Human Umbilical Cord-Derived Mesenchymal Stem Cells Improve Learning and Memory Function in Hypoxic-Ischemic Brain-Damaged Rats via an IL-8-Mediated Secretion Mechanism Rather than Differentiation Pattern Induction

Along with improvements in medical technology and therapeutic ability, the morbidity of hypoxic-ischemic brain damage (HIBD) has increased each year [1], demonstrating the need for improved treatment methods. Studies by other groups [2,3,4] and our research group [5,6] have consistently demonstrated that mesenchymal stem cells (MSCs) transplanted from bone marrow represent a useful therapeutic strategy for treating neurological disorders. However, in many studies, the sources of MSCs have been limited to animals, with a long path to human clinical applications. Human umbilical cords are an attractive source of MSCs, as they display many advantages compared to other stem cell types, such as embryonic stem cells, in terms of reduced tumourigenicity, immunogenicity, and ethical issues. They also exhibit increased proliferation compared to bone marrow derived-MSCs and are considered more valuable for transplantation in stem cell therapy and research studies [7,8]. However, to the best of our knowledge, it is unclear whether hUC-MSCs exhibit better biological characteristics and a higher potential for neural differentiation than other MSCs.

According to the accumulated evidence, the two primary mechanisms considered to underlie the therapeutic effects of MSCs in neurological disorders include the facilitation of the paracrine secretion of cytokines and the induction of neural differentiation patterns. Some studies have demonstrated that MSCs directly release or promote the release of various cytokines, including chemokines and growth and trophic factors, from the injured tissues [9]. These factors have been shown to ameliorate the inflammatory response [10], and promote neurogenesis and angiogenesis in the injured microenvironment [11,12], and improve neurological function in the brains of injured animals. However, some studies have supported a different perspective regarding the potency of MSCs to induce neuronal differentiation. It has been proposed that implanted MSCs can differentiate into mature neurons and play a role in cell replacement. Nevertheless, fewer than 20% of MSCs undergo neural differentiation in vivo [13,14]. Therefore, some research groups have attempted to improve this differentiation rate and thus increase the therapeutic potential of MSCs in various animal models by inducing these cells in vitro [15,16,17]. Some studies have suggested that hUC-MSCs improve neurological disease by either enhancing a paracrine effect or inducing differentiation patterns [18,19]. However, to the best of our knowledge, very few reports have addressed which mechanism participates in the therapeutic effects of these cells during HIBD.

In the current study, we first measured the biological characteristics and the neural differentiation and redifferentiation (neural differentiation for the second time) potential of hUC-MSCs in vitro. We then transplanted undifferentiated, differentiated and redifferentiated hUC-MSCs into HIBD rats to identify any therapeutic advantages in vivo. Finally, we investigated a possible mechanism by which hUC-MSC-secreted IL-8 might help promote the recovery of learning and spatial memory capabilities in HIBD rats.

hUC-MSCs were supplied by the Chongqing stem cell bank. All protocols involving human subjects were approved by the Ethics Committee of the Children's Hospital of Chongqing Medical University (Approval number: 024/2013). The cells were cultured in Dulbecco's modified Eagle's medium/ F12 (DMEM/F12; 1:1; Gibco, Carlsbad, CA, USA) supplemented with 10% FBS (Gibco), 100 units/ml penicillin, and 100 µg/ml streptomycin at 37°C in a humidified incubator containing 5% CO₂. When the cells reached 90% confluence, they were digested using trypsin/EDTA (Gibco) and reseeded in new culture dishes. Unless indicated otherwise, all hUC-MSCs used in the subsequent experiments were from passage 5 to passage 10.

For the cytogenetic analysis, hUC-MSCs with 50% confluence in 75-cm² flasks (Corning Costar Corp, NY, USA) at various passages (P2, P10, and P15) were incubated at 37°C in colcemid (Irvine Scientific, CA, USA) at a final concentration of 1µg/ml for 3.5 h to induce cell cycle arrest at metaphase. The cells in metaphase were karyotyped in accordance with the recommendations of the International System for Human Cytogenetic Nomenclature (ISCN 2009).

To stain the hUC-MSCs, cultured cells at passage 2 were trypsinised using trypsin/EDTA, washed with PBS and resuspended at a concentration of 10⁶ cells /ml. The cell suspensions were incubated with different antibodies (1:100 dilution) for 30 min at RT in the dark and analysed using a FACSCanto II system (BD Biosciences). The CD34-PE, CD29-PE, CD45-FITC, HLA-ABC PE and HLA-DR-FITC primary antibodies were purchased from eBioscience (San Diego, CA, USA). The CD44-FITC, CD90-PE and CD105-PE primary antibodies were purchased from BD Biosciences (San Jose, CA, USA), and the recommended isotype control for PE and FITC were purchased from 4A Biotech Co., Ltd. (Beijing, China).

All chemicals used for the modified neuronal medium (MNM) were purchased from Sigma (St. Louis, MO, USA). Neuronal differentiation and redifferentiation were induced according to a previously described method [20]. Briefly, hUC-MSCs at passage 5 were seeded at 4x10⁵/well in a 6-well culture plate (Corning Costar Corp). Cells were pre-induced in all-trans-retinoic acid (ATRA) (1 µmol/L) for 24 hours, and the undifferentiated hUC-MSCs were induced using MNM for 24 hours to achieve neuronal differentiation status (Dif). The MNM was replaced with complete culture medium (DMEM/F12 containing 10% FBS) for 24 hours following neuronal differentiation. Cells were then transferred to MNM for another 24 hours to achieve neuronal redifferentiation status (Re-Dif). Cells presenting neuron-like morphology were identified as positive when calculating differentiation efficiency, and the neural induction rate was measured using Image J software. Three independent experiments were performed and evaluated.

In brief, total RNA was extracted using an RNA Extraction Kit (Bioteke Co., Ltd, Beijing, China). RNA was reverse-transcribed using a PrimeScript RT reagent Kit according to the manufacturer's protocol. Real-time PCR reactions were performed using a RealMasterMix kit (SYBR Green; Takara Biomedicals, Osaka, Japan) and a Bio-Rad Real-Time PCR system. All samples were normalised to the expression level of β-actin. The primer sequences are listed in Table 1.

Total protein was extracted using RIPA buffer containing PMSF (Bioteke Co. Ltd). SDS-PAGE gels of different concentrations (Beyotime, Shanghai, China) were prepared, and proteins of interest were electrophoretically separated according to their molecular weight. The anti-NSE, anti-MAP2, anti-β-tubulin-III and anti-IL-8 primary antibodies were purchased from Abcam (Cambridge, MA, USA). The anti-p-JNK, anti-JNK, anti-PECAM-1 and anti-β-actin primary antibodies and the HRP-conjugated secondary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The proteins of interest were visualised using Luminata Crescendo Western HRP Substrate (Millipore, Billerica, MA, USA) and the Syngene G-Box Imaging System.

IL-8, TIMP-1, CCL2, CXCL1 and VEGF production were examined using commercially available ELISA kits. The IL-8, CXCL1 and VEGF commercial ELISA kits were purchased from Raybiotech (Norcross, GA, USA), and the TIMP-1 and CCL2 commercial ELISA kits were purchased from 4A Biotech Co., Ltd. The experimental procedure was performed in accordance with the corresponding reagent kit protocol. The optical density at 450 nm was recorded using a microtiter plate reader (Thermo).

The siIL-8 lentivirus was targeted to human chemokine ligand 8 (CXCL8, NM-000584.3) and was constructed by Neuron Biotech (Shanghai, China). hUC-MSCs at a confluence of 50-60% in a 6-well plate were infected with the siIL-8 lentivirus at a multiplicity of infection (MOI) of 2.5. After 48 h, green fluorescent protein (GFP) was detected under a microscope, and 0.5µg/ml puromycin was added to select for silL-8-infected hUC-MSCs. A single siIL-8-infected hUC-MSC was seeded into a 96-well plate to obtain stable cell lines for further experiments. The siIL-8-hUC-MSC stable cell line carried a sequence specific for hIL-8 (gccaaggagtgctaaagaact) for hIL-8.

Total RNA was extracted from different experimental groups using TRIzol reagent (Invitrogen, Carlsbad, CA). After purification and quality assessment, the bio-tagged cRNAs were fragmented into strands of 35-200 bases in length according to the manufacturer's protocol. Then, the cRNAs were hybridised to the Human Genome U133 Plus 2.0 array, which contains approximately 54675 probes, for 16 h at 45°C. All samples were prepared in triplicate. The hybridisation data were analysed using Affymetrix® GeneChip® Command Console® Software. The expression profiles were examined via hierarchical cluster analysis using Cluster version 2.11 software. Affymetrix® NetAffx™Analysis Center software was used for gene ontology analysis. The data were considered to be significantly different when the expression was altered by at least two-fold, and the differentially expressed genes were selected for further analyses. The microarray data was submitted to Gene Expression Omnibus (GEO), and the submission number is GSE65631.

All animal experiments were approved by the Ethical Committee of Chongqing Medical University. Sprague Dawley (SD) rats were purchased from the Experimental Animal Center of Chongqing Medical University [Certificate: SCXK (Yu) 2012-0001]. All rats were housed in a specific pathogen-free (SPF) laboratory. The animal housing rooms were maintained at a constant temperature (25°C) in a 12 h/12 h light-dark cycle, and the animals were provided with access to food and water ad libitum.

The HIBD model was established according to the Rice model [21]. After 5 days, 2×10⁵ cells suspended in 5 µL of sterile PBS were transplanted into the HIBD rats (n=76, 12 days old) via intracerebroventricular injection. The injection site was located 1.2 mm posterior to the bregma, 1.4 mm to the left of the lambdoid suture and at a needle depth of 3.6 mm. The injection rate was 1µL/min, and a microinjection instrument was used to control the injection, which lasted a total of 5 min. The needle remained in the lateral ventricle for 2 min. The HIBD control group (n=27) was injected with the same volume of sterile PBS. The sham group (n=18) received the sham surgery without hypoxia-ischemia or cell transplantation.

The pups were sacrificed two days after cell transplantation, and the brain tissues from the sham group (n=6), HIBD group (n=6), hUC-MSC group (n=6), GFP-hUC-MSC group (n=3) and siIL-8-hUC-MSC group (n=3) were collected. For haematoxylin-eosin (HE) staining, 4-µm continuous coronary slices on glass slides were stained with HE, and the pathology and morphology of the cells were analysed using a light microscope (Nikon). For double-immunofluorescence labelling, 30% sucrose paraformaldehyde was applied to dehydrate the tissue sections. These sections were immunostained with monoclonal mouse anti-BrdU (1:500 dilution, Sigma) and rabbit anti-PECAM-1 (1:100 dilution) antibodies at 4°C for 20 h. The primary antibodies were visualised using Alexa Fluor 488 chicken anti-mouse IgG and Cy3-conjugated Affinipure Goat Anti-rabbit IgG. The fluorescent antibodies against BrdU and PECAM-1 were purchased from Invitrogen or Proteintech (Wuhan, China). Images were captured via laser scanning confocal microscopy (Nikon).

The Morris water maze test was performed as previously described [5,6]. Briefly, four weeks after cell transplantation, the rats from the sham group (n=9, 6 weeks old), HIBD group (n=18), hUC-MSC group (n=17), Dif group (n=12), Re-Dif group (n=10), GFP-hUC-MSC group (n=8) and siIL-8-hUC-MSC group (n=8) were subjected to the Morris water maze test. The procedure lasted a total of six days. The first day consisted of the visual platform test. Each experimental rat underwent four trials of 60 s in four quadrants, and the time (escape latency) and distance (path length) required to locate the platform were recorded. During the next four days of training, the average time to locate the platform on the four trials was recorded. On the final day, the probe trial was performed by removing the platform, and the number of times that the rats crossed the former platform location in 60 s was recorded.

This behavioural test was performed in a 1 m² open-topped box constructed of black Plexiglas. During the pretraining phase, the animals (n=42, 6 weeks old) were placed in the empty box to adapt to the environment for 10 min for two consecutive days. During the sample phase (day 3), four different shaped toys were placed at the four corners of the box, and the animals were allowed to explore the novel objects for 5 min. During the delay period (5 min), all objects were cleaned with alcohol to remove olfactory cues and any sawdust, and two adjacent toys were exchanged (excluding the two diagonal toys). During the test phase (3 min), the animals were placed in the box again, and the time spent exploring the four toys was recorded. In the subsequent analysis, the ratio of the time spent exploring the two novel objects (changed position) to the time spent exploring the two familiar objects (unchanged position) was compared.

All chemicals used for the electrophysiology experiments were purchased from Sigma. The rats (n=17, 8 weeks old) were anesthetised via intraperitoneal injection of pentobarbitone (60 mg/kg). The hippocampal slices were transferred to a chamber placed on the stage of an upright microscope (Nikon E600FN, Tokyo, Japan) and perfused with oxygen-saturated artificial cerebrospinal fluid (ACSF) at 35°C. The recording and stimulating electrodes were placed in the CA1 area of the hippocampal slice. Extracellular field excitatory postsynaptic potentials (fEPSPs) were evoked using a single test stimulus, and the test stimulus current was adjusted to half of the maximal amplitude to establish a steady baseline (≥20 min) before tetanic stimulation. To induce LTP, high frequency stimulation (HFS) was applied, and the effects were recorded every 2 min for 60 min.

The results are presented as the means ± S.E.M. One-way analysis of variance (ANOVA) was used for multivariate data analysis, and Student's t test was used for statistical analysis. A value of P≤0.05 was considered to be significant.

The hUC-MSCs cultured on a culture dish surface exhibited a spindle-shaped fibroblast-like morphology and were arranged in a spiral (Fig. 1A). The cells at passages 10 and 15 demonstrated no apparent morphological differences compared to the cells at passage 2 (Fig. 1A middle and right). To determine the karyotypic stability of increasing cell passages, trypsin and Wright's stain-banding analysis were performed on hUC-MSCs at passages 2, 10 and 15. As shown in Fig. 1B, chromosomal aberration of the hUC-MSCs was not observed during 15 generations, suggesting that these hUC-MSCs maintained a normal chromosomal structure for at least 15 passages. Subsequently, cell-surface antigen expression was further examined via flow cytometry. The data displayed in Fig. 1C show that the hUC-MSCs positively expressed CD29, CD44, CD90, CD105, and HLA-ABC but not CD34, CD45 or HLA-DR. Taken together, these findings suggest that the hUC-MSCs in the present study exhibit the morphological characteristics of MSCs and express MSC-related surface markers, which verified their suitability as transplantation cells for subsequent experiments.

To investigate whether the hUC-MSCs displayed neuronal differentiation and redifferentiation potency, we subjected hUC-MSCs to a differentiation- and redifferentiation-inducing protocol. As shown in Fig. 2A (middle panel), following neuronal differentiation (Dif group), the cells exhibited morphology similar to hUC-MSCs. However, the cells subjected to neuronal redifferentiation (Re-Dif group) displayed a dramatically neuron-like morphology, with a round, bright cell body and neurite-like extensions. The neural induction rate was nearly 80% (Fig. 2A right).

Next, the expression levels of neural-specific markers, including NSE, MAP2 and β-tubulin-III, were examined in the three induced groups. Fig. 2B, 2D and 2F reveal that the mRNA expression levels of NSE, MAP2 and β-tubulin-III were gradually increased following the induction of differentiation and redifferentiation. In response to the redifferentiation treatment, we found that the levels of the three markers were sharply increased (**P<0.01, ***P<0.001, one-way ANOVA). The changes in the protein expression levels of these three neural-specific markers were completely consistent with the changes in their mRNA expression (Fig. 2C, 2E, 2G), suggesting that the hUC-MSCs in the present study were successfully induced to differentiate into neuron-like cells expressing high levels of neural markers via an in vitro redifferentiation method.

To confirm the therapeutic efficacy of hUC-MSC transplantation, comparison experiments consisting of brain histopathological evaluation, rat behavioural tests and electrophysiological examination of brain slices were performed in vivo. Based on HE staining, the periventricular brain tissue displayed uneven, diffuse staining and nuclear condensation (Fig. 3A middle), and the cell morphology in the HIBD pups was altered compared to the sham rats. In contrast, hUC-MSC transplantation partially alleviated these changes (Fig. 3A right). To analyse the spatial learning and memory ability of the HIBD rats, the Morris water maze test was performed four weeks following hUC-MSC transplantation. On the visible platform tests (Day 1), no difference in the escape latency or the path length was detected (p>0.05, one-way ANOVA), which indicated that the HIBD model and cell transplantation did not affect the movement or visual ability of the rats (Fig. 3B-3C). During the hidden platform training (Day 2-5), all of the groups demonstrated gradual decreases in the escape latency. The HIBD group exhibited a longer escape latency than the sham group (*P<0.05, **P<0.01, ***P<0.001, Sham vs. HIBD, one-way ANOVA), whereas the escape latency was significantly reduced in the HIBD group transplanted with hUC-MSCs (#P<0.05, HIBD vs. hUC-MSCs, one-way ANOVA) (Fig. 3D). On the probe trial tests (Day 6), the hUC-MSC-treated group entered the platform quadrant significantly more than the HIBD group, although less than the sham group (***P<0.001, Sham vs. HIBD; *P<0.05, HIBD vs. hUC-MSCs, Sham vs. hUC-MSCs) (Fig. 3E). We also examined the spatial configuration memory of the rats for objects using an object-in-place task. The data in Fig. 3F show that the HIBD rats spent less time investigating the two objects at the new location than the sham rats (***P<0.001, Sham vs. HIBD); however, the novel object exploration time was significantly increased following hUC-MSC treatment (*P<0.05, HIBD vs. hUC-MSCs).

To examine whether hUC-MSC transplantation facilitates hippocampal synaptic plasticity, we recorded fEPSPs in the hippocampal CA1 region of brain slices. The basal level of fEPSPs during the 20 min pre-HFS and the 60 min post-HFS were recorded in the three groups. At 30 min post-HFS application, the incremental mean slope of the fEPSPs in response to HFS in the HIBD group (110.1%±11.7% of baseline, n=5) was lower than that in both the hUC-MSC (153.1%±17.5% of baseline, n=6) and sham groups (167%±21.8% of baseline, n=6) (*P<0.05, HIBD vs. hUC-MSCs; ***P<0.001, HIBD vs. Sham) (Fig. 3G). The fEPSP level induced by HFS remained constant until the end of the experimental recording (60 min) in the three groups. The above results suggest that hUC-MSC transplantation improved spatial learning and memory function and enhanced hippocampal synaptic plasticity in HIBD rats.

To evaluate the effects of the differentiated and redifferentiated cells on biological function in vivo, we transplanted differentiated or redifferentiated hUC-MSCs into the HIBD rats. On the Morris water maze test, no significant difference in the escape latency or the path length during the Day 1 test was detected between the three groups (P>0.05, one-way ANOVA) (Fig. 4A-4B). During the hidden platform acquisition training phase (from Day 2-5), the escape latency of all three groups gradually decreased, demonstrating that the rats in each group showed spatial memory for the underwater platform, and no difference between the hUC-MSC, Dif and Re-Dif groups was detected (P>0.05, one-way ANOVA) (Fig. 4C). During the probe trial (Day 6), although the number of platform quadrant entries displayed a slight increase in both the Dif and Re-Dif groups, no significant difference between the three groups was detected (P>0.05, one-way ANOVA) (Fig. 4D). On the object-in-place task, the three groups exhibited similar memory for the novel objects, and no significant differences were observed (P>0.05, one-way ANOVA) (Fig. 4E). These results suggested that similar to undifferentiated hUC-MSCs, both differentiated and redifferentiated hUC-MSCs promote neural function recovery in HIBD rats; however, the Dif and Re-Dif cells did not confer greater therapeutic advantages than the undifferentiated hUC-MSCs, even though they expressed higher levels of neuronal markers in vitro.

To explain the high in vitro neural potential and the lack of therapeutic advantages of Dif and Re-Dif hUC-MSCs in vivo, we employed a gene expression profiling microarray to determine the differences in gene expression between the undifferentiated, Dif and Re-Dif hUC-MSCs. According to genome cluster analysis and the scatter plot, the Dif and Re-Dif groups displayed similar gene expression patterns, which were very different from that of the undifferentiated hUC-MSCs (Fig. 5A-5B). Then, we screened the consistently altered genes in the three groups (consistently increased or decreased) and selected 483 genes that displayed significantly different expression according to the cluster analysis. As shown in Fig. 5C, following neural differentiation and redifferentiation, the Dif group exhibited little change, and the Re-Dif group showed the opposite gene expression pattern to the undifferentiated hUC-MSCs group. Furthermore, these differentially expressed genes were analysed via gene ontology and biological function analyses. The histogram in Fig. 5D shows the top 15 biological functions of the differentially expressed genes, which included signal transduction, development, negative regulation of cell proliferation, cell cycle, and nervous system development. KEGG-pathway analysis was applied to determine the possible pathways that participate in the neural differentiation and redifferentiation process. Of the many possible signalling pathways, the cytokine-cytokine receptor interaction pathway displayed the most significant difference (Fig. 5E). The above data indicate that the induction of differentiation and redifferentiation in hUC-MSCs is a gradual process that results in distinct gene expression patterns and that the cytokine-cytokine receptor interaction pathway is primarily involved in this process.

The data from the gene expression profiling chip assay suggested that cytokines from the hUC-MSCs play an important role in improving the recovery of learning and memory function in HIBD rats following hUC-MSC transplantation. Therefore, to explore the molecular mechanism underlying the therapeutic effect of hUC-MSCs, the expression levels of various cytokines were analysed according to the gene expression profiling chip assay data. We found that the hUC-MSCs expressed high levels of many cytokine genes, such as TIMP-1, IL-8, and CXCL1 (Fig. 6A). Furthermore, ELISA was applied to verify the gene chip results. As shown in Fig. 6B, IL-8 was the cytokine that showed the highest level of secretion by hUC-MSCs (***P<0.001, IL-8 vs. TIMP-1, one-way ANOVA). In addition, we detected a change in the IL-8 protein level in rat brain tissue and, surprisingly, found that both the levels of hIL-8 secretion and IL-8 protein expression were significantly increased in the hUC-MSC-transplanted brain tissue (*P<0.05, HIBD vs. hUC-MSCs, Sham vs. hUC-MSCs, one-way ANOVA) (Fig. 6C-6D), suggesting that IL-8 secretion by hUC-MSCs may participate in the restoration of learning and memory function in the HIBD rats.

IL-8 is an important factor that promotes angiogenesis and participates in many physiological processes. Angiogenesis was evaluated to elucidate the role of IL-8 in vivo. As shown in Fig. 7A-7B, the levels of CD31 protein expression and VEGF secretion in the hUC-MSC group were significantly higher than those in the HIBD group (**P<0.01, HIBD vs. hUC-MSCs; **P<0.01, Sham vs. HIBD). Moreover, the immunofluorescence results revealed that the number of endothelial cells displaying CD31 and BrdU double-labelling was significantly increased in the hUC-MSC group compared to the HIBD group (*P<0.05, HIBD vs. hUC-MSCs, Sham vs. HIBD, one-way ANOVA) (Fig. 7C-7D). Moreover, the level of p-JNK protein expression was clearly increased in the HIBD brain tissue following hUC-MSC transplantation compared to the sham and HIBD brain tissue, even though no significant difference in the level of total JNK protein expression was detected between the three groups (Fig. 7E). These data indicate that IL-8 secreted by hUC-MSCs may regulates angiogenesis in the hippocampus of HIBD rats via the JNK signalling pathway.

To verify the hypothesis that IL-8 secreted by hUC-MSCs regulates angiogenesis in the hippocampus of HIBD rats via the JNK signalling pathway, we generated stable cell lines of siIL-8-hUC-MSCs and GFP-hUC-MSCs and transplanted them into the HIBD rats. For two days following transplantation, the levels of hIL-8 secretion and IL-8 protein expression in rat brain tissue were statistically significantly lower in the siIL-8-hUC-MSC group than in the GFP-hUC-MSC group (***p<0.001, Student's t test) (Fig. 8A-8B). During subsequent follow-up, the levels of VEGF secretion and CD31 protein expression were significantly reduced in the siIL-8-hUC-MSC group compared to the GFP-hUC-MSC group (*P<0.05, Student's t test) (Fig. 8C-8D). Immunofluorescence results revealed that the number of endothelial cells displaying CD31 and BrdU double-labelling was significantly decreased in the siIL-8-hUC-MSC group (*P<0.05, Student's t test) (Fig. 8E-8F). In addition, the levels of p-JNK protein expression clearly decreased following siIL-8-hUC-MSC transplantation, although total JNK protein expression was not significantly different (Fig. 8G). On the Morris water maze test, the siIL-8-hUC-MSC group exhibited a longer escape latency (Fig. 8H lower left), although these rats spent less time in the platform quadrant than the GFP-hUC-MSC group (*P<0.05, Student's t test) (Fig. 8H lower right). The above results further support our hypothesis that IL-8 secreted by hUC-MSCs regulates angiogenesis in the hippocampus of HIBD rats via the JNK signalling pathway.

hUC-MSCs are a promising type of MSC with clinical applications. According to the ISCT standards, which include the criteria for defining human MSCs [22], the hUC-MSCs in the present study exhibited the morphological characteristics and immunophenotype of hMSCs, which was consistent with other studies [23,24]. Furthermore, the hUC-MSCs exhibited excellent proliferative capacity and chromosomal stability until passage 15, indicating that they are safe for clinical application. The multi-lineage differentiation potential of these was verified previously [25]; therefore, we did not repeat these experiments in the current study.

MSCs display the ability to cross the germinal layer to differentiate into neuron-like cells and produce greater functional results than undifferentiated stem cells, suggesting that MSCs represent excellent transplantation cells for cell-based therapy and tissue engineering [26,27]. Additionally, our previous studies have demonstrated that differentiated and redifferentiated BMSCs from rats exhibited neural phenotypes and biological functions [20,28]. Whether hUC-MSCs, a special type of MSC, display neuronal differentiation potency was unclear. To address this issue, hUC-MSCs were subjected to a differentiation induction programme. Surprisingly, the differentiated cells did not exhibit a neuron-like morphology, although their neural marker levels were higher than those of the undifferentiated hUC-MSCs. This interesting finding prompted us to conduct a redifferentiation induction protocol on the hUC-MSCs. Interestingly, the redifferentiated cells exhibited dramatic morphological changes and expressed much higher levels of neural markers than the differentiated hUC-MSCs, suggesting that the hUC-MSCs used in the present study display in vitro neuronal differentiation potency. Moreover, because the differentiated hUC-MSCs maintained a similar morphology to the undifferentiated hUC-MSCs, hUC-MSCs from humans may display more strictly conservative properties than MSCs derived from rats. The initial differentiation was not sufficient to counteract the conservative properties and alter the cytoskeleton, although redifferentiation induced morphological changes and increased neural marker expression. Further studies should be conducted to elucidate the precise mechanism underlying these changes.

To evaluate the therapeutic effects of the three states of hUC-MSCs, differentiated, redifferentiated or undifferentiated hUC-MSCs were transplanted into HIBD rats. Because previous studies have reported that ischemia-related gene expression occurs in a time-dependent manner at 72 h or later following ischemia [29], we defined an effective therapeutic window as consisting of no more than 10 days [30,31]. Intracerebral transplantation may serve as the most appropriate method for treating nervous system diseases because this method results in the most extensive grafting of the implanted cells to the injured tissue. Four weeks after transplantation, neurobehavioral function was assessed. Our data revealed that the hUC-MSCs in all the three states promoted the recovery of neural function in HIBD rats, indicating that hUC-MSC transplantation showed an advantage in HIBD rats, which was consistent with a previous study [18]. However, the results also showed that the differentiated and redifferentiated cells were not able to produce more significant therapeutic advantages than the undifferentiated hUC-MSCs. Some studies have shown that inducing differentiation is not necessary; for example, undifferentiated adipose-derived stem cells (ASCs) were effective in promoting nerve regeneration [32], and the undifferentiated and differentiated cells improved function to a similar extent [33]. Our results verified that inducing differentiation and redifferentiation neural patterns may not be a necessary component of the treatment process.

To explain the high in vitro neural potential and the lack of therapeutic advantages of hUC-MSCs in vivo, a gene expression profiling microarray was used in the present study. Based on analysis of the chip results, the induction of differentiation and redifferentiation in hUC-MSCs was clearly a gradual process that produced distinct gene expression patterns. Importantly, the cytokine-cytokine receptor interaction pathway primarily participated in this process.

The regulation of the injury microenvironment by secreted cytokines and the induction of differentiation are the two primary mechanisms thought to contribute to the therapeutic effects of MSCs [34,35]. As mentioned previously, the differentiated and redifferentiated cells did not further restore biological functions compared to undifferentiated hUC-MSCs in the current study, and the results of gene expression profiling suggested that cytokine production in hUC-MSCs merited further investigation. We found that IL-8 was the most abundantly secreted cytokine by hUC-MSCs in vitro. Surprisingly, the levels of both hIL-8 secretion and IL-8 protein expression were significantly increased in the hUC-MSC-transplanted brain tissue, suggesting that IL-8 secretion by hUC-MSCs may improve the restoration of learning and memory function in HIBD rats. IL-8 is an important angiogenic factor [36] that enhances endothelial cell proliferation and survival and that regulates angiogenesis [37,38]. Several independent reports have demonstrated that IL-8 stimulates VEGF expression via the MAPK signalling pathway [19,39]. In our study, we confirmed that hUC-MSC treatment significantly increased the expression levels of CD31 and VEGF in the HIBD brain and promoted angiogenesis in the hippocampus. Moreover, the level of p-JNK protein expression was clearly increased in the hUC-MSC-treated group. These data indicate that the IL-8 secreted by hUC-MSCs may regulates angiogenesis in the hippocampus of HIBD rats via the JNK signalling pathway.

To validate the effects of IL-8 from hUC-MSCs on hippocampal angiogenesis via the JNK pathway, we established a stable siIL-8-hUC-MSC cell line. Following transplantation into the HIBD rats, we found that levels of both hIL-8 secretion and IL-8 protein expression were significantly decreased in the brain tissue. Moreover, levels of CD31 expression, VEGF release, p-JNK protein expression, and hippocampal angiogenesis were all significantly decreased by siIL-8-hUC-MSC transplantation, thereby suggesting that the decrease in hUC-MSC-secreted IL-8 can attenuate angiogenesis in the hippocampus of HIBD rats via the JNK pathway. Furthermore, the GFP-hUC-MSC transplantation group demonstrated greater benefits for treatment than the siIL-8-hUC-MSC group, which verified the hypothesis that hUC-MSC transplantation improves the recovery of learning and memory functions via IL-8 secretion.

In summary, we demonstrated that hUC-MSCs exhibit the biological characteristics of stem cells and display neural directional differentiation potential. hUC-MSC transplantation improved the recovery of learning and memory functions via IL-8 secretion, which enhanced angiogenesis in the hippocampus of HIBD rats, but this treatment did not induce a differentiation pattern. These findings might provide the experimental foundation for the clinical application of hUC-MSCs to hypoxic-ischemic disorders.

HIBD (Hypoxic-ischemic brain damage); MSCs (Mesenchymal stem cells); hUC-MSCs (Human umbilical cord-derived mesenchymal stem cells); IL-8 (Interleukin-8); Dif (Differentiation status); Re-Dif (Redifferentiation status); fEPSPs (Field excitatory postsynaptic potentials); HFS (High frequency stimulation).

The reported work was supported by grants from the National Natural Science Foundation of China (Nos. 30830106 and 81271385) and the Stem Cell Therapy for Special Study of Children's Hospital of Chongqing Medical University (SCT-201203).

The authors declare no conflict of interest.