Mesenchymal stem cells reduce long-term cognitive deficits and attenuate myelin disintegration and microglia activation following repetitive traumatic brain injury

Abstract

The underlying mechanisms for the beneficial effects exerted by bone marrow-mesenchymal stem cells (BM-MSCs) in treating repetitive traumatic brain injury (rTBI)-induced long-term sensorimotor/cognitive impairments are not fully elucidated. Herein, we aimed to explore whether BM-MSCs therapy protects against rTBI-induced long-term neurobehavioral disorders in rats via normalizing white matter integrity and gray matter microglial response. Rats were subjected to repeated mild lateral fluid percussion on day 0 and day 3. On the fourth day post-surgery, MSCs groups received MSCs (4 × 10⁶ cells/ml/kg, intravenously) and were assessed by the radial maze, Y maze, passive avoidance tests, and modified neurological severity scores. Hematoxylin & eosin, and Luxol fast blue stainings were used to examine the histopathology and white matter thickness. At the same time, immunofluorescence staining was used to investigate the numbers of tumor necrosis factor-alpha (TNF-α)-containing microglia in gray matter. Three to nine months after neurotrauma, rats displayed sensorimotor and cognitive impairments, reduced thickness in white matter, and over-accumulation of TNF-α-containing microglia and cellular damage in gray matter. Therapy with BM-MSCs significantly attenuated the rTBI-induced sensorimotor and cognitive impairments and all their complications. Mesenchymal stem cell therapy might accelerate the recovery of sensorimotor and cognitive impairments in rats with rTBI via normalizing myelin integrity and microglia response.

Introduction

Epidemiological studies revealed that repetitive head injury during sports might contribute to clinical syndromes of cognitive and sensorimotor impairments.^1–3 Imaging studies further showed that the location of white matter injury could predict cognitive function in patients after traumatic brain injury (TBI). ⁴ Mice with chronic TBI also displayed progressive white matter degeneration. ⁵ White matter injury contributes significantly to long-term cognitive and sensorimotor deficits. ⁶ In addition to white matter degeneration, TBI displayed reactive microglia in patients.^2,3,7 Our previous studies further demonstrated that microglia expression of tumor necrosis factor-alpha (TNF-α) in the gray matter was associated with cognitive and sensorimotor impairments in rats with TBI.^8,9 It can be derived from these observations that both the white matter integrity and gray matter expression of TNF-α-containing microglia play a role in the pathogenesis of TBI-induced neurobehavioral disorders.

The administration of bone marrow-derived mesenchymal stem cells (BM-MSCs) through systemic or central routes has shown potential in mitigating the motor and cognitive impairments caused by TBI.^10–13 This therapeutic approach is effective during acute, sub-acute, or chronic phases of TBI. The positive outcomes of BM-MSC therapy are thought to arise from its ability to enhance the endogenous repair and remodeling processes in the brain of the recipient. ¹⁴ Additionally, this therapy has demonstrated a notable decrease in the number of microglia, macrophages, neutrophils, and apoptotic neurons in the affected cortical area, along with a reduction in pro-inflammatory cytokine levels. ¹⁵ This raises the possibility that BM-MSC therapy protects against TBI-induced cognitive and sensorimotor impairments by affecting both the white matter integrity and gray expression of TNF-α-containing microglia.

To address the aforementioned inquiries, our study employed a combination of neurobehavioral and histopathological analyses, utilizing the same group of animals. The purpose of these assessments was to evaluate the extent of neurological deficits and the degree of brain injury in both hemispheres, focusing on two distinct stages: the sub-acute phase (7 days post-injury) and the chronic phase (ranging from 1 to 9 months post-injury).

Materials and methods

Animals

Adult male Wistar rats, specific pathogen-free and aged 10 weeks with body weights ranging from 350 to 380 grams, were acquired from BioLASCO Taiwan Co., Ltd in Taipei, Taiwan. These rats were accommodated in the Central Animal Facility at Chi Mei Medical Center, housing in groups of four. The environmental conditions were maintained at a constant temperature of 22 ± 1 °C and followed a 12-h light-dark cycle. Pellet rat chow and tap water were available ad libitum, and their body weights were recorded every month. The experimental protocols were approved by the Institutional Animal Care and Use Committee at Chi Mei Medical Center (approval number 108101805). All animal experiments were designed following ARRIVE guidelines. Using randomization list methods (https://www.randomizer.org/), two hundred male Wistar rats were divided into sham operation with normal saline (NS) treatment (Sham + NS), sham operation with mesenchymal stem cell treatment (Sham + MSC), repetitive TBI (rTBI) with NS treatment (rTBI + NS), and rTBI with MSC treatment (rTBI + MSC) groups. Rats received NS (1 ml/kg) or MSC (4 × 10⁶ cells/ml/kg) at day four post-surgery. Individual rats were identified using Stoelting™ Rat Ear Tags, supplied by Stoelting Co., Illinois, USA.

A rat model of rTBI

All animals were anesthetized with Zoletil®100 (20 mg/kg, intraperitoneally; Virbac, Carros, France). Following the previous methodology, ¹⁶ rats were surgically prepared for lateral fluid percussion (LFP). Rats were subjected to repeated mild (1.5 atm) LFP on days 0 and 3. Under anesthesia, each rat was positioned in a stereotaxic frame, and a sagittal incision was made in the scalp. A craniotomy of 4.8 mm diameter was then performed at coordinates A/P −3 mm, M/L 4 mm from the bregma, specifically over the right parietal cortex, 3.0 mm to the right of the midline. ¹⁷ For the LFP injury, a modified leur-lock connector (trauma cannula) with an inner diameter of 2.6 mm was fixed into the craniotomy using dental acrylic. This connector was then attached to a fluid percussion device (sourced from VCU Biomedical Engineering, Richmond, VA, USA). Post-impact, approximately 30% of rats exhibited seizures immediately after LFP induction, lasting 10–15 s. LFP also replicated other TBI features like hematoma formation, acute apnea (usually 10–60 s), and bradycardia. ¹⁸ On day 3, rats were re-anesthetized for either a sham operation or mild LFP. Post-surgery, the cannula was removed, and the incision was sutured. The sham group underwent the same surgical steps, except the LFP, and remained under anesthesia for the duration of experiments. A heating pad maintained the rats’ body temperature during surgery and recovery at 36.5 ^oC. Afterward, the rats were returned to their cages, kept at an ambient temperature of 26 °C, and given appropriate care, including food and hydration. Ketoprofen (5 mg/kg/BID, subcutaneous injection) was administered for postoperative analgesia.

In this research, neurobehavioral testing and immunohistochemical analysis were utilized to examine the pathophysiological alterations in the cerebral microenvironment as part of a longitudinal assessment of the effects of rTBI. These evaluations were systematically conducted during both the acute phase (7 days post-injury) and the extended chronic phase (from 1 to 9 months post-injury). Continuous daily monitoring of the rats was implemented to track their survival status, alongside meticulous observations of their nutritional intake and waste output. Inclusion criteria for rTBI rats in this study are a righting reflex time (325 s), indicative of a mild to moderate injury, ¹⁹ and no disruption to the underlying dura. Sham rats spontaneously righted (∼20 s). Notably, a small percentage of the rats (less than 3%), which succumbed shortly after the surgical intervention, were excluded from the final data set and analysis of the study.

Mesenchymal stem cell preparation

Human BM-MSCs were purchased from EMD Millipore (#SCC034, Billerica, MA, USA) and cultured in low glucose DMEM containing 10% fetal bovine serum and 1% penicillin and streptomycin, both sourced from Invitrogen in Paisley, UK, and Thermo Scientific in Loughborough, UK, respectively. This protocol aligns with previously described methods. ⁸ The cells, utilized between passages 3–7, underwent phenotyping using Novocyte flow cytometry from ACEA Biosciences (San Diego, USA) and a range of monoclonal antibodies (CD34, CD45, CD90, CD73, and CD105), all purchased from BD Pharmingen (San Diego, CA, USA). MSCs were identified based on the presence (CD90+, CD73+, CD105+) and absence (CD31−, CD34−, CD45−) of specific markers. In vitro differentiation assays (Lonza Walkersville, Inc., Walkersville, MD, USA) are performed according to the manufacturer's instructions. The capability of trilineage differentiation (osteocytes, chondrocytes, and adipocytes) revealed positive results.

Study design

Before behavioral testing, each rat underwent a 7-day acclimation period in the research facility. Following this, they were familiarized with the Y maze, radial maze, and passive avoidance apparatus over 3 days, exploring each for 10 min daily. One day before surgery, rats received behavioral pre-testing. At indicated times following the surgery, the animals were subjected to behavioral tests from day 7 to month 9 following the first impact or surgery. Following the final behavioral assessment, rats were thoroughly anesthetized using Zoletil®100 (intraperitoneally administered). They then underwent transcardial perfusion, initially with 0.9% sodium chloride solution, followed by 4% paraformaldehyde obtained from Merck & Co. Inc., Kenilworth, NJ, USA. Subsequently, their brains were carefully collected and placed in formalin for an overnight fixation process. Serial brain sections were processed for staining, ranging from +0.8 mm to −5.3 mm relative to the bregma and spanning from the cortex to the hypothalamus. These sections were treated with hematoxylin and eosin (H&E) for general tissue structure visualization, Luxol fast blue (LFB) for myelin staining, or immunohistochemical procedures for specific antigen detection.

We used 6–10 rats/group for the behavioral tests and 10 rats/time point/group for histology stain for statistical validity. Researchers who were unknown to the experimental groups did behavioral and histological assessments.

Histological stainings

Serial coronal sections, each 5 μm thick, were prepared using a microtome and adhered to positively charged glass slides. Prior to staining, the sections were subjected to deparaffinization using xylene and then rehydrated progressively through a graded ethanol-to-water series. The sections were stained with H&E using conventional histological methods to evaluate cellular morphology. To observe white matter, adjacent sections were treated with LFB staining. Observations of each slide were made using a bright-field microscope from Zeiss Gmbh, Gottingen, Germany.

For immunofluorescence studies, other adjacent sections were stained with ionized calcium-binding adaptor molecule (Iba1, 1:500; sourced from Abcam, Cambridge, MA, USA) and TNF-α (1:1000; from Millipore, Billerica, MA, USA). Visualization of Iba1 + and TNF-α+ cells was achieved using Alexa Fluor 488-conjugated and Alexa Fluor 568-conjugated goat anti-rabbit IgG (both at 1:400, from Invitrogen, CA, USA), with excitation and emission wavelengths of 495/525 nm and 578/603 nm, respectively. Additionally, DAPI staining (1: 25,000, from ThermoFisher, #62247) was applied to highlight nucleated cells, using excitation and emission wavelengths of 360 nm and 460 nm. Control sections were incubated with all reagents except the primary antibody to serve as negative controls. The Iba1/TNF-α double-positive cells were captured at 100× magnification using the Axiovision microscope system (Zeiss Gmbh, Gottingen, Germany and counted by Axiovision image analysis software (Zeiss). This counting was performed by an investigator who was not aware of the treatment status of the animals. Regions of interest (ROIs) were assessed bilaterally for each image at 100× magnification. These ROIs in TBI animals were then compared with those on the ipsilateral side of the injury in Sham animals.

Neurobehavioral tests

The neurobehavioral tests used in this study include the radial maze assay, ²⁰ the Y maze, ²¹ passive avoidance assay, ²² and the modified neurological severity score (mNSS). ²³ In brief, during the radial maze assessments, each rat underwent daily training with one trial per day for 5 consecutive days before surgery. Finally, the number of long-term memory (reference memory) errors and short-term memory (working memory) errors were counted. The spontaneous alternation behaviors requiring attenuation and working memory were assessed in a Y-maze test. In passive avoidance tests, both the waiting time and the number of errors were counted to evaluate the ability of learning and memory. The mNSS integrates motor, sensory, and reflex assessments to evaluate neurological impairment following rTBI.

Statistical analysis

No outliers were excluded from the statistical analysis. The analysts were masked to the condition in any experiment in which the data points were manually obtained. Statistical analyses were performed using Graph Pad Prism 7.01 (Graph Pad Software Inc., CA, USA. RRID: SCR-002798). Survival rates were compared with the Kaplan–Meier survival analysis followed by the log-rank significance test. ²⁴ Behavioral performance data were analyzed using two-way ANOVA, followed by Tukey's multiple comparisons test. ²⁵ The Kruskal–Wallis tests with Dunn's post-hoc tests analyzed the histological staining and immunostaining data. ²⁶ Data were presented as mean ± SD. p-values <0.05 were considered statistically significant.

Results

Our previous results ²⁷ showed that the TBI + 1 × 10⁶ MSc group and TBI + 4 × 10⁶ MSC group rats had significantly lower cerebral contusion volumes and contusion areas than did the TBI + vehicle group of rats up to 28 days post-TBI. However, the beneficial effects on cerebral contusion of the higher dose (4 × 10⁶ MSCs) were not superior to those of the dose (1 × 10⁶ MSCs). In the present study, described below, only the 4 × 10⁶ MSCs were used for long-term effects examination.

Mesenchymal stem cell therapy increases the survival ratio in adult male rats with rTBI

The survival ratios were first analyzed to assess the potential reduction in the lethality of experimental rTBI caused by MSCs treatment. As demonstrated in Figure 1, up to 9 months post-rTBI, rTBI significantly decreased the percent of survival from 95% in the sham + NS group rats to 62% in the rTBI + NS group rats (p < 0.01). However, MSC treatment significantly increased the percentage of survival from 62% in the rTBI + NS group rats to 71% in the rTBI + MSCs group rats (Figure 1).

Figure 1.

Kaplan–Mater analysis followed by log-rank tests was performed to determine the percent survival (or survival ratio) in each group. Sham + NS, sham operation with normal saline-treated; sham + MSCs, sham operation with MSCs-treated; rTBI + NS, repeated TBI with NS-treated; rTBI + MSCs, repeated TBI with MSCs-treated. rTBI: repetitive traumatic brain injury; NS: normal saline.

Mesenchymal stem cell therapy accelerates recovery of neurobehavioral disorders in adult male rats with rTBI

Our study encompassed various assessments to evaluate different aspects of neurological function in rats. These included the radial maze, Y maze, and passive avoidance tests for cognitive function assessment, the mNSS for overall neurological function evaluation. In the radial maze assay, it was observed that the rats in the rTBI + NS group exhibited significantly increased retention times (or latency in seconds; analysis with a mixed effects two-way ANOVA model found significant effects of the group [F(3195) = 84.52, p < 0.0001], time [F(5195) = 8.824, p < 0.0001], and group × time interaction [F(15,195) = 4.746, p < 0.0001]) and higher counts of both working memory errors (analysis with a mixed effects two-way ANOVA model found significant effects of the group [F(3195) = 328.6, p < 0.0001], time [F(5195) = 24.64, p < 0.0001], and group × time interaction [F(15,195) = 15.8, p < 0.0001]) compared to the Sham + NS group rats. Post-hoc comparisons with Tukey's multiple comparison test indicated significant increase in latency and number of working memory errors in rTBI on month day 7 (D7), day 28 (D28), month 3 (M3), month 6 (M6), and months 9 (M9) compared to Sham + NS D7 to M9 values (p < 0.05), as illustrated in Figure 2a and 2b.

Figure 2.

The recovery of rTBI-induced neurobehavioral disorders can be accelerated by MSCs therapy. The behavioral test battery consisted of three memory tests (radial maze, Y-maze, and passive avoidance), one motor test (rotarod), and one modified neurological severity score (mNSS) was performed before the surgery (pre) and at 7 days (D7) to 9 months (M9) after surgery. For the radial-arm maze test, the retention time (latency period, second) (a) and working memory errors (b) were determined. For the Y-maze test, the levels of alternation were determined (c). The retention time (d) for the passive avoidance test was determined. The mNSS assayed the neurological deficits (e). Data are presented as means ± SD (n = 12-18 per group). *p < 0.05, Sham + NS vs. rTBI + NS group; and ⁺p < 0.05 rTBI + NS vs. rTBI + MSC group. rTBI: repetitive traumatic brain injury; NS: normal saline.

In the Y-maze test, the rTBI + NS group of rats displayed a notable decrease in the percentage of alternations compared to the Sham + NS group, indicating a significant difference (p < 0.05) at time points D7, D28, M3, M6, and M9, as shown in Figure 1c (analysis with a mixed effects two-way ANOVA model found significant effects of the group [F(3195) = 73.37, p < 0.0001], time [F(5195) = 24.63, p < 0.0001], and group × time interaction [F(15,195) = 5.287, p < 0.0001]). Conversely, the rTBI + MSCs group exhibited a significantly higher percentage of alternations than the rTBI + NS group during the same intervals (p < 0.05, Figure 2c).

In passive avoidance testing, the rTBI + NS group demonstrated notably shorter retention times than the Sham + NS group from D7 to M9, a statistically significant difference (Figure 2d; analysis with a mixed effects two-way ANOVA model found significant effects of the group [F(3195) = 193.3, p < 0.0001], time [F(5195) = 17.69, p < 0.0001], and group × time interaction [F(15,195) = 11.47, p < 0.0001]). In contrast, the rTBI + MSCs group showed significantly longer retention times compared to the rTBI + NS group (p < 0.01, Figure 2d).

Furthermore, the mNSS results indicated that the rTBI + NS group had significantly higher scores compared to the Sham + NS group from D7 to M9 (Figure 2e; analysis with a mixed effects two-way ANOVA model found significant effects of the group [F(3195) = 520.4, p < 0.0001], time [F(5195) = 74.1, p < 0.0001], and group × time interaction [F(15,195) = 26.23, p < 0.0001]). However, the rTBI + MSCs group exhibited significantly lower mNSS scores than the rTBI + NS group, particularly from D7 to D28, as depicted in Figure 2e.

MSCs therapy attenuates the cellular damage scores of gray matter in rTBI rats

H&E stainings revealed that compared to the Sham + NS group rats, the rTBI + NS group rats had significantly higher levels of damage scores in the hippocampus cornu ammonis (CA) CA1, CA2, CA3, and DG regions, and cortex evaluated at D7 to M9 (Figure 3a–e; [KW test: H = 162.1 in CA1 region, H = 165.9 in CA2 region, H = 156.8 in CA3 region, H = 152.3 in DG region, and H = 159.5 in cortex region, all p < 0.001]). However, compared to the rTBI + NS group rats evaluated at D7 to M9, scores rTBI + MSC group rats had significantly lower damage values in these brain regions (p < 0.05, Figure 3a–e). In the hypothalamus, there were insignificant differences between each group at D7 to M9 post-surgery (Figure 3d; KW test: H = 18.82, p = 0.469). In the striatum region, the TBI-induced brain damage was shown only at M9, which could be reduced following MSC therapy (Figure 3e; KW test: H = 43.05, p = 0.001).

Figure 3.

MSCs attenuate rTBI-induced cellular damage in gray matter. (a) Representative photographs of H & E stained sections for different groups of rat brains. The boxed areas in the full-size brain image on the top left graphic are magnified in the representative images. Following brain injury, cell shrinkage with dark stained pyknotic nuclei and perineuronal vacuolization in the (b) hippocampus, (c) cortex, (d) hypothalamus (Hypo.), and (e) striatum (Stri.) were noted. Each bar graph represents the mean ± SD. (n = 10 each group) of neuronal damage scores at different time points after brain trauma for each experimental group of rats. *p < 0.05 vs. the Sham + NS group; ⁺p < 0.05, vs. the rTBI + NS group. CA = cornu ammonis; CA1, CA2, and CA3 are hippocampus subfields; DG = dentate gyrus; D = day; M = month. rTBI: repetitive traumatic brain injury; NS: normal saline.

MSCs therapy attenuates the reduced thickness of white matter of brain hemispheres in rTBI rats

Luxol blue staining, as shown in Figure 4a, indicated that the thickness of the ipsilateral and contralateral corpus callosum body (CCb), as well as the ipsilateral and contralateral angular bundle (ab), was significantly lower ([KW test: H = 168.3 in CCb and H = 155.5 in ab, all p < 0.001]) in the rTBI + NS group compared to the Sham + NS group. This difference was consistent across the time points from D7 to M9, as detailed in Figure 4b, 4c, 4d, and 4e. In contrast, the rTBI + MSCs group exhibited significantly higher thickness values in these same regions, as presented in Figure 4b–e.

Figure 4.

MSCs attenuate rTBI-induced reduced thickness of white matter. (a) Representative photographs of Luxol fast blue stained sections for different groups of rats’ brains. The boxed areas in the full-size brain image on the top left graphic are magnified in the representative images. Each bar represents the mean ± SD (n = 10 for each group) of the thickness of white matter or intensity of white matter regions at different time points after rTBI for each group of rats, including (b) ipsilateral CCb, (c) contralateral CCb, (d) ipsilateral ab, (e) contralateral ab, (f) LD, (g)VPL/VPM, and (h) ic. *p < 0.05 vs. the sham + NS group; ⁺p < 0.05 vs. rTBI + NS group. CCb = corpus callosum body; ab = angular bundle; LD = laterodorsal thalamic nuclei; VPL = ventroposterolateral; VPM = ventroposteromedial; ic = internal capsule. rTBI: repetitive traumatic brain injury; NS: normal saline.

Additionally, Figure 4f–h showed that, in comparison to the Sham + NS group, the rTBI + NS group did not exhibit significant changes in the intensity (as a percentage of the sham controls) of three white matter regions: the laterodorsal thalamic nuclei (LD), ventroposterolateral/ventroposteromedial (VPL/VPM), and internal capsule (ic), with the insignificant difference ([KW test: H = 18.18 in LD, H = 25.34 in VPL/VPM, and H = 21.03 in ic, all p > 0.05]).

MSCs attenuate rTBI-induced increased numbers of TNF-α-containing microglia in gray matter

Immunofluorescence staining (Figure 5) revealed that compared to Sham + NS group rats, rTBI + NS group rats had significantly higher numbers of TNF-α-containing microglia in many gray matters including hippocampus (KW test: H = 126.5, p < 0.0001), cortex (KW test: H = 149.2, p < 0.0001), hypothalamus (KW test: H = 76.3, p < 0.0001), and striatum (KW test: H = 80.57, p < 0.0001) (Figure 5b–e) at D7 to M9. Again, compared to rTBI + NS group rats, rTBI + MSCs group rats had significantly lower numbers of TNF-α-containing microglia in these gray matters (Figure 5).

Figure 5.

MSCs attenuate rTBI-induced numbers of TNF-α-containing microglia cells of the gray matter in rat brains. (a) Representative photographs of the immunofluorescence staining sections for different groups of rats’ brains. Each bar represents the mean ± SD (n = 10 for each group) of the number of the TNF-α-containing microglial cells in (b) hippocampus, (c) cortex, (d) hypothalamus, and (e) striatum at different time points after rTBI for each group of rats. *p < 0.05 vs. the Sham group; and ⁺p < 0.05 vs. rTBI + NS group. rTBI: repetitive traumatic brain injury; TNF-α: tumor necrosis factor-alpha; NS: normal saline.

Discussion

Following an accident, TBI patients display initial neurological manifestations, including an altered level of consciousness, convulsions, coma, and confusion. ²⁸ Subsequently, these patients frequently display cognitive and sensorimotor impairments, manifesting as difficulties with memory, attention, language comprehension, reading, writing, and spatial orientation tasks. ²⁸ In addition, a TBI causes blood–brain barrier disruption, diffuse axonal injuries, ²⁹ mitochondrial distress, and free radical generation within the injured brain tissues.^30,31 In the present study, we performed radial maze, Y maze, and passive avoidance to evaluate cognitive function, and mNSS to assess neurological function. Our results showed that rats with rTBI shared similar cognitive and sensorimotor deficits with the TBI patients. These neurobehavioral disorders and neuropathology lasted 3–9 months at least. After that, the rTBI-induced neurobehavioral disorders may be mediated by cellular damage to gray matters of the hippocampus, cortex, hypothalamus, and striatum, reduced thickness of corpus callosum ipsilateral, corpus callosum contralateral, ab ipsilateral, and ab contralateral, and increased numbers of TNF-α-containing microglia in the gray. Again, our present results are consistent with many previous studies. For example, patients with a single moderate to severe or repetitive moderate TBI days to years after head concussions have found white matter degeneration, reduction in corpus callosum thickness, and neuroinflammation with reactive microglia.^2,3,7,32,33 Animals with a single or repetitive moderate TBI have also shown oligodendroglial apoptosis, axonal injury, myelin loss, and persistent microglial activation.^34–36 The most striking finding of the present study is that cognitive and sensorimotor impairments with all their long-term sequelae could be ameliorated in TBI rats with MSC therapy. Intravenous administration of BM-MSCs has shown therapeutic benefits for various brain conditions, including TBI.^37,38 In experimental TBI models, the secretome from BM-MSCs has been found to facilitate the repair of damaged brain tissue by supporting the survival and proliferation of neural stem cells. ³⁹ Our previous studies have revealed that intravenous injection of the BM-MSCs secretome not only reduces neuronal loss and apoptosis but also stimulates the production of neurotrophic factors such as VEGF, leading to functional improvements in a rat model of TBI. ⁴⁰ The current findings suggest that BM-MSCs may offer a promising approach to TBI recovery. Our results indicate that BM-MSCs could potentially rectify cognitive and sensorimotor deficits in TBI-affected rats by reversing cellular damage in gray matter, decreasing the reduction in white matter thickness, and reducing the accumulation of TNF-α-containing microglia in the gray matter. The similar mechanism of MSCs involved in anti-inflammatory and anti-immunomodulatory properties after TBI has been promoted by Zhou et al.(2019). ⁴¹ In addition to their secretory ability, MSCs can selectively migrate to the injured brain tissue of the TBI rat and then differentiate into neurons and glial cells to repair damaged tissue, thereby improving sensorimotor and cognitive function.

In response to TBI, elevated TNF-α was noted in human serum, cerebrospinal fluid, and rodent serum.^42,43 Microglia activation was promoted to be a biomarker for TBI. ⁴⁴ Following a TBI, the microglia and astrocytes were activated, accompanied by the secretion of pro-inflammatory cytokines, including TNF-α, interleukin-1, and interleukin-6. In particular, microglial responses lasted for several months after a TBI. ⁴⁵ The injured and dead cells that elicit microglial responses caused tissue damage by releasing several pro-inflammatory cytokines.^36,46 Our previous findings indicated that the excessive accumulation of TNF-α-containing microglia in brain areas affected by TBI is linked to the development of cerebral contusions and motor neurological deficits. Treatments with etanercept ⁴² or astragaloside ⁴⁷ could mitigate these adverse effects. In our present study, immunofluorescence staining revealed that chronic TBI caused over-accumulation of TNF-α-containing microglia within the injured gray matters, including the hippocampus, cortex, hypothalamus, and striatum, which could be attenuated by MSCs therapy. It is likely that MSCs, like etanercept or astragaloside, improve outcomes of TBI in rats by reducing the over-accumulation of TNF-α-containing microglia within the gray matter.

Intravenous administration of BM-MSCs was seen to lower microglia activity at damaged sites in TBI rats, ¹⁵ inhibit pro-inflammatory cytokines, and elevate anti-inflammatory cytokines ⁴⁸ in TBI rats. BM-MSC treatment also improved neural regeneration and functional recovery in TBI rats.^49–52

Although our present results and others ⁴¹ have confirmed that MSCs can improve sensorimotor and cognitive impairments and the prognosis of TBI, MSCs used in clinical practice are still difficult. First, MSCs have the potential to promote tumor growth. ⁵³ Second, there are still many problems that need to be solved about the complexity of pathophysiology and the application of stem cell therapy. ⁴¹ Several clinical studies on stem cell treatment of TBI have achieved good therapeutic results, ⁴¹ but the sample size is not large enough, and there is no control group. Future studies need to elucidate the safety of stem cells, the route of injection, the time of injection, and the specific mechanisms that affect the clinical application of stem cells.

More recently, Witcher and colleagues ⁹ explored the role of microglia in neuropathological processes at various stages following TBI in mice. They focused on the acute (1-day post-injury), sub-acute (7 days post-injury), and chronic (30 days post-injury) phases. Their findings revealed a substantial increase in gene expression related to inflammation and neuropathology just 1 day after the injury, which appeared to be independent of microglial activity. However, using PLX5622 (a colony-stimulating factor 1 receptor antagonist) to deplete microglia before TBI resulted in reversing the TBI-induced changes in gene expression associated with neuroinflammation and neuropathology. They concluded that TBI leads to persistent neuroinflammation and neuronal dysfunction, which are mediated by microglial activity. Our present study showed that at 7 days to 9 months post-injury, MSCs reversed rTBI-related neurobehavioral disorders, neuronal damage, white matter disintegration, and accumulation of TNF-a-containing microglia in rats. MSCs therapy likely reverses the rTBI-related neurobehavioral disorders and neuropathology via modulating the microglial responses.

Conclusions

In conclusion, as depicted in Figure 6, rats suffering from chronic rTBI exhibited sensorimotor and cognitive deficits, decreased white matter thickness, and increased cellular damage. There was also an accumulation of TNF-α-containing microglia in the cortical, hippocampal, hypothalamic, and striatal regions. Treatment with BM-MSCs effectively mitigated these rTBI-induced impairments and associated complications. The findings suggest that MSCs therapy promotes recovery of sensorimotor and cognitive functions in rTBI-affected rats, likely by restoring myelin integrity and normalizing the microglial response.

Figure 6.

Graphs depict temporal associations driving the genesis of cognitive and sensory-motor deficits caused by a rTBI. The sequences include cellular damage and TNF-α-containing microglia within the gray matter and reduced thickness of white matter following rTBI. MSCs-mediated paracrine factors improve cognitive and sensory-motor deficits in rTBI rats by reducing gray matter affectation and white matter injury. TNF-α: tumor necrosis factor-alpha; rTBI: repetitive traumatic brain injury; MSCs: mesenchymal stem cells.

Author biographies

Lan-Wan Wang (MD, PhD) is director of Pediatric Neurology Division at Chi Mei Medical Center. Her work involves investigating neurodevelopmental disorders for autism and in very preterm children, microglia and neuron interaction in brain trauma, and the efficacy of mesenchymal stem cells in treating brain injury.

Chung-Ching Chio (MD) is honorary superintendent of Chi Mei Medical Center and former President of the Taiwan Neurosurgical Society. His research interests are pathogenesis of neurotrauma, exercise rehabilitation, and stem cell therapy.

Chien-Ming Chao (MD, MS) is attending physicians of Department of Intensive Care Medicine at Chi Mei Medical Center. His studies focus on thermoregulation, neuroinflammation, cardiology, and heat stress-related disorders.

Pi-Yu Chao (MS) is a research assistant in Cerebro-Cardiovascular Laboratory in Department of Medical Research of Chi Mei Medical Center. She conducts research on the effects of physical exercise and stem cell therapy on neurobehavioral function after traumatic brain injury in rat models.

Mao-Tsun Lin (MD, PhD) is former president of Taiwanese Basic Neuroscience Association and Taiwanese Physiological Association. He used to be the chair professor at Chi Mei Medical Center, National Cheng Kung University, and Taipei Medical University. Now, he is an advisor for Chi Mei Medical Center. He is also a pioneer in the field of heat stroke research.

Ching-Ping Chang (PhD) is head and principal investigator of Cerebro-Cardiovascular Laboratory at Chi Mei Medical Center. She conducts research on the research fields of heat stroke, traumatic brain injury, ischemic stroke, cardiac ischemia, and neurodegenerative diseases.

Hung-Jung Lin (MD) is superintendent of Chi Mei Medical Center, former president of Taiwan Emergency Medicine Society, and former CEO of Joint Commission of Taiwan. He has researched the epidemiology and basic and clinical studies of hemorrhagic shock, traumatic brain injury, and carbon monoxide poisoning. His research interests include real-time artificial intelligence to predict adverse outcomes in disease and the underlying mechanisms of brain damage.