Mesenchymal stem cells of Systemic Sclerosis patients, derived from different sources, show a profibrotic microRNA profiling

Systemic Sclerosis (SSc) is a disease with limited therapeutic possibilities. Mesenchymal stem cells (MSCs)-therapy could be a promising therapeutic option, however the ideal MSCs source has not yet been found. To address this problem, we perform comparison between bone marrow (BM)-MSCs and adipose (A)-MSCs, by the miRs expression profile, to identify the gene modulation in these two MSCs source. MicroRNAs (miRs) are RNAs sequences, regulating gene expression and MSCs, derived from different tissues, may differently respond to the SSc microenvironment. The miRs array was used for the miRs profiling and by DIANA-mirPath tool we identified the biological functions of the dysregulated miRs. In SSc-BM-MSCs, 6 miRs were significantly down-regulated and 4 miRs up-regulated. In SSc-A-MSCs, 11 miRs were significantly down-regulated and 3 miRs up-regulated. Interestingly, in both the sources, the involved pathways included the senescence mechanisms and the pro-fibrotic behaviour. Furthermore, both the MSCs sources showed potential compensatory ability. A deeper knowledge of this miRs signature might give more information about some pathogenic steps of the disease and in the same time clarify the possible therapeutic role of autologous MSCs in the regenerative therapy in SSc.


Identification of miR expression profiling discriminating SSc-BM-MSCs and SSc-A-MSCs. The
SSc-MSCs were isolated from SSc patients and the demographic and clinical characteristics of the patients are showed in Table 1.
MiRs target prediction analysis. In silico analysis, predicting the putative biological functions of the miRs, suggested their involvement in specific Kyoto Encyclopedia of Genes and Genomes (KEGG) signalling pathways. www.nature.com/scientificreports www.nature.com/scientificreports/ The KEGG database identifies significantly enriched metabolic pathways or signal transduction pathways from lists of candidate target genes and compares this enrichment with a reference background.
The results obtained were included in a Neo4J graph database 28 to provide a graphical representation of miRNAs-targets interactions (Fig. 3). This graphical representation highlighted that both the Profiles shared, 2 important pathways in this clinical setting: the TGF-beta Signalling Pathway and Signalling Pathways Regulating Pluripotency Of Stem Cells.

Discussion
Considering the regenerative potential of MSC, displaying immunomodulatory, angiogenic and antifibrotic capabilities, MSC-based therapy could counteract the three main pathogenic axes of SSc 17,22,29,30 , a clinical setting lacking effective therapy. This is the first paper reporting in silico comparative analysis of miRs profile of MSCs isolated from different sources (BM and A) of SSc patients. We showed that, independent from the source, www.nature.com/scientificreports www.nature.com/scientificreports/ an upregulation of the pathways regulating the senescence and the profibrotic phenotype, may be observed in the MSCs of patients affected by SSc, a disease characterized by diffuse fibrosis of skin and internal organs. Furthermore, both BM-and A-SSc-MSCs display a down regulation of the miRs controlling the genes related to cells survival, thus suggesting the ability of these cells to protect themselves, activating specific pathways in response to the critical conditions, found in the scleroderma microenvironment 31,32 , such as hypoxia and inflammation 14,33 . At present, one third of the clinical trials using engrafting MSCs are designed to evaluate their therapeutic role in autoimmune diseases (for the latest update, see ttp://www.clinicaltrials.gov). Considering the immunomodulatory, pro-angiogenic and antifibrotic capabilities of MSCs, their transplant may have a potential application in cell-based therapies for SSc and results obtained in preclinical models 3,6,14,19,20,23 , as well as the few cases reported in the SSc 34,35 seem very promising. However, many in vitro experiments showed some differences in the biologic functions of SSc-MSCs, deriving from different tissues 3,6,22,23 . These data raised some concern about their efficacy in transplant strategy, since the profibrotic signature of SSc-MSCs may thwart their potential beneficial effects and suggested further refinements in the molecular and genetic knowledge of MSCs.  www.nature.com/scientificreports www.nature.com/scientificreports/ It has been reported that long-term culture evokes continuous changes in MSCs: proliferation rate decays, the cell size increases, differentiation potential is affected, chromosomal instabilities may arise, and molecular changes are acquired. In fact, early passage MSCs were preferred for therapeutic efficacy in many clinical trials 36 and, on this basis, we choose to characterize the miR profiling of P3 MSCs, providing a profiling of in vitro expanded MSCs at the time in which they may be used for transplantation. We observed that P3 SSc-BM-MSCs expressed 6 miRs significantly down-regulated and 4 miRs significantly up-regulated when compared with HC-BM-MSCs. As far as, A-tissue is concerned, P3 SSc-MSCs expressed 11 down-regulated miRs and 3 up-regulated, when compared with HC. These data identified 2 different profiles and of note, the miRs listed in one Profile were not present in the other Profile. This datum confirms that miRs constitute a family of regulators for gene expression, which may be largely tissue-dependent 37 .
To assess the biological significance of miRs listed in both Profiles, we relied on in silico analysis for predicting potential miRs targets. In both Profiles, we observed two specific patterns and by the "Diana miRPath" software, we try to identify the signalling pathways, significantly associated with each pattern in both Profiles.
In the Profile BM, 6 miRs were down-regulated, targeting 32 KEGG pathways, considering the high number of these pathways, we focused our discussion on the KEGG pathways that may play a specific relevance in SSc, such as "Signalling Pathways Regulating Pluripotency Of Stem Cells", "MAPK Signalling Pathway", "HIF-1 Signalling Pathway", "TGF-Beta Signalling Pathway", "P53 Signalling Pathway" And "PI3K-Akt Signalling Pathway". The "Signalling Pathways Regulating Pluripotency Of Stem Cells" may involve transcription factors and their downstream target genes promoting self-renewal and pluripotency of stem cells. Using microT-CDS, among the validated target genes of this pathway, we observed the genes regulating for: SMAD-2 and -5, thus suggesting a profibrotic signature of MSCs 5 and KLF4 38,39 , which is involved in the cellular senescence. The "MAPK Signalling Pathway" is involved in various cellular functions, including cell response to different stimulation. This pathway includes several validated target genes: HSPA8, TAOK1, RPS6KA3 and PAK2, regulating the adaptation to different stress stimuli and cells survival 29,[40][41][42] , and CRKL, promoting activities including cytoskeletal remodelling, cell motility, cell proliferation and mitosis 38 . Furthermore, several genes involved in this pathway display pro-apoptotic activity, such as ELK4, JUN, RASA2, DUSP1 and ZAK, and proliferation, such as RAP1B 38 .
The "HIF-1 Signalling Pathway" includes gene encoding proteins, mediating adaptive responses to nitric oxide and reduced oxygen availability, a common condition during SSc, in which a desertification of vascular tree is present, and it is not surprising to observe that the validated target genes for this pathway are: PLCG1, FLT1 and EIF4E2, which may mediate MSCs proliferation, angiogenesis and survival in hypoxia condition 38,43,44 , suggesting a compensatory mechanism in these patients. The "TGF-beta Signalling Pathway" includes genes with a specific relevance during fibrotic process. The experimentally validated target genes are: SMAD2 and 5 and PPP2R1B, promoting the TGF-beta signalling 5,45 . The "p53 Signalling Pathway" may regulate genes involved in DNA damage and oxidative stress response. The experimentally validated target genes included in this pathway were: CDK2, CCNE2 46 and SESN1, stimulating proliferation, and ZMAT3, PMAIP1, SIAH1, and PTEN 38 , contributing to apoptosis or cell-cycle inhibitor. The "PI3K-Akt Signalling Pathway" is activated by many cellular stimuli or toxic insults and regulates fundamental cellular functions such as transcription, translation, proliferation, growth, and survival. The validated target genes are: BRCA1 and EIF4E2, promoting cells response to damage and hypoxia 32,44 ; YWHAG, promoting a more mature phenotype in MSCs 47 , CDK2 and CCNE2, playing a role in cell cycle G1/S transition 46 ; PTK2, promoting proliferation; ITGAV and PPP2R1B, promoting the TGF-beta signalling in both normoxia and hypoxia; PTEN, PPP2R5E, regulating the balance between survival and apoptosis; EFNA1, KDR and FLT1, encoding member of the vascular endothelial growth factor family 38 .
In the Profile A, we listed 11 down-regulated miRs, targeting 18 KEGG pathways. Interestingly, 2 relevant target pathways, the "TGF-beta Signalling Pathway" and "Signalling Pathways Regulating Pluripotency Of Stem Cells" are the same observed in the Profile BM, although these pathways, in Profile A, are regulated by different miRs. The validated target genes, included in "TGF-beta Signalling Pathway" in Profile A are: SP1, SMAD4 and RPS6KB1, which may promote a more mature and profibrotic phenotype in MSCs 48,49 and RBL1, which may inhibit the cells proliferation by arresting the cells in G1 50 . The validated target genes, included in "Signalling Pathways Regulating Pluripotency Of Stem Cells" are: JARID2, promoting more mature phenotype in MSCs; SMAD4, pro-fibrotic 3,51 ; KAT6A and GSK3B, promoting cell proliferation 52,53 ; MEIS1 and FZD1, promoting senescence 54,55 and KRAS, promoting cell survival 56 and SMARCAD1, which may play a role during DNA  Continued www.nature.com/scientificreports www.nature.com/scientificreports/ repair 38 . Some down-regulated miRs, in Profile A, target specific pathways, which have not been identified in Profile BM. "The Regulation Of Actin Cytoskeleton", involving the validated target gene: ARPC5, ITGA5, WASF2, ACTN1, GIT1, IQGAP1 and DIAPH1 promoting actin cytoskeleton reorganization for cell invasion; PAK2/7, FGF9, PPP1CC and KRAS, promoting proliferation 38 , migration, and survival 42,56 , and of note, the target gene PAK2, promoting the cellular survival, was shared between BM-and A-MSCs. The "Wnt Signalling Pathway" including the validated genes: CHD8, GSK3B, TBL1XR1 and CCND2, promoting proliferation, migration and invasion; FZD1, promoting senescence; CSNK1A1, promoting self-renewal; and SMAD4 promoting a more mature phenotype in MSCs and SENP2, that may contribute to escape to TGF-beta signalling and PRKACA a survival kinase 38 .
In conclusion, the comparison of miR profiles of MSCs isolated from BM and A-tissue obtained from SSc patients, identified 2 different signatures and sharing similar functional alterations. The in silico prediction showed that both the BM-and A-MSCs display a similar senescent and profibrotic signature. Interestingly, both the MSCs source showed increased activity of the pathways related to survival ability and activation of compensatory mechanism, and these pathways may help MSCs to survive under the pathological constraints specific of the disease, without reverse the profibrotic phenotype, but this in silico approach is not able to define a hierarchy among these pathways and further studies need to define this setting. It must be pointed out that among the validated genes, potentially up-regulated in these cells, there are some genes involved in the angiogenic mechanisms. The activation of these genes may be involved in the improvement of the vascular conditions observed in the clinical trials, after infusion of these cells in SSc patients. In addition, "TGF-beta Signalling Pathway" is potentially up-regulated in both BM-and A-SSc-MSCs, correlating with previous work suggesting an activation of this pathway during SSc 20, 30 . It has been showed that the increase of TGF-beta in SSc-MSCs may stimulate fibrotic process and also promote Tregs induction 20 , however future studies are ongoing in order to confirm the involvement of these down-regulated miRs in controlling TGF-beta function. In fact, the limitation of any in silico study is that the gene target interaction is a bioinformatics prediction, but it is the first necessary step, to select the potential genes, before starting focused experiments, thus restricting the field of interest and saving from unnecessary experiments. Further studies are ongoing in our laboratory to evaluate in vitro the ability of dysregulated miRs, to modulate the expression of the gene target and to evaluate the biological impact of this miRs on therapeutic function of SSc-MSCs. However, the main benefit of this in silico approach is the possibility to analyse all the identified miRs as a group, more than considering these molecules, separately, providing a larger vision of the biological pathways induced by the whole signature and give an integrated representation of the possible functional differences between BM-and A-MSCs. This bioinformatics approaches substantiate the concept of disease inherent abnormalities of SSc-MSCs, suggesting that these cells might contribute to the disease progression, as already reported in previous papers 3,22,23 . From a translational point of view, a better knowledge of MSCs signature, might allow us, in the future, to better select and potentially manipulate the MSCs to improve the development of MSC-based therapy for SSc.

Methods
All methods were performed in accordance with the relevant guidelines and regulations.
Patients, controls. We enrolled in this study 6 SSc patients, with a very similar clinical pattern. All patients fulfilled the 2013 classification criteria for SSc 58 . Our patients fulfilled the classification criteria in less than one year from the onset of Raynaud's Phenomenon (RP). All SSc patients underwent 20-day washout from any immunosuppressive treatment and one month from intravenous prostanoids. During this period, only proton-pump inhibitors and clebopride were allowed. Patients who could not undergo therapeutic washout, due to severe organ complications, were not enrolled in the study. The study was approved by our local ethics committee (ASL Avezzano Sulmona L' Aquila, protocol number 1092). Demographic and clinical characteristics of the patients are showed in Table 1.
Isolation, culture and immunophenotyping of BM-MSCs. After approval of local ethics committee (ASL Avezzano Sulmona L' Aquila n.1092) and written informed consent from patients, the BM was obtained by aspiration from the posterior superior iliac crest from 3 out of the 6 patients enrolled in the study. Three frozen  www.nature.com/scientificreports www.nature.com/scientificreports/ BM-MSCs samples obtained from age-and sex-matched healthy donors (HC) were purchased from Lonza (USA) and used as control. Samples were placed into tubes containing ethylenediamine tetra acetic acid (EDTA) and the BM cells were obtained by density gradient sedimentation on 12% hydroxyethyl amide. The upper phase was harvested, centrifuged at 700 g for 10 min and plated at a concentration of 5 × 10 3 cells/cm 2 in Dulbecco's modified  www.nature.com/scientificreports www.nature.com/scientificreports/ Eagle's medium (DMEM; Gibco, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, USA), 2 mmol/l L-glutamine (EuroClone, Italy) and 100 U penicillin, 1000 U streptomycin (Biochrom AG, Germany). Both SSc and healthy control (HC) cultures were grown and expanded in flasks at 37 °C, 5% CO2 in a humidified thermostat for 2 weeks until confluence and used for the following experiments at third passage (P3). The International Society for Stem Cell Therapy (ISCT) drew the surface antigen pattern for defining MSCs 59 . According, the P3 isolated cells were analyzed for the surface antigen by flow cytometry (FACSCAN, Becton Dickinson, USA), to assess their purity (CD105, CD73 and CD90 > 95% positive and CD34-, CD45-, CD14-, CD79-, HLA class II (DR) ≤ 2% positive). This assessment was done concurrently with miRs evaluation.
Isolation, culture and immunophenotyping of A-MSCs. Adipose tissue was obtained from the other 3 patients enrolled in the study, after approval of local ethics committee (ASL Avezzano Sulmona L' Aquila n.1092) and written informed consent from patients. Three frozen A-MSCs samples obtained from age-and sex-matched healthy donors (HC) were purchased from Lonza (USA) and used as control. Lipoaspirate (500 mg) was washed to remove excess blood by mixing with an equal volume of phosphate-buffered saline (PBS; Euroclone, Italy). The fat is placed in a sterile tissue culture plate with 0.075% Collagenase Type I prepared in PBS containing 2% P/S for tissue digestion (lysis buffer). Furthermore, the pellet, containing the A-MSCs, was obtained by centrifuging the sample at 2000 rpm for 5 min. After spinning, all the collagenase solution was aspirated, and the pellet resuspended in 1 ml lysis buffer, incubated for 10 min on ice, washed with 20 ml of lysis buffer and centrifuged at 2000 rpm for 5 min. The cell pellet was resuspended in a maximum of 3 ml of stromal medium (DMEM + 20%FBS + glutamine + penicillin/streptomycin), the cell suspension is filtered through 70 μm cell strainer and incubate at 37 °C with 5% humidified CO 2 . The outgrown cells were cultured in fresh medium for another 7-10 days until a confluent monolayer was obtained.
These initial cells, referred to as passage 1 (P1), were further sub-cultured at a seeding density of 100-500 cells/cm 2 and serially passaged in a humidified thermostat for 2 weeks until confluence and used for the following experiments at third passage (P3). The P3 isolated cells were analyzed for the surface antigen by flow cytometry (FACSCAN, Becton Dickinson, USA), to assess their purity (CD105, CD73 and CD90 > 95% positive and CD34-, CD45-, CD14-, CD79-, HLA class II (DR) ≤ 2% positive). This assessment was done concurrently with miRs evaluation.
MiRs profiling. Total RNAs were extracted from P3-MSCs using the mirVana miRs isolation kit (Ambion, Thermo Fisher Scientific, USA) by following the manufacturer's instructions. RNA concentrations and qualities were evaluated by using NanoDrop 2000 (Thermo scientific, USA). Total RNA samples, containing the fraction less than 200 nucleotides, were used for miRs profiling studies. Identical amounts of RNAs extracted from each patient and healthy control were pooled together and subjected (700 ng per RNAs' pool) to qRT-PCR by using the TaqMan MicroRNA reverse transcription kit (Applied Biosystems, USA) and the Megaplex RT primers human pool (Applied Biosystems, USA). Subsequently, microfluidic cards TaqMan array human microRNA A + B v3.0 (Applied Biosystems, USA) were used, according to the manufacturer's instructions. Three replicates for each pooled sample were analysed. MiRs' expression levels were evaluated by comparative assay: samples were analysed on a ViiA7 (Applied Biosystems, USA) and data were processed by ViiA7 software and further elaborated by Expression Suite (v.1.0.3, Applied Biosystems, USA) also at the statistical level. 2 −ΔΔCt method was used to