Silencing BRE Expression in Human Umbilical Cord Perivascular (HUCPV) Progenitor Cells Accelerates Osteogenic and Chondrogenic Differentiation

BRE is a multifunctional adapter protein involved in DNA repair, cell survival and stress response. To date, most studies of this protein have been focused in the tumor model. The role of BRE in stem cell biology has never been investigated. Therefore, we have used HUCPV progenitor cells to elucidate the function of BRE. HUCPV cells are multipotent fetal progenitor cells which possess the ability to differentiate into a multitude of mesenchymal cell lineages when chemically induced and can be more easily amplified in culture. In this study, we have established that BRE expression was normally expressed in HUCPV cells but become down-regulated when the cells were induced to differentiate. In addition, silencing BRE expression, using BRE-siRNAs, in HUCPV cells could accelerate induced chondrogenic and osteogenic differentiation. Hence, we postulated that BRE played an important role in maintaining the stemness of HUCPV cells. We used microarray analysis to examine the transcriptome of BRE-silenced cells. BRE-silencing negatively regulated OCT4, FGF5 and FOXO1A. BRE-silencing also altered the expression of epigenetic genes and components of the TGF-β/BMP and FGF signaling pathways which are crucially involved in maintaining stem cell self-renewal. Comparative proteomic profiling also revealed that BRE-silencing resulted in decreased expressions of actin-binding proteins. In sum, we propose that BRE acts like an adaptor protein that promotes stemness and at the same time inhibits the differentiation of HUCPV cells.


Introduction
BRE is also known as TNFRSF1A modulator and BRCC45. This highly conserved gene was initially identified as a stressresponsive gene -whose expression was inhibited following DNA damage and retinoic acid treatment [1]. BRE shares no homology with other known gene products [2]. The protein contains two putative ubiquitin E2 variant domains but lacks critical cysteine which is required for ubiquitination [3]. BRE protein is present in both the cytosol and nucleus. In the cytoplasm, BRE binds to the cytoplasmic region of p55 TNF receptors to suppress TNF-a induced activation of NF-kB [4]. The protein can also bind to Fas to inhibit the mitochondrial apoptotic pathway [5]. BRE is also found to be a component in BRISC complex that specifically cleaves lysine63-linked ubiquitin [6]. BRE may function as a key adaptor protein which assembles the different components of the BRISC complex [7]. In the nucleus, BRE is a component of the DNA damage responsive BRCA1-RAP80 complex. BRE protein acts as an adapter that links the interaction between NBA1 and the rest of the complex. This adapter modulates the ubiquitin E3 ligase activity of the BRCA1/BARD1 complex by interacting with MERIT 40 which enhances cellular survival following DNA damage [7,8,9].
BRE has been extensively studied in lung tumour, hepatocellular carcinoma and oesophageal carcinoma [10,11,12,13] showing that it promoted cell survival. Nevertheless, the function of BRE in stem cells has never been investigated. BRE is expressed very early on in embryonic development; at the 2-cell stage, in the inner cell mass cells of blastocysts and even in embryonic stem cells. In this context, we want to establish why BRE was expressed so early in development and whether it was involved in maintaining stemness and cell differentiation. It is generally accepted that stress-responsive genes, like BRE, play a crucial role in biological processes such as cell survival, differentiation, apoptosis and regeneration. To address our questions, we employed HUCPV progenitor cells as our experimental cell model [14]. These cells are normally found in the perivascular regions of human umbilical arteries and vein and contain a rich source of commercially valuable mesenchymal stem cells (MSCs). The HUCPV cells were found to have a colony forming unitfibroblast (CFU-F) frequency of about 1:300, which is far higher than that of bone marrow (1:10 4 -1:10 6 , depending on age) or umbilical cord blood (1:200 million). The HUCPV cells also showed a higher proliferative potential and expressed higher levels of CD146 (a putative MSC marker) in comparison to MSCs obtained from bone marrow [15]. These cells also express surface antigens CD44, CD73, CD90, CD105 and CD106 but do not express CD34 or CD45 [14,15,16]. HUCPV cells are multipotent and capable of differentiating into all mesenchymal lineages in vitro [15,16,17]. Notably, these cells contribute to rapid connective tissue healing in vivo by producing bone and fibrous stroma [18]. Besides being multipotent, these cells are immunoprivileged making them less likely to be immune-rejected as allografts [16,19]. Furthermore, these foetal cells express the embryonic cell markers SSEA-4, RUNX1 and OCT4 [20]. HUCPV cells can be obtained non-invasively, making them an ideal source for stem cells therapies.

Cell culture
The human umbilical cords were obtained from the Department of Obstetrics and Gynecology and were approved by Chinese University of Hong Kong Clinical Research Ethics committee (project reference number CRE 2011.116). This is centrally registered with the Hong Kong Health Authority. The physician obtained verbal informed consent from the mother for use of the umbilical cord in research. The cords were then dissected aseptically with the aid of a dissecting microscope as described by Sarugaser et al. [16]. Briefly, the amniotic epithelium was removed from the cord with forceps and scissors ( Figure 1A). The two umbilical arteries and umbilical vein were then separated from the cord using forceps. The vessels were then tied in a loop at each ends using sutures as shown in Figure 1B and the placed in 80 U/mL type I collagenase (Gibco) and 0.01 U/mL in a 50 mL Falcon tube. The digestion was carried out in a shaker for four hours at 37uC. The extracted cells were then centrifuged at 500 rpm for 5 min at room temperature. The supernatant was next centrifuged at 1,500 rpm for 3 min at room temperature. The pellet of cells were resuspended in regular growth medium consists of DMEM/F12 supplemented with 15% embryonic stem cellqualified fetal bovine serum (ESQ-FBS), 100 units/mL penicillin and 100 mg/mL streptomycin (all from Gibco) and seeded into culture dish coated with 1% gelatin in ddH 2 O. The cells are maintained in a 5% CO 2 humidified incubator (Thermo Scientific). After one week culture, the isolated HUCPV progenitor cells became confluent for the analysis ( Figure 1C).

Flow cytometry
The crude HUCPV progenitor cells were purified by flow cytometry. Briefly, the confluent culture was trypsinized into suspension and incubated with anti-human CD44, CD90, CD105 and CD146 conjugated PE antibodies for positive selections and anti-human CD34 and CD45 conjugated PE antibodies for negative selection. All antibodies were purchased from BD Biosciences. The immune reactions were performed at 4uC for 20 min. The cells were analyzed and sorted using a FACSAria flow cytometer (BD Biosciences) with FACSDiva software (BD Biosciences).

Mouse embryonic stem cell (ESC) culture
Mouse ES cell line (AINV15, obtained from ATTC) was cultured on 13 mm glass coverslips in 1,400 U/ml of LIF (Millipore) and expanded by co-culture with 10 mg/ml mitomycin C-inactivated mouse embryonic fibroblasts to inhibit differentiation. To induce ESC differentiation, LIF was withdrawn from the culture medium for 24 hours and then the cells were fixed in 10% formalin. Along with undifferentiated ESC cultures, they were processed for immunofluorescent staining with BRE and OCT4 antibodies. Each immunofluorescent staining analysis was performed in triplicate.

Immunofluorescence microscopy
HUCPV cells, cultured on glass coverslips, were fixed in 10% formalin and permeabilized with 0.5% Triton X-100 (Sigma) with 0.1% SDS (Sigma) for 30 min. The samples were then washed three times with PBS and blocked with 2% BSA with 5% normal horse serum for 1 hour. Afterward, the samples were incubated with primary antibody overnight. Primary antibodies used in this study include: CD146 (Zymed, Invitrogen), Ki-67 (Santa Cruz), SOX9 (Abcam), type I collagen (Millipore), type II collagen (Millipore). Non-specific antibody binding were then washed with PBS with 0.05% Tween-20 (PBST) three times for 10 min and PBS for 5 min. Then secondary antibody (Jackson ImmunoResearch Laboratories) was added and incubated for 1 hour. The unbound antibodies were washed with PBST three times for 10 min and PBS for 5 min. DAPI was added for visualizing the nucleus. The fluorescent signal was detected using a confocal microscope (Olympus, FluoView 1000) under a Fluequipped with 406 Zeiss PlanNeofluo objectives.

RT-qPCR and statistical analysis
Total RNA was isolated from cells using TRIzolH reagent (Invitrogen) according to the manufacturer's instruction. The purity, integrity, and concentrations of RNA were evaluated using the Nanodrop Bioanalyzer (Nanodrop Technologies). First strand cDNA was obtained from 1 mg total RNA reverse transcribed using Ready-To-Go You-Prime First-Strand beads and oligo(dT) following the manufacturer's instruction (GE Healthcare). The resulting cDNAs were amplified using SYBRH Premix Ex Taq mix (Takara Biotechnology) and analyzed by a 7900HT Fast Real-Time PCR System (Applied Biosystems). The shuttle PCR protocol included initial denaturation (95uC for 30 second), annealing (40 cycles of 95uC for 5 second and 60uC for 30 second) and final extension (95uC for 15 second, 60uC for 1 minute, and 95uC for 15 second). Genes of interest were analyzed and normalized against house-keeping gene, GAPDH. The other internal house-keeping gene, glucuronidase-b (gusb) was selected to check the stability of GAPDH. Experiments were performed in triplicates. The expression levels were calculated by 2 2DDCT method [21]. For Statistical analysis, the data was presented as mean 6 standard deviation. Statistical significance was determined by Student's t-test. The sequences of the primer sets used for the RT-qPCR reactions are listed in Table S1. The primers were designed with PrimerDepot database (http://primerdepot. nci.nih.gov/) or Primer3Plus software (http://www.pubmed.de/ cgi-bin/primer3/primer3plus.cgi).

Transfection with siRNA in HUCPV cells
Transient BRE silencing was performed using siRNAs and Lipofectamine TM RNAiMAX (Invitrogen) transfection reagent. The transfection was conducted following the manufacturer's protocol using 12 pmol of siRNA together with 1 mL transfection reagent in Opti-MEM no serum medium (Invitrogen). The target sequence of BRE-siRNA was: AAC TGG ACT GGT GAA TTT TCA. The control (Ctl) is a scrambled siRNA was purchased from Healthcare and Co., Hong Kong. The experiments were performed in triplicate.

Microarray analysis
Total RNA was extracted from HUCPV cells using NucleoS-pinH RNA II (Macherey-Nagel) according to manufacturer's instructions. The RNA integrity was determined by Nimblegen bioanalyzer prior to microarray hybridization (Roche). Then, 10 mg of total RNA was reverse transcribed by SuperScript TM Double-Stranded cDNA Synthesis Kit (Invitrogen). For Cy3 cDNA labelling, 1 mg of double-stranded cDNA sample was labelled using the NimbleGen One-Color DNA labelling kit (Roche) following the manufacturer's manual. Briefly, 1 mg cDNA was labelled with Cy3 random nanomers using Klenow fragment. Followed by incubation at 37uC for 2 hours, the labelling reaction was next terminated using EDTA. After isopropanol precipitation, the Cy3 labelled cDNA pellet was rinsed with 80% ethanol and vacuum dried on low heat. Cy3 labelled cDNA was rehydrated in 25 ml PCR grade water and quantified using Quan-iT PicoGreen dsDNA assay kit (Invitrogen). The analysis was performed in triplicate.
Array hybridization was performed on NimbleGen human 126 135K v5.1 high-density single channel oligonucleotide arrays (Roche). Briefly, 4 mg of Cy3 labelled cDNA was vacuumed dried and resuspended in 3.3 ml sample tracking controls and hybridized at 42uC in NumbleGen Hybrization System 4 for 16 hours. Next, the array was washed using the NimbleGen wash buffer kits, the array slide was spin dried and scanned using the NimbelGen MS 200 Microarray Scanner. Scanned images were analyzed using a DEVA software Version 1.0.1 (Roche) and Partek Genomics Suite version 6.12.0103 (Partek). Array data has been deposited in NCBI Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/ geo/) number GSE39948.

Comparative Proteomics
Total protein lysates were extracted from HUCPV cells transfected with BRE-siRNA or Ctl-siRNAs. Extraction condition, two-dimensional gel electrophoresis (2-DE), MALDI-TOF mass spectrometry and bioinformatics search were performed as we previously reported [22,23,24]. Briefly, first dimensional electrophosis (1-DE) was performed on an IPGphor IEF system using 11cm long IPG electrode strip with pH 3-10 gradient (Amersham Biosciences, UK) and an Ettan IPGphor Strip Holder (Amersham Biosciences, UK). 150 mg of protein was applied for each IPG strip. The total volume of protein sample with rehydration buffer (8M Urea, 2% CHAPS (w/v), 1% IPG buffer (v/v), 40 mM DTT) loaded onto the strip holder was 210 ml. The sufficient volume of IPG Cover Fluid was applied to each strip so as to minimize evaporation and urea crystallization. The rehydration step was done under voltage and followed by a separation process. The electrophoresis condition for step 1 was 30 V for 13 hrs; step 2 was 500 V for 1 hr; step 3 was 2000 V for 1 hr and step 4 was 5000 V for 20 hrs. The program was stopped when the total volt-hours reached 40000. After the 1-DE was completed, the sample strips were removed from the strip holders. Each strip was then treated with 1% DTT in 6.5 ml of equilibration buffer (50 mM Tris, 6M of urea, 30% glycerol, 2% SDS, 0.1% bromophenol blue) for 30 min. The medium was then changed to 1% iodoacetamide (IAA, w/v, Sigma-Aldrich, USA) dissolved in the 6.5 ml of the same equilibration buffer. The strips were treated in the solution for 30 min and then loaded onto 12% SDS-polyacrylamide gels with 0.2% agarose in electrophoresis running buffer (25 mM Tris, 192 mM glycine, 0.1%SDS, adjust to pH 8.3). Protein markers (20 to 120 kDa, Fermentas Life Sciences) were also loaded into the gel for determining the size of all the proteins resolved in the gel. The 2-DE was performed in an ISO-DALT apparatus (Hoefer Scientific Instruments) at room temperature under constant voltage 100 V till the dye front reached the bottom of the gel. The gels were then fixed in 50% methanol, 12% acetic acid and 0.5 ml 37% formaldehyde for 1 hour. After fixation, the gels were washed in MilliQ water (46), 50% ethanol (v/v) for 20 min (26), 0.02% sodium thiosulphate (w/v, Merck, UK) for 10 min and distilled water (36). Subsequently, the gel was stained in silver solution (0.15% silver nitrate in 0.75 ml 37% formaldehyde) at 4uC for 1 hr. After several brief washes, the gels were developed in developer solution (1 ml 37% formaldehyde, 30 g sodium carbonate and 2 mg sodium thiosulphate in one liter buffer) until the desired staining intensity was attained. The gels were then immersed in 5% acetic acid (v/v, BDH Chemicals Ltd., UK) for 5 mins to terminate the staining process. Finally, the silver stained gels were scanned using a GS 800 Densitometer (Bio-Rad Laboratories, USA) and the images captured were used for image analysis. The protein spots on the gel were analysed using a PDQuest 2D Analysis Software version 7.13 PC (The Discovery Series, Bio-Rad Laboratories, USA). Each experiment was performed in triplicate.

Protein identification by mass fingerprinting
All protein spots of interest were isolated from the 2D-gel and processed for silver destaining. The gel pieces were first washed in MilliQ water, immersed in 200 ml of destaining solution (15 mM potassium ferricyanide and 50 mM sodium thiosulphate) and then incubated at room temperature until they turned colourless. Each gel pieces was then washed in 400 ml of MilliQ water for 15 min (36). The destained gel pieces were then equilibrated in 200 ml of 10 mM ammonium bicarbonate/ 50% acetonitrile for 15 min. The gel was dehydrated in 200 ml of acetonitrile for 15 min and dried at 30uC for 5 min. The gels were digested with 15 mg/ml of trypsin in 40 mM ammonium bicarbonate/50% acetonitrile (v/v) at 35uC for 16 hrs. Three ml of extraction solution (50% acetonitrile (v/v) and 5% trifluoroacetic acid (Fluka Chemika, Switzerland) were used to stop the reaction. Three ml of reaction mixture was mixed with a-cyano-4-hydroxycinnamic acid matrix and then spotted onto a sample plate for ESI-MS/MS analysis (Bruker Daltonics, USA). The mass spectrums generated were analysed using a Bruker Daltonics software and by mass fingerprinting, which were submitted to the SwissPort bioinformation stations using MASCOT 2.2.07 engine search.

Immunoprecipitation (IP) Assay
Ctl-siRNAs and BRE-siRNA treated cell samples were collected and lysed by in denature buffer (50 mM Tris pH 8.0, 150 mM NaCl, 1% NP-40 and 0.5% sodium deoxycholate) inside a microcentrifuge tube. After the samples were centrifuged for 20 mins at 12,000 rpm at 4uC, the supernatant were aspirated and place in a fresh tube. The lysates were pre-cleared with Sepharose beads for 1 hr on ice and spun in centrifuge to discard the bead pellet. The pre-clear lysates were incubated with the appropriate concentration of antibody in the Sepharose beads, overnight. When the incubation was completed, the lysates were washed in lysis buffer 36. Finally, the lysates were added the 26 loading buffer and boiled at 95uC for 5 mins. The lysate samples were resolved and the proteins were analysed by routine western blot analysis using the OdysseyH CLx infrared imaging system. The actin, ANAXII and BRE were used for the IP assay. The Assay was performed in triplicate.

Scratch cell migration assay
The scratch assay was performed according to methods described by Liang et al. [56]. The Ctl-siRNAs and BRE-siRNA treated HUCPV cells were seeded onto 4-well culture plates. The cells were then treated with 2 mg/ml mitomycin C (Sigma, USA) for 1 hr to prevent further cell proliferation (which could confound our interpretation of the cell migration analysis). After the treatment, a sterile p10 pipet tip was used to scrap off cells from the center of the monolayer to create a gap in the culture. Reference lines were etched onto the bottom of the plastic culture dishes to define the position of the gap/wound. Photos were taken of the Ctl-siRNAs and BRE-siRNA treated cells migrating into the wound area at day 0-day 3 incubation. The photographic images captured were quantitatively analysed by establishing the number of cells that have migrated into the gap. The assay was performed in triplicate.

HUCPV cells are multipotent
HUCPV cells are normally found in the perivascular regions of human umbilical arteries and veins, which we isolated according to procedures described by Sarugaser et al. [16]. Briefly, both ends of the umbilical vessels were first ligated and then the perivascular region containing the HUCPV cells was digested using collagenase ( Figures 1A-C). The crude cell extracts produced were then sorted and purified for HUCPV cells using flow cytometery for CD34 2 , CD45 2 , CD44 + , CD90 + , CD105 + and CD146 + surface markers ( Figure 2A). Immunofluorescence staining revealed that our HUCPV cells strongly expressed CD146, which is a marker for mesenchymal stem cell ( Figure 2B). The cells were negatively stained for CD34, which is a hematopoietic stem cell marker ( Figure 2C).
To assess the developmental potential of HUCPV cells, they were treated with osteogenic and chondrogenic inducing media. These cells were able to differentiate into osteoblasts after 3 weeks induction. Alizarin red S staining revealed the presence of calcified bone matrix in the cell culture as the HUCPV cells underwent osteogenesis ( Figure 3A). These osteogenic induced HUCPV cells also expressed collagen type I while the control cultures only expressed background levels of collagen ( Figure 3B). Similarly, the HUCPV cells were tested for their chondrogenic potential and were treated with chondrogenic induction medium. Alcian blue staining revealed that the induced HUCPV cells deposited sulfated proteoglycans 4 weeks after treatment ( Figure 3C). These cells also expressed the key transcription factor SOX9 and formed cell aggregates not found in control cultures ( Figure 3D).

BRE expression during HUCPV and embryonic stem cell (ESC) differentiation
It has been reported that BRE expression was down-regulated when tumor cells were treated with retinoic acid, a differentiating agent [1]. Therefore, we examined BRE expression in HUCPV cells when they are induced to differentiate. We established that BRE expression was down-regulated when HUCPV cells were induced in osteogenic medium after 3 weeks or chondrogenic medium after 4 weeks as compared with control medium ( Figure 4A). Immunofluorescence staining using BRE antibody also confirmed that BRE protein expression was decreased as HUCPV cells differentiated into osteoblasts and chondrocytes ( Figure 4B). Therefore, we hypothesize that BRE suppresses HUCPV cell differentiation.
Besides HUCPV cells, we also examined the co-localization of Oct4 (pluripotent marker) and BRE in mouse embryonic stem cells (ESCs). In the presence of LIF, we found that BRE and OCT4 were strongly co-expressed in undifferentiated ESCs ( Figure 5 A-C). When LIF was withdrawn from the culture medium to allow the ESCs to differentiate, it resulted in a synchronized reduction of BRE and OCT4 expression ( Figure 5D).

Silencing BRE accelerates chondrogenic and osteogenic induction
We have used siRNAs to silence BRE expression to test our hypothesis that BRE inhibits HUCPV cell differentiation. HUCPV cells were transfected twice with either BREor ctl-(scrambled control) siRNAs at 24-hour intervals, prior to inducing the cells with differentiation medium. We first confirmed that transfection with our BRE-siRNA was able to significantly reduce both BRE mRNA and protein expressions in HUCPV cells ( Figure 6A). In addition, the growth rate and survival rate of BREsilenced HUCPV cells were unaffected when compared with the control. This is evident by the duration required for BRE-silenced and control cells to reach confluency and the total number of DAPI labeled cells present in the cultures ( Figure 6B).
In this study, HUCPV cells were transfected with BRE-siRNA or Ctl-siRNAs prior to exposure to osteogenic inducing medium for 3 days. QRT-PCR analysis of these cultures revealed that there were significant increase in osteopontin (OPN), osteocalcin (OC), collagen I (COL1) and RUNX2 expression in the induced BREsilenced cells compared with induced cells transfected with Ctl-siRNAs ( Figure 7A). Correspondingly, Alizarin red S staining revealed that the induced BRE-silenced HUCPV cultures contained calcified bone matrix whereas no staining was found in the control cultures ( Figure 7B). Normally, 3 weeks were required to induce HUCPV cells to undergo osteogenesis in osteogenic inducing medium. Our results revealed that silencing BRE expression significantly accelerated this response as evident by the HUCPV cells expressing osteogenic gene markers and calcified matrix within 3 days of osteogenic induction.
The same approach was used to assay whether chondrogenesis could be speeded up by transfecting HUCPV cells with BRE-siRNA prior to exposure to chondrogenic inducing medium. RT-qPCR analysis revealed that the BRE-silenced HUCPV cells strongly expressed chondrogenic marker SOX9 and extracellular matrix proteoglycan versican (VCAN) after 10 days incubation in chondrogenic inducing medium, as compared with cells transfected with Ctl-siRNAs ( Figure 8A). However, there was no significant up-regulation of biglycan (BGN), a small leucine-rich proteoglycan found in extracellular matrix ( Figure 8A). Immunofluorescence staining confirmed that BRE-silenced HUCPV cells strongly expressed SOX9 after 10 days induction but not by HUCPV

Microarray analysis of BRE-silenced HUCPV cells
Since silencing BRE expression could accelerate induced chondrogenic and osteogenic differentiation, we postulated that BRE might play an important role in regulating stemness and differentiation in HUCPV cells. In this context, we performed microarray expression profiling to ascertain the global changes in gene expression following BRE-silencing. HUCPV cell cultures were transfected twice with either BREor Ctl-siRNAs over a 24hour interval. The transcriptome of BRE-silenced cells were then compared with cells transfected with Ctl-siRNAs. Only genes whose expression differed by at least 1.4-fold with a P value less than 5% were considered significant and deemed genes that are differentially expressed. We determined that there were 1301 genes upregulated and 1433 genes down-regulated in BRE-silenced cells compared with the control cells. Hierarchical clustering performed on differentially expressed genes are shown in Figure 9A. These genes were enriched by gene ontology according to biological processes ( Figures 9B). Cell proliferation, growth, and metabolic processes categories were the top three enriched scores. To specifically elucidate how BRE-silencing modulates stem cell function, gene sets that were significantly associated with developmental processes were investigated.

BRE and Epigenetic regulation
Expressions of polycomb complex genes, which are key epigenetic regulators, were also found to be altered by BREsilencing. There are two main Polycomb-repressive complexes (PRCs): PRC1 and PRC2 which are both capable of repressing hundreds of genes in mESCs. Components of the PRC1 and PRC2 genes that we found down-regulated by BRE-silencing include PHC3, SCML2, EZH2, SUZ12P and PHF19; whereas, the genes that were up-regulated include L3MBTL and JARID2 (Figures S2A and B). Other genes involved in chromatin modifications, such as DNA methylation and histone modifications, were also altered in response to BRE-silencing. In our study, the expression of HDAC9 was up-regulated following BRE-silencing. In addition, non-histone chromatin associated genes CHAF1A, CHAF1B, and POLE2 were all down-regulated in response to BRE-silencing.

BRE and TGF-b/BMP signaling
TGF-b/BMP signaling pathway has been extensively reported to play a crucial role in maintaining hESCs and hMSCs in an undifferentiated state and also in suppressing cell differentiation [32,33,34,35,36,37]. Consistent with these reports, we have found that components of this signaling pathway were down-regulated by BRE-silencing. Specifically, SMAD6, SMAD8, and their downstream target genes ID1, ID2, ID2B, and ID3 were all downregulated ( Figure S3A-C).

BRE and chemokines
The effects of BRE-silencing on chemokines and cytokines expression by HUCPV cells were investigated. The chemokines that were up-regulated include CCL3, CXCL3, CXCL5, CXCL10, CXCL11, and CXCL12 ( Figure S5A and B). A number of cytokines and receptors such as TNFSF9, TNFSF15, TNFSF18, TNFRSF19, , collagen I (COL1) and RUNX2 gene expression patterns were strongly up-regulated in BRE-silenced HUCPV cells compared with cells treated with Ctl-siRNAs. All transfected cells were induced in osteogenic medium (ost) for 3-day. As a negative control, the BRE-siRNA and Ctl-siRNAs transfected cells were maintained in control medium (Ctl) 3-day for comparison. Gene expression was normalized against GAPDH. As a control, the stability of GAPDH during osteogenic differentiation was also investigated using GUSB for normalization. The results revealed that GAPDH expression did not alter significantly during differentiation -supporting our use of GAPDH as an internal control.  IFNA1, IFNA2, IFNA4, IFNA5,  IFNA13, IFNA14, TNFAIP8, IL11, IL16, IL20, HTR2A, IFNA

BRE-silencing alters the proteome of HUCPV cells
We have also examined the proteome of BRE-silenced HUCPV cells and compared it with the proteome of HUCPV cells transfected with Ctl-siRNAs. The Protein lysates from both specimens were first resolved by 2-DE. Proteins that were determined to be differentially expressed were then identified by MALDI-TOF mass spectrometry. Seventy differentially expressed protein spots were detected. There were 19 protein spots that were up-regulated while 51 protein spots were down-regulated as a consequence of BRE-silencing. These proteins were categorized into their role in terms of molecular functions which was achieved by using the Database for Annotation, Visualization and Integrated Discovery (DAVID) program (Table S2) [44,45].
Functional annotation clustering pointed to alterations in cytoskeletal protein binding in response to BRE-silencing in HUCPV cells. Furthermore, BRE-silencing response proteins appear to be involved in bindings of DNA, RNA, ATP and unfolded protein. In addition, the differentially expressed proteins have been annotated to regulate enzymatic activities such as isomerase and oxidoreductase activities.
We selected actin and ANXA2 for further investigation. Immunofluorescence staining microscopy confirmed our proteomic results that actin and ANXA2 expressions were reduced in BRE-silenced HUCPV cells (Figures 11A and B). Furthermore, in co-immunofluorescence staining of cultured HUCPV cells, we found actin and BRE were was co-localized in the cytoplasm. Likewise, ANXA2 was also co-localized with BRE. In addition, phalloidin staining revealed that the highly organized F-actin structures seen in control HUCPV cells were totally disassembled in BRE-silenced HUCPV cells ( Figure 12A). We also observed striation patterns when BRE antibody was used in immunofluorescence studies. Therefore, we performed immunoprecipitation study to determine whether actin and ANXA2 can directly bind endogenous BRE. We used monoclonal BRE specific antibody to pull down the proteins in cell lysates. We determined that BRE can directly associate with actin and ANXA2 ( Figure 12B). Since actin and ANXA2 are also involved in regulating cell motility, we investigated whether cell migration was altered in BRE-silenced HUCPV cells using a Scratch assay. We observed that BREsilenced cells migrated at a slower rate than cells transfected with Ctl-siRNAs in the Scratch assay. Approximately 4068.6% of the area of the gap was filled up with BRE-silenced cells compared with 96610% for Ctl-siRNAs treated cells ( Figure 12C). The experiment was repeated in triplicate. The statistical difference of P values were determined by t-test; p,0.05, was considered significantly different.

BRE-silencing alters expression of stemness-related proteins
We have closely examined the proteomic results for proteins that maintain stemness and regulate stem cell differentiation. The results revealed that calreticulin (CALR), far upstream elementbinding protein 1 (FUBP1) and pyruvate kinase isozymes M1/M2 (PKM2) expression were all down-regulated when BRE was silenced in HUCPV cells (Table S4).

Discussion
In this study, we demonstrated that HUCPV cells transient transfected with BRE-siRNA could accelerate osteogenic and chondrogenic differentiation when they are cultured in the appropriate inducing medium. Normally, osteogenic induction of HUCPV cells requires approximately three weeks to take place while chondrogenic induction requires even longer, four weeks. In contrast, when BRE expression was silenced, HUCPV cell differentiation was greatly speeded up -3 days instead of 3 weeks were required for osteogenic induction and 10 days instead of 4 weeks for chondrogenic induction. Previous studies have demonstrated that using BMP-2 and dexamethasone, it was possible to rapidly induce rat bone marrow stromal cells to become osteoblasts, requiring only 8 days [47]. Presently, we revealed that silencing BRE expression prior to osteogenic induction could enhance osteogenesis even faster than by adding exogenous BMP-2. This observation is particularly important in the field of regenerative medicine, especially in the context of non-union bone and cartilage fractures when rapid repair is highly desirable to prevent later complications.
To elucidate the function of BRE in HUCPV cells, we employed microarray analysis. We identified a number of BRE targeted genes in HUCPV cells that have important roles in maintaining stemness. We grouped these genes into different categories. In BRE-silenced HUCPV cells, we observed OCT4, FGF5, and FOXO1A expression was down-regulated and these transcriptional genes are essential for maintaining pluripotency in ES cells. Especially OCT4, which is part of the core transcriptional regulatory circuit that maintains stem cell pluripotency by activating self-renewal genes and suppressing genes that promote differentiation [26,27]. This finding is further collaborated by the fact that HMGB2, a co-activator of OCT4 activity, was also correspondingly down-regulated following BRE silencing [48]. In addition, we also established that pluripotent epiblast marker FGF5 and the recently identified essential hESCs pluripotent gene FOXO1A were also inhibited in response to BRE-silencing [49,50,51,52]. Expressions of epigenetic and chromatin modifier genes were also affected -with HDAC9 expressions enhanced following BRE-silencing. Histone deactylase genes perform diverse stem cell function [53]. In agreement with our observation where BRE-silencing promoted osteogenenic differentiation, it has been shown that HDAC9 played a crucial role in enhancing hMSCs osteogenesis [54]. These results indicate that BRE-silencing participates in modulating the genes involved in maintaining stemness and chromatin modification.
We also observed expression of genes encoding for components of the TGF-b/BMP, and FGF signaling pathways repressed in response to BRE-silencing. These pathways work to promote selfrenewal and suppress differentiation [34,35,36,37]. SMAD6, SMAD8, ID1, ID2, and ID3, all of them were negatively regulated by BRE-silencing, have been shown to inhibit osteogenic differentiation in MSCs [55,56]. FGF2 has also been identified as one of the self-renewal factors which can sustain the growth of hESCs and hMSCs on a feeder-independent and serum-free condition [57,58]. Moreover, FGF9 has been implicated in promoting survival of germ cells in fetal testis and it is a highly specific ligand for FGF receptor 2 and 3 which maintain pluripotency in mouse germ cells [59,60,61]. It has been reported that FGF signaling through FGFR2 inhibits mouse osteoblast differentiation [62]. A few studies have indicated that FGF2 could also inhibit multi-lineage MSC differentiations, while other studies have suggested FGF2 was a positive regulator of MSCs differentiation [63,64,65,66]. The contradictory effects may be dependent on the developmental status of the stem cells as it was reported that FGF2 stimulates growth of immature osteoblast progenitors but induces osteogenic differentiation of more mature precursors [67]. Our study suggests that FGF signaling inhibits HUCPV cells differentiation. BRE-silencing negatively regulated genes such as FGF2, FGF9, FGFR2 and FGFR3; thereby permitting HUCPV cells to differentiate into osteoblasts more rapidly when induced.
HOX transcription factors are important regulators of cell proliferation, self-renewal and differentiation in hematopoietic stem cells (HSCs) [41,43]. Overexpression studies have reported that HOXB4 could enhance mHSCs expansion and mESCs differentiation [68,69]. Furthermore, overexpression of HOXB4 up-regulates the transcriptional expression of BRE in mHSCs [68]. In our microarray study, however, we did not observe a significant change of HOXB4 expression in response to BRE-silencing. This observation implied that HOXB4 regulates the expression of BRE but not vice versa. Therefore, it will be interesting to establish whether HOXB4 overexpression can up-regulate BRE expression in mesenchymal stem cells or whether the protein regulation is lineage specific.
BRE-silencing resulted in the alteration of a number of chemokines. Amongst the chemokines that were up-regulated includes, TNFSF15, TNFRSF9, GDF5 and IL7. These chemokines have been implicated in their ability to promote osteogenesis of various types of progenitor cells [70,71,72,73]. Other cytokines that have been enhanced in response to BRE-silencing were TNFRSF9, IL7, IL15RA and IL21R which are up-regulated as hESCs differentiate into embryoid bodies [74]. Furthermore, CXCL3, CXCL10 and TNFSF9, which were positively regulated by BRE-silencing, have been classified as inhibitory factors in a study of intercellular network which modulated hematopoietic stem cell fate non-autonomously [75]. Therefore, BRE-silencing modulates the expression of cytokines which provided a favorable environment for differentiation to progress.
We also employed comparative proteomics approach to elucidate all the proteins that are differentially expressed after BRE-silencing in HUCPV cells which provided valuable insight on BRE targeted post-translational proteins to complement the microarray study. Functional annotation clustering analysis revealed that BRE function was associated with the organization of the cytoskeleton. BRE-silencing resulted in decreased presence of actin, ANXA2 and other cytoskeletal binding proteins. We also demonstrated that actin and ANXA2 could be immunoprecipitated with BRE. ANXA2 has been shown to bind to F-actin and plays a critical role in Rho/ROCK signaling pathway [76,77]. ANXA2 along with other actin-binding proteins are capable of regulating actin dynamics, organization and turnover in migratory cells [76,78,79,80,81,82]. Actins, together with intermediate filaments and microtubules, play an important role in cytoskeletal organization and response to external mechanical stimuli that affect cell differentiation [83,84,85]. During differentiation, the cell morphology is almost always altered in respond to integrins, cadherins and cytoskeletal proteins being differentially expressed [86]. When MSCs undergo osteogenesis, the cells become flatten and spread out; while the cells acquire a spherical shape when they undergo chondrogenesis [87,88]. It has been demonstrated when actin is disrupted by cytochalasin B (an inhibitor of actin polymerization), chondrogenesis is greatly enhanced in ESCs and MSCs. In contrast, chondrogenesis is hindered under conditions that favour actin assembly [89,90,91,92]. In our study, we demonstrated that the structural integrity of F-actin became dissembled in BRE-silenced HUCPV cells while F-actin structures were very distinctive in cells transfected with Ctl-siRNAs. The reduction of actin protein in BRE-silenced cells might provide a favourable environment for differentiation to take place, which agrees with our experimental observation that BRE-silencing accelerated chondrogenic differentiation of HUCPV cells.
Because actin and ANXA2 play a role in cell migration, we investigated the effect of BRE-silencing on cell movement. The HUCPV cells showed reduced ability to migrate after BREsilencing. The reduction is probably associated with the downregulation of actin protein which affected the cytoskeleton to change dynamically during movement -since migration requires a synchronized interactions of actin and its interacting partners such as actin-binding proteins and integrins [93]. ANXA2 has also been implicated in the regulation of HSC binding to osteoblasts and homing to the bone marrow niche [94]. Further studies will be required to establish whether BRE has a role in cell homing.
BRE-silencing could also suppress CALR, FUBP1 and PKM2 expression. In mESCs, down-regulation of CALR resulted in reduction of calcium ion level to enhance adipogenic differentiation [95]. It has also been reported that CALR may be involved in the regulation of osteogenesis and chondrogenesis in MSCs. CALR expression is reduced in the early stage of osteoblast differentiation in MC-3T3-E1 cells. Overexpression of CALR inhibits both the basal and vitamin D-induced expression of osteocalcin and calcium ion accumulation in the extracellular matrix and mineralization of bone nodules in cultures [96]. These findings may explain our observation of why BRE-silencing resulted in enhanced osteogenic and chondrogenic differentiation in HUCPV cells. BRE-silencing also negatively regulates FUBP1 and PKM2. FUBP1 is present in undifferentiated cells but not in differentiated cells and regulates c-myc expression by binding to a single-stranded far-upstream element upstream of the c-myc promoter [97]. Previous report showed that down-regulation of FUBP1 and subsequent down-regulation of c-myc were needed for lung cell differentiation in mice [98]. This protein may act both as an activator and repressor of transcription. Likewise, PKM2 has been determined to stimulate Oct4-mediated transcriptional activation [99]. Oct4 and c-myc constitute are two of the Yamanaka transcription factors crucial for maintaining pluripotency of stem cells [31,100]. BRE-silencing resulted in decreased expression of pluripotency factors and may explain our observation of enhanced osteogenic and chondrogenic differentiation in HUCPV cells. Further studies are required to examine the effect of BRE-silencing on other mesenchymal lineage differentiations.
In sum, our findings suggest a multifunctional role for BRE in maintaining stemness and the cytoskeletal architecture of HUCPV cells. In addition, BRE expression can be manipulated to accelerate induced chondrogenic and osteogenic differentiation in HUCPV cells ( Figure 13). Figure S1 Microarray analyses of differentially expressed stemness-related genes in response to BREsilencing. (A) Heat map displaying the patterns of differentially expressed stemness genes in response to BRE-silencing. The red boxes indicate that gene signals that are higher than the background signal (grey boxes), whereas blue boxes indicate signals that are lower than background. (B) RT-qPCR was performed to confirm the microarray data, specifically genes associated with stemness. Gene expression was normalized to GAPDH. The statistics of P values were determined by t-test; *p,0.05, **p,0.01, ***p,0.001 were considered significantly different and #contrast with microarray results. (TIF) Figure S2 Microarray analyses of differentially expressed epigenetic-related genes in response to BRE-silencing. (A) Heat map displaying the patterns of differentially expressed epigenetic genes in response to BREsilencing. (B) RT-qPCR was performed to confirm the microarray data, specifically genes associated with epigenetics. Gene expression was normalized to GAPDH. The statistics of P values were determined by t-test; *p,0.05, **p,0.01, ***p,0.001 were considered significantly different and #contrast with microarray results. N = 3 independent experiments. (TIF) Figure S3 Microarray analyses of differentially expressed genes associated with TGF-b signalling in response to BRE-silencing. (A) Heat map displaying the patterns of differentially expressed genes associated with TGF-b signalling in response to BRE-silencing. (B) RT-qPCR was performed to confirm the microarray data, specifically genes associated with TGF-b signalling. Gene expression was normalized to GAPDH. The statistics of P values were determined by ttest; *p,0.05, **p,0.01, ***p,0.001 were considered significantly different. N = 3 independent experiments. (C) Illustrating the cascade associated with the TGF-b signalling pathway (adapted from the KEGG database). Genes within the green boxes are significantly affected by BRE silencing while genes within the dark boxes are considered insignificantly affected. (TIF) Figure S4 Microarray analyses of differentially expressed genes associated with FGF signalling in response to BRE-silencing. (A) Heat map displaying the patterns of differentially expressed genes associated with FGF signalling in response to BRE-silencing. (B) RT-qPCR confirming the microarray data. Gene expression was normalized to GAPDH. The statistics of P values were determined by t-test; *p,0.05, **p,0.01, ***p,0.001 were considered significantly different.   Table S2 The top 20 proteins identified by comparing BRE-silenced and control human umbilical cord perivascular cells, grouped based on molecular function using Gene Ontology tools. (JPG)