Exogenous Oligodendrocyte Progenitor Cell Transplantation Results in Structural and Functional Recovery in a Preterm Infant Cerebral White Matter Injury Rat Model


 BackgroundCerebral white matter injury (WMI) is the most common brain injury in preterm infants; it leads to motor and developmental deficits and is often accompanied by cognitive impairment. WMI is characterized by the loss of pre-myelinating oligodendrocytes. Regeneration therapies for preterm neonates with WMI are still in the preclinical phase, among which oligodendrocyte progenitor cell (OPC) transplantation is a promising approach. One promising approach for treating preterm infants is cell replacement therapy, in which lost cells are replaced by human OPCs (hOPCs) derived from human neural stem cells (hNSCs). MethodsIn this study, we developed a method to induce the differentiation of hNSCs into hOPCs. OLIG2+/NG2+/PDGFRα+/O4+ hOPCs were enriched and transplanted into the corpus callosum of a preterm infant WMI rat model. ResultsTransplanted hOPCs survived and migrated throughout the major white matter tracts. Morphological differentiation of transplanted hOPCs was observed. Histology and Magnetic resonance imaging (MRI) revealed lesioned structural repair. Electron microscopy revealed the re-myelination of the axons in the corpus callosum. The Morris water maze test revealed a recovery of cognitive function. ConclusionsOur study showed that transplantation of hOPCs derived from hNSCs is a viable therapeutic strategy for cerebral WMI. The results of our study contribute to the further development of cell therapeutic strategies.


Abstract Background
Cerebral white matter injury (WMI) is the most common brain injury in preterm infants; it leads to motor and developmental de cits and is often accompanied by cognitive impairment. WMI is characterized by the loss of pre-myelinating oligodendrocytes. Regeneration therapies for preterm neonates with WMI are still in the preclinical phase, among which oligodendrocyte progenitor cell (OPC) transplantation is a promising approach. One promising approach for treating preterm infants is cell replacement therapy, in which lost cells are replaced by human OPCs (hOPCs) derived from human neural stem cells (hNSCs).

Methods
In this study, we developed a method to induce the differentiation of hNSCs into hOPCs.

Results
Transplanted hOPCs survived and migrated throughout the major white matter tracts. Morphological differentiation of transplanted hOPCs was observed. Histology and Magnetic resonance imaging (MRI) revealed lesioned structural repair. Electron microscopy revealed the re-myelination of the axons in the corpus callosum. The Morris water maze test revealed a recovery of cognitive function.

Conclusions
Our study showed that transplantation of hOPCs derived from hNSCs is a viable therapeutic strategy for cerebral WMI. The results of our study contribute to the further development of cell therapeutic strategies.

Background
Cerebral white matter injury (WMI) is the most common pathological alteration-related brain injury in preterm infants. It is characterized by the loss of white matter [1]. WMI can lead to pediatric neurodevelopmental disorders, resulting in motor and developmental defects accompanied by cognitive de cits. This situation is always persistent from adolescence to adulthood [2]. Clinical evidence indicates that maternal-fetal infection and in ammation are mainly associated with WMI, where hypoxia-ischemia is thought to be the primary cause of neural cell damage [3].
Therapeutic strategies to promote neural regeneration for WMI are still under development. Previous studies have shown that cell transplantation, an approach to transplant neural cells to replenish or replace damaged or dysfunctional tissue, is a promising therapeutic strategy for WMI [4]. Human oligodendrocyte progenitor cells (OPCs; hOPCs) are considered potential candidates for cellular transplantation [5]. In the periventricular leukomalacia (PVL) rat model, transplanted hOPCs survived longer than 5 weeks and differentiated into oligodendrocytes expressing myelin basic protein (MBP).
Both motor and cognitive recovery were also observed in the rat model [6]. Several studies have shown that transplantation of hOPCs derived from the fetal brain, human induced pluripotent stem cells, or embryonic stem cells leads to myelination in animal models [7][8][9][10].
In this study, we established a neonatal rat model of WMI that mirrors the characteristics of WMI in preterm infants [11] to determine whether transplantation of hOPCs to neonatal rats with WMI improves neurobehavioral and cognitive functions as well as neuroprotection.

Methods
Human neural stem cell (NSC; hNSC) culture Aborted human fetuses, aged 7-9 weeks post-conception, were obtained from the Department of Obstetrics and Gynecology of the Sixth Medical Center of PLA General Hospital, Beijing, China, who had requested to terminate gestation and consented to donate the aborted fetus after being fully informed of the study, according to the guidelines approved by the hospital's ethics committee.

Animal models
This study was conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All protocols were approved by the Clinical Pharmacology Ethics Committee of the Sixth Medical Center of PLA General Hospital, Pediatrics Ethics Committee in Beijing (SCXK20170005). Sprague-Dawley rats (8-9 g) at postnatal day (P) 3 were randomly assigned into three groups (n = 9 in each group): normal control group rats, preterm infant WMI (PWMI) model group rats, and graft group rats (model + graft). Hypoxic ischemia (HI) was induced in model rats and graft rats, as previously described [14][15][16]. Brie y, the pups were cryoanesthetized by incubation on ice for 7 min, the right common carotid artery was severed from the accompanying vagus and sympathetic nerves and permanently ligated using a microelectrocauterizer, and the pups were returned to the dam for 1 h and then exposed to 6% O 2 for 90 min. Litter-matched sham controls were not subjected to carotid ligation or hypoxia.
Hematoxylin and eosin staining Seven days after the induction of HI, three rats were randomly selected from the model group. After anesthetization, the rat brain was perfusion-xed with 4% paraformaldehyde. Brain tissue was harvested, xed, and embedded in para n for hematoxylin and eosin staining. Pathologic conditions were observed in the brain dehydrated with gradient ethanol (70% ethanol for 2 h, 80% ethanol overnight, 90% ethanol for 2 h, 100% ethanol for 1 h, ), vitri ed with xylene, waxed, embedded in para n, and conventionally sliced into 5 µm tissue sections. Subsequently, the sections were stained with hematoxylin for 10 min, washed with water for 15 min, then stained with eosin for 1 min, and washed with water for 10 s.

OPC transplantation
Seven days after the induction of HI, OPCs were harvested from the ask and resuspended in phosphate balanced solution (PBS) at a density of 1 × 10 5 cells/µL, a total of 18 µL. Then, 3 µL (3 × 10 5 cells) of the suspension was injected into the right corpus callosum of each P9 Sprague-Dawley rat at the following stereotaxic coordinates from bregma: anteroposterior, -1.0 mm; mediolateral, 1.5 mm; deep level, -1.5 mm, using a 5 mL syringe for more than 10 min. The syringe was left in place for an additional 5 min, then gradually withdrawn, and the scalp was sutured and disinfected with 75% ethanol. The pups were anesthetized with ketamine and xylazine and placed in the stereotactic frame during the surgical procedures. After surgery, pups were placed on heating pads in individual cages until they recovered from anesthesia and returned to their mother until weaning on the 21st day. At that point, the male and female rats were moved to new cages and raised separately until they were sacri ced. Immunosuppressants were administered to all rats 3 days before surgery, and their administration continued throughout the study. The immunosuppressant cyclosporine (Sandimmun) was administered to all rats intraperitoneally once daily at a concentration of 10 mg/kg, beginning 3 days before surgery.

Immunohistochemistry of tissue sections
Three months after cell transplantation, after anesthetization, all rats were perfused with 0.1 M PBS and 4% PFA. The brains were removed, post-xed in 4% PFA overnight, and cryoprotected in 30% sucrose; frozen and coronal brain sections were cut using a Leica CM 1950 cryostat (Leica Microsystems; Leica CM, 1850) at a thickness of 14 µm. The immunostained slides were washed with PBS three times; brain sections were incubated in blocking solution (PBS + 10% normal goat serum + 0.3% Triton X-100) for 1 h at 37°C and then incubated with a primary antibody diluted in blocking solution overnight at 4°C. The next day, brain sections were washed three times in PBS, incubated with secondary antibodies (Alexa

Morris water maze test
The water maze testing took place in a 1.7 m (diameter) × 50 cm (height) hard black plastic tub that was lled with water (25 ± 1°C), with the bottom 45 cm above ground level. The black platform (12 cm diameter × 30 cm height) was submerged 1.5-2.0 cm below the water surface. The test was performed 3 months after transplantation (n = 6 per group). For the navigation trial, rats underwent two trials per day, and there were four different starting points for each trial for a total of 5 days. Rats escaping the platform were restricted to within 60 s and allowed to rest on the platform for 10 s. The space probe trial, with the platform removed, was conducted on the 6th day over 60 s. The escape latency, number of platform crossings, percentage of time spent in the platform quadrant, and percent distance traveled in the platform quadrant were recorded. Data were analyzed using a tracking program (ZS-001moriss Morris Water Maze Video Analysis System, ZS Dichuang, www.zslab1.com).
Mapping of human cell engraftment -The positions of all anti-human nuclei + cells were mapped on 14 µm coronal sections from − 5.04 to 2.04 bregma (random selection of a representative case of a graft rat; a total of seven sites and six slices were cut at each site).
Cell counting -A total of six rats in the graft group (six sections per rat, 14 µm coronal section samples of the corpus callosum, from − 3.5 to 1.5 bregma) were examined for cells expressing HNA together with either MBP, GFAP, OLIG2, or Ki67. All data shown represent means ± standard error of the mean (SEM)

Myelinated axon counts
Six samples from each group were analyzed using a quantitative evaluation method (G-ratio and ultrastructural grading scores). One hundred myelinated axons from each sample and 600 myelinated axons from each group were calculated. All data shown represent means ± SD.

Statistics
Analysis of variance (ANOVA) was performed using GraphPad Prism v5.0c for Macintosh (GraphPad Software, San Diego, CA, USA). Statistical tests for behavioral tasks were performed using one-way ANOVA followed by the Tukey's Honestly Signi cant Difference (HSD) test with Scheffe's, Bonferroni, and Holm's multiple comparison post-hoc tests. The data met the assumptions of the statistical tests used, and P-values less than 0.05 were considered signi cant.

Results
Histological analysis of the brains of PWMI model rats The PWMI model was established using neonatal rats 3 days after birth because they can simulate the clinical features of PWMI [11]. Brains were harvested 7 days after PWMI was established and were then prepared for analysis. Compared to the control, the brains of PWMI rats were atrophied. The white matter and corpus callosum of PWMI rats were thinner, and abnormal arrangements of neural bers were observed, indicating net necrosis. In the lesioned area, glial cell karyopyknosis and necrosis were obvious, and the number of glial cells and microglia increased (Figure 1).
MBP is mainly located in the myelin sheath, and its expression can re ect the damage of the white matter. Immunohistochemical staining showed that MBP expression in the corpus callosum and ipsilateral capsula externa was signi cantly decreased in the hemisphere subjected to a hypoxic-ischemic injury (right hemisphere) compared to the contralateral hemisphere (Figure 2a). Quantitative analysis based on staining area and signal intensity showed a signi cantly decreased MBP expression in the hemisphere with WMI (Figure 2b and 2c).
In vitro induction of hNSCs to obtain hOPCs hNSCs were isolated from the human fetal brain (gestational age, 8-10 weeks) and suspension cultured, allowing neurosphere formation (Figure 3a). We then induced differentiation of hNSCs into hOPCs and detected OPC-speci c markers to assess induction e cacy. The morphology of hOPCs derived from hNSCs showed a "bipolar" structure ( Figure 3d). As positive staining of OLIG2 represents the e cacy of induction [17], we detected the expression of OLIG2. Besides OLIG2, other OPC-speci c markers, including NG2, PDGFR-α, and O4, were detected. Immuno uorescence analysis showed positive staining of NG2/OLIG2/PDGFR-α/O4 (Figure 3b, 3c, 3e, and 3f), indicating the differentiation of hNSCs to hOPCs and, thus, that they were appropriate for cellular transplantation [18].
Survival, migration, and morphological differentiation of transplanted hOPCs in the rat brain Transplantation of hOPCs in PWMI rats was conducted via unilateral stereotaxic injection in the right hemisphere corpus callosum, and immunosuppression agents were administered to control the rejection reaction. The brains of PWMI rats were harvested 3 months after transplantation. Because of the relatively high rate of O4 + cells in our study, we did not purify the induced hOPCs with anti-O4 antibodies, as usually done in prior studies before transplantation [9,19].
The brains were further prepared for immunohistochemistry analysis using the human-speci c markers stem121 (cytosolic) and anti-human nuclear antigen (hNA) . Transplanted hOPCs migrated from the injection site to white matter tracts in the corpus callosum, external capsule, striatopallidal striae, and cortical layer 6 of both hemispheres (Figure 4a-4c). Furthermore, the morphology of engrafted cells developed from bipolar to multipolar shapes, similar to mature oligodendrocytes (Figure 4c). Quantitative analysis showed that the density of hNA + cells in the corpus callosum was 3986 ± 255 cells/mm 2 , and the number of hNA + cells decreased in the rostral and caudal regions (Figure 4d and 4e).
We further explored whether hOPC transplantation leads to myelination in PWMI rats. Transmission electron microscopy was used to analyze axonal ensheathment and myelin compaction. Compared to axons in normal rats, demyelination and deformed myelin sheaths were observed in most axons of the brains of PWMI rats (Figure 6a Model). In the hOPC transplantation group, demyelination was less common, although most myelin sheaths presented multilayers (Figure 6a Graft). In addition, we compared the G-ratio between different groups. G-ratio was de ned as the ratio of axonal diameter to total myelin-ensheathed ber diameter, and it indicated the severity of myelin sheath damage; the higher the score, the worse the myelin injury [21]. The results showed limited damage in the hOPC-transplanted group, as the G-ratio was signi cantly lower than that in the PWMI group (P < 0.05) (Figure 6c-6e).
We also implemented a grading system [22] for the quantitative evaluation of myelin damage (Table 1). A high score indicates damage severity. Consistent with the G-ratio, the scores of the hOPC-transplanted group were signi cantly lower than those of the PWMI group without hOPC transplantation (P < 0.05) (Figure 6b). Taken together, hOPC transplantation led to re-myelination in PWMI rats. Separation in myelin con guration 1 2 Interruption in myelin con guration 2 3 Honeycomb appearance 3 4 Collapsed myelin forming ovoids 4 5 Effects of hOPCS on spatial learning and memory in PWMI rats.
To determine whether hOPC transplantation rescues neurobehavioral de cits in PWMI rats, we performed the Morris water maze test to compare spatial learning and memory capabilities between the control, model, and transplantation groups. In the navigation trials, the transplantation group showed a decreased escape latency compared to the model group (Figure 7c). In the probe trials, PWMI rats showed fewer platform crossings, less time spent in the target quadrant, and longer distance to nd platforms than the control and graft groups, but there was no signi cant difference between the control and graft groups (Figure 7a, 7b, and 7d). The results above showed that hOPC transplantation could rescue spatial learning and memory de cits in PWMI rats.

Discussion
WMI is associated with preterm birth, and hypoxia-ischemia is thought to be one of the underlying pathological causes [23]. PWMI is characterized by the degeneration of pre-oligodendrocytes, leading to a lack of oligodendrocytes [3]. Therefore, supplementation with exogenous hOPCs is a reasonable therapeutic strategy.
The potential of OPC transplantation to restore neural functions was rst demonstrated in PVL model rats [24]. Several studies have shown the therapeutic effects of hOPC transplantation in various central nervous system (CNS) injury models, such as leukodystrophies, multiple sclerosis, and Pelizaeus-Merzbacher disease [4,6,19,25,26]. OPC transplantation showed therapeutic potential in a CNS injury animal model, mimicking brain hypoxia-ischemia [16].
Evidence of the effects of hOPC transplantation in a PWMI animal model is limited. In the present study, we determined the therapeutic potential of hOPC transplantation in PWMI rats. We found that hOPC transplantation can lead to re-myelination and neurobehavioral recovery. Consistent with previous reports, transplanted hOPCs showed explicit migration ability and differentiation capacity, but the mitotic potential was limited. The Morris water maze test revealed functional recovery following hOPC transplantation in PWMI rats. Exogenous OPC transplantation is always accompanied by a rejection reaction, resulting in the immunity of the host to engrafted cells and even death. We demonstrated that the application of immunosuppressive agents can lower the risk of rejection reactions in exogenous OPC transplantation. Kim et al. [6] showed that after intracerebroventricular transplantation, genetically modi ed hOPCs survived for more than 5 weeks in vivo and ameliorated neurobehavioral and cognitive de cits in PVL rats, further suggesting that hOPC replacement has the best therapeutic prospects for PWMI. Nevertheless, the technical operations required to obtain hOPCs are relatively complex. We wanted to obtain hOPCs through simple, fast, and e cient methods and intensively study their migration and differentiation trajectories in vivo to explore neuro-repair improvement in ultrastructure and behavior in PWMI.
In clinical trials, the safety and effectiveness of transplanted cells are crucial. First, we presented a reliable strategy that differentiates hOPCs from hNSCs. hOPCs derived using our protocol expressed OLIG2, NG2, PDGFRα, and O4, close to current de nitions of oligodendrocyte progenitors [27,28]. The acquisition of highly enriched OPC preparations allowed us to transplant cells in the absence of isolation, ensuring cell viability and health as much as possible.
In the rst 3 days of the model, a large number of in ammatory factors were released, and brain edema did not subside [29,30], which is not conducive for the normal survival of the transplanted cells, and research has shown that transplantation attenuates brain injury at 3-10 days [31,32]. In the present study, hOPC transplantation was performed 7 days after HI induction. Based on the injury sites responsible for critical symptoms, hOPCs were injected into the injured lateral corpus callosum using a stereotactic apparatus, and transplantation was performed in situ to achieve accurate repair of the damaged site.
We examined the fate of engrafted hOPCs in the rat corpus callosum after 3 months of transplantation; the cells were effectively implanted, and hOPC populations survived well and were successfully integrated into the host tissue, consistent with the ndings of previous studies [26,33]. We found that engrafted hOPCs could greatly migrate, and the trajectory of migration was consistent with the lesions of PWMI.
Transplanted cells develop to have a more mature multipolar morphology. The overwhelming majority of the hOPCs developed into more mature stages and expressed CC1, and a small fraction expressed MBP. Ultrastructural electron microscopy showed a compact myelin sheath that wrapped the axon, while some cells remained in the oligoprogenitor stage and expressed OLIG2, hardly differentiating into astrocytes, which corroborates the results of previous studies [24,34]. Less than 7% of hOPCs within the migratory domains were positive for the mitotic marker Ki67 after 3 months, indicating that cells could still differentiate and proliferate. There was no evidence of tumorigenesis or areas of graft overgrowth in implanted hNSC-derived OPCs.
Three months post-transplantation, rats were tested using the Morris water maze to assess behavioral consequences. Consistent with the organizational analysis, PWMI rats showed de ciencies in spatial learning and memory, which were substantially ameliorated by hOPC transplantation. This implies that hOPCs play an important role in white matter recovery, resulting in improved cognitive function.

Conclusions
Our study clearly showed the key role of hOPCs in white matter recovery and their contribution to the restoration of the cognitive function. Therefore, it is suggested that transplantation of hOPCs restores neurobehavioral disorders as well as cognitive de cits of PWMI animals by preventing neuron demyelination in periventricular white matter. Our ndings will further contribute to the improvement of cellular therapeutic strategies.      plot graph of the three groups showing the mean G-ratio distribution. One-way ANOVA test. *P < 0.05, ***P < 0.001.

Figure 7
The effects of transplanted OPCs on behavioral recovery. (A) Representative sample paths from the maze trials. (B) Probe trials were performed 4 h after the last maze trails on navigation day 5; time spent in the target quadrant was monitored. One way ANOVA test, *P < 0.05; ns, not signi cant. (C) Morris water maze task was performed three months later to test the spatial learning ability of control (n = 6), model (n = 6), and graft (n = 6) groups, as shown by the time (escape latency) to nd the submerged platform at navigation days 1 to 5. Two-way ANOVA test, *P < 0.05, **P < 0.01, ***P < 0.001. (D) Probe trials were conducted to monitor the number of crossing platforms in the three groups. One way ANOVA test, *P < 0.05, ***P < 0.001; ns, not signi cant.