Human Neural Stem Cell Extracellular Vesicles Improve Tissue and Functional Recovery in the Murine Thromboembolic Stroke Model

To eliminate the potential confounding variable of genetic differences, NSC and MSC were isogenically derived from H9 pluripotent stem cells using processes previously developed [24‐26]. NSC and MSC EVs were quantified and evaluated for size differences using Nanosight’s nanoparticle tracking analysis. NSC and MSC EVs have overlapping, but distinct size and concentration profiles, with a broader peak present in the MSC EV profile indicating presence of a range of vesicles up to 300 nm in size, while the vast majority of NSC EVs were under 200 nm (Fig. 1a). Evaluation of NSC EVs by electron microscopy (EM) revealed the presence of disperse multivesicular bodies (MVBs; Fig. 1b, left panel) and purified vesicles (Fig. 1b, right panel) could be visualized by EM after transfer to the electron microscopy grid. Differentiated neural cells were cultured with NSC EVs labeled with DiI and EVs were taken up by the neural cells in vitro, as shown in super resolution confocal microscopy projection images (Fig. 1c and enlarged inset). Analysis of EVs by flow cytometry revealed that both cell types produced EVs that contained similar amounts of commonly reported EV markers such as CD63 and CD81, which are both members of the highly conserved tetraspanin superfamily.

In order to compare the therapeutic efficacy of isogenically derived NSC and MSC EVs side by side, EV biodistribution was first evaluated. Indium-111 (In-111)-labeled EVs were injected 1 h post-TE-MCAO. Animals were imaged by single photon emission computed tomography (SPECT) at 1 and 24 h post-injection (Fig. 2b) [27]. SPECT results demonstrated systemic distribution not only in the lungs, liver, and spleen, as reported in other EV biodistribution studies [16, 28], but were also present in the infarcted hemisphere by 1 h post-TE-MCAO. By 24 h, EVs were largely cleared from the infarct site, although still present in the other organs. These results suggest that EVs preferentially accumulate in the penumbra of the injury. Based on this clearance from the infarct, animals received a three-dose treatment regimen of either EVs or PBS vehicle by tail vein injection at 2, 14, and 38 h post-TE-MCAO. Animals were evaluated (after confirming no difference in cerebral blood flow; Fig. S1, a) by neurological deficit score (NDS) at 48 h and adhesive tape test (ATT) at 96 h post-TE-MCAO followed by blood collection and tissue analysis (Fig. 2a). NSC EV-treated animals during NDS assessment demonstrated a decrease in deficits compared to controls as evaluated by lower scores (p ≤ 0.055) Fig. 2c). NSC EV-treated animals performed significantly (p ≤ 0.001) faster on ATT (96.17 ± 11.57 vs. 162.53 ± 6.3 s, respectively), indicating enhanced sensorimotor function, when compared to controls or MSC EV-treated animals (Fig. 2d). Analysis of metabolically active tissue by 2,3,5-triphenyltetrazolium chloride (TTC) staining versus dead tissue (colorless) indicated significantly decreased tissue loss in NSC EV-treated animals compared to the MSC EV treatment group (27.97 ± 2.78 vs. 48.19 ± 5.79 mm², Fig. 2e, f). Since EVs are present in bodily fluids and they could affect the systemic immune response via both direct and indirect antigen presentation, we next checked the peripheral immune response after EV treatment. Quantitative flow cytometry analysis of freshly collected blood samples at 96 h post-stroke indicated that NSC EV treatment significantly promoted macrophage polarization toward an anti-inflammatory M2 phenotype (Fig. 3a–c, j) and increased the regulatory T cell (Fig. 3d–f, k) population resulting in the downregulation of pro-inflammatory effector Th17 cells (Fig. 3g–i, l). Thus, our data indicates that NSC EV treatment after stroke is capable of dampening injury responses while augmenting a reparative systemic immune response (Fig. 3). In summary, this data indicates PSC-derived NSC EVs provide molecular and behavioral benefits, while PSC-derived MSC EV treatment resulted in more variable results in both infarct size and behavioral outcome assessment indicating a clear NSC EV benefit in the middle-aged embolic model. While overall survival was not significantly different between the groups, 55% of animals in the MSC EV and PBS groups survived to the endpoint, while 65% of NSC EV-treated mice survived (Fig. S1). For these reasons, NSC EVs were further explored as a candidate treatment, while evaluation of MSC EVs was discontinued.

Stroke therapeutics are often tested in young animals within a narrow time range post-stroke. NSC EVs were further explored in aged mice (18 ± 1 months), starting approximately 6 h post-stroke, to fall outside the time window of traditional tPA administration in humans. Dosage in the embolic model was maintained constant; however, the administration window was shifted to 6, 24, and 48 h post-stroke. (Fig. 4a). Blinded investigators randomly divided mice into non-stroked (sham) and stroked with either PBS vehicle (control) or NSC EV in PBS treatment groups (N = 24 animals/group). Analysis of T2-weighted (T2W) sequences 2 days post-TE-MCAO indicated a significant decrease in lesion volume in NSC EV-treated animals (58.2 ± 5.03 and 37.9 ± 2.84 mm³, respectively) (Fig. 4b, c), while ex vivo Q-ball MRI (performed on the fixed brain post-euthanasia) indicated that NSC EV treatment attenuated the post-stroke cerebral atrophy and significantly decreased it compared to the vehicle-treated group (22.8 ± 0.40 and 10.6 ± 1.94% of contralateral hemisphere) (Fig. 4d). Diffusion tensor imaging (DTI) and fractional anisotropy (FA) analysis was also performed after Q-ball imaging; however, no significant differences in diffusivity or white matter integrity were observed between the two groups subjected to TE stroke, which is likely due to less white matter content in small rodents.

Behavioral characteristics and motor function were evaluated 14 days post-TE-MCAO. NSC EV-treated animals exhibited significantly improved coordination on the balance beam relative to control, with NSC EV-treated animals crossing in 18.9 ± 1.36 s and control animals crossing in 28.0 ± 0.45 s (Fig. 4e). Significantly fewer foot slips while crossing the beam (2.21 ± 0.18 vs. 1.25 ± 0.21 foot slips) were also observed in NSC EV-treated animals (Fig. 4f). Grasping ability and forelimb strength were evaluated by the hanging wire test. NSC EV-treated animals could hang an average of 28.47 ± 1.18 s, while control animals grasping was significantly shorter (5.1 ± 0.91 s) (Fig. 4g). Episodic memory was evaluated by novel object recognition (NOR) testing. NSC EV-treated mice spent significantly more time exploring the novel object (NO; 36.92 ± 1.48 s) than the control group that spent only 26.50 ± 3.29 s on average with the NO. There were no significant differences in time spent with the familiar object between groups. Novel object discrimination index (DI) indicated NSC EV-treated animals performed significantly better than control group (0.26 ± 0.04 and 0.0005 ± 0.05, respectively; Fig. 4i). Finally, depressive phenotype was assessed by tail suspension test 28 days post-TE-MCAO. Controls were immobile for a significantly longer time period (178.13 ± 9.96 s) as compared to NSC EV-treated animals (123.08 ± 9.58 s) (Fig. 4h). The NSC EV group was not statistically different from the sham group in survival rates, while fewer animals survived to the endpoint in the control group (Fig. S1 a; p ≤ 0.319). Collectively, this data indicates an early neuroprotective effect of NSC EV in aged mice as indicated by reduced lesion volume and improvements in functional outcomes as measured by grasping ability, forelimb strength, motor coordination, and memory consolidation.