crucial role in the onset of the Ca 2+ response to chemical stimulation in circulating ECFCs and vascular endothelial cells. 28 Figure 3C,D). Similarly, the spiking Ca 2+ signal was prevented F I G U R E 2 Extracellular Ca 2+ entry triggers hypoxic human amniotic fluid-derived stem cell (hAFS) secretome-induced intracellular Ca 2+ oscillations in circulating endothelial colony-forming cells (ECFCs). A, intracellular Ca 2+ oscillations induced by hypoxic hAFS secretome (hAFS-CM Hypo ) in the presence of external Ca 2+ . The horizontal bar above the Ca 2+ tracings indicates the application period of hAFS-CM Hypo . B, extracellular Ca 2+ was removed (0 Ca 2+ ) at 200 s, and hAFS-CM Hypo was applied at 300 s from the beginning of the recording. Intracellular Ca 2+ oscillations were not recorded under 0Ca 2+ conditions. Extracellular Ca 2+ was restored at 1500 s, thereby resuming the spiking Ca 2+ response. Horizontal bars above the Ca 2+ tracings indicate the application period of hAFS-CM Hypo and of physiological salt solution (PSS) supplemented (Ca 2+ ) or not (0 Ca 2+ ) with Ca 2+ . C, 10-min pre-incubation with Pyr6 (10 μmol/L), a selective inhibitor of store-operated Ca 2+ entry (SOCE), curtailed, but did not prevent, the onset of the Ca 2+ response to hAFS-CM Hypo . The horizontal bars above the Ca 2+ tracings indicate the application period of hAFS-CM Hypo and Pyr6. D, mean ± SE of the percentage of cells displaying a Ca 2+ response to hAFS-CM Hypo under the designated treatments. E, mean ± SE of the percentage of cells displaying intracellular Ca 2+ oscillations (ie more than one Ca 2+ transient) in response to hAFS-CM Hypo under the designated treatments. F, mean ± SE of the amplitude of the 1st Ca 2+ spike elicited by hAFS-CM Hypo under the designated treatments. G, mean ± SE of the intracellular Ca 2+ transients elicited by hAFS-CM Hypo under the designated treatments. * indicates P < .05 (Student's t test). NR, no response by pre-treating the cells with ruthenium red (10 µmol/L, 10 minutes) ( Figure 3E), a less specific TRPV4 inhibitor. 34 Collectively, these data provide the evidence that, upon PLC engagement, DAG gates TRPV4 to mediate the influx of extracellular Ca 2+ that triggers the rhythmical Ca 2+ -dependent recruitment of InsP 3 Rs. To further corroborate this model, hAFS-CM Hypo was administered to circulating ECFCs

| hAFS-CM Hypo induces the nuclear translocation of NF-κB in a Ca 2+ -dependent manner
The transcription factor NF-κB has long been known to translate intracellular Ca 2+ signals in a pro-angiogenic output in ECFCs. 12,22,40 In quiescent cells, the p65 NF-κB subunit is retained in the cytoplasm by the physical association with the inhibitory IκB protein, but is released from inhibition and primed to translocate into the nucleus by an oscillatory increase in [Ca 2+ ] i . 15,40 Immunofluorescence revealed that p65 NF-κB displayed a cytosolic distribution in non-stimulated F I G U R E 3 The role of transient receptor potential vanilloid 4 (TRPV4), InsP 3 receptors (InsP 3 Rs) and two-pore channel 1 (TPC1) in hypoxic human amniotic fluid-derived stem cell (hAFS) (hAFS) secretome-induced intracellular Ca 2+ oscillations in circulating endothelial colony-forming cells (ECFCs). A, intracellular Ca 2+ oscillations induced by the hypoxic hAFS secretome (hAFS-CM Hypo ) in circulating ECFCs under control conditions, that is in the presence of extracellular Ca 2+ and in the absence of inhibitors of the Ca 2+ signalosome. The horizontal bar above the Ca 2+ tracings indicates the application period of hAFS-CM Hypo . B, 30-min pre-incubation with U73122 (10 μmol/L), an antagonist of phospholipase C (PLC), and 10-min pre-incubation with Xestospongin C (XeC; 1 μmol/L, 10 min), a blocker of InsP 3 Rs, prevented the oscillatory response to hAFS secretome. The horizontal bar above the Ca 2+ tracings indicates the application period of hAFS-CM Hypo . C, 10-min pre-incubation with RHC-80267 (RHC; 50 μmol/L), which selectively interferes with diacylglycerol (DAG) lipase activity, suppressed hAFS-CM Hypo -induced intracellular Ca 2+ oscillations in circulating ECFCs. The horizontal bars above the Ca 2+ tracings indicate the application period of hAFS-CM Hypo and RHC. D, 30-min pre-incubation with RN-1734 (20 μmol/L), a selective TRPV4 blocker, prevented the oscillatory Ca 2+ response to hAFS-CM Hypo in circulating ECFCs. The horizontal bars above the Ca 2+ tracings indicate the application period of hAFS-CM Hypo and RN-1734. E, 10-min preincubation with ruthenium red (10 μmol/L), a pan-specific inhibitor of TRPV channels, also suppressed hAFS-CM Hypo -induced intracellular Ca 2+ oscillations in circulating ECFCs. The horizontal bars above the Ca 2+ tracings indicate the application period of hAFS-CM Hypo and ruthenium red. F, ECFCs were pretreated for 30 min with cyclopiazonic acid (CPA; 10 μmol/L) to fully deplete the endoplasmic reticulum (ER) Ca 2+ reservoir and activate store-operated Ca 2+ entry (SOCE). Thereafter, hAFS-CM Hypo was added and caused a transient reduction in intracellular Ca 2+ levels followed by a sustained increase in [Ca 2+ ] i . 30-min pre-incubation with RN-1734 (20 μmol/L) to block TRPV4 prevented this [Ca 2+ ] i rise and unmasked the progressive decrease in Fura-2 fluorescence, which reflects AA-dependent SOCE inhibition (please, see the text for a wider explanation). The horizontal bars above the Ca 2+ tracings indicate the application period of hAFS-CM Hypo and CPA. G, 30-min pre-incubation with NED-19 (100 µmol/L), a selective two-pore channel (TPC) blocker, and 30-min pre-incubation with the lysosomotropic agent, glycyl-lphenylalanine 2-naphthylamide (GPN; 200 µmol/L), inhibited the oscillatory Ca 2+ response to hAFS-CM Hypo . The horizontal bar above the Ca 2+ tracings indicates the application period of hAFS-CM Hypo F I G U R E 4 Statistical analysis of phospholipase C (PLC) and nicotinic acid adenine dinucleotide phosphate (NAADP) signalling. Mean ± SE of the percentage of endothelial colony-forming cells (ECFCs) presenting a Ca 2+ response (A) and, among these, of ECFCs presenting intracellular Ca 2+ oscillations (ie more than one Ca 2+ transient) (B) upon exposure to the hypoxic human amniotic fluid-derived stem cell (hAFS) secretome (hAFS-CM Hypo ) under the designated treatments. Mean ± SE of the amplitude of the 1st Ca 2+ spike (C) and of the intracellular Ca 2+ transients (D) elicited by hAFS-CM Hypo under the designated treatments. Mean ± SE of the percentage of ECFCs presenting a Ca 2+ response (E) and, among these, of ECFCs presenting intracellular Ca 2+ oscillations (ie more than one Ca 2+ transient) (F) when exposed to hAFS-CM Hypo in the absence (Ctrl) and presence of blockers of transient receptor potential vanilloid 4 (TRPV4) signalling. * indicates P < .05 (Student's t test). NR, no response ECFCs (Ctrl; Figure 5A,B), whereas it was mainly accumulated in the nucleus upon exposure to hAFS-CM Hypo (Figure 5A,B). Conversely, pre-treating the cells with RN-1734 (20 µmol/L, 10 minutes), which suppresses the Ca 2+ spikes, or BAPTA (30 µmol/L, 2 hours), a membrane-permeable Ca 2+ buffer, 9,12 significantly (P < .05) inhibited hAFS-CM Hypo -induced nuclear translocation of p65 NF-κB ( Figure 5A,B). These data, therefore, demonstrate that intracellular Ca 2+ oscillations drive hAFS-CM Hypo -induced p65 NF-κB translocation into the nucleus in circulating ECFCs.

| NF-κB drives hAFS-CM Hypo -induced ECFC tubulogenesis
A recent report from our group demonstrated that hAFS-CM collected under hypoxic conditions specifically induced ECFC tubulogenesis in vitro. 8,9 Unlike other in vitro assays, for example migration and invasion, the Matrigel-based tube formation assay involves all the main physiological steps of the angiogenic process, including endothelial cell proliferation, adhesion, migration and differentiation. 41 We, therefore, evaluated both topologic (number of meshes and junctions per picture) and dimensional (total number of tubules per picture) of the capillary-like networks formed by circulating ECFCs placed in a Matrigel scaffold, as shown elsewhere. 19,21 Preventing the oscillatory increase in [Ca 2+ ] i with BAPTA has previously been shown to interfere with hAFS-CM Hypo -induced ECFC assembly in a bidimensional tubular network. 8,9 Likewise, ECFCs did not originate  (Figure 6), a selective NF-κB blocker. 12,42 Collectively, these data show that hAFS-CM Hypo requires TRPV4mediated extracellular Ca 2+ entry to trigger the nuclear translocation of p65 NF-κB and promote ECFC tubulogenesis.

| D ISCUSS I ON
Paracrine therapy through stem cells-secreted mediators is emerging as an alternative, promising strategy to treat AMI by instructing resident cells, for example cardiac stromal cells, cardiomyocytes and endothelial cells, to optimize endogenous mechanism of cardiac repair. 10

| hAFS-CM Hypo induces intracellular Ca 2+ oscillations by activating TRPV4
Human amniotic fluid-derived stem cells -conditioned medium entry and/or endogenous Ca 2+ release. 24,25,50,51 Removal of extracellular Ca 2+ prevented the onset of hAFS-CM Hypo -induced intracellular Ca 2+ oscillations, which promptly resumed upon restitution of external Ca 2+ . This finding further supports the notion that VEGF is not involved in the spiking Ca 2+ signal. Indeed, VEGF is still able to trigger 1-4 intracellular Ca 2+ spikes in the absence of extracellular Ca 2+ entry in circulating ECFCs. 12 Nonetheless, early reports demonstrated that extracellular Ca 2+ entry through two distinct DAG-sensitive pathways, that is TRPC1 and TRPC3, induced intracellular Ca 2+ oscillations, respectively, in primary myelofibrosis-derived ECFCs 21 and UCB-derived ECFCs. 13 The circulating ECFCs employed in the present investigation do not express TRPC3. 13 Furthermore, in these cells, TRPC1 is not sensitive to DAG, 21 but is part of a super-molecular complex including also Orai1 and STIM1, which is assembled upon depletion of the ER Ca 2+ store. 28,52 However, DAG may be converted by DAG lipase in AA, 53 which selectively gates extracellular Ca 2+ entry through TRPV4 in circulating ECFCs 31,54 and vascular endothelial cells. 23,25 The following pieces of evidence demonstrate that TRPV4-mediated extracellular Ca 2+ entry triggers hAFS-CM Hypo -induced intracellular Ca 2+ oscillations in circulating ECFCs. First, the pharmacological blockade of TRPV4 with two structurally distinct inhibitors, RN-1734 and ruthenium red, suppressed the onset of the oscillatory F I G U R E 6 Hypoxic human amniotic fluid-derived stem cell (hAFS) secretome induces endothelial colony-forming cell (ECFC) tubulogenesis in Ca 2+ -and NF-κB-dependent manner. Tubulogenesis assay on ECFCs plated in the presence of EBM-2 supplemented with 2% foetal bovine serum (FBS) and treated with or without 80 µg/mL of the secretome obtained from hAFS after hypoxic conditioning (hAFS-CM Hypo ) supplemented or not with RN-1734 (RN; 20 μmol/L, 30-min pre-incubation before stimulation) or thymoquinone (25 µmol/L, 10-min pre-incubation before stimulation). Digital images of endothelial tubes were obtained by bright-field light microscopy 10 h after plating cells on Matrigel-coated wells; scale bar: 50 µm. B, mean ± SE of the following parameters evaluated from digital images: number of master tubules (TLSs)/picture (Ba), number of meshes/picture (Bb), number of master junctions/picture (Bc). *** indicate P < .0001 (oneway ANOVA followed by the post hoc Dunnett's test). hAFS-CM Hypo , RN and thymoquinone were maintained during the tubulogenic assay signal. RN-1734 is a selective TRPV4 antagonist, 34 and it does not inhibit the other TRPV isoform expressed in circulating ECFCs, that is TRPV1. 22,54 Second, U73122 and RHC-80267, which, respectively, inhibit PLC activity 12 and DAG lipase, 53 also prevented the oscillatory Ca 2+ response to hAFS-CM Hypo . Interestingly, the intracellular Ca 2+ signals evoked by MCP-1, which is quite abundant in hAFS-CM, 8 also require DAG metabolism by DAG lipase in human monocytes. 49 Third, the pharmacological blockade of Orai1, which contributes to SOCE, 28,52 curtailed, but did not abrogate, the number of Ca 2+ transients evoked by hAFS-CM Hypo . Therefore, while SOCE is required to maintain the intracellular Ca 2+ oscillations by reloading the ER with Ca 2+ in a SERCA-dependent manner, 12

| hAFS-CM Hypo promotes in vitro tubulogenesis by inducing the nuclear translocation of NF-κB
Endothelial Ca 2+ oscillations support in vitro tubulogenesis and neovessel formation in vivo. 24  Of note, preventing the intracellular Ca 2+ oscillations with RN-1734 and the downstream recruitment of NF-κB with thymoquinone potently inhibited hAFS-CM Hypo -induced ECFC tube formation. This finding is strongly supported by the evidence that multiple pro-angiogenic genes, for example intercellular adhesion molecule 1, selectin E and various matrix metalloproteinases, are expressed upon the Ca 2+ -dependent activation of NF-κB in circulating ECFCs. 22,59 Furthermore, NF-κB regulates the expression of a large array of pro-angiogenic genes, for example those encoding for growth factors (eg VEGF), chemokines and cell adhesion molecules. 59 Thus, the Ca 2+ -dependent engagement of NF-κB is likely to play a crucial role in hAFS-CM Hypo -induced revascularization in murine models of AMI 8 and hindlimb ischaemia. 11 The involvement of other pro-angiogenic signalling pathways, such as phosphoinositide 3-kinase/Akt, which can also be activated by hAFS-CM Hypo20 and is sensitive to Ca 2+ in ECFCs, 18 cannot be ruled out.

| CON CLUS IONS
The present investigation reveals for the first time the signalling pathways whereby hAFS-CM Hypo induces pro-angiogenic Ca 2+ oscillations in circulating ECFCs, which represent the most suitable cellular substrate to achieve therapeutic angiogenesis in ischae-

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

F I G U R E 7
The mechanisms leading to the onset of hypoxic human amniotic fluid-derived stem cell (hAFS) secretome-induced intracellular Ca 2+ oscillations in endothelial colony-forming cells (ECFCs). Exposure of circulating ECFCs to hypoxic hAFS secretome (hAFS-CM Hypo ) results in phospholipase C (PLC) engagement, followed by production of diacylglycerol (DAG) and inositol-1,4,5-trisphosphate (InsP 3 ). DAG is converted by DAG lipase into arachidonic acid (AA), which gates transient receptor potential vanilloid 4 (TRPV4) to mediate extracellular Ca 2+ entry through the plasma membrane. InsP 3 primes ER-embedded to be activated by the incoming Ca 2+ . Nicotinic acid adenine dinucleotide phosphate (NAADP)-induced endolysosomal (EL) Ca 2+ release mediated by two-pore channel 1 (TPC1) is also likely to contribute to the Ca 2+ -dependent recruitment of InsP 3 receptors (InsP 3 Rs). Endoplasmic reticulum (ER) Ca 2+ depletion, in turn, leads to store-operated Ca 2+ entry (SOCE) activation and maintenance of intracellular Ca

DATA AVA I L A B I L I T Y S TAT E M E N T
All the data are fully available upon reasonable request.