Micro bicinchoninic acid (BCA) protein assay reagents were from Pierce (Micro BCA™ Protein Assay Kit; Thermo Fisher Scientific, Inc.). The polyclonal antibodies against Snail (3879S), FOXC2 (12974S), and Fibronectin (30903S), as well as the monoclonal antibody against GAPDH (D4C6R) were all from Cell Signaling Technology. HRP-conjugated donkey anti-rabbit and anti-mouse immunoglobulin (Ig) G secondary antibodies were from Jackson ImmunoResearch Laboratories. All other reagents were from Sigma-Aldrich; Merck KGaA.

Human bone marrow-derived mesenchymal stromal/stem cells (MSCs, PCS-500-012) were purchased from the American Type Culture Collection. Cell culture media was from Life Technologies Corp. Cells were plated in high glucose aMEM supplemented with 10% FBS and 50 units/ml penicillin/streptomycin and cultured in a humidified incubator at 37°C with 5% CO₂. MSCs were kept subconfluent and expanded for not more than 10 passages by a 1:2 split on a weekly basis. VM was assessed in vitro using Cultrex (3432-010-01, R&D Systems) to monitor 3D capillary-like structures formation (24). In brief, each well of a 96-well plate was pre-coated with 50 µl of Cultrex. MSCs suspension in culture media (10⁴ cells/100 µl) was then seeded on top of polymerized Cultrex and incubated at 37°C in a CO₂ incubator for different time points of vascular network formations. Phase contrast pictures were taken over time using a digital camera coupled to an inverted microscope. For each loop and tube measurement, the pixels that belong to its edge are considered its border or perimeter. The number of loops and tubes, as well as tube branching formed by the cells were quantified using the Wimasis analysis software (https://www.wimasis.com; Cordoba, Spain) or the ImageJ software (https://imagej.net) (25).

Total RNA was extracted from cell monolayers using 1 ml of TRIzol reagent for a maximum of 3×10⁶ cells as recommended by the manufacturer (Life Technologies). For cDNA synthesis, 2 µg of total RNA was reverse-transcribed using a high-capacity cDNA reverse transcription kit (Applied Biosystems). The cDNA was stored at −20°C prior to PCR. Gene expression was quantified qPCR using iQ SYBR-Green Supermix (Bio-Rad Laboratories). DNA amplification was carried out using an Icycler iQ5 (Bio-Rad Laboratories) and product detection was performed by measuring the binding of the fluorescent dye SYBR-Green I to double-stranded DNA. The following primer sets were from Qiagen: FOXC2 (Hs_FOXC2_1_SG, QT00220871), SNAI1 (Hs_SNAI1_SG, QT00010010), Fibronectin (Hs_FN1_1_SG, QT00038024), TGFβ (Hs_TGFB1_1_SG, QT00000728), GAPDH (Hs_GAPDH_1_SG, QT00079247) and Peptidylprolyl Isomerase A (PPIA) (Hs_PPIA_4_SG, QT01866137). The relative quantities of target gene mRNA were normalized against internal housekeeping genes PPIA and GAPDH. The RNA was measured by following a ∆C_T method employing an amplification plot (fluorescence signal vs. cycle number). The difference (∆C_T) between the mean values in the triplicate samples of the target gene and the housekeeping genes was calculated with the CFX manager Software version 2.1 (Bio-Rad Laboratories) and the relative quantified value (RQV) was expressed as 2^−ΔΔCq (26). Single amplicons and appropriate melting curves were indicative of specific qPCR conditions and efficacy (not shown).

For gene silencing experiments, MSCs were transiently transfected with siRNA sequences using Lipofectamine-2000 transfection reagent (Thermo Fisher Scientific, Inc.). Gene silencing was performed over 24 h using 20 nM siRNA against TGFβ (Hs_TGFB1_2 FlexiTube siRNA Geneglobe ID: SI00013601), Smad2 (Hs_SMAD2_1 FlexiTube siRNA GeneGlobe ID: SI00082460), Smad3 (Hs_SMAD3_1 FlexiTube siRNA GeneGlobe ID: SI00082481), or scrambled sequences (AllStar Negative Control siRNA, 1027281). The above small interfering RNA and mismatch siRNA were all synthesized by Qiagen and annealed to form duplexes. Gene silencing efficacy was assessed by RT-qPCR as described above.

Nuclear extraction was performed as described by us previously (27). Briefly, cell monolayers were first lysed with a cytoplasmic buffer and then with a nuclear buffer according to the manufacturer's instructions (Invent Biotechnologies, SC-003). In the case of the cells cultured on Cultrex, they were first detached from the matrix using a non-enzymatic Cultrex organoid harvesting and dissociation solution (3700-100-01) from R&D Systems. Nucleus enrichment was assessed upon Fibrillarin protein expression, whereas protein GAPDH protein expression was used to assess cytosolic purity/contamination of the nuclear fraction.

Electrophoresis reagents origin, total cell lysis procedure, SDS-polyacrylamide gel electrophoresis, electro transfer to low-fluorescence polyvinylidene difluoride membranes, and immunodetection were conducted as described in detail previously (28). Immunoreactive material was visualized by enhanced chemiluminescence.

Cell migration assays were carried out using the Real-Time Cell Analyzer (RTCA) Dual-Plate (DP) Instrument of the xCELLigence system (Roche Diagnostics). Adherent MSC monolayers were trypsinized and seeded (30,000 cells/well) onto CIM-Plates 16 (Roche Diagnostics). These migration plates are similar to conventional transwells (8 µm pore size) but have gold electrode arrays on their bottom side of the membrane to provide real-time data acquisition of cell migration. Prior to cell seeding, the underside of the wells from the upper chamber were coated with 25 µl of 0.15% gelatin in PBS and incubated for 1 h at 37°C. Cell migration was continuously monitored for up to 6 h using serum-free media, in the presence or absence of 30 ng/ml TGFβ. In all cases, the impedance values were measured by the RTCA DP Instrument software and were expressed as Normalized Cell Migration Index. Each experiment was performed two times in triplicates.

All statistical analyses were conducted using the GraphPad Prism 7 software (Dotmatics). Data and error bars are presented as the mean ± standard error of the mean from three or more independent experiments, unless otherwise specified. Hypothesis testing was performed using the Kruskal-Wallis test followed by a Dunn Tukey's post-test (>2 groups). P<0.05 was considered to indicate a statistically significant difference.

The ability of MSCs to form in vitro 3D capillary-like structures mimicking VM was first assessed as described in the Methods section. Mature structures were formed upon 6 h (Fig. 1A, middle panels) when compared to 2D monolayers (Fig. 1A, upper panels). VM parameters analysis, namely mean loop area and perimeter, were performed as described in the Methods section (Fig. 1A, lower panels) and reflected such in vitro maturation in time (Fig. 1B). Cell lysates were isolated from 2D and 3D cultures and immunoblotting performed in an attempt to characterize the acquisition of an epithelial-to-mesenchymal transition (EMT) molecular phenotype as well as TGFβ signaling (Fig. 1C). VM structures were found to significantly trigger the expression of EMT biomarker Snail and FOXC2 as reported elsewhere (24), whereas inductions in TGFβ and of the phosphorylated states of Smad2/3, but not that of total Smad2/3 or β-Actin, were also observed (Fig. 1D, black bars). Harmonized densitometric normalization was performed to β-Actin for the expression of all those tested biomarkers that were found to be changed only. Altogether, these data suggest that a potential TGFβ signaling axis appears to correlate with EMT and be involved during VM in MSCs.

The contribution of the TGFβ signaling axis involving the TGFβR was next assessed using Galunisertib, a pharmacological inhibitor well known to alter the kinase activity of TGFβR (29–31). Moreover, the requirement of TGFβ, as induced upon VM (Fig. 1), was also addressed. Pre-transient silencing of TGFβ was performed for 24 h, then cells seeded on top of Cultex for 6 h. VM structures were reduced when TGFβ was silenced (Fig. 2, middle panels). This suggests a potential requirement for an autocrine regulation process to take place in order to trigger VM. Similarly, when cells were treated with Galunisertib, VM formation was also inhibited (Fig. 2, lower panels). Altogether, these data confirm that a TGFβ signaling axis requiring the kinase activity of the TGFβR is a prerequisite to VM and further support the increase in downstream phosphorylation status of Smad2/3 observed (Fig. 1). Whether TGFβ could further solely and specifically regulate any downstream transcriptional activity was next assessed.

MSCs were treated or not with TGFβ and nuclear fractionation performed as described in the Methods section. Cell lysates were harvested along cytoplasmic and nuclear fractions and immunoblotting performed. The purity control of the cytosolic fraction and of the nuclear material was respectively attested when GAPDH and Fibrillarin were immunoblotted (Fig. 3A). While TGFβ effects in total cell lysates were confirmed (Fig. 3B, Lysate), Snail as well as the phosphorylated Smad2/3 proteins were found to significantly translocate to the Fibrillarin-enriched nucleus fraction (Fig. 3B, Nucleus). Collectively, efficient nuclear translocation in response to TGFβ treatment prompts for the exploration of gene regulation processes. Total RNA was therefore extracted from treated cells and genes of interest assessed by RT-qPCR. TGFβ indeed significantly increased Snail, Fibronectin, and TGFβ gene expression levels confirming increased transcriptional activity in treated MSC (Fig. 3C). Galunisertib pharmacological inhibition of the TGFβR activity prevented those TGFβ-mediated inductions and confirms the necessity of an active signal transducing process (Fig. 3C).

Given the active transcriptional process induced in TGFβ-treated cells, total RNA was extracted from cell monolayers and from cells forming VM on Cultrex. While increased gene expression of Snail, FOXC2, and TGFβ upon VM recapitulated that increase observed in response to TGFβ, gene expression of Fibronectin remained however unchanged (Fig. 4A). As FOXC2 and SNAIL have been previously reported to alter VM formation in MSC (24), transient gene silencing was performed to repress TGFβ, as well as Smad2/3 and cells subsequently seeded on top of Cultrex (Fig. 4B). Accordingly with the downstream effect of Galunisertib on Smad2/3 phosphorylation, silencing of Smad3 reduced all the VM parameters associated with vascular structure formation, including tube length, branching points, total loops, and total tubes (Fig. 4C). This was strongly associated with significant reduction in total loops, and a tendency to reduction of all other parameters in Smad2-silenced cells. Intriguingly, silencing of TGFβ also only altered total loops and tubes formation without affecting other VM parameters (Fig. 4C). Collectively, this evidence suggests that common signaling cues are triggered upon either TGFβ treatment or VM formation. Moreover, given the selective regulation of EMT biomarkers expression, namely that of Fibronectin, complex interplay between these cues will require further investigation although evidence suggests that possible autocrine regulation by TGFβ may regulate VM.

Given some of the common acquisition of an EMT phenotype between TGFβ treatment and VM formation, the involvement of the Smad2/3 signaling required for MSCs to migrate and form 3D capillary-like structures was next assessed. Coupled to the increased TGFβ expression and a possible autocrine regulation, MSC chemotaxis in response to TGFβ was performed in siRNA transiently silenced cells for Smad2 and Smad3 as described in the Methods section and validated (Fig. 5A). Real-time cell migration was monitored for up to 6 h and found to significantly increase in response to TGFβ (Fig. 5B, left panel, closed circles). When gene silencing was performed to suppress either Smad2 or Smad3, TGFβ chemotaxis was significantly reduced in both conditions (Fig. 5B, middle and right panels respectively). This evidence supports the hypothesis that an autocrine TGFβ-mediated process could regulate in vitro VM formation. More importantly, and along their role in VM formation described above, this represents strong evidence for the involvement of Smad2/3 transducing events in response to such autocrine regulation.