A Novel OsteomiRs Expression Signature for Osteoblast Differentiation of Human Amniotic Membrane-Derived Mesenchymal Stem Cells

Human amniotic membrane-derived mesenchymal stem cells (hAM-MSCs) are a potential source of cells for therapeutic applications in bone regeneration. Recent evidence reveals a role for microRNAs (miRNAs) in the fine-tuning regulation of osteogenesis (osteomiRs) suggesting that they can be potential targets for skeleton diseases treatment. However, the functions of osteomiRs during differentiation of hAM-MSCs to osteogenic lineage are poorly understood. In this investigation, we discovered a novel miRNAs expression signature corresponding to the matrix maturation (preosteoblast) and mineralization (mature osteoblast) stages of dexamethasone-induced osteoblastic differentiation of hAM-MSCs. Comprehensive miRNAs profiling using TaqMan Low Density Arrays showed that 18 miRNAs were significantly downregulated, whereas 3 were upregulated in the matrix maturation stage (7 days after osteogenic induction) in comparison to undifferentiated cells used as control. Likewise, 47 miRNAs were suppressed and 25 were overexpressed at mineralization stage (14 days after osteogenic induction) in comparison to osteoprogenitors cells. Five out 93 miRNAs (miR-19b-3p, miR-335-3p, miR-197-3p, miR-34b-39, and miR-576-3p) were regulated at both 7 and 14 days suggesting a role in coordinated guidance of osteoblastic differentiation. Exhaustive bioinformatic predictions showed that the set of modulated miRNAs may target multiple genes involved in regulatory networks driving osteogenesis including key members of BMP, TGF-β, and WNT/β-catenin signaling pathways. Of these miRNAs, we selected miR-204, a noncoding small RNA that was expressed at matrix maturation phase and downregulated at maturation stage, for further functional studies. Interestingly, gain-of-function analysis showed that restoration of miR-204 using RNA mimics at the onset of mineralization stage dramatically inhibited deposition of calcium and osteogenic maturation of hAM-MSCs. Moreover in silico analysis detected a conserved miR-204 binding site at the 3′UTR of TGF-βR2 receptor gene. Using luciferase assays we confirmed that TGF-βR2 is a downstream effector of miR-204. In conclusion, we have identified a miRNAs signature for osteoblast differentiation of hAM-MSCs. The results from this study suggested that these miRNAs may act as potential inhibitors or activators of osteogenesis. Our findings also points towards the idea that miR-204/TGF-βR2 axis has a regulatory role in differentiation of hAM-MSCs committed to osteoblastic lineage.


Introduction
Osteoporosis disease and bone fractures are among the major public health concerns worldwide, as they produce a decline in patient mobility and a significant increase in medical care costs [1]. A recent review on this subject summarize the evidence indicating that both skeletal development and bone regeneration depend on the proper functional differentiation of mesenchymal stem cells to osteoblasts [2]. Human mesenchymal stem cells are pluripotent stem cells that undergo a multistage differentiation process in which they proliferate and differentiate into osteocytes, adipocytes, and chondrocytes [3]. From the past decade, human amniotic membrane-derived mesenchymal stem cells (hAM-MSCs) have been used for in vivo construction of tissue-engineered cartilage, bone, and other soft tissues, and they have become a promising tool for the treatment of bone diseases [4]. Osteoblastic differentiation from hMSCs is a highly specialized process involving the coordinated activation of WNT/catenin, FGF, and BMPs/TGF-signaling pathways [5,6], which activate runt-related transcription factor 2 (RUNX2) and Osterix (OSX/Sp7) transcription factors, among others, that in turn activate a complex transcriptional program driving the differentiation of osteoprogenitors towards functional mature osteocytes [7]. Therefore, elucidating the molecular mechanisms that regulate the osteoblastic differentiation is essential to help understand the molecular mechanisms of osteogenesis, but it also may be a guide for the development of novel therapies for the treatment of bone lesions and osteoporosis.
MicroRNAs (miRNAs) are evolutionary conserved single-stranded small RNA molecules of 21-25 nucleotides in length that function as negative regulators of gene expression [8]. MiRNAs function as guide molecules in posttranscriptional gene silencing acting by complementary binding with the 3 untranslated region (UTR) of specific transcripts leading to target mRNA degradation and/or translational repression in P-bodies [9]. Recent evidences suggests an important role for miRNAs in the fine-tuning regulation of genes involved in osteogenesis indicating that these so-called osteomiRs can be promising therapeutic targets for skeleton diseases [10][11][12]. However, the functions of miRNAs during differentiation of hAM-MSCs to osteogenic lineage are still poorly understood. Despite this interest, comprehensive miRNAs profiling studies associated with differentiation of hAM-MSCs to osteoblasts are scarce; thus the relevance of small RNAs in osteogenesis is not yet known. In this investigation, we reported a novel miRNAs expression signature in two stages of hAM-MSCs differentiation to osteoblasts and functionally characterized the role of miR-204 in this cellular process. Implications on the molecular mechanisms regulating the differentiation of hAM-MSCs and the potential therapeutic applications are discussed.

Isolation of Stem Cells from the Human Amniotic
Membranes. Caesarean-delivered term placentas (n=4) were collected from healthy donor mothers. Procedures were approved by the Ethics Committee of the University Hospital "Dr. José Eleuterio González", and each donor gave her consent. Pieces of amniotic membrane (10 x10 cm) were subjected to two enzymatic digestions by adding (i) 0.125% trypsin/0.5 mM EDTA solution at 37 ∘ C for 30 min and (ii) 100 U/mL collagenase type II and 3 mM calcium chloride in the Dulbecco's Modified Eagle's Medium (DMEM) for 2 h at 37 ∘ C, followed by washing with PBS. The resulting cell suspension was filtered, and cells were seeded in 25cm 2 flasks in DMEM containing 10% fetal bovine serum (FBS) and 100 U/mL penicillin, 100 mg/mL streptomycin, and 0.25 mg/mL amphotericin B. The resulting hAM-MSCs were cultured in a humidified 5% CO 2 atmosphere at 37 ∘ C until 90% confluence. All experiments were carried out using hAM-MSCs below 10 passages.

Osteogenic Differentiation of hAM-MSCs.
The osteogenic differentiation was induced as previously described [13]. Briefly, isolates of hAM-MSCs (2x10 3 cells/cm 2 ) were grown in osteogenic medium (DMEM supplemented with 10% FBS, 10 mM -glycerophosphate, 0.25 mM ascorbic acid, and 10 −8 M dexamethasone). Cultures were maintained with medium changes every 2-3 days. Cells were stained with Alizarin Red S dye reactive (pH 4.3) to evaluate calciumrich deposits and monitor the differentiation process as described [14]. Cells were imaged and birefringence staining was quantified using the FluorChem 9900 system.

MicroRNAs Profiling Using TaqMan Low Density Arrays.
Total RNA (70 ng) was isolated from undifferentiated hAM-MSCs (time 0), and cells collected after 7 and 14 days of osteogenic induction using the TRIzol reagent (Invitrogen). MiRNAs expression profiles were obtained using the stem-loop qRT-PCR-based TaqMan Low Density Arrays (TLDAs) covering 677 human mature miR-NAs (https://www.thermofisher.com/order/catalog/product/ 4398965). Briefly, expression of miRNAs was performed by reverse transcription and quantitative real-time polymerase chain reaction using the Megaplex TaqMan Low Density Arrays v2.0 system 4398965 (Applied Biosystems. Foster City, CA) as described by the manufacturer. In order to detect low abundant miRNAs a preamplification step was included. For this, 12.5 l TaqMan PreAmp Master Mix was incubated with 2.5 l Megaplex PreAmp Primers and water up to 22.5 l. Then, 2.5 l of RT reaction was added to the PreAmp mix and sequences were amplified following the next program: 95 ∘ C for 10 min (enzyme activation), 55 ∘ C for 2 min (annealing), 72 ∘ C for 2 min (extension), 95 ∘ C for15 s (denaturation), 69 ∘ C for 4 min (annealing/extension), and 99 ∘ C for 10 min (enzyme inactivation). Then, preamplified products were loaded into the TLDAs for PCR reactions and amplification signal detection was performed in the 7900 FAST real-time thermal cycler (ABI). Tests were normalized using RNU44 as control.

Reverse Transcription and Real-Time Polymerase Chain
Reaction. Quantitative real-time RT-PCR (qRT-PCR) analysis for miRNAs expression was performed using the Taq-Man MicroRNA Assay kits (ThermoFisher). Total RNA (100 ng) was reverse transcribed using a looped RT specific primer, dNTPs (100 mM), reverse transcriptase MultiScribe (50 U/ l), 10X buffer, RNase inhibitor (20 U/ l), and 4.16 l RNase-free water. Retrotranscription reaction (1:15) was mixed with master mix TaqMan (Universal PCR Master Mix, No AmpErase UNG, 2X), and the corresponding specific TaqMan PCR probe. PCR reaction was performed in a GeneAmp System 9700 (Applied Biosystems) as follows: 95 ∘ C for 10 min and 40 cycles at 95 ∘ C for 15 s and 60 ∘ C for 1 min. Tests were normalized using RNU44 as endogenous control.

Prediction of Gene Targets and Gene Ontology Analysis.
Target genes of miRNAs were predicted using TargetScan, PicTar5, miRanda, and DIANA microT software. Only targets predicted by two algorithms were included in further analyses. Cellular pathways and processes potentially affected by let-7c-3p were predicted using DAVID 6.7 software.

Transfection of miR-204
Mimics. Seven days after induction of osteogenic differentiation, hAM-MSCs were transfected with miR-204 mimics (40 nM) and scramble (40 nM) sequence (AM17110, ThermoFisher) as negative control using siPORT amine transfection agent (Ambion). Briefly, miR-204 mimics was diluted in 25 l of Opti-MEM to 40 nM concentrations and then added to wells containing cultured cells in 450 l of DMEM. After 48 h incubation, total RNA was extracted using TRIzol and efficacy of miR-204 restoration was evaluated by qRT-PCR using specific stem-looped RT oligonucleotide and TaqMan probe as implemented in the TaqMan MicroRNA assays protocol.

Effects of miR-204
Mimics on Osteogenic Differentiation of hAM-MSCc. Briefly, 11 days after of osteoblastic differentiation the cells were transfected with 30 nM of miR-204 mimics using siPORT amine transfection agent (Ambion), and 48 h after transfection the calcium-rich deposits were evaluated with Alizarin Red S dye reactive (pH 4.3) which binds to intracellular calcium ions. Cells were incubated with dye for 15 min at room temperature and washed tree times with PBS 1x, and then cells were imaged and birefringent staining was quantified using FluorChem 9900 for comparisons of pixel intensity between control nontransfected and miR-204 mimics transfected hAM-MSCs.

Luciferase Gene Reporter Assays.
A DNA fragment of the 3 UTR of TGF-R2 gene containing the predicted miR-204 binding site and as well as a mutated version of the seed region were cloned into the p-miR-report vector (Ambion) downstream of the luciferase gene. Constructs were verified through automatic sequencing. Then, recombinant pmiR-LUC-TGF R2 wild type and pmiR-LUC-TGF-R2 mutant plasmids were transfected into hAM-MSCs cells using lipofectamine 2000 (Invitrogen). At 24 h after transfection, cells were cotransfected with miR-204 mimics and scramble (40 nM) and incubated for 24 h. Finally, firefly and Renilla reniformis luciferase activities were measured by the Dual-Glo luciferase Assay (Promega) using a Fluoroskan Ascent6 Microplate Fluorometer, and firefly luciferase activity was normalized with Renilla reniformis luciferase data.

Statistical Analyses.
Experiments were performed three times by triplicate and results were represented as mean ± SD. One-way analysis of variance (ANOVA) was used to compare the differences between means. A p<0.05 was considered as statistically significant.

Osteoblastic Differentiation of hAM-MCS In Vitro.
Previously, we have isolated and characterized the biological events underlying the osteoblastic differentiation of the hAM-MSCs subpopulation studied here and established the temporality of proliferation (undifferentiated cells, time 0), matrix maturation (preosteoblast, 7 days), and mineralization (mature osteoblast, 14 days) stages [13,14]. To assess the ability of hAM-MSCs isolates to differentiate into osteogenic lineage, the calcium deposition and expression of the mineralization stage marker collagen type I-alpha (COL1A2) were measured. The hAM-MSCs were induced to differentiation in osteogenic media and maintained for three weeks at 37 ∘ C. Cultures were stained with Alizarin Red at 0, 7, and 14 days after induction and imaged. Results showed a gradual and significant time-dependent increase of calcium-rich deposits in preosteoblasts and osteoblasts in comparison to undifferentiated control cells (time 0) (Figures 1(a) and 1(b)). Congruently, immunoblot assays revealed that the COL1A2 protein expression was significantly increased at 7 and 14 days after induction relative to undifferentiated control cells (Figures 1(c) and 1(d)). These data confirmed that the isolates of hAM-MSCs used in this study efficiently initiated the process of differentiation to osteoblasts as previously described [13,14].

MicroRNAs Expression Signatures of hAM-MSCs Differentiation to Osteoblastic Lineage.
To evaluate the potential contribution of miRNAs in osteoblast maturation of hAM-MSCs, we comprehensively analyzed 667 miRNAs using stemloop qRT-PCR in TaqMan Low Density Arrays (TLDAs) as previously described [15]. Undifferentiated hAM-MSCs were induced to osteoblastic lineage in differentiation media and global miRNAs profiles were assessed at matrix maturation (7 days after osteogenic induction) and mineralization (14 day after osteogenic induction) stages. After comparative 2-ΔΔCt analyses, a total of 93 miRNAs significantly modulated (| log 2(T/N)| > 1.0; p<0.05) were identified. Of these, 18 miRNAs showed a significant downregulation, whereas three were upregulated at matrix maturation stage in comparison to undifferentiated cells (Table 1). Likewise, 47 miRNAs were significantly suppressed and 25 were overexpressed at mineralization stage ( Table 2). This panel of up-and downregulated miRNAs was able to separate the preosteoblast and mature osteoblast groups relative to undifferentiated cells as depicted in the 2-way unsupervised hierarchical clustering shown in Figures 2(a) and 2(b). A volcano plot representation of miRNAs deregulated at 7 and 14 days is shown for easy visualization in Figures 2(c) and 2(d), respectively. Interestingly, five out 93 miRNAs were modulated at both times: miR-19b-3p was upregulated, whereas miR-335-3p, miR-197-3p, miR-34b-3p, and miR-576-3p were downregulated, suggesting an important role in the coordinated guidance of the osteoblastic differentiation process. Then, we proceeded to validate the changes in expression of four selected miRNAs using specific probes in qRT-PCR assays. Representative data at 14 days after induction showed that levels of miR-21, miR-125b, miR-204, and miR-123 detected by qRT-PCR were similar to those found in TLDAs platform (Figure 3).
To understand the biological impact of miRNAs variations during early and late stages of osteogenesis, we computationally predicted their gene targets using a combination of TargetScan, PicTar5, miRanda, and DIANA microT prediction tools. Exhaustive bioinformatic analysis yielded a large list of putative gene targets. In order to associate miRNAs function to specific cellular pathways and processes, the complete list of putative gene targets was filtered using the Gene Ontology (GO) terms function as implemented in DAVID databases. In Tables 3 and 4, we present an overview of the main signaling pathways impacted by the complete dataset of miRNAs regulated at both 7 and 14 days after osteogenic differentiation. For instance, predictions showed that 17 miRNAs modulated at 7 days may be targeting at least 42 different genes involved in the osteogenesis-related TGF-signaling pathway. In addition, 20 miRNAs were predicted to target 65 genes involved in signaling pathways regulating pluripotency of stem cells (Table 3). These data confirmed that an important number of modulated miRNAs are related to known cellular pathways associated with hAM-MSCs osteogenic differentiation.

Overexpressed OsteomiRs Target Multiple Genes Involved in TGF-and WNT/ -Catenin Pathways.
In order to better characterize the impact of modulated miRNAs on the      MSX2 which function as repressors (Figure 4). For instance, miR-19b-3p may stimulate differentiation by targeting FIAT which is a repressor of the osteogenesis-activator ATF4 (Figure 4). Likewise, at both early and late times of osteogenic differentiation, miR-19b-3p may also enhance osteogenesis by targeting AXIN2 which is a repressor of -catenin. A more complex regulation may be occurring for some genes. At late stages of osteogenic differentiation, let-7f-5p, miR-145-5p, and miR-148b-3p were predicted to cooperatively target p21, a repressor of RUNX2, which could stimulate osteoblasts mineralization. In contrast, miR-150-5p could inhibit osteoblastic differentiation by targeting -catenin in undifferentiated  cells. Similarly, at late stage of maturation, the dual targeting of Osteryx (OST) by miR-449b-5p and miR-339b-5p could result in osteogenesis impairment (Figure 4). In addition, we found that a number of target genes could be recognized by more than one miRNA indicating an extensive target gene redundancy and at the same time revealing complex and intricate miRNA/mRNA regulation networks. As an example, in Figure 5 we depicted a miRNA/mRNA interactions network for miRNAs upregulated at 14 days and their predicted targets with known inhibitory functions in TGF-and WNT/ -catenin pathways.

MiR-204 Inhibits the Osteoblast Differentiation by Targeting TGF-R2.
To initiate the study of the molecular mechanisms underlying the regulation of hAM-MSCs differentiation, we focused on miR-204, a noncoding small RNA overexpressed at the matrix maturation stage and downregulated at the maturation stage. Bioinformatics predictions showed that miR-204 may target multiple key genes involved in osteogenesis promotion including ACVR1C, ACVR2B, TGBR2, BMPR1A, BMPR2, SMAD2, and SMAD5 among other; thus it could have a relevant role in osteoblast differentiation. To evaluate the role of miR-204 in osteogenic   differentiation of hAM-MSCs, we first restored its expression using RNA mimics at the onset of mineralization stage (Figure 6(a)). Data showed that miR-204 inhibited deposition of calcium and osteogenic maturation of hAM-MSCs ( Figure 6(b)). Then, we searched for potential downstream targets of miR-204 using different miRNA target prediction tools. In silico analysis detected a conserved miR-204 binding site at the 3 UTR of TGF-R2 receptor gene. Thus it was reliable to propose that restoration of miR-204 levels may negatively regulate osteoblast differentiation through direct targeting of TGF-signaling. To corroborate this hypothesis, we performed luciferase reporter gene assays. A DNA fragment of the 3 UTR containing the TGF-R2 binding site for miR204 was cloned downstream of the luciferase gene into the pmiR report vector (Figure 6(d)). In addition, point mutations in the predicted miR-204 binding site of the 3 UTR were included in the analysis. Recombinant pmiR-LUC-TGF-R2 plasmid was transfected during differentiation of hAM-MSCs and luciferase activity was analyzed after 24 h of transfection. Data showed that ectopic expression of miR-204 and cotransfection of pmiR-LUC-TGF-R2 construct resulted in a significantly reduction of the relative luciferase activity in comparison with controls ( Figure 6(d)). When mutated sequences of the 3 UTR TGF-R2 were assayed, no significant changes in luciferase activity were found, indicating that miR-204 binding was specific.

Discussion
Human MSCs were initially isolated from bone marrow [18], and they have been later identified in several fetal tissues such as the liver, bone marrow, and pancreas, in  the endothelial umbilical vein and in the preterm blood of the fetus [19]. The International Society of Cellular Therapy (ISCT) has described MSCs as cells that express the surface markers CD73, CD105, CD90, and CD44 in the absence of hematopoietic stem cell markers and differentiate into mesenchymal lineages such as osteoblasts, adipocytes, and chondroblasts in vitro [4]. Several miRNAs modulate osteogenic differentiation; thus they have been dubbed as osteomiRs. For instance, miR-125b regulates osteoblastic differentiation of human MSCs through targeting of BMPR1b [20], whereas miR-223 suppresses osteoblast differentiation by inhibiting DHRS3 [21]. However, miRNAs functions during differentiation of hAM-MSCs to osteogenic lineage have not been completely explored. In order to contribute to the understanding of miRNAs roles in osteogenic differentiation of hAM-MSCs, here we performed a miRNAs profiling in MSCs obtained from human placenta. Our findings indicated a novel miR-NAs expression signature corresponding to preosteoblast and mature osteoblast stages of dexamethasone-induced osteoblastic differentiation of hAM-MSCs. Bioinformatic predictions showed that these miRNAs may target multiple genes involved in regulatory networks driving osteogenesis including members of BMP, TGF-, and WNT/ -catenin signaling pathways. Importantly, our results confirmed that a number of miRNAs have well-known functions in osteogenesis. For instance at 7 days of differentiation, we found a significant repression of miR-29b relative to osteo-progenitors cells (Table 1). Likewise, ectopic expression of miR-29b resulted in low levels of osteogenesis inhibitors CDK6 and HDAC4 and stimulated the osteogenic differentiation of unrestricted somatic stem cells from human cord blood [22]. At 14 days, we identified that miR-119a-5p and miR-22-3p previously reported in osteogenesis are overexpressed in hAM-MSCs (Table 2). Similarly, it was described that upregulation of miR-22 promotes the osteogenic differentiation of human adipose tissue-derived mesenchymal stem cells by repressing the osteogenesis repressor HDAC6 [23]. These data suggested that miRNAs detected here as regulated could be regulating the differentiation of hAM-MSCs, importantly; most of miR-NAs detected here as modulated have not been previously connected to osteoblastic differentiation, suggesting that they may represent novel miRNAs with potential therapeutic applications.
In addition, we found that the set of upregulated miRNAs may regulate a large list of key genes involved in TGFand WNT/ -catenin pathways, which are critical for osteogenic differentiation and bone formation. This panel of deregulated miRNAs represents a guide for further functional characterization of the osteogenesis process. To go beyond the miRNAs profiling, we initiated the functional analysis of one deregulated miRNAs related to osteogenesis. Our data indicated that restoration of miR-204, a noncoding small RNA that was expressed at matrix maturation stage and downregulated in maturation stage, impaired differentiation at least by targeting TGF-R2 which is regulator of RUNX2 gene, the master regulator of stem cells differentiation.

Conclusions
In summary, here we reported a novel miRNAs signature during differentiation of hAM-MSCs to osteogenic lineage which may target multiple genes involved in regulatory networks driving osteogenesis including members of BMP, TGF-, and WNT/ -catenin signaling pathways. Our findings suggested an unexpected regulatory role for the miR-204/TGF-R2 axis in differentiation of hAM-MSCs committed to osteoblastic lineage. These data may serve as a guide for further functional analysis focused in the implementation of osteomiRs as novel therapeutic tools in bone diseases.

Data Availability
All data used to support the findings of this study are included within the article. The raw data obtained from the miRNAs BioMed Research International 13 profiling in TLDAs platform used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare no conflicts of interest.