Intracerebroventricular Administration of hNSCs Improves Neurological Recovery after Cardiac Arrest in rats

Abstract

Irreversible brain injury and neurological dysfunction induced by cardiac arrest (CA) have long been a clinical challenge due to lack of effective therapeutic interventions to reverse neuronal loss and prevent secondary reperfusion injury. The neuronal regenerative potential of neural stem cells (NSCs) provides a possible solution to this clinical deficit. We investigated the neuronal recovery potential of human neural stem cells (hNSCs) via intracerebroventricular (ICV) xenotransplantation after CA in rats and the effects of transplanted NSCs on the proliferation and migration of endogenous NSCs. Outcome measures included neurological functional recovery measured by neurological deficit score (NDS), electrophysiologic analysis of EEG, and assessment of proliferation and migration at the cellular level and the Wnt/β-catenin pathway at the molecular level. Neurological functional assessment based on aggregate neurological deficit score (NDS) showed better recovery of function after hNSCs therapy (P<0.05). Tracking of stem cells’ proliferation with Ki67 antibody suggested that the NSCs group had more prominent proliferation compared to control group (number of Ki67+ cells, Control VS. NSC: 89.0±31.6 VS. 352.7±97.3, P<0.05). In addition, cell migration tracked by Dcx antibody showed more Dcx+ cells migrated to the far distance zone from SVZ in the treatment group (P<0.05). Further immunofluorescence staining confirmed that the expression of the Wnt signaling pathway protein (β-catenin) was upregulated in the NSC group (P<0.05). ICV delivery of hNSCs promotes endogenous NSC proliferation and migration and ultimately enhances neuronal survival and neurological functional recovery. Wnt/β-catenin pathway may be involved in the initiation and maintenance of this enhancement.

Graphic abstract

INTRODUCTION

Cardiac arrest (CA) has remained a focus of clinical and preclinical research due to high mortality and morbidity worldwide. It is estimated that 290,000 in-hospital CAs and 350,000 out-of-hospital CAs occur in the U.S. per year [1]. Despite decades of study, targeted temperature management remains the only medical intervention available to improve neurological function after CA [2]. Even still, no post-CA intervention has been identified to reverse neuronal loss secondary to CA, and 92.6% of CA survivors are discharged with neurological dysfunction [3].

Neural stem cells (NSCs), which persist in selected regions of the brain throughout the life span [4], are pluripotent and self-renewing cells with the ability to generate all of the major cell types of the central nervous system (CNS) in the adult animal [5], and they play regenerative and reparative roles in response to brain injury or disease, making them an exciting new tool in the treatment of ischemic brain injury [6–8]. Under different conditions, NSCs hold the potential for differentiating into neurons and glial cells, and projecting onto the host neurons after global ischemia [9]. Improved functional recovery in neonatal hypoxic ischemic encephalopathy (HIE) and in stroke has been demonstrated by exogenous administration of neural stem cells in preclinical studies [10, 11]. It has also been proposed that NSC administration is capable of activating the proliferation and migration of endogenous neural stem cells, which exist in the dentate gyrus (DG) and subventricular zone (SVZ) [12].

Evidence in the ischemic stroke model showed that SVZ cells contributed to repair after regional ischemia injury [13]. However, different from regional ischemic injury of stroke, global ischemic injury after CA may have a negative impact on SVZ cells. Targeting the activation of SVZ cells with NSCs transplantation may ameliorate or reverse this impact and may achieve a more promising therapeutic effect than ischemic stroke. ICV injection of stem cells enables widespread cerebral engraftment of cells along the entire neuroaxis [14]. Given the proximity of the SVZ to the cerebral ventricles, the ICV route may have advantages in activating SVZ cells.

While the functional recovery potential of stem cell therapy in ischemic brain injury has been promising within the preclinical literature, the underlying mechanisms of this therapy remain to be elucidated. The Wnt family is a complex network of nineteen secreted glycolipoproteins [15]. Wnt-related proteins are active in a diversity of stem cells, including neural, mammary, and embryonic stem cells [16]. Wnt signaling has been implicated in stem cell self-renewal because of its crucial roles in regulating cell proliferation, differentiation, migration, and patterning in embryonic development and tissue homeostasis [16–18]. Accumulated evidence also suggested that Wnt signaling pathway activation was involved in the repair of various brain and spinal cord injuries [19] and enhanced expression of β-catenin closely related to the protective effect on motor function as well as limiting apoptosis in mice ischemic stroke models [20]. Given the seemingly critical role of Wnt signaling in stem cell pathways, we hypothesize that this pathway is also implicated in the underlying mechanisms of exogenous neural stem cell therapy.

The role of xenotransplantation of exogenous neural stem cells in the proliferation and migration of endogenous neural stem cells and timing of such therapies, as well as their effects on functional outcomes in brain injury after cardiac arrest, has not been reported in the current literature base. In this study, we hypothesize that ICV administration of xenotransplantated exogenous NSCs will enhance endogenous NSC proliferation and migration, promote neuronal survival, and ultimately improve the neurological functional outcome after CA. Furthermore, the possible neuroprotective mechanisms of NSCs in brain resuscitation via regulation of the Wnt signaling pathway was explored.

Materials and Methods

Experimental procedures of hNSCs therapy after cardiac arrest are shown in Figure 1.

Fig 1.

Experimental procedures of hNSCs therapy after cardiac arrest. (N = 11/group)

Animal model

All animals were maintained in accordance with NIH guidelines for the humane care of animals. Experimental protocols were reviewed and approved by the IACUC of University of Maryland School of Medicine Animal Care and Use Committee. Twenty-two adult Long–Evans rats (14 Male and 8 female, average weight 325.6 ± 5.5g, Charles River, Wilmington, MA) were randomly assigned into the NSC treatment group (NSC, N=11, 7 male/4 female) and control group (Control, N=11, 7 male/4 female). All rats were housed in a 12-hour light-dark cycle space with free access to water and food. CA was induced by 8 min asphyxia according to the detailed model outlined in our previous studies [21]. Briefly, rats were intubated following anesthetization with isoflurane. Anesthesia was maintained with 1.5% isoflurane mixed in gas with 50% O₂ through a ventilator after intubation. The femoral artery was cannulated to collect blood samples used in analysis of arterial blood gases (ABG) and to observe and record dynamic changes of mean arterial pressure (MAP). The ipsilateral femoral vein was used to inject drugs and fluids. Heart rate, body temperature, and depth of anesthesia were monitored in real-time. EEG baseline was recorded for 5-min, followed by a 5-min washout period (ventilated with air without isoflurane) to minimize the influence of anesthesia on the EEG results. Muscle relaxant, vecuronium (2 mg/kg), was injected into the rat’s femoral vein to inhibit respiratory muscle movement 3 minutes prior to ventilator discontinuation, while the animals remained anesthetized. Asphyxia was achieved by clamping the ventilator circuit and turning off the ventilator for 8 min. The time point of CA was confirmed and recorded when ECG waveform became flat and the MAP was <10 mmHg. Subsequently, ventilator support and chest compressions with a frequency of 200bpm were instantly performed. Drugs that supported resuscitation (epinephrine 5μg/kg, sodium bicarbonate 0.84mg/kg) were administrated to reverse CA and balance metabolic acidosis. Return of spontaneous circulation (ROSC) was defined as MAP>60 mmHg. 100% oxygen was supplied, and no further isoflurane was given after ROSC. ABG was tested 20 minutes after ROSC and ventilator parameters were adjusted according to ABG results. Throughout the duration of the experiment, the temperature of rats was continuously monitored via a rectal temperature probe and was controlled between 36 °C and 37.5 °C by physical methods (heating with heating pad or cooling by fan). All rats were included in the study unless they met the following two exclusion criteria: 1. asphyxia induction failure (without cardiac arrest); 2. ROSC was not achieved within 2 minutes CPR.

EEG record and analysis

EEG signals were continuously recorded using a four-channel EEG system (Tucker-Davis Technologies, USA). Five electrodes were implanted in the skull of rats (2 on left, 2 on right and 1 as reference) 24h before CA surgery. Complete EEG data consisted of two durations: initial 5 min baseline signal prior to CA, followed by 360 min continuous signal post-ROSC. A custom algorithm using Matlab (MathWorks, Natick, MA) was used to analyze all of the EEG data and transfer the graphical data into the quantitative form. Finally, the quantitative results were exported in the form of qEEG Information Quantify (qEEG-IQ) values to represent modified entropy, which was published in our previous studies [21–23].

Human Neural Stem Cells (hNSCs) administration

The hNSCs (ReNcellVM Human Neural Progenitor Cell Line, EMD Millipore SCC008) were maintained in NSC Maintenance Medium (Millipore, Cat# SCC008) supplemented with 20ng/ml basic fibroblast growth factor (bFGF, Thermo Fisher, Cat# 13256029) and 20ng/ml epidermal growth factor (EGF, Thermo Fisher Cat# PHG0314). All cell culture vessels for maintenance of hNSCs were coated with 20μg/ml Laminin (Sigma, L2020) overnight before use [24]. The rats were anesthetized and placed on the stereotaxic instrument. To expose the skull, a 1–1.5cm incision was made in the skin. An entry point (P, 1.2mm, L 1.5mm, D 4mm) was then drilled into the skull using a highspeed drill bit (0.6mm in diameter). A vertical 24-gauge needle was penetrated into the left lateral ventricle to slowly administer 2.0×10⁵ hNSCs in 5 ul saline solution in the NSC group or equivalent volume of saline without hNSCs in the control group at 3h after ROSC. The needle was kept in place for 1 min to avoid leakage of the solution. The needle was slowly removed, and the skin was closed with a 4–0 suture.

Neurological function evaluation

Assessment of neurological recovery was performed using well-validated neurological deficit scores (NDS) with a total score of 80 [21, 25]. The evaluations were conducted at four-time points of 6, 24, 48 and 72h from ROSC. We evaluated 7 aspects of arousal, brainstem function, motor assessment, sensory assessment, motor behavior, behavior, and seizures as the overall NDS. Furthermore, a sub-NDS analysis including motor behavior and behavior was used to evaluate the advanced motor function and behavior of rats [26]. The total score at different time points was taken as an independent evaluation index, and the overall NDS at 72h was defined as the primary endpoint of neurological outcome. In addition, the survival times of animals that died within 72 hours were recorded and the survival time of survival rats was recorded as 72 hours.

Cresyl violet staining

Cresyl violet staining was chosen to evaluate the viability of the cells by observing nuclear and nucleolar structures of neurons and was used to determine the degree of ischemic injury. The preparation of each frozen slice was conducted in accordance with a standard rat brain atlas. Four characteristics, including pyknosis, karyorrhexis, karyolysis and cytoplasmic changes in form and color [22], were used to identify ischemic injured neurons under ×400 magnification in a bright field of the Leica DMi8 microscope (Leica Microsystems). Histopathological damage scoring (HDS), as a quantitative index of neuronal damage, was defined as the percentage of ischemic or dead neurons in the whole area of region of interest (including hippocampus CA2, CA3) 72 hours after ROSC.

Tracking of transplanted NSCs and characterization of SVZ cells proliferation and migration

The rats were sacrificed 72 hours after ROSC and brains were sliced in a coronal plane after the rats were perfused with phosphate-buffered saline and fixed with 4% paraformaldehyde. For the immunofluorescence (IF) staining, brains were cut into 10 μm coronal slices on a freezing microtome (Leica, Wetzlar, Germany). Mouse anti-NeuN (1:100, Alexa 488 conjugate, Millipore Sigma) and mouse anti-Ku86 (1:100, Santa Cruz Biotechnology, Dallas, TX) were selected as the primary antibodies to detect the neurons and transplanted hNSCs in the region of hippocampus (CA2, CA3, CA4, DG). For the quantification of surviving neurons, the number of NeuN-positive cells (200X field) was divided by field area, and output as cell density (count/mm²). The mouse anti-doublecortin (Dcx) (1:100, Abcam, San Francisco, CA) antibody was chosen to track the migration of SVZ cells. Relative migration, which was quantified by the number of Dcx-positive cells in the unit distance (25 μm as a distance unit) divided by the number of all Dcx-positive cells in the microscope field, and exported as a percentage, was used to describe the tendency of cells departing from the SVZ. Absolute migration was calculated by the absolute value of the number of Dcx-positive cells in the unit distance (25 μm) and was used to describe the number of cells involved in migration. Mouse antiKi67 (1:50, Santa Cruz Biotechnology, Dallas, TX) and rabbit anti-β-catenin (1:200, Sigma-Aldrich, St. Louis, MO) were used to mark the proliferation and Wnt/β-catenin pathway of SVZ cells respectively. Ku86 and Ki67/β-catenin/Dcx double staining were performed in SVZ to determine whether the grafted stem cells were involved in proliferation and migration. Dcx and Ki67 staining were performed in the hippocampus, and quantitative comparison was made between groups. Each group included five animals for the immunofluorescence quantitative analysis and each animal was used to provide one random SVZ for calculating all positive cells (Ki67+, β-catenin+, and Dcx+) in the entire SVZ. All fluorescence micrographs were obtained with Leica DMi8 microscope (Leica Microsystems). The number of positive cells in identical areas was measured using ImageJ software (National Institute of Health, Bethesda, MD) [27].

Statistical Methods

Parametric data (qEEG-IQ) were shown as mean ± S.E.M. and non-parametric data (NDS) were presented as median and 25–75 interquartile range. One-way ANOVA was used to analyze qEEG-IQ. Repeated measurement was used to analyze overall NDS. Cell numbers and HDS are presented as Mean ± SEM and univariate tests were used in the analysis of HDS and cell numbers. Differences were considered significant at P < 0.05. All statistical analyses were performed using SPSS (Version 22, IBM SPSS Statistics, Armonk, NY).

RESULTS

Baseline and EEG Characteristics

The baseline ABG data are shown in Table 1. There was no significant difference between the Control and NSC group in the baseline indices related to respiration (PH, PCO₂, PO₂, HCO3⁻, Base excess, SO₂) and organ perfusion (Lactate).

Table 1.

ABG baselines in Control and NSCs group

	Control	NSCs	P-value
pH	7.43±0.02	7.43±0.01	0.922
PCO2, mmHg	39.1±1.4	39.7±1.4	0.759
PO2, mmHg	176.4±6.5	191.4±7.7	0.161
HCO3−, mmol/L	25.8±0.8	26.3±0.6	0.631
Base excess	1.9±1.0	2.0±0.6	0.908
Lactate, mmol/L	1.01±0.12	1.04±0.18	0.884
SO2, %	100±0.0	100±0.0

EEG signals fell to flat waveforms (qEEG-IQ=0) from baseline (qEEG-IQ=1) soon after CA, and gradually recovered after ROSC. There was no significant difference in IQ value between the two groups before hNSC treatment (Control VS. NSC: 0.92 ± 0.04 VS. 0.88 ± 0.04, p>0.05). The comparison of the original EEG tracings at 3h post-ROSC found a greater proportion of abnormal waveform bursts with high amplitudes (1–1.8mV) lasting 2–6 seconds in the control group (Supplementary Figure 1).

Neurological functional recovery and survival

NDS at 4-time points showed that both groups gradually recovered, with increasing NDS over time (Supplementary Table 1). Aggregated NDS analysis showed the NSC group has significantly better NDS than the control group (P<0.05). Among them, there were significant differences of NDS between treatment and control groups at 48h (median, 25^th–75^th percentile, Control VS. NSC: 50.0, 47.0–55.0 VS. 67.0, 54.50–71.25; P=0.013) and 72h (Control VS. NSC: 53.0, 48.25–63.25 VS. 72.5, 63.75–75.75; P=0.027) (Fig 2A). The NSC group had better recovery in aggregated sub-NDS in motor behavior and behavior when compared to the Control group (Supplementary Table 2, P = 0.017). The analysis at four time points between groups showed a similar trend as overall-NDS, with improvement of sub-NDS most prominent at 48h (Median, 25^th–75^th percentile, Control VS. NSC: 0, 0 −2.0 VS. 7.0, 0 −14.0, P = 0.009) and 72h (Control VS. NSC: 0, 0 – 4.0 VS. 7.0, 0 −16.0, P = 0.026) (Fig 2B). Cumulative survival time showed longer survival in the treatment group, though without statistical significance (68.3 ± 3.2h VS. 65.3 ± 3.2h, P >0.05). The Kaplan Meier survival curve showed there was no difference between groups (8/11, 72.7% for both groups; Supplementary Fig 2).

Fig 2.

NDS and sub-NDS comparison in the NSC and Control groups. Aggregated NDS analysis showed the NSC group has significantly better NDS than the control group (P<0.05). (A) NSC groups had a significant neurological functional improvement at 48 (P = 0.013) and 72h (P = 0.014), compared to the control group. (B) Sub-NDS comparison at four time points between groups. The improvement of sub-NDS was prominent in the NSC group at 48h (P = 0.009) and 72h (P = 0.026). (*P<0.05; ** P<0.01).

Cresyl Violet staining and neuronal damage

We found a high density of ischemic injured neurons in CA2 compared to CA3 region in both groups (Fig 3A, B). Based on HDS, the control group had a higher percentage of neuronal damage in CA2 (Control VS. NSC: 51.3%±9.7% VS. 32.3%±3.8%; P<0.05) and CA3 (Control VS. NSC: 14.7%±4.5% VS. 10.0%± 2.6%; P=0.197), compared to the treatment group (Fig 3C).

Fig 3.

Cresyl-violet staining and quantification of ischemic neurons with histopathological damage scoring (HDS). 200X, scale bar=25μm. (A) (B) Cresyl-violet staining 3d after CA in the Control and NSC group. Yellow arrows indicate ischemic injured neurons in CA2 and CA3. (C) HDS quantification comparison of control and NSC groups in CA2 and CA3 showed a higher proportion of damaged neurons in CA2 of control group (*, P<0.05).

hNSCs tracking and neuronal survival

We used human nuclei Ku-86 antibody and neuronal nuclei NeuN antibody to recognize hNSCs and neurons in the injured brain. The Ku-86 positive cells were primarily located in the hippocampus but not SVZ 3 days after hNSC administration (Fig 4A, C, E). Compared to the Control group, the NSC group had more NeuN positive cells in CA2 (Control VS. NSC: 673.0±94.4/mm² VS. 1328.8±156.2/mm², P < 0.01, Fig 4B), CA3 (Control VS. NSC: 361.5±89.8/mm² VS. 926.8±98.3/mm², P < 0.01, Fig 4D), and DG (Control VS. NSC: 2548.7±433.3/mm² VS. 4575.2±709.4/mm², P < 0.05, Fig 4F). It is notable that, immunofluorescence staining of Ku86 from the same rats showed rich Ku86 positive cells were found in the hippocampus, while no Ku86 positive cells were found in SVZ. The migration of grafted NSCs might have occurred earlier than the host NSCs, which meant that measurement at the later time point within the SVZ, where migration begins, did not capture grafted cells that had already migrated (Supplementary Fig 3).

Fig 4.

Grafted hNSC tracking and neuronal survival at 3 days post-ROSC. Neuronal (NeuN+, green) survival and hNSC (Ku86+, red) tracking in CA2 (A), CA3 (C), CA4, and DG (E). Quantitative analysis showed the NSC group had a better neuronal survival than the control group in CA2 (B), CA3 (D), and DG (F). (*P < 0.05, **P < 0.01). 400X, Scale bar = 25μm.

SVZ cells proliferation and Wnt/β-catenin signal pathway

As a nuclear protein that is expressed in proliferating cells, which may be required for maintaining cell proliferation, Ki67 was used to mark and quantify the proliferation of cells in the SVZ [4]. β-catenin was chosen to mark the Wnt/β-catenin pathway based on its key role in the downstream components of the canonical Wnt/β-catenin signaling pathway [28]. More Ki67 positive cells were found in the SVZ of the lateral ventricles (LV) in the NSC group, which was accompanied by the high expression of β-catenin, compared with the control group (Fig 5A). Further quantitative analysis of cell numbers showed that the expression of the two proteins was statistically different between the two groups (Ki67, Control VS. NSC: 89.0±31.6 VS. 352.7±97.3, P = 0.042, Fig 5B; β-catenin, Control VS. NSC: 80.8±49.4 VS. 426.2±121.7; P = 0.025, Fig 5C). In order to exclude the possible interference of grafted NSCs vs. endogenous SVZ cells in Ki67/β-catenin expression, we used Ki67/β-catenin and Ku86 double labeling respectively. The results showed that a small number of Ku86 positive cells had no significant effect on the above results (Fig 6A/B). A consistent trend was also observed when we compared the number of Ki67 positive cells in the hippocampus represented by CA2 (Ki67, Control VS. NSC: 19.6±4.9 VS. 80.4±19.2, P = 0.015, Supplementary Figure 4).

Fig 5.

Expression of Ki67 (green) and β-catenin (red) in SVZ (left) with immunofluorescence. Compared with the control group, more Ki67 and β-catenin (A) positive cells were found in the NSC group 3 days after CA. The quantitative analysis showed that the number of Ki67 (B) and β-catenin (C) positive cells were significantly higher in the NSC group compared to the Control group (*P < 0.05). 400X, Scale bar = 25μm

Fig 6.

Representative figures showed double labeling with Ku86 and Ki67/Dcx in SVZ. (A) Grafted cells (Ku86 positive, indicated by red arrows) were scattered in the area far away from the edge of the ventricle and did not overlap with most Ki67 positive cells. (B) There was no overlap between grafted cells (Ku86 positive cells, indicated by green arrows) and region with β-catenin signals.

Migration of SVZ cells

We then tracked and analyzed the migration of SVZ cells using the Dcx antibody. Dcx is a protein which appears to direct neuronal migration by regulating the organization and stability of microtubules [29]. More Dcx-positive cells were in the area proximal to the SVZ in the NSC group (Fig 7A). In the control group, more relative migration existed in the area of 0~50μm from the SVZ (Control VS. NSC: 0–25μm, 28.89% ± 2.62% VS. 8.76%± 1.32%; 25–50μm, 20.3% ±2.23% VS. 13.130 ± 0.89%). On the contrary, in the area of 75~150μm from the SVZ, there was a relatively high percentage of positive cells in the NSC group (Control VS. NSC: 75–100μm, 13.67% ± 1.29% VS. 23.26% ± 1.83%; 100–125μm, 9.3% ±1.86% VS.22.4% ± 2.57%; 125–150μm, 5.61% ± 0.85% VS. 14.57% ± 2.12%, Fig 7B). The number of cells was dominant on absolute migration in the NSC group at all distances, with significantly higher absolute migration of positive cells in the distant area (>75μm), compared to the control group (Fig 7C). In order to determine the extent and proportion of grafted cells involved in the migration in the SVZ, we used Ku86 and Dcx co-staining in the SVZ and found that, although there were several overlaps between Ku86 positive cells and Dcx positive cells (red arrows), the grafted cells were not the main group involved in migration (Fig 7A, lower right). In the hippocampus (CA2, CA3 and DG), quantification results showed that the absolute number of Dcx positive cells involved in migration in the NSC group was significantly more than that in the control group (Control VS. NSC: 86.6±16.8 VS. 193.4±33.5, P = 0.021, Supplementary Figure 5).

Fig 7.

Migration of SVZ cells marked by Dcx and Ku86 co-staining. (A) The number of migration cells (Dcx+) in SVZ were dominant in NSC group; Cells double-labeled by Ku86 and Dcx (indicated by red arrows) accounted for a very low proportion of all Dcx positive cells, suggesting that the grafted cells were not the main component of the migration group. (B) Quantified relative migration (proportion of Dcx+ cells in a unit region/whole area) showed a high percentage of positive cells in the NSC group. (*P < 0.05; **P < 0.01); (C) Quantified absolute migration (number of Dcx+ cells at a different distance from the SVZ) showed the cell number was dominant on absolute migration in the NSC group at all regions and significantly higher in the distant regions (>75μm), compared to the control group. (*P < 0.05; **P < 0.01) 400X, Scale bar = 25μm

DISCUSSION

Stem cell therapy, which targets neuronal proliferation, migration and redistribution, is a promising therapeutic strategy, given that the loss and death of neurons are the pathological basis of brain injury after CA [10]. The search for pharmacologic interventions to improve neuronal survival and neurological recovery persists, and this study presents ICV xenotransplantation of hNSCs as a potential therapy to fill this void. Prior in vivo studies have shown the therapeutic efficacy of NSCs transplantation in both animal models [12] and clinical studies [30, 31] of CNS disorders. Our study demonstrates rapid engraftment of exogenous NSCs directly infused by the ICV route, and we present the first preclinical study demonstrating promotion of endogenous NSC proliferation and migration with resultant improvement in neuronal survival and early neurological functional recovery after cardiac arrest. These findings suggest that NSC transplantation can partially reverse the neuronal injury caused by CA.

At the cellular level, the treatment group showed greater expression of NeuN in the hippocampus, indicating that treatment with hNSCs led to better survival of neurons, presumably contributing to significantly improved functional recovery. These functional improvements were most notable at 48h and 72h after CA, suggesting that early delivery of therapy (within 3h of CA) provides a subsequent early therapeutic effect and may impact recovery in the first days, starting at 2 days following ROSC. This early timing is consistent with a stroke model, in which rats receiving a hippocampal hNSC injection 24 hours after middle cerebral artery occlusion (MCAO) showed an improvement in the foot-fault and adhesive removal tests administered 72h later [32]. Stem cell therapy for global cerebral ischemia models involving 5–6 minutes of cardiac arrest, which leads to a mild brain injury, has been demonstrated beneficial at 3 days with adipose-derived stem cells [33] or bone marrow stem cells [34]. Intraventricular implantation of mesenchymal stem cells was reported to increase neurogenesis in the SVZ and the DG [35]. In cardiac arrest, a therapy with potential to provide early improvements in functional recovery could have dramatic implications in neuro-prognostication and post-CA care. Despite significant improvement in functional outcome with the NSC therapy, there is no significant difference between the two groups in in survival rate or cumulative survival time. This is possibly due to small sample size with limited power, limited observation time (72 hours after resuscitation), and/or the brain microenvironment post-ischemic injury with increased neuroinflammation [36], resulting in neuronal cell death.

We aimed to further understand the cellular mechanisms involved in endogenous NSC proliferation and migration in the setting of exogenous NSC transplantation. Our study showed human NSC transplantation promoted endogenous cell proliferation and migration in SVZ. The treatment of hypoxic-ischemic brain injury using stem cells is mainly based on one of two basic mechanisms: direct delivery of exogenous stem cells or activation of endogenous stem cells [12]. Proposed mechanisms of exogenous stem cells treatment in hypoxic-ischemic brain injury have included direct neuronal replacement, immunomodulation, secretion of neurotrophic factors to support neurogenesis, and the promotion of neurogenic potential of endogenous NSCs [37, 38]. Endogenous NSCs remain in a phase of quiescence for prolonged periods within the brain [39], and are able to re-enter the cell cycle and proliferate under particular conditions [40]. Experimental models of ischemic brain injury, predominantly stroke, have demonstrated that the proliferation and migration of endogenous NSCs, particularly in the SVZ, are key spontaneous regenerative processes that occur within the ischemic brain [12], via vessel scaffolds, generation of neuroblasts, replacement of apoptotic neurons, and improved functional outcomes [41, 42], with crosstalk between transplanted exogenous cells and endogenous NSCs related to neurogenesis [43]. The migrated newly generated cells, mostly derived from stem cells located in the DG and SVZ, ultimately replaced dying neurons in damaged regions [41]. Intravenously administered bone marrow stromal cells (MSCs) accelerated the proliferation of endogenous cells, and improved neurological outcome after stroke [44].

Our results demonstrate that SVZ cells retain and exhibit the ability to proliferate, even after global cerebral ischemia, and this ability was more pronounced after exogenous NSC transplantation. It is notable that SVZ cells are composed of different cell types, including Type E (ependymal cells), Type B (B1 and B2, both had astrocyte characteristics while B2 cells derived from B1 cells but do not possess an apical contact), Type C (transient-amplifying cells generate from Type B cells), and Type A (young neurons that raised from Type B cells) cells [45]. NSCs in the SVZ and DG show the characteristic astroglial properties [46, 47], which makes it wildly accepted that this subtype of astrocytes in the SVZ (B cells) serve as endogenous NSCs [48–50]. For ependymal cells (E cells), studies have shown that they do not divide in the adult [51], and recent studies working on single-cell transcriptomic and lineage analysis have proven that E cells have no progenitor properties or function as NSCs [52]. Based on SVZ cell characteristics, using Ki67 or BrdU to label neurogenesis in the SVZ is a wildly accepted method [13, 35]. Microtubule-binding protein Dcx, which is transiently expressed in proliferating progenitor cells and newly generated neuroblasts, serves as a marker of migration [53]. The existing research results show that NSCs with impaired Dcx expression show reduced migration and delayed differentiation [54]. We found that, on the third day, the distribution of the Ku86 labeled NSCs (grafted) in the SVZ was very rare (Supplementary Figure 3).

Our evaluation of relative migration assessed the trend of relative distances in Dcx+ cell migration over a period of time (3 days), which can be considered to be the aggregate ability of all cells to independently migrate. Absolute migration measured the total number of Dcx+ cells participating in migration within a certain distance from the SVZ and contained the total quantity of cell resources homing in a unit area. This evaluation, based on the two dimensions of the individual and the whole, displayed consistent advantages in the treatment group. This study demonstrated that cells migrated further from the SVZ following hNSC administration and that more cells migrated and homed per unit area in the treatment group at all distances, suggesting that treatment with hNSCs promotes proliferation and accelerates the neurogenesis of SVZ cells, facilitating robust migration of these cells away from the SVZ and toward regions of injury. Proposed mechanisms of this facilitation include secretion of growth and other soluble factors released by transplanted cells, which promote neurogenesis and contribute to the evolution of NSCs after brain injury [55] and regulation of inflammation and the cellular microenvironment by transplanted cells [56]. We also propose that NSCs delivered by the ICV route bypass the blood-brain barrier (BBB) and, driven by the flow of cerebrospinal fluid, these transplanted hNSCs have direct interaction with SVZ cells. The potential mechanism proposed is shown in Fig 8.

Fig 8.

Possible mechanisms of hNSC transplantation by ICV on promoting endogenous SVZ NSC proliferation and migration and ultimately enhancing neuronal survival and neurological functional recovery. The grafted NSCs promoted the neurogenesis of SVZ cells through the complex “cell-cell crosstalk” dependent on Wnt / β-catenin signaling pathway in the whole SVZ cell environment.

We next propose the potential role of the Wnt/β-catenin pathway in the proliferation and migration of NSCs. Immunofluorescence staining exhibited increased Ki67 expression in the treatment group, indicating increased proliferation, and increased β-catenin expression in the treatment group, indicating the possible downstream role of the Wnt/β-catenin pathway in neural stem cell proliferation. Wnt signaling has been considered a critical factor in the plasticity of the nervous system by influencing fate, polarity, and migration of cells [57]. Exogenous expression of Wnt3a by lentivirus vectors enhanced the Wnt signaling pathway and promoted adult hippocampal neurogenesis [58]. In addition, β-catenin was expressed in a wider area of the SVZ compared to Ki67 (Fig 5. A), which indicated that β-catenin was upregulated not only in endogenous NSCs but also in the whole SVZ cell environment. This demonstrates the possibility that the interaction between grafted NSCs and endogenous NSCs may not be a simple “direct stimulation”, but rather may be achieved in a specific cell environment through a variety of cell-cell crosstalk dependent on the Wnt/β-catenin pathway. Stem cell-based immunomodulation has been demonstrated in ischemic stroke models [59] and remains a novel inquiry as to whether grafted NSCs create a more suitable environment for neurogenesis within the SVZ. These cellular level investigations are worthy of further exploration.

EEG monitoring and ABG results confirmed that there was no significant difference in baseline between two groups. After evaluation of the original EEG tracings prior to qEEG-IQ analysis, we discovered a greater proportion of bursts of abnormal waveforms with high amplitudes (1–1.8mV) lasting 2–6 seconds in the control group (Supplementary Figure 1). Evidence has shown that excessive glial-mediated immunity can cause inflammatory changes known to be epileptogenic [60], and some reports have demonstrated that microglia activated within several minutes after the ischemic attack produce multiple inflammatory cytokines, which may cause cortical irritability [61, 62]. We can reasonably propose that hNSC transplantation may have been protective against these epileptiform changes via modulation of post-CA acute inflammation. Evidence from similar brain ischemic models does support the anti-inflammatory effects of transplanted NSCs via inhibition of microglial activation and reduction of proinflammatory factors in the acute phase of stroke [63].

Although the immune rejection of xenotransplantation is a common concern, previous studies have shown that this effect is limited [26, 64]. It was reported that ReNcell VM cells grafted into the striatum of rats can be detected even four weeks after transplantation, without negative immune response [65]. Since we started treatment and evaluated the end point much earlier than the studies mentioned above, the effect of immune rejection in our study will be even smaller. Acute brain injury changes immune homeostasis and leads to immunosuppression [66], when cells were transplanted in the acute phase. Stem cells have been reported to regulate immunity in the treatment of ischemic brain injury [59], and the modulation effect may help to weaken immune rejection and benefit the survival of stem cells. Finally, the cell delivery route (ICV) can avoid direct contact with immune organs (such as the spleen, lymphatic system) caused by intravascular delivery, and may also be beneficial for minimizing immune rejection.

Apart from highlighting the molecular mechanisms and clinical outcome implications of this therapeutic strategy, our study also draws attention to delivery route for stem cell therapy after CA. We have recently reviewed the administration routes of stem cell therapy after ischemic brain injuries, including ICV, intraparenchymal, intravascular, and intranasal routes of stem cell delivery [8]. ICV administration holds the benefits of bypassing the blood brain barrier and allows for delivery to multiple sites of the CNS [67] via the cerebrospinal fluid. The ICV route has been used worldwide for decades, with clinical applications in various central nervous system (CNS) diseases [68], and the proximity of the ventricles to the SVZ makes it a particularly appealing route in stem cell applications and cardiac arrest therapies. However, this route of administration carries unique risks, given its invasive nature, which has raised concerns regarding translation into clinical practice. Evidence suggests that regular training and strict aseptic procedures greatly reduce the risk of complications [69]. It has been reported that the occurrence of ICV related complications is variable, and the variability in safety may be due to a lack of consensus on best practices of device use [70]. Practices for ICV device implantation were recommended, including: proper preoperative examination, prophylactic use of antibiotics 3 to 5 days in advance, establishment of a core team to perform the procedure, et cetera [70]. More importantly, clinical stem cell therapies using this administration route have been initiated in hemorrhagic stroke with noted safety and improved neurological outcomes [71, 72]. The ICV delivery in stem cell therapy for brain injury after CA deserves more future exploration.

Conclusion

ICV delivery of hNSCs promotes endogenous SVZ NSC proliferation and migration and ultimately enhances neuronal survival and neurological functional recovery. The Wnt/β-catenin pathway may be involved in the initiation and maintenance of this enhancement.

Supplementary Material

12015_2020_10067_MOESM1_ESM

Supplementary Fig 1 EEG Characteristics at 6h after ROSC. Number of topical bursts of abnormal waveforms with high amplitudes (1–1.8mV) lasting 2–6 seconds was found in the Control group.

12015_2020_10067_MOESM2_ESM

Supplementary Fig 2 Survival analysis at 72h. (A) Comparison of mean survival time showed a longer cumulative survival time in the treatment group, though without statistical significance (Control VS. NSC: 65.3 ± 3.2h VS. 68.3 ± 3.2h, P >0.05). (B) Kaplan Meier survival curve showed there was no difference in survival rates between groups (8/11, 72.7% for both groups).

12015_2020_10067_MOESM3_ESM

Supplementary Fig 3 Immunofluorescence staining with Ku86 in hippocampus and SVZ of the same rats. More Ku86 positive cells were found in hippocampus while no Ku86 positive cells were found in SVZ. 200X, Scale bar = 50μm

12015_2020_10067_MOESM4_ESM

Supplementary Fig 4 Cell proliferation in hippocampus. (A) As a typical ischemic targeting region, we selected CA2 for Ki67 (green) and Ku86 (red) double staining. The red arrow points to Ku86 positive cell. (B) The number of Ki67 positive cells in NSC group was significantly higher than in control group (*P < 0.05). 400X, Scale bar = 25μm

12015_2020_10067_MOESM5_ESM

Supplementary Fig 5 Cells migration in hippocampus. (A), (B) The co-staining of Dcx (green) and Ku86 (red) was performed in hippocampus, CA2, CA3 and DG. The red arrow points to Ku86 positive cell. (C) The density of Dcx positive cells in the hippocampus was lower than that in SVZ, but the comparison between two groups showed that Dcx signal was expressed in more cells in the hippocampus of NSCs treated animals (*P < 0.05). 400X, Scale bar = 25μm

12015_2020_10067_MOESM6_ESM

Supplementary Table 1 NDS assessment at 4-time points (Median and 25-75 interquartile)

Supplementary Table 2 sub-NDS assessment at 4-time points (Median and 25-75 interquartile)