Intravenous Human Umbilical Cord-Derived Mesenchymal Stromal Cells Promote Functional Recovery after Experimental Intracerebral Hemorrhage Via Local and Systemic Immunomodulation

Abstract

Intracerebral hemorrhage (ICH) is associated with high mortality and disability, and current treatments offer limited benefits for functional recovery. Human umbilical cord–derived mesenchymal stromal cell (hUCMSC) have strong proliferative, neuroprotective, and immunomodulatory properties, making them attractive for clinical translation. We evaluated the therapeutic effects of intravenously administered hUCMSCs in a collagenase-induced ICH model using male mice. Mice received low or high doses of hUCMSC once or twice within 72 h after ICH. Repeated high-dose administration significantly improved motor, cognitive, and affective behaviors. Although repeated administration of high-dose hUCMSCs produced the most pronounced behavioral recovery, most subsequent analyses were performed using the single-dose groups. Histological analysis showed reduced neuronal apoptosis and microglial activation, consistent with neuroprotection. In vitro assays demonstrated suppression of inflammatory gene expression and promotion of an anti-inflammatory phenotype in immune cells. Flow cytometry revealed selective reduction of pro-inflammatory macrophages and microglia, increased reparative subsets, and systemic modulation of myeloid dynamics. Our results suggest that intravenous hUCMSC administration at a higher dose confers robust neuroprotection through coordinated local and systemic immunomodulation, providing translational insights for clinical MSC therapy in ICH.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12975-025-01408-3.

Introduction

Intracerebral hemorrhage (ICH) represents 10–20% of all strokes and is one of the most severe types. It has a one-month mortality rate of approximately 40% and a disability rate of about 60% three months after onset [1, 2]. Although surgical interventions and acute medical treatments can be life-saving, their impact on improving long-term neurobehavioral outcomes is limited [3–5]. Rehabilitation has been identified as a key strategy for functional recovery in ICH patients; however, its therapeutic efficacy remains modest [6].

In recent years, stem cell–based therapies have been actively investigated as potential regenerative approaches for neurological disorders [7–9]. Stem cells possess the ability to proliferate and self-renew under specific conditions [10]. They are broadly categorized into two groups based on their developmental stage: embryonic stem cells, derived from the inner cell mass of the blastocyst, and somatic stem cells, which originate from various adult tissues. Among somatic stem cells, mesenchymal stromal cells (MSCs) have attracted considerable attention for their unique biological properties, including secretion of growth factors and cytokines, immunomodulation, neurogenesis, and angiogenesis [11]. MSCs can be isolated from diverse tissues such as bone marrow, adipose tissue, and perinatal tissues, making them attractive candidates for clinical applications [12, 13].

Human bone marrow–derived MSC (hBMSC) are the most extensively studied MSC type and have been approved in many countries for treating graft-versus-host disease (GVHD) [14, 15]. Clinical trials are also underway for various conditions, including stroke [16]. Several animal studies have demonstrated that intravenous administration of MSC can improve neurobehavioral outcomes following ICH [17, 18]. However, the procurement of hBMSC requires invasive bone marrow aspiration, which carries a complication risk of approximately 0.05%, including bleeding and infection [19].

To overcome this limitation, our group has developed human adipose-derived stem cells (hADSC), which can be obtained less invasively and have comparable or superior cytokine secretion profiles compared to hBMSC [20]. Subsequently, we established human amnion-derived amniotic-derived MSC (hAMSC) from fetal membranes obtained during cesarean section, a procedure that avoids additional invasive harvesting [21]. hAMSC secrete a wide array of cytokines at levels equal to or greater than those of hBMSC [22] and have been reported to exceed hADSC in their cytokine secretion profile [23]. Our previous studies showed that hAMSC produced greater functional improvement in ICH-bearing mice compared to hADSC [24]. Preclinical in vivo studies have reported no significant adverse events following hAMSC administration [25], and no significant safety concerns have emerged in human trials [26]. Indeed, clinical trials of hAMSC are ongoing for GVHD and Crohn’s disease [25, 27].

Human umbilical cord–derived MSC (hUCMSC) shares many advantages with hAMSC, as they can be isolated from perinatal tissues without additional invasive procedures [28, 29]. Notably, hUCMSC exhibit enhanced proliferative capacity compared to bone marrow–derived MSCs, which may be advantageous for large-scale clinical applications [28, 30].

Based on these findings, we hypothesized that intravenous administration of hUCMSC would enhance recovery from neurobehavioral deficits induced by modulating both local and systemic inflammatory responses. To assess this hypothesis, we induced ICH in mice, administered hUCMSC, and evaluated functional recovery utilizing a comprehensive battery of neurobehavioral assessments. Furthermore, to investigate the underlying mechanisms, we performed quantitative PCR (qPCR), immunostaining, TUNEL assays, and flow cytometry.

Materials and Methods

Cell Preparation of hUCMSC

The procedures for hUCMSC isolation and expansion were approved by the Institutional Ethics Committee of Hyogo Medical University (approval number: 0325). They were performed in accordance with a previous report [31]. Written informed consent was obtained from all donors. Briefly, umbilical cords were obtained from full-term cesarean sections. The cords were mechanically separated from the chorion and amnion and digested with a collagenase/dispase solution at 37 °C for 1 h. The resulting cell suspensions were filtered through a 100-µm mesh, resuspended in R: STEM medium (ROHTO Pharmaceutical Co., Ltd., Japan) supplemented with proprietary growth factors, and cultured in uncoated plastic dishes under standard conditions (37 °C, 5% CO₂, humidified atmosphere). No coating materials were used for cell adhesion. hUCMSCs were cryopreserved and thawed for use at passage 7. For cryopreservation, hUCMSCs were suspended in freezing medium and stored in a − 80 °C deep freezer until use. For experiments, frozen vials were rapidly thawed in a 37 °C water bath, gently mixed to ensure complete thawing, and immediately transferred into prewarmed culture medium for recovery before administration. All hUCMSCs used in this study were derived from a single donor to ensure consistency across experiments.

Mice and ICH Induction

The Institutional Animal Care Ethical Committee approved all the animal experiments of this study (Approval numbers: 21–012 A and 24–047 A). The procedure has been described previously [32, 33]. Male C57BL/6J mice (7–9 weeks old, 20–25 g) were anesthetized with 1.5–2.0% isoflurane and secured in a stereotaxic apparatus. A 30-gauge needle attached to a Hamilton syringe was inserted into the right striatum (0.2 mm anterior and 2.0 mm lateral to the bregma, 3.5 mm ventral from the dura). Collagenase Type A (AOF grade; Worthington Biochemical Corp., Lakewood, NJ, USA; Cat. No. CLSAFA) was dissolved in 0.5 µL of sterile saline, and 0.4 U was injected at a rate of 0.1 µL/min over 5 min. The needle was left in place for 10 min after injection to prevent backflow, then slowly withdrawn. After surgery, mice were allowed to recover on a heating pad and returned to their home cages with free access to food and water. All mice were monitored at least twice daily for signs of pain, dehydration, or neurologic deterioration, and additional supportive care (fluid supplementation or soft food) was provided when necessary. No animals met predefined exclusion criteria, which included death during surgery.

Mice were randomly assigned to treatment groups using a random-number generator. All behavioral, histological, and flow cytometry analyses were performed by investigators blinded to group allocation.

Administration of hUCMSC to ICH-bearing Mice

We conducted one experiment to optimize the dosage and times of the hUCMSC administrations. To optimize the dosage and times of hUCMSCs, mice were randomly divided into six groups as follows: the ICH-high*1 hUCMSC group (n = 17), the ICH-high*2 hUCMSC group (n = 14), the ICH-low*1 hUCMSC group (n = 11), the ICH-low*2 hUCMSC group (n = 11), the ICH-Vehicle group (n = 17), and the sham control group (n = 16). Here, the ICH-high*1 hUCMSC group and ICH-low*1 hUCMSC group received intravenous injections of hUCMSC at doses of 2.5 × 10⁵ cells/50 µl and 0.5 × 10⁵ cells/50 µl, respectively, at 24 h after the ICH induction. The ICH-high*2 hUCMSC group and ICH-low*2 hUCMSC group received intravenous injections of hUCMSC at doses of 2.5 × 10⁵ cells/50 µl and 0.5 × 10⁵ cells/50 µl, respectively, at 24 and 72 h after the ICH induction. The ICH-Vehicle group received an intravenous injection of 50 µl of vehicle control. The sham group underwent a scalp incision alone.

The group sizes were based on our previous studies demonstrating statistically detectable differences in neurobehavioral outcomes using similar sample sizes [32, 33]. No interim analyses were conducted.

Neurobehavioral Tests

The procedure has been described previously [33, 34]. Briefly, the following neurobehavioral tests were started at Day 29 to assess the phenotypic differences in the mice. All behavioral assessments were performed by experimenters blinded to group identity. Researchers who were blinded to the experiments evaluated the results of the tests. Details of each test are described below. A detailed schedule of behavioral testing was followed from Day 29 to Day 51 after ICH. The sequence of tests was as follows: Open Field Test (Days 29–31), Forced Swim Test (Day 38), Morris Water Maze acquisition and probe testing (Days 42–46), and Passive Avoidance Test (Days 48–50).

Open Field Test

A cubic open-field box made of transparent acrylic panels, without a ceiling (30 × 30 × 30 cm), was placed inside a ventilated soundproof chamber (Taiyo Electric). An overhead incandescent bulb provided room lighting intensity (approximately 110 lx) at the center of the test arena. A fan mounted on the upper part of one wall produced masking noise of 45 dB inside the chamber. On each lateral side of the open-field box, two infrared beams were set 2 cm above the floor at 10 cm apart. A flip-flop circuit was installed between the two beams. The total number of circuit breaks was recorded as a locomotive behavior. Each mouse was allowed to explore the open-field arena freely for 10 min per day for three consecutive days.

Open Space Swimming Test

The open space swimming test was conducted to evaluate their neuromuscular strength and depression-like symptoms in the mice. A circular pool (inside diameter: 95 cm, depth: 35 cm) surrounded by a white wall (width: 130 cm, height: 120 cm) was filled with water to a depth of 20 cm. The water was made opaque with titanium oxide to enable a video-tracking system (Be Chase ver.1.3, ISONIX) to trace the behavior of mice. The water temperature was maintained at 22 ± 1 °C. In each test, a mouse was placed in the pool with its head facing the outer edge of the pool and allowed to swim freely for 10 min. All tests were recorded with an overhead CCD camera. The total swimming length of the subject mouse was calculated using the video-tracking system.

Passive Avoidance Learning Test

The passive avoidance learning test was conducted to evaluate the long-term memory function of the mice. An apparatus consisted of light and dark compartments with the same dimensions (15 × 15 × 15 cm) with a grid floor. A guillotine door separated the two compartments. In the conditioning test, a mouse was placed in the light compartment. Ten seconds later, the guillotine door was opened. When the mouse moved into the dark compartment, the guillotine door was closed. Ten, five seconds later, a scrambled electrical shock (at 120 V for 5 s) was delivered through the grid floor.

24 and 48 hours later, a retention test was conducted without the electrical shock. The subject mouse was placed in the light compartment, and the duration to enter the dark compartment was recorded up to 180 s. This duration (passive avoidance latency) reflected the memory functions of the subjected mice to electrical shocks at 24-hour intervals.

Forced Swim Test

The forced swim test was carried out using a procedure modified from the original report [35]. An infrared sensor was mounted in the center of the ceiling of a cubic chamber (45 cm on each side) with soundproofing and ventilation functions, and can detect mouse activity one second in advance. Directly below this chamber, mice were individually placed into a glass cylinder (25 cm height, 10 cm diameter) containing 15 cm of water maintained at 23–25 °C. Every minute for a total of 6 min, the mice were rated for activity counts, defined as the presence of active, escape-oriented behaviors such as swimming, jumping, rearing, sniffing, or diving. At each time point, the mice were observed for 1 min, and activity counts were recorded.

Morris Water Maze Learning Test

The Morris water maze learning test was conducted to evaluate the spatial recognition of the mice. A circular pool (inside diameter: 95 cm, depth: 35 cm) surrounded by a white wall (width: 150 cm, height: 120 cm) was filled with water to a depth of 22 cm. The water was made opaque with titanium oxide to enable the video-tracking system to trace the behavior of mice. The temperature of the water was maintained at 24 ± 1 °C. The pool was divided into four virtual quadrants: north, south, east, and west. A round white platform (diameter: 10 cm) was placed at the center of the north quadrant of the pool and submerged 0.5 cm below the surface of the water. In each test, a mouse was randomly released into the water with its head facing the edge of the pool in either the south, east, or west quadrants. The test ended when the subject mouse reached the platform and stayed on it for 10 s. If the mouse could not find the platform within 60 s, it was guided to the platform by the experimenter and kept there for 10 s. Each mouse was tested 5 times per day with 30-second intervals for five consecutive days. All tests were recorded with an overhead CCD camera. The duration between the release of the subjected mouse into the water and its arrival on the platform was measured as the escape latency using the video-tracking system.

Immunostaining

The procedure has been described previously [33]. Mice were divided into two groups: the hUCMSC group, which received an intravenous injection of 2.5 × 10⁵ hUCMSCs suspended in 50 µL of vehicle at 24 h after ICH induction, and the ICH group, which received an intravenous injection of 50 µL of vehicle alone. Tissue collection was performed on Day 3 after ICH, as previous studies have shown that perihematomal inflammation and neuronal apoptosis peak within the first 72 h after hemorrhage [36]. Under deep anesthesia with 3–4% isoflurane, mice were perfused transcardially with normal saline followed by 4% paraformaldehyde in PBS. The brains were post-fixed in 4% paraformaldehyde for 1 day and then cryoprotected in 30% sucrose at 4 °C. Brain tissues were sectioned coronally at 10 μm using a cryostat (CM1950, Leica Biosystems, Wetzlar, Germany).

Brain sections were subjected to immunofluorescence staining using the following primary antibodies: anti-Iba1 (rabbit polyclonal; Abcam, Cambridge, UK; Cat. No. ab178847; 1:250), anti-GDNF (mouse monoclonal; Santa Cruz Biotechnology, Dallas, TX, USA; Cat. No. sc-13147; 1:100), anti-TNF-α (rabbit monoclonal; Abcam, Cambridge, UK; Cat. No. ab6671; 1:100) and anti-CD11b (rat monoclonal; BioLegend, San Diego, CA, USA; Cat. No. 101201; 1:100). The appropriate secondary antibodies were applied, including Alexa Fluor 488-conjugated anti-rabbit IgG (Cell Signaling Technology, Danvers, MA, USA; Cat. No. 4412 S; 1:500), Alexa Fluor 555-conjugated anti-mouse IgG (Cell Signaling Technology; Cat. No. 4409 S; 1:500), and Alexa Fluor 647-conjugated anti-rat IgG (Cell Signaling Technology; Cat. No. 4418 S; 1:500). Nuclei were counterstained with DAPI (Sigma-Aldrich, St. Louis, MO, USA; Cat. No. D9542). Images were acquired using a fluorescence microscope (BZ-X710, Keyence, Osaka, Japan) under identical exposure settings for all groups, allowing for quantitative comparison.

To evaluate apoptosis, TUNEL staining was performed on frozen brain sections using the In Situ Cell Death Detection Kit, Fluorescein (Sigma-Aldrich; Cat. No. 11684795910) according to the manufacturer’s instructions. Briefly, sections were incubated for 1 h at room temperature with the TUNEL reaction mixture, which contained terminal deoxynucleotidyl transferase (TdT) and fluorescein-labeled dUTP. Sections were then washed in PBS and counterstained with DAPI. All stained sections were visualized using a fluorescence microscope (BZ-X710, Keyence). Alexa Fluor 488 signals from TUNEL and Iba1 were detected in the green channel, while DAPI was detected in the blue channel.

For quantification, two coronal sections per mouse were analyzed, and three perihematomal fields were imaged in each section, yielding six fields per mouse. Sections were collected at approximately + 0.5 to − 0.5 mm relative to bregma. The size of each region of interest corresponded to the field of view at the acquisition magnification of the fluorescence microscope. Sample sizes were as follows: TUNEL (hUCMSC, n = 8; ICH control, n = 5), Iba1 (hUCMSC, n = 5; control, n = 5), and TNFα, CD11b, and GDNF (hUCMSC, n = 5; control, n = 5). A schematic of the perihematomal region analyzed is provided in Supplementary Fig. 1.

RNA Extraction and Quantitative Real-Time PCR

For in vivo analysis, perihematomal brain tissue was collected at Days 3, 4, and 8 post-ICH. Brains were removed under deep anesthesia, rinsed in PBS, and approximately 2-mm-thick coronal blocks encompassing the hematoma and surrounding tissue were dissected, snap-frozen in liquid nitrogen, and stored at −80 °C.

For in vitro analysis, immune cells (peripheral blood leukocytes, splenocytes, and bone marrow cells) were stimulated with LPS (100 µg/mL) and cocultured with hUCMSCs in Transwell inserts (0.4 μm pore size), as described in the following section. After 24 h, cells were harvested and washed in PBS. Peripheral blood cells were prepared after RBC lysis and seeded at 4 × 10⁵/insert, splenocytes at 5 × 10⁶/insert, and bone marrow cells at 4 × 10⁵/insert. hUCMSCs were cultured in Rstem medium until confluent before coculture.

RNA from both brain and coculture samples was extracted using the RNeasy Mini Kit (Qiagen) with on-column DNase treatment. RNA concentration and purity were measured (SpectraMax^® ABS Plus), and 1 µg of RNA was reverse-transcribed into cDNA using ReverTra Ace^® qPCR RT Master Mix (TOYOBO). Quantitative PCR was performed on a QuantStudio 12 K Flex (Applied Biosystems) with THUNDERBIRD^® SYBR^® qPCR Mix (TOYOBO). Cycling parameters were: 95 °C for 1 min, then 40 cycles of 95 °C for 15 s, 60 °C for 15 s, and 72 °C for 35 s. Melt-curve analysis confirmed specificity.

Relative expression was determined using the 2^−ΔΔCt method with β-actin as the reference gene. Target genes included TNFα, TSG6, GDNF, and VEGFα (brain tissue) and CCL2, F4/80, Ly6c, IL6, TNFα, CD163, and CD204 (coculture). Primer sequences are provided in Supplementary Table S1.

In Vitro Coculture Assay with LPS-Stimulated Immune Cells

To assess the immunomodulatory effects of hUCMSCs under inflammatory conditions, we established an in vitro coculture system using a Transwell setup. Immune cells were derived from peripheral blood, spleen, and bone marrow of 7–9-week-old C57BL/6J mice and stimulated with lipopolysaccharide (LPS, 100 µg/mL). hUCMSCs were seeded in the lower chamber until confluence, and LPS-stimulated immune cells were placed in the upper insert for 24 h, allowing only paracrine interactions.

Peripheral blood leukocytes were collected using heparinized syringes. After red blood cell lysis (RBC Lysis Buffer, BioLegend), leukocytes were resuspended in RPMI-1640 supplemented with 10% fetal bovine serum (FBS) and seeded into Transwell inserts at 4 × 10⁵ cells/insert.

Splenocytes were prepared by mincing spleens, filtering through a 70-µm strainer, and lysing RBCs. Cells were suspended in RPMI-1640 + 10% FBS and seeded at 5 × 10⁶ cells/insert.

Bone marrow cells were isolated by flushing femurs with PBS, followed by RBC lysis. Nucleated cells were seeded into inserts at 4 × 10⁵ cells/insert in RPMI-1640 + 10% FBS.

hUCMSC culture was performed in Rstem medium (ROHTO Pharmaceutical) in the lower wells of 24-well Transwell plates (0.4 μm pore size). Once confluent, the LPS-stimulated immune cells were added to the inserts and cocultured for 24 h at 37 °C in 5% CO₂.

Flow Cytometry

All flow cytometry analyses were conducted by operators blinded to treatment groups. The same gating strategy was applied uniformly across all samples, as shown in Fig. 5B, 6 A. The procedure has been described previously [33]. Briefly, mice were deeply anesthetized with 3–4% isoflurane and perfused with ice-cold PBS. Their brains and spleens were removed; femurs and tibias were dissected to harvest bone marrow. Peripheral blood was collected via cardiac puncture into heparinized syringes or tubes (MOCHIDA PHARMACEUTICAL CO., LTD, Tokyo, Japan) and kept on ice until further processing. Spleens were mechanically dissociated by gently pressing through a 70-µm cell strainer using the plunger of a syringe. Cells were centrifuged at 400 × g for 5 min at 4 °C. Pellets were resuspended in 1–2 mL of red blood cell (RBC) lysis buffer (eBioscience or equivalent), incubated for 2–3 min at room temperature with gentle agitation, quenched with PBS, and washed once with PBS. Leukocytes were resuspended in FACS buffer.

Fig. 5.

Flow cytometric analysis of brain-infiltrating myeloid subsets after hUCMSC administration in ICH mice. (A) Schematic timeline of the experimental protocol. ICH was induced on Day 0, followed by intravenous injection of hUCMSC or vehicle on Day 1. Mice were sacrificed on Days 2, 3, and 7 for flow cytometric analysis of perihematomal brain tissue. (B) Representative gating strategy of the hUCMSC group at Day 3. Live cells were identified by FSC/SSC and Zombie viability dye. Infiltrating myeloid populations were distinguished as CD45^hiCD11b^hi (green gate) and CD45^intCD11b^int (red gate). Within each population, expression of CD86, CD163 and CD204 markers was assessed. (C) Quantification of macrophage subsets. In the CD45^hiCD11b^hi fraction, hUCMSC treatment significantly reduced CD86⁺ cells and increased CD163⁺ cells on Day 7, with no changes in CD204⁺ cells. In the CD45^intCD11b^int fraction, hUCMSC administration decreased CD86⁺ cells and elevated CD163⁺ cells at later time points, while CD204⁺ cells remained unchanged. Data are presented as mean ± SEM (n = 5-7 per group). *P < 0.05, **P < 0.01 vs. ICH-vehicle

Fig. 6.

Flow cytometric analysis of myeloid cell subsets in the spleen, blood and bone marrow after hUCMSC administration. (A) Gating strategy for identifying myeloid cell subsets. Live single cells were selected based on forward and side scatter (FSC-A vs. SSC-A), Zombie-negative staining, and FSC-A vs. FSC-H gating. Lineage-negative cells (CD45R/B220⁻, CD90.2⁻, Ly6G, NK1.1, CD49b) were gated for CD11b⁺ populations. Myeloid subsets were then further classified based on Ly6C and Ly6G expression (B) Quantification of myeloid cell subsets in the spleen (n = 9 (Day 2 in both groups), n = 7, 5 (Day 3, in control and hUCMSC groups), n = 7 (Day 7 in both groups)). (C) Quantification in peripheral blood (n = 9 (Day 2 in both groups), n = 7, 6 (Day 3, in control and hUCMSC groups), n = 7 (Day 7 in both groups)). (D) Quantification in bone marrow (n = 6, 5 (Day 2, in control and hUCMSC groups), n = 7, 5 (Day 3, in control and hUCMSC groups), n = 7, 4 (Day 7, in control and hUCMSC groups)). Data are shown as mean ± SEM. *P < 0.05, **P < 0.01, vs. ICH-vehicle

Bone marrow was flushed from femurs and tibias with cold RPMI-1640 using a 25-gauge needle, filtered through a 70-µm strainer, and centrifuged at 400 × g for 5 min at 4 °C. Pellets were treated with 1–2 mL RBC lysing buffer (BioLegend) for 2–3 min at room temperature with gentle agitation, quenched with PBS, washed once with PBS, and resuspended in FACS buffer.

Whole blood was diluted 1:1 with cold PBS and incubated with RBC lysing buffer (BioLegend) for 10 min at room temperature with gentle mixing. Lysis was quenched by excess cold PBS, followed by centrifugation at 400 × g for 5 min at 4 °C. Pellets were washed once with PBS and resuspended in FACS buffer.

Brains were minced and gently sheared by passing through 18- and 20-gauge needles. Tissue was washed, resuspended in 25% Percoll (GE Healthcare), and centrifuged for 20 min at 500 × g. Supernatants were discarded and pellets were resuspended in Histopaque (Sigma). After a second centrifugation (20 min, 500 × g), enriched immune cells were collected from the Histopaque interface, washed once in PBS, and resuspended in FACS buffer (PBS + 2% FBS).

Cells were incubated with fluorochrome-conjugated antibodies for 20–30 min at 4 °C in the dark, washed with FACS buffer, and filtered through 40–70 μm strainers before acquisition. The list of antibodies and viability dye, including fluorochromes, suppliers, and final working concentrations, is provided in Supplementary Table S2. A viability dye was used according to the manufacturer’s instructions. Fluorescence was acquired on an LSRFortessa X-20 (BD Biosciences) using BD FACSDiva software. Data were analyzed with FlowJo v10.5 (BD Biosciences).

All centrifugations were performed with the brake off unless otherwise indicated. After lysis steps, samples were immediately diluted with excess cold PBS to terminate RBC lysis.

Statistical Analysis

All results are expressed as mean ± SEM. For the statistical analysis of the neurobehavioral tests, repeated-measures ANOVA was performed with group as the between-subject factor and repeated measures (e.g., session, trial, or time) as the within-subject factor. When ANOVA revealed significant effects, post hoc comparisons were conducted using the Tukey–Kramer method. For the statistical analyses of flow cytometry data and immunostaining data, the Wilcoxon test was applied for each day. Statistical significance was set at P < 0.05 (two-tailed). All statistical analyses were performed using JMP version 17 (SAS Institute, Cary, NC, USA). Blinding, randomization, and exclusion criteria were incorporated into all analyses as described above in accordance with ARRIVE 2.0 guidelines.

Results

Intravenous Administration of hUCMSC Dose-Dependently Improves ICH-Induced Neurobehavioral Deficits during the Subacute Phase

To evaluate the therapeutic efficacy and optimal dosing and frequency of administration of hUCMSC, we induced ICH in mice. Two different doses were administered: 0.5 × 10⁵ or 2.5 × 10⁵ cells per mouse, either once at 24 h post-ICH or twice (at 24 and 72 h post-ICH). Neurobehavioral assessments—including open field, passive avoidance learning, forced swim, and Morris water maze tests—were conducted from Day 29 onward (Fig. 1A). In the open field test, there was a significant group effect on locomotion counts (Fig. 1B; F_{5, 270}= 21.9002, P < 0.0001). The ICH-high*2 group displayed significantly lower locomotion counts compared to the ICH-vehicle group at Day 2 (P = 0.022) and Day 3 (P = 0.0059), while the ICH-high1 group showed a similar trend at Day 3 (P = 0.088).

Fig. 1.

Figure 1 Behavioral assessment of functional recovery following intravenous administration of hUCMSCs after ICH. (n=14 in ICH-high*2 group, n=17 in ICH-high*1 group, n=11 in ICH-low*2 group, n=17 in ICH-low*1 group, n=17 in ICH-vehicle group, n=16 in Sham group). (A) Experimental timeline and group assignments. ICH was induced on Day 0, followed by intravenous administration of hUCMSC at high dose (2.5 × 10⁵ cells/mouse) or low dose (0.5 × 10⁵ cells/mouse) either once (at 24 h) or twice (at 24 h and 72 h post-ICH). Neurobehavioral tests were conducted beginning on Day 29. ICH-vehicle and sham groups served as controls. (B) Open field test. Locomotion counts over 10 minutes were significantly reduced in the ICH-high*2 hUCMSC group compared to the vehicle group on Days 2 and 3, indicating reduced hyperactivity or anxiety-like behavior. (C) Passive avoidance learning test. Step-through latency was significantly prolonged in all hUCMSC-treated groups compared to the vehicle group on Days 2 and 3, suggesting improved learning and memory performance. (D) Forced swim test over 6 minutes. hUCMSC-treated mice, particularly in the ICH-high*2 group, exhibited higher activity counts than the vehicle group, indicating reduced behavioral despair. (E) Quantification of activity counts during the initial one minute of the forced swim test. The ICH-high*2, ICH-high*1, and ICH-low*2 groups showed significantly higher activity compared to the vehicle group. (F) Morris water maze learning test. The ICH-high*2, ICH-high*1, and ICH-low*2 groups demonstrated significantly reduced escape latency over repeated trials compared to vehicle-treated ICH mice, indicating enhanced spatial learning and memory. Data are presented as mean ± SEM. * P < 0.05, ** P < 0.01 vs. ICH-vehicle group

In the passive avoidance learning test, group effects were also significant (Fig. 1C; F_{5, 270} = 9.7170, P < 0.0001). The ICH-high*2 group exhibited significantly longer step-through latency than the ICH-vehicle group at 24 h and 48 h (P < 0.0001 for both). The ICH-high*1 group showed longer latency at 48 h (P < 0.0001), and the ICH-low*2 group showed longer latency at 24 h (P < 0.0001).

In the Morris water maze, escape latency significantly differed among groups (Fig. 1D; F_{5, 86} = 8.6717, P < 0.0001). The ICH-high*2 group showed shorter latency than ICH-vehicle at Day 2 (P = 0.0398), Day 4 (P = 0.0489), and Day 5 (P = 0.0030); the ICH-high*1 group at Day 4 (P = 0.0135) and Day 5 (P = 0.0092); and the ICH-low*2 group at Day 4 (P = 0.0490) and Day 5 (P = 0.0252).

In the forced swim test, significant group effects were observed for overall activity counts (Fig. 1E; F_{30, 425} = 3.4666, P < 0.0001). During the first minute (Fig. 1 F), the ICH-high2, ICH-high1, and ICH-low*2 groups had significantly higher activity counts than the ICH-vehicle group (P < 0.0001).

Taken together, these results indicate that hUCMSC treatment, especially with repeated high-dose administration, significantly improves locomotor activity, learning, and memory, which ICH damaged.

Intravenous Administration of hUCMSC Reduces Apoptotic Cell Death in the Perihematomal Region

ICH is known to induce neuroinflammation via activation of macrophages and microglia, subsequently triggering apoptotic pathways in the brain [37, 38]. As shown in the experimental timeline (Fig. 2A), hUCMSCs were intravenously administered 24 h after ICH induction, and brains were harvested on Day 3 for histological analysis. To examine apoptotic cell death, TUNEL staining was performed on perihematomal sections. In untreated mice, numerous TUNEL-positive cells were detected (Fig. 2B, top row), whereas hUCMSC-treated mice exhibited markedly fewer TUNEL-positive cells (Fig. 2B, bottom row). Quantitative analysis confirmed that hUCMSC administration significantly reduced TUNEL-positive cell density compared with untreated mice (P < 0.01; Fig. 2B, right panel).

Fig. 2.

hUCMSC administration reduces apoptosis, suppresses microglial activation, and modulates inflammatory and neurotrophic markers in the perihematomal brain after ICH. (A) Experimental timeline. Intracerebral hemorrhage (ICH) was induced on Day 0, followed by intravenous injection of hUCMSC or vehicle at 24 h. Brains were collected on Day 3 for histological analyses. (B) Representative TUNEL staining (green) with DAPI nuclear counterstain (blue) and quantification of TUNEL⁺ cells, showing a significant reduction of apoptotic cells in the hUCMSC group compared with ICH-vehicle. (C) Representative Iba1 staining (green) with DAPI (blue) and quantification of Iba1⁺ cells, demonstrating decreased microglial activation in the hUCMSC group. (D) Representative immunofluorescence images for TNFα (green), CD11b (cyan), and GDNF (red) with DAPI (blue), and quantification of positive cells. hUCMSC treatment significantly reduced TNFα⁺ cells and increased GDNF⁺ cells, while CD11b⁺ cell numbers showed no significant change. Scale bar = 50 µm. Data are presented as mean ± SEM. * P < 0.05, ** P < 0.01 versus ICH-vehicle

hUCMSC Treatment Attenuates Microglial Activation

Under the same timeline (Fig. 2A), immunofluorescence staining for Iba1 was conducted to evaluate microglial activation. Vehicle-treated mice exhibited numerous Iba1-positive cells in the perihematomal region, while hUCMSC-treated mice showed a marked reduction (Fig. 2C). Quantitative analysis demonstrated a significant decrease in Iba1-positive cell density after hUCMSC administration (P < 0.05; Fig. 2C, right panel).

hUCMSC Attenuates Local Neuroinflammation in the Brain

To examine the effects of hUCMSC administration on neuroinflammatory responses after ICH, immunofluorescence staining for TNFα, CD11b, and GDNF was performed following the same schedule (Fig. 2A). In the ICH-vehicle group, abundant TNFα-positive cells were observed, whereas hUCMSC-treated mice exhibited markedly fewer TNFα-positive cells (Fig. 2D). In contrast, GDNF-positive cells were visibly increased in the hUCMSC group compared with the vehicle group (Fig. 2D). CD11b immunoreactivity showed a tendency toward reduction following hUCMSC treatment, although the difference was not statistically significant. Quantitative analysis confirmed a significant decrease in TNFα-positive cells (P < 0.01) and a significant increase in GDNF-positive cells (P < 0.01) in the hUCMSC group. These findings suggest that hUCMSC administration suppresses the expression of acute-phase pro-inflammatory cytokines while enhancing neurotrophic factor production in the perihematomal area following ICH.

Intravenous Administration of hUCMSC Modulates Inflammatory and Neuroprotective Gene Expression in the Perihematomal Brain

To evaluate the effects of hUCMSC treatment on inflammatory and neuroprotective mediators after ICH, we performed quantitative PCR analysis of perihematomal brain tissue at Days 3, 4, and 8 following intravenous administration of hUCMSC or vehicle (Fig. 3A). Expression of TNFα, mTSG6, GDNF, and VEGFα was quantified to capture both proinflammatory and reparative responses. In the present study, we found that intravenous administration of hUCMSC significantly modulated this pathological milieu in the perihematomal brain.

Fig. 3.

Intravenous administration of hUCMSC modulates inflammatory and neuroprotective gene expression in the perihematomal brain. (A) Timeline of the experimental procedures. ICH was induced at Day 0. hUCMSC or vehicle was intravenously injected at 24 hours after ICH induction. Mice were sacrificed at Days 3, 4, and 8 for qPCR analyses. (B-E) Quantitative real-time PCR analyses of gene expression in ICH brain tissues. Relative mRNA expression of (B) TNFα, (C) mouse TSG6 (mTSG6), (D) GDNF, and (E) VEGFα in Control and hUCMSC-treated groups at indicated time points. Data are shown as mean ± SEM (n = 10 in each group). * P < 0.05, ** P < 0.01, vs. ICH-vehicle

The proinflammatory cytokine TNFα was significantly decreased in the hUCMSC-treated group compared to the untreated ICH group at Day 8 (P < 0.05) (Fig. 3B), indicating a delayed but sustained anti-inflammatory effect. In contrast, mTSG6, a potent immunomodulatory factor known to interact with CD44⁺ on activated microglia and astrocytes, was significantly upregulated at Day 4 (P < 0.01) and Day 8 (P < 0.05) in the hUCMSC group (Fig. 3C). This suggests that hUCMSC-derived TSG6 may suppress NF-κB–driven transcription of proinflammatory genes, thereby attenuating microglial activation and cytokine production.

GDNF expression was significantly upregulated in hUCMSC-treated brains at Days 4 and 8 (P < 0.05) (Fig. 3D), while levels at Day 3 showed no significant difference from controls. This pattern suggests a delayed but sustained increase in neurotrophic support that may facilitate neuronal survival and repair. VEGFα, a critical angiogenic mediator, was also significantly upregulated at Days 3 (P < 0.05) and 4 (P < 0.01) (Fig. 3E), potentially facilitating vascular remodeling and improving tissue perfusion during the subacute phase.

These temporal expression patterns are consistent with previous reports that MSC-derived TSG6 reduces neuroinflammation [39]. Still, our findings extend these observations to the perihematomal region. Such a molecular profile—characterized by sustained suppression of proinflammatory signals alongside activation of reparative pathways—may contribute to improved functional recovery, as observed in our neurobehavioral assessments. Collectively, these results support the concept that hUCMSC create a favorable post-ICH microenvironment through both cellular and molecular mechanisms, with potential relevance for human ICH treatment.

hUCMSC Attenuate Pro-Inflammatory Gene Expression and Partially Enhance M2 Markers in Immune Cells from Spleen, Peripheral Blood, Bone Marrow, and Brain in an In Vitro Coculture Model

Given that intravenous hUCMSC administration modulated immune responses in vivo, we next investigated whether hUCMSC could directly influence the inflammatory phenotype of immune cells in vitro. Splenocytes, peripheral blood leukocytes, bone marrow cells, and brain-derived cells were stimulated with LPS and cocultured with hUCMSC for 24 h using a Transwell system (Fig. 4A). Gene expression analysis by qPCR revealed that, in splenocytes, F4/80 and Ly6c were significantly downregulated in the LPS + hUCMSC group (P < 0.01) compared to LPS alone, whereas CCL2 expression was markedly upregulated (P < 0.01) (Fig. 4B). In peripheral blood leukocytes, F4/80 and IL6 were reduced (P < 0.05), while CCL2 was significantly increased (P < 0.01) (Fig. 4C). In bone marrow–derived cells, Ly6c expression was significantly decreased (P < 0.05), with TNFα and IL6 showing a downward trend in the hUCMSC group (Fig. 4D). In brain-derived cells, the expression of the pro-inflammatory marker CD86 tended to decrease, whereas the anti-inflammatory/reparative marker CD163 was significantly upregulated (P < 0.05) and CD204 levels remained unchanged (Fig. 4E). This shift suggests a transition toward a more anti-inflammatory phenotype, although recent studies highlight that macrophage/microglial activation states exist along a continuum rather than a strict M1/M2 dichotomy [40, 41].

Fig. 4.

Immunomodulatory effects of hUCMSC on LPS-stimulated immune cells derived from different tissues. (A) Schematic of the co-culture experiment. hUCMSC were seeded in the lower chamber of a 24-well Transwell system (0.4 μm pore size), and immune cells derived from mouse spleen, peripheral blood, bone marrow, or brain were seeded in the upper chamber with or without 100 μg/mL LPS stimulation. After 24 hours of co-culture, total RNA was extracted from the immune cells, followed by reverse transcription and qPCR to quantify expression of immune-related genes. (B) In splenocytes, the expression of F4/80 and Ly6c was significantly suppressed in the LPS+hUCMSC group compared to LPS alone, while CCL2 was upregulated. (C) In peripheral blood leukocytes, hUCMSC co-culture increased CCL2 expression, while F4/80 and IL-6 were downregulated compared with LPS alone. (D) In bone marrow–derived cells, pro-inflammatory mediators TNFα and IL-6 showed a decreasing trend, and Ly6c expression was reduced, suggesting attenuation of inflammatory monocyte activation. (E) In brain-derived immune cells, CD86 expression showed a modest reduction following hUCMSC treatment, whereas CD163 expression was significantly increased. CD204 expression did not differ significantly between groups. These changes reflect alterations in immune-related gene expression induced by hUCMSCs, without implying classification into discrete M1/M2 polarization states, in line with current understanding that myeloid cell activation exists along a continuum rather than a binary spectrum. Data are presented as mean ± SEM (n = 6 per group). * P < 0.05, ** P < 0.01 versus the LPS group

These findings indicate that hUCMSC broadly suppress LPS-induced pro-inflammatory gene expression across multiple immune compartments while selectively enhancing M2-associated markers, supporting their potential to modulate systemic and central immune responses after LPS-induction.

hUCMSC Treatment Differentially Modulates Pro- and Anti-Inflammatory Polarization in Infiltrating Macrophages and resident Microglia after ICH

Flow cytometric profiling of perihematomal brain tissue (Fig. 5) revealed that hUCMSC therapy significantly modulated both infiltrating macrophages (CD45^hiCD11b^hi) and resident microglia (CD45^intCD11b^int) after ICH.

In the CD45^hiCD11b^hi macrophage population, CD86⁺ pro-inflammatory cells were markedly reduced in the hUCMSC-treated group compared with the ICH-vehicle group at Day 7 (P < 0.01), whereas no significant differences were observed at Days 2 or 3. CD163⁺ cells were increased in the hUCMSC group on Day 3, although without statistical significance, and returned to baseline levels by Day 7. CD204⁺ subsets showed a tendency to increase on Day 3 in the hUCMSC group, but no significant intergroup differences were detected at any time point.

In the CD45^intCD11b^int microglial population, hUCMSC treatment significantly reduced CD86⁺ cells at both Day 2 (P < 0.05) and Day 7 (P < 0.05), while no significant differences were observed at Day 3. Conversely, CD163⁺ reparative cells were significantly increased in the hUCMSC group at Day 7 (P < 0.05), with a trend toward elevation on Day 3. CD204⁺ microglia showed no significant changes between groups at any of the examined time points.

Taken together, these findings indicate that hUCMSC administration selectively suppressed CD86⁺ pro-inflammatory subsets and promoted CD163⁺ reparative subsets in both infiltrating macrophages and resident microglia, while CD204⁺ cells remained unaffected.

Systemic Myeloid Dynamics Reflect Transient Mobilization Followed by Suppression

To further characterize in vivo immune alterations induced by hUCMSC therapy, we next assessed systemic responses across the spleen, blood, and bone marrow by flow cytometric analysis (Fig. 6). In the spleen, hUCMSC-treated mice showed a transient increase in CD11b⁺Ly6G⁺ neutrophils at Day 2 (P < 0.05), followed by a significant decline at Day 3 (P < 0.05). CD11b⁺Ly6C^low monocytes were also reduced on Day 3 (P < 0.05), with no significant differences in the CD11b⁺Ly6C^hi subset. Peripheral blood mirrored these dynamics: both neutrophils and Ly6C^low monocytes peaked at Day 2 (neutrophils: P < 0.01; Ly6C^low: P < 0.05), whereas inflammatory monocytes declined by Day 3 (P < 0.05). In the bone marrow, hUCMSC therapy caused significant reductions in CD11b⁺Ly6G⁺ cells by Day 7 (P < 0.01) and suppressed both Ly6C^low and Ly6C^hi monocyte populations at Days 3 and 7 (P < 0.05 and P < 0.01, respectively). These findings demonstrate a rapid immunomodulatory response characterized by early mobilization, followed by suppression of inflammatory subsets—suggesting an early shift toward anti-inflammatory regulation after ICH.

Discussion

In this study, we demonstrated that intravenous administration of hUCMSCs confers multifaceted therapeutic effects in a murine model of collagenase-induced ICH. Behavioral assessments showed that high-dose repeated administration (Fig. 1) significantly improved motor and cognitive recovery, as evidenced by shortened escape latency in the Morris water maze, increased exploratory activity in the open field, and attenuated depression-like behavior in the forced swim test. These findings suggest that hUCMSC therapy can improve not only motor and cognitive deficits but also affective disturbances, which are increasingly recognized as important consequences of ICH. Histological analyses revealed that hUCMSC treatment reduced neuronal apoptosis and attenuated perihematomal microglial activation (Fig. 2), consistent with a direct neuroprotective effect. At the molecular level, hUCMSC suppressed the pro-inflammatory cytokine TNFα while inducing the anti-inflammatory mediator TSG6 and upregulating neurotrophic and angiogenic factors such as GDNF and VEGFα (Fig. 3), suggesting coordinated immunomodulatory and reparative actions. Complementary in vitro co-culture assays further demonstrated that hUCMSC downregulated inflammatory transcripts and promoted an anti-inflammatory macrophage phenotype characterized by increased CD206 expression (Fig. 4). Flow cytometric profiling of immune cell dynamics provided mechanistic insight into these effects. In the brain, hUCMSC administration significantly modulated both infiltrating macrophages (CD45^hiCD11b^hi) and resident microglia (CD45^intCD11b^int). In both populations, hUCMSCs suppressed the accumulation of CD86⁺ pro-inflammatory cells while promoting CD163⁺ reparative subsets, whereas CD204⁺ cells remained unaffected (Fig. 5). Systemically, hUCMSC transiently mobilized myeloid cells but subsequently suppressed neutrophils and Ly6C^low monocytes in the spleen, blood, and bone marrow (Fig. 6), highlighting their role in orchestrating both central and peripheral immune responses.

Taken together, these findings indicate that intravenous hUCMSC therapy promotes functional recovery after ICH through dual mechanisms of neuroprotection and immune modulation. hUCMSCs reduced neuronal apoptosis and microglial activation while enhancing trophic factor expression, thereby supporting tissue repair. In parallel, they selectively suppressed CD86⁺ pro-inflammatory macrophages and microglia and increased CD163⁺ reparative subsets, while systemically limiting neutrophil and Ly6C^low monocyte expansion. Such combined neuroprotective and immunomodulatory effects suggest that hUCMSCs may represent a promising adjunctive therapy to improve neurological and neuropsychiatric outcomes in ICH.

ICH remains one of the most devastating forms of stroke, accounting for 10–20% of all cases and associated with high mortality and disability rates despite advances in acute medical and surgical interventions [2, 42]. While prompt hematoma evacuation, hemostatic therapy, and rigorous neurocritical care can reduce early mortality, these measures have shown limited efficacy in ameliorating long-term neurological deficits [1, 2]. Functional recovery after ICH largely depends on the brain’s capacity to repair itself. Yet, the endogenous repair mechanisms are insufficient to reverse secondary injury cascades—including neuroinflammation, apoptosis, oxidative stress, and disruption of neurovascular integrity [43–45]. Neuroinflammation plays a central role in triggering these pathological cascades, including iron-mediated oxidative damage and immune cell infiltration, all of which contribute significantly to neurological deterioration after ICH [46–48]. The early activation of microglia, release of pro-inflammatory mediators, and peripheral leukocyte infiltration constitute key contributors to secondary brain damage [46, 48].

Over the past two decades, MSC have emerged as promising candidates for regenerative therapies in neurological diseases [9, 11, 49, 50]. MSC have a wide range of therapeutic effects: they secrete trophic factors that support neuronal survival and synaptic plasticity, influence immune responses by shifting microglia and macrophages toward anti-inflammatory phenotypes, and promote angiogenesis and neurogenesis [49, 51]. Among MSC sources, hBMSC are well studied, but their clinical use is limited by the invasiveness of bone marrow aspiration and donor variability [12, 13, 52, 53]. Alternative sources such as adipose tissue and perinatal tissues (e.g., amnion, umbilical cord) offer less invasive harvesting, greater proliferative capacity, and in some cases, higher cytokine secretion profiles [33, 54].

hUCMSC is anattractive for clinical translation. They can be obtained non-invasively from postnatal tissue otherwise discarded, expanded efficiently in vitro, and have demonstrated potent immunomodulatory and neuroprotective effects in preclinical models of ischemic stroke and neurodegeneration [55, 56]. Several studies have also explored the therapeutic potential of hUCMSCs in experimental ICH. Intravenous or intracerebral transplantation of hUCMSCs has been reported to reduce hematoma volume, suppress neuroinflammation, and promote neurological recovery in both mouse and rat ICH models [57, 58]. These findings collectively support the neuroprotective and immunomodulatory potential of hUCMSCs in hemorrhagic stroke. However, few studies have comprehensively examined their role in ICH, especially in relation to both brain and peripheral immune responses. Prior work suggests that peripheral immune modulation can significantly influence CNS inflammation and recovery after stroke [59], but direct evidence in the context of hUCMSC therapy for ICH has been limited.

Against this background, the present study was designed to test the hypothesis that intravenous administration of hUCMSC could improve neurobehavioral outcomes after ICH through coordinated modulation of local and systemic inflammation. By integrating behavioral assays, molecular profiling, histological analyses, and systemic immune cell characterization, we aimed to elucidate both the neuroprotective and immunoregulatory mechanisms underlying hUCMSC therapy, and to identify potential translational markers that could guide future clinical applications.

At the molecular level, hUCMSC treatment modulated inflammatory and neuroprotective gene expression in the perihematomal region (Fig. 3). Notably, mTSG6 expression was significantly upregulated. At the same time, TNFα was reduced, suggesting that hUCMSC may exert anti-inflammatory effects partly through the induction of endogenous TSG6, which is known to suppress TNFα-mediated inflammatory cascades [54, 56]. Immunohistochemistry further revealed that TNFα-positive cells were already reduced at Day 3 post-ICH in the perihematomal region (Fig. 2), even though whole-brain qPCR differences were not significant at this time point, implying a localized early anti-inflammatory response. In parallel, hUCMSC treatment enhanced GDNF and VEGFα expression in the perihematomal brain, both of which are known to promote neuronal survival, synaptic plasticity, and angiogenesis [60, 61]. Although these factors are not directly linked to inflammation, their upregulation, alongside the suppression of microglial activation (Fig. 2), suggests that hUCMSC therapy orchestrates both neuroprotective and anti-inflammatory mechanisms that converge to improve behavioral outcomes.

Histological analyses confirmed these molecular findings: TUNEL staining demonstrated a significant reduction in apoptotic cells in the perihematomal region at Day 3 (Fig. 2), indicating that hUCMSC attenuates cell death during the early injury phase. This aligns with prior evidence that MSC therapy reduces apoptosis through paracrine-mediated neurotrophic support [24, 59].

Beyond the CNS, our in vitro coculture experiments (Fig. 4) showed that hUCMSCs suppress LPS-induced pro-inflammatory gene expression (CCL2, F4/80, Ly6C, IL-6) and enhance anti-inflammatory CD206 expression in immune cells from spleen, blood, bone marrow, and brain. These results indicate that hUCMSC directly attenuate inflammatory responses while promoting anti-inflammatory programs, thereby exerting a cell-intrinsic immunomodulatory effect across multiple immune compartments. Importantly, flow cytometry analysis (Figs. 5 and 6) demonstrated that hUCMSC administration alters systemic myeloid cell dynamics in vivo, including transient mobilization followed by suppression of Ly6G⁺ neutrophils and a reduction of inflammatory Ly6C^hi monocytes. In parallel, hUCMSC treatment in the brain reduced CD86⁺ myeloid cells across multiple time points and selectively increased CD163⁺ macrophages, indicating a shift toward an anti-inflammatory phenotype. Together, these systemic and central immunomodulatory effects likely contribute to the observed reduction in CNS inflammation, consistent with prior studies linking peripheral immune suppression and microglial polarization to neuroprotection in ICH and stroke [24, 55, 56].

Taken together, our findings support a dual mechanism by which hUCMSC therapy ameliorates ICH-induced damage: (1) local suppression of neuroinflammation and apoptosis within the perihematomal brain, and (2) systemic immunomodulation that limits peripheral myeloid activation and infiltration into the CNS. This integrated regulation of both central and peripheral immune responses may explain the robust functional recovery observed, particularly with repeated high-dose administration. Clinically, these results suggest that optimizing cell dose and timing could maximize the therapeutic potential of hUCMSC for ICH patients, warranting further translational and clinical studies.

In addition to these findings, this study presents several novel contributions. The most distinctive feature is the simultaneous evaluation of both central (brain) and peripheral (spleen, blood, bone marrow) immune responses within the same experimental framework, demonstrating that hUCMSCs exert coordinated immunomodulatory effects at both sites. Notably, high-dose, repeated intravenous administration yielded the most pronounced functional recovery, consistent with previous reports of dose-dependent benefits in MSC therapy [33, 55]. Furthermore, the temporal and spatial correspondence between increased TSG6 expression and decreased TNFα, together with reduced microglial activation and diminished apoptotic cell death, supports a working model in which hUCMSCs promote resolution of inflammation and confer neuroprotection via paracrine mechanisms, possibly involving the TSG6–CD44 axis.

Several findings regarding VEGF expression are consistent with previous studies. In our model, VEGFα mRNA exhibited a transient rise during the early subacute phase (Days 3–4), suggesting an angiogenic response shortly after hUCMSC administration. This pattern aligns with previous reports showing that MSC transplantation enhances angiogenesis after ICH. For instance, Bao et al. demonstrated that Flk-1⁺ human bone marrow–derived MSCs significantly increased vessel density in the perihematomal region of rat ICH models [62]. Broader reviews have also described that MSCs promote angiogenesis and vascular remodeling through the secretion of angiogenic factors such as VEGF and bFGF in hemorrhagic stroke contexts [63]. Moreover, in a non-stroke setting, Wu et al. showed that hUCMSCs stimulate endothelial migration, proliferation, and tube formation via the integrin β1/ERK1/2/HIF-1α/VEGF-A signaling axis [64]. Underscoring the inherent angiogenic potential of these cells. Taken together, these data support the notion that hUCMSCs contribute to neurovascular repair following ICH by transiently upregulating VEGF and related trophic factors. Second, TNFα and GDNF displayed a “local versus whole-brain” discrepancy: immunohistochemistry detected significant perihematomal changes at Day 3, whereas whole-brain qPCR often failed to reach statistical significance. This may reflect dilution of localized effects in whole-tissue analyses and underscores that hUCMSC actions may initiate in peri-lesional niches. Third, the transient early increases in Ly6G⁺ and Ly6C^low cells—followed by subsequent declines—differ from studies reporting immediate suppression. This biphasic pattern could represent an initial mobilization phase facilitating debris clearance, followed by immune resolution, and may be specific to the acute-phase high-dose regimen used here.

From a translational perspective, these findings support three important strategies for clinical development. First, optimizing dose and schedule—especially combining an initial high-dose bolus with early repeated dosing—may enhance functional recovery. Second, identifying early pharmacodynamic markers, such as TSG6 upregulation, TNFα suppression, and reduction of microglial activation, could inform the individualized timing of administration. Third, the dual central–peripheral mechanism observed here suggests that hUCMSC therapy could synergize with conventional surgical and medical interventions by dampening systemic immune activation while concurrently suppressing local neuroinflammation. The off-the-shelf availability and intravenous delivery of hUCMSC further enhance the feasibility of translation to clinical practice.

The translational potential of these findings is considerable. The observed dual modulation of systemic and local immune responses by hUCMSC suggests that similar therapeutic benefits may be achievable in human ICH, where systemic inflammation and local neuroinflammation are both key determinants of outcome [56, 59]. The intravenous delivery route—minimally invasive and suitable for rapid administration—aligns well with the acute stroke care pathway, facilitating integration with standard treatments such as hematoma evacuation, blood pressure management, and neurocritical care monitoring. The dose-dependent effects observed here, with high-dose, repeated administration yielding maximal functional recovery, support the rationale for early-phase clinical trials testing escalating-dose regimens in patients with ICH. Biomarkers identified in this study, including perihematomal TNFα suppression, TSG6 upregulation, reduction of microglial activation, and modulation of circulating myeloid subsets, could serve as early pharmacodynamic endpoints to guide dosing schedules and patient selection.

Furthermore, the anti-apoptotic and neurotrophic effects observed—exemplified by fewer TUNEL-positive cells and increased GDNF expression—suggest that hUCMSC therapy could expand its benefit window beyond the hyperacute phase, targeting ongoing secondary injury processes. This is especially relevant in clinical ICH, where perihematomal edema, microglial activation, and delayed neuronal death continue for days to weeks post-onset [54, 60].

While differences between rodent models and human pathology must be acknowledged, the robust and multifaceted mechanisms demonstrated here—including paracrine-mediated immunomodulation, trophic support, and apoptosis inhibition—enhance the plausibility of successful translation. Indeed, a recent Phase I dose-escalation trial using intravenous MSCs in acute ICH patients demonstrated feasibility and safety, with patient inclusion stratified by hematoma volume (< 60 mL) and neurological status, and utilized cytokine-based biomarkers to assess therapeutic response [65]. Furthermore, preclinical meta-analyses have shown that the timing of MSC administration (e.g., within 8 h vs. 24 h post-ICH) substantially influences efficacy in terms of both structural and functional outcomes [66]. Therefore, future trials should stratify by hematoma volume, timing of cell administration, and baseline systemic inflammatory status to optimize therapeutic efficacy. Ultimately, these data reinforce the idea that hUCMSC therapy, delivered intravenously with precise timing, could act as an adjunct treatment to enhance neurological recovery and decrease disability in patients with ICH.

Conclusion

In conclusion, intravenous administration of hUCMSC provides neuroprotection in a murine ICH model by simultaneously reducing local neuroinflammation and cell death in the perihematomal brain and modulating systemic myeloid responses. These combined central and peripheral immune effects lead to notable improvements in motor, cognitive, and emotional functions. The findings emphasize the need to optimize dose and timing to maximize treatment benefits, supporting the potential for clinical translation, especially as an adjunct to standard treatments for ICH.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

S.T., Y.K., Y.T. and N.D. performed experiments. S.T. and Y.K. wrote the main manuscript text and prepared figures. All authors reviewed the manuscript.

Funding

This work was supported by JSPS KAKENHI Grant Number JP 22K09271.

Data Availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics Approval and Consent to Participate

Informed consent was obtained from all women according to our institutional ethics committee (approval number 325).

All experiments were approved and confirmed by the Ethics Review Board of Hyogo Medical University (Approval numbers: 21 − 012 and 24–047 A).

Competing interests

The authors declare no competing interests.