Human neural stem cells improve early stage stroke outcome in delayed tissue plasminogen activator-treated aged stroke brains

Abstract

Introduction:

Clinically, significant stroke injury results from ischemia-reperfusion (IR), which induces a deleterious biphasic opening of the blood-brain barrier (BBB). Tissue plasminogen activator (tPA) remains the sole pharmacological agent to treat ischemic stroke. However, major limitations of tPA treatment include a narrow effective therapeutic window of 4.5 h in most patients after initial stroke onset and off-target non-thrombolytic effects (e.g., the risk of increased IR injury). We hypothesized that ameliorating BBB damage with exogenous human neural stem cells (hNSCs) would improve stroke outcome to a greater extent than treatment with delayed tPA alone in aged stroke mice.

Methods:

We employed middle cerebral artery occlusion to produce focal ischemia with subsequent reperfusion (MCAO/R) in aged mice and administered tPA at a delayed time point (6 h post-stroke) via tail vein. We transplanted hNSCs intracranially in the subacute phase of stroke (24 h post-stroke). We assessed the outcomes of hNSC transplantation on pathophysiological markers of stroke 48 h post-stroke (24 h post-transplant).

Results:

Delayed tPA treatment resulted in more extensive BBB damage and inflammation relative to MCAO controls. Notably, transplantation of hNSCs ameliorated delayed tPA-induced escalated stroke damage; decreased expression of proinflammatory factors (tumor necrosis factor-alpha (TNF-α) and interleukin (IL)-6), decreased the level of matrix metalloprotease-9 (MMP-9), increased the level of brain-derived neurotrophic factor (BDNF), and reduced BBB damage.

Conclusions:

Aged stroke mice that received delayed tPA treatment in combination with hNSC transplantation exhibited reduced stroke pathophysiology in comparison to non-transplanted stroke mice with delayed tPA. This suggests that hNSC transplantation may synergize with already existing stroke therapies to benefit a larger stroke patient population.

1. Introduction

Stroke often results in long-term neurological disability and is a leading cause of mortality worldwide. In the United States, data projections estimate that by 2030, an additional 3.4 million adults (3.88% of the adult population) will experience a stroke (Ovbiagele et al., 2013). This is an alarming 20.5% prevalence increase from the year 2012 (Benjamin et al., 2018). Most strokes are ischemic and caused by cerebral thrombosis. Thrombolysis in occluded cerebral arteries remains the standard treatment strategy to restore blood flow and rescue cells in the ischemic penumbra. However, fast reperfusion challenges the blood-brain barrier (BBB) with oxidative stress and disrupts cerebral microvascular endothelial cell tight junctions of the BBB (Lochhead et al., 2010; Yemisci et al., 2009). Reperfusion following ischemic stroke results in a biphasic opening of the BBB (Kuroiwa et al., 1985). The initial opening of the BBB is transient and occurs within several hours post-stroke (Shi et al., 2016), while the second opening of the BBB is permanent and occurs in the chronic stroke phase 24–72 h post-reperfusion (Yang and Rosenberg, 2011). The second opening of the BBB contributes greatly to brain cell death. For example, consequential loss of BBB integrity allows for extravasation of fluids, intravascular proteins, peripheral immune cells and potentially neurotoxic molecules into the brain parenchyma, thereby contributing to secondary neuronal injury following the initial ischemic insult. BBB disruption in ischemic stroke is complex and mediated by several factors, including upregulation of proinflammatory cytokines, increased activity of matrix metalloproteinases (MMPs), and disruption of tight junctions (Defilippi et al., 1992; Gasche et al., 1999a; Lochhead et al., 2010; Lu et al., 1998; Pettigrew et al., 2008; Rossi et al., 2011; Wolpe et al., 1988). Therefore, preservation of BBB integrity is an attractive therapeutic strategy for ischemic stroke.

Tissue plasminogen activator (tPA) remains the only U.S. Food and Drug Administration–approved clot-dissolving pharmacological agent to treat ischemic stroke. Unfortunately, effective administration of tPA has a narrow time window (FDA approval <3 h; ECASS-3 Trial <4.5 h) for the overall population (Roth, 2011) and the risk of increased ischemic/reperfusion (IR) injury (Sumii and Lo, 2002). Common ischemic stroke comorbidities, such as diabetes mellitus, increase the risk of BBB damage and poor outcome following intravenous thrombolytic therapy (Poppe et al., 2009). Consequently, many ischemic stroke patients do not receive tPA treatment. New therapeutic approaches are greatly needed in order to minimize off-target effects of tPA.

The Stroke Treatment Academic Industry Roundtable (STAIR) suggests that stroke therapies “focus on drugs/devices/treatments with multiple mechanisms of action and that target multiple pathways”. Neural stem cells (NSCs) are attractive candidates for stroke treatment because they exhibit multiple therapeutic capabilities. For instance, NSCs can participate in functional neural replacement in different central nervous system (CNS) regions (Yandava et al., 1999), or exhibit bystander effects in which NSCs synthesize therapeutic gene products that dampen inflammation, enhance endogenous repair mechanisms, or provide neurotrophic support to host brain tissue following injury (Jeyakumar et al., 2009; Lee et al., 2007; Ourednik et al., 2002; Park et al., 2002).

Clinically, stem cell therapy is currently aiming for stroke rehabilitation by transplanting stem cells during later recovery phases. However, stem cell therapy during the subacute stage may benefit more stroke patients by ameliorating early-phase stroke injury and subsequently reducing later complications of secondary stroke damage (Boese et al., 2018). The focus of this study is to assess pleiotropic actions of NSCs targeting early stroke injury.

Rodent models of transient focal ischemia induced by middle cerebral artery occlusion with subsequent reperfusion (MCAO/R) are well-established and widely used to study ischemic stroke (O’Neill and Clemens, 2001). In this study, we used a transient MCAO/R model as a proof-of-concept IR stroke model and administered tPA at a delayed time point (6 h post-injury) to induce severe BBB damage characteristic of tPA’s off-target (non-thrombolytic) actions beyond the 4.5 h therapeutic window. We transplanted well-characterized human NSCs (Huang et al., 2014) into the ipsilesional hippocampus 24 h after MCAO/R. We investigated the effects of human NSC transplantation on ischemic lesion volumes, expression of inflammatory mediators, and integrity of the BBB 48 h post-stroke. Of note, age is a major risk factor for stroke and correlates with worse prognosis (Saposnik et al., 2008). Given that age is associated with chronic inflammation (Franceschi and Campisi, 2014) and impaired BBB function (Montagne et al., 2015), we employed aged mice in our study to more appropriately model ischemic stroke and recapitulate the deleterious non-thrombolytic effects of tPA.

2. Materials and methods

2.1. Animals

50–52 week old C57BL/6 J male mice were acquired from Jackson Laboratory (Bay Harbor, ME, USA) and kept at 18–22°C and provided necessary food and water on a 12 h light-dark cycle.

2.2. Animal model of stroke

The Institutional Animal Care and Use Committee of Tulane University (New Orleans, LA, USA) reviewed and authorized all experimental procedures. Treatment and management of mice were executed in congruence with the principles of the American Veterinary Medical Association, Tulane University Protocol, and the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

We have chosen 50–52 week old mice (middle aged) because 1) delayed tPA administration results in very high mortality in male mice older than 68 weeks and 2) NOX2-dependent disruption of neurovascular coupling is already prominent by 50–52 weeks (Park et al., 2007). Using intraluminal MCAO, targeted cerebral ischemia was induced (Clark et al., 1997). After anesthesia was performed with 1% isoflurane in 30% oxygen, an incision was created along the midline of the neck to expose the left common carotid artery and external artery. Into the left internal artery, a silicon rubber-coated 6–0 nylon monofilament (Doccol Corporation, Sharon, MA, USA) was inserted to obstruct the origin of the MCA. Next, reperfusion was commenced by filament withdrawal 60 min post-occlusion. Control mice sustained a sham-operation, performed similarly but with the filament being removed immediately after insertion, thus effecting neither occlusion nor reperfusion. Animal temperature was constantly measured. Utilizing laser Doppler technology (Perimed, Stockholm, Sweden), regional cerebral blood flow (rCBF) was evaluated. MCAOs with a ≥ 80% diminution in rCBF following MCA occlusion with post-recovery rCBF to 90.1% ± 4.0 were considered successful.

2.3. tPA treatment

6 h following MCAO, mice were administered 10 mg/kg over 30 min of intravenous tPA (Genentech) or an equivalent amount of PBS (vehicle) (Zhang et al., 2002).

2.4. Human neural stem cell culture

We employed well-characterized hNSCs. The cells were sourced from primary hNSCs isolated from the telencephalon of late first-trimester human fetuses (Lee et al., 2007). An established medium was used to amplify and sustain the hNSCs and contained the following: human recombinant EGF (20 ng/mL, Invitrogen, Carlsbad, CA, USA), recombinant bFGF (20 ng/mL, Invitrogen), heparin (8 μg/mL, Sigma, St Louis, MO, USA), antibiotics (penicillin, amphotericin, and streptomycin), L-glutamine (Invitrogen), Leukemia Inhibitory Factor (LIF, 10 ng/mL, Millipore, Billerica, MA, USA), and DMEM/F12 high glucose (Invitrogen). Cells were subcultured once weekly via enzymatic digestion at a minimum of 48–72 h before transplantation. mRNA and protein expression of Sox2, nestin, doublecortin, GFAP, EGF receptor, and nucleostemin were confirmed during hNSC expansion using RT-PCR and immunostaining. Additionally, flow cytometry analysis of hNSCs confirmed high levels of NSC markers (e.g., CD133). The hNSCs involved in this study are proliferated solely with mitogens and have no transgene introduction or genetic manipulation. Previous research has demonstrated successful engraftment (xenograft) and integration of hNSCs into the mouse brain (Tennstaedt et al., 2015) and no incidents of tumorigenesis occurred.

2.5. Human neural stem cell transplantation

24 h post-MCAO/R, hNSCs were transplanted into the hippocampus, a neurogenic region that naturally hosts signals for stem cell migration. Following anesthesia with isoflurane, 50–52 week old C57BL/6 J mice were positioned in a stereotaxic frame (Stoelting, Co., Wood Dale, IL, USA). Ocular ointment was administered to the eyes in order to prevent drying. The head of each mouse was sterilized with 70% ethanol before being incised at the midline. A tiny burr hole through the skull (0.5-nm diameter, F.S.T) was created (2-mm posterior to bregma; 1.5-mm lateral to sagittal suture). Over a 3-min interval, 2 μL of PBS carrying ~100,000 viable cells were introduced into the ipsilesional hemisphere of the post-MCAO/R brain (depth; 2–2.5-mm dorsal). Control mice were administered an equal volume of PBS. The lesion was sealed using cyanoacrylate glue (Vector Laboratories, Burlingame, CA, USA), and brains were collected 48 h after MCAO/R for examination.

2.6. Quantification of infarct volume

48 h after MCAO/R, triphenyl tetrazolium chloride (TTC) staining of brain slices was performed. 1-mm coronal sections were cut and then incubated in 2% TTC solution (Sigma). In functional brain tissue, mitochondria transmute TTC into a red substance, in contrast to the colorlessness of the ischemia-damaged region. The NIH program ImageJ was employed to measure infarct size. The following equation was used to calculate infarct volume as a percent volume of the contralateral hemisphere in order to compensate for edema: [volume of contralateral hemisphere – (volume of total ipsilesional hemisphere – volume of infarct area)]/volume of contralateral hemisphere.

2.7. Reverse transcriptase polymerase chain reaction

Total RNA was taken from ipsilesional MCAO/R mouse brain tissue. The brain tissue (~20–30 mg), maintained in RNAlater (Ambion, Inc.), was submerged in Trizol reagent (Invitrogen) before being homogenized via MagNA Lyser (Roche Diagnostics, USA). RNA fractions were combined with proper amounts of 70% ethanol after chloroform extraction and then placed into a RNeasy spin column (Qiagen, Valencia, CA). DNase I (Qiagen) digested the remaining material. 2 μg of total RNA was reverse-transcribed with ABI high-capacity cDNA reverse transcriptase and random primers (4368814, Invitrogen) as a means of producing first-strand cDNA. 2 μL of reverse transcriptase reaction mixture in 10 μL of SsoFast Probes Supermix with Rox (BioRad, CA) was utilized to perform PCR amplification in a CFX96 Real-time thermal cylinder (BioRad). The PCR protocol included a 30-second hold at 95 °C as well as 35 cycles of activation for 5 s at 95 °C and annealing/extending for 10 s at 60 °C. TaqMan® Gene Expression Assays were utilized (TNF-α: Mm00443258_m1; IL-6: Mm00446190_m1; MMP-9: Mm00442991_m1; GAPDH: Mm99999915_g1). Ct values were then standardized comparatively to GAPDH. The Livak (2^−ΔΔCt) procedure was employed to calculate variations in specific gene expression compared to the control group.

2.8. Western blot analysis

48 h following MCAO/R, the ipsilesional brain regions were extracted after the mice underwent heavy anesthesia. Total protein was separated via homogenization in cold RIPA lysis buffer (Santa Cruz Biotechnology, Inc., Dallas, TX, USA). After 15 min of 12,000 rpm centrifugation, supernatants were harvested. 50 μg protein was loaded onto 4% to 12% Bis-Tris NuPAGE Novex gels (Invitrogen) and transferred onto PVDF membranes (Thermo Fisher Scientific Pierce, USA). Primary antibodies that were utilized included: β-actin (1:2500, Thermo Fisher Scientific), brain-derived neurotrophic factor (BDNF) (1:1000, Abcam), MMP-9 (1:1000, Abcam), and zona occluden-1 (ZO-1) (1:500, Invitrogen). Defined horseradish peroxidase-conjugated secondary antibodies (1:3000, Invitrogen) were then administered to the blots and detection was performed utilizing the ECL Plus Western Blotting Substrate (Thermo Fisher Scientific) and ImageQuant LAS 500 Imager (GE Healthcare Life Sciences, Piscataway, NJ, USA). ImageJ was used for evaluation of protein band intensities and quantification.

2.9. Gelatin zymography

As previously outlined (Hu and Beeton, 2010), SDS-PAGE zymography allowed for the detection of functional MMP-9. Brain samples were prepared for western blotting, however supernatants were not denatured before gel loading. Protein (40 μg/well) was loaded onto a Novex 10% gelatin zymogram gel (Invitrogen) and MMP-9 mouse recombinant protein was included as a positive gelatinase activity control (Invitrogen). The gel was then renatured with renaturing buffer (Invitrogen) following electrophoresis and was then incubated in developing buffer (Invitrogen) at 37 °C overnight. The gel next underwent staining of 0.5% SimplyBlue SafeStain (Invitrogen) for an hour, then was destained. Using the ChemiDoc™ MP Imaging System (Bio-Rad, Hercules, CA, USA), MMP-9 protease activity was visualized indirectly as clear bands against a dark background. Images were analyzed using ImageJ.

2.10. Statistical analysis

GraphPad Prism 6 and SPSS 19 software were utilized to execute statistical analyses. One-way ANOVA with Fisher’s LSD post-hoc test evaluated variances between numerous groups. If P-values were < 0.05, tests were deemed statistically significant. Data are presented as mean ± SEM.

3. Results

3.1. hNSC transplantation reduces infarct volume

We first assessed the effects of hNSC transplantation on infarct volume in aged stroke brains. We measured ischemic lesion volume 48 h post-MCAO/R (24 h post-engraftment) by staining brain sections with TTC (white, infarct; red, viable) (Fig. 1). In agreement with prior reports, 60 min of filamentous MCAO induced ischemic lesions in the cortex and caudoputamen (Nagasawa and Kogure, 1989). Compared to MCAO/R mice, the size of infarct tissue in transplanted MCAO/R brains that received delayed tPA (MCAO/R+tPA+Tx) was significantly smaller. The mean infarct volume (white area) of the ipsilesional hemisphere in MCAO/R brains was 46.05 ± 3.1% (***P < .001 vs. sham), while mean infarct volume in MCAO/R+tPA brains was 49.75 ± 2.4% (***P < .001 vs. sham) (Fig. 1A and B). In stroke mice that were engrafted with hNSCs (MCAO/R+tPA+Tx), the mean infarct volume was 33.80 ± 3.38% (***P < .001 vs. sham), which was significantly smaller than in the MCAO/R (^##P < .01 vs. MCAO/R) or MCAO/R+tPA group (^§§P < .01 vs. MCAO/R + tPA; Fig. 1A and B).

Figure 1. Human NSCs reduce infarct volume of delayed tPA-treated stroke mouse brains.

(A) Infarct volume was significantly reduced in hNSC-engrafted brains that received delayed tPA (MCAO/R+tPA+Tx). ***P < .001 vs. sham group; ^##P < .01 vs. MCAO/R group; ^§§P < .01 vs. MCAO/R + tPA group. (B) TTC staining (white, infarct). Shown are two representative samples of different mouse brains. (n = 8, Sham; n = 6, MCAO/R; n = 7, MCAO/R+tPA; n = 12, MCAO/R+tPA+Tx). Scale bar = 2 mm. Data are presented as mean ± SEM.

3.2. hNSC transplantation reduces gene expression of pro-inflammatory cytokines and MMP-9

IR injury in the CNS triggers the release of inflammatory cytokines and chemokines, and contributes to further tissue damage through BBB disruption and extravasation of peripheral immune cells into the brain (Clausen et al., 2008; Inose et al., 2015; Jin et al., 2010; Pan et al., 2011; Rochfort and Cummins, 2015). We assessed whether exogenous hNSCs were able to influence proinflammatory gene expression at 48 h post-MCAO/R using RT-PCR (Fig. 2A). mRNA expression of inflammatory mediators, TNF-α and IL-6, was significantly elevated in the ipsilesional tissue of MCAO/R brains. Specifically, the mean expression levels of TNF-α and IL-6 in MCAO/R brains were 46.15 ± 4.0 and 29.02 ± 8.3 fold (***P < .001 vs. sham), respectively, while mean TNF-α and IL-6 mRNA expression levels in non-transplanted MCAO/R+tPA brains were 51.28 ± 2.57 and 34.01 ± 6.5 fold, (***P < .001 vs. sham), respectively. By contrast, hNSC transplantation (MCAO/R+tPA+Tx) significantly reduced mRNA expression of TNF-α and IL-6 (29.08 ± 5.2 and 9.99 ± 4.8 fold, respectively) when compared to both MCAO/R (^#P < .05, ^##P < .01 vs. MCAO/R, respectively) and non-transplanted MCAO/R+tPA brains (^§P < .05, ^§§P < .01 vs. MCAO/R+tPA, respectively; Fig. 2A).

Figure 2. Human NSCs reduce expression of MMP-9 and proinflammatory genes.

(A) Diminished expression of inflammatory markers within the ipsilesional hemisphere of transplanted MCAO/R+tPA+Tx brains. RT-PCR was utilized to evaluate inflammatory gene expression, then normalized to that of GAPDH. Transcript levels of proinflammatory cytokines (TNF-α and IL-6) are upregulated in MCAO/R and MCAO/R+tPA brains; in transplanted MCAO/R+tPA+Tx brains, however, these levels are downregulated in comparison to the MCAO/R+tPA group. ***P < .001 vs. sham; ^#P < .05, ^##P < .01 vs. MCAO/R; ^§P < .05, ^§§P < .01 vs. MCAO/R+tPA group (n = 5). (B) In MCAO/R and non-transplanted MCAO/R+tPA brains, MMP-9 is significantly upregulated. When hNSCs are transplanted (MCAO/R+tPA+Tx), however, MMP-9 is significantly reduced. ***P < .001 vs. sham; ^#P < .05 vs. MCAO/R; ^§P < .05 vs. MCAO/R+tPA group (n = 5). Data are presented as mean ± SEM.

Following stroke, MMP levels are elevated in ischemic tissue as they begin to degrade tight junction proteins that are crucial for the BBB (Bauer et al., 2010; Romanic et al., 1998). Accordingly, inhibition of MMPs is reported to reduce infarct size (Romanic et al., 1998; Yang et al., 2013). Since MMP-9 is upregulated by tPA and associated with BBB breakdown after stroke (Asahi et al., 2001; Gasche et al., 1999b; Jin et al., 2015), we assessed whether treatment with hNSCs could reduce MMP-9 expression in delayed tPA treatment in aged stroke brains. mRNA expression of MMP-9 was significantly elevated in the ipsilesional MCAO/R brains (3.62 ± 0.7 fold, ***P < .001 vs. sham; Fig. 2B) and MCAO/R+tPA brains (4.28 ± 0.5 fold, ***P < .001 vs. sham; Fig. 2B). By contrast, hNSC transplantation (MCAO/R+tPA+Tx) significantly reduced mRNA expression of MMP-9 (2.20 ± 0.3 fold) when compared to both MCAO/R (^#P < .05 vs. MCAO/R; Fig. 2B) and non-transplanted MCAO/R+tPA brains (^§P < .05 vs. MCAO/R+tPA; Fig. 2B).

3.3. hNSC transplantation increases brain-derived neurotrophic factor expression

Next, we assessed the relative levels of the neuroprotective protein, BDNF between treatment groups to test whether the beneficial effect of hNSCs could also be attributed to the secretion of neuroprotective and anti-apoptotic gene products. BDNF is a major neurotrophin that promotes the survival, differentiation and synapse formation of neurons in the CNS, thereby contributing to neuroprotection and recovery in the context of ischemic stroke (Beck et al., 1994; Ploughman et al., 2009). Previously, we have shown that transplanted hNSCs secrete BDNF in mouse stroke brains (Huang et al., 2014). Through western blotting, we found reduced BDNF levels in aged MCAO/R+tPA brains compared to sham controls at 48 h post-injury (****P < .0001 vs. sham; Fig. 3A and B). However, transplantation of hNSCs significantly increased BDNF expression in engrafted MCAO/R+tPA (MCAO/R+tPA+Tx) brains (^§§P < .01 vs. MCAO/R+tPA; Fig. 3A and B).

Figure 3. hNSC transplantation increases brain-derived neurotrophic factor level.

(A) BDNF levels are shown via western blotting in the ipsilesional hemisphere of sham, MCAO/R, MCAO/R+tPA, and MCAO/R+tPA+Tx mice. (B) Quantification of A (n = 4, ****P < .0001, **P < .01 vs. sham group; ^§§P < .01 vs. MCAO/R+tPA group). Data are expressed as mean ± SEM.

3.4. hNSC transplantation ameliorates BBB damage

Biphasic opening of the BBB from IR injury allows for the extravasation of foreign intravascular proteins into the brain parenchyma. In order to assess the effects of hNSC transplantation on BBB integrity in aged brains, we quantified the amount of blood-derived IgG present in the parenchyma 48 h post-MCAO/R. Western blot analysis revealed significantly higher amounts of IgG in the ipsilesional brain of MCAO/R (**P < .01 vs. sham; Fig. 4A and B) and MCAO/R+tPA (****P < .0001 vs. sham; ^##P < .01 vs. MCAO/R; Fig. 4A and B) brains, with MCAO/R+tPA brains containing the highest amount of blood-derived IgG. Notably, we observed significantly reduced levels of IgG in the parenchyma of hNSC-transplanted (MCAO/R+tPA+Tx) brains in comparison to non-transplanted brains (^§§§§P < .0001 vs. MCAO/R+tPA; ^##P < .01 vs. MCAO/R; Fig. 4A and B), suggesting that hNSCs preserve BBB integrity in the subacute phase of stroke after delayed tPA treatment.

Figure 4. BBB leakage is reduced by hNSC transplantation.

(A) IgG levels are displayed via western blotting in the ipsilesional hemispheres of sham, MCAO/R, MCAO/R+tPA, and MCAO/R+tPA+Tx brains. (B) Quantification of A (n = 4, **P < .01, ****P < .0001 vs. sham group; ^##P < .01 vs. MCAO/R group; ^§§§§P < .0001 vs. MCAO/R+tPA group). (C) Western blot evaluation of the MMP-9 protein level in the ipsilesional hemisphere. (D) Quantification of C (n = 4, *P < .05, ****P < .0001 vs. sham group; ^#P < .05 vs. MCAO/R group; ^§§§§P < .0001 vs. MCAO/R+tPA group). (E) Zymography assay displays MMP-9 activity in the ipsilesional hemisphere. (F) Quantification of E (n = 4, ^##P < .01 vs. MCAO/R group; ^§§§§P < .0001 vs. MCAO/R+tPA group). (G) Western blot evaluation of ZO-1. (H) Quantification of G (n = 4, *P < .05, ****P < .0001 vs. sham group; ^#P < .05 vs. MCAO/R group; ^§P < .05 vs. MCAO/R+tPA group). Data are expressed as mean ± SEM.

In order to explore the mechanism by which transplanted hNSCs protect BBB integrity, we next examined the level of MMP-9, which contributes to BBB breakdown in ischemic stroke. Western blotting confirmed elevated MMP-9 levels at 48 h post-injury in the MCAO/R group in comparison to sham controls (*P< .05 vs. sham; Fig. 4C and D), while MMP-9 levels were the highest in the MCAO/R+tPA group (****P< .0001 vs. sham; ^#P < .05 vs. MCAO/R; Fig. 4C and D). However, engraftment of hNSCs after delayed tPA treatment significantly reduced MMP-9 levels (^§§§§P< .0001 vs. MCAO/R+tPA; ^#P < .05 vs. MCAO/R; Fig. 4C and D). To further confirm these observations, we used gel zymography to measure MMP-9 enzyme activity on gelatin substrate. At 48 h post-injury, we found significantly increased MMP-9 activity in both MCAO/R and non-transplanted MCAO/R+tPA groups compared to sham controls, with MCAO/R+tPA brains showing the highest activity (^##P < .01 vs. MCAO/R; Fig. 4E and F). Consistent with our western blot analysis, transplantation of hNSCs following delayed tPA treatment significantly reduced MMP-9 activity as measured by gel zymography compared to the non-transplanted MCAO/R+tPA group (^§§§§P< .0001 vs. MCAO/R+tPA; ^##P < .01 vs. MCAO/R; Fig. 4E and F). Our results suggest that transplantation of hNSCs following ischemic stroke may protect BBB integrity by reducing MMP-9 levels and activity.

Tight junctions are a crucial component of the BBB as they restrict paracellular molecular trafficking between the CNS and the circulatory system (Bauer and Traweger, 2016). It is noteworthy that BBB leakage has been reported in vivo at 24 h post-MCAO without disruption of claudin-5 and occludin (Kuntz et al., 2014). However, ZO-1, a tight junction protein that connects occludin to the actin cytoskeleton (Fanning et al., 1998), is known to be degraded by MMP-9 (Asahi et al., 2001; Bauer et al., 2010). Given that hNSC transplantation reduced the MMP-9 level and activity after delayed tPA treatment (Fig. 4C – F), we assessed the effect of hNSC transplantation on ZO-1 levels in aged stroke brains. Western blot analysis showed decreased levels of ZO-1 in both MCAO/R (*P < .05 vs. sham; Fig. 4G and H) and MCAO/R+tPA (****P < .0001 vs. sham; ^#P < .05 vs. MCAO/R; Fig 4G and H) brains. However, transplantation of hNSCs after delayed tPA treatment ameliorated tPA-associated ZO-1 degradation (^§P < .05 vs. MCAO/R+tPA; Fig. 4G and H). Our results imply that hNSCs preserve BBB integrity in delayed tPA-treated stroke brains through downregulation of proinflammatory factors, MMP-9, and reducing proteolytic cleavage of ZO-1. These data support the hypothesis that NSCs act through multiple mechanisms to improve stroke pathophysiology.

4. Discussion

Although tPA’s thrombolytic actions within cerebral blood vessels have proven to be beneficial following stroke, the narrow therapeutic window and off-target effects on the BBB and brain parenchyma severely limit tPA’s therapeutic potential (Su et al., 2008; Wang et al., 2004; Wang et al., 1998). Consequently, few stroke patients benefit from tPA treatment (Marler and Goldstein, 2003) and tPA by itself cannot completely repair neurological damage from ischemic stroke. Therefore, extensive research is needed to discover new therapeutic agents that synergize with tPA and minimize its off-target effects on the BBB and brain parenchyma.

Filamentous tMCAO is an established method to model ischemic stroke (Engel et al., 2011), but baseline physiological characteristics of animal subjects also need to be considered in order to closely mirror stroke pathophysiology as it occurs in humans. Aging increases the risk of ischemic stroke and influences both inflammation and BBB function (Franceschi and Campisi, 2014; Montagne et al., 2015). While age alone does not factor into clinical tPA eligibility within the 4.5 h therapeutic window (Derex, 2010; Mishra et al., 2010), the effect of delayed tPA (>4.5 h) treatment on aged stroke brains has not been well-studied. Therefore, we chose to use aged mice to test whether hNSC transplantation in the subacute phase of stroke could ameliorate tPA-associated BBB damage.

The pathophysiology of ischemic stroke is complex and mediated by many factors. Accordingly, STAIR recommends that future stroke therapies focus on targeting multiple pathways. NSCs participate in CNS homeostasis and repair through several mechanisms, which makes exogenous NSC transplantation an attractive therapeutic strategy for ischemic stroke. In order to maximize their paracrine therapeutic effects during the subacute phase, exogenous NSCs must quickly navigate to the ischemic lesion. We have previously demonstrated that the ipsilesional hippocampus is an ideal site for transplantation because the hippocampus is a natural neurogenic niche that already contains a plethora of signals for NSC proliferation, differentiation and migration, even under normal physiological conditions (Boldrini et al., 2018; Eckert et al., 2015; Huang et al., 2014). Other studies have transplanted exogenous NSCs into CNS regions other than the hippocampus post-ischemia, (Darsalia et al., 2011; Kelly et al., 2004) but this results in limited migration of NSCs to the site of injury. Here, 24 h post-transplant (48 h post-MCAO/R), we assessed the effects of hNSC engraftment in aged mice that underwent MCAO/R and delayed tPA treatment. We found that hNSC-engrafted brains showed reduced infarct volume, increases an enhanced therapeutic BDNF level, reduced hallmarks of tPA-exacerbated IR injury and neuroinflammation.

The BBB plays important roles in maintenance of brain cell health and proper signaling and forms a physical barrier that restricts trafficking of cells and molecules between the circulatory system and the CNS. IR injury leads to an initial opening of the BBB, which, if left unchecked, can result in a permanent second opening of the BBB 24–72 h after stroke onset (Kuroiwa et al., 1985). The second opening of the BBB contributes greatly to neuronal injury and is mediated by degradation of endothelial tight junctions through oxidative stress, upregulation of MMP activity, and inflammation (Lochhead et al., 2010; Pan et al., 2011; Rochfort and Cummins, 2015). While post-stroke inflammation contributes to the progression of BBB damage, it is noteworthy that neuroinflammation attracts transplanted NSCs to the infarct lesion (Andres et al., 2011; Imitola et al., 2004; Ohira et al., 2010). Previous work has demonstrated that neuroinflammation after stroke attracts NSCs to the infarct lesion (Imitola et al., 2004), but NSCs display limited migration when transplanted within 6 h or 12 h post-MCAO/R (Huang et al., 2014). Therefore, we chose to engraft exogenous hNSCs 24 h post-MCAO/R at the time of abundant inflammation.

We previously reported that engrafted hNSCs exhibit anti-inflammatory actions in the subacute phase of stroke (Eckert et al., 2015; Huang et al., 2014). In this study, we demonstrate that hNSCs dampen tPA-exacerbated inflammation and BBB damage in aged stroke brains. Proinflammatory cytokines are upregulated in the ischemic brain and have been shown to potentiate BBB disruption (Pan et al., 2011; Rochfort and Cummins, 2015). Therefore, downregulation of proinflammatory mediators could be an effective therapy to mitigate BBB breakdown and additional IR injury. Indeed, we demonstrated in this study that engraftment of hNSCs dampened the expression of pro-inflammatory cytokines (TNF-α and IL-6) induced by tPA-exacerbated IR injury.

In addition to dampening inflammation, we demonstrated that hNSCs were able to preserve BBB integrity in aged brains during the subacute phase of ischemic stroke, as indicated by significantly reduced blood-derived IgG leakage into the brain parenchyma. Previous studies have reported that MMP-9 activity is positively correlated with BBB breakdown after stroke (Jin et al., 2015) and linked to tPA-induced hemorrhage in both stroke patients (Castellanos et al., 2003) and animal models (Mishiro et al., 2012; Wang et al., 2003). In this study we showed that hNSCs significantly reduced the MMP-9 level and overall MMP-9 activity in the ipsilesional hemisphere in aged stroke brains treated with delayed tPA. This suggests that hNSCs may protect BBB tight junction proteins from proteolytic cleavage during the subacute phase of stroke.

Tight junction proteins, such as claudin-5, occludin, and ZO-1 are crucial for BBB integrity (Fanning et al., 1998; Jiao et al., 2011). Previous research has demonstrated that at least one of the mechanisms by which MMP-9 participates in BBB damage is through the degradation of specific tight junction proteins (Rosell and Lo, 2008). Of note, BBB leakage begins to occur 24 h post-MCAO/R, regardless of claudin-5 or occludin disruption (Kuntz et al., 2014). However, MCAO/R does result in the destruction of ZO-1, a protein that is important for connecting the actin cytoskeleton of microvascular endothelial cells to occludin (Fanning et al., 1998). Given that MMP-9 is known to degrade ZO-1 (Asahi et al., 2001; Bauer et al., 2010), this suggested that hNSC-mediated downregulation of MMP-9 may ameliorate proteolytic degradation of ZO-1 and preserve BBB integrity. To test this mechanism in our tMCAO/R mouse model, we quantified the relative amount of ZO-1 and found hNSC-engrafted stroke brains showed increased amounts of ZO-1 compared to brains treated with delayed tPA alone. This highlights a possible role of hNSCs in ameliorating proteolytic degradation of ZO-1 in the early phase of stroke. Thus, adjuvant therapy with hNSCs following tPA treatment may be an effective strategy to retain thrombolytic activity and reduce BBB damage in the subacute phase of stroke.

Since NSCs are known to secrete neurotrophins to support survival, differentiation, and synapse formation in their progeny (Lu et al., 2003), we assessed whether the beneficial effect of hNSCs could partially be explained by provision of neurotrophic factors by hNSCs. We demonstrated in this study that ipsilesional brains transplanted with hNSCs contain significantly higher amounts of BDNF. Still, further studies using genetically modified NSCs that overexpress or suppress BDNF are needed in order to determine if BDNF is indispensable for hNSC therapy in the subacute phase of stroke.

The mammalian CNS has a limited repair capacity, and endogenous NSCs are known to participate in angiogenesis and neurogenesis following stroke. Although spontaneous stroke recovery can occur, endogenous neurogenesis and angiogenesis are often insufficient to fully repair brain damage caused by ischemic stroke. Thus, our study addresses an important need for an effective stroke treatment by early subacute administration of exogenous hNSCs post-tPA treatment for potentially improving long-term stroke outcome. Previously, we have reported extensive migration of hNSCs 24 h post-transplant and improved neurological outcomes at 28 days post-MCAO/R (Huang et al., 2014). One defining criterion of NSCs is their ability to generate neurons, glia, and oligodendrocytes or self-renew. We anticipate that engrafted hNSCs in delayed tPA-treated brains will differentiate into multiple types of proper neural lineages in the brain and increase neurogenesis. Furthermore, NSCs have been shown to improve angiogenesis via trophic support such as paracrine stimulation with vascular endothelial growth factor (VEGF) and can influence capillary blood flow in the CNS (Lacar et al., 2012; Teng et al., 2008). Although it is beyond the scope of the current study, we anticipate that hNSCs in tPA-treated brains will increase vascular density. This creates a unique opportunity for future studies to examine whether the long-term actions of exogenous NSCs are affected by delayed tPA treatment.

Multiple clinical trials have investigated whether the therapeutic window for tPA can be extended to benefit a greater number of stroke patients. For instance, the EPITHET trial reported that, when administered 3–6 h after stroke onset, intravenous tPA was associated with reduced infarct growth and improved clinical outcomes (Davis et al., 2008). The EXTEND trial and WAKE-UP trial have reported that a subpopulation of patients who had ischemic stroke but salvageable brain tissue benefitted from receiving tPA up to 9 h after stroke onset (Ma et al., 2019; Thomalla et al., 2018). In addition, a later meta-analysis that examined data from the EXTEND, ECASS4-EXTEND, and EPITHET stroke clinical trials also came to the conclusion that treatment with tPA beyond the 4.5 h window and up to 9 h after stroke onset resulted in better functional outcomes at 3 months compared to placebo (adjusted odds ratio [OR] 1.86, 95% CI 1.15–2.99, p = .011) (Campbell et al., 2019). Although the rate of intracerebral hemorrhage was higher in patients treated with tPA, this increase did not negate the overall net benefit of thrombolysis past the 4.5 h window. Other retrospective studies have come to similar conclusions. A post hoc case-control study that analyzed the pooled EPITHET-DEFUSE combined dataset reported that patients receiving tPA in the 3–6 h time window following stroke had significantly attenuated infarct growth and an increased perfusion rate compared with placebo (Ogata et al., 2013).

Several challenges exist for establishing safety and efficacy of stem cell-based therapies for ischemic stroke. Extensive research need to be conducted in order to determine the optimal cell type, dosage, route of administration, and pathological stage of delivery (Kenmuir and Wechsler, 2017; Lees et al., 2012). Stroke severity, infarct size, and location are also likely important factors that should be considered since cell type and delivery method may affect homing ability of the cells to the site of injury. Regarding clinical trials of hNSCs in stroke patients, the safety and efficacy of hNSC transplantation for the treatment of ischemic stroke have been evaluated. For instance, a phase 1 clinical trial on men ages 60 years or older with stable disability 6–60 months after ischemic stroke demonstrated that direct intracranial injection of CTX-DP immortalized hNSCs did not result in any adverse events and was associated with improved neurological function at 2 years post-treatment (Kalladka et al., 2016). Results from our study further support the notion that tPA may be used for stroke treatment past the 4.5 h window in combination with other therapeutic agents such as stem cells. Our findings showed that a combination of tPA with hNSC transplantation reduced inflammation and BBB damage associated with delayed tPA treatment.

Extending the therapeutic time window for tPA can be achieved by combining a pharmacological or cell-based approach as demonstrated here. Of note, prior treatment with neuroprotectants before tPA, such as minocycline or cyclosporine A (CsA), may synergize with hNSC engraftment to greatly improve outcome. For example, minocycline has been shown to enhance neurorestoration following stroke through various mechanisms, such as protecting BBB integrity and reducing inflammation (Lampl et al., 2007; Machado et al., 2006; Yang et al., 2015a; Yang et al., 2015b). CsA is clinically proven to be safe in patients (Nighoghossian et al., 2015; Piot et al., 2008), and a neuroprotective role has also been reported against acute IR injury in animal stroke models (Borlongan et al., 2005; Cho et al., 2013; Domanska-Janik et al., 2004; Griffiths and Halestrap, 1993; Matsumoto et al., 2002; Muramatsu et al., 2007; Murozono et al., 2004; Osman et al., 2011; Tuladhar et al., 2015; Uchino et al., 1998; Uchino et al., 2002; Vachon et al., 2002; Yoshimoto and Siesjo, 1999; Yoshimoto et al., 2001; Yu et al., 2004; Yuen et al., 2011). Therefore, it is possible that extending the therapeutic time window and efficacy for tPA can be achieved by combining neuroprotection with stem cells. In addition, brain imaging techniques can further assist in selecting patients with salvageable tissue who can be candidates for stroke combination therapies.

In summary, NSCs are gaining recognition for therapeutic mechanisms other than direct cell replacement in the context of neurodegenerative and neurovascular diseases. Indeed, our study supports the notion that NSCs participate greatly in CNS health and repair through BBB support, downregulation of neuroinflammation, and provision of neurotrophic factors (Goldberg et al., 2015; Huang et al., 2014; Lee et al., 2007). These bystander effects of NSCs, as we have shown in this study, may be equally or more important than direct cell replacement for improving stroke outcome.

5. Conclusions

Our study demonstrates that transplantation of hNSCs into the hippocampus 24 h post-IR protects the BBB and reduces infarct volume in aged brains that receive delayed tPA administration. hNSC transplantation ameliorated ischemic-reperfusion cerebral injury by downregulating proinflammatory cytokines that are upregulated post-stroke. Our findings demonstrate intrinsic properties of hNSCs that make them attractive candidates for future cell-based treatment of IR injury.

Abbreviations

BBB

blood-brain barrier

CNS

central nervous system

CsA

cyclosporine A

DEFUSE-3

Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke

DMEM

Dulbecco’s Modified Eagle’s Medium

ECASS-4

European Cooperative Acute Stroke Study 4

EGF

epidermal growth factor

EPITHET

Echoplanar Imaging Thrombolytic Evaluation Trial

EXTEND

Extending the Time for Thrombolysis in Emergency Neurological Deficits

FGF

fibroblast growth factor

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

GFAP

glial fibrillary acidic protein

hNSCs

human neural stem cells

IL-6

interleukin-6

LIF

leukemia inhibitory factor

MCAO/R

middle cerebral artery occlusion/reperfusion

MMP

matrix metalloproteinase

PBS

phosphate buffered saline

PFA

paraformaldehyde

rCBF

regional cerebral blood flow

RT-PCR

reverse transcription polymerase chain reaction

TNF-α

tumor necrosis factor-alpha

tPA

tissue plasminogen activator

TTC

triphenyl tetrazolium chloride

VEGF

vascular endothelial growth factor

WAKE-UP trial

Efficacy and Safety of MRI-Based Thrombolysis in Wake-Up Stroke