Identification of peroxiredoxin II and its related molecules as potential biomarkers of dermal mesenchymal stem cell homing using network analysis

Abstract

In this study, we performed RNA sequencing of Prx II^+/+ and Prx II^−/− dermal mesenchymal stem cells (DMSCs) to identify differentially expressed genes (DEGs). To explore the role of Prx II in DMSCs, we performed Gene Ontology analysis of the DEGs. The results showed that the DEGs were mainly involved in the biological processes of cell migration, intercellular adhesion, and coordination of the regulation of stem cell homing. Through the construction of protein–protein interaction network, four hub genes Cd274, Ccl5, Il1b, and Stat1 involved in cell adhesion and cell homing were screened. Quantitative reverse transcription PCR analysis showed that Cd274, Ccl5, Il1b, and Stat1 were down regulated in Prx II^−/− DMSCs. miRwalk and Starbase databases were further used to screen the upstream molecules miRNA and lncRNA regulating hub gene. Prx II was found to be involved in the regulation of stem cell homing via the Tctn2/miR-351/Stat1/Il1b axis. Thus, we demonstrated that Prx II is a key molecule in the regulation of the homing ability of DMSCs. Our results provide a theoretical foundation for improving the homing ability of DMSCs by targeting Prx II.

Introduction

Mesenchymal stem cells (MSCs) are considered as a promising stem cell source for regenerative medicine [1]. Their strong differentiation ability and immune regulation activity play key roles in tissue repair. The therapeutic effect of MSCs mainly depends on their specific homing to the injured site. Several studies have shown that when the body experiences ischemia, hypoxia, or other injuries, endogenous or exogenous MSCs with limited efficiency exhibit homing to the inflamed area and injured tissue [2, 3]. After homing to the site of injury, MSCs exert a therapeutic role either by leading to tissue differentiation and injured cell replacement or by secreting various cytokines to promote angiogenesis and regulate the inflammatory response. Although MSCs have an inherent ability to migrate to the injured or inflamed site, owing to the limitations of the signal pathways that regulate cell movement, their homing efficiency is limited [4, 5]. In fact, most injected MSCs cannot reach the site of injury, making it difficult to achieve the expected clinical effect. Therefore, identifying methods that can improve the homing rate of MSCs and their clinical efficacy has become an important research focus.

MSCs can be mobilized and enter into the circulation. Alternatively, MSCs can be recruited from the circulation to their original location. Studies have shown that chemokines and their receptors control MSC homing mobilization and migration [6, 7]. The chemokine (C–C motif) ligand 5 (CCL5) can enhance the migration ability of adipose and dermal MSCs (DMSCs) [8]. In an Alzheimer’s disease mouse model, researchers detected an increase in CCL5 expression in bone marrow-derived MSCs in the mouse brain [9]. Furthermore, some researchers found that CX3CL1 expression was upregulated in injured brain tissue. CX3CL1 can bind to its receptor, CX3CR1, to induce MSC migration to the brain [10]. In addition, the chemokines CXCL12, CXCL13, CXCL16, CCL11, and CCL2 can promote two-way migration of bone marrow MSCs and human bone marrow endothelial cells [6]. Hence, chemokines and their receptors play important roles in MSC migration and homing.

After MSC mobilization, migration, and homing, MSCs start to adhere to capillary wall endothelial cells. In this process, cell adhesion molecules expressed by MSCs bind with adhesion molecule ligands present in the extracellular matrix. The main adhesion molecules expressed by MSCs and bind to endothelial cells are integrins and selectins. The integrin vascular cell adhesion molecule-1/very late antigen-4 (VLA-4) plays an important role in tight adhesion [11]. The adhesion between endothelial cells and human MSCs treated with VLA-4 antibody was significantly decreased in shear blood flow, suggesting that adhesion of MSCs depends on VLA-4 [12]. Thus, adhesion molecules are indispensable components of the MSC homing process.

Peroxiredoxins (Prx) belong to the peroxidase family of antioxidant proteins, which are widely present in the body. Peroxiredoxin II (Prx II) contains two highly conserved cysteine residues with redox activity, which can rapidly remove low concentrations of active oxygen in vivo [13]. Recent studies have shown that Prx II regulates various stem cells. Silencing of Prx II in colon cancer stem cells led to decreased Nanog expression and cell proliferation [14]. Furthermore, Prx II regulates embryonic stem cell differentiation into neural cells [15]. Studies have demonstrated that the expression of SOX2 and OCT4 was decreased in Prx II-silenced Huh7-H-RasG12V cells [16, 17]. Furthermore, Prx II was found to protect DMSCs from premature senescence [18,19,20]. These results indicate that Prx II regulates the proliferation, differentiation, and senescence of stem cells, and that it may be a key regulator of stem cell function.

In this study, we used RNA sequencing to analyze the differentially expressed genes (DEGs) between Prx II^+/+ and Prx II^−/− DMSCs derived from mice. Bioinformatics tools were used to perform Gene Ontology (GO) biological process enrichment analysis. The GO term enrichment analysis showed that the DEGs in Prx II^−/− DMSCs were mainly related to stem cell migration. Using the STRING database and Cytoscape software, we screened hub genes related to the Prx II-regulated homing of DMSCs, which revealed the homing mechanism of stem cells at the molecular level based on potential key micro-RNAs (miRNAs) and long non-coding RNAs (lncRNAs) included in the miRWalk and Starbase databases. Our study provides insights into stem cell treatment for various diseases.

Materials and methods

Ethics statement

This study was approved by the Animal Ethics Committee of Heilongjiang Bayi Agricultural University (Approval number: TDJH202116) and was strictly carried out under the guide for the care and use of laboratory animals.

Chemicals

Culture dishes (100 mm) were purchased from NEST Biotechnology (Wuxi, China), fetal bovine serum and penicillin/streptomycin were purchased from Solarbio (Beijing, China), and TRIzol was purchased from Sigma (St. Louis, MO, USA). We used the inNovaUscript II All in One First strand cDNA Synthesis SuperMix kit (Innovagene, Hunan, China), inNova Taq SYBR Green qPCR Mix (Innovagene), miRNA First Strand cDNA Synthesis (Tailing Reaction) (B532451-0010; Sangon Biotech, Shanghai, China), miRNA qPCR Kit (SYBR Green Method) (Sangon Biotech), and a CFX96 real-time PCR system (Bio-Rad, Hercules, CA, USA).

Isolation of DMSCs

DMSCs were isolated and characterized according to standard protocols [21], which our laboratory has also reported on previously [22, 23]. Prx II-rescued cells (referred to as WT) and Prx II gene knockout 129/SVJ mice were maintained in a pathogen-free facility (20–22 °C, humidity 50–60%, and 12-h-dark/light cycle). Briefly, neonatal mice were anesthetized with isoflurane (1.5–2% for induction anesthesia; 1–1.5% for maintenance of anesthesia). We isolated the skin from neonatal 129/SvJ mice and Prx II knockout 129/SvJ mice (aged 1–2 days) for 5 min. After removing the skin, the mice were placed in a euthanasia box that received CO₂ at a rate of 10–30% of the volume of the euthanasia box every minute. The mice were confirmed to be motionless, non-breathing, and to have dilated pupils for 30 s. After ceasing the CO₂ input, the mice were observed for another minute to confirm their death. The skin samples were digested in 0.25% trypsin-EDTA so that the dermal layer was isolated from the epidermal layer. The dermis was cut into approximately 1 cm² pieces and then digested with 0.25% trypsin-EDTA for about 1 h and seeded in DMEM/F12 medium.

Cell culture

The cells were cultured in standard DMEM/F12 medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at 37 °C in a 95% air/5% CO₂ incubator. The culture medium was refreshed every day. When cells reached 90% confluence, they were subcultured after treatment with a 0.25% trypsin-EDTA mixture.

RNA sequencing

mRNA was isolated from total RNA by separating the poly-A containing molecules using poly-T primers. RNA fragmentation, first- and second-strand cDNA syntheses, end repair, adaptor ligation, and PCR amplification were performed. The average size of the cDNA in the sequencing libraries was approximately 160 bp (excluding the adapters). The RNA integrity and quantity of the cDNA libraries were validated before sequencing. The cDNA libraries were clustered onto a TruSeq paired-end flow cell and sequenced using an Illumina HiSeq 2000 Sequencer, producing 100-bp paired-end reads (2 × 100). RNA sequencing was performed at Gminix.

Functional enrichment analysis

GO analysis is widely used in functional studies and gene set enrichment analysis. The DAVID gene annotation tool [24] was used to analyze GO term enrichment among the DEGs. Targeting graphs of salient functions in three categories, i.e., biological process, molecular function, and cellular component, were constructed to understand the biological functions of the DEGs. We selected P < 0.05 and FC > 1.2 or < 0.83333 to indicate that genes with a difference of > 1.2 times were DEGs. The biological process terms were analyzed using ClueGo in Cytoscape v3.8.2 to visualize DEG enrichment [25].

Intercellular adhesion and cell migration in GO, construction of a protein–protein interaction (PPI) network, and identification of hub genes

The STRING 10 database provides known and predicted protein interactions based on confidence. These interactions include direct (physical) and indirect (functional) associations originating from the genomic background, experiments, co-expression, and previous knowledge [26]. Genes involved in regulating cell migration and the cell adhesion module were considered as seed nodes and mapped to the STRING database to construct an extended PPI network with a medium confidence score of 0.4. All PPI networks were visualized using Cytoscape v3.8.2 software and a network analyzer tool to calculate the parameters of intermediate betweenness centrality (BC) and degree. Betweenness centrality reflects the number of shortest paths through a node, which is very important in node importance analysis. The degree indicates the number of specific protein interactions. In the final network, the height node is displayed in a large circle, whereas green and blue shadows indicate the BC value of the node from high to low. In this study, nodes with a BC threshold > 0.05 and a degree greater than the average were considered hub genes.

Prediction of pivotal miRNAs and construction of an mRNA–miRNA interaction network

The hub genes in key pathways were selected, and their targeted miRNAs were predicted using miRWalk v2.0 [27]. To verify the accuracy of the results, five databases were used, including TargetScan, miRanda, miRDB, miRWalk, and RNA22. To obtain the target miRNAs of hub genes, an mRNA–miRNA visualization network was constructed using Cytoscape v3.8.2, and target miRNAs for more than two hub genes were selected for further analysis.

miRNA–lncRNA prediction

The Starbase v2.0 tool was used to predict lncRNAs upstream of the selected miRNAs to identify lncRNAs regulating these miRNAs [28]. An lncRNAs–miRNA visual network was constructed using Cytoscape. Finally, the intersection of each miRNA prediction result was obtained.

Lentivirus-mediated rescue of Prx II

To construct lentiviral vectors to rescue Prx II expression in Prx II^−/− DMSCs, the murine full-length Prx II cDNA was used as a PCR template to clone Prx II into the LV5 plasmid, and an empty vector was used as the control. Vectors were purchased from Suzhou GenePharma (Suzhou, China). DMSCs were seeded into 6-well plates at 3 × 10⁵ cells/well and grown to 70–80%. Viral stock at a multiplicity of infection of 60 and 10 mg/mL polybrene were added to the DMSCs. WT cells and control cells (referred to as NC) were harvested.

Verification of hub gene, miRNA, and lncRNA expression using quantitative reverse transcription (RT-q) PCR

Total RNA was extracted from DMSCs using TRIzol reagent. RNA was quantified using a Nanodrop 2000 system (Wilmington, DE, USA) and reverse-transcribed using a cDNA synthesis kit. qPCRs were run using SYBR Green Master Mix on a CFX96 real-time PCR system. Oligonucleotide primers were synthesized and designed at Sangon Biotech and are listed in Table 1. The relative expression levels of hub genes, miRNAs, and lncRNAs were calculated using the 2^−ΔΔCq method.

Table 1 Primer sequences

Statistical analysis

Results are expressed as mean ± standard deviation (SD) of three independent experiments and were analyzed by GraphPad Prism 6.0 (GraphPad Software Inc., La Jolla, CA, USA). Student’s unpaired t-tests were performed. A P-value less than 0.05 was considered significant. IBM SPSS Statistics Version 25 was used for statistical analysis.

Results

Functional enrichment analysis

In our previous study, we used RNA sequencing to analyze and compare somatic Prx II^+/+ and Prx II^−/− DMSCs. The results showed that 472 genes were differentially expressed (176 upregulated and 296 downregulated) [29]. We used the DAVID software for GO term enrichment analysis of the DEGs, using a threshold of P < 0.05. The results showed that the upregulated genes were significantly enriched in the biological process terms positive regulation of the MAPK cascade (GO:0043410) and cell adhesion mediated by integrin (GO:0033631) (Fig. 1A). Downregulated genes were significantly enriched in the biological process terms positive regulation of cell migration (GO:0030335) and regulation of cell adhesion (GO:0022407) (Fig. 1B). Interestingly, these DEG-enriched GO terms were mainly related to cell migration and intercellular adhesion. We used the Cytoscape plug-in ClueGo to visualize the interaction network of biological processes, as shown in Fig. 1C, D. Each node represents a GO term, and nodes of the same color represent shared or similar genes among them. Therefore, nodes of the same color represent shared or similar genes, and different colors indicate that they are enriched with different genes. These results indicated that Prx II is involved in regulating DMSC homing.

Fig. 1

Gene ontology (GO) term analysis. A GO term analysis of upregulated genes between peroxiredoxin (Prx) II^+/+ and Prx II^−/− dermal mesenchymal stem cells (DMSCs) derived from mice. B GO term analysis of the downregulated genes between Prx II^+/+ and Prx II^−/− DMSCs derived from mice. The horizontal axis represents the gene numbers. The vertical axis represents enriched GO terms. We divided differential GO terms into cell migration (dark blue) and cell-to-cell adhesion (light blue) to facilitate differentiation. C, D ClueGo was used to analyze the interaction networks in enriched biological processes. Multi-colored dots indicate the involvement of multiple biological processes

PPI network analysis of cell migration and intercellular adhesion

To investigate the molecular mechanism underlying Prx II-regulated DMSC homing, we screened GO terms (Tables 2, 3) related to cell migration and intercellular adhesion, as they are closely associated with stem cell homing. The cell migration and cell adhesion genes were separately used to construct PPI networks using the STRING database. The networks were visualized and analyzed using Cytoscape. Analysis of the nodal properties in each network revealed an average degree of 10.372 for cell migration nodes and of 2.6154 for intercellular adhesion nodes. Degree refers to how many nodes this node is connected to. The more connected nodes, the more this gene is in a central position of the regulatory network (i.e., a hub gene). Here, we calculated the average degree of DEGs in cell migration and cell adhesion, which were 10.372 and 2.6154, respectively. We then considered higher than average degree for DEGs as a condition for hub gene screening. Three genes were identified as hub genes for cell migration: Cd274 antigen (Cd274), interleukin 1 beta (Il1B), and signal transducer and activator of transcription 1 (Stat1). Three genes were identified as hub genes for intercellular adhesion: Cd274, Il1B, and Ccl5 (Fig. 2A, B and Table 4). Next, we detected the mRNA expression levels of Cd274, Il1b, Stat1, and Ccl5 in the two cell types using RT-qPCR, and the results were consistent with the RNA sequencing results (Fig. 2C). The expression of Prx II in the NC and WT cells was determined using immunoblotting. Prx II protein was successfully expressed in the WT group (Fig. 2D). When Prx II was expressed in Prx II^−/− DMSCs cells, the expression levels of Cd274, Il1b, Stat1, and Ccl5 in WT cells were significantly higher than those in NC cells (Fig. 2E). Together, these results suggested that Prx II regulates DMSC homing by regulating these hub genes.

Table 2 GO terms associated with cell migration

Table 3 GO terms associated with cell–cell adhesion

Fig. 2

Protein–protein interaction (PPI) networks of cell migration and intercellular adhesion biological processes. A, B PPI network constructed to screen hub genes. Node color: shades of green to blue represent high to low values of betweenness centrality (BC). Node size: circle size represents the node degrees. Larger and darker-colored nodes represent genes with more links. C, D Expression levels of hub genes detected using RT-qPCR. Data are presented as the mean ± SD from at least three independent experiments (*P < 0.05, **P < 0.01, ***P < 0.001)

Table 4 Key genes selected based on visualize parameters like BC and degree

Further miRNA mining and interaction network analysis

As miRNAs recognize their target mRNAs through base complementary pairing, we predicted that Prx II regulates mRNAs via miRNAs and therefore explored how Prx II regulates hub genes. Four genes related to DMSC homing were screened, and gene-miRNA analysis was performed using miRWalk. The intersection of the miRNAs predicted by the TargetScan, miRanda, miRDB, miRWalk, and RNA22 databases was selected as the prediction result. The selection threshold was set to P < 0.05. The minimum sequence length was a 7-mer, and the target gene-binding region was the 3′ untranslated region. Cytoscape was used to draw the interaction network, as shown in Fig. 3A. We selected miRNAs with gene cross-linking (> 2) for further study (Fig. 3A and Table 5). Three miRNAs were predicted, and RT-qPCR was used to verify their expression levels in Prx II^+/+and Prx II^−/− DMSCs. Only miR-351 was found to be altered (Fig. 3B). After Prx II rescue in Prx II^−/− DMSCs, the miR-351 expression levels in WT cells were significantly higher than those in NC cells (Fig. 3C). These findings indicated that Prx II may regulate Il1b/Stat1 through miR-351 to regulate stem cell homing.

Fig. 3

Interaction network of hub genes and their target miRNAs. A Network of genes and miRNAs. B, C Genes are colored in blue, and the node size represents the number of targeted miRNAs. miRNAs are colored in red; miRNAs targeting more than two genes are colored in green. The expression levels of miRNAs targeting more than two hub genes were detected using RT-qPCR. Data are presented as the mean ± SD from at least three independent experiments (*P < 0.05, **P < 0.01, ***P < 0.001)

Table 5 miRNAs and its target genes

LncRNA prediction

To explore the specific underlying molecular mechanism, we searched for molecules upstream of miRNAs. We evaluated lncRNAs, and the lncRNA of muu-miR-351 was predicted using Starbase. Cytoscape was used to draw the interaction network shown in Fig. 4A. Next, we analyzed differentially expressed lncRNAs between Prx II^+/+ and Prx II^−/− DMSCs using lncRNA sequencing. We identified 464 differentially expressed lncRNAs (262 upregulated and 202 downregulated). The differentially expressed lncRNAs were plotted in a heat map (Fig. 4B). Interestingly, only lncRNA Tctn2 was significantly different (P < 0.001) (Table 6). The results showed that after Prx II knockout, Tctn2 expression decreased, which was consistent with the sequencing results (Fig. 4C). Similarly, Tctn2 expression was measured in NC and WT cells; it was increased in WT cells when compared with NC cells (Fig. 4D).

Fig. 4

Interaction network between micro-RNAs (miRNAs) and long non-coding RNAs (lncRNAs). A lncRNAs–miRNA network. miRNAs are colored in blue, lncRNAs are colored in red, and differentially expressed lncRNAs are colored in green. B Heat map of the differentially expressed lncRNAs of three Prx II^+/+ dermal mesenchymal stem cells (DMSCs) and peroxiredoxin (Prx) II^−/− DMSCs. Each row represents a differential lncRNA, lncRNAs shown in red are upregulated and those shown in green are downregulated. C Expression levels of the lncRNA Tctn 2 in Prx II^+/+ DMSCs and Prx II^−/− DMSCs as determined by RT-qPCR. D Expression levels of Tctn 2 in NC and WT cells by RT-qPCR. Data are presented as the mean ± SD from at least three independent experiments (*P < 0.05, **P < 0.01, ***P < 0.001)

Table 6 List of differentially expressed LncRNAs

Discussion

During tissue injury, various chemokines, adhesion factors, and growth factors are locally produced and secreted, thus initiating MSC homing [30, 31]. The interactions between chemical factors and their receptors in the internal environment guide stem cells to the lesion site [30, 31]. Several studies have shown that MSC homing involves substances very similar to those involved in white blood cell homing to inflamed tissues [32]. MSCs show strong adhesion, crawling, diffusion, and transendothelial cell migration in the treatment of various diseases [32]. Furthermore, the homing rate of MSCs directly determines their therapeutic effect. Therefore, identifying methods to enhance or mediate the migration ability of stem cells is a promising approach for improving the recruitment of stem cells to benefit patients.

The migration and recruitment of MSCs to the target site is essential for the success of MSC-based therapies [32]. In this study, we investigated Prx II^+/+ and Prx II^−/− DMSCs by RNA sequencing analysis. We found that the DEGs were mainly enriched in cell migration and intercellular adhesion. In a PPI network, Cd274, Il1b, Stat1, and Ccl5 were identified as hub genes. Studies have shown that NKX2-1-AS1 transcripts control cell adhesion and migration of lung tumor cell lines by regulating CD274 [33]. Oviduct epithelial cell culture medium supplemented with IL1B significantly induced polymorph nuclear neutrophil migration [34]. The CCL5/CCR5 axis has been shown to enhance tumor local invasiveness by inducing intracellular calcium-dependent signaling cascades and matrix metalloproteinase [35]. The transcription factor STAT1 promotes the migration of nasopharyngeal carcinoma cells [36]. In addition, STAT1 regulates CD274 to participate in pancreatic cancer progression [37]. We found that the expression level of miR-351 decreased after Prx II knockout, and miR-351 was predicted to regulate the expression of Stat1, indicating a positive regulatory relationship between miR-351 and Stat1. Through further analysis of the underlying mechanism, we identified an lncRNA upstream of miR-351 (Tctn2) and found that Prx II is involved in the regulation of this lncRNA. In summary, these data indicate that Prx II regulates the homing of DMSCs by regulating the Tctn2/mir-351/Stat1/Il1b axis (Fig. 5). The study findings provide a new approach for treating various diseases using stem cells.

Fig. 5

Peroxiredoxin (Prx) II regulates dermal mesenchymal stem cell (DMSC) homing. Prx II participates in DMSC homing by downregulating the expression of hub genes, including Cd274, Il1b, Stat1, and Ccl5, related to stem cell migration and adhesion. Further, Prx II regulates the Tctn2/miR-351/Stat1/Ccl5 axis involved in DMSC homing so as to repair damaged tissues

Prx II, as a peroxidase, plays an important role in maintaining the concentration of hydrogen peroxide [38]. When the TNF-α receptor is activated, Prx I and Prx II modulate the intracellular hydrogen peroxide level, thus regulating cell signal transduction [39]. Prx II not only acts as a peroxide reductase, but also regulates signal transduction by regulating the redox state of protein kinase [40]. We previously found that Prx II can regulate the senescence of DMSCs and determined the regulatory mechanism in detail, providing a foundation for analysis of the relationship between Prx II and the regulation of stem cell function [19, 20]. In this study, Prx II was identified as an important regulator of stem cell homing. Although numerous studies have investigated the use of stem cells in the treatment of various diseases, including cancers, no clinically significant treatment results have been obtained, likely because stem cells have different proliferation mechanisms in the body. As this study was based on in vitro and in silico analyses, the clinical relevance of the findings requires further study.

We identified 464 differentially expressed lncRNAs after knocking down Prx II, suggesting that Prx II is a key target for lncRNA regulation. Studies have shown that lncRNAs have many mechanisms of action. LncRNAs (1) induce transcription in the upstream promoter region of the protein-coding gene, thus interfering with the expression of adjacent protein-coding genes, (2) inhibit RNA polymerase II or mediate chromatin remodeling and histone modification, thus affecting gene expression, and (3) bind to specific proteins to regulate their activity [41]. Thus, there may be a mutual regulatory relationship between genes and lncRNAs. Elucidating the specific regulatory mechanism of Prx II and lncRNAs in stem cell homing will be the focus of our further research.

In conclusion, using RNA sequencing and bioinformatics analyses, we found that DEGs were mainly enriched in the GO terms cell migration and intercellular adhesion after Prx II knockout, and that Prx II regulates the homing of DMSCs by regulating the lncRNA Tctn2/mir-351/Stat1/Il1b axis. It is of great significance to clarify the specific molecular mechanism of stem cell homing and external regulation of the target to improve the homing efficiency of MSCs.