hUC-MSCs transplantation promotes neurological recovery by ameliorating spinal cord injury and hippocampal synaptic remodeling in female mice

Abstract

Introduction

Spinal cord injury (SCI) induces not only local spinal cord pathology but also distal brain dysfunction, causing multi-dimensional impairments in motor, emotional, and cognitive functions. Human umbilical cord mesenchymal stem cells (hUC-MSCs) transplantation has emerged as a promising therapeutic strategy for SCI, but whether it concurrently improves pathological alterations in both the spinal cord and brain remains unclear.

Methods

After establishing a mouse model of spinal cord contusion, hUC-MSCs were transplanted intrathecally starting 7 d post-injury and administered weekly for 4 wk. This study evaluated the effectiveness of hUC-MSCs transplantation in promoting neurological functional recovery in female mice with SCI through multi-dimensional behavioral tests. In addition, histopathological staining, immunofluorescence staining, and western blotting were performed to evaluate pathological changes in the spinal cord and hippocampus after hUC-MSCs transplantation.

Results

hUC-MSCs transplantation significantly improved hindlimb locomotor function and alleviated anxiety-like and depression-like behaviors, as well as long-term memory deficits in female mice with SCI. At the spinal cord level, hUC-MSCs increased IL-10 expression, reduced glial activation, and promoted axonal regeneration. In the hippocampus, hUC-MSCs alleviated neuroinflammation and region-specific abnormalities in synaptic remodeling and microglial synaptic pruning.

Conclusion

This study demonstrated that hUC-MSCs transplantation concurrently improved pathological alterations in both the spinal cord and hippocampus after SCI, providing experimental evidence for comprehensive neurorestoration following SCI.

Translational potential

This preclinical study demonstrated that subacute sequential hUC-MSCs transplantation is a promising therapeutic strategy for SCI. By improving both motor, emotional, and cognitive outcomes, it supports the concept of “brain-spinal co-repair” and provides a rationale for optimizing cell transplantation regimens to facilitate the clinical translation of hUC-MSC-based therapies for SCI.

Graphical abstract

1. Introduction

Spinal cord injury (SCI) stands as one of the most catastrophic traumatic diseases of the central nervous system (CNS). It is typified by high mortality in the acute phase, permanent functional disability, and multiple systemic complications [1,2]. Epidemiological data indicate that traumatic SCI accounts for over 85% of clinical cases, predominantly triggered by high-energy mechanical factors such as falls, traffic accidents, and violent trauma. Its core clinical manifestations are sensory and motor deficits below the injury level, imposing substantial economic burdens on patients’ families and society [3]. Notably, complications such as emotional dysregulation, including anxiety and depression, and cognitive impairments are highly prevalent in patients with chronic SCI and significantly impact prognosis [4]. In addition, there is a significant sex difference in emotional disorders among SCI patients, with a higher incidence in females, which may be closely related to functional differences in the hypothalamic-pituitary-adrenal (HPA) axis [5]. These neuropsychiatric complications profoundly impaired quality of life and social function, further complicating rehabilitation efforts.

From the pathophysiological perspective, the spinal cord, as a central relay station for neural signaling from the brain, triggers a series of complex pathophysiological responses following SCI [6,7]. In rodent models, external mechanical forces cause vertebral fractures or dislocations, leading to spinal cord compression or transection. Structural and functional disruptions, such as axonal breakage and blood-spinal cord barrier damage, further induce persistent molecular cascades, which can be broadly divided into three stages: the acute phase (0–7 d post-injury), the subacute phase (7–28 d post-injury), and the chronic phase (> 28 d post-injury) [8,9]. More critically, pathological changes in the spinal cord could induce distal cerebral abnormalities. Clinical neuroimaging and rodent studies have confirmed that SCI triggers structural atrophy in the primary sensory and motor cortex, characterized by axonal degeneration of the corticospinal tract and reduced myelin integrity. This neural remodeling of cortical circuits directly hindered motor function recovery [10]. Functional MRI studies further revealed abnormal functional connectivity between the prefrontal cortex and the hippocampus in SCI patients, and such neural network dysregulation was associated with severe cognitive impairment [11]. Concurrently, the limbic system which plays a central role in emotional regulation, encompassing brain areas like the hippocampus and amygdala, has been shown to undergo structural and functional alterations [12]. Emerging evidence indicated that region-specific alterations in glial reaction within these brain regions after SCI drove neuroinflammatory cascades that underlay anxiety, depression, and memory decline [13]. Hu et al. [14] demonstrated that SCI impaired hippocampal neurogenesis and induced cognitive deficits in mice, with microglia-derived extracellular vesicles serving as key mediators of spinal cord-hippocampal crosstalk during this process. Furthermore, persistent neuroinflammation not only exerted acute-phase neurotoxic effects on the brain but also exacerbated chronic-phase cerebral neurodegeneration, thereby hindering the rehabilitation after SCI [15].

In recent years, therapeutic strategies for SCI, including surgery, medication, rehabilitation therapy, and electrical stimulation of the spinal cord, remain insufficient to achieve complete functional recovery [16]. As an emerging regenerative medicine approach, cell transplantation has become one of the most promising modalities in SCI repair, and cell type selection and transplantation timing are key determinants of therapeutic efficacy [17]. Currently, multiple cell types have been applied in phase I clinical trials for SCI repair, including neural stem cells (NSCs), mesenchymal stem cells (MSCs), supportive glial cells, and macrophages [18]. Among them, MSCs derived from various origins such as bone marrow, adipose tissue, and umbilical cord exhibit low immunogenicity, enabling their use in both autologous and allogeneic transplantation [19]. Human umbilical cord mesenchymal stem cells (hUC-MSCs) are endowed with superior self-renewal and differentiation capabilities, and exert neurorestorative effects by secreting neurotrophic factors, reducing inflammatory responses, and alleviating oxidative stress, thus emerging as an attractive candidate for SCI repair [20]. Regarding transplantation timing, studies have shown that the pro-inflammatory microenvironment in the acute phase and inhibitory scar formation in the chronic phase collectively compromised the survival of transplanted cells [21]. In contrast, the subacute phase was characterized by attenuated inflammation and preserved neuroplasticity, making it a favorable window for hUC-MSCs transplantation. Animal studies have further demonstrated that subacute hUC-MSCs transplantation not only reconstructed neural circuits to facilitate motor function recovery, but also improved bladder function [22]. Despite accumulating evidence supporting the efficacy of hUC-MSCs in promoting spinal cord repair, whether hUC-MSCs transplantation also ameliorates SCI-induced pathological alterations in the brain remains largely unknown. Given the growing recognition of brain-spinal crosstalk after SCI, it remains important to determine whether cell therapy can concurrently target both spinal and cerebral dysfunction.

This study aimed to determine whether hUC-MSCs transplantation could simultaneously ameliorate pathological alterations in both the spinal cord and hippocampus after SCI, thereby providing experimental support for the concept of “brain–spinal co-repair”. Histological and molecular analyses were performed to characterize pathological changes in both the spinal cord and hippocampus following hUC-MSCs transplantation. Axonal regeneration, glial scar formation and the inflammatory microenvironment were evaluated at the spinal cord level. Synaptic remodeling in the CA1 region together with microglial morphology and synaptic pruning were assessed at the hippocampal level, so as to clarify whether hUC-MSCs transplantation promoted functional recovery by alleviating pathological synaptic remodeling.

2. Materials and methods

2.1. Acquisition and culture of hUC-MSCs

The hUC-MSCs used in this study were obtained from Tianjin Changhe Biotechnology Co., Ltd. (Batch No. 59873131-7). Before distribution, comprehensive quality control assessments were conducted by the manufacturer's accredited testing facility, including viability (> 95%), sterility, and phenotypic characterization. Cryopreserved cells were transported under validated cold-chain conditions (on dry ice) and subjected to immediate visual inspection for vial integrity and contamination upon receipt. All cell aliquots were subsequently stored in liquid nitrogen until experimental use. This entire procedure adhered to ethical guidelines and relevant operational protocols. For subsequent experiments, hUC-MSCs were rapidly thawed at 37°C and then resuspended in MEM-α culture medium (Gibco, C12571500BT) containing 10% fetal bovine serum (FBS, Vivacell, C04001-500) and 1% penicillin-streptomycin (Gibco, 15140-122). The hUC-MSCs were cultured under standard conditions (37°C, 5% CO₂, and 95% humidity), with medium changed every 48 h. Morphological homogeneity and adherence of hUC-MSCs were confirmed via phase-contrast microscopy. When cell confluence reached 80-90%, passaging was performed with 0.25% trypsin-EDTA (Gibco, C25200-072). To guarantee functional stability, hUC-MSCs from passages 3 to 5 were used for transplantation.

2.2. Lentiviral transduction of hUC-MSCs

Luciferase was used exclusively as a reporter gene for in vivo tracking of transplanted hUC-MSCs. hUC-MSCs were seeded into a six-well plate at a density of 5 × 10⁵ cells per well for culture. After 24 h, the regular medium was exchanged for serum-free medium containing Luciferase lentivirus for transduction. According to the previous research and manufacturer's instructions [23], an appropriate volume of lentiviral solution was added to each well for transduction, with a multiplicity of infection of 50 (MOI = viral titer × viral volume/number of cells). After 12 h, the medium was replaced with the complete medium. Three days after transduction, hUC-MSCs were detached using Accutase and counted. Subsequently, hUC-MSCs were resuspended in phosphate-buffered saline (PBS) for intrathecal transplantation.

2.3. Animals

The rodents used in the experiment were female C57BL/6J mice (8-9 wk old; body weight 18 ± 2 g), purchased from Beijing Vital River Experimental Animal Technology Co., Ltd. (Beijing, China, license: SCXK (Jing) 2021-0006). Female mice were selected because their shorter urethra facilitated postoperative bladder care, thereby reducing urinary complications and improving model consistency. Mice were group-housed, with 3 to 5 mice per cage, under environmental conditions of constant temperature (20-25°C), a fixed humidity range (50%-55%), and a 12 h/12 h light/dark cycle, with free access to food and water. Before the experiments, mice were subjected to an acclimation phase to minimize the impact of environmental stressors and then randomized into three groups. The Sham + PBS group underwent laminectomy and intrathecal injection of PBS; the SCI + PBS group received spinal cord contusion followed by intrathecal injection of PBS; the SCI + hUC-MSCs group received spinal cord contusion followed by intrathecal injection of hUC-MSCs. During the chronic phase after SCI, mice successively underwent CatWalk gait analysis, open field test (OFT), novel object recognition task (NORT), sucrose preference test (SPT), forced swimming test (FST) and tail suspension test (TST) in order of increasing stimulation intensity. To minimize potential confounding effects of acute stress, a 24 h interval was maintained between tests, and all procedures were conducted under standardized conditions.

2.4. Establishment of the SCI model

After acclimating to the housing environment for 1 wk, a spinal cord contusion model was established by the Impactor Model III SCI impact system. Mice were weighed and anesthetized with isoflurane. After anesthesia was induced, isoflurane was maintained at 1.0%-1.5%. The dorsal surgical area was shaved and disinfected with iodophor. A midline posterior incision was made over the thoracic vertebrae, with longitudinal incisions made through the skin, subcutaneous tissue, and fascial layer. The bilateral paravertebral muscles were retracted to expose the T₁₀ vertebral lamina, and then the spinal cord was exposed after laminectomy. The mice were secured in the spinal stabilization device, and a 5.0 g impactor was dropped from a height of 12.5 mm. Successful impact was indicated by local swelling and congestion of the spinal cord. The incision was sutured and disinfected. The mice were then placed in a warming box until recovery from anesthesia, followed by return to their cages. All surgical procedures were conducted under sterile conditions. Immediately after injury, the BMS score was recorded as 0, confirming successful model establishment. Mice with a score higher than 0, indicating failure in establishing the SCI model, were excluded. Manual bladder expression was performed twice daily to alleviate urinary retention until spontaneous voiding was restored.

2.5. Intrathecal injection of hUC-MSCs

Sequential intrathecal transplantation of hUC-MSCs was performed in the subacute phase of SCI (on 7, 14, 21, and 28 d post-injury). Referring to existing studies [24], the tail region of mice was shaved and disinfected, and the L₅-L₆ intervertebral space was located at the midline between the bilateral iliac crests. The needle was inserted from this point, and tail-flick phenomenon indicated successful intrathecal access, after which cell suspension was slowly injected. A total of 3 × 10⁵ hUC-MSCs in 10 μl were administered. After injection, the Hamilton needle was left in place for 1 min before being slowly and gently pulled out, and the skin was sterilized. Isoflurane was maintained throughout the procedure. The Sham + PBS and SCI + PBS groups were injected with the same volume of PBS at the corresponding time points following the above procedure.

2.6. In vivo bioluminescence imaging

Based on a previous study [25], hUC-MSCs were transduced with Luciferase lentivirus before intrathecal injection in mice. In vivo bioluminescence imaging was performed immediately after cell transplantation, as well as 3 and 7 d post-transplantation. D-Luciferin sodium salt (150 mg/kg) was injected intraperitoneally 10 min before imaging. The dorsal side of the mice was cleaned, and anesthesia was induced with isoflurane. The anesthetized mice were placed in a prone orientation within the dark chamber of the small animal imaging system (IVIS SPECTRUM, USA). The mice were positioned with limbs extended, trunk upright, and the injection site facing upward. Bioluminescence imaging was performed to detect luciferase signals at the transplantation site.

2.7. Behavioral assessment

2.7.1. Basso Mouse Scale (BMS) score

The Basso Mouse Scale (BMS) is a widely used method for evaluating hindlimb locomotor function in SCI mice [26]. During the experiment, the mice were placed in an open field for 5 min. Double-blind and independent observations by two individuals were carried out to ensure reliability. Hindlimb motor function was evaluated before injury and weekly until 9 wk post-injury. BMS scores ranged from 0 to 9, where 0 indicated no detectable joint movement, and 9 represented normal movement based on hindlimb joint movement and support, motor coordination, and trunk stability. A higher score indicated better locomotor recovery in mice.

2.7.2. Catwalk gait analysis

As described previously [27], the CatWalk XT system (Noldus, Wageningen, Netherlands) was used to record the footprints of mice, and to evaluate the effect of hUC-MSCs transplantation on the motor coordination of SCI mice. The experiment was conducted in a darkroom, and mice received adaptive training in advance to ensure they could pass through the glass walkway smoothly. During the formal assessment, the walkway and camera settings were adjusted to optimize the clarity of movement trajectories and footprint recordings. Mice were placed on the glass walkway and allowed to move from one end to the other. Trials in which mice stopped or turned around were excluded. After each test, the walkway was thoroughly cleaned to eliminate odors. Gait parameters, including the regularity index (RI), print position, and footprint length, were automatically recorded via the CatWalk XT 10.6 software. Gait patterns served as sensitive indicators of motor coordination. The RI, which represents the proportion of normal gait, ideally approaches 100% in healthy mice.

2.7.3. Open field test (OFT)

The OFT was used to detect both locomotor activity and anxiety-like behavior in mice. As previously described [28], mice were positioned in the center (20 × 20 cm) of the test box (40 × 40 × 40 cm), which was divided into a center and a peripheral area. They were permitted to freely explore for 5 min. The computer tracking system was used to automatically record parameters including the time spent in the center, the number of entries into the center, and the total distance traveled, and the Visu Track software was used for statistical analysis. Following each test, the test box was sanitized with 75% alcohol to remove any residual odors.

2.7.4. Sucrose preference test (SPT)

The SPT is a widely used behavioral assay for assessing anhedonia-like behavior in mice. Reduced preference for sucrose solution is considered an indicator of depressive-like behavior. As outlined in previous studies [29], mice were provided with two bottles of 1% sucrose solution 48 h before the experiment. Subsequently, mice were individually housed and deprived of food and water for 24 h. During the test, each mouse was given one bottle of 1% sucrose solution and one bottle of water. After 12 h, the positions of the two bottles were swapped to prevent the mice from developing the position preference. After 24 h, the consumption of sucrose solution and pure water was recorded separately to calculate sucrose preference.

2.7.5. Forced swimming test (FST)

As outlined in previous studies [30], the FST was conducted to study depressive-like behaviors in rodents. During the test, a transparent plastic cylinder (10 × 35 cm) was filled with water at 23-25°C to a depth of 25 cm to prevent the mice from supporting themselves on the bottom of the cylinder. During the test, the floating time and latency to floating of mice were recorded for 6 min. Prolonged floating time and shortened latency to immobility are indicators of depressive-like behaviors.

2.7.6. Tail suspension test (TST)

As previously described [31], the TST was performed to evaluate depressive-like behaviors in mice. During the test, mice were suspended by adhesive tape in a visually isolated area, 40 cm above the ground, for 6 min. The immobility time and latency to immobility were recorded. Immobility was defined as a state of despair in which mice remained completely motionless during suspension. Prolonged immobility time and shortened latency to immobility are indicators of depressive-like behaviors. To ensure accurate results, the tape was fixed 1 cm from the tip of the mouse's tail to prevent tail climbing or entanglement with the rod. Mice that fell from the device during the experiment were excluded.

2.7.7. Novel object recognition task (NORT)

Under normal conditions, mice display an innate preference for exploring novel objects, reflecting intact recognition memory for familiar objects [32]. Therefore, the NORT was utilized to evaluate learning and memory function in mice. The test consisted of four phases: habituation, training, long-term memory test, and location memory test. In the habituation stage, mice were placed in the testing arena and allowed to explore freely for 5 min/d for 3 consecutive days. The training phase was conducted 24 h after completion of habituation. Two identical objects (object A and object B) were symmetrically positioned in the arena, and mice were allowed to explore freely for 5 min. The long-term memory test was performed 24 h after training, during which object A was replaced with a novel object (object C), and mice were allowed to explore for 5 min. Another 24 h later, the object location memory test was conducted, where object C was relocated to the diagonal position relative to object B, and mice were allowed to explore freely for an additional 5 min. During the training phase, long-term memory test, and object location memory test, the time spent sniffing each object was recorded separately. Sniffing time was defined as the duration when the mouse's nose was within 2 cm of the object or when the mouse actively explored the object, excluding time spent climbing on or leaning against the object. After each test, both the objects and the testing arena were thoroughly cleaned with 75% alcohol to eliminate residual odors.

2.8. Enzyme-linked immunosorbent assay (ELISA)

Blood samples were collected from the medial canthus of the eye at the same time point across all groups. Serum was obtained by centrifugation at 4°C and 4000 rpm for 10 min and then stored at −20°C until analysis. Serum levels of corticotropin-releasing hormone (CRH), adrenocorticotropic hormone (ACTH), and corticosterone (CORT) were measured using commercial ELISA kits (JONLNBIO, JL11988-48T, JL12373-48T, and JL11918-48T) according to the manufacturer's instructions. Before the assay, all reagents were equilibrated to room temperature. Briefly, samples and standards were added to microplate wells in duplicate, followed by incubation with HRP-conjugated reagents at 37°C in the dark for 60 min. After washing, TMB substrate solution was added to each well and incubated for color development at 37°C in the dark for 15 min. The reaction was terminated with stop solution. The optical density (OD) was measured at 450 nm using a microplate reader, and hormone concentrations were calculated based on the standard curves.

2.9. Tissue processing

Mice were anesthetized by intraperitoneal injection of pentobarbital sodium (40 mg/kg), followed by transcardiac perfusion with ice-cold PBS and subsequently 4% paraformaldehyde (PFA). The injured spinal cord tissue and brain were harvested, post-fixed in 4% PFA overnight, and then transferred to 30% sucrose solution in PBS for cryoprotection. After 3 d, tissues were embedded in optimal cutting temperature (OCT) compound and stored at −80°C. Spinal cord tissues were sectioned sagittally at 30 μm, and brain tissues were sectioned coronally at 50 μm using the cryostat (Leica, Germany) for subsequent histological staining.

2.10. Immunofluorescence staining

Frozen sections of the spinal cord and brain were allowed to reach room temperature, after which they were rinsed three times with Tris-buffered saline supplemented with Tween 20 (TBST). For permeabilization and blocking, the sections were treated with 0.25% Triton X-100 and 5% bovine serum albumin (BSA) for 1 h. Following blocking, the sections were rinsed three times with TBST and then incubated with primary antibodies at 4°C overnight. The next day, the sections were rinsed three times with TBST, followed by incubation with secondary antibodies at room temperature in the dark for 1 h. Following this, sections were rinsed three times with TBST in the dark and were mounted with antifade medium containing 4′,6-Diamidino-2-phenylindole (DAPI, Abcam, ab104139). Images of spinal cord and brain sections were randomly captured on an automated quantitative pathology imaging system (Vectra Polaris, USA) and an LSM 900 laser scanning confocal microscope (Zeiss, Germany), respectively. For spinal cord sections, imaging focused on the epicenter, as well as the rostral and caudal regions, with concurrent local magnification observations. Immunofluorescence intensity was quantified by the ImageJ software, followed by normalization to the corresponding intact spinal cord region. For brain tissue sections, imaging primarily targeted the hippocampal CA1 region.

To assess the phagocytic function of microglia, Z-stacks imaging was performed on microglia in the hippocampal CA1 region using a 63× oil immersion objective of the LSM 900 laser scanning confocal microscope (Zeiss, Germany), with 0.8 μm z-steps. To visualize the phagocytic state of microglia, we converted the Z-stack three-dimensional (3D) image sequences into two-dimensional (2D) images through maximum intensity projection (MIP). Following established protocols [33], the acquired images were processed and analyzed by Imaris software (Bitplane, Zurich, Switzerland) to construct 3D models of CD68-positive structures and Synapsin Ⅰ (SYN1) puncta within microglia. The phagocytic function was quantified as the volume of SYN1 puncta or CD68-positive structures relative to the total volume of microglia. In addition, SYN1 puncta were reconstructed as spots in Imaris, and the number of puncta contained within IBA-1-positive microglia was quantified. Microglial morphology was analyzed by the ImageJ software to quantify branch complexity.

For immunofluorescence staining of cultured hUC-MSCs, cells were cultured in 24-well plates, washed three times with TBST, and fixed with 4% PFA (Solarbio, P1110) for 15 min at room temperature. Subsequently, immunofluorescence staining was performed using the same protocol described for tissue sections. Nuclei were counterstained with DAPI (Beyotime, BL105A) for 15 min, followed by three additional washes with TBST. Fluorescence images were acquired using the inverted fluorescence microscope (Leica, Germany).

The antibodies used were as follows: Rabbit anti-Ionized calcium-binding adapter molecule 1 (IBA-1, 1:200, FUJIFILM, 019-19741), Goat anti-IBA-1 (1:200, Abcam, ab5076), Goat anti-Glial fibrillary acidic protein (GFAP, 1:200, Abcam, ab53554), Mouse anti-Neurofilament protein-200 (NF-200, 1:200, Invitrogen, 13-1300), Mouse anti-CD68 (1:200, PTM BIO, PTM5185), Rabbit anti-Synapsin Ⅰ (SYN1, 1: 200, Abcam, ab254349), Mouse anti-Chondroitin sulfate proteoglycans (CSPGs, 1:200, Sigma–Aldrich, C8035), Mouse anti-Human Nuclear Antigen (HuNu, 1:200, Abcam, ab191181), Donkey anti-Rabbit IgG H&L Alexa Fluor 488 (1:200, Abcam, ab150061), Donkey anti-Rabbit IgG H&L Alexa Fluor 647 (1:200, Abcam, ab150075), Donkey anti-Goat IgG H&L Cy3 (1:200, Beyotime, A0502), Donkey anti-Mouse IgG H&L Alexa Fluor 488 (1:200, Abcam, ab150105).

2.11. Nissl staining

In accordance with the protocols outlined in the Nissl staining kit (Cresyl Violet Method, Solarbio, G1430), 10 μm-thick serial frozen sections of the spinal cord were prepared. The sections were immersed in staining solution and incubated at 56°C for 30 min. Subsequently, the sections were rinsed with double-distilled water to remove excess staining solution. Nissl differentiation solution was added dropwise, and the sections were observed under the microscope until the background was nearly colorless. Following dehydration with gradient ethanol and clearing with xylene, the sections were mounted with neutral resin. Images were captured by the digital pathology imaging analysis system (Leica, Germany) and analyzed using the ImageJ software.

2.12. Golgi staining

Golgi staining was employed to characterize neuronal morphology in the dorsal hippocampus (dHPC) and ventral hippocampus (vHPC) CA1 regions, including the assessment of dendritic complexity and spine density. Brain tissues were fixed by transcardiac perfusion with 4% PFA and post-fixed for 24 h, followed by immersion in GENMED fixative for 14 d. Tissue blocks were rinsed with PBS and transferred to 30% sucrose solution in PBS at 4°C for 48 h. Sections (80-100 μm) were prepared in 6% sucrose solution using a vibratome (Leica, Germany) and mounted onto gelatin-coated slides. Golgi staining was carried out in line with manufacturer's instructions (GENMED, GMS 80020.1). After staining, sections were dehydrated through graded ethanol, cleared with xylene, and coverslipped with neutral resin in the dark. This method capitalizes on the argyrophilic nature of neurons to effectively visualize their morphology [34]. Neurons in the CA1 region of the dHPC and vHPC were identified under an upright optical microscope (Olympus, Japan) using the 20× objective. Pyramidal neurons were distinguished by their typical soma with multiple dendritic branches at both apical and basal ends. Sholl analysis was applied to assess dendritic complexity. Concentric circles at 40 μm intervals were centered on the soma, and the number of dendritic intersections, as well as total dendritic branch number and length, were quantified. Secondary dendrites were imaged under the 100× oil immersion objective to assess spine density (defined as the number of spines per 10 μm). Only neurons with complete morphology and minimal overlap with neighboring dendrites were selected for analysis. Quantitative analysis was conducted by the ImageJ software. Importantly, cells were sampled from independent animals across different groups rather than repeatedly measured from a single tissue field.

2.13. Western blotting

Mice were deeply anesthetized with an overdose of sodium pentobarbital, and tissues from the injured spinal cord, dHPC, and vHPC were rapidly dissected, frozen in liquid nitrogen, and stored at −80°C. Western blotting was performed following standard protocols. Tissues were lysed in RIPA buffer (Beyotime, P0013B) supplemented with a protease inhibitor (Beyotime, P1045-1) and phosphatase inhibitor mixture (Beyotime, P1045-2). Protein concentrations were determined using a NanoDrop micro-spectrophotometer (Thermo Scientific, USA).

Protein samples were mixed with loading buffer (Biosharp, BL502B) at a 4:1 ratio and heated at 100°C for 10-20 min for denaturation. Equal amounts of protein were separated by 10% or 12.5% SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked with 5% non-fat milk in TBST for 1 h, then incubated overnight with primary antibody at 4°C. After washing with TBST, membranes were incubated with secondary antibody at room temperature for 1 h. Protein bands were visualized using an ECL chemiluminescence kit (EMD Millipore, WBKLS0500) and captured with the GelView 6000Plus imaging system (Biolight Biotechnology, China). Band intensities were quantified using the ImageJ software.

The antibodies used were as follows: Rabbit anti-Interleukin 1β (IL-1β, 1:1000, Proteintech, 26048-1-AP), Rabbit anti-Tumor necrosis factor-α (TNF-α, 1:1000, Wanleibio, WL01581), Rabbit anti-Interleukin 10 (IL-10, 1:1000, Abcam, ab9969), Rabbit anti-Synapsin Ⅰ (SYN1, 1: 1000, Abcam, ab64581), Rabbit anti-Glutamate receptor 1 (GluR1, 1: 1000, Abcam, ab183797), Rabbit anti-Postsynaptic density protein 95 (PSD95, 1: 2000, Abclonal, A7889PM), Rabbit anti-Brain-derived neurotrophic factor (BDNF, 1: 1000, Abcam, ab6201), Rabbit anti-Glyceraldehyde 3-phosphate dehydrogenase (GAPDH, 1: 5000, Abcam, ab181602), Rabbit anti-β-actin (1: 5000, Bioworld, AP0060), Goat anti-Rabbit IgG HRP-linked Antibody (1: 10000, Beyotime, A0208).

2.14. Statistical analysis

Normally distributed data were presented as Mean ± SEM. All statistical analyses were performed by GraphPad Prism 8 software. Prior to parametric analyses, the Shapiro–Wilk test was used to assess data normality, and the Brown–Forsythe test was used to evaluate homogeneity of variance. For comparisons among multiple groups, one-way analysis of variance (ANOVA) followed by Tukey's post hoc test was applied for normally distributed data with equal variance. Data that failed to meet the assumptions of normality or homogeneity of variance were analyzed using the Kruskal–Wallis test followed by Dunn's multiple comparisons test. Because this study was a hypothesis-driven study with predefined primary endpoints, no global false discovery rate (FDR) correction across all experiments was applied. Body weight and BMS scores were analyzed using two-way repeated-measures ANOVA. Sholl analysis data were analyzed using two-way ANOVA with repeated measures across distance from the soma. P < 0.05 was regarded as statistically significant. All experiments were independently replicated at least three times.

3. Results

3.1. Characterization and in vivo fate of transplanted hUC-MSCs following intrathecal administration

To characterize the biological properties of the hUC-MSCs used in this study, their surface marker expression was first analyzed by flow cytometry. The results showed that the percentages of CD14- and CD34-positive cells were 0.67% and 0%, respectively, both below 2%, whereas the percentages of cells positive for the MSC markers CD73 and CD105 were both greater than 95% (Fig. S1A). Morphological observation revealed that hUC-MSCs displayed a spindle-shaped appearance and possessed strong proliferative capacity. Furthermore, under lineage-specific induction conditions, the cells underwent osteogenic, adipogenic, and chondrogenic differentiation, which were verified by Alizarin Red S, Oil Red O, and Alcian Blue staining, respectively. The results demonstrated that hUC-MSCs could successfully differentiate into osteoblasts, adipocytes, and chondrocytes (Fig. S1B), confirming their typical trilineage differentiation potential. To further verify the human origin of the cultured cells, immunofluorescence staining for the human nuclear marker (HuNu) was performed. The results showed robust HuNu-positive staining in hUC-MSCs, confirming their human origin and supporting the subsequent in vivo cell-tracking experiments (Fig. S1C). Collectively, these findings indicated that hUC-MSCs used in this study possessed the characteristic surface marker profile and multipotent differentiation capacity, meeting the established criteria for MSCs set by the International Society for Cellular Therapy (ISCT) [35].

Intrathecal delivery was used to transplant hUC-MSCs into the subarachnoid space, potentially facilitating their distribution within the cerebrospinal fluid (CSF). This method facilitates widespread distribution of transplanted cells and trophic factors through the CSF. Additionally, it avoids secondary injury associated with in situ transplantation and overcomes the low delivery efficiency of intravenous injection [36]. Accordingly, mice received intrathecal transplantation of hUC-MSCs (3 × 10⁵ cells in 10 μl) at 7 d post-injury following establishment of the spinal cord contusion model. To evaluate the in vivo fate of transplanted cells, hUC-MSCs were transduced with a lentiviral vector carrying the luciferase gene 3 d before transplantation, and in vivo bioluminescence imaging was performed to monitor cell survival after repeated intrathecal transplantation. Bioluminescent signals were detected at 0, 1, 3, and 7 d after both the first and second transplantation (Fig. S1D). The results showed that hUC-MSCs survived for approximately 3-7 d after each transplantation in SCI mice. Following both the first and second administrations, the bioluminescent signal gradually decreased over time, and by 7 d post-transplantation, little or no detectable signal remained, indicating few or no surviving cells at the transplantation site. In addition, no obvious migration of transplanted cells toward the spinal cord lesion was observed (Fig. S1E and F). Importantly, the similar survival patterns after the second transplantation suggested that repeated intrathecal administration of hUC-MSCs did not induce overt immune rejection that markedly affected subsequent cell survival. Nevertheless, the survival curve indicated that transplanted hUC-MSCs did not persist long-term after a single administration. Therefore, these findings supported the rationale for adopting a sequential transplantation strategy to prolong the duration of cell-mediated repair effects in SCI mice.

To further determine whether intrathecally transplanted hUC-MSCs migrated into the brain, HuNu immunofluorescence staining was performed in the dHPC and vHPC at 7 and 56 d after transplantation (Fig. S1D and G). No HuNu-positive cells were detected in either hippocampal region at either time point (Fig. S1G), suggesting that intrathecally transplanted hUC-MSCs were unlikely to directly engraft within the dHPC or vHPC.

3.2. hUC-MSCs transplantation facilitated motor function recovery in female mice with SCI

After establishing the spinal cord contusion model, intrathecal transplantation of hUC-MSCs was initiated at 7 d post-injury (subacute phase) and continued for 4 wk. Four weeks after sequential transplantation, the reparative effects of hUC-MSCs transplantation were evaluated using multidimensional behavioral assessments, including motor, emotional, and cognitive function (Fig. 1A). Body weight was measured weekly in all groups. Compared with the Sham + PBS group, mice in the SCI + PBS and SCI + hUC-MSCs groups exhibited significant weight loss (P < 0.0001), while there was no statistically significant difference in body weight between the SCI + PBS and SCI + hUC-MSCs groups (Fig. 1B). This indicated that hUC-MSCs transplantation had no substantial impact on the body weight of SCI mice.

Fig. 1.

Effects of hUC-MSCs transplantation on motor function recovery in female mice with SCI.

(A) Experimental procedure. Following the establishment of a mouse spinal cord contusion model, sequential intrathecal transplantation of hUC-MSCs was initiated during the subacute phase of SCI. During the chronic phase, a battery of standardized behavioral tests was conducted to comprehensively assess neurological recovery. (B-C) Body weight and BMS score in each group were continuously monitored before injury and weekly until 9 wk post-injury (Body weight: group effect: F_{2, 247} = 215.3, P < 0.0001, time: F_{9, 247} = 36.18, P < 0.0001, interaction: F_{18, 247} = 3.774, P < 0.0001; BMS score: group effect: F_{2, 252} = 1213, P < 0.0001, time: F_{11, 252} = 65.67, P < 0.0001, interaction: F_{22, 252} = 17.43, P < 0.0001; BMS subscore: group effect: F_{2, 252} = 5810, P < 0.0001, time: F_{11, 252} = 198.8, P < 0.0001, interaction: F_{22, 252} = 50.69, P < 0.0001). n = 8-11 mice per group. (D-F) Representative footprint images and sequential views of limb movements were recorded using the CatWalk XT system at 56 d post-injury. Quantitative analysis of gait pattern parameters, including regularity index (F_2,21 = 79.07, P < 0.0001) and hindlimb print length (LH: F_2,21 = 17.19, P < 0.0001; RH: F_2,21 = 6.368, P = 0.0069). n = 8 mice per group. Data were expressed as Mean ± SEM. Data on body weight and BMS score were analyzed by two-way repeated measure ANOVA, while other results were performed using one-way ANOVA followed by Tukey's post hoc test. Significance: **P < 0.01, ****P < 0.0001 versus the Sham + PBS group; #P < 0.05, ##P < 0.01, ###P < 0.001 versus the SCI + PBS group. LH, left hindlimb. RH, right hindlimb.

To determine the impact of hUC-MSCs transplantation on locomotor recovery, hindlimb motor function was evaluated before injury and weekly for 9 wk after SCI. The results showed that from 21 to 63 d post-injury, the SCI + hUC-MSCs group exhibited significantly higher scores than the SCI + PBS group (21 d post-injury: P = 0.0279; 28 d post-injury: P = 0.0392; 35 d post-injury: P = 0.0392; 42 d post-injury: P = 0.0195; 49 d post-injury: P = 0.0061; 56 d post-injury: P = 0.0011; 63 d post-injury: P = 0.0017; Fig. 1C). Similar improvements were also observed in the BMS subscore from 42 to 63 d post-injury (42 d post-injury: P = 0.0395; 49 d post-injury: P = 0.0395; 56 d post-injury: P = 0.0016; 63 d post-injury: P = 0.0016; Fig. 1C). At 56 d post-injury, catwalk gait analysis further demonstrated improved hindlimb motor coordination in the SCI + hUC-MSCs group. Specifically, compared with the SCI + PBS group, the SCI + hUC-MSCs group showed a marked decrease in hindlimb dragging (Fig. 1D) and a notable shortening of footprint length on both the left and right hindlimbs (LH: P = 0.0038, RH: P = 0.0240, Fig. 1F). Notably, the RI in the SCI + hUC-MSCs group was significantly increased, approaching 60% (P = 0.0007, Fig. 1E). Collectively, these findings indicated that hUC-MSCs transplantation facilitated motor function recovery in female mice with SCI.

3.3. hUC-MSCs transplantation ameliorated neurological impairments and improved chronic HPA axis dysregulation in female mice with SCI

To investigate the effects of hUC-MSCs transplantation on cerebral neurological function in SCI mice, behavioral tests, including the OFT, NORT, SPT, FST, and TST, were performed sequentially from 57 to 68 d post-injury to evaluate emotional and cognitive phenotypes (Fig. 2A). Previous studies have validated the applicability of these behavioral paradigms for assessing behavioral phenotypes in SCI models. Moreover, conducting these tests during the chronic phase may reduce potential confounding effects associated with motor dysfunction [13].

Fig. 2.

Effects of hUC-MSCs transplantation on emotional and cognitive function and the HPA axis in female mice with SCI.

(A) Experimental procedure. At 57-68 d post-injury, the emotional and cognitive phenotypes of mice were sequentially assessed by OFT, NORT, SPT, FST, and TST. (B) Trajectory plots of OFT and detection indicators, including time spent in center, total distance traveled and number of entries into center (time spent in center: F_{2, 21} = 56.31, P < 0.0001; total distance traveled: F_{2, 21} = 22.02, P < 0.0001; entries in the central zone: H(2) = 16.24, P = 0.0003). n = 8-10 mice per group. (C) Schematic diagram of SPT and detection indicators, including sucrose solution consumption and sucrose preference (sucrose solution: F_{2, 21} = 17.27, P < 0.0001; sucrose preference: F_{2, 21} = 9.955, P = 0.0009). n = 8-10 mice per group. (D) Schematic diagram of FST and detection indicators, including latency to floating (F_{2, 21} = 10.94, P = 0.0006) and floating time (F_{2, 21} = 12.93, P = 0.0002). n = 8-10 mice per group. (E) Schematic diagram of TST and detection indicators, including latency to immobility (F_{2, 22} = 6.204, P = 0.0073) and immobility time (F_{2, 22} = 5.940, P = 0.0087). n = 8-10 mice per group. (F) Schematic diagram of NORT. (G) Sniffing time for objects A and B during the training period (object A: F_{2, 19} = 5.767, P = 0.0110; object B: F_{2, 19} = 6.578, P = 0.0067). n = 7-8 mice per group. (H) Sniffing time for objects C and B during the long-term memory test (object C: F_{2, 19} = 30.18, P < 0.0001; object B: F_{2, 19} = 1.017, P = 0.3805). n = 7-8 mice per group. (I) Sniffing time for objects C and B during the location memory test (object C: F_{2, 19} = 21.78, P < 0.0001; object B: F_{2, 19} = 6.417, P = 0.0074). n = 7-8 mice per group. (J) Schematic diagram of the HPA axis. (K-M) At 68 d post-injury, levels of serum CRH, ACTH, and CORT in mice were detected by ELISA (CRH: H(2) = 8.763, P = 0.0058; ACTH: F_{2, 15} = 11.63, P = 0.0009; CORT: F_{2, 14} = 22.36, P < 0.0001). n = 5-7 mice per group. Data were expressed as Mean ± SEM. Data on entries in the central zone and CRH level were analyzed by the Kruskal–Wallis test, while other results were performed using one-way ANOVA followed by Tukey's post hoc test. Significance: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 versus the Sham + PBS group; ns P > 0.05, #P < 0.05, ##P < 0.01, ###P < 0.001 versus the SCI + PBS group.

Regarding emotion-related behaviors, the OFT was employed to assess anxiety-like behaviors in SCI mice and to determine the effects of hUC-MSCs transplantation on these behavioral traits. The results revealed that the SCI + PBS group exhibited less time spent in the center, shorter travel distance, and fewer central entries than the Sham + PBS group (time spent in center: P < 0.0001; total distance: P < 0.0001; entries into the center: P = 0.0002). In contrast, the SCI + hUC-MSCs group displayed increased time spent in the center (P = 0.0331, Fig. 2B), indicating that hUC-MSCs transplantation alleviated anxiety-like behavior in SCI mice. While no statistically significant differences were found in total distance traveled or the count of central entries between the SCI + PBS and SCI + hUC-MSCs groups, both the trajectory heatmap and the outcomes of Tukey's multiple comparisons suggested a trend towards enhanced exploratory activity in the SCI + hUC-MSCs group (P = 0.09, Fig. 2B). This further reinforced the notion that hUC-MSCs transplantation facilitated motor function recovery in SCI mice. Subsequently, the SPT, FST, and TST were applied to assess the effects of hUC-MSCs transplantation on depressive-like behaviors in SCI mice. We found that hUC-MSCs transplantation reversed the SCI-induced decrease in sucrose preference in the SPT (sucrose solution: P = 0.0018; sucrose preference: P = 0.0465; Fig. 2C), prevented the SCI-induced shortening of latency to floating and immobility (latency to floating: P = 0.0405; latency to immobility: P = 0.0335) in the FST and TST, and attenuated the prolongation of floating and immobility time (floating time: P = 0.0009; immobility time: P = 0.0403; Fig. 2D and E). These findings suggested that hUC-MSCs transplantation ameliorated depressive-like behaviors in female mice with SCI.

The NORT was performed from 58 to 61 d post-injury to evaluate recognition memory in mice, including long-term memory and object recognition-associated spatial memory (Fig. 2F). In the training phase, the SCI + PBS group spent less time sniffing the two identical objects than the Sham + PBS group (object A: P = 0.0091; object B: P = 0.0062). Nevertheless, there was no statistical difference between the SCI + PBS and SCI + hUC-MSCs groups (Fig. 2G). The long-term memory test was implemented 24 h after the training phase. The results revealed that the SCI + PBS group had a significantly shorter sniffing time for the novel object than the Sham + PBS group (object C: P < 0.0001). Conversely, the SCI + hUC-MSCs group showed a significant preference for the novel object (object C: P = 0.0099; Fig. 2H). Meanwhile, among the three groups, there was no significant difference in the sniffing time for the familiar object (Fig. 2H). These results suggested that hUC-MSCs transplantation ameliorated SCI-induced long-term memory deficits. Subsequently, the object location memory test was performed 24 h after the long-term memory test. After object C was relocated, compared with the Sham + PBS group, both the SCI + PBS and SCI + hUC-MSCs groups displayed a significant decrease in the sniffing time for the object C (P < 0.0001; P = 0.0004), with no statistical difference observed between the two groups (Fig. 2I). The above results indicated that hUC-MSCs transplantation had no significant effect on SCI-induced spatial memory deficits.

To further determine whether the improvements in emotional and cognitive behaviors were merely secondary to enhanced locomotor recovery, correlation analyses were performed between BMS scores and neurobehavioral outcomes. Scatter plots showed no significant correlations between BMS scores and the time spent in the center of the OFT (R² = 0.2784, P = 0.0525), sucrose preference of the SPT (R² = 0.1040, P = 0.2609), latency to floating (R² = 0.1423, P = 0.1837) and floating time (R² = 0.0809, P = 0.3243) of the FST, latency to immobility (R² = 0.0157, P = 0.6695) and immobility time (R² = 0.0026, P = 0.8613) of the TST, as well as time spent exploring the novel object in the NORT (R² = 0.2206, P = 0.0901; Fig. S2A–G). Although locomotor recovery might contribute to behavioral performance to some extent, the lack of significant correlations between BMS scores and emotional/cognitive indices suggested that the observed behavioral improvements were unlikely to be solely attributable to enhanced motor function.

The HPA axis is a major neuroendocrine system involved in stress responses, emotional regulation, and multiple physiological processes [37]. As illustrated in Fig. 2J, CRH secreted by the hypothalamus stimulates the release of ACTH from the anterior pituitary gland, which subsequently induces the secretion of glucocorticoids, including cortisol in humans and CORT in rodents. Elevated glucocorticoid levels in turn exert negative feedback on the hypothalamus and pituitary to suppress further CRH and ACTH release. Previous studies have demonstrated that the HPA axis underwent chronic dysregulation after SCI, leading to persistently elevated cortisol levels in patients [38]. To further investigate the effect of hUC-MSCs transplantation on the HPA axis after SCI, serum concentrations of CRH, ACTH, and CORT were examined in mice. The results indicated that the SCI + PBS group showed significantly higher serum CORT levels than the Sham + PBS group (P < 0.0001). Notably, this phenomenon was reversed by hUC-MSCs transplantation (P = 0.0029, Fig. 2M). Similarly, the results revealed that both the SCI + PBS and SCI + hUC-MSCs groups had lower CRH and ACTH levels than the Sham + PBS group (CRH: P = 0.0388, P = 0.0206; ACTH: P = 0.0011, P = 0.6923). However, compared with the SCI + PBS group, the SCI + hUC-MSCs group exhibited partial recovery of ACTH level (P = 0.0093; Fig. 2K and L). These findings suggested that hUC-MSCs transplantation improved SCI-induced chronic dysregulation of the HPA axis.

3.4. hUC-MSCs transplantation promoted axonal regeneration, inhibited scar formation and inflammatory response in female mice with SCI

To elucidate the mechanisms underlying SCI repair mediated by hUC-MSCs, immunofluorescence staining for NF-200 (a neurofilament marker), IBA1 (a microglia marker), GFAP (an astrocyte marker), and CSPGs (a glial scar marker) was performed at 68 d post-injury. In addition, the expression levels of pro-inflammatory cytokines IL-1β and TNF-α, as well as the anti-inflammatory cytokine IL-10, were assessed in the lesion. In the SCI + PBS group, extensive axonal fragmentation and atrophy were observed throughout the lesion, including the epicenter and the adjacent rostral and caudal regions. Correspondingly, NF-200 immunoreactivity was significantly reduced, whereas GFAP immunoreactivity was markedly increased compared with the Sham + PBS group (NF-200: lesion P < 0.0001; rostral P < 0.0001; caudal P < 0.0001; GFAP: lesion P < 0.0001; rostral P = 0.0001; caudal P = 0.0006; Fig. S3A, E and F). However, hUC-MSCs transplantation significantly increased NF-200 fluorescence intensity and reduced GFAP immunoreactivity compared with the SCI + PBS group (NF-200: lesion P = 0.0334; rostral P = 0.0718; caudal P = 0.0335; GFAP: lesion P = 0.0042; rostral P = 0.0178; caudal P = 0.0164; Fig. S3A, E and F). To further evaluate glial scar formation, immunofluorescence staining for CSPGs was performed. The expression of CSPGs was markedly elevated following SCI (P = 0.0150) and was predominantly localized within GFAP positive regions. hUC-MSCs transplantation markedly reduced CSPGs deposition at the lesion site (P = 0.0407; Fig. S4A and B), suggesting an attenuation of glial scar formation. Together, these findings indicated that hUC-MSCs transplantation promoted neurofilament preservation and inhibited scar formation in the injured spinal cord.

Similarly, the SCI + PBS group exhibited significantly increased IBA-1 fluorescence intensity in the lesion, rostral, and caudal regions compared with the Sham + PBS group (lesion: P = 0.0220; rostral: P = 0.0091; caudal: P = 0.0106; Fig. S3B and G). In contrast, hUC-MSCs transplantation significantly reduced microglial accumulation following SCI. Specifically, the SCI + hUC-MSCs group showed a significant decrease in IBA-1-positive microglia compared with the SCI + PBS group, although no statistical difference was detected at the rostral end (lesion: P < 0.0001; rostral: P = 0.5346; caudal: P = 0.0370; Fig. S3B and G). Furthermore, hUC-MSCs transplantation upregulated IL-10 expression at the lesion site (P = 0.0153, Fig. S5A–B). In contrast, no statistical differences were observed in TNF-α or IL-1β expression among the three groups (Fig. S5A, C and D). Although TNF-α expression showed a decreasing trend following hUC-MSCs transplantation, the difference was not statistically significant compared with the SCI + PBS group (P > 0.05). Collectively, these findings suggested that hUC-MSCs transplantation attenuated the inflammatory response after SCI, primarily by enhancing anti-inflammatory signaling rather than suppressing TNF-α or IL-1β expression.

Furthermore, Nissl staining revealed a significantly greater number of Nissl bodies in the lesion area of the SCI + hUC-MSCs group than in the SCI + PBS group, with neuronal morphology appearing relatively intact (P = 0.0114, Fig. S3C–D). These results indicated that hUC-MSCs transplantation promoted neuronal survival at the injury epicenter, further supporting its therapeutic potential for SCI repair.

3.5. hUC-MSCs transplantation restored dendritic structure and synaptic maturation in hippocampal CA1 neurons after SCI

The dHPC and vHPC exhibit distinct functional specializations, with the dHPC mainly involved in memory and spatial processing and the vHPC associated with emotional regulation [39]. As the CA1 region serves as a key hub integrating cognitive and emotional information [40], we further examined synaptic structure and function in the CA1 region of both hippocampal subregions to investigate the hippocampal changes associated with hUC-MSC-mediated neurological recovery after SCI (Fig. 3A and I).

Fig. 3.

Effects of hUC-MSCs transplantation on dendritic complexity in hippocampal CA1 neurons of female mice with SCI.

(A and I) Schematic diagrams of neuronal complexity and secondary dendrites in the CA1 region of the dHPC and vHPC. Hippocampal coronal sections were processed for Golgi staining, and images were obtained under a 20× objective and a 100× oil immersion objective (n = 3 mice per group). For each animal, five well-impregnated CA1 pyramidal neurons were randomly selected for analysis, resulting in a total of 15 neurons per group for Sholl analysis and dendritic morphology quantification. Scale bar for neuronal complexity images: 50 μm. Secondary dendrites from the analyzed neurons were further magnified for dendritic spine density quantification. Scale bar for secondary dendrite images: 10 μm. (B and J) The number of intersections of neurons in the dHPC and vHPC CA1 region with concentric circles at different distances from the soma (dHPC: group effect: F_{2, 1070} = 0.4044, P = 0.6674, distance: F_{29, 1070} = 22.52, P < 0.0001, interaction: F_{58, 1070} = 2.402, P < 0.0001; vHPC: group effect: F_{2, 1260} = 168.1, P < 0.0001, distance: F_29,1260 = 56.38, P < 0.0001, interaction: F_{58, 1260} = 0.9842, P = 0.5102). (C and K) The number of dendritic branches in dHPC and vHPC CA1 neurons (dHPC: F_2,42 = 25.44, P < 0.0001; vHPC: F_2,42 = 6.426, P = 0.0037). (D and L) The total dendritic length in dHPC and vHPC CA1 neurons (dHPC: F_2,42 = 0.1729, P = 0.8418; vHPC: F_2,42 = 14.66, P < 0.0001). (E and M) The maximum dendritic length in dHPC and vHPC CA1 neurons (dHPC: F_2,42 = 8.249, P = 0.001; vHPC: F_2,42 = 5.849, P = 0.0057). (F and N) Dendritic spine density in dHPC and vHPC CA1 neurons (dHPC: F_{2, 42} = 7.022, P = 0.0023; vHPC: F_{2, 42} = 8.833, P = 0.0006). (G and O) Representative images of dendritic spine morphology, including thin, mushroom, stubby, and filopodial spines. (H and P) Quantitative analysis of each type of dendritic spine in dHPC and vHPC CA1 neurons (dHPC: thin F_2,42 = 23.67, P < 0.0001; stubby F_2,42 = 14.09, P < 0.0001; mushroom F_2,42 = 0.5502, P = 0.5809; filopodia F_2,42 = 0.4190, P = 0.6604; vHPC: thin: F_2,42 = 31.12, P < 0.0001; stubby: F_2,42 = 0.6076, P = 0.5494; mushroom: F_2,42 = 8.500, P = 0.0008; filopodia: F_2,42 = 0.7165, P = 0.4943). Data were expressed as Mean ± SEM. Data on Sholl analysis were analyzed by two-way repeated measures ANOVA, while other results were performed using one-way ANOVA followed by Tukey's post hoc test. Significance: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 versus the Sham + PBS group; ns P > 0.05, #P < 0.05, ##P < 0.01, ###P < 0.001, ####P < 0.0001 versus the SCI + PBS group.

In the dHPC, Sholl analysis showed that SCI induced a significant increase in dendritic intersections within 0–60 μm of the soma and in the total number of dendritic branches compared with the Sham + PBS group (P < 0.0001). However, hUC-MSCs transplantation significantly attenuated these SCI-induced changes (P < 0.0001; Fig. 3A–C). Interestingly, there were no statistical differences in the total dendritic length among the three groups (Fig. 3D). These findings might reflect SCI-induced dendritic remodeling in dHPC CA1 neurons, characterized by restricted dendritic growth and compensatory increases in dendritic branching. hUC-MSCs transplantation partially reversed these alterations, as evidenced by the significantly greater maximum dendritic length in the SCI + hUC-MSCs group compared with the SCI + PBS group (P = 0.0010, Fig. 3E). We further examined the number and morphology of dendritic spines in dHPC CA1 neurons, and found that hUC-MSCs transplantation alleviated the SCI-mediated reactive increase in dendritic spine density (P = 0.0081, Fig. 3F). Morphologically, dendritic spines were categorized into thin, mushroom, stubby, and filopodia types (Fig. 3G), which was consistent with previous studies. Thin and filopodia spines are considered immature, while stubby and mushroom spines are regarded as mature [41]. Our study showed no statistical differences in the density of mushroom and filopodia spines among the three groups. The SCI + PBS group had a higher density of thin spines and a lower density of stubby spines than the Sham + PBS group, indicating that weaker or immature spines were predominant (thin: P < 0.0001; stubby: P < 0.0001). However, hUC-MSCs transplantation prevented this shift toward immature spines in dHPC CA1 region neurons (thin: P < 0.0001; stubby: P = 0.0007; Fig. 3H).

Sholl analysis of vHPC CA1 neurons revealed a pattern opposite to that observed in the dHPC (Fig. 3I). Specifically, SCI significantly reduced dendritic intersections (50–300 μm from the soma), dendritic branching, total dendritic length, maximum dendritic length, and dendritic spine density compared with the Sham + PBS group (total dendritic branching points: P = 0.0038; total dendritic length: P = 0.0001; maximum dendritic length: P = 0.0291; spine density: P = 0.0004). These impairments were markedly ameliorated by hUC-MSCs transplantation, which increased dendritic intersections (80–300 μm from the soma) and restored overall dendritic complexity and spine density relative to the SCI + PBS group (total dendritic branching points: P = 0.0324; total dendritic length: P < 0.0001; maximum dendritic length: P = 0.0074; spine density: P = 0.0434; Fig. 3I–N). Moreover, hUC-MSCs transplantation significantly increased the proportion of mushroom spines and decreased the proportion of thin spines, whereas no significant differences were observed in stubby or filopodial spines (mushroom: P = 0.0011; thin: P < 0.0001; Fig. 3O and P). These results suggested that similar to the changes observed in the dHPC, hUC-MSCs transplantation inhibited excessive formation of immature dendritic spines in vHPC CA1 neurons triggered by SCI.

In summary, SCI led to region-specific dendritic remodeling in hippocampal CA1 neurons, accompanied by impaired synaptic maturation. hUC-MSCs transplantation largely reversed these structural abnormalities, restoring dendritic complexity and promoting synaptic maturation. Such hippocampal synaptic remodeling may underlie the beneficial effects of hUC-MSCs transplantation on emotional and cognitive dysfunction after SCI.

3.6. hUC-MSCs transplantation improved synaptic function remodeling in the hippocampal CA1 region induced by SCI

Western blotting was used to detect changes in synaptic protein expression of the hippocampus, including the presynaptic marker SYN1, and the postsynaptic markers GluR1 and PSD95 (Fig. 4A and B). We found that hUC-MSCs transplantation differentially modulated SYN1 expression in the dHPC and vHPC, without affecting the postsynaptic markers. Specifically, the results showed that the SCI + PBS group exhibited higher SYN1 expression in the dHPC than the Sham + PBS group (P = 0.0075). However, hUC-MSCs transplantation downregulated the expression of SYN1 (P = 0.0106, Fig. 4E). Conversely, the SCI + PBS group exhibited lower SYN1 expression in the vHPC than the Sham + PBS group (P = 0.0147), while hUC-MSCs transplantation significantly upregulated the expression of SYN1 (P = 0.0012, Fig. 4F). For postsynaptic markers, there were no statistical differences in the expression levels of GluR1 (Fig. 4C and D) and PSD95 (Fig. 4G and H) among the three groups in both the dHPC and vHPC, suggesting that hUC-MSCs transplantation primarily improved presynaptic remodeling.

Fig. 4.

Effects of hUC-MSCs transplantation on the expression of synapse-related proteins in the hippocampus of female mice with SCI.

(A and B) Western blotting was used to evaluate the expression of synapse-related proteins in the dHPC and vHPC in each group, including the presynaptic marker SYN1, the postsynaptic markers GluR1 and PSD95, and the neurotrophic factor BDNF. (C and D) Bar graphs showed the expression levels of GluR1 relative to GAPDH in the dHPC and vHPC, respectively (dHPC: F_2,6 = 1.598, P = 0.2777; vHPC: F_2,6 = 1.364, P = 0.3250). n = 3 mice per group. (E and F) Bar graphs showed the expression levels of SYN1 relative to GAPDH in the dHPC and vHPC, respectively (dHPC: F_{2, 6} = 10.36, P = 0.0113; vHPC: F_{2, 6} = 23.48, P = 0.0015). n = 3 mice per group. (G and H) Bar graphs showed the expression levels of PSD95 relative to GAPDH in the dHPC and vHPC, respectively (dHPC: F_2,6 = 0.2381, P = 0.7952; vHPC: F_2,6 = 0.0683, P = 0.9347). n = 3 mice per group. (I and J) Bar graphs showed the expression levels of BDNF relative to β-actin in the dHPC and vHPC, respectively (dHPC: F_2,6 = 8.414, P = 0.0182; vHPC: F_2,6 = 9.649, P = 0.0133). n = 3 mice per group. Data were expressed as Mean ± SEM. Statistical analysis was performed using one-way ANOVA followed by Tukey's post hoc test. Significance: *P < 0.05 versus the Sham + PBS group; #P < 0.05, ##P < 0.01 versus the SCI + PBS group.

Studies have reported that BDNF plays a central role in regulating dendritic spine maturation and synaptic transmission, particularly within hippocampal circuits associated with learning, memory, and emotional behaviors [42,43]. Therefore, the expression of BDNF was examined in both the dHPC and vHPC. The results showed that the SCI + PBS group exhibited lower BDNF expression in both the dHPC and vHPC than the Sham + PBS group (dHPC: P = 0.0254; vHPC: P = 0.0114). However, hUC-MSCs transplantation did not affect BDNF expression in the dHPC (Fig. 4I). In contrast, it significantly upregulated BDNF expression in the vHPC (P = 0.0268; Fig. 4J), suggesting that hUC-MSCs transplantation may contribute to the restoration of synaptic function by specifically regulating BDNF expression in the vHPC.

In summary, hUC-MSCs transplantation ameliorated SCI-induced hippocampal presynaptic remodeling by modulating SYN1 expression and increasing BDNF levels in the vHPC, providing synaptic-level evidence for the observed behavioral improvements.

3.7. hUC-MSCs transplantation inhibited glial cell activation in the hippocampal CA1 region induced by SCI

Having characterized the regulatory effects of hUC-MSCs transplantation on neuronal structure and function in the dHPC and vHPC CA1 regions of SCI mice, we further focused on the hippocampal inflammatory microenvironment. Through immunofluorescence staining for IBA-1 and GFAP, the activation pattern of glial cells was evaluated (Fig. 5A and D). The results indicated that the SCI + PBS group exhibited significantly higher mean fluorescence intensities of both IBA-1 and GFAP in the dHPC and vHPC than the Sham + PBS group (IBA-1: dHPC P = 0.0038, vHPC: P = 0.0003; GFAP: dHPC P = 0.0026, vHPC: P = 0.0106). Notably, hUC-MSCs transplantation markedly suppressed microglial activation (dHPC: P = 0.0051; vHPC: P = 0.0030; Fig. 5B and E) and astrocytic reactivity (dHPC: P = 0.0318; vHPC: P = 0.0300; Fig. 5C and F). These findings demonstrated that hUC-MSCs transplantation attenuated SCI-mediated hippocampal neuroinflammation.

Fig. 5.

hUC-MSCs transplantation suppressed glial cell activation in the hippocampal CA1 region of female mice with SCI.

(A) Representative panoramic images of immunofluorescence staining for GFAP (red) and IBA-1 (green) in the dHPC CA1 region were shown on the right side of the image. Scale bar: 200 μm. The high-resolution version of the boxed region on the right image was shown on the left. Scale bar: 50 μm. (B) Quantitative analysis of the relative mean fluorescence intensity of IBA-1 in the dHPC CA1 region (F_2,6 = 18.78, P = 0.0026). n = 3 mice per group. (C) Quantitative analysis of the relative mean fluorescence intensity of GFAP in the dHPC CA1 region (F_2,6 = 17.49, P = 0.0031). n = 3 mice per group. (D) Representative panoramic images of immunofluorescence staining for GFAP (red) and IBA-1 (green) in the vHPC CA1 region were shown on the right side of the image. Scale bar: 200 μm. The high-resolution version of the boxed region on the right image was shown on the left. Scale bar: 50 μm. (E) Quantitative analysis of the relative mean fluorescence intensity of IBA-1 in the vHPC CA1 region (F_2,6 = 39.77, P = 0.0003). n = 3 mice per group. (F) Quantitative analysis of the relative mean fluorescence intensity of GFAP in the vHPC CA1 region (F_2,6 = 10.87, P = 0.0101). n = 3 mice per group. Data were expressed as Mean ± SEM. Statistical analysis was performed using one-way ANOVA followed by Tukey's post hoc test. Significance: *P < 0.05, **P < 0.01, ***P < 0.001 versus the Sham + PBS group; #P < 0.05, ##P < 0.01 versus the SCI + PBS group.

3.8. hUC-MSCs transplantation enhanced microglia-mediated synaptic pruning in the dHPC CA1 region after SCI

To further investigate whether the regulatory effect of hUC-MSCs transplantation on synaptic plasticity in the hippocampal CA1 region of SCI mice was associated with microglia-mediated synaptic pruning, both the volume and the number of SYN1 puncta engulfed by microglia were quantitatively analyzed (Fig. 6A). The results showed that compared with the Sham + PBS group, the volume of engulfed SYN1 puncta within individual microglia in the dHPC CA1 region was significantly decreased in the SCI + PBS group (P = 0.0083), while hUC-MSCs transplantation markedly reversed this reduction (P = 0.0448; Fig. 6C). Consistent with the volumetric analysis, the number of SYN1 puncta contained within microglia was also significantly reduced after SCI (P = 0.0068) and was restored following hUC-MSCs transplantation in the dHPC CA1 region (P = 0.0123; Fig. 6D). These findings indicated that hUC-MSCs transplantation significantly ameliorated SCI-induced impairment of microglial synaptic pruning in the dHPC CA1 region. In contrast, no significant differences were observed among the three groups in either the volume of SYN1 within microglia or the number of SYN1 puncta engulfed by microglia in the vHPC CA1 region (Fig. 6F and G), suggesting that neither SCI nor hUC-MSCs transplantation affected microglia-mediated synaptic homeostasis in the vHPC CA1 region.

Fig. 6.

Regulatory effects of hUC-MSCs transplantation on microglia-mediated synaptic pruning in the hippocampal CA1 region of female mice with SCI.

(A) Representative images showing synaptic pruning by microglia in the dHPC and vHPC CA1 regions of each group, as revealed by immunofluorescent co-localization staining of IBA-1 (red) and SYN1 (green). n = 3-4 mice per group. And 13-16 cells per group were analyzed. 2D images are presented as MIP, scale bar = 10 μm. Corresponding 3D reconstructions of microglia reconstructed generated using Imaris are shown on the right. Scale bar = 2 μm. (B) Representative images showing the phagocytic capacity of microglia in the dHPC and vHPC CA1 regions of each group, as demonstrated by immunofluorescent co-localization staining of IBA-1 (red) and CD68 (green). n = 3-4 mice per group. And 13-16 cells per group were analyzed. 2D images are presented as MIP. Scale bar = 10 μm. Corresponding 3D reconstructions generated using Imaris are shown on the right. Scale bar = 2 μm. (C and F) Quantitative analysis of the volume of SYN1 puncta within microglia in the dHPC and vHPC CA1 regions, respectively (dHPC: F₂,₄₂ = 5.499, P = 0.0076; vHPC: F₂,₃₈ = 0.9773, P = 0.3856). (D and G) Quantitative analysis of the number of SYN1 puncta contained within microglia in the dHPC and vHPC CA1 regions, respectively (dHPC: F₂,₄₂ = 6.479, P = 0.0035; vHPC: F₂, 38 = 0.5223, P = 0.5973). (E and H) Quantitative analysis of the volume of CD68-positive structures within microglia in the dHPC and vHPC CA1 regions, respectively (dHPC: F₂,₃₈ = 4.819, P = 0.0136; vHPC: F₂,₄₃ = 1.828, P = 0.1730). Data were expressed as Mean ± SEM. Statistical analysis was performed using one-way ANOVA followed by Tukey's post hoc test. Significance: *P < 0.05, **P < 0.01 versus the Sham + PBS group; #P < 0.05 versus the SCI + PBS group.

Subsequently, the CD68-positive structures within microglia were evaluated using a similar approach (Fig. 6B). CD68, as a lysosomal marker, is widely used to reflect the phagocytic activity of microglia [44]. Quantitative analysis revealed that compared with the Sham + PBS group, the volume of CD68-positive structures within individual microglia in the dHPC CA1 region was significantly reduced in the SCI + PBS group (P = 0.0425). Notably, hUC-MSCs transplantation significantly increased the volume of CD68-positive structures relative to the SCI + PBS group (P = 0.0187; Fig. 6E), indicating a restoration of microglial phagocytic capacity in the dHPC. In contrast, no significant differences were observed among the three groups in the vHPC CA1 region (Fig. 6H), consistent with the results obtained for microglia-mediated synaptic pruning. Taken together, these findings indicated that hUC-MSCs transplantation enhanced the phagocytic capacity of microglia and ameliorated SCI-induced impairment of synaptic pruning in the dHPC CA1 region, which may contribute to improved synaptic homeostasis. Notably, this regulatory effect exhibited clear regional specificity and was not observed in the vHPC CA1 region.

3.9. hUC-MSCs transplantation increased branch complexity of microglia in the dHPC CA1 region after SCI

To further investigate the relationship between microglial morphology and function, microglia were reconstructed, and their branch number and total branch length were quantified to evaluate the effects of hUC-MSCs transplantation on microglial morphology in the hippocampal CA1 region after SCI (Fig. 7A). Greater branch complexity is generally associated with enhanced homeostatic functions, including phagocytosis and synaptic pruning [45]. We found that both branch number and total branch length of microglia in the dHPC CA1 region in the SCI + PBS group were significantly lower than those in the Sham + PBS group (branch number: P = 0.0014; total branch length: P = 0.0364). hUC-MSCs transplantation could significantly reverse these alterations, and microglial morphological complexity in the SCI + hUC-MSCs group was significantly higher than that in the SCI + PBS group (branch number: P = 0.0407; total branch length: P = 0.0262; Fig. 7B and C). However, in the vHPC CA1 region, there were no significant differences in the branch number and total branch length of microglia among the three groups (Fig. 7D and E). These results indicated that SCI induced functional and structural deficits of microglia in the dHPC CA1 region without affecting the vHPC. hUC-MSCs transplantation not only improved the synaptic pruning function of microglia in this region but also increased their branch complexity. These findings were consistent with reduced microglial activation observed following hUC-MSCs transplantation after SCI.

Fig. 7.

Regulatory effects of hUC-MSCs transplantation on microglial morphology in the hippocampal CA1 region of female mice with SCI.

(A) Representative images showing morphological characteristics of microglia in the dHPC and vHPC CA1 regions of each group (IBA-1, red; skeletonized images, white; n = 3-4 mice per group). And 14-15 cells per group were selected for analysis. 2D images of MIP. Scale bar = 10 μm. (B and C) Quantitative analysis of the microglial branch number and total branch length in the dHPC CA1 region of each group (branch number: F_{2, 40} = 7.477, P = 0.0017; total branch length: F_{2, 40} = 4.710, P = 0.0146). (D and E) Quantitative analysis of the number of microglial branches and total branch length in the vHPC CA1 region of each group (branch number: F_{2, 41} = 0.3326, P = 0.7190; total branch length: F_{2, 41} = 0.9718, P = 0.3870). Data were expressed as Mean ± SEM. Statistical analysis was performed using one-way ANOVA followed by Tukey's post hoc test. Significance: *P < 0.05, **P < 0.01 versus the Sham + PBS group; ns P > 0.05, #P < 0.05 versus the SCI + PBS group.

4. Discussion

The present study systematically evaluated the therapeutic effects of hUC-MSCs transplantation on pathological alterations in both the spinal cord and hippocampus following SCI, extending its role beyond traditional spinal cord repair. Sequential intrathecal transplantation of hUC-MSCs improved locomotor recovery and ameliorated emotional and cognitive dysfunction. At the spinal cord level, it attenuated neuroinflammation and promoted axonal regeneration. In parallel, hUC-MSCs transplantation ameliorated hippocampal abnormalities, including microglial activation, synaptic remodeling, and impaired synaptic pruning, accompanied by improved hippocampus-dependent behaviors. Collectively, these findings supported the concept of “brain–spinal co-repair” as a concurrent therapeutic phenomenon in which hUC-MSCs transplantation ameliorates pathological alterations in both the spinal cord and hippocampus following SCI.

4.1. Rationale for sequential intrathecal hUC-MSCs transplantation after SCI

MSCs have been widely investigated for SCI repair due to their neuroprotective and regenerative properties. Compared with bone marrow- and adipose-derived MSCs, hUC-MSCs exhibited superior repair efficacy in nerve injury models [46]. Consistent with these characteristics, the hUC-MSCs used in this study displayed typical MSC characteristics and were transplanted intrathecally at 7 d post-injury, a time window previously shown to support tissue repair after SCI [47]. Although intrathecal delivery was minimally invasive and allowed repeated injections, cell survival may be limited by the post-injury microenvironment [48,49]. Indeed, longitudinal bioluminescence imaging revealed that transplanted hUC-MSCs survived for only 3–7 d after each administration. Therefore, we adopted a sequential transplantation strategy involving weekly injections for 4 consecutive weeks, as reported previously [50]. Notably, similar survival kinetics following the first and second transplantations suggested that repeated intrathecal administration did not induce overt immune rejection, consistent with the low immunogenicity of MSCs [51]. Furthermore, no HuNu-positive cells were detected in either the dHPC or vHPC, suggesting that transplanted hUC-MSCs did not establish long-term residence in this brain region. However, transient migration to other brain regions or indirect actions mediated through CSF circulation cannot be completely excluded.

4.2. hUC-MSCs transplantation promoted motor function recovery by regulating the local inflammatory microenvironment after SCI

Studies have confirmed that reactive astrocytes and activated microglia contributed to the hostile microenvironment after SCI by promoting glial scar formation, chronic neuroinflammation, and neuronal damage [52,53]. Therefore, modulation of the inflammatory microenvironment was a critical strategy for SCI repair. MSCs responded to inflammatory cues by secreting anti-inflammatory and immunomodulatory factors, including IL-10, transforming growth factor-β (TGF-β), and prostaglandin E2 (PGE2), thereby limiting scar formation and promoting axonal regeneration [54]. Consistent with these studies, hUC-MSCs transplantation enhanced hindlimb motor recovery in SCI mice, upregulated IL-10 expression, restored NF-200 immunoreactivity, and reduced GFAP, CSPGs, and IBA1 expression at the lesion site. These changes suggested that hUC-MSCs promoted a more favorable microenvironment for axonal regeneration by attenuating glial scar formation and excessive microglial accumulation. Although IL-10 likely contributed to these effects, the therapeutic benefits were unlikely to be mediated by IL-10 alone, as MSCs regulate inflammation through multiple paracrine mechanisms [55,56]. Notably, IL-1β and TNF-α levels had largely returned to baseline during the chronic phase of SCI, suggesting that hUC-MSCs primarily enhanced the regenerative microenvironment rather than directly suppressing pro-inflammatory cytokines. This interpretation was supported by the increased density of Nissl bodies observed at the lesion. Importantly, these findings were consistent with clinical studies reporting improved neurological outcomes following hUC-MSCs transplantation in SCI patients [57]. Collectively, our results highlight the potential role for hUC-MSCs in microenvironment regulation and functional recovery after SCI.

4.3. hUC-MSCs transplantation promoted neurological function recovery in female mice with SCI by improving hippocampal synaptic remodeling

Clinical studies have shown that SCI was frequently accompanied by persistent emotional and cognitive impairments in addition to motor dysfunction, with female patients being particularly susceptible to these complications [58]. Consistent with these observations, female mice with SCI exhibited marked emotional disturbances and cognitive deficits in this study. Importantly, hUC-MSCs transplantation alleviated these behavioral abnormalities. Additionally, the lack of significant correlations between BMS scores and emotional/cognitive indices suggested that these benefits were unlikely to be merely secondary to locomotor recovery, but may instead reflect direct modulation of brain dysfunction by hUC-MSCs.

Studies showed that MSCs transplantation exerted neuroprotective and restorative effects in neuropsychiatric disorders by promoting endogenous neurogenesis in the hippocampal dentate gyrus through the secretion of neurotrophic factors [59]. Li et al. [60] found that hUC-MSCs transplantation significantly ameliorated anxiety-like and depression-like phenotypes in chronic unpredictable mild stress (CUMS) models by modulating the inflammatory microenvironment within the CNS. Regarding cognitive function, MSCs transplantation improved memory deficits after traumatic brain injury (TBI) by enhancing the expression of synaptic plasticity-related proteins in the hippocampus [61]. Consistent with these findings, this study suggested that hUC-MSCs transplantation was associated with improved emotional and cognitive function, accompanied by partial restoration of HPA axis homeostasis and region-specific normalization of hippocampal synaptic remodeling. Specifically, SCI resulted in elevated serum CORT levels accompanied by reduced CRH and ACTH levels, a pattern consistent with HPA axis dysregulation under chronic stress conditions [62]. Furthermore, impaired pituitary responsiveness and altered sensitivity to glucocorticoid feedback led to asynchronous changes in CRH, ACTH, and CORT levels, resulting in hormonal imbalance during disease progression or recovery [63]. Consistent with these observations, hUC-MSCs transplantation reduced CORT levels and restored ACTH secretion, whereas CRH levels remained unchanged. The mechanisms underlying this hormonal response require further investigation.

Synaptic plasticity is fundamental to the regulation of emotional and cognitive functions. Previous studies have shown that HPA axis dysfunction was often accompanied by hippocampal synaptic remodeling, which was characterized by synaptic dysfunction and impaired dendritic spine maturation, thereby contributing to the development of neuropsychiatric disorders [[64], [65], [66]]. In the present study, SCI induced region-specific synaptic remodeling within the hippocampal CA1 subregion. In the dHPC, excessive dendritic branching, an increased proportion of immature dendritic spines, and elevated SYN1 expression suggested the presence of aberrant compensatory synaptic remodeling. In contrast, the vHPC exhibited reduced dendritic complexity and decreased SYN1 expression, indicating substantial synaptic loss. Consistent with our findings, Kim et al. [66] reported that SCI increased the proportion of immature dendritic spines in the motor cortex, suggesting that SCI induced widespread synaptic remodeling throughout the CNS. Notably, hUC-MSCs transplantation markedly reversed these pathological alterations and restored the homeostasis of hippocampal synaptic structure and synapse-related proteins. Previous studies have demonstrated that hUC-MSCs transplantation improved emotional and cognitive function by attenuating hippocampal neuroinflammation, oxidative stress, and neuronal loss, while MSC-derived neurotrophic factors further promoted synaptic repair [67,68]. Moreover, synaptic morphology is closely linked to synaptic function, and its homeostasis is regulated by neurotrophic factors [69]. Therefore, we speculated that region-specific changes in BDNF expression may be associated with hippocampal synaptic remodeling after hUC-MSCs transplantation, although the causal relationship remains to be determined. In the vHPC, hUC-MSCs transplantation may promote dendritic complexity and the formation of mature dendritic spines in CA1 neurons by upregulating BDNF expression, thereby improving emotional function. Supporting this notion, enhanced BDNF signaling has been shown to increase dendritic spine density and alleviate emotional disturbances [70]. In addition, MSC transplantation has been reported to upregulate synaptophysin expression and facilitate neural network reconstruction [71]. Accordingly, the increased SYN1 expression observed in the vHPC may contribute to enhanced presynaptic function and the remodeling of emotion-related neural circuits. Interestingly, hUC-MSCs transplantation did not affect BDNF expression in the dHPC. This phenomenon may be associated with the aberrant compensatory synaptic remodeling observed in this region. Combined with our findings, the restoration of synaptic structure in the dHPC may rely more heavily on the recovery of synaptic pruning function rather than BDNF-mediated neurotrophic effects, which may partly explain why spatial memory deficits were not completely rescued in SCI mice. Furthermore, hUC-MSCs transplantation did not alter the expression of postsynaptic markers, suggesting that its regulatory effects on hippocampal synaptic plasticity after SCI were primarily mediated through presynaptic structures and related signaling pathways. Studies showed that multiple neurotrophic factors, including neurotrophin-3 (NT-3) and glial cell line-derived neurotrophic factor (GDNF), may also contribute to neural repair after SCI [72]. Future studies are needed to further elucidate the roles of other neurotrophic factors and their potential interactions with BDNF in regulating hippocampal synaptic plasticity following SCI.

4.4. hUC-MSCs transplantation improved synaptic remodeling by enhancing microglial synaptic pruning in the dHPC CA1 region

Glial cells are essential components of the CNS that maintain neuronal function and synaptic homeostasis [73]. Previous studies have shown that sustained activation of microglia and astrocytes following CNS injury triggered chronic neuroinflammation and contributed to neurodegeneration through mechanisms including oxidative stress, synaptic damage, and neuronal apoptosis [74,75]. In the present study, SCI induced persistent activation of glial cells in the hippocampus. Consistent with our findings, Jure et al. [76] reported that SCI activated hippocampal astrocytes and microglia, impaired neuronal survival, and consequently led to cognitive deficits and depression-like behaviors. Notably, hUC-MSCs transplantation effectively attenuated SCI-induced neuroinflammation in the hippocampus. Previous studies have likewise demonstrated that MSCs alleviated emotional and cognitive disorders in TBI and CUMS models by suppressing excessive glial activation [60,77].

Studies have shown that microglia maintained neural network homeostasis by engulfing and pruning excessive or dysfunctional synaptic connections, a process that depended on lysosomal activity and dynamic morphological changes [78,79]. Impairment of synaptic pruning induced aberrant synaptic remodeling and neuroimmune dysregulation, thereby disrupting the establishment of mature hippocampal circuits and contributing to the development of emotional and cognitive dysfunction [80]. In this study, SCI specifically reduced microglial branch complexity and phagocytic capacity in the dHPC CA1 region, accompanied by excessive dendritic branching and increased SYN1 expression. These findings suggested that impaired microglia-mediated synaptic pruning may be an important contributor to synaptic homeostasis disruption in this region. Similar phenomena have also been reported in other chronic stress models [81]. Notably, hUC-MSCs transplantation significantly restored microglial branch complexity and the pruning of presynaptic inputs in the dHPC CA1 region, thereby providing functional support for the correction of aberrant compensatory synaptic remodeling. In contrast, no comparable changes were observed in the vHPC, which was consistent with consistent with the synaptic loss observed in this region. Emerging evidence has indicated that astrocytes regulated microglia-mediated engulfment of presynaptic inputs through complement-dependent signaling, particularly the C3/C3aR pathway [82]. Reactive astrocytes released C3, which bound to C3aR expressed on microglia and modulated their phagocytic activity, thereby influencing synaptic remodeling and neuroinflammation [83]. Studies have shown that hUC-MSCs suppressed excessive activation of this pathway and improved synaptic remodeling [60]. Although our study did not directly investigate this pathway, the concurrent attenuation of astrocytic and microglial activation after hUC-MSCs transplantation suggested that astrocyte–microglia crosstalk may contribute to the observed hippocampal remodeling after SCI. Interestingly, whether microglia directly engulf intact synaptic structures remains controversial. Some studies suggested that microglia regulated synaptic plasticity through the selective degradation of presynaptic terminals rather than whole synapses, and this process may occur independently of complement signaling pathways [84], highlighting the complexity and specificity of this regulatory mechanism. Collectively, these findings indicated that the region-specific effects of hUC-MSCs transplantation on the hippocampus after SCI were reflected not only in synaptic remodeling but also in the differential regulation of microglia-mediated synaptic pruning between the dHPC and vHPC.

4.5. Limitations

This study has several limitations. First, consistent with previous reports in SCI and TBI models [85], transplanted hUC-MSCs survived for only 3–7 d and showed no evidence of migration or long-term engraftment in either the injured spinal cord or the hippocampus. Therefore, the therapeutic effects were more likely associated with paracrine and immunomodulatory mechanisms than with direct cell replacement, although direct evidence for this interpretation remains lacking. Previous studies have demonstrated that hUC-MSC-derived trophic factors and extracellular vesicles contribute to neural repair [86,87]. For example, MSC-derived extracellular vesicles carrying miR-124 suppressed excessive glial activation and neuroinflammation, whereas hUC-MSC-derived extracellular vesicles carrying miR-149 promoted neuronal survival and preserved BSCB integrity through neurotrophic receptor-related signaling pathways [88,89]. However, the key mediators responsible for the observed “brain–spinal co-repair” phenomenon remain unidentified. Second, although the hUC-MSCs used in this study met ISCT criteria, potential biological variations among different donors and production batches were not evaluated. In addition, xenogeneic transplantation was performed without immunosuppression. Repeated intrathecal injections did not result in accelerated cell clearance and MSCs were generally considered to possess low immunogenicity, but host adaptive immune responses were not directly assessed. Therefore, the potential influence of local immune responses on cell survival and therapeutic efficacy cannot be excluded.

Furthermore, only female mice were used as research subjects. Although hUC-MSCs transplantation improved emotional and cognitive impairments following SCI, the generalizability of these results requires further validation. Existing studies have reported sex differences in HPA axis function, with female animals exhibiting higher and more prolonged ACTH and CORT secretion [37]. Sex-dependent differences have also been observed in neuroinflammation and behavioral outcomes after stress exposure, with microglia-mediated inflammation being more prominent in females [90]. Consistently, a previous study demonstrated sex-related differences in emotional outcomes following SCI in different rat strains [91]. Therefore, subsequent studies should evaluate the neurorestorative effects of hUC-MSCs on SCI in rodents of different sexes and strains to improve translational relevance. Finally, although hUC-MSCs regulated microglia-mediated synaptic remodeling in the hippocampal CA1 region after SCI, the underlying molecular mechanisms and the interactions among microglia, astrocytes, and neurons remain to be elucidated.

4.6. Prospects for future research

Although alterations in microglial synaptic pruning and BDNF expression were observed following hUC-MSCs transplantation, the causal relationships among these changes were not directly established. Future studies employing pathway-specific interventions or genetic approaches will be required to determine whether these processes are mechanistically linked. In addition, optimization of the transplantation window and dosage regimen, together with validation in non-human primate models, will be essential for clinical translation. Advanced approaches such as single-cell RNA sequencing, spatial transcriptomics, and secretome profiling may further clarify the long-term fate of transplanted cells and elucidate the molecular mechanisms underlying the proposed “brain–spinal co-repair” effect. Comparative analyses of hUC-MSCs derived from different donors and production batches may also facilitate the development of more effective cell-based or cell-free therapies for SCI.

5. Conclusions

This study highlight that effective SCI treatment may require simultaneous modulation of pathological processes in both the injured spinal cord and brain regions. Sequential intrathecal transplantation of hUC-MSCs during the subacute phase promoted multidimensional neurological recovery in female mice with SCI. These therapeutic effects were accompanied by coordinated pathological improvements in both the spinal cord and hippocampus, including attenuation of neuroinflammation, amelioration of HPA axis dysregulation, and region-specific regulation of microglia-mediated synaptic remodeling. This work extended the current therapeutic framework beyond conventional spinal cord repair toward an integrated paradigm, thereby providing new insights into comprehensive neurological rehabilitation after SCI.

Data availability

The data in this study were shown in the main text and supplementary. The data that support the findings of this study are available from the corresponding author upon reasonable request.

CRediT authorship contribution statement

Boya Huang: Investigation, Data curation, Methodology, Validation, Writing-original draft. Jiahao Ren: Methodology, Data curation, Formal analysis, Visualization. Xinjie Liu: Methodology, Data curation, Formal analysis, Validation. Mi Zhou: Investigation, Methodology, Validation. Yilin Pang: Investigation, Methodology, Validation. Zemin Cao: Methodology, Validation. Xiaofeng Lin: Methodology, Validation. Xiang Gao: Methodology, Software. Baoyou Fan: Methodology. Yan Hao: Methodology. Shen Liu: Methodology. Hui Shen: Methodology. Guangzhi Ning: Writing-review & editing, Supervision, Project administration. Xue Yao: Writing-review & editing, Validation, Supervision, Project administration. Shiqing Feng: Writing-review & editing, Resources, Validation, Supervision, Project administration, Funding acquisition.

Ethical approval and informed consent

All experiments adhered to animal welfare standards and ethical requirements, and were approved by the Experimental Animal Welfare and Ethics Committee of Tianjin Medical University General Hospital (Tianjin, China; Approval No: IRB2023-DW-132).

Declaration of generative AI in scientific writing

No generative artificial intelligence (AI) or AI-assisted technologies were used in the preparation of this manuscript.

Funding statement

This work was supported by grants from the National Natural Science Foundation of China (82272470, 82472414), the Tianjin Key Medical Discipline Construct Project (TJYXZDXK-3-005A), the Key Discipline Special Project of Tianjin Health and Technology Project (TJWJ2022XK002), the Beijing-Tianjin-Hebei Basic Research Cooperation Special Project (22JCZXJC00050), the Tianjin Science and Technology Project (21JCYBJC01780).

Declaration of competing interests

The authors have no conflicts of interest relevant to this article.

Appendix A. Supplementary data

The following is the Supplementary data to this article: