Human multipotent mesenchymal stromal cell transcriptome changes in response to PDGF-BB

Activation of multipotent mesenchymal stromal cells (MSCs) is a central part of tissue response to the damage. Platelet-derived growth factor (PDGF-BB), which is abundantly released in the damaged area, potently stimulates the proliferation and migration of MSCs. Recent evidences indicate that tissue injury associates with the accumulation of senescent cells, including ones of MSC origin. Therefore, we hypothesized that PDGF-BB induces MSC senescence. To evaluate mechanisms of early activation of MSCs by PDGF-BB, we performed transcriptome proling of human MSCs isolated from adipose tissue after exposure to PDGF-BB for 12 and 24 hours. We demonstrated that PDGF-BB induced the expression of senescence markers in MSCs. However, further experimental validation of transcriptomic data clearly indicated that PDGF-BB exerts mitogenic and pro-migratory effects on MSCs, but did not signicantly stimulate their senescence.


Introduction
Recently, tissue injury was associated with the accumulation of senescent cells, including those originating from multipotent mesenchymal stem/stromal cells (MSCs) (Neri and Borzì, 2020). Numerous studies have implicated MSCs to the tissue regeneration upon damage, with paracrine activity being a central mechanism of their action (Pittenger, 2019;Sagaradze, 2020;Basalova, 2020). Signals from damaged tissues could both activate MSCs and cause the acquisition of a senescent phenotype characterized by cell cycle arrest with p16 INK4a , p21 Waf1/Cip1 and p53 expression activation, resistance to apoptosis, and speci c senescence-associated secretory phenotype (SASP) activated by transcription factors NF-κB and C/EBPβ (Coppe, 2010; Campisi, 2019; Fa án-Labora, 2020). SASP is characterized by a shift towards a proin ammatory, mitogenic cytokines as well as changes in the extracellular matrix composition. MSC senescence caused the crucial impairment of their regenerative capacity and regulatory functions.
Upon a damage MSCs respond to a multitude of factors, including abundantly presented platelet-derived growth factor (PDGF). These cells express PDGFRβ receptor, which has the highest a nity for PDGF-BB dimer (Heidaran, 1991). PDGF can activate both migration and proliferation of stromal cells (De Donatis, 2008), therefore stimulating tissue regeneration. However, PDGF-mediated activation of stromal cells could also contribute to brogenesis (Klinkhammer, 2018).
Since proliferation and senescence are closely related, we hypothesized that PDGF-BB could simultaneously cause MSCs senescence. To test this hypothesis, we performed transcriptomic pro ling of human adiposederived MSCs after treatment with PDGF-BB by RNA sequencing. Acquisition of senescent phenotype of MSCs treated by PDGF-BB was evaluated by beta-galactosidase and p21 expression as well as secretion of SASP components (IL-6, MCP-1, PAI-1). We con rmed mitogenic and pro-migratory effects of PDGF-BB on MSCs but did not observe the signi cant stimulation of MSC senescence in these conditions.

Materials And Methods
Human MSC isolation and culture Human adipose-derived MSCs were isolated from subcutaneous adipose tissue obtained from 4 healthy donors during abdominal surgery. All donors or, if donors were under 18, a parent and/or legal guardian gave their informed consent and the local ethics committee of the Medical Research and Education Center of Lomonosov Moscow State University (Moscow, Russia) approved the study protocol (#4, 04.06.2018). All methods were performed in accordance with the relevant guidelines and regulations.
Subcutaneous adipose tissue samples (0,5-5 ml) harvested during surgery were homogenized and digested in collagenase I (200 U/ml, Worthington Biochemical) and dispase (40 U/ml, Sigma) solution under agitation for 30-40 min at 37°C. Then tissue was centrifuged at 200 g for 10 min and the supernatant discarded. The pellet containing ADSC was lysed to destroy erythrocytes, ltered through a sieve (BD Falcon Cell Strainer, 100 mkm) and centrifuged at 200 g for 10 min. The nal pellet was resuspended in culture medium. Cells were cultured in AdvanceSTEM™ Mesenchymal Stem Cell Media containing 10 % AdvanceSTEM™ Supplement (HyClone, USA), 1 % antibiotic-antimycotic solution (HyClone) at 37 °C in 5 % CO2 incubator. Cells were passaged at 70 % con uency using HyQTase solution (HyClone). For the experiments, MSCs of the third or fourth passages were seeded at a density of 7x10^3 cells/cm2 on uncoated culture plates (Corning, USA) a day before the experiment. At these passages phenotypic characterization and functional evaluation of mesenchymal properties were performed as described below.
MSCs were washed thoroughly with HBSS (HyClone, USA) to remove any residues of AdvanceSTEM™ Supplement. Then HBSS was replaced by serum-free AdvanceSTEM™ Mesenchymal Stem Cell Media and MSCs were cultivated overnight. In selected experiments 10 ng/ml PDGF-BB were added 18 hours after deprivation. For assessment of transcriptome content MSCs were cultured under standard conditions (5 % CO2 and 21 % O2) at 37 °C for 12 or 24 h with or without PDGF-BB. At the end of the incubation cells were washed with HBSS and were lysed in the RLT buffer for further isolation of the total RNA. MSCs quantity and viability were assessed at the end of experiment using Countess cell counter (Invitrogen).

MSC differentiation assays
MSC abilities to differentiate into osteogenic, adipogenic and chondrogenic directions were tested in vitro using standard differentiation and analysis protocols described elsewhere. Brie y, osteogenic differentiation was induced by plating 6х10^4 of MSCs into 24-well plate and incubating in growth medium containing 10^-8 М dexametasone, 10 mМ β-glycerol-2-phosphate, 0.2 mМ 2-phospho-L-ascorbic acid (Invitrogen, USA) during 3 weeks. Differentiation e ciency was analyzed using Alizarin Red S staining. Adipogenic differentiation was induced by 3 cycles of consecutive incubation for 6 days in the growth medium containing 10^-6 М dexametasone, 10µM insulin, 200µM indometacin and 0.5mM 3-isobutyl-1-methylxantine (Invitrogen, USA) followed by 3 days incubation in growth medium containing 10µM insulin. Accumulation of intracellular lipids was assessed by Oil-Red-O staining. Chondrogenic differentiation was induced by incubation of pelleted MSCs in AdvanceSTEM Chondrogenic Differentiation Medium (HyClone) for 28 days and stained with toluidine blue.

HUVEC isolation and culture
Human umbilical cord endothelial cells (HUVEC) were obtained from cell culture Depository of Live Systems (Lomonosov Moscow State University), ID of the collection MSU_HUVEC. HUVEC were cultured in EGM™ Media containing 2 % fetal bovine serum (FBS) and recombinant growth factors (CC-4176, Lonza, Switzerland), at 37 °C in 5 % CO2 incubator. Cells were passaged at 70 % con uency using HyQTase solution (HyClone). For migration assay cells were used between passage 3-6.

RNA isolation and transcriptome analysis
Total cellular RNA was isolated from cultured MSCs using RNeasy Kit (Qiagen, cat # 74104) according to manufacturer instructions. Five hundred nanograms of total RNA was used for transcriptome analysis.

Whole-Genome Expression Analysis Using the Sentrix® Human-6 Expression BeadChips
Whole-genome gene expression was assessed using a direct hybridization approach. To amplify starting material, a single-round IVT ampli cation was carried out using the Illumina RNA Ampli cation Kit, manufactured by Ambion®, Inc.

Data preprocessing
Illumina BeadStudio output le was processed using lumi package available in Bioconductor software collection. The processing included following steps: Quality control of raw data; Background subtraction; Variance stabilization using algorithm VST, developed for Illumina data. Values after this processing step should be treated as log-transformed. Quantile normalization.
First, pairwise correlations of intensities were consistently high (>0.95) for the samples.
Additionally, we excluded all probes that had detection p-value > 0.05 in all 3 samples. 4 columns (samples) and 11073 rows (probes) corresponding to 9659 non-redundant gene.

Fold change calculation
Calculated fold changes of gene expression relative to the expression in control sample.

Enrichment analysis
Enrichment analysis was conducted with the GOstats package (4) of Bioconductor. This analysis aims to identify biological processes (in our case, "BP" branch of Gene Ontology (5)

Immunohistochemistry
Human adipose-derived MSCs from 3 healthy donors were seeded at density 10^4 per ml. Cell were used between passage 3 and 4. PDGF BB were added at concentration 10 ng/ml for 24 hours after 18 hours of serum deprivation. MSCs were xed with 4% paraformaldehyde solution (Panreac, Spain) at room temperature for 10 min and incubated with 0,2% triton X100 (Sigma, USA) solution at RT for 10 min. Further, MSCs were incubated for 1 hour in 1% bovine serum albumin (BSA, Sigma, USA), 10% normal goat serum (Abcam, UK) solution at room temperature to block the non-speci c interaction of antibodies. Subsequently, the samples were incubated with primary polyclonal rabbit antibody for p21 Waf1/Cip1 (Cell Signalling, 2947S) or rabbit polyclonal IgG (Biolegend, 910801) in 1% BSA solution at +4° overnight. Then samples were incubated with uorescence labeled goat-anti-rabbit secondary antibodies (A11034, Invitrogen) at room temperature for 1 h. Cell nuclei were labeled with DAPI (DAKO). Samples were analyzed with a Leica DM6000B uorescent microscope equipped with a Leica DFC 360FX camera (Leica Microsystems GmbH, Germany). The percent of p21-positive MSCs was evaluated.
β-galactosidase staining Evaluation of β-galactosidase activity was proceed with the Senescence β-Galactosidase Staining Kit (Cell Signalling, 9860) according to manufacturer's recommendations.

ELISA assay
MSCs from 3 healthy donors were used between passages 3-4. Cells were seeded at density 10^4 per ml and allowed to attach to plastic. After 18 hours of serum deprivation PDGF BB were added at concentration 10 ng/ml for 24 hours. MSCs were washed thoroughly with HBSS (HyClone, USA), then HBSS was replaced with DMEM low glucose (Gibco, USA) with 1% BSA and MSCs were cultivated for 72 hours. Condition medium was collected and centrifuged at 300g for 10 min to remove cellular debris. Supernatant was transferred to clear tube with with 1% protease inhibitor cocktail x100 (Sigma Aldrich). The concentration of IL-6, MCP-1 (CCL-2), PAI-1, HGF, VEGF in the samples of conditioned medium was evaluated with ELISA kit (R&D Systems) according to manufacturer's recommendations and normalized to cell number.
Migration assay HUVEC migration assay was performed in automated RTCA xCELLigence™ system (ACEA Bioscienses, USA) according to manufacturer's recommendations. Brie y, HUVEC media was changed to basal media EBM (Lonza, Switzerland) with 0,5% bovine serum albumin (BSA, Imtek, Russia) to harvest serum growth factors 4-6 hours before experiment. The lower chamber of CIM-plate was lled with EBM with 0,5% BSA as a negative control, EGM2, containing 2% of fetal bovine serum (FBS) (Lonza, Switzerland), conditioned media from MSC (CM MSC) or conditioned media from MSC pretreated with 10 ng/ml PDGF BB for 24 hours. As a control of possible chemotactic activity of PDGF itself it was added to one of the chambers in concentration 10 ng/ml in EBM 0,5% BSA. 20 000 of HUVEC were placed in the upper chamber. After placing the CIM plate into RTCA xCELLigence system bioimpedance on each well was xed every 15 minutes. Data were processed with RTCA Software Pro (ACEA Biosciences, USA).

Statistical analysis
We have performed at least three independent experiments for each measurement. Statistical data processing was performed using the program GraphPad Prizm 8.0. Means ± standard deviation (SD) were used to present quantitative data. The comparison of independent groups was performed by Mann-Whitney test. Differences were considered statistically signi cant at the signi cance level of p < 0.05.

Results
PDGF-BB up-regulates genes associated with proliferation and cell cycle and down-regulates immune response-related pathways in MSCs MSCs isolated from the adipose tissue of healthy donors comply with recommendations of the International Society for Cellular Therapy (Dominici et al., 2006). These cells adhere to the culture plastic, possessed CD73 + / CD90 + / CD105 + / CD45 − / CD34 − / CD14 − / CD19 − immunophenotype and differentiated into osteogenic, chondrogenic and adipogenic lineages upon culturing in the induction medium (Fig. 1). Most cells were also positive for PDGFRβ (87-95%).
1. Evaluation of MSC markers by ow cytometry. More than 99% of MSCs express CD73, CD90, CD105 on cell surface.
To evaluate the effect of PDGF-BB on global changes of gene expression in MSCs, we performed expression pro ling of cells treated by PDGF for 12 of 24 hours using RNA sequencing. Differential expression analysis revealed numerous expression changes induced by PDGF treatment (152 and 127 genes were upregulated at least 2-fold after 12-and 24-hour exposure to PDGF-BB, correspondingly; 173 and 163 genes were downregulated at least 2-fold after 12-and 24-hour exposure to PDGF-BB).
Functional enrichment analysis was performed using clusterPro ler Bioconductor package. As a result, each differentially expressed gene list was associated with quantitatively ranked lists of Gene Ontology biological processes and KEGG pathways summarizing its effects at a systems level.
Downregulated genes were associated with immune response-related pathways and processes, most prominently gamma-interferon response. Signi cant decrease in expression was observed for chemokine genes, including CCL8 (C-C motif chemokine ligand 8, IL-8), CXCL10 (C-X-C motif chemokine ligand 10), CCL2 (C-C motif chemokine ligand, MCP-1), CXCL12 (C-X-C motif chemokine ligand). In addition, there was a strong decrease in IL6 gene expression, a high level of secretion of which is a characteristic feature of SASP. Among genes downregulated in MSCs after PDGF-BB treatment we also found THBS2 (thrombospondine-2). Figure 3 illustrates similarities in enriched biological processes between 12-and 24hour treatments.
PDGF-BB-treated MSCs did not acquire senescence biomarkers Senescent cells are characterized by accumulation of cell cycle inhibitors, such as p16 INK4a , p21 Waf1/Cip1 and p53. We evaluated the number of MSCs with p21-positive nuclei with and without PDGF-BB treatment for 24 hours. We did not nd signi cant differences between treated vs non-treated MSCs (Fig. 4, A). Furthermore, we evaluated the accumulation of senescent cells by activity of β-galactosidase. There was no signi cant elevation of β-gal activity in PDGF-BB treated MSCs (Fig. 4, B), which indicated that PDFG-BB did not induce cellular senescence.
Senescent associated secretory phenotype (SASP) is characterized by the elevated production of cytokines and interleukins, which mediates proin ammatory microenvironment of senescent cells. Among possible SASP factors in conditioned medium of MSCs we measured such crucial factors as IL-6, MCP-1 (CCL2) and PAI-1 by ELISA. Obtained protein concentrations were normalized by cell number. In case of IL-6 we found signi cant decrease in conditioned medium of PDGF-BB treated MSCs comparing to non-treated control.
Concerning MCP-1 and PAI-1 there was no signi cant changes in levels of secreted proteins between treated vs non-treated cells (Fig. 4, C).

Immunocytochemical evaluation of p21 Waf1/Cip1 in control (CTRL) MSCs and after incubation with
PDGF-BB for 24 hours. Fluorescent microscopy, green staining -p21 Waf1/Cip1 , blue staining -nuclei stained with DAPI. The percentage of p21 + cells was not changed signi cantly after PDGF-BB treatment.

Evaluation of β-galactosidase activity (blue staining) in control (CTRL) MSCs and after incubation with
PDGF-BB. Phase contrast microscopy. The percentage of β-gal + cells was not changed signi cantly after PDGF-BB treatment.
3. MSC secretion of IL-6 decreased, while secretion of MCP-1 and PAI-1 was not changed signi cantly between control (CTRL) and after PDGF-BB treatment.
Data are presented as mean value of three independent experiments, error bars -standard deviation SD (B) or standart error SE (A, C), n.s. -no signi cant difference. Scale bar 100 µm.

PDGF-BB stimulated MSC pro-angiogenic activity
Angiogenic properties of MSCs mediated by cell paracrine factors are well described and have been shown to decrease with aging. We evaluated whether pro-angiogenic activity of MSC conditioned media is changed after PDGF-BB treatment using migration endothelial cells (HUVEC) test. HUVEC migrated in xCELLigence system towards basal media without additives as a negative control, EGM2 containing 2% of FBS as a positive control, conditioned media (CM) from MSCs or MSCs pretreated with PDGF-BB for 24 hours. We observed an increase in HUVEC migration towards CM from MSCs comparable with positive control (Fig. 5,  A). Moreover, CM from MSCs pretreated with PDGF-BB stimulated the velocity of endothelial cells even stronger, while PDGF-BB itself decreased it (data not shown). The evaluation of the key pro-angiogenic growth factors, hepatocyte growth factor (HGF) and vascular endothelial growth factor (VEGF), demonstrated a signi cant increase in concentration of these factors in CM from PDGF-BB-treated MSCs comparing to CM from control MSCs (Fig. 5, B).

Discussion
PDGF-BB is a well-known mitogen for stromal cells, such as broblasts and MSCs. Produced in high concentrations during the acute phase of tissue damage, it activates stromal cell proliferation and migration (De Donatis, 2008). PDGFRβ receptor, which has the highest a nity for PDGF-BB dimer (Heidaran, 1991), is detected on many stromal cells, including MSCs. PDGF is an essential mitogen for many types of stromal cells involved in wound healing as it allows the cell to skip the G1 check point in the cell cycle (Song, 2007). During regeneration, PDGF attracts broblasts, monocytes, and macrophages into the damage zone, stimulates their proliferation and deposition of the temporary extracellular matrix, and then collagen I type, which forms the basis of the scar . This process is essential for healing. However, apparently PDGF can trigger the secretion of endogenous factor by macrophages (and possibly other cells), which leads to the formation of an autocrine stimulation loop. Thus, both PDGF-mediated activation of the migration and proliferation of stromal cells into the lesion zone and the switching of the secretory cell phenotype can contribute to brogenesis and associated senescence (Klinkhammer, 2018).
Since aggressive surrounding in damaged zone may lead to induction of cellular senescence, we evaluated whether PDGF-BB could up-regulate senescence-related genes within human MSC transcriptional pro le.
Groups of up-regulated genes were mainly associated with activation of the cell cycle, stimulation of DNA replication, and cell division indicating that MSCs were actively proliferating after PDGF-BB treatment. It correlated with decreased accumulation of CDK inhibitors, such as p16INK4a, p21Waf1/Cip1 and p53, the markers of senescent cells. Indeed, in validation tests we have not found any signi cant elevation in the amount of p21-positive MSCs after PDGF-BB treatment. Many of the proteins coding by the up-regulated genes also modulate the reparative response of cells to DNA damage. Such results correlate with the general position on the role of PDGF-BB in the initial stages of response to damage as a crucial mitogen for MSCs.
Importantly, we observed signi cant increase in the expression of genes regulating MSC migration such as PLAU and its receptor PLAUR, MMP-1 and SERPINE1 (PAI-1). These genes are also found to be associated with increased cellular senescence. As it was described previously, accelerated aging is associated with Nevertheless, the increased secretion of PAI-1 by PDGF-BB-treated MSCs was not con rmed by ELISA in our validation tests.
Additionally, an increase in the expression of tubulin alpha and beta chain isoform genes stimulated by PDGF-BB treatment of MSCs indicates both the activation of proliferation and the possible rearrangement of the cytoskeleton during migration. Also, an increase in the expression of the HS3ST3A1 gene (heparin sulfate-glucosamine 3-sulfotransferase), an enzyme involved in the synthesis of heparan sulfate, which plays the role of an adapter in the stimulation of haptotaxis of MSCs through the activation of tyrosine kinase receptors, was detected (Veevers-Lowe et al., 2011). These data indicate the migration response of MSCs on the rst day of PDGF-BB stimulation and the formation of the supporting mechanisms for homing stimulation.
Among genes down-regulated in human MSCs after PDGF-BB treatment we found THBS2 (thrombospondine-2), which was de ned as a protein secreted by senescent cells. It inhibits endothelial migration and prevents cells from senescence escape (Guillon et al., 2019). However, we received data indicating that conditioned medium from MSC after PDGF-BB treatment had stimulatory effect on HUVEC directional migration comparing to initial MSC secretome. This may be due to the up-regulation of proangiogenic growth factor secretion by MSCs as well as down-regulation of THBS2 expression.
Analysis of clusters of down-regulated genes with 2-fold changes in expression in PDGF-BB-treated human MSCs showed a decrease in the activity of transcription of pro-in ammatory factors involved in the response to interferons, the innate immune response, and the response to viruses. In addition, there was a strong decrease in IL6 gene expression, a high level of secretion of which is a characteristic feature of SASP. These effects for IL6 and MCP-1 were con rmed by ELISA in the experimental validation tests. Obtained data support the acquisition of anti-in ammatory phenotype for MSCs under PDGF-BB, which is consistent with the idea of the formation of the immunosuppressive MSC phenotype under acute damage. According to the evaluation of widely recognized senescence biomarkers (p21 + and β-gal+) in MSCs we did not observe any signi cant difference between control cells and cells after PDGF-BB treatment. ELISA assay showed no signi cant increase in the level of SASP components, including IL-6, MCP-1 or PAI-1 after MSCs incubation with PDGF-BB.
It is important to note that cellular aging is a controversial process, and the development of senescent phenotype could have some bene ts. In a young organism, after damage senescence of myo broblasts promote the inhibition of their proliferation after healing, as well as the involvement of immune cells that eliminate myo broblast. The critical contribution of senescent cells in wound healing, injury repair and tissue regeneration was demonstrated in several studies and related to so-called early senescence (Ritschka 2017; Campisi 2019). However, if the immune system does not cope with the elimination of senescent cells, they can accumulate in tissues promoting chronic pro-in ammatory microenvironment, brosis, or malignancy of adjacent healthy tissues (late senescence). Switching the normal MSC phenotype to senescent one can be determined by the entry of cells into the hostile environment after tissue damage, and PDGF-BB could play a critical role in this process. Summarizing our data, at short periods of exposure to PDGF-BB (up to 24 hours), the proliferation of MSCs is activated, the expression of genes associated with migration increases, the level of secretion of growth factors increases, due to which MSCs paracrine stimulate endothelial migration and angiogenesis. However, analysis of senescent biomarkers in MSCs in such conditions indicate in favour the acquisition of early senescent phenotype, with bene ts for tissue regeneration, but not the signs of late senescence.