Stroke Recovery Boosted by 7-Day Neural Progenitor Cell Transplant

Transplantation of hiPSC-NPCs seven days after stroke results in long-term graft survival

To generate a cell graft with translational potential we differentiated human iPSCs into neural progenitor cells (hiPSC-NPCs, hereafter called “NPCs”) under transgene- and xeno-free conditions that can be smoothly adapted to GMP-grade for clinical applications as previously described23. The cell source has previously been characterized in-depth to spontaneously differentiate into astrocytes and functional, mature neurons and is compatible with safety switches for safe in vivo applications23. First, we confirmed the presence of canonical NPC markers including Nestin and Pax6 and the absence of pluripotent markers such as Nanog in NPCs (Fig. 1A, and Supplementary Fig. 1). Successful spontaneous differentiation upon growth factor withdrawal over a 4-week period was validated by upregulation of differentiation markers such as neuronal βIII-Tubulin and astrocytic S100β (Fig. 1B).

All experimental animal groups underwent a photothrombotic stroke or sham surgery in the right sensorimotor cortex. At 7 days post-injury (dpi), one group of mice received a local transplantation of NPCs expressing a dual-reporter (red firefly luciferase and eGFP, rFluc-eGFP, Supplementary Fig. 2), while the other groups (vehicle and control) were subjected to a sham transplantation (Fig. 1C). Cerebral blood perfusion was assessed in all animals by Laser-Doppler Imaging (LDI) to confirm stroke induction after surgery (Fig. 1D). LDI data demonstrated a 60-70% reduction in cerebral blood flow in both groups compared to non-stroked animals 30 min after stroke induction (vehicle: −61%, NPC: −69% Fig. 1E). There was no significant change in perfusion between treatment and control group (p > 0.05). Five weeks after stroke induction, infarcted tissue extended from +2.5 to –2.5 mm anterior-posterior relative to bregma (Fig. 1F). A non-significant decrease in lesion volume was observed in mice that received NPCs compared to the vehicle group (NPC: 2.24 mm3; vehicle: 3.66 mm3, p = 0.09, Fig. 1G, H).

Transplanted cell survival was confirmed over a 35-day time course using luciferase bioluminescence imaging (Fig. 1I). The bioluminescence signal remained relatively constant during the first two weeks after implantation (dpi 7–21), followed by a significant increase in bioluminescent signal until dpi 35 (p < 0.05; Fig. 1J). No increase in bioluminescence signal has been observed between 35 and 43 dpi (p = 0.99). As expected, no signal was detected in vehicle and sham control treatment groups that had not received cells.

To quantify cell graft size and migration, confocal microscopy images of serial brain sections stained with an antibody against human nuclei (HuNu) were evaluated. As NPCs can migrate within the brain30, we delineated the graft distribution in two distinct areas: the “graft center”, defined as the region with densely populated cells ( > 3x HuNu signal to background), and the “graft periphery”, where cell density was reduced (HuNu signal threshold between >1.5x and <3x to background; Fig. 1K). Areas of graft center and graft periphery were measured for all brain sections and identified with their stereotaxic coordinates (Fig. 1L). The largest distribution of peripheral graft cells was observed at bregma (graft distribution area: center = 0.32 mm2; periphery = 0.8 mm2); only minor migration of cells was found on the most anterior and posterior brain sections (0.12 mm2 and 0.25 mm2, respectively) (Fig. 1M). Together, this data indicates that NPCs grafts are capable of surviving for extended periods after implantation and are likely to migrate towards neighboring cortical regions.

NPC transplantation affects microglia response

After stroke, resting microglia react to the inflammatory stroke milieu and change their morphology from highly ramified into a more amoeboid type, depending on the distance to the lesion site (Fig. 2A). We therefore quantified three morphological features (branching, circularity, ramification) in addition to Iba1 fluorescence intensity to study the effect of NPC therapy on microglia activation (Fig. 2B). First, relative immunoreactivity levels of Iba1 were compared between treatment groups in the stroke core of the ipsilateral hemisphere and a corresponding region of interest (ROI) in the contralesional hemisphere for each mouse. Overall, we found an increased immunoreactivity of Iba1 in the ipsilateral cortex, with a strong signal in the stroke core, compared to the contralesional cortex in both groups. However, mice that received NPC exhibited a significant 32% reduction in Iba1 immunoreactivity within the stroke core relative to the vehicle group (Iba1 fluorescence intensity: NPC: +280%, vehicle: +410%, p = 0.02, Fig. 2C). As expected, no significant differences in Iba1 immunoreactivity were detected between the groups on the contralesional cortex (p = 0.42, Fig. 2C, and Supplementary Fig. 3A). Microglial branching index, a measure of microglial branching complexity, and the ramification index, a measure of the ratio of the cell’s perimeter to its area, was significantly reduced in the ischemic border zone (IBZ) compared to their respective contralateral ROIs. However, mice that received NPC therapy showed increased numbers in both, the branching index (vehicle: 19.1 ± 23.8, NPC: 29 ± 37.7, p < 0.05) and the ramification index (vehicle: 1.09 ± 0.16, NPC: 1.37 ± 0.35, p < 0.001) compared to vehicle animals (Fig. 2D). Importantly, no significant differences in branching index and ramification index were detected between the groups on the contralesional cortex (p > 0.05), and there were no significant differences in circularity between the treatment groups in the contralesional and the ipsilesional area of interest (p > 0.05). Elevated levels of GFAP immunoreactivity, indicating glial scar formation, were observed in the IBZ compared to the contralesional cortex in both groups (Fig. 2E, and Supplementary Fig. 3A). The increase in GFAP immunoreactivity was comparable between both treatment groups (GFAP fluorescence intensity: NPC: +1150%, vehicle: +1350%, p = 0.65, Fig. 2F).

To explore apoptopic cells in the ischemic border zone and the stroke core, capsase-3 immunohistochemical stainings were performed at 43 days after stroke induction (Fig. 2G). Numerous caspase-3+ cells were found in the ischemic border zone, as well as in the stroke core, compared to the uninjured contralesional side (Fig. 2H). However, we did not find any significant differences in caspase-3+ counts between NPC and vehicle treatment groups. To further characterize apoptotic cells and investigate potential differences between the two treatment conditions, we performed co-labeling with markers for endothelial cells (CD31, Supplementary Fig. 3B) and mature neurons (NeuN, Supplementary Fig. 3C). In the IBZ, 18.8 ± 6% of caspase-3+ cells co-expressed CD31 in the NPC group, compared to 13.1 ± 4% in the vehicle group (p = 0.46). In the stroke core, 24.8 ± 9% of caspase-3+ cells were CD31+ in the NPC group versus 18.3 ± 6% in the vehicle group (p = 0.57). Co-labeling with NeuN revealed that in the IBZ, 27 ± 7% of caspase-3+ cells were NeuN+ in the NPC group and 17.5 ± 4% in the vehicle group (p = 0.26). In the stroke core, we observed 15.1 ± 3% NeuN/caspase-3 double-positive cells in the NPC group and 23.9 ± 8% in the vehicle group, indicating a trend toward a higher proportion of apoptotic neurons in vehicle-treated animals, although this difference did not reach statistical significance (p = 0.34).

These results suggest that NPC therapy reduces microglial activation but does not affect glial scarring or apoptosis.

Increased neurite outgrowth and endogenous neurogenesis in NPC-receiving animals

The immunoreactivity levels of neurofilament light (NF-L) and heavy chain (NF-H), major proteins of the neuronal cytoskeleton, were quantified to determine overall axon regeneration in the stroke core and the surrounding IBZ and were normalized to the unaffected contralesional cortex (Fig. 3A, B). We observed significantly higher levels of NF-H in the IBZ of the NPC group (NPC: +48%, p < 0.05) compared to the vehicle group, but no differences in in the stroke core (Fig. 3C). For NF-L, we detected a significant increase in relative fluorescence intensities in the stroke core of the NPC group (NPC: +63%, p < 0.005) compared to animals that did not receive cell therapy. No changes have been observed between the groups for NF-L expression in the IBZ (Fig. 3D).

To visualize axons, the anterograde axonal tracer biotinylated dextran amine (BDA) was microinjected into the ischemic border zone in ipsilesional cortex adjacent to the stroke area of NPC- and vehicle-treated mice one week prior to sacrifice. Axonal density was quantified using fluorescence intensities normalized to BDA-labelled cell bodies and compared between NPC- and vehicle-receiving animals (Fig. 3E). NPC-grafted animals showed increased axonal connection density within the IBZ, compared to the vehicle group (1.83% ± 0.342%; NPC: 3.4% ± 0.36%, p = 0.0055, Fig. 3F). Together this data indicates that engrafted NPCs promote neuro- and axonogenesis in and around the infarct region.

Neurite extensions derived from the NPC graft were evaluated using human-specific neural cell adhesion molecule (hNCAM) staining (Fig. 3G–J). Among animals receiving cell transplants, 71% exhibited axonal extensions to the ipsilateral corpus callosum (cc), and 57% exhibited their neurites along the cc to the contralateral hemisphere (Fig. 3K). In 57% of the animals that received NPC grafts, we observed axonal extensions to the cingulate cortex. Notably, in 90% of all transplanted animals, cells extended their neurites along the ipsilateral cortex to the primary and secondary motor cortex and the primary somatosensory cortex.

NPC transplantation has previously been shown to promote endogenous neurogenesis in the subventricular zone (SVZ) and subsequent migration to the ischemic lesion site31. We therefore assessed if NPC-transplantation positively influenced endogenous neurogenesis in the ischemic cortex. We injected NPC- and vehicle-receiving animals with the nucleoside analog 5-ethynyl-2′-deoxyuridine (EdU) at day 7 following transplantation and, after sacrifice, identified EdU+/NeuN+ double-positive neurons (endogenous neurogenesis) and EdU+/NeuN+/HuNu+ tripple-positive neurons (exogenous neurogenesis) on a subset of coronal brain sections in the infarct core and the subventricular zone (Fig. 3L, M). While we could not detect any differences in endogenous neurogenesis in the ischemic core, we found a significant increase in EdU+/NeuN+ double-positive neurons in the SVZ in NPC-treated animals compared to the vehicle group (vehicle: 1.5 ± 0.577, NPC: 20 ± 7.638, p < 0.01, Fig. 3N).

These findings suggest that NPC transplantation enhances axonal sprouting, neurogenesis in the SVZ, and neurite outgrowth across multiple brain regions.

Increased vasculature density and reduced vessel leakage in mice treated with NPCs

To visualize blood vessel distribution in brain sections, coronal brain sections were stained for CD31 (Fig. 4A, B). 43 days after stroke induction, an ischemic border zone, characterized by hypovascularization, extended up to 300μm around the stroke core in vehicle-receiving animals (Fig. 4C). Vascular density, quantified here as vascular area fraction (NPC: 17%, vehicle: 9%, p < 0.001), length (NPC: 29 ± 5.7, vehicle: 18 ± 9, mm/mm2, p < 0.001), and number of branches (NPC: 651 ± 171, vehicle: 326 ± 219, mm-2, p < 0.001) were found significantly increased in the ischemic zone of the NPC group compared to the vehicle group (Fig. 4D). By comparison, no difference was observed between the two treatment groups in the contralateral cortex (Supplementary Fig. 4A, B). Interestingly, vascular area fraction and blood vessel length in the ischemic cortex recovered to similar levels as found in the contralateral intact cortex of NPC-transplanted animals.

To evaluate the number of newly-formed blood vessels in the ischemic border zone, the nucleotide analogue 5-ethynyl-2′-deoxyuridine (EdU) was systemically applied 7 days after cell transplantation (Fig. 4E, F). NPC-receiving mice showed an increase in EdU+ blood vessels per mm2 in the ischemic border zone compared to vehicle-receiving animals (NPC: 49.2 ± 14.3 mm-2; vehicle: 29.8 ± 7.4 mm-2; p < 0.05, Fig. 4G).

Blood vessel integrity was evaluated by utilizing vessel leakage as a parameter of vessel permeability. Immature newly formed blood vessels may cause leakage, which is linked to poor outcome after stroke32. To assess vessel leakage, coronal brain slices were co-stained with CD31- and fibrinogen antibodies, and sections of the ischemic border zone, ischemic core and intact contralateral cortex were analyzed in which cross-sections of blood vessels were visible (Fig. 4H, I). Based on the fibrinogen expression in proximity to these cross-sections, a vessel leakage profile was generated for each zone. A significant decrease in relative leakage was found in the stroke core of mice that received NPCs compared to those that had not received cells (NPC: +240%, vehicle: +435% leakage compared to contralateral side, p < 0.05) (Fig. 4J), while no difference was observed in the IBZ of both treatment groups. As expected, there was no change in vessel leakage in the contralesional cortex of both treatment groups.

To further address blood brain barrier (BBB) integrity, animals of both treatment groups were systemically injected with Evans blue (EB), a BBB permeability marker 24 h before sacrifice (Fig. 4K). EB has a high affinity for serum albumin, which does not cross the intact BBB to the brain parenchyma33. 43 days after stroke induction, coronal brain sections of both groups exhibited a very strong EB signal in the stroke core, and a lower signal intensity in the ischemic border zone (Fig. 4L, M). Compared to the contralateral cortex, both treatment groups showed a similar signal increase in both ROIs (IBZ; vehicle: +300%, NPC: +238%, stroke; vehicle: +462%, NPC: +340%, p > 0.5).