The migration and differentiation of hUC-MSCsCXCR4/GFP encapsulated in BDNF/chitosan scaffolds for brain tissue engineering

We previously developed a biomaterial scaffold that could effectively provide seed cells to a lesion cavity resulting from traumatic brain injury. However, we subsequently found that few transplanted human umbilical cord mesenchymal stem cells (hUC-MSCs) are able to migrate from the scaffold to the lesion boundary. Stromal derived-cell factor-1α and its receptor chemokine (C-X-C motif) receptor (CXCR)4 are chemotactic factors that control cell migration and stem cell recruitment to target areas. Given the low expression level of CXCR4 on the hUC-MSC membrane, lentiviral vectors were used to generate hUC-MSCs stably expressing CXCR4 fused to green fluorescent protein (GFP) (hUC-MSCsCXCR4/GFP). We constructed a scaffold in which recombinant human brain-derived neurotrophic factor (BDNF) was linked to chitosan scaffolds with the crosslinking agent genipin (CGB scaffold). The scaffold containing hUC-MSCsCXCR4/GFP was transplanted into the lesion cavity of a rat brain, providing exogenous hUC-MSCs to both lesion boundary and cavity. These results demonstrate a novel strategy for inducing tissue regeneration after traumatic brain injury.


Introduction
Traumatic brain injury (TBI) often results in mortality or long-term disabilities such as dysphasia or cognitive deficits [1]. The primary destruction and secondary injuries such as ischemia and inflammation caused by TBI may cause extensive tissue loss of cerebral parenchyma which results in cavities formation [2]. To date, there are no effective clinical treatments for repairing cerebral parenchyma following TBI [3]. Recently, stem/ progenitor cell transplantation has shown great promise for this purpose [4]. However, nerve repair remains a challenge, and new treatment methods are required to minimize neuronal death and provide exogenous seed cells to replenish lost neurons and promote the recovery of neurological function after TBI [5].
Recent advances in stem cell research have included the use of human umbilical cord mesenchymal stem cells (hUC-MSCs) in disease models [6][7][8][9][10]. Owing to their immunosuppressive capacity and ability to differentiate into various cell types [11,12], hUC-MSCs are a major candidate for cell-based therapies, particularly in regenerative medicine. Our previous work demonstrated that chitosan-based scaffolds can provide a suitable microenviron ment for hUC-MSC attachment and proliferation [5] and may therefore serve as a basis for TBI treatment.
Secondary injury resulting from trauma, including delayed release of inflammatory/biochemical mediators and ischaemia, can cause continuous neuronal damage at the lesion cavity edge. Cell migration is a critical aspect of stem cell recruitment to lesioned The migration and differentiation of hUC-MSCs CXCR4/GFP encapsulated in BDNF/chitosan scaffolds for brain tissue engineering areas; however, there are currently no effective methods to induce the migration of transplanted seed cells to lesion sites so as to replace cells lost through injury. Stromal cell-derived factor (SDF)-1α and its receptor chemokine (C-X-C motif) receptor (CXCR)4 are important chemotactic factors for stem cells [13]. Together they enhance proliferation, promote chain migration and transmigration, and activate intracellular pathways modulating neural stem cells [14]; SDF-1α has been shown to be critical for MSC migration [15], and high levels of SDF-1α/CXCR4 binding at injury sites help to retain mobilised CXCR4-positive cells [16]. However, only a small proportion (~1%) of cultured MSCs express CXCR4 [17][18][19][20][21][22].
To overcome this limitation and to determine whether SDF-1α/CXCR4 can induce the migration of transplanted hUC-MSCs, in this study we transfected hUC-MSCs with a lentiviral vector encoding CXCR4 and examined the effects of CXCR4 overexpression on cell migration in vitro and in vivo. hUC-MSCs CXCR4/GFP were induced to differentiate into neurons in the presence of recombinant human brain-derived neurotrophic factor (BDNF) released from cellular scaffolds. Our findings provide a potential strategy for the therapeutic application of SDF-1α/CXCR4 to promote the migration of transplanted hUC-MSCs to the site of brain injury following TBI.

Detection of cell surface marker expression
To confirm the identity of transfected hUC-MSCs CXCR4/GFP , surface antigen expression was evaluated by flow cytometry using phycoerythrinconjugated antibodies against cluster of differentiation CD34, CD73, and CD105; fluorescein isothiocyanateconjugated antibodies against CD45, CD90, and CD14; and an antibody against human leukocyte antigen HLA-DR (all from BD Pharmingen. San Jose, CA, USA). After incubation with the antibodies, cells were centrifuged and washed twice with PBS before they were sorted with a BD FACscan instrument (BD Pharmingen).

Fabrication of chitosan scaffolds with genipinimmobilized BDNF (CGB scaffolds)
Scaffolds were constructed as previously described [4]. Crosslinked scaffold solution was generated by adding 9 mg genipin (Yuanyebio, Shanghai, China) and 2 mg recombinant human BDNF (PeproTech, Rocky Hill, NJ, USA) to 10 ml acetic acid (1%) containing 0.2 g chitosan. The solution was stored in a freeze dryer (Thermo Scientific) at −56 °C for 24 h after vigorous mixing; the scaffolds were then immersed in 0.1 mol/l NaOH to neutralize the remaining acetic acid. After overnight drying at room temperature, scaffolds were repeatedly rinsed with ultra-pure water and sterilized by ultraviolet radiation for 1 h.
Adhesion and activity assays hUC-MSCs CXCR4/GFP (1 × 10 4 ) were cultured in medium containing 1% penicillin, 1% streptomycin (Solarbio, Beijing, China), and 10% FBS and added to CGB scaffolds in 24-well plates. Samples were incubated at 37 °C in 5% CO 2 /95% air. Scaffolds with adherent hUC-MSCs CXCR4/GFP were examined on day 1 by scanning electron microscopy. In briefly, the CGB scaffolds with or without hUC-MSCs CXCR4/GFP were fixed in 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB) for 30 min, washed in ice-cold PBS, blocked in O.C.T. and processed in a cryostat (Leica CM1950, Germany). Standard 25 mm thick sections were cut, then affixed to poly-lysine coated glass slides, then store in drying case before scanning. The images of scanning electron microscope (SEM) were randomly collected of each group (N = 3). Cell viability in the CGB and control groups (cells incubated with or without CGB scaffolds, respectively) was assessed with the CCK-8 assay. After adding 10 μl CCK-solution to each well for 1 h at 37 °C, absorbance was measured at 450 nm on a microplate reader. Cell viability was calculated using the equation: viability (%) = (absorbance of experimental group-absorbance of control group)/absorbance of control group × 100%.

Transwell migration assay
The role of the SDF-1α/CXCR4 signaling axis in hUC-MSC migration was assessed with the transwell assay using 8 μm inserts (BD Pharmingen). A total of 1 × 10 4 cells were resuspended in 100 μl serum-free DMEM/ F12 medium and loaded into the upper wells while the lower chambers were filled with 500 μl complete medium (DMEM/F12 supplemented with 10% FBS). For migration assays, SDF-1α was added to the lower chamber at a concentration of 50, 75, or 100 ng ml −1 ; another group was also treated with AMD3100 blocking reagent. After incubation in a humidified chamber of 5% CO 2 /95% air at 37 °C for 24 h, cells were fixed with 500 μl methanol for 15 min. For the migration assay, the inner surface of the upper chamber was wiped with cotton swabs to remove non-migrating cells; the chambers were then washed with 500 μl PBS and stained with 500 μl haematoxylin for 1 min at room temperature. After another wash with 500 μl PBS, transwell membranes were removed and placed on slides and stained cells in five random microscopic fields at 40 × magnification were counted using Image J software (National Institutes of Health, Bethesda, MD, USA).

Animal experiments
Animal experiments were approved by the Nantong University Animal Experimentation Committee and were carried out in accordance with their ethical guidelines, which were in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. A TBI model was generated as follows. Male Sprague-Dawley rats (12 weeks old; 380-450 g) were anesthetised with zoletil 50 (55 mg kg −1 body weight). A piece of skull over the left frontal cortex was removed by drilling. A dural incision was made to expose the forebrain, and defects with a diameter and thickness of 2 mm were created with a scalpel in the cortex at 2 mm to the right and left of bregma. After hemostasis, CGB scaffolds containing hUC-MSCs CXCR4/GFP or hUC-MSCs GFP were implanted into the injury sites. For control animals, saline was instead injected into the site (N = 6). After 2 weeks and 6 weeks, brains were harvested after transcardial perfusion with 250 ml saline and 250 ml 4% paraformaldehyde in 0.1 M PB and postfixed with 4% paraformaldehyde (Sigma) in 0.1 M PB (pH 7.4) for 2 h. After dehydration in 20% and 30% sucrose solutions, brains were sectioned and embedded in optimum cutting temperature compound (Surgipath, Richmond, IL, USA) and sectioned on a cryostat (CM1950; Leica, Wetzlar, Germany). The 10 μm thick sections were collected on poly-L-lysine-coated glass slides and stained for hematoxylin-eosin (HE) stain and immunocytochemical analysis. In brief, to identify differentiating neurons derived from hUC-MSC, sections of 6 weeks were incubated overnight at 4 °C with a mouse antibody against microtubule-associated protein (MAP)-2 (1:300, ab11267; Abcam), after blocking nonspecific binding in 10% goat serum for 30 min at RT, then labelled with iFluor 594 goat antimouse IgG (1:300, 16468; AAT Bioquest, Sunnyvale, CA, USA). Nuclei were counterstained with DAPI. Sections were mounted with mounting medium and analysed by fluorescence microscopy (Axiostar scope A1, Carl Zeiss, Jena, Germany).

Statistical analysis
All data are presented as mean ± standard deviation and were analysed with SPSS 10.0 software (SPSS Inc., Chicago, IL, USA). Differences between groups were evaluated by one-way analysis of variance (ANOVA) and were considered significant at P < 0.05 (designated as * in figures).

CXCR4 is stably overexpressed in hUC-MSCs CXCR4/ GFP
The efficiency of CXCR4 transfection was evaluated by qRT-PCR and western blotting. Protein localisation was determined by confocal microscopy ( figure 1(A)), while flow cytometry was used to quantify transfection efficiency ( figure 1(B)). CXCR4 expression was higher in hUC-MSCs CXCR4/GFP than in control cells (hUC-MSCs GFP ), as determined by qRT-PCR (figure 1(C)) and western blotting (figures 1(D) and (E)). These results indicate that transfected hUC-MSCs CXCR4/GFP stably express CXCR4.

hUC-MSCs CXCR4/GFP express stromal and MSC markers
Flow cytometry was used to identify immunophenotype of hUC-MSCs CXCR4/GFP . From the results, we can confirm that hUC-MSCs CXCR4/GFP expressed the stromal/MSC markers CD73, CD105, and CD90 but not the haematopoietic or endothelial markers CD14, CD34, CD45, CD79a, or HLA-DR (figure 2). These results demonstrate that the immunophenotype of hUC-MSCs CXCR4/GFP was unaffected by transfection.
hUC-MSCs CXCR4/GFP are compatible with CGB scaffolds There was no difference in viability between transfected cells and negative controls (P > 0.05) ( figure 3(A)), suggesting that the transfection was only mildly cytotoxic to hUC-MSCs CXCR4/GFP , which were available in large numbers for the repair of lesion cavities after TBI. The CCK-8 assay also showed no differences in viability between hUC-MSCs CXCR4/GFP grown in the presence or absence of a CGB scaffold (P > 0.05) ( figure  3(B)). The scaffolds were only mildly cytotoxic to hUC-MSCs CXCR4/GFP , indicating excellent biocompatibility. Scanning electron micrographs revealed a smooth scaffold surface with adherent and proliferating hUC-MSCs CXCR4/GFP (figure 4), suggesting that the scaffold can serve as a carrier for hUC-MSC CXCR4/GFP .
CGB scaffolds induce hUC-MSCs CXCR4/GFP differentiation hUC-MSCs CXCR4/GFP were cultured in neuronal differentiation medium with or without BDNF for 12 d and examined by immunocytochemistry to detect MAP-2 and NeuN double-positive neurons. The percentage of neurons was 20.5% ± 2.40% in BDNFinduced cells as compared to 4.3% ± 1.18% in the control group (P < 0.05; figure 6). These results indicate that BDNF released from CGB scaffolds enhances the differentiation of hUC-MSC into neurons.

Discussion
There are few effective treatments for TBI [1], which is associated with the loss of neurons and a high mortality  rate. Cell-based therapies have shown promise as a potential therapeutic strategy; biocompatible scaffolds can provide 3D support and influence cell properties, thereby improving their survival [23,24]. hUC-MSCs have many advantages-including their abundance and low immunogenicity-for the treatment of TBI.  In our previous work, we designed a mold based on the size of lesion cavities in a TBI model and used it to construct a chitosan scaffold for hUC-MSCs [5]. The chemokine receptor CXCR4 along with its ligand SDF-1α are important chemotactic factors for stem cells that determine their retention and/or migration [13,17]. Implanted MSCs have been shown to migrate to injury sites in ischemic brain [25], infarcted myocardium [26], and bone fractures [27]; the local concentration of SDF-1α is upregulated after tissue injury, and its interaction with CXCR4 is critical for this process. Several studies have reported that MSCs lost CXCR4 expression after prolonged culturing or repeated subculturing [18][19][20][21]. We therefore hypothesized that enhancing CXCR4 expression could improve the therapeutic efficacy of transplanted MSCs by stimulating their migration to sites of injury.
We demonstrated that CXCR4 overexpression did not affect hUC-MSC surface marker expression or differentiation potential, but enhanced their migration in response to SDF-1α both in vitro and in vivo. We also showed that hUC-MSC overexpressing CXCR4 adhered to and proliferated on the scaffolds. Moreover, BDNF released from CGB scaffolds induced the differentiation of CXCR4-overexpressing hUC-MSCs into neurons, which contributed to the regeneration of brain tissue at the lesion boundary, also with the potential benefit for proliferation and differentiation of neural stem cell.
In conclusion, the present results indicate that hUC-MSCs overexpressing CXCR4 migrate in response to SDF-1α from CGB scaffolds to the site of brain injury. The transplanted CGB scaffold with hUC-MSCs CXCR4/GFP encapsulated can match the lsssion cavity of injured brain caused by TBI which was shown by HE staining. It was evidenced that the BDNF released from CGB scaffolds induced the differentiation of these hUC-MSCs into neurons. Also, in vivo, the encapsulated hUC-MSCs CXCR4/GFP can not only differentiated into neurons in the scaffold site but also in the lesion site border to replenish the apoptosis of neurons. These findings suggest that introducing hUC-MSCs CXCR4/GFP to the injury site via CGB scaffolds may be a potential therapeutic strategy for the treatment of TBI.