Integrated analysis of DNA methylome and transcriptome reveals the differences in biological characteristics of porcine mesenchymal stem cells from bone marrow and umbilical cord

Background Bone marrow (BM) and umbilical cord (UC) are the main sources of mesenchymal stem cells (MSCs). These two MSCs display signicant differences in many biological characteristics, yet the underlying molecular mechanisms need to be explored. Results In this study, to better understanding the biological features of MSCs, we isolated BMMSCs and UCMSCs from inbred Wuzhishan miniature pigs and generated the rst global DNA methylation and gene expression proles of porcine MSCs. The results showed that the osteogenic and adipogenic differentiation ability of porcine BMMSCs is stronger than that of UCMSCs. Stem cell surface marker CD90 were positively detected in both BMMSCs and UCMSCs. 587 genes were differentially methylated (280 hypermethylated and 307 hypomethylated) at the promoter regions between BMMSCs and UCMSCs. Meanwhile, 1,979 differentially expressed genes (1,407 up-regulated and 572 down-regulated) were identied between BMMSCs and UCMSCs. Integrative analysis reveals that 120 genes displayed differences in both gene expression and promoter methylation. Gene Ontology enrichment analysis revealed that these differential genes were associated with cell differentiation, cell migration, and immunogenicity properties. Remarkably, skeletal system development related genes were signicantly hypomethylated and up-regulated in UCMSCs, while cell cycle genes were signicantly higher down-regulated and hypermethylated, implying UCMSCs have higher cell proliferative activity and lower osteogenic differentiation potential than BMMSCs. Conclusions Our results indicate that DNA methylation plays an important role in regulating the biological characteristics differences between BMMSCs and UCMSCs. The study might provide a molecular theory basis for the application of porcine MSCs in human. Cells 0.05 % 3 xation. MSCs were resuspended in 1% (w/v) bovine serum albumin (Sigma) for 30 min in room temperature to block the non-specic binding sites. After blocking, the BMMSCs were incubated with CD29 (VMRD), CD44 (VMRD), CD45 (VMRD) and FITC-anti-human CD34/PE-anti human CD90 (eBioscience) monoclonal antibodies at room temperature for 20 min. The UCMSCs were incubated with CD31, CD45 (Veterinary Medical Research & Development, VMRD) and FITC-anti-human CD34/PE-anti human CD90 (eBioscience) monoclonal antibodies, respectively, at room temperature for 20 min. The CD29, CD44 and CD45 groups were then stained with rat anti-mouse IgG1-FITC (IVGN), goat anti-mouse IgG2a-PE secondary antibody (IVGN) and anti-mouse IgM-PE (eBioscience) at room temperature for 20 min, respectively. Flow cytometric acquisition and data analysis were performed with BD FACS Calibur Flow Cytometer and Cell Quest software. As a negative control, the cells were incubated only with DPBS. Each ow cytometry experiment was performed in triplicate.

reveals that 120 genes displayed differences in both gene expression and promoter methylation. Gene Ontology enrichment analysis revealed that these differential genes were associated with cell differentiation, cell migration, and immunogenicity properties. Remarkably, skeletal system development related genes were signi cantly hypomethylated and up-regulated in UCMSCs, while cell cycle genes were signi cantly higher down-regulated and hypermethylated, implying UCMSCs have higher cell proliferative activity and lower osteogenic differentiation potential than BMMSCs. Conclusions Our results indicate that DNA methylation plays an important role in regulating the biological characteristics differences between BMMSCs and UCMSCs. The study might provide a molecular theory basis for the application of porcine MSCs in human.

Background
Mesenchymal stem cells (MSCs), also known as seed cells, are widely used in tissue repair and regeneration because of their self-renewal and differentiation capacity together important immunosuppressive properties and low immunogenicity (Kolf et al., 2007;Shi et al., 2010;Li and Hua, 2017). MSCs were originally isolated from bone marrow (BM). However, the use of BMMSCs is not always acceptable due to the highly invasive donation procedure and the signi cant decline in cell number and proliferative/differentiation capacity with age (Romanov et al., 2003). In recent years, studies have discovered other MSC sources from almost every tissue of the body, such as adult adipose (AT), placenta and amniotic uid (Semenov et al., 2010;Ivana et al., 2011;Urrutia et al., 2019). Additionally, umbilical cord MSCs (UCMSCs) have been introduced as an promising source of MSCs and applied in preliminary clinical treatments, because they can be easily obtained, display less negative effects on the donor and avoid certain ethical questions (Lindenmair et al., 2012;Chandravanshi and Bhonde, 2018).
Though MSCs derived from different sources share many similar biological characteristics, they also exhibit distinct and unique gene expression and functional properties (Si et al., 2011;Cho et al., 2017).
The miniature pig (Sus scrofa) is an attractive and realistic large animal model, because of the anatomical, physiological and genomic similarities to human (Vodička et al., 2010;Wang et al., 2010).
Over 25 years, the inbred Wuzhishan miniature pig was developed by the Institute of Animal Sciences, Chinese Academy of Agricultural Sciences. The inbred WZSP line has high genetic stability (Fang et al., 2012) and the inbreeding coe cient at the twenty-fourth generations reached 0.994 in 2013 (Mu et al., 2015), which has been widely used for studying human diseases, including atherosclerosis, cardiovascular disease, xenotransplantation, and diabetes (Dong et al., 2014;Zhao et al., 2018). Due to the source and availability of a large quality of human MSCs is limited, the therapeutic potentialities of MSCs derived from animal sources have acquired wide attention (Bai et al., 2012;Lu et al., 2014;Khatri and Richardson, 2017). Porcine MSCs can be easily obtained and share similar morphology and multilineage differentiation potential to those of human MSCs (Groth et al., 2012). MSCs derived from inbred WZSP are highly stable and conducive to establish a reliable system to evaluate the biological characteristics of porcine MSCs.
DNA methylation is a stable epigenetic modi cation that regulates many biological processes, including genomic imprinting, X-inactivation, genome stability and gene regulation (Ambrosi et al., 2017). However, there is limited information about the DNA methylation and gene expression regulation of porcine MSCs.
In this study, to reveal the molecular mechanism of the biological characteristic differences of MSCs, we isolated BMMSCs and UCMSCs from inbred WZSP. The genome-wide DNA methylome and transcriptome maps of BMMSCs and UCMSCs were generated by methylated DNA immunoprecipitation sequencing (MeDIP-Seq) and RNA sequencing (RNA-seq), respectively. We identi ed a set of genes displaying expression and methylation differences between these two MSCs, which are critical for regulating the biological functions of porcine MSCs. This study provides a molecular theory basis of the application of porcine MSCs in clinical therapy.

Methods
Isolation and culture of porcine mesenchymal stem cells The WZSP littermates were purchased from the National Germplasm Resources Center of Laboratory Miniature Pig, Beijing, China. All animal procedures were approved by the Animal Care and Use Committee of the Institute of Animal Sciences, Chinese Academy of Agricultural Sciences. The pigs were injected intravenously with propofol (2 mg/kg) to induce full anesthesia. UCMSCs were isolated from the umbilical cord of four WZSP littermates on the day of birth, and BMMSCs were isolated from the bone marrow of the same individuals on 42 days after birth. To isolate BMMSCs, the bone marrow was extracted and centrifuged for 5 min at 1000 rpm. To isolate UCMSCs, the umbilical cords were cut into 1-2 mm 2 pieces, attached and cultured. The isolated MSCs were cultured in DMEM/F12 medium (Gibco) with 20 % fetal bovine serum (Gibco), 50 units/ml penicillin G, and 50μg/ml streptomycin and incubated at 37°C in 5% CO 2 in a humidi ed incubator; the medium was replaced every 3 days.
Flow cytometric analysis of cell surface antigen expression Flow cytometry was used to analyze the surface marker phenotype of the MSCs, similar to our previous report. Cells were harvested by 0.05 % trypsin-EDTA for 3 minutes at 37 °C, followed by washing and xation. MSCs were resuspended in 1% (w/v) bovine serum albumin (Sigma) for 30 min in room temperature to block the non-speci c binding sites. After blocking, the BMMSCs were incubated with CD29 (VMRD), CD44 (VMRD), CD45 (VMRD) and FITC-anti-human CD34/PE-anti human CD90 (eBioscience) monoclonal antibodies at room temperature for 20 min. The UCMSCs were incubated with CD31, CD45 (Veterinary Medical Research & Development, VMRD) and FITC-anti-human CD34/PE-anti human CD90 (eBioscience) monoclonal antibodies, respectively, at room temperature for 20 min. The CD29, CD44 and CD45 groups were then stained with rat anti-mouse IgG1-FITC (IVGN), goat anti-mouse IgG2a-PE secondary antibody (IVGN) and anti-mouse IgM-PE (eBioscience) at room temperature for 20 min, respectively. Flow cytometric acquisition and data analysis were performed with BD FACS Calibur Flow Cytometer and Cell Quest software. As a negative control, the cells were incubated only with DPBS. Each ow cytometry experiment was performed in triplicate.

Adipogenic and osteogenic differentiation of porcine BMMSCs and UCMSCs
To evaluate the differentiation ability of MSCs in vitro, we replaced the DMEM/F12 medium with an adipogenic/osteogenic differentiating medium when cells reached 80 % con uency. Cells were cultured for 2 or 3 weeks before collection and the medium were changed every 3 days. After 2 or 3 weeks, Oil red O was used in the analysis of adipogenic differentiation and alizarin red staining was used in the analysis of osteogenic differentiation.

Methylated DNA immunoprecipitation sequencing
Genomic DNA was isolated using E.Z.N.A. HP Tissue DNA Midi Kit (Omega) and was sonicated to 100-500 bp fragments with a Bioruptor Sonicator (Diagenode). Four BMMSCs DNA samples were pooled and four UCMSCs DNA samples were pooled by homogeneous mixing prior to methylated DNA immunoprecipitation sequencing. The libraries were constructed following the manufacturer's instructions, same to our previous reports (Yang et al., 2016;Yang et al., 2017), and sequenced on an Illumina HiSeq 2000 with 49 bp paired-end reads.

MeDIP-seq data analysis
After ltering the low-quality reads that contained more than 5 'N's or over 50 % of the sequence with low quality value (Phred score < 5), the clean reads were aligned to the pig reference genome downloaded from the USCS database, allowing up to two mismatches using SOAP2 (v2.21) (Li et al., 2009). Reads mapping to the same genomic location were regarded as potential clonal duplicates due to PCR ampli cation biases. To avoid stochastic sampling drift, we ltered out CpG sites that were covered by less than a 10 read depth . Model-based Analysis of ChIP-Seq (MACS v1.4.2) (http://liulab.dfci.harvard.edu/MACS/) was used to scan the methylated peaks in pig genome with default parameters (Zhang et al., 2008). The methylation level of each peak was calculated using the RPMK method. Differentially methylated regions (DMRs) were identi ed using the exact test for negative binomial distribution with a signi cance threshold of FDR < 0.001 and |log2 FC| ≥ 1. We de ned the region 2 kb upstream of TSS as promoter and the region from the TSS to TTS as gene body. The promoters that contained one or more DMRs were considered as differentially methylated promoters for further analysis.
Transcriptome sequencing and data analysis RNA from BMMSCs and UCMSCs was isolated using the Trizol reagent (Invitrogen, Carlsbad, CA, USA), treated with DNase I (Qiagen) and cleaned using the RNAeasy MiniElute Cleanup kit (Qiagen, Basel, Switzerland). The integrity of total RNA was checked by an Agilent 2100 Bioanalyze (Agilent Technologies, Palo Alto, CA, USA), and only RNA samples with a RNA Integrity Number (RIN) score > 8 were used for RNA sequencing. Equal amounts of RNA from four samples of BMMSCs and UCMSCs were combined into a pool, respectively, Beads with oligo (dT) were used to isolate poly (A) mRNA after total RNA was collected. Fragmentation buffer was added to fragment the mRNA. Taking these short fragments as templates, a random hexamerprimer was used to synthesize the rst-strand cDNA. The second-strand cDNA was synthesized using buffer, dNTPs, RNaseH and DNA polymerase I, respectively. Short fragments were puri ed with QiaQuick PCR extraction kit and resolved with EB buffer for end reparation and adding poly (A). The short fragments were then connected with sequencing adaptors. And, for ampli cation with PCR, we selected suitable fragments, as templates, with respect to the result of agarose gel electrophoresis. The libraries were sequenced using Illumina HiSeq 2000 to generate 90 bp paired-end reads.
After trimming the adaptor sequences and removing low-quality reads, clean reads were mapped to Sus scrofa reference genome using SOAP2 (v2.21) allowing up to three mismatches (Li et al., 2009). RPKM value was used to measure the expression level of each gene. Differentially expressed genes between BMMSCs and UCMSCs were identi ed using the exact test for negative binomial distribution. Genes with FDR < 0.001 and |log2 FC| ≥ 1 were considered as differentially expressed.

GO enrichment analysis
Functional enrichment analysis was performed using the DAVID (Database for Annotation, Visualization and Integrated Discovery) web server (http://david.abcc.ncifcrf.gov/) (Da et al., 2009). Genes with DMRs in promoters were mapped to their human orthologs and were submitted to DAVID for GO enrichment analysis.

RT-qPCR
Total RNA was extracted using the RNA Extraction Kit (BioTeke). First Strand cDNA was synthesized using the oligo (dT)18 primer provided in the RevertAid First Strand cDNA synthesis kit (Thermo). q-PCR was performed on an ABI 7500 machine using the SYBR Premix Ex Taq kit (TaKaRa) and the glyceraldehyde-3-phosphate dehydrogenase gene (GAPDH) was used as endogenous control gene.
Relative expression levels of objective mRNAs were calculated using the 2 -ΔΔCt method. Primer sequences are shown in Additional File 1: Table S4.
The DNA isolated from UCMSCs and BMMSCs were treated with sodium bisul te using an EZ DNA Methylation-Gold Kit (ZYMO Research) according to the manufacturer's instructions. Quantitative methylation analysis of the DMRs was performed using the Sequenom MassARRAY platform (CapitalBio, Beijing, China) (Mathias et al., 2005). Speci c primers were designed using the EpiDesigner software (Sequenom). The quantitative results for each CpG or multiple CpGs were analyzed with EpiTyper software v1.0 (Sequenom). The primer sequences are shown in Additional File 1: Table S4.

Isolation and identi cation of porcine BMMSCs and UCMSCs
We isolated the BMMSCs and UCMSCs from the inbred WZSPs. Adhesion of BMMSCs to the plastic ask was observed at 24h after isolation. As the process continued, adherent cells displayed a scatter distribution and grew in isolated clones. UCMSCs gradually grew outward from the umbilical cord tissues after 7 days. The cell morphology of UCMSCs was similar to BMMSCs: The majority of the cells were fusiform and their nucleoli were clear. The passage cells reached 90 % con uency after approximately 3 days ( Figure 1A).
Flow cytometry (FCM) analysis was performed to con rm the surface marker characteristics of MSCs. In BMMSCs and UCMSCs. the stem cell surface markers (CD29, CD44 and CD90) were positively detected, whereas leucocyte marker CD45 and hematopoietic lineage marker CD34 were negative ( Figure 1B). The UCMSCs were positive for CD90, but negative for CD34, CD45 and endothelial marker CD31 ( Figure 1C). The in vitro differentiation potential of BMMSCs and UCMSCs for osteogenic and adipogenic lineages were further detected. We observed that the calci ed nodules on the cell surface of MSCs increased with the induction of osteoblast differentiation. On the 21 st day of osteogenic induction, the morphology of the cells signi cantly changed by displaying substantial accumulation of orange sediment ( Figure 1D), the calci ed nodules of BMMSCs were much more obvious than UCMSCs. On the 21 st day after adipogenic differentiation induction, numerous intracellular lipid droplets were formed ( Figure 1E), and the lipid droplets of BMMSCs were much more obvious than UCMSCs. The results showed that both MSCs we obtained have the potentials for osteogenic and adipogenic differentiation, while the differentiation ability of BMMSCs is stronger than that of UCMSCs.
DNA methylomes and transcriptome pro ling of porcine BMMSCs and UCMSCs We carried out MeDIP-seq and RNA-seq to pro le the genome-wide DNA methylome and transcriptome of were used for the following analyses.

Methylome characteristics of the porcine BMMSCs and UCMSCs
We rst analyzed the genome-wide DNA methylation pattern of porcine MSCs ( Figure 2) and found that the methylation level was negatively correlated with repeat length (Pearson's r = -0.248, p < 0.001), and positively correlated with gene number (Pearson's r = 0.335, p < 0.001), CpG islands (CGIs) length (Pearson's r = 0.482, p < 0.001), CpG sites number (Pearson's r = 0.777, p < 0.001) and especially with observed over expected number of CpG (CpGo/e) ratio (Pearson's r = 0.790, p < 0.001). We further analyzed the methylation of the upstream 2 kb regions of the transcription start sites (TSS), the gene body and the downstream 2 kb regions of the transcription termination site (TTS) for MSCs (Figure 3). The TSS of both MSCs displayed low methylation, while the DNA methylation level of gene body regions was relatively constant, which was much higher than the 5' and 3' anking regions. These results were consistent with previous reports (Yang et al., 2016;Yang et al., 2017).

Promoter methylation and transcriptional repression in MSCs
Methylated peaks were detected across different genomic elements. Reads per Kilo bases per million reads (RPKM) value was used to evaluate the methylation level of each peak. A total of 150,690 and 161,105 methylated peaks were generated with average lengths of 1,462 and 1,466 bp in BMMSCs and UCMSCs, respectively, covering 9.74 % and 10.44 % of the Sus scrofa genome. We classi ed genes into four groups according their methylation modi cation: (I) only promoter was modi ed; (II) only genebody was modi ed; (III) both were modi ed; (IV) neither promoter nor genebody were modi ed. The number of genes classi ed into these four methylation types in BMMSCs was 1,134, 8,424, 2,213 and 8,656, respectively ( Figure 4A); and the number in UCMSCs was 1,187, 8,106, 2,520 and 8,614, respectively ( Figure 4B). The expression level of genes in group IV was signi cantly higher than that of genes in other three groups, while the genes in group I exhibited the lowest expression levels ( Figure 4C). These results implied that both promoter and genebody methylation could affect gene expression. We analyzed the effects of the promoter CpG islands (CGI) on gene expression and found that the expression level of genes without promoter CGI was signi cantly lower than that of genes with promoter CGI ( Figure 4D). Meanwhile, we found genes with low methylation modi cation at promoter CGI had signi cantly higher expression levels than genes with high methylation modi cation at promoter CGI ( Figure 4E), implying the methylation of CpGI also regulated the gene expression in MSCs.

Differential genes between BMMSCs and UCMSCs
We next compared the DNA methylation and gene expression differences between porcine BMMSCs and UCMSCs. 587 genes showing differentially methylated at the promoter region were identi ed, 280 genes were hypermethylated and 307 genes were hypomethylated in UCMSCs (Additional File 1: Table S1). Gene Ontology (GO) enrichment analysis revealed that the hypermethylated genes were signi cantly associated with skeletal system development, pattern speci cation process, and chordate embryonic development ( Figure 5A). While the hypomethylated genes were signi cantly enriched in regulation of amine transport, regulation of catecholamine secretion, regulation of system process, and G-protein signaling, coupled to cyclic nucleotide second messenger ( Figure 5B).
We also identi ed 1,979 differentially expressed genes (DEGs) between BMMSCs and UCMSCs (Additional File 1: Table S2). Compared with BMMSCs, 1,407 genes were up-regulated while 572 genes were down-regulated in UCMSCs. GO enrichment analysis revealed that the up-regulated genes were signi cant enriched in nuclear division, mitosis, organelle ssion and cell cycle ( Figure 5C), implying that UCMSCs have higher cell proliferative activity than BMMSCs. The down-regulated genes were signi cant enriched in skeletal system development, translational elongation cell migration, cell adhesion, ossi cation, and metabolic related processes ( Figure 5D). These differential genes further revealed the characteristics of MSCs that depended on different cellular sources.
We found 102 genes that have both expression and promoter methylation differences. 36 of these genes were hypermethylated and down-regulated in BMMSCs, including C8ORF73, AOC3, FGF21, AC005841.

Validation of the MeDIP-seq and RNA-seq data
The methylate degree of 31 differentially methylated regions in the promoter of 15 genes was veri ed by Sequenom MassArray methylation analysis ( Figure 6 and Additional File 1: Table S3), the expression level of 12 DEGs was validated by real-time quantitative PCR (RT-qPCR, Figure 6 and Additional File 2: Figure S1). These results were in accordance with the MeDIP-seq and RNA-seq results, con rming the reliability of our omic data.

Discussion
The biological characteristics of mesenchymal stem cells (MSCs) derived from different sources are different in proliferation, differentiation and migration abilities that affect the their tissue repair capacity (Kolf et al., 2007;Shi et al., 2010;Li and Hua, 2017). Porcine MSCs can be easily obtained and share similar morphology and differentiation potential with human MSCs. The inbred WZSP line is an ideal large animal model with high genetic stability (Fang et al., 2012), which provides excellent materials for understanding the molecular characteristics of MSCs. To explore the biological characteristics and regulation mechanism of MSCs derived from different sources, we isolated BMMSCs and UCMSCs from WZSP, and pro led the genome-wide DNA methylome and transcriptome maps of these two MSCs.
Our results indicated that the porcine MSCs had similar DNA methylation patterns with other pig tissues Yang et al., 2016;Yang et al., 2017): TSS maintained a low methylation status and gene body exhibited a much higher level of DNA methylation than the 5' and 3' anking regions. Genome-wide integrated maps of DNA methylation and transcriptome of porcine MSCs showed that the expression of genes was affected by both promoter and genebody methylation, con rming that promoter methylation repress gene expression (Jones, 2012;Smith et al., 2014). Most CpGs in the mammalian genomes are methylated, whereas CpGs within CGIs are usually unmethylated. While methylated CGIs are also observed during normal biological processes, such as X chromosome inactivation and gene imprinting (Cottrell, 2004) . In this study, we found that the expression level of gene without promoter CGI was signi cantly lower than that of gene with promoter CGI. Additionally, methylate level of the promoter CGI had a negative correlation with gene expression level. These results indicated the methylation of CpGI might regulate the gene expression in MSCs, but their regulation mechanism still need to be further explored.
MSCs derived from different sources also manifested unique molecular characteristics. We identi ed 587 genes displaying promoter methylation differences and 1,979 genes displaying expression differences were compared between BMMSCs and UCMSCs. 102 genes had both expression and promoter methylation differences. Enrichment analysis revealed these differential genes were functionally related to the biological characteristics of MSCs. Skeletal system development was the most signi cantly enriched biological process for both the hypermethylated genes (such as Homeobox genes) and the down-regulated genes (such as PTN, RBP4) in UCMSCs. Homeobox genes are master developmental control genes that act at the top of genetic hierarchies regulating aspects of morphogenesis and cell differentiation in animals (Mark et al., 1997). Pleiotrophin (PTN) gene has a higher expression level and lower promoter methylation degree in BMMSCs, this gene plays an important role in bone formation by mediating the recruitment and attachment of osteoblasts/osteoblast precursors to the appropriate substrates for the deposition of new bone (Erlandsen et al., 2012). These results indicated that BMMSCs have much higher osteogenic differentiation potential than UCMSCs. Previous study also showed that osteoblast differentiation of UCMSCs was less e cient, even after addition of 1.25-dihydroxyvitamin D3, a potent osteoinductive substance (Majore, 2011).
Compared with UCMSCs, Inter-alpha (globulin) inhibitor H5 (ITIH5) gene had a higher expression level and lower promoter methylation degree in BMMSCs. ITIH5 was highly expressed in human adipocytes and adipose tissue, its expression was higher in obese subjects and reduced during diet-induced weight loss (Anveden et al., 2012). Fibroblast growth factor 21 (FGF21), an endocrine regulator in lipid metabolism, caused a dramatic decline in fasting plasma glucose, fructosamine, triglycerides, insulin, and glucagon when administered daily for 6 weeks to diabetic rhesus monkeys (Murata et al., 2011) (Alexei et al., 2007. Compared with BMMSCs, ITIH5 and FGF21 had a higher expression level and lower promoter methylation degree in UCMSCs. The results illustrated that the adipogenic differentiation ability of BMMSCs was stronger than UCMSCs. Meanwhile, we observed that cell cycle genes were signi cantly up-regulated and hypomethyalted in UCMSCs, such as CTF1, DAB2IP and CACNA1G. Cardiotrophin 1 (CTF1) can stimulate the proliferation of cardiomyocytes (Stejskal and Ruzicka, 2008) and plays an important role in cardiac repair in the infarcted heart (Freed et al., 2005). DAB2 interacting protein (DAB2IP) is a newly described member of the Ras GTPase-activating protein family and plays an important role in maintaining cell homeostasis and regulating cell proliferation, survival, and death (Daxing et al., 2009). Calcium channel, voltage-dependent, T type, alpha 1G subunit (CACNA1G) is a T-type calcium channel gene, and the hypermethylation of CACNA1G has been shown in various human tumors, which potentially affects cell proliferation and apoptosis (Toyota et al., 1999). These results implied that UCMSCs have higher cell proliferative activity than BMMSCs.
The extent of tight junction formation is one of many factors that regulate motility, invasion, and metastasis. As a member of Claudins, claudin 4 (CLDN4) is required for the formation and maintenance of tight junction (Lin et al., 2013). The forkhead box L1 (FOXL1) protein belongs to the forkhead box (Fox) family of transcription factors. Over-expression of FOXL1 inhibits tumor cell growth, migration and invasion of renal and pancreatic cancer cells (Geng et al., 2013;Feng-Qiang et al., 2014). Compared with BMMSCs, both CLDN4 and FOXL1 displayed a higher expression level and lower promoter methylation degree in UCMSCs, demonstrating that these differential genes regulated the migration ability of procine MSCs.
Additionally, G protein-coupled receptor 44 (GPR44) plays a major role in the activation and chemotaxis of Th2 cells, eosinophils, and basophils (Ishii et al., 2012). G-protein signaling modulator-3 (GPSM3) is known to bind to Gαi·GDP subunits and free Gβ subunits during their biosynthetic path toward Gγ dimer formation. GPSM3 is an important regulator of monocyte function involved in the regulation of differentiation, chemotaxis, and survival in vitro and in vivo, its de ciency is protective in acute in ammatory arthritis (Giguere et al., 2013). UL16 binding protein 3 (ULBP3), as a MHC class I-related molecule, can bind to human cytomegalovirus glycoprotein UL16 and activate natural killer cells (Kubin et al., 2001). The lower expression and higher methylation of GPR44, GPSM3 and ULBP3 in UCMSCs compared with BMMSCs suggested that the two MSCs have different immunogenicity properties.

Declarations
Ethics approval and consent to participate All animal procedures were approved by the Animal Care and Use Committee of the Institute of Animal Sciences, Chinese Academy of Agricultural Sciences.

Consent for publication
Not applicable Availability of data and material The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.   promoter CGI and genes without promoter CGI. (E) Gene expression comparison between genes with different methylation level at promoter CGI.   Figure S1 and Additional File 1: Table S3.