Umbilical Cord Mesenchymal Stem Cells Arrest Intervertebral Disc Degeneration by Enhancing Cell Migration

Meng Zhang China-Japan Union Hospital of Jilin University Butain Zhang China-Japan Union Hospital of Jilin University Ran Li China-Japan Union Hospital of Jilin University Te Liu China-Japan Union Hospital of Jilin University Jun Zhang China-Japan Union Hospital of Jilin University Ying Wang China-Japan Union Hospital of Jilin University Fuqiang Zhang China-Japan Union Hospital of Jilin University Xupeng Mu (  muxupeng@jlu.edu.cn ) China-Japan Union Hospital of Jilin University Jinlan Jiang China-Japan Union Hospital of Jilin University


Background
Intervertebral disc degeneration (IDD) is characterized by a decrease in the number of nucleus pulposus cells and abnormalities in the brous ring structure, followed by the disruption of the balance between matrix anabolism and catabolism. 1,2 The reduction of water content, glycoprotein, and type II collagen in the intervertebral disc (IVD) leads to changes in the structure and morphology of the IVD. These changes eventually affect the function of the IVD and cause lower back pain, which brings huge treatment costs, causing a serious economic burden. 3 Pharmacological treatment and surgery can relieve symptomatic pain of IDD but fail to reverse disc degeneration. 4 Therefore, various kinds of biological therapeutic methods have been studied in the hope of regenerating the dysfunctional discs. 5 Among these methods, stem cell therapy is of high interest. 6 Due to the extensive tissue distribution of mesenchymal stem cells (MSCs), along with their pluripotency and non-immunogenicity, they play an important role in damaged tissue repair, which makes MSCs attractive for IVD regeneration. 7 In vitro, co-culture of MSCs and IVD cells can differentiate into nucleus pulposus-like cells and synthesize glycosaminoglycans and type II collagen. 8 In vivo, MSCs are capable of differentiating into nucleus pulposus-like cells, increasing the extracellular matrix (ECM), as well as disc hydration, disc height, and glycosaminoglycan production. 9 Compared with other types of cells for transplantation, umbilical cord mesenchymal stem cells (UCMSCs) have several advantages. 10 UCMSCs are derived from umbilical cord discarded after birth. Separation and cultivation are simple and effective, which do not require an invasive procedure. 11 In addition, UCMSCs can expand in vitro and still retain pluripotency after many generations of expansion. 12 However, in most previous studies, MSCs were transplanted to the degenerative disc site through invasive intradiscal injection. During intradiscal injection, it is necessary to enter the nucleus pulposus cavity through annulus brosus, but this process destroys the physiological structure of the disc that causes disc degeneration. 13 This local transplantation effect relies on MSCs to differentiate into IVD-like cells to repair the damaged tissue. 14 However, the microenvironment of the IVD tissue includes low oxygen, low nutrition, acidic pH and high mechanical load, which poses a huge threat to the survival of stem cells. 15 Also, cell leakage remains a side effect to be considered when using intradiscal injection. 16 More recently, some studies have addressed the role of MSCs in the injury site homing following intravenous injection. 17 Utilizing the homing ability of systemically delivered MSCs can bypass the increased complications associated with intradiscal injection into the IVD. In bovine whole organ culture, MSC recruitment occurs in response to IVD degrading conditions. 18 A study comparing intravenous stem cells to intradiscal disc stem cells showed that regardless of the delivery method used, MSC implantation is not observed at the injured site. Both delivery methods increase the expression of glycosaminoglycan (GAG) and Acan, suggesting that there may be paracrine effects. 19 Another study proved that systemic perfusion of MSCs in the mouse disc degeneration group increases the number of stem cell transplantation in endplates, brous rings, and even the NP compared with the normal control group. 20 Cunha et al. found that systemic transplantation of MSCs has a bene cial effect on IVD regeneration in situ, and stated that tissue regeneration is achieved through complex interactions between stromal cells and immune system cells. 21 Although the exact mechanism of MSCs homing to damaged tissues has not been resolved, there is increasing evidence that chemokines and their receptors are important factors in controlling cell migration. 22 In particular, interactions between stromal cell-derived factor-1α (SDF-1α, also known as CXC motif chemokine 12, CXCL12) and its receptor CXC chemokine receptor type 4 (CXCR4) are essential in this process. 23 Additionally, hepatocyte growth factor, via interactions with its receptor c-met, and monocyte chemoattractant protein-1 (MCP-1), via interactions with chemokine receptor 2 (CCR2), are involved in stem cell migration 24 . However, the expression levels of these receptors are low in MSCs in vitro expansion. 25 Therefore, systemic delivery of MSCs can repair damaged intervertebral tissue, but signi cant challenges remain in effectively implanting them into the target site to enhance successful implantation. 26 Promoting and improving cell homing remains a major challenge in regenerative medicine.
Nanotechnology and nanoparticles (NPs) are widely used in regenerative medicine and have good potential in the diagnosis and treatment of different diseases. 27 One nanoparticle is iron oxide (Fe 3 O 4 ), which has been widely used as a contrast agent and drug carrier in preclinical and clinical environments. 28 Due to its good biocompatibility and response to magnetic elds, many studies have recently used it as a tracer for stem cells. 29 Some studies have shown that NPs do not alter the viability, proliferation, phenotype, function, and migration characteristics of MSCs in vitro. 30 However, some other studies suggest that Fe 3 O 4 nanoparticle labeling does not affect stem cell viability, differentiation, and proliferation, but improves stem cell migration to tumor, in ammation, and degenerative disease sites. 31 Although the molecular mechanism by which Fe 3 O 4 nanoparticle-labeled MSCs have enhanced migration capacity has not been elucidated, it has been suggested that Fe 3 O 4 nanoparticles enhance stem cell migration by up-regulating the expression of CXCR4 on MSCs. 32 Increasing evidence shows that cell responses to Fe 3 O 4 nanoparticles have indeed been observed, although the inherent characteristics of Fe 3 O 4 nanoparticle-labeled MSCs have not been fully explored. 33 Therefore, the strategy of using Fe 3 O 4 nanoparticles to enhance stem cell migration capacity can improve the effectiveness of stem cell therapy. 34 In this study, we synthesized polydopamine-capped Fe 3 O 4 (Fe 3 O 4 @PDA) and explored the role of UCMSCs labeled with NPs in enhancing stem cell regeneration in disc degeneration (Fig. 1A). Moreover, we tested our hypothesis that Fe 3 O 4 @PDA NPs stem cell labeling strategies to increase their migration capacity optimize IDD therapy.

Transmission electron microscopy (TEM) of Fe 3 O 4 @PDA NP localization in UCMSCs
We used TEM to assess the uptake and localization of Fe 3 O 4 @PDA NPs s by UCMSCs and their possible in uence on cell ultrastructure. MSCs treated with Fe 3 O 4 @PDA NPs (50 µg/mL) were washed twice with PBS and xed in phosphate-buffered Karnofsky's solution, followed by staining with 2% osmiumtetroxide at 4 °C overnight. Then cells were scraped off the plastic and dehydrated in an ethanol series. The specimens were embedded in Araldite (Serva). Ultrathin sections were established using an ultramicrotome (Leica, Bensheim, Germany), mounted on pioloform-coated copper grids, stained with uranyl acetate and lead hydroxide, and analyzed using a Tecnai Spirit system (120 kV; FEI, Hillsboro, OR, USA).

Effects of Fe 3 O 4 @PDA NPs on UCMSC viability and proliferation
CCK-8 assay was used to measure cell viability and proliferation. Cells were seeded into 96-well plates (2×10 4 cells/well) and grown overnight, and then incubated with Fe 3 O 4 @PDA NPs at different concentrations (0, 25, 50, 75, 100, and 150 µg/mL). After 24-h incubation, CCK-8 reagent (Sigma-Aldrich) was added to each well, and the cells were incubated for 2 h at 37 °C. The optimum density (OD) of cells in each well was measured at 450 nm (OD 450 ) using a microplate reader (Bio-Rad Laboratories Inc., Hercules, CA, USA).
Cells were incubated with 50 µg/mL nanoparticles for 24 h. Fe 3 O 4 @PDA NP-labeled cells were digested and seeded in 96-well plates (4×10 3 / well). At different time points (days 1, 3, 5 and 7), the CCK-8 reagent was added to measure the OD 450 on a microplate reader to assess cell proliferation capacity.

Effects of Fe 3 O 4 @PDA NPs on UCMSC surface markers and differentiation
After incubation with Fe 3 O 4 @PDA NPs, the labeled and control cell populations (1×10 6 ) were washed three times with PBS and harvested. Then, cells were incubated with antibody cocktails against FITC CD90, PerCP-Cy™5.5 CD105, APC CD73, PE CD45, PE CD34, PE CD11b, PE CD19 and PE HLA-DR in the dark for 30 min at 4 °C. After incubation and two washing steps, the probe was measured using ow cytometry (FC500; Beckman Coulter Inc., Brea, CA, USA) with CXP software (Beckman Coulter Inc.). All antibodies were purchased from BD Biosciences.
To evaluate the osteogenic and adipogenic differentiation capacity of stem cells labeled with Fe 3 O 4 @PDA NPs, we performed differentiation induction experiments on labeled and unlabeled stem cells. UCMSCs were incubated with Fe 3 O 4 @PDA NPs (50 µg/mL) for 24 h, then the cells were digested and centrifuged. They were seeded into a 12-well plate (5×10 3 cells/cm 2 ). After 3 days of cell culture, different differentiation induction media were added. For the adipogenic induction, MSCs were grown in an adipogenic differentiation medium (StemPro Adipogenesis Differentiation Kit; Invitrogen Australia Pty Ltd; Australia). Cells were cultured for 14 days and the media were changed every 3 days. Then cells were stained with Oil Red O for further analysis of adipogenic differentiation. For the osteogenic induction, UCMSCs were grown in an osteogenic differentiation medium (StemPro Osteogenesis Differentiation Kit; Invitrogen Australia Pty Ltd; Australia). Cells were replaced with new medium every 3 days. After 21 days, cells were stained with 2% Alizarin Red S solution and visualized under a light microscope and images were captured for analysis.

Migration assays
The migratory ability of labeled and control UCMSCs was tested in a 24-well Transwell plate (FluoroBlok, 8.0 µm colored polyester membrane; BD Biosciences). Precisely, 1×10 6 cells in growth medium supplemented with 1% FBS were added to the top chambers and the lower chambers were lled with the growth medium containing 10% FBS. After incubating the plates for 24 h at 37 °C in a humidi ed 5% CO 2 atmosphere, cells on the lower side of the membranes were xed with 4% paraformaldehyde for 15 min and counted under an optical microscope (X51; Olympus, Tokyo, Japan) after staining with 0.5% crystal violet (Sigma-Aldrich).

Animal experiments
The use of animals was approved by the Welfare and Research Ethics Committee on Animal Experiments of Jilin University. A total of 45 Sprague Dawley rats aged 6-8 weeks (nine rats/group) were used in this study. The IDD model was established by caudal needle puncture, as previously described. 36 Brie y, rats were anesthetized with 10% chloral hydrate (3.5 mL/kg) to identify and mark the six and seventh coccygeal vertebrae (Co6/Co7) via X-ray radiography (Faxitron X-ray, Lincolnshire, IL, USA) under appropriate parameters (35 kV, 300 mA and 1 min). A 22G needle was applied to percutaneously penetrate the IVD, rotated 360° and held in position for 30 s before extraction. Rats were subjected to standard postoperative procedures. Rats were randomly divided into 4 groups; control group, saline group, UCMSC group (MSC unlabeled with Fe 3 O 4 @PDA), and UCMSC + Fe 3 O 4 @PDA group (UCMSC labeled with Fe 3 O 4 @PDA). Twenty-four hours after surgery, cells (1×10 6 ) were resuspended in saline and then injected into the tail vein. Before injection, labeled and unlabeled UCMSCs were stained using CM-Dil (Sigma-Aldrich) according to manufacturer's instructions. All rats were euthanized after 2 weeks of treatment.

Calculation of disc height index (DHI)
Lateral plain radiographs of the rat tails were taken before and 2 weeks after injury using a cabinet X-ray system with an exposure time of 60 s and penetration power of 35 kV. The percentage of DHI was calculated using RadiAnt software as previously described. 37 The change in DHI was expressed as the %DHI [%DHI = (postpunctured DHI/prepunctured DHI) ×100]. Three independent observers independently measured and interpreted the radiographic images in a blinded manner.

Real-time quantitative polymerase chain reaction (RT-qPCR)
Total RNA was isolated from the cells of nucleus pulposus or Co6/7 using the Trizol reagent according to the manufacturer's instructions. RNA was reverse transcribed into cDNA according to the manufacturer's instructions. RT-qPCR was carried out using the ABI StepOnePlus RT-PCR System (Applied Biosystems, Foster City, CA, USA) with SYBR® Green PCR Master Mix (Applied Biosystems, Warrington, UK). The relative expressions of aggrecan, type II collage, Sox-9, Mmp-13, Tnf-α, and Il-1β were determined and normalized to the GAPDH housekeeping gene. The primers are presented in Table. 1. The real-time PCR thermo-cycler parameters consisted of an initial enzyme activation step at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s, 55 °C for 15 s, and 72 °C for 30 s. The cycle threshold (Ct) value of each sample was calculated in triplicate. The 2 −ΔΔCt value was then used to calculate the target gene relative expression against GAPDH mRNA levels. 38 2.13 Histological and immunohistochemical analysis of the IVD Two weeks after cell transplantation, Co6/7 discs were harvested with adjacent vertebrae. Samples were xed in 4% paraformaldehyde, decalci ed in Morse's solution (22.5% formic acid, 10% sodium citrate in deionized water) and embedded in para n. Disks were sectioned at a thickness of 5 µm and then stained with Safranin-O-fast green as well as hematoxylin and eosin to analyze the histological grading score. 39 The above sections were then used for immunohistochemistry to detect aggrecan, type II collagen, Sox-9, Mmp-13, Tnf-α and Il-1β. The sections were left at room temperature for 60 min and dewaxed twice with xylene (10 min each). Then, they were treated with 100, 95, 80, and 70% ethanol for 5 min, and washed three times with PBS. Endogenous peroxidase was inactivated with 3% hydrogen peroxide. Antigen retrieval was performed by heating the sample at 95 °C for 20 min in 10 mM sodium citrate (pH 6.0). The sections were observed using an optical microscope (Olympus, Tokyo, Japan).

Evaluation migration of transplanted UCMSCs
We used immuno uorescence technology to study the fate of CM-Dil-labeled UCMSCs after transplantation. Two weeks after cell transplantation, Co6/7 discs were harvested, xed and decalci ed with adjacent vertebrae. The intervertebral disc tissue was sliced into 5-µm sections by a freezing microtome (LEICA, Germany). The nuclei were stained with 4', 6'-diamino-2-phenylindole (DAPI, HARVEY, USA). Sections were observed using a uorescence microscope (Olympus Corporation, Tokyo, Japan) to analyze the positive of CM-Dil-labeled cells.

Characterization of Fe 3 O 4 @PDA NPs
In the current study, water-dispersible, uniformly distributed PDA-coated Fe 3 O 4 NPs were synthesized.
TEM images showed that the prepared Fe 3 O 4 NPs was very uniformly dispersed (Fig. 1B), and we observed an average particle size of about 50 nm. The average diamBeter was about 53.50 nm (Fig. 1C).

Internalization of Fe 3 O 4 @PDA NPs
To study the cellular uptake of Fe 3 O 4 @PDA NPs, we used ICP-OES and quanti ed iron in cell lysates after a 24-h co-incubation. ICP-OES results revealed that Fe 3 O 4 @PDA NPs were taken up by UCMSCs in a dosedependent manner ( Fig. 2A). Furthermore, we used TEM to further observe the fate of the particles after being phagocytosed by the cells, and the results showed that the particles mainly existed in the cytoplasm (Fig. 2B). To observe the internalization more directly and labeling e ciency of Fe3O4@PDA NPs, Prussian blue staining was used. The results were consistent with ICP-OES measurements; after the higher dose treatment, more NPs were internalized by the cells (Fig. 2C). When the NP concentration was not less than 50 µg/mL, the cells could be effectively labeled.  . 2D). There was no statistical signi cance among different concentrations. At the NP concentration of 50 µg/mL, the cell viability was not affected and the cells could effectively be labeled; therefore, we chose 50 µg/mL Fe 3 O 4 @PDA NP concentration for further experiments. To assess the potential in uence of NPs on UCMSC proliferation, we evaluated cell proliferation (1, 3, 5 and 7 days) after the stem cells were labeled with Fe 3 O 4 @PDA NPs. Compared with the control, NP-labeled UCMSCs showed similar proliferation rates without statistical signi cance (Fig. 2E).

Impact of Fe 3 O 4 @PDA NPs on UCMSC functionality
To determine whether Fe3O4@PDA NP labeling changed stem cell characteristics in UCMSCs, we used ow cytometry and analyzed the expression of characteristic markers on the surface of unlabeled and labeled stem cells. NP labeling did not alter the expression of typical surface markers on UCMSCs (Fig. 3A). In addition, NP-labeled and unlabeled UCMSCs were cultured in osteogenic and adipogenic differentiation media to evaluate the effect of NP labeling on stem cell differentiation. As shown in Fig. 3B, unlabeled and labeled cells had good osteogenic and adipogenic differentiation abilities.

Enhanced migration capacity of Fe 3 O 4 @PDA NP-labeled MSCs in vitro
Transwell experiments showed that nanoparticle labeling enhanced the migration of UCMSCs. There were signi cantly more migrated cells in the NP-labeled stem cell group than in the unlabeled group (Fig. 4A-B). We examined the mRNA and protein expression to study whether NPs increased UCMSC migration via CXCR4. As shown in Fig. 4C-D, the gene and protein expression levels of CXCR4 in the labeled group were higher than unlabeled group.

IVD height radiographic changes and histological evaluation
Radiographs were taken for each disc before and 2 weeks after the puncture. The %DHI remained relatively stable in the control group whereas it decreased continuously in all three operated groups (saline, unlabeled UCMSC, and labeled UCMSC groups). The %DHI of the saline and unlabeled UCMSC groups continued to decrease during the entire experimental period. The %DHI of the labeled UCMSC group was higher than that of the saline and unlabeled UCMSC groups (P < 0.01) (Fig. 5A-B).
To quantify the degree of disc degeneration, we used a histological grading scale to assess the number and morphology of annulus brosus AF cells, the boundary between AF and nucleus pulposus, and the number and morphology of nucleus pulposus cells. As shown in Fig. 5C-D, H&E and Safranin-O/fast green staining revealed that the annulus brosus was composed of regularly arranged slices and NP appeared as an organized structure with abundant cellular components in the control group. In the saline group, the nucleus pulposus tissue had completely disappeared and the annulus brosus structures were destroyed. The groups treated with unlabeled and labeled UCMSCs improved the regeneration of nucleus pulposus and annulus brosus, whereas Safranin-O staining of the nucleus pulposus tissue was deeper and that of the AF tissue structure was more complete in the labeled UCMSCs group.

Fe 3 O 4 @PDA NP labeling enhanced UCMSC migration in degenerated IVD
To track cell migration into the coccygeal IVDs in rats in vivo, CM-Dil-labeled UCMSCs were detected in IVD after 2 weeks of cell transplantation. A large number of systemic injections of Dil-labeled UCMSCs were observed around the IVD, but fewer cells appeared at the center of the IVD. As shown in Fig. 6A, the number of labeled UCMSCs entering the disc increased. These results meant that much more UCMSCs in the labeled group can migrated into IVD than in the unlabeled group, and Fe 3 O 4 @PDA NPs upregulated enhanced the migration ability of UCMSCs to homing into degenerated IVD.

Gene expression and immunohistochemical analysis
To con rm the effect of UCMSCs transplantation on the disc matrix, the expression of acan, Col 2, Sox-9, Mmp-13, Tnf-α, and Il-1β in the disc were measured using real-time PCR. Two weeks after injury, the mRNA expression of acan, Col 2, and Sox-9 was decreased in the saline, unlabeled UCMSC, and labeled UCMSC groups compared with the control group, while Mmp-13, Tnf-α, and Il-1β expression was signi cantly increased in the saline, unlabeled UCMSC, and labeled UCMSC groups. Compared with the unlabeled UCMSC group, the expression of acan, Col 2, and Sox-9 mRNAs in the labeled UCMSC group was signi cantly increased, while that of Mmp-13, Tnf-α, and Il-1β was signi cantly decreased (Fig. 6B). The results showed that the labeled stem cells had increased aggrecan, type 2 collagen and Sox-9, and reduced Mmp-13, Tnf-α, and Il-1β expression. Figure 6C showed the immunohistochemical analysis of the expression of aggrecan, type 2 collagen, Sox-9, Mmp-13, Tnf-α and Il-1β in the nucleus pulposus tissue in different groups. Compared with the control group, aggrecan, type 2 collagen and Sox-9 expressions signi cantly decreased in the saline, unlabeled UCMSC, and labeled UCMSC groups. Mmp-13, Tnf-α and Il-1β expressions signi cantly increased in the saline, unlabeled UCMSC, and labeled UCMSC groups compared with the contol group. Aggrecan, type 2 collagen, and Sox-9 expressions were signi cantly higher in the labeled group than in the unlabeled group. In comparison with the unlabeled group, the Mmp-13, Tnf-α and Il-1β expression decreased drastically in the labeled group.

In vivo toxicity Test
Exactly 14 days after treatment, H&E staining was performed on the heart, liver, spleen, lung, and kidney of each group to analyze the toxicity of NPs to rats. H&E staining histological analysis did not reveal any morphological changes in the heart, liver, spleen, lungs, or kidneys in each group, further con rming the low toxicity of Fe 3 O 4 @PDA NPs in vivo (Fig. 7).

Discussion
The purpose of this study was to prove the hypothesis that nanoparticles enhance the ability of UCMSCs to repair IDD by increasing the migration of implanted cells. We have demonstrated the e ciency and safety of Fe 3 O 4 @PDA NP-labeled UCMSCs, which increased stem cell migration in vitro and effectively repaired degenerated discs in a rat caudal disc degeneration model after systemic injection.
MSC transplantation for the treatment of IDD has been recognized as a promising strategy. 40 One of the main obstacles of cell therapy to treat IVD degeneration is the method of delivery, which is largely limited to direct injection into IVD due to its avascular nature. 41 Whether the regeneration potential of stem cells in degenerated IVD tissue is due to the transdifferentiation and replacement of dead cells, or due to the secretion of soluble factors to stimulate local progenitor cell survival, proliferation, and differentiation is still a subject of debate. 42 Increasing evidence shows that the e ciency of MSC homing to injured tissues is an important parameter that affects the outcome of MSC treatment. Although the exact molecular mechanism by which MSCs migrate and enter the injury site is not fully understood, pretreatment of MSCs with genetic modi cation, speci c cytokines, or hypoxia to overexpress CXCR4 can promote the migration of MSCs to damaged tissues. 43 Therefore, using the homing ability of stem cells may bypass the risks associated with direct injection.
The characteristic of Fe 3 O 4 @PDA NPs in enhancing the expression of CXCR4 in stem cells will undoubtedly expand the application of them in nanomedicine. Unlike the complex process and potential safety concerns of altering the cell surface through conjugation of speci c proteins and retroviral vectors encoding homing receptors, this simple method was implemented by adding the appropriate concentration of Fe 3 O 4 @PDA NPs to the medium. After being internalized by the cell, Fe 3 O 4 @PDA NPs may be degraded into iron ions in the lysosome by hydrolytic enzymes. 44 After entering the mitochondrial membrane, iron ions can react with hydrogen peroxide and oxygen produced by mitochondria. Apopa et al. reported that reactive oxygen species (ROS) produced by iron NPs increase human microvascular endothelial cell permeability. 45 Yun et al. found that CXCR4 expression in MSCs is increased by internalizing Fe 3 O 4 NPs into cells and the level of ROS in MSCs also is increased accordingly. CXCR4 expression and the ROS level can be induced by Fe 3 O 4 NPs in a time-and concentration-dependent manner. 46 Xu et al. showed that the migration of MSCs via the up-regulation of CXCR4 is induced by ironcontaining NPs, and iron-containing NP-loaded MSCs with high CXCR4 expression maintain good biological activity and can migrate to cancer cells in vivo and in vitro. 47 Other studies have shown that iron overload can lead to harmful cellular consequences and enhanced cell proliferation by changing the expressions of genes involved in cell cycle control and reducing intracellular hydrogen peroxide. 48 Therefore, the choice of NP concentration is also very important for the migration of cells. Consistent with these studies, 50 µg/mL of Fe 3 O 4 @PDA NPs increased the migration of stem cells in this study. These results indicated that the regulation of CXCR4 by Fe 3 O 4 @PDA NPs might be a complex process involving various mechanisms in different cellular environments. The detailed mechanism will be further explored in future research.
Although the special microenvironment without blood vessels in inner AF and NP leads us to speculate that MSCs will not migrate into IVD, recent research suggested that MSCs can be recruited to respond to IVD lesions. 40 Illien-Jünger et al. demonstrated that degenerative conditions induce the release of factors promoting MSC recruitment in an ex vivo organ culture. 49 Zhang et al. prepared nanomaterials for a sustained release of stromal cell-derived factor-1α (SDF-1α), to recruit MSCs and in turn achieve regeneration of IVD. 50 Sakai et al found that during degeneration of the IVD, cells from the bone marrow can migrate to the IVD. The frequency of the cells is related to the blood vessel distribution of the IVD. As the distance from the blood vessel increases, fewer cells appear, and the migration effect becomes limited. 51 SDF-1α, also known as CXC motif chemokine 12 (CXCL12), is a powerful chemokine cytokine that plays a key role in the recruitment, proliferation and differentiation of stem cells. CXCR4 is a speci c receptor for SDF-1, which is widely expressed on stem cells. 52 SDF-1α are reportedly upregulated during the degeneration of IVD. 53 CXCR4 overexpression promotes MSC retention within the IVD and enhances the stem cell-based IVD regeneration. 54 Consistently, our results showed that labeling UCMSCs with Fe 3 O 4 @PDA NPs increased stem cell CXCR4 expression and effectively repaired degenerated discs. In addition to homing to the lesion, MSCs also participate in the repair process through their paracrine function. MSCs produce large amounts of growth factors, and communication with the in ammatory microenvironment is an important part of this process. 55 It has been demonstrated that the expression of proin ammatory mediators, TNF-α and IL-1β, increase in the degenerated disc tissue. 56 Some reports have shown that the immunomodulatory and anti-in ammatory effects of MSCs may help modulate the IVD repair process. 57 In the present study, we found that UCMSCs labeled with Fe 3 O 4 @PDA NPs decreased the expression of Mmp-13, Tnf-α and Il-1β and promoted the expression of aggrecan, type 2 collagen, and Sox-9 in the degenerated disc tissue of rats. These ndings might suggest that the SDF-1/CXCR4 axis played an important role in mediating MSC homing to IDD and paracrine function of MSCs might contribute to facilitating IDD regeneration. Our results showed UCMSCs labeled with Fe 3 O 4 @PDA NPs increased homing ability and could be transported through the tail vein of rats to the IVD, where there was a physical barrier and had the additional advantage of repeated treatment when necessary. However, the environment and degenerate model of IVD is different between rats and humans. Further transplantation studies should be performed using UCMSCs labeled with Fe 3 O 4 @PDA NPs in clinical applications.
The timing of cell therapy for disc degeneration is controversial. During disc degeneration, homeostasis gradually becomes disturbed and eventually leads to a pathological condition. Some studies suggested that the best time to treat disc degeneration is 2 or 4 weeks after disc injury and they concluded that the therapeutic effect of MSCs injection is more optimal in a later stage of degeneration rather than in early stages. 58 Other studies reported that for cell therapy, immediately or 24 hours after disc injury is the best time to treat disc degeneration and they postulated that a narrow window of opportunity may exist for a maximum therapeutic effect during the early stages of degeneration. 59 Within the scope of this study, we did not use magnetic resonance imaging to track the fate of stem cells labeled with Fe 3 O 4 @PDA NPs, because Fe 3 O 4 NPs are negative contrast agents and affect our evaluation of disc. 60

Conclusions
In conclusion, systemic transplantation of UCMSCs labeled with Fe 3 O 4 @PDA NPs retards IDD via enhancing migration and decreasing proin ammatory mediators in rats, which provides an e cient way to improve the effects of UCMSC transplantation in IDD treatment. Availability of data and materials Not applicable.

Ethics approval and consent to participate
This study was approval by the Ethic Committee at Jilin university. The use of animals for research purposes was in accordance with the Declaration of Helsinki, and the research did not contain any individual person's data in any form.