Retraction Note: Chemokine receptor 7 overexpression promotes mesenchymal stem cell migration and proliferation via secreting Chemokine ligand 12

Great interest has been shown in mesenchymal stem cell (MSC) therapy in a wide variety of clinical domains. However, the therapeutic efficiency depends on the proliferation and migration of MSCs. Chemokine receptors are involved in regulating the proliferation and migration to the specific organs of MSCs in different microenvironments. CXC receptor seven (CXCR7), a newly discovered Chemokine ligand 12 (CXCL12) receptor, has organ specificity for tumour migration. We hypothesized that CXCR7 expression affects proliferation and migration of MSCs. In present study, we constructed long-term and stable mMSCs lines overexpressing and suppressing CXCR7 modifications with lentiviral vectors. The transduction efficiencies, mRNA and protein expression of CXCR7 were significantly regulated. CXCR7 gene overexpression promoted mMSCs proliferation and migration, whereas suppressing CXCR7 had the opposite effect. Additional CXCL12 improved the vertical migration of mMSCs. The overexpression of CXCR7 increased the MSC-secreted CXCL12, VCAM-1, CD44 and MMP2 levels, which contributed to the improvement of mMSC proliferation and migration. Therefore, overexpressing CXCR7 improved the proliferation and migration of mMSCs, which may be attributable to the CXCL12 secreted by MSCs, leading to a positive feedback loop for CXCL12/CXCR7 axis. Our results may provide a potential method for improving the treatment effectiveness of mMSCs by overexpressing CXCR7. Mesenchymal stem cells (MSCs), with their multipotent capacity self-renewal potential

Considering the low expression of CXCR7 10,12 , it is inefficient to temporarily alter CXCR7 expression with its blockers CCX754 or CCX771, or to induce short-term genetic modification for in vivo therapy. Thus, it is necessary to construct MSCs such that the overexpression or suppression of CXCR7 is permanent and stable, as a goal of in vivo treatment. What's more, we used a very similar approach to address a question in regards to the effects of modulation of the beta-catenin/Wnt pathway in stem cells 13 . The aim of this study was to construct mMSCs with long-term and stable CXCR7 regulation via lentiviral vector transduction and to evaluate the effects of CXCR7 overexpression on the proliferation and migration of mMSCs as well as the underlying mechanism.

Materials and Methods
Cell culture. Mouse MSCs isolated from the bone marrow of C57BL/6 mice and 293FT cells were obtained from Cyagen Biosciences, Inc. (Guangzhou, China). The identification of mMSCs according to their cell surface phenotypes and their multipotency for differentiation along the adipogenic, osteogenic, and chondrogenic lineages was performed by the supplier, as described by Liu and Cai 13,14 . Either mMSCs or 293FT cells were cultured in a 1:1 mix of Dulbecco's modified Eagle media/nutrient mixture F-12 (DMEM/F12) (Wisent, Inc., St-Bruno, Quebec, Canada) containing 10% FBS (Wisent, Inc.) and 1% antibiotics (streptomycin and penicillin) and incubated at 37 °C in a humidified atmosphere of 5% CO 2 .
Lentiviral vector-mediated CXCR7 overexpression and knockdown in mMSCs. mMSCs with passage number 4-6 were used for transduction. The recombinant lentivirus vector over-expressing the murine CXCR7 was constructed by using a CMVIE promoter-dependent lentiviral expression vector 15 . The methods used in the present study were the same as methods we have previously used 13 . Briefly, the full-length coding sequence (CDS) of CXCR7 (NM_001271607, 1088 bp), which was purchased from Hanbio Biotechnology Co., LTD (Shanghai, China), was cloned into the lentiviral expression vector between the CMVIE promoter and ZsGreen, followed by T2A-Puromycin. Finally, the lentivectors pHBLV-CMVIE-CXCR7-ZsGreen-T2A-Puromycin (over-expressing CXCR7) were obtained, and the empty vector pHBLV-CMVIE-ZsGreen-T2A-Puromycin was used as an empty control, termed NC (normal control).
A murine CXCR7 knockdown construct expressing short hairpin RNA (shRNA) targeting endogenous CXCR7 was generated in a lentivirus-based shRNA vector driven by the U6 promoter containing ZsGreen 16 . Target sequences were selected by software available on the Invitrogen website. Additionally, a negative control was generated containing non-specific shRNA (pHBLV-U6-ShRNA-ZsGreen-Puromycin).
The recombinant plasmids pHBLV-CMVIE-CXCR7-ZsGreen-T2A-Puromycin, pHBLV-CMVIE-ZsGreen-T2A-Puromycin, pHBLV-U6-ShRNA-ZsGreen-Puromycin or pHBLV-U6-ZsGreen-Puromycin were separately co-transfected with two packaging plasmids pSPAX2 and pMD2G into 293FT cells at the indicated concentrations using Lipofectamine 2000 (Hanbio Biotechnology) according to the manufacturer's instruction. mMSCs (2 × 10 5 /well) were seeded in 6-well cell culture plates, grown overnight and transduced with lentiviral vectors in a minimal volume of medium. Two days after the lentivirus vectors over-expressed the target gene, the stable cell lines were selected with 2 μg/ml puromycin. When all of the non-transfected cells disappeared and isolated colonies began to appear, the colonies with the greatest number of cells were chosen for expansion. A limiting dilution assay was used to obtain transduced cell clones expressing ZsGreen. The mMSCs were cultured in normal culture media for 20 passages after transduction to assay their long-term transfection efficiency. The transfection efficiency of the lentivirus vectors in passage 20 transduced-mMSCs was identified using fluorescence microscopy, and the percentage of ZsGreen-positive cells was determined by flow cytometry analysis using a Becton Dickinson FACS Calibur flow cytometer (FACS Calibur, Becton-Dickinson, Franklin Lakes, NJ, USA).

RNA isolation and quantitative real-time PCR (qRT-PCR).
Total RNA was extracted from the cells using TRIzol reagent (Ambion, Austin, TX, USA). The 260/280 absorbance ratio was measured to verify the purity and the concentration of the RNA. The qRT-PCR reaction was performed using the SYBR Green Realtime PCR Master Mix (Toyobo Co., Ltd., Osaka, Japan) and the Applied Biosystems (ABI) 7500 real-time PCR system (Applied Biosystems, Inc., Foster City, CA). Relative changes in the gene expression were normalized to the expression of MGAPDH levels and calculated using the 2( −ΔΔ Ct) method. The primer sequences used for PCR amplification in our study were designed as Table 1.
Flow cytometric analysis of CXCR7 expression on mMSCs. The CXCR7 expression on the surface of mMSCs was assessed by flow cytometry analysis. After harvested and washed in PBS, mMSCs were suspended in PBS at a concentration of 1 × 10 6 cells/ml and then incubated for 60 min with 10 µl of the primary antibody of CXCR7 (PE-CXCR7, biolegend) per 100 µl following analysis with a flow cytometer (BD Biosciences). Then the CXCR7 protein (relative expression) as CXCR7 (PE-A)-positive cells/total cells was calculated.
Cell proliferation assessment. The proliferation of mMSCs was analysed using Cell Counting Kit-8 (CCK8) (Dojindo Laboratory, Japan) according to the manufacturer's instructions. mMSCs were seeded in 96-well plates at 5 × 10 3 cells per well in 100 μl of growth medium. The cells were incubated for 2 h at 37 °C. Absorbance was assessed at 450 nm using a microplate reader (Bio-Tek, USA). The mean optical density (OD) of the three wells in each group was used to reflect the percentage of cell proliferation.
In vitro scratch assay. mMSCs were seeded in 6-well culture plates. When the cells reached approximately 100% confluence, a scratch was made with a 10 μl sterile pipette tip, and the cells were cultured in serum-free DMEM/F12 for an additional 24 h 17 . The images of the wound were recorded using a light microscope immediately after scratching and 18 h later. The horizontal migration abilities of the cells were quantified by measuring the wound widths of five different wound surfaces in each group using the Image-J analysis software. The experiment was performed three times. Transwell migration assay. The vertical migration of mMSCs was determined using the transwell migration assay. Transwell inserts (6.5 mm diameter and 8 μm pore size; Millipore) were loaded with 1 × 10 4 mMSCs in 200 μl of serum-free DMEM/F12, and 600 μl of DMEM/F12 supplemented with 10% FBS was added to the lower chambers. In two other groups, 50 ng/ml of CXCL12 (Abcam, Britain) was added to the lower chambers to evaluate the overexpression and suppression of CXCR7 separately. The cells were allowed to migrate at 37 °C in a humidified CO 2 incubator for 12 h. The cells remaining on the upper surface of the filter were then removed with cotton swabs, and the cells that migrated to the lower surface were stained with crystal violet (Beyotime Institute of Biotechnology, Haimen, China) for 20 min. Stained cells from five randomly chosen fields were counted under a light microscope.

Measurement of surface protein and cytokines by ELISA. After 18 h of culture in a transwell
assay, mMSCs and supernatants were collected and centrifuged to remove debris. Then, mMSCs were lysed by Membrane Protein Extraction Kit (Mem-PER ™ Plus, Thermo Scientific, USA). CXCL12 (SDF-1α), vascular cell adhesion molecule-1 (VCAM-1), Clusters of Differentiation-44 (CD-44), matrix metalloproteinase2 (MMP2), Collagen-I, tumour necrosis factor (TNF-α) and IL-10 were measured by enzyme-linked immunosorbent assay (ELISA) kits (RayBiotech, USA; Bio-Techne, USA; Cusabio Biotech, China; ExCell Biology, China). The content of CXCL12 was determined using a semi-quantitative standard curve, and the additional CXCL12 was removed. ELISA was performed according to the manufacturer's instructions. All samples were measured three times.
Statistical analysis. The data were presented as the means ± standard deviation (SD). Comparisons amongst multiple groups were performed by one-way ANOVA followed by Bonferroni's post hoc test. A p-value less than 0.05 was considered statistically significant.

Lentiviral vector transduction efficiency in mMSCs.
The ZsGreen-positive cell ratio was used in our study to reflect the transduction efficiency of the target gene in mMSCs. The efficiencies of the lentiviral vector transduction of MSC-OE-CXCR7, MSC-OENC-CXCR7, MSC-Sh-CXCR7 and MSC-ShNC-CXCR7 after 20 passages in mMSCs were 91.29%, 91.39%, 91.69% and 91.28%, respectively (Fig. 1a). Moreover, the RT-PCR results showed that CXCR7 mRNA expression were significantly higher in the MSC-OE-CXCR7 cells than in the MSC-OENC-CXCR7 cells (p < 0.05) and were significantly lower in the MSC-Sh-CXCR7 cells compared with the MSC-ShNC-CXCR7 cells (p < 0.05). There was no significant difference in CXCR7 mRNA expression between the MSC-Blank and MSC-OENC-CXCR7 or MSC-ShNC-CXCR7 groups (Fig. 1b). Similar results for CXCR7 protein expression were observed by flow cytometry analysis ( Fig. 1c-f). These results suggest that lentivirus-mediated transduction is efficient and stable.  after day 3 (Fig. 2a). In contrast, the suppression of CXCR7 (MSC-Sh-CXCR7 group) inhibited cell proliferation compared with the normal control group of suppression of CXCR7 (MSC-ShNC-CXCR7 group) (p < 0.05). The difference reached statistical significance on days 5 to 7 (Fig. 2b).
Overexpression of CXCR7 and additional CXCL12 promoted the migration of mMSCs. The effects of overexpressing CXCR7 on the horizontal and vertical migration of mMSCs were measured by the scratch and transwell assays, respectively. The scratch healing percentage increased in the MSC-OE-CXCR7 group compared with the MSC-Blank group at both 12 h and 18 h (p < 0.05), but it decreased in the MSC-Sh-CXCR7 group compared with the MSC-ShNC-CXCR7 group both at 12 h and 18 h (p < 0.05). No significant differences in the scratch healing percentage between the MSC-Blank and MSC-OENC-CXCR7 or MSC-ShNC-CXCR7 groups were observed (Fig. 3). More cells migrated in the MSC-OE-CXCR7 group than in the MSC-OENC-CXCR7 group at both 12 h and 18 h (p < 0.05). Conversely, fewer cells migrated in the MSC-Sh-CXCR7 group at both 12 h and 18 h compared with the MSC-ShNC-CXCR7 group (p < 0.05). There was no significant difference in number of cells that migrated between the MSC-Blank and MSC-OENC-CXCR7 or MSC-ShNC-CXCR7 groups (Fig. 4).
Additional CXCL12 improved the vertical migration ability of mMSCs. More cells migrated in the MSC-OE-CXCR7 + CXCL12 group than in the MSC-OE-CXCR7 group at both 12 h and 18 h (p < 0.05), and more migration was observed in the MSC-Sh-CXCR7 + CXCL12 group than in the MSC-Sh-CXCR7 group (p < 0.05) (Fig. 4).
These results suggested that overexpressing CXCR7 could improve the migratory ability of mMSCs, whereas suppressing CXCR7 does the opposite. Overexpression of CXCR7 increased MSC-secreted CXCL12. To explore the effects of modulating CXCR7 on MSC-secreted CXCL12, we further investigated the expression of CXCL12 mRNA by qRT-PCR, the expression of CXCL12 protein in mMSCs by western blot and the concentration of CXCL12 in the supernatant of transwell chambers by ELISA. The expression level of CXCL12 mRNA was significantly higher in the MSC-OE-CXCR7 and MSC-OE-CXCR7 + CXCL12 groups than in the MSC-OENC-CXCR7 group (p < 0.05) but was significantly lower in the MSC-Sh-CXCR7 and MSC-Sh-CXCR7 + CXCL12 groups than in the MSC-ShNC-CXCR7 group (p < 0.05). Furthermore, additional CXCL12 further improved the expression of CXCL12 mRNA in MSC-OE-CXCR7 + CXCL12 and MSC-Sh-CXCR7 + CXCL12 groups compared with the MSC-OE-CXCR7 and MSC-Sh-CXCR7 groups, respectively (p < 0.05). However, there was no significant difference between the MSC-BLANK and MSC-OENC-CXCR7 groups (p > 0.05) (Fig. 5b). Similar trends were shown in CXCL12 protein expression in mMSCs by western blot analysis (Fig. 5a).
The concentration of CXCL12 in the supernatant of transwell chambers increased in the MSC-OE-CXCR7 and MSC-OE-CXCR7 + CXCL12 groups compared to the MSC-OENC-CXCR7 group (p < 0.05) by ELISA. Despite excluding the interference of exogenous CXCL12, the concentration of CXCL12 was still significantly higher in the MSC-OE-CXCR7 + CXCL12 group than in the MSC-OE-CXCR7 group (p < 0.05). CXCL12 was lower in the MSC-Sh-CXCR7 group compared with that in MSC-ShNC-CXCR7 group (p < 0.05). However, there was no significant difference in CXCL12 between the MSC-ShNC-CXCR7 and MSC-Sh-CXCR7 + CXCL12 groups (p > 0.05) (Fig. 5c).

The effect of CXCR7 on mMSC secretion of Collagen-I.
To evaluate the detailed mechanism by which CXCR7 regulated the migration of mMSCs, ELISA was used to detect the Collagen-I in the supernatant of transwell chambers. However, there was no difference in Collagen-I among all groups (p > 0.05) (Fig. 7).

Overexpression of CXCR7 had no significant impact on TNF-α and IL-10.
To determine the effects of overexpressing CXCR7 on inflammation, the concentrations of TNF-α and IL-10 in the supernatant of transwell Chambers were measured by ELISA. There were no differences in the concentrations of TNF-α and IL-10 among all groups (p > 0.05) (Fig. 8a,b).

Discussion
Great interest has been shown in MSC therapy for a wide variety of clinical domains, such as moderate-to-severe ARDS [19][20][21][22] . However, its therapeutic efficiency depends on the proliferation and migration of MSCs to a target tissue 23 . The major findings of the present study were as follows: (1) the overexpression of CXCR7 significantly improved the proliferation and migration of mMSCs; (2) the positive feedback regulation of CXCL12 and the increased expression of  The third generation of self-inactivation (SIN) human immunodeficiency virus type 1 (HIV-1) was chosen as a vector of target gene overexpression in this study. Several studies have confirmed that HIV-1-based SIN lentivirus vectors sustain high levels of transgene expression in human embryonic stem cells, although their application in MSCs is still unknown 24,25 . Therefore, lentivirus vectors pHBLV-CMVIE-ZsGreen-T2A-Puromycin and pHBLV-U6-ZsGreen-Puromycin were successfully constructed for gene transfection and interference, which contains CMV promoter and U6 promoter.
Improving the proliferation and migration of MSCs to a target tissue is the key issue for improving the efficiency of the therapy. A concentration gradient of CXCL12 can specifically regulate different effects of chemotaxis in stem cells [26][27][28] . In a previous study, CXCR4 was considered the unique receptor of CXCL12 and was shown to play an important role in chemotaxis 11,29 . Balabanian and Burns found that there was another binding site of CXCL12, named RCD1 (CXCR7) in 2006. The affinity of CXCR7 to CXCL12 is 10 times greater than that of CXCR4 11 . However, the lower cell surface expression of CXCR7 than of CXCR4 has seriously restricted the role of the CXCL12/CXCR7 axis on the migration of mMSCs. The present study demonstrates that the overexpression of CXCR7 significantly improved and that the low expression of CXCR7 decreased the proliferation and migratory ability of mMSCs. Additional CXCL12 was added to simulate inflammation microenvironments, which further increased the migratory ability of mMSCs. Our results suggest that the overexpression of CXCR7 have a positive effect on MSC proliferation and migration. However, Liu et al. 30 found that CXCR7 was shown to have no influence on MSC migration in renal ischaemia/reperfusion injury models. Gheisari et al. 31 found that CXCR7-transfected MSCs were trapped in the lungs, whereas native MSCs home very rarely to kidneys and bone marrow. In addition, the activation of CXCR7 may improve the migration of colorectal cancer cells in the lung but not in the liver 9 . Thus, it is suspected that MSCs overexpressing CXCR7 may have organ specificity, especially in the lungs.
Paracrine adhesion factors have been proposed to be the primary mechanism for MSC migration 32 . To explore the detailed mechanism of increased migration of mMSCs by overexpressing CXCR7, the levels of VCAM-1, CD44, MMP2 and Collagen-I, which are important MSC paracrine factors that regulate the vascular adhesion of MSCs 33,34 , were measured in the supernatant of transwell chambers. However, VCAM-1, CD44, and MMP2 increased and decreased significantly after overexpression and suppression of CXCR7, respectively. It has been demonstrated that blocking VCAM-1 by its antibody significantly reduced the migration of MSCs 28 . Further, CD44 and MMP2 play important roles in the migration of MSCs. Thus, this study showed that the increased mMSC secretion of VCAM-1, CD44, and MMP2 is one of the primary mechanisms related to mMSC migration that are promoted by the overexpression of CXCR7.
In previous studies, the expression of CXCL12 increased significantly in injured and hypoxic tissues 10,35,36 . Thus, additional CXCL12 was added to the lower chambers of transwell assays to simulate the microenvironment of injured and hypoxic tissues. We found that the overexpression of CXCR7 in mMSCs not only improved their migration and proliferation but also increased their CXCL12 secretion. The mechanism is possibly described based on the following two phenomena. On one hand, more CXCL12 combined with its ligand CXCR7 when CXCR7 was overexpressed in mMSCs, which may have increased the ability of these cells to migrate. The supplemented CXCL12 may combine with more CXCR7 and CXCR4 on mMSCs in a positive feedback loop, which was also reported in Wang's study 37 . On the other hand, the high expression of CXCR7 may activate the CXCL12/ CXCR7 axis, thus making mMSCs secrete more CXCL12. It was supported by the increasing protein and mRNA level of CXCL12. If true, this would be a novel discovery, however, further study is needed.
One concern is that gene transfection may influence the function of mMSCs or promote inflammation 38,39 . There were no significant changes in TNF-α or IL-10 among the groups in the supernatant of transwell chambers in this study 40 . Our results may prove that MSCs that migrate after CXCR7 gene transduction are not associated with inflammation, which ensures the safety of overexpressing CXCR7 in mMSCs that are transferred to animals for further research.
One limitation of our study should be noted. In our experiments, lipopolysaccharide was not added, which cannot simulate endotoxin-related injury. This will be explored further in follow-up studies.
In conclusion, we demonstrated that lentiviral vector transduction successfully facilitated the sustained and efficient gene modification of CXCR7 in mMSCs. The overexpression of CXCR7 significantly improved the proliferation and migration of mMSCs without inducing inflammation. The mechanism is the positive feedback regulation of CXCL12 and is related to the increased expression of VCAM-1, CD44, and MMP2 but not Collagen-I. Our results provide a potential method for improving the treatment effectiveness of mMSCs by overexpressing CXCR7. Further animal research will be necessary to explore the treatment effects of overexpressing CXCR7 in mMSCs.