Translating intracarotid artery transplantation of bone marrow‐derived NCS‐01 cells for ischemic stroke: Behavioral and histological readouts and mechanistic insights into stem cell therapy

Abstract The present study used in vitro and in vivo stroke models to demonstrate the safety, efficacy, and mechanism of action of adult human bone marrow‐derived NCS‐01 cells. Coculture with NCS‐01 cells protected primary rat cortical cells or human neural progenitor cells from oxygen glucose deprivation. Adult rats that were subjected to middle cerebral artery occlusion, transiently or permanently, and subsequently received intracarotid artery or intravenous transplants of NCS‐01 cells displayed dose‐dependent improvements in motor and neurological behaviors, and reductions in infarct area and peri‐infarct cell loss, much better than intravenous administration. The optimal dose was 7.5 × 106 cells/mL when delivered via the intracarotid artery within 3 days poststroke, although therapeutic effects persisted even when administered at 1 week after stroke. Compared with other mesenchymal stem cells, NCS‐01 cells ameliorated both the structural and functional deficits after stroke through a broad therapeutic window. NCS‐01 cells secreted therapeutic molecules, such as basic fibroblast growth factor and interleukin‐6, but equally importantly we observed for the first time the formation of filopodia by NCS‐01 cells under stroke conditions, characterized by cadherin‐positive processes extending from the stem cells toward the ischemic cells. Collectively, the present efficacy readouts and the novel filopodia‐mediated mechanism of action provide solid lab‐to‐clinic evidence supporting the use of NCS‐01 cells for treatment of stroke in the clinical setting.


| INTRODUCTION
Stroke remains as one of the most prevalent causes of disability and death among adult populations around the world, 1 significantly costing the United States billions of dollars each year. 2 Tissue plasminogen activator (tPA) is the sole FDA-approved drug to treat acute ischemic stroke, which accounts for roughly 87% of all strokes. 3,4 tPA is most effective when administered intravenously (IV) within 4.5 hours of stroke onset, 4 but is toxic outside this therapeutic window, causing hemorrhagic transformation. 4 Mechanical thrombectomy serves as an alternative treatment for ischemic stroke, but it too encounters challenges such as a limited therapeutic window (6-24 hours post stroke), 5 bleeding, coagulation abnormalities, and intracranial hemorrhage. 6 Since most stroke patients do not have access to tPA therapy or qualify for mechanical thrombectomy within the limited therapeutic windows, novel treatments are warranted. Cell-based regenerative medicine has emerged as a safe and effective experimental treatment for stroke and has reached clinical trials. The central nervous system has long been considered as incapable of regeneration. Stem cell research has challenged this paradigm with compelling evidence of exogenous and endogenous repair processes. 7 Transplantation of embryonic, fetal, umbilical, amnion, and induced pluripotent stem cells shows functional improvements in experimental stroke. 8,9 Adult bone marrow-derived stem cells, such as endothelial progenitor cells, hematopoietic, and mesenchymal stem cells (MSCs), have expedited the translation of lab-to-clinic stem cell therapy due to their logistical ease in isolation and amplification, and being relatively free from ethical concerns. 10,11 MSCs have been explored as transplantable donor cells for many experimental models of neurological diseases, 12,13 such as Parkinson's disease, [14][15][16] amyotrophic lateral sclerosis, [17][18][19] Alzheimer's disease, 20,21 and stroke. [22][23][24][25][26][27][28] Cell replacement was initially implicated in MSCs' therapeutic effects, yet with only modest graft survival despite robust functional outcomes, 29 the currently accepted mechanism of action involves bystander repair processes primarily via stem cell-secreted therapeutic factors. [30][31][32][33] This promising preclinical evidence, as well as solid safety record in treating hematological diseases, provides concrete grounds for clinical translation of MSC therapy in stroke. However, two clinical trials examining MSCs upheld their safety, but did not reveal efficacy. 34,35 IV transplantation of autologous bone marrow MSCs at 4 weeks following stroke reputedly demonstrated enhanced neurological outcomes, yet these functional improvements diminished by 12 months after transplantation. 34 An inconsistent adherence may have contributed to strict doses and therapeutic time windows. For example, preclinical investigations employ approximately 4 million cells for IV administration for a 250 g stroke rat, corresponding to 840 million cells in a human of 75 kg, whereas clinical trials have employed doses well below those deemed optimal in preclinical animal models. 34,35 This lack of translation of optimal laboratory parameters may explain these failed clinical efforts despite overwhelming experimental evidence. 23 Recognizing these critical translational gaps, while acknowledging the long-standing safety profile and the abundance of preclinical studies demonstrating neuroprotective effects, provides a solid rationale for testing the efficacy of an MSC line toward advancing stem cell therapy in stroke.
In identifying transplantable bone marrow-derived MSCs for clinical application, we used lab-to-clinic translational research criteria, namely the cells need to be of human origin, clinical grade, ample supply, and with well-defined phenotypic markers. To this end, the adult

| METHODS
The data that support the findings of this study are available from the corresponding author upon reasonable request. This series of highly translational studies utilized NCS-01 cells, which are human bone marrow-derived MSCs produced by Progenitor Cell Therapy

Significance statement
The present study recognizes critical translational gaps in stem cell transplant dose, route, and timing after stroke, and acknowledges solid safety profile of mesenchymal stem cells. The study tested a human bone marrow-derived mesenchymal stem cell line called NCS-01 in oxygen glucose deprivation and middle cerebral artery occlusion models, which revealed the optimal dose of 7.5 × 106 cells/mL via the intracarotid artery within 3 days poststroke. Secretion of cytokines, specifically bFGF and IL-6, and filopodia formation, are potential mechanisms. Based on these preclinical data, the FDA in July 2019 approved intracarotid NCS-01 cell transplantation in ischemic stroke patients.
(Mountain View, California). The release criteria included phenotypic characterization of these cells via fluorescence-activated cell sorting, indicating that these cells are CD105+, CD73+, CD90+, CD34−, CD45−, and CD14−. Subsequent release criterion added the capacity of these cells to secrete high amounts of basic fibroblast growth factor (bFGF) and interleukin-6 (IL-6). Cell viability of NCS-01 cells was confirmed at least >85% prior to starting each experiment. Engraftment was also confirmed for each transplant experiment and revealed modest graft survival, with graft persistence almost nondetectable by day 3 post-transplantation (Supplemental Figure S1). No ectopic tissue or tumor formation was detected in any study. Maryland) to better mimic the human condition. Immediately after thawing, cells (4 × 10 4 cells/well) were seeded and grown in 96-well plate coated by poly-L-lysine in Neurobasal media (GIBCO, California) containing 2 mM L-glutamine, 2% B27 (GIBCO) and 50 U/mL penicillin and streptomycin for 7-10 days at 37 C in humidified atmosphere containing 5% CO 2 . To further probe the mechanism of action of NCS-01 cells, a vis-a-vis in vitro efficacy test was performed comparing the targeted cell viability and NCS-01 cells' filopodia formation when cocultured with a specific neural cell lineage, including primary rat cortical neurons (CAMBREX), primary rat astrocytes (Fisher), and primary rat endothelial progenitor cells (EPCs) (Harvard, Massachusetts). Neurons were seeded and grown in culture as above. Astrocytes were seeded and grown in high glucose DMEM (Fisher), 10% fetal bovine serum (Neuromics), and 1% penicillin and streptomycin.

| OGD model
The cultured primary rat cortical cells or human neural progenitor cells were exposed to OGD as described previously 36  Cultured cells were placed in humidified chamber, and then equilibrated with continuous flow of 95% N 2 and 5% CO 2 gas for 15 minutes. After this equilibrium, the chamber was sealed and placed into the incubator at 37 C for 90 minutes. Thereafter, OGD was terminated by adding glucose to the culture medium and returning the cultures to the standard 20% O 2 and 5% CO 2 incubator. A 1-or 2-hour period of "reperfusion" in standard medium and normoxic condition was allowed.

| Cell survival
The viability of primary rat cortical cells or human neural progenitor cells was evaluated immediately after OGD (no NCS-01 cells added) and at 5 hours after OGD using standard 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) and trypan blue exclusion assays. Following reperfusion, reduction of MTT by cellular dehydrogenases was used as a measure of mitochondrial activity as previously described. 37 In addition, trypan blue (0.2%) exclusion method was conducted and mean viable cell counts were calculated in four randomly selected areas (1 mm 2 , n = 10) to reveal the cell viability after the ischemic-reperfusion condition. Briefly, within 5 min after adding trypan blue, we digitally captured under microscope (×200) 10 pictures

| Cytokine analysis
Cytokines were measured in the supernatant of cultured cells at 5 hours after OGD using commercially available ELISA kits to detect initially human brain derived neurotrophic factor (BDNF), beta-nerve growth factor (β-NGF), IGF-1, VEGF, bFGF, and IL-6, but subsequently just focused on bFGF, and IL-6, which were chosen based on our pilot studies demonstrating that the expression levels of these two cytokines were consistently upregulated in the supernatant of cultured NCS-01 cells. Briefly, the supernatant was collected from the culture medium of NCS-01 cells and transferred into a centrifuge tube which was then centrifuged at 1500 rpm for 10 minutes at 4 C.
Thereafter, aliquots of the supernatant were collected and stored at −80 C until use. Concentrations of cytokines were detected using ELISA kits (Abcam, United Kingdom) according to the manufacturer's instructions.

| Filopodia formation assay and IL-6 and bFGF treatment
This set of experiment was designed to reveal a potential regenerative mechanism mediating the therapeutic effects of NCS-01 cells on ischemic cells. Primary rat cortical neurons (which may also include a small percentage of glial cells) were subjected to OGD-reperfusion paradigm as described above. Next, NCS-01 cells were cocultured with the host cells for 3 hours at 37 C using a two-chamber system (Fisher), with NCS-01 cells suspended on the upper chambers at different distances (ranging from 0 to 2.04 mm) above host cells which were seeded on the lower chambers containing DMEM medium and poly-L-lysine coated glass coverslips. The lower chambers that contained host rat cells only (since the upper chambers small pore size did not allow migration of human NCS-01 cells toward the lower chambers, as confirmed by lack of HuMito stained cells in these chambers) were then processed for immunocytochemical analyses of formation of human filopodia using N-cadherin antibodies (Abcam). Double labeling with a neuronal cell death marker using propidium iodide (Abcam) staining was also carried out to assess any interaction between NCS-01 cell-derived filopodia formation and primary rat cortical neuron viability. A control group was also established comprising cortical neurons that were not subjected to OGD nor received any treatment. In the additional mechanistic probe test, primary rat cortical neurons, primary rat astrocytes, and primary rat EPCs were subjected to OGD and reperfusion as described above.  were housed under normal conditions, with two animals housed per cage (20 C temperature, 50% relative humidity, and 12-hour light/ dark cycle). The rats had free access to water and food. All necessary precautions were taken in order to reduce pain and suffering of the animals throughout the study. Animals were closely checked twice per day. All studies were performed by personnel blinded to the treatment condition.

| MCAO model
As previously described, stroke surgery was performed using the MCAO technique. 28,39,40 Animals were anesthetized with a mixture of 1% to 2% isoflurane in nitric oxide/oxygen (69%/30%) via a face mask, and body temperature was maintained at 37 ± 0.3 C during the surgical procedures. A midline skin incision was made in the neck with subsequent isolation of the left common carotid artery, the external carotid artery (ECA), and internal carotid artery. Thereafter, a 4-0 monofilament nylon suture (15.0-17.0 mm) was advanced from the common carotid artery bifurcation until it blocked the origin of the MCA. The skin incision was closed with surgical clips. Animals were allowed to recover from anesthesia during MCAO. At 1 hour after MCAO, animals were re-anesthetized, and reperfusion commenced with the withdrawal of the nylon suture. We have standardized the MCAO model, with stroke animals showing ≥80% reduction in regional cerebral blood flow (CBF) during the occlusion period as determined by laser Doppler (Perimed, Periflux System 5000). For baseline regional CBF measurement, the laser Doppler probe was placed over the right frontoparietal cortical area supplied by the MCA.
We also found no significant differences in physiological parameters, including PaO 2 , PaCO 2 , and plasma pH measurements. Rats that reached the 80% regional CBF reduction and >75% biased swing activity (see below) during occlusion were enrolled in these studies.
Thereafter, incisions were closed, and animals were allowed to recover from anesthesia. Whereas for transient MCAO, the nylon suture blocked the MCA for 60 minutes, for permanent MCAO, the nylon suture was not removed. butterfly needle and a 1-mL syringe was inserted into the stump of right ECA through an incision on the vessel and was tightened to the ECA by sutures. The catheter was positioned forward into ICA and passed the pterygopalatine artery, the extracranial branch of ICA, in order to enhance the delivery of the saline/fibroblast cells or stem cells toward the brain. The catheter was filled with saline/fibroblast cells or stem cells to prevent formation of potential harmful air bubble. The syringe was also filled with saline/fibroblast cells or stem cells. After dosing, the needle was removed without any flush and the CCA was closed with sutures to prevent any bleeding, and the skin wound reclosed with surgical clips. NCS-01 cell doses ranged from 7.5 × 10 4 to 3.75 × 10 7 total in 0.1 to 5 mL with 7.5 × 10 6 cell/mL concentration. In subsequent studies, the ICA route was vis-à-vis compared with IV administration (via the jugular vein).

| Behavioral tests
All investigators testing the animals were blinded to the treatment condition. Each rat was subjected to a series of behavioral tests to reveal motor and neurological performances at different time points before and after stroke and transplantation of NCS-01 cells. The tests included the modified Bederson neurological test (composite score of contralateral hind limb retraction, beam walking ability, and bilateral forepaw grasp), while motor function was assessed by the Elevated Body Swing Test (EBST). The Bederson neurological test is an evaluation of the animal's sensorimotor function, which consists of three distinctive phases, performed over approximately 10 minutes per rat.
Each phase of the test was conducted sequentially and was each scored 0 to 3. These three evaluations were: 1. contralateral hind limb retraction; 2. bilateral forepaw grasp; and 3. beam walking ability.
Contralateral hind limb retraction measured the ability of the animal to replace the hind limb after it was displaced laterally by 2 to 3 cm.
Grades were as follows: 0 for immediate smooth replacement, 1 for slow replacement, 2 for partial rigid replacement, and 3 for no replacement. Bilateral forepaw grasp measured the ability of a rat to hold onto a 2-mm diameter steel rod. Grade 0 was used for rats with normal forepaw grasping behavior, 1 for rapid grasping but with rigidity, 2 for slow grasping with rigidity, and 3 for a rat unable to grasp with the forepaw. Beam walking ability used a beam apparatus that was 80 cm long with a flat surface of 2-4 cm width resting at least 40 cm above the table/surface top on two poles. The animal was placed at one end of the beam then the ability to traverse the beam and reach the other end was assessed. The grades were as follows: 0 for a rat that easily traversed the beam, 1 if the rat slowly traversed the beam, 2 for partially traversing the beam but falls off, and 3 for a rat unable to stay on the beam for 10 seconds. The scores from all three tests were added to give a total neurologic deficit score (maximum possible score is 9 with mean composite neurologic score of 3). A score of 2.5 was set as a criterion to be considered a "stroke" animal. EBST is a measure of asymmetrical motor behavior that does not require animal training or drug injection. 41 The rats were held, in the vertical axis, approximately 1 in from the base of its tail and then elevated to an inch above the surface on which it has been resting. The frequency and direction of the swing behavior were recorded for over 20 tail elevations. A swing was counted when the head of the rat moved more than 10 from the vertical axis to either side. Normally, intact rats display a 50% swing bias, that is, the same number of swings to the left and to the right. A 75% swing bias toward one direction was used as criterion of motor deficit. 41 The total number of swings made to the biased side was added per group and divided by the n, providing the average number of swings per treatment group.

| Euthanasia and perfusion
Under deep anesthesia, rats were euthanized for immunofluorescent and protein analysis. For immunofluorescent analysis, briefly, animals were perfused through the ascending aorta with 200 mL of cold PBS, followed by 200 mL of 4% paraformaldehyde in phosphate buffer (PB). Brains were harvested and postfixed in the same fixative for 72 hours, followed by 30% sucrose in PB until completely sunk. Series of coronal sections were cut at a thickness of 40 μm using a cryostat and stored at −20 C. Brains were harvested and coronal sections were collected at a thickness of 2 mm using a brain matrix.
Staining for human mitochondria-positive cells was conducted on every 1 of 6 sections, 40 mm thick in brain. All sections were washed three times for 5 minutes in PBS. Sections were incubated with saline sodium citrate solution at pH 6 for 40 minutes at 80 C for antigen retrieval. Next, samples were blocked for 60 minutes at room temperature with 8% normal goat serum (Invitrogen, California) in PBS containing 0.1% Tween 20 (PBST; Sigma). Sections were then incubated overnight at 4 C with mouse anti-HuMito (Abcam) with 4% normal goat serum. Thereafter, the sections were washed five times for 10 minutes in PBST and soaked in 4% normal goat serum in PBST containing corresponding secondary antibodies, goat anti-rabbit IgG-Alexa Fluor 594 (red; 1:500; ThermoFisher Scientific) for 90 minutes.
Finally, sections were washed five times for 10 minutes in PBST and three times for 5 minutes in PBS, cover-slipped with Vectashield hardset with DAPI (H-1500, Vector Laboratories, California). All sections were examined using a confocal microscope (Zeiss). Control studies included exclusion of primary antibody substituted with 5% normal goat serum in PBS. No immunoreactivity was observed in these controls.

| Histology
Alternate brain tissue sections were processed for Nissl staining, which was performed with 0.1% cresyl violet solution (Sigma-Aldrich) using a standard protocol to evaluate the peri-infarct injury of our MCAO model. From each perfused brain, six coronal sections between the anterior edge and posterior edge of the MCAO infarct area were collected and processed for Nissl staining. Every sixth coronal tissue section was chosen at random to quantify cell survival in the peri-infarct area. Brain sections were examined using a light microscope (Keyence). Neuronal survival in the peri-infarct area of the brain was quantified using a computer-assisted image analysis system (NIH Image Software) and was expressed as a percentage of the ipsilateral hemisphere compared to the contralateral hemisphere.

| Statistical analyses
The data were evaluated statistically using two-way ANOVA followed by post hoc Bonferroni corrected pairwise comparison t-tests. Statistical significance was preset at P < .05 (Statview). The Kolmogorov-Smirnov test was performed to assess normality and the resulting values were less than 5% of the critical values. 3.1.2 | Route type favors ICA over IV cell delivery for reducing infarct area, but both routes comparably improve neurological function Based on the above in vitro dose finding study, in vivo experiments were conducted to find the best route to most effectively deliver NCS-01 cells into the target lesion. This study compared ICA and IV routes in a 2 × 2 between groups design. Groups of three (for IV F I G U R E 1 NCS-01 cells protect cultured primary rat cortical cells against oxygen glucose deprivation (OGD). Host cell viability was measured at different dosages relative to the number of host cells. Cell survival was assessed via trypan blue for control (0%) and at NCS-01 cell amounts ranging from 5% to 100%. Ratios of 1:4 and 1:1 displayed the greatest survival (#P < .05 vs 0%) (A). Cell survival was also assessed via trypan blue for control (0%) and higher amounts of NCS-01 cells, ranging from 100% to 400% (*P < .01, #P < .001 vs 0%) (C). All three of the higher ratios exhibited robust survival in comparison to the control, *P < .01. Mitochondrial activity was measured via MTT assay for the control and for NCS-01 cell amounts ranging from 25% to 100% (*P < .01, **P < .001, #P < .0001 vs 0%) (B) and from 100% to 400% (*P < .05, #P < .0001 vs 0%) (D), with another control. All ratios improved absorbance, with 4:1 providing the best results administration) or six (for ICA administration) male rats were subjected to 1 hour transient MCAO and were treated with either saline or 7.5 × 10 6 NCS-01 cells in 1 mL. Animals were followed-up for up to 7 days post-transplantation.
IV-delivered or ICA-delivered NCS-01 cells provided substantial neurologic and pathologic benefit as compared to saline when administered in rats with transient MCAO (Figure 2A,B). Furthermore, at 7 days poststroke, the ICA-administered NCS-01 cells reduced the infarct size almost twice as smaller as that seen with IV administration (P < .0001). Thus, ICA delivery was more effective than the IV route in ameliorating brain damage. In addition, neurological deficit was decreased with either route of NCS-01 cell administrations compared to saline-treated stroke rats. By day 7, NCS-01 cell treatment groups scored approximately 40% lower than the saline groups (P < .05, P < .0001). This measure was unaffected by route type, presumably indicating that the clinical measurements may not always reflect enhanced neurological function in all affected brain regions.
Taken together, these results suggest that ICA delivery of NCS-01 cells produced better stroke outcomes than IV delivery, at least in one aspect of brain damage. In addition, these results provided evidence that the therapeutic effects of NCS-01 cells seen in vitro translated to functional benefits in the in vivo stroke model.

| ICA infusion of NCS-01 cells poses no greater risk than no treatment or placebo
After determining logistical factors such as dosage and route, it is important to address safety concerns when considering eventual F I G U R E 2 Intracarotid artery (ICA) transplants of NCS-01 cells attenuate stroke-induced behavioral and histological deficits. Route efficacy was measured by comparing changes in infarct volume and neurological deficit after administration of either NCS-01 cells or saline (A, B). Both routes of NCS-01 cell administration decreased both infarction (*P < .05, #P < .0001) and neurological deficit scores (*P < .05, #P < .0001 vs saline) as compared to saline, but a difference was observed between ICA and IV delivery only in measures of infarct volume. In this case, ICA-administration was more effective than IV-administration. No difference was observed between routes for neurological deficit. In addition, two dosages of NCS-01 cells, saline, or no infusion (MCAO only) displayed no substantial differences when using cerebral blood flow assessment as a measure of safety indicating that ICA NCS-01 cells did not alter blood flow and was safe (albeit did not lead to overt embolism) (C) clinical translation of NCS-01 cells. To evaluate ICA administration safety, CBF assessment was conducted. Male rats (n = 6 per group)   All NCS-01 cell treatment groups significantly improved over the saline groups on all test outcomes by day 28 or even at earlier time points ( Figure 5). Infarct volumes ( Figure 5A,C,E) varied between approximately 60% and 80% reduction in all NCS-01 cell treatment groups compared to saline placebos (P < .05, P < .0001). The greatest reductions were found in rats treated with NCS-01 cells at day 1 and day 3 poststroke (P < .0001). While still significant, infarct reduction was somewhat attenuated for rats treated with NCS-01 cells more than 3 days post-MCAO.
A similar pattern of results was observed in neurological deficit scores ( Figure 5B,D,G). Neurological function improved in all NCS-01 cell treatment groups over the saline group, but this improvement was again slightly attenuated in rats treated more than 3 days post-MCAO (P < .05, P < .0001). Lastly, the third experiment also measured host cell survival in the contralateral (not directly affected by MCAO) and ipsilateral (directly affected by MCAO) peri-infarct area ( Figure 5F). Again, all NCS-01 cell treatment groups exhibited better host cell survival than the saline group, but this effect was attenuated for rats treated more than 3 days post-MCAO. Taken together, the data showed that delivery of NCS-01 cells ameliorated brain damage and neurological deficit, but that delaying treatment after 3 days post-MCAO may not provide as much functional benefits. This time-dependent therapeutic outcome of NCS-01 cells is a significant finding as it suggests the ideal clinical target would likely be patients less than 3 days after onset of ischemic stroke.

| Other MSCs (Li cells) and NCS-01 cells display comparable therapeutic effects on host cell viability
In an effort to determine whether NCS-01 cells performed equally or better than other MSCs, we initially embarked on in vitro tests of

| NCS-01 cells produce greater amounts of IL-6 and bFGF in vitro
Additionally, potential mechanisms of action for the neurorestorative effects of NCS-01 cells on cultures of primary rat cortical neurons and astrocytes were explored. In the brain, MSCs may secrete trophic factors and cytokines such as BDNF, β-NGF, IGF-1, VEGF, bFGF, and IL-F I G U R E 8 Legend on next page. 6. Since these trophic factors and cytokines have the potential to rescue cells against the deleterious effects of ischemic insult, a prelimi-   68 and 1.80 mm), and, to a lesser degree, the two farther distances (1.92 and 2.04 mm), also significantly improved cell viability and mitochondrial activity compared to the non-stem cell condition. Taken together, these results indicate direct cell-to-cell contact as optimally effective but suggest the potential to achieve equally robust ameliorative effects against OGD via indirect long distance rescue via filopodia formation (Supplemental Figure S2).
In order to probe the exact role of IL-6, bFGF, and NCS-01-derived filopodia in the rescue of different types of host cells, an additional measure was performed by exposing primary cortical neurons, primary rat astrocytes, and primary rat EPCs to OGD and reperfusion and treating with cell media only (OGD only control), IL-6 + bFGF, NCS-01 cell coculture, and a combination of IL-6 + bFGF and NCS-01 cell coculture. MTT assay reveals that compared to OGD only, all treatments improved mitochondrial activity in a similar pattern among cell types, with the greatest activity in the groups treated with NCS-01 cell coculture only and IL-6 + bFGF only ( Figure 8H). F I G U R E 8 NCS-01 cells, in part, employ filopodia-mediated long distance rescue of cultured primary rat cortical cells against oxygen glucose deprivation (OGD). Primary rat cortical cells were subjected to OGD and reperfusion then cocultured with NCS-01 cells at different distances, as depicted by procedural timeline and diagram (A and A 0 ). Propidium iodide and N-cadherin staining was conducted for the different distances: (B and B 0 ) 1.68 mm, (C and C 0 ) 1.80 mm, (D and D 0 ) 1.92 mm, and (E and E 0 ) 2.04 mm. Propidium iodide staining (red) reveals fewer dead primary rat cortical cells (indicated by red color) at closer distances than at farther distances between upper and lower chambers (B-E). N-cadherin staining (green) indicates the presence of filopodia extending from NCS-01 cells (B 0 -E 0 ) Scale bar = 50 μm. Quantitative analysis of primary rat cortical neurons reveals that cell viability correlates inversely with increased distance between NCS-01 cells and primary rat cortical neurons, when measured by trypan blue (TB) assay (F). Similarly, MTT assay reveals that greater mitochondrial activity correlates inversely with greater distance between NCS-01 cells and primary rat cortical neurons (G). Significance depicted as a-d at P < .05, a: greater than OGD only and all treatment groups; b: greater than all groups except control, non-OGD group; c: greater than OGD only and 1.92 and 2.04 mm groups; d: greater than OGD only (F, G). Additionally, primary rat cortical neurons, primary rat astrocytes, and primary rat EPCs were subjected to OGD and reperfusion then treated with IL-6 + bFGF and/or cocultured with NCS-01 cells at 0.8 mm distance. MTT assay displays greatest mitochondrial activity in all cell types when cocultured with NCS-01 cells only or treated with IL-6 + bFGF only (*P < .05, #P < .01, &P < .001, $P < .0001) (H). Moreover, N-cadherin staining demonstrates no significant difference in filopodia formation in the two cocultured groups and a similar pattern among different cell types (I) Neurons display the greatest recovery and are the only cell types to exhibit a slight but significant difference between IL-6 + bFGF treatment only and NCS-01 cell coculture only, with the latter performing slightly better than the former treatment (P < .05). There was no significant difference between IL-6 + bFGF-mediated and NCS-01-mediated improvements for astrocytes and EPCs. Interestingly, the combined treatment groups, while significantly improved compared to OGD only, performed substantially worse than either of their respective stand-alone groups in every case (P < .01, P < .001, P < .0001), with the minor exception of astrocytes treated with IL-6 + bFGF and those treated with combined, which did not differ significantly (P > .05). Furthermore, across all three neural cell lineages, Ncadherin staining indicates 35% to 55% increments in fluorescence intensity in stand-alone NCS-01 cell coculture or combined treatment compared to OGD only, indicating filopodia formation accompanied the improved cell viability and mitochondrial activity in these treatments ( Figure 8I). The observed increment in N-cadherin staining did not significantly differ between NCS-01 cell coculture only and combined treatment, suggesting similar robust filopodia formation in both treatments.
Overall, these results bolster our claim that improvements in mitochondrial activity via IL-6 and bFGF release and filopodia formation may be the primary modes of action of NCS-01 cells' rescue of host cells.

| DISCUSSION
The present study assessed the potential of NCS-01 cells as a cell- Finally, we observed for the first time a novel mechanism involving filopodia formation in stem cells under stroke conditions. That stem cells may propel cell processes in long distances toward the ischemic cells suggests the potential for transplanted cells to engraft in a conducive environment remote from the injured tissue, but still rescue ischemic cells. Filopodia formation has been previously observed in neuroprotective action by Rho kinase inhibition on organotypic hippocampal slices against in vitro ischemia. 44 Similarly, overexpression of transmembrane glycoprotein CD44 in vitro promotes the elongation, spread, and number of filopodia of cultured neural precursor cells, while in vivo accelerates the transendothelial migration and facilitates the invasion of certain perivascular sites. 45 Filopodia formation and cell motility, especially transendothelial migration, may be facilitated by adhesion molecules, such as Ninjurin 1, 46 and transcription factors, including serum response factor. 47 Understanding the putative role of these transmembrane glycoproteins, adhesion molecules, and regulatory factors may improve filopodia formation, as well as the resulting therapeutic outcome of NCS-01 cells in stroke.
Stroke is one of the main causes of disability among adults worldwide, with risk factors such as aging, hypertension, hyperglycemia, diabetes mellitus, and obesity. 48,49 The only FDA-approved drug is tPA with limited treatment window and high risk of hemorrhagic transformation; hence, there is a significant need for finding a novel treatment for stroke. Recently, stem cell therapy has emerged as a promising experimental treatment, even reaching clinical trials, for neurological diseases, including stroke, 49 Despite our extensive study designs and findings, there are certain limitations in the present study. In particular, future studies may consider using large animal models, such as porcine and/or nonhuman primate, to better mimic the human clinical pathologies (ie, white matter injury) and to further elucidate the current optimal dose, treatment timing, and route of administration of NCS-01 cells. In addition, since NCS-01 cells are derived from human cells and would be used for humans, the current human-to-rat paradigm may not fully capture the envisioned human-tohuman clinical product. Another limitation is our relatively young stroke animals did not exhibit comorbidities associated with stroke, warranting the need to test this cell therapy in a stroke model with comorbidity factors such as aging and hypertension. 53 Furthermore, while the secretion of bFGF and IL-6 coupled with the filopodia formation accompanied the therapeutic effects of NCS-01 cells, further manipulation of these phenotypic functions (eg, upregulating or downregulating cytokine release, and facilitating or inhibiting filopodia formation) may reveal more detailed brain repair machinery of this cell therapy. 54  Additionally, direct cell-to-cell contact, including filopodia interacting with other protrusive cell membranes may be required for the observed therapeutic activities. The ICA infusion appeared to improve cell delivery to the ischemic brain due to its capacity to directly deliver the highest number of cells to the lesion area without losing a high percentage of cell concentration to other circulatory areas. We further demonstrated that NCS-01 cells maintained a therapeutically active state under a dosedependent mechanism both in vitro and in vivo paradigms. The therapeutic window indicated that even though the best time to administer the NCS-01 cells was within 3 days post-MCAO, beneficial effects were still recognized when administered at 1 week after MCAO. Compared with other MSCs, NCS-01 cells ameliorated both neurostructural and functional deficits after stroke through a broad therapeutic window. Collectively, these translational observations provided solid lab-to-clinic evidence supporting NCS-01 cell therapy for stroke in the clinical setting.

ACKNOWLEDGMENT
The authors thank the entire Borlongan Neural Transplantation Laboratory for excellent technical assistance.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.