Phenotypic, morphological, and metabolic characterization of vascular‐spheres from human vascular mesenchymal stem cells

Abstract The ability to form spheroids under non‐adherent conditions is a well‐known property of human mesenchymal stem cells (hMSCs), in addition to stemness and multilineage differentiation features. In the present study, we tested the ability of hMSCs isolated from the vascular wall (hVW‐MSCs) to grow as spheres, and provide a characterization of this 3D model. hVW‐MSCs were isolated from femoral arteries through enzymatic digestion. Spheres were obtained using ultra‐low attachment and hanging drop methods. Immunophenotype and pluripotent genes (SOX‐2, OCT‐4, NANOG) were analyzed by immunocytochemistry and real‐time PCR, respectively. Spheres histological and ultrastructural architecture were examined. Cell viability and proliferative capacity were measured using LIVE/DEATH assay and ki‐67 proliferation marker. Metabolomic profile was obtained with liquid chromatography–mass spectrometry. In 2D, hVW‐MSCs were spindle‐shaped, expressed mesenchymal antigens, and displayed mesengenic potential. 3D cultures of hVW‐MSCs were CD44+, CD105low, CD90low, exhibited a low propensity to enter the cell cycle as indicated by low percentage of ki‐67 expression and accumulated intermediate metabolites pointing to slowed metabolism. The 3D model of hVW‐MSCs exhibits stemness, dormancy and slow metabolism, typically observed in stem cell niches. This culture strategy can represent an accurate model to investigate hMSCs features for future clinical applications in the vascular field.


| INTRODUCTION
Human mesenchymal stem cells (hMSCs) are multipotent fibroblastlike cells, residing in human tissues including fresh and cadaveric arteries (Pasquinelli et al., 2010;Valente et al., 2014) and endowed of specific skills such as self-renewal and multilineage plasticity. In vivo, they live in a three-dimensional (3D) hypoxic microenvironment identified in multiple adult organs and tissues (Li & Xie, 2005) as a reservoir of stem cells in an undifferentiated and quiescent state. In vitro, hMSCs are cultured in a two-dimensional (2D) plastic adherent monolayer according to International Society for Cellular Therapy (ISCT) minimal criteria (Dominici et al., 2006). MSCs are promising candidates in the regenerative medicine field, and hence they are characterized by multilineage differentiation, migration, immunosuppressive and anti-inflammatory properties (Baer et al., 2010;Baraniak & McDevitt, 2010). The 3D culture strategy resembles the physiological microenvironment of the native niche by keeping the cell-cell interactions (Page, Flood, & Reynaud, 2013) and preserving the stemness features (Cesarz & Tamama, 2016).
Spheroids, so called for their spherical shape, are 3D symmetric cellular aggregates floating in the culture medium. Multiple approaches have been developed to generate spheroids based on cellular adhesiveness promotion and preventing contact with the substrate (Foty, 2011;Liu et al., 2013;Tsai, Liu, Yuan, & Ma, 2015;Wu, Di Carlo, & Lee, 2008;Yuhas, Li, Martinez, & Ladman, 1977;Zimmermann & McDevitt, 2014), including: spinner flasks using constant agitation; liquid overlay technique employing plastic surface precoated with organic matrix; hanging drops exploiting the gravity; nonadherent surfaces such as plastic ultra-low attachment culture substrates; microfluidic system applying a continuous perfusion; more recently, chitosan biomaterials embedding in gel.
Spheroids may be transplanted and enhance bone and cartilage regeneration as well as they may improve the wound healing, angiogenesis and cardiac functions (Sart, Tsai, Li, & Ma, 2014). Recent studies showed an enhanced therapeutic potential, clonogenicity (Guo, Zhou, Wang, & Wu, 2014), multipotent differentiation (Wang et al., 2009), cell survival, angiogenic potential (Bhang, Lee, Shin, Lee, & Kim, 2012) and anti-inflammatory properties (Bartosh et al., 2010) in hMSC spheres. These studies address mainly the beneficial properties of spheres, whereas a broad characterization of hMSC under 3D culture conditions remains largely incomplete. In this study, we characterized hVW-MSCs-derived spheres, including immunophenotypic and molecular profile, morphology, histological and ultrastructural investigation, proliferative capacity and metabolomic profile. We demonstrate that hVW-MSC spheres gain a dormant proliferative and metabolic status that hypothetically approximates the native niche progenitors.

| Immunophenotype characterization
For flow cytometry, hVW-MSCc were detached, counted and incubated with 1 μg antibody/10 6 cells of unconjugated CD44, CD90, CD105, CD166, CD146, Platelet derived growth factor receptor-β (PDGFR-β), Neuron glial antigen 2 (NG2), CD133 and CD34, and conjugate (CD45) antibodies according to datasheet's instructions. Cells were washed with PBS and labeled with appropriate fluorescent probes in the dark, rinsed again before analysis in a flow cytometer (FACSAria, Becton Dickinson); the negative control was labeled with secondary antibody only; about 10,000 events were collected and data were elaborated using the FACSDiva Software (Becton Dickinson). Each antibody was diluted in 1% bovine serum albumin (BSA) in PBS and incubated for 40 min at 4 C. In addition, 6 Â 10 5 hVW-MSCs were seeded on glass, fixed in 2% paraformaldehyde for 4 min at room temperature (rt). For nuclear or cytoplasmic antigens, cells were permeabilized with 1% Tryton X-100 in PBS for 4 min at rt. Cells were blocked with 1% BSA for 30 min at rt and stained with ki-67, Notch-1 and Runx-1 antibodies for 1 hr at 37 C in a wet chamber.
After washing, samples were stained with AlexaFluor-488 or AlexaFluor-546 secondary antibodies in the dark, counterstained with Pro Long anti-fade reagent with DAPI (Molecular Probes, Milano, Italy). Antibodies and respective dilutions are reported in Table 1. Negative controls were performed omitting the primary antibody.
Samples were observed in a Leica DMI6000 B inverted fluorescence microscope (Leica Micro-systems; Wetzlar, Germany).

| hVW-MSCs spheroids: cell number composition, size and growth curve
To generate spheroids, hVW-MSCs were detached, pelleted, counted and filtered through a 40 μm cell strainer before plating in ultra-low attachment 6-well plates at the density of 1 Â 10 5 cells/well or establishing hanging drops of 35 μL with 2 Â 10 4 cells each drop both cultured in DMEM 10% FBS. After 3 days, spheroid formation was observed and pictured with the inverted light microscope (LM) before the collection for morphological, immunophenotypic and metabolomic analysis. To calculate the cell composition of individual spheroid, they were mechanically and enzymatically dissociated with trypsin in single cells and seeded on glass cover slips to allow cell adhesion. After washing with PBS, the samples were methanol fixed for 10 min, rinsed again and stained with crystal violet dye for 30 min at RT. Each sample was manually counted under a microscope equipped with an ocular micrometer and using a hemocytometer. Other spheroids (n = 5) were transferred in single ultralow attachment 24-well plate and let floating in DMEM medium with 10% FBS for up 18 days capturing images every 3 days. For each sample, diameters were measured and expressed as average values.

| hVW-MSCs spheroids: immunophenotype examination
The immunophenotype of 3D cultures hVW-MSCs was investigated through immunocytochemistry using a non-biotin amplified method (NovoLink Polymer Detection Kit; Novocastra, Newcastle upon Tyne, UK) according to manufacturer's instructions. Spheres were formalin fixed and embedded in paraffin; 3 μm thick sections were cut, dewaxed in xylene and rehydrated in graded alcohols. The antigen retrieval was performed using citrate buffer at pH 6, at 120 C, 1 atm for 21 min. After endogenous peroxidase activity neutralization, specimens were labeled with a broad panel of antibodies against CD44, CD105, CD90, ki-67, Runt related transcription factor-1 (Runx-1), Alpha-smooth muscle actin (α-SMA), PDGFR-β, Stromal cell surface marker-1 (Stro-1), CD34,Nestin, overnight at 4 C in a wet chamber; all antibodies were diluted in 1% BSA in PBS as listed in Table 1. Further, the sections were exposed to 3,3 0 -diaminobenzidine (DAB) substrate/chromogen, counterstained with hematoxylin, dehydrated, coverslipped, and viewed in a LM. Digital images were acquired at 10Â magnification using Image-Pro Plus 6 software (Media Cybernetics). To determinate the cycling cells and their position inside the spheroids, the ki-67 intensely stained cells were manually counted and the values were expressed as absolute values.

| Pluripotent gene expression
Total RNA was extracted from hVW-MSCs spheres using TRIreagent, according to the manufacturer's indications (TRIzol reagent; Invitrogen). One μg of RNA was reverse transcribed in a 20 μL volume of reaction using a High Capacity Reverse Transcription Kit (Applied Biosystems, Carlsbad, CA). Real-Time PCR analysis was performed using the SYBR green approach (Power Sybr Green PCR Master Mix) using specific couples of primers (Sigma-Aldrich; listed in Table 2) and carried out in an ABI Prism 7000 Sequence Detection System (Applied Biosystems). Each assay was run in triplicate and target gene expression was normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Relative quantification of mRNA expression was calculated with 2 ÀΔΔCt method. Results were expressed as fold changes relative to hVW-MSC grown in 2D condition.  2.9 | hVW-MSCs spheroids: metabolomic activity 1 Â 10 5 hWV-MSCs were cultured in basal medium with 10% FBS both in adhesion to plastic dish (2D) and 3D condition as vascular-spheres, recovered and processed for metabolomics analysis. For 2D, the cells were cultured until confluence while vascular-spheres were obtained after 3 days of culture in ultralow attachment 6-well plates. To normalize the metabolomics data, the total cellular proteins of monolayer and spheroids were extracted using lysis buffer (KH 2 PO 4 0.

| Statistical analysis
Each experiment was executed in triplicate. Results were analyzed by GraphPad Prism 6 statistical software and expressed as mean ± SD.
Statistical analysis was performed using unpaired Student's t test and results with p value < .05 were considered statistically significant.

| Metabolomic analysis of vascular-sphere and their parental hVW-MSC cells
The metabolomics analysis mirrored a differential metabolism between the 2D and the 3D growth conditions. We analyzed the third passage of hVW-MSC cultured both on 2D monolayer and 3D vascular-spheres. Metabolite concentrations, p-values and foldchange are shown in Table 3. As seen in Figure 6a

| DISCUSSION
The present study was aimed at characterizing vascular-spheres obtained from 3D culture of hVW-MSCs, combining the reliability of primary cultures with the accuracy of the 3D environment, which mimics the stemness niche.
Femoral artery hVW-MSCs were able to rapidly and spontaneously aggregate into vascular-spheres, composed of CD44 + , CD105 low , CD90 low mesenchymal cells. Accordingly, a CD44 + multipotent stem cell population with multilineage mesodermal plasticity, including smooth muscle cells and pericytes, was identified in the adventitia of adult fresh human internal thoracic artery (Klein et al., 2011). Further, we found a significant up-regulation of NANOG in vascular-spheres, hypothesizing that 3D culture enhances stemness and pluripotency features of hVW-MSCs. This condition is also supported by the ability of hVW-MSCs to survive anoikis, a form of programmed cell death that occurs when anchorage dependent cells, like mesenchymal cells, grow under non-adherent conditions without extracellular matrix support. Indeed, the resistance to lethal or sublethal stresses is suggestive of an elevated hierarchical status (Ciavarella et al., 2015), therefore it could be speculated that hVW-MSCs within spheres may be ancestor MSCs.
An interesting question concerns the methods of spheroid formation, considered to derive from tight cell aggregates (Rajcevic et al., 2014); theoretically, vascular-spheres should originate from cell division or cell adhesion. We observed only few and peripheral cells positive to ki-67, a well-known proliferation marker, in vascular spheres; therefore, we hypothesized that more than 80% of hVW-MSC  (Haust, 1987). A recent study demonstrated increased proliferation in MSCs following siRNA knockdown of the ciliary proteins IFT 172 and KIF 3 A, and the down-regulation of OCT-4, SOX-2 and NANOG, confirming that primary cilium is crucial to MSC proliferation and stemness (Ma et al., 2020). Further, primary cilia regulate many signaling pathways associated with cell differentiation and lineage specification, suggesting their potential application as novel therapeutic target (Bodle & Loboa, 2016).
In addition, we observed a time-dependent reduction of vascularspheres, thus indicating that cell loss prevailed over cell proliferation.
Possible explanations for the high rate of quiescent cells seen in mesenchymal spheroids could be cell selection through anchorage impediment, aggregation modality, remodeling of cell shape, gravity or other physical force influence, floating, low nutrient support or oxygen availability.
Another interesting finding of this study regards how anchorageindependent mesenchymal cells aggregate to form vascular-spheres.
Since these cells are predominantly quiescent, it could be that physical forces trigger their aggregation; this may also explain why variously sized spheres were generated in relation to the culture method Histology and ultrastructural analysis revealed that hVW-MSC cells had a loose arrangement coherent with their mesenchymal nature; they were joined by subplasmalemmal linear densities (mesenchymal-type junctions) and delivered extracellular matrix focally. As previously reported (Bartosh et al., 2010), also vascularspheres are histologically composed by two distinct areas with a peripheral zone of multilayered spindle hVW-MSC cells and a core zone of rounded cells. Previous studies reported that the central core of the vascular-spheres contained mostly dead cells as a consequence of low oxygen availability, the entity of dead cells correlating with spheroids size (more than 200-250 μm in diameter) (Curcio et al., 2007). Also our study showed a mixture of live and dead cells; confocal fluorescent microscopy of 300-350 μm vascular-spheres localized dead cells in the inner core. TEM provided evidence that a significant proportion of dead cells were in apoptosis, suggesting that poor access to nutrients or oxygen are not exclusive determinants of core necrosis; in fact, SEM and TEM showed multiple 100 nm sized nanotubular projections as those described in a previous study by our group (Valente, Rossi, Resta, & Pasquinelli, 2015). These thin membranous channels represent a novel communication system of exchanging molecules and cytoplasmic organelles among cells also placed at a significant distance from the point of origin of the projection itself; in addition to being a structural requirement for the exchange of biological information (Simons & Raposo, 2009), nanotubular projections can constitute a system for feeding cells located in an unfavorable positions or eliminating waste products. In our view, the presence of core apoptosis may reflect the effect due to the absence of nutrient and the alteration of the catabolic process. Further, metabolic analysis supports the quiescent status of hVW-MSCs composing the vascular-spheres. The regulation of central metabolic pathways is considered an important modulator of stem cell quiescence and pluripotency. By the inhibition or enhancement of key processes, stem cells regulate energy production using nutrientsensing pathways (Ochocki & Simon, 2013). The accumulation of metabolites from glycolysis, tricarboxylic acid cycle and some related to amino acid catabolism may indicate a slower metabolism within spheroids. To sustain cell functions, higher amounts of energy are necessary, while there is a lower need for anabolic precursors. In this case, all metabolic pathways are redirected to complement tricarboxylic acid cycle, to obtain energy for cell survival. This characteristic is supported by other metabolomic experiments, in which citrate, glutamate and malate were decreased in embryonic stem cells and induced pluripotent stem cells when compared to somatic cells (Vacanti & Metallo, 2013). Results in embryonic stem cells showed glycolysis is one of the most active components of their metabolism, as they produce more lactate and consume more glucose than differentiated cells (Ochocki & Simon, 2013;Zhang et al., 2011).
Spheroids showed high values of glutamine, although glutamate values were lower. This can be explained because, apart from glucose, glutamine can be a substrate for energy production. Glutamine provides a source of nitrogen to proliferating cells for amino acid synthesis. Though, if cells are not proliferating, glutamine is not being consumed, and then there is no production of glutamate (Li, Zhang, Zhao, Ma, & Chen, 2016;Vacanti & Metallo, 2013). To maintain the metabolite flux in the tricarboxylic acid cycle, anaplerotic pathways such as glutaminolysis are often necessary. In this case, glutamine is an important substrate (Vacanti & Metallo, 2013).

| CONCLUSION
This study reports the characteristics of hVW-MSCs derived from human femoral arteries when cultured as vascular-spheres under nonadherent conditions. This 3D model is a surrogate of the physiological vascular niche microenvironment in which the hVW-MSCs in vivo reside. The culture methods employed in the generation of vascularsphere reversibly select a CD44 + , CD105 low , CD90 low vascular mesenchymal cell population that is highly resistant to anchorageimpairment stress and, when driven by physical forces, acquires the ability to aggregate rearranging the contractile cytoskeleton without losing the mesenchymal immunophenotypic and stemness markers.