Brain endothelium-derived extracellular vesicles containing amyloid-beta induce mitochondrial alterations in neural progenitor cells

Aim: Elevated brain deposits of amyloid beta (Aβ40) contribute to neuropathology and cognitive dysfunction in Alzheimer’s disease (AD). However, the role of the blood-brain barrier (BBB) as an interface for the transfer of Aβ40 from the periphery into the brain is not well characterized. In addition, a substantial population of neural progenitor cells (NPCs) resides in close proximity to brain capillaries that form the BBB. The aim of this study is to understand the impact of brain endothelium-derived extracellular vesicles (EV) containing Aβ40 on metabolic functions and differentiation of NPCs. Methods: Endothelial EVs were derived from an in vitro model of the brain endothelium treated with 100 nM Aβ40 or PBS. We then analyzed the impact of these EVs on mitochondrial morphology and bioenergetic disruption of NPCs. In addition, NPCs were differentiated and neurite development upon exposure to EVs was assessed using the IncuCyte Zoom live cell imaging system. Results: We demonstrate that physiological concentrations of Aβ40 can be transferred to accumulate in NPCs via endothelial EVs. This transfer results in mitochondrial dysfunction, disrupting crista morphology, metabolic rates, fusion and fission dynamics of NPCs, as well as their neurite development. Conclusion: Intercellular transfer of Aβ40 is carried out by brain endothelium-derived EVs, which can affect NPC differentiation and induce mitochondrial dysfunction, leading to aberrant neurogenesis. This has pathological implications because NPCs growing into neurons are incorporated into cerebral structures involved in learning and memory, two common phenotypes affected in AD and related dementias.


INTRODUCTION
Alzheimer's disease (AD) is the most common neurodegenerative disease, affecting one in every eight people over the age of 65 and around half of the individuals over the age of 85 [1] . Elevated brain deposits of beta-amyloid (Aβ) are now accepted to contribute to the neuropathology and cognitive dysfunction seen in AD [2] . The levels of Aβ in the central nervous system (CNS) result from an equilibrium between production in the brain, influx from plasma, and efflux via blood-brain barrier (BBB) efflux transporters [3] . Although these processes contribute to the formation of Aβ deposits and plaques in AD, emerging evidence indicates that the BBB is a critical interface in the transfer of Aβ from the periphery into the brain [4,5] . However, the mechanisms underlying this process remain unclear. Recent evidence indicates that it is important to consider the role of the BBB in the transfer of Aβ when analyzing neuropathologies associated with AD [6][7][8][9] . We previously demonstrated that exposure to EVs carrying Aβ 40 (EV-Aβ 40 ) negatively affects the differentiation of NPCs into mature neurons; however, the mechanisms behind these effects are poorly understood [10,11] . In the present study, we hypothesize that an underlying cause of altered neurogenesis, which is a prominent feature of AD, may be mitochondrial dysfunction that results from EV-mediated Aβ 40 exposure. The effect of exposing NPCs to Aβ 40 that may be trafficked across the BBB, compounded with endogenous production of Aβ 40 in the brain, may exacerbate the mitochondrial dysfunction observed in progenitor cells, leading to aberrant neurogenesis.
The BBB provides a physiological, immunological, and structural interface between the brain parenchyma and the peripheral circulation. It also represents a crucial interface between the circulatory system and brain tissue, helping to maintain the overall homeostasis of the CNS. The brain endothelium, formed by cerebral endothelial cells, is the crucial element of the BBB that is responsible for its structural and metabolic properties. These endothelial cells produce EVs that can travel and interact with other cell types in the brain to deliver various biologically active cargo materials, such as Aβ 40 [10,11] . Aβ 40 is produced by the cleavage of amyloid precursor protein (APP), which is ubiquitously expressed on the membrane of neurons as well as endothelial cells that make up the BBB [12,13] . CNS levels of Aβ 40 result from an equilibrium established between production in the brain, influx from plasma, and efflux by BBB transporters. We here study the involvement of Aβ 40 because it is the main isoform produced by the cleavage of APP, making up 90% of all Aβ 40 produced in the brain. Increased Aβ 40 deposition is also associated with vascular dysfunction and is likely to compromise the integrity of the BBB during AD [14,15] . Therefore, our studies focus on this amyloidogenic species along with its ability to cross the BBB and impair neurogenesis via EV-mediated transfer.
EVs have recently been postulated to be significantly involved in various neurodegenerative diseases, including Aβ 40 pathology [16,17] . It was demonstrated that Aβ 40 could be packaged into EVs and shed into the microenvironment and that human amyloid plaques are enriched with EV proteins [9] . Additionally, we previously showed that exposure of brain endothelial cells to Aβ 40 results in effective packaging of Aβ 40 in EVs, leading to the transfer of Aβ 40 into neighboring cells, including NPCs. EVs are heterogeneous in their shape, size, origin, and content. They include exosomes with a size of approximately 30-100 nm and other larger vesicles [18,19] . EVs are formed in a stepwise process whereby the vesicles bud from the endosomal membranes generating intra-endosomal vesicles, followed by fusion of the endosome membrane with the plasma membrane and release of the EVs into the extracellular environment. Their content is diverse and includes mRNAs, miRNAs, lipids, and proteins [20] . EV cargo can be released into the immediate microenvironment or peripheral sites when transported via the circulatory system [21] . EVs, being of endosomal origin, have been shown to be released from multiple cell types creating a unique cell-to-cell communication system, which may play important physiological and pathological roles in the CNS [22,23] . However, few studies have focused on EVs derived from brain endothelial cells, or on the effect of their cargo on neurogenesis in the brain. This process is of physiological significance because neurogenic niches are predominantly localized around the brain microvasculature that forms the BBB [24] . Although Aβ 40 accumulation has a detrimental effect on neuronal growth in diseases such as AD, the mechanism of action that leads to aberrant neurogenesis remains largely unknown. Therefore, the objective of the present study is twofold: (1) to determine the mechanistic effect of Aβ 40 on NPCs that may lead to aberrant neurogenesis; and (2) to delineate whether vesicle-mediated trafficking of Aβ 40 can potentiate the neurogenic deficiency seen in AD patients. Our results demonstrate that brain endothelium-derived EVs can transfer cargo to NPCs and may serve as a vehicle for cell-to-cell communication that regulates mitochondrial metabolic dynamics in NPCs and, ultimately, decreases their neurogenic potential.
genotype. The iPSC line was differentiated into NPCs in accordance with our previously reported approach [25] . Cells were plated on day in vitro (DIV) 20 on poly-L-ornithine/ laminin plates in D/NPEN medium [Supplementary

Transfection of hCMEC/D3
hCMEC/D3 were transfected with the CD9 Cyto-Tracker constructs to label EVs. EVs were labeled with RFP through the expression of pCT-CD9-RFP Cyto-Tracker vector (pCMV, Exosome/Secretory, CD9 Tetraspanin Tag). The cells were transduced using a transfection kit (System Biosciences, Cat# CYTO123-PA-1) and PureFection Transfection Reagent (System Biosciences, Cat# LV750A-1), in accordance with the manufacturer's protocol. Twenty-four hours after transfection, hCMEC were exposed to 100 nM fluorescent Aβ 40 HiLyte and the EV-containing supernatant was harvested 48 h later. EVs were isolated from the tissue culture supernatant as outlined below. Once isolated, EVs were pipetted onto glass slides, heat-fixed at 95 °C for 10 min, and fixed again with 100% ethanol for 30 min at 4 °C. Slides were washed three times with PBS and mounted using ProLong Gold Antifade reagent with DAPI (Invitrogen). Specimens were covered with coverslips and fluorescent images [green fluorescence (Aβ HiLyte Alexa Fluor488) and red fluorescence (CD9-RFP)] were acquired using a Nikon Eclipse Ti-U inverted fluorescent microscope. The settings for fluorescence intensity and contrast were optimized and the same settings were used for the analysis of EVs. Intensity plot images were acquired using Nikon-Elements imaging software (NIS).

Aβ 40 -loaded EVs
EVs were loaded with Aβ 40 as described earlier. Briefly, hCMEC/D3 cells were grown as a monolayer and split when greater than 90% confluent in exosome-depleted complete medium incubated at 37 °C and 5% CO 2 . Freshly solubilized Aβ 40 (Anaspec, Cat# AS-20698) or Aβ 40 HiLyte (Anaspec, Cat# AS-60491-01) was used in the following experiments because this species has been shown to induce proinflammatory reactions in brain microvessels and endothelial cells and is the most abundant of the proamyloidogenic species produced from APP [26,27] . Aβ 40 was dissolved in 0.1 M NH 4 OH basic buffer (Anaspec, Cat# AS-61322) and further diluted in PBS, in accordance with the manufacturer's instructions. Then, hCMEC/D3 were treated with fresh EBM-2 medium supplemented with growth factors and 100 nM Aβ 40 for 48 h. This dosage of Aβ 40 was chosen since the Aβ 40 levels in AD patients' CSF are in the nanomolar range. This concentration has also been shown to result in maximal EV loading [28] . Control cultures received fresh medium supplemented with PBS as a vehicle control treatment. Treatment was terminated by removal of the medium and EVs were isolated from ~1.7 × 10 7 cells using the ExoQuick-TC kit (System Biosciences, Cat# EXOTC50A-1). Briefly, 10 mL of hCMEC/D3 culture supernatant was mixed with 2 mL of ExoQuick Exosome precipitation solution. This mixture was incubated overnight at 4 °C to facilitate even coating of the EVs. The next day, the samples were centrifuged at 1500 × g for 30 min, the supernatants were removed, and the pellets were centrifuged further at 1500 × g for 5 min. The resulting EV pellets were resuspended in PBS for characterization or use in our experiments. The whole fraction of isolated EVs was used in our experiments to account for both small EVs (exosomes) as well as larger particles, both of which can contain Aβ 40 .

Nanoparticle tracking analysis
EVs isolated from the supernatant of hCMEC/D3 cells were characterized and analyzed by nanoparticle tracking analysis (NTA) on the Nanosight NS300 (Malvern Instruments Company, Malvern, UK). Optimal detection settings were acquired with the screen gain set to 11 and a detection threshold level of 5 using Nanosight NTA 2.

Enzyme-linked immunosorbent assay
Human Amyloid Beta 40 enzyme-linked immunosorbent assay (ELISA) (Invitrogen, Cat# KHB3481) was used to quantify the levels of Aβ 40 loaded into EVs. First, EVs were lysed at a 1:1 ratio with RIPA buffer (Thermo Fisher Scientific, Cat# 89900) containing protease inhibitor (Thermo Fisher Scientific, Cat# 78431). The resulting lysate was loaded into a precoated 96-well plate, in accordance with the manufacturer's instructions, and absorbance was measured using the Colorimetric SpectraMax 190 plate reader and normalized to total protein levels as measured by a BCA assay (Thermo Fisher Scientific, Cat #23223).

IncuCyte zoom
iPSC-derived NPCs were seeded at a cell density of 40,000 cells per well in a sixwell culture plate coated with poly-L-lysine and allowed to differentiate into neurons. Alternatively, ReN cells were seeded at 60,000 cells per well in a 12-well plate coated with laminin. The IncuCyte Zoom automated cell imaging system (Sartorius) was used to automatically acquire kinetic data every hour for either 24 h (iPSC-derived neurons) or 14 days of NPC differentiation (ReN cells) to assess neurite outgrowth and neurogenesis, respectively. iPSC-derived neurons were treated once with EVs, Aβ 40 , or medium alone. Images were taken every hour for 24 h. ReN cells were treated with EVs or Aβ 40 every other day along with fresh medium over the course of 14 days of differentiation. Neurotrack software was used to quantify neuron cell-body clusters, neurite branch points, and neurite branch points per cell-body area.

Quantitative PCR
Total RNA was isolated from cell culture lysates using RNeasy mini kit (Qiagen, Cat# 74104), in accordance with the manufacturer's instructions. Total nucleic acid concentrations were quantified using the Nanodrop 2000 (Thermo Fisher Scientific). RT-PCR was performed with a total of 100 ng of RNA using the qScript XLT 1-Step RT-qPCR ToughMix Low ROX (Quantabio, Cat# 89236-676) reaction mix and the Applied Biosystems 7500 system (Applied Biosystems, Foster City, CA, USA). TaqMan Gene Expression Assays and the following primers were used: Hs00208382_m1: Mfn2 primer; Hs00250475_m1: Mfn1 primer; Hs00953477_m1: SIRT3 primer; Hs00697394_g1: Mff primer; Hs01047018_m1: OPA1 primer; Hs00608023_m1: Bcl-2 primer; and DNM1L-Hs01552605_m1: Drp-1 primer. Human GAPDH (ThermoFisher Scientific, Cat# 4310884E) was used as an endogenous control for sample normalization. PCR product specificity was determined using melting curve assessment and gene expression differences were determined using the ΔΔCt method.

Transmission electron microscopy
EVs-EV samples from both control and Aβ 40 -treated groups were visualized using transmission electron microscopy (TEM). EV suspensions were fixed in 2% paraformaldehyde for 1 h and then loaded on a copper mesh grid and allowed to incubate for 45 min. The grids were allowed to dry overnight. Then, the EV suspension was stained using 1% uranyl acetate. Once the grids were dry, EVs were imaged using a JEOL JEM-1400 transmission electron microscope with images acquired using an AMT BioSprint digital camera.
ReN cells-Cells grown on glass coverslips were fixed overnight at 4 °C in 2% glutaraldehyde in 0.1 M phosphate buffer, before being incubated for 1 h in 2% osmium tetroxide in 0.1 M phosphate buffer, dehydrated through a series of graded ethanol concentrations, and embedded in EM-bed (Electron Microscopy Sciences). The glass coverslips were removed by treatment with hydrofluoric acid. Next, 100 nm sections were cut on a Leica Ultracut EM UC7 ultramicrotome and stained with uranyl acetate and lead citrate. The grids were viewed at 80 kV in a JEOL JEM-1400 transmission electron microscope and images were captured by an AMT BioSprint digital camera. For quantitative analysis, fields were chosen at random, and acquisition and quantification were performed using FIJI (NIH) and GraphPad Prism 9.3.1 software, respectively. Morphometric analyses of the mitochondria and substructures were performed by three investigators in a randomized, double-blind manner, which included taking the images, quantifying the data, and running the statistics separately. Images were obtained at random from three independent experiments [29] .

Western blot
Treated cells were washed with PBS and lysed with radioimmunoprecipitation assay (RIPA) buffer (Thermo Fisher Scientific, Cat# 89900) supplemented with protease inhibitors (Thermo Fisher Scientific, Cat# 78431). Total protein concentration was determined using the BCA protein assay kit (Thermo Fisher Scientific, Cat# 23223). Samples were prepared for immunoblotting with RIPA diluent and sodium dodecyl sulfate (SDS). Then, 15-20 μg of total protein was loaded per lane on 4%-20% midi-PROTEAN TGX Stain-Free polyacrylamide precast gels (Bio-Rad Laboratories, Cat# 4568096) and transferred to a nitrocellulose membrane using the Trans-Blot Turbo Transfer System (Bio-Rad Laboratories, Cat# 1704159). Membranes were blocked using Intercept TBS Blocking Buffer (LICOR, Cat# 927-60001) for 1 h at room temperature. The next day, blots were incubated overnight at 4 °C with primary antibodies [Supplementary Table 2] diluted in 5% bovine serum albumin (BSA). Blots were washed three times using Tris-buffered saline with 0.1% Tween-20 (TBS-T) before incubation with secondary antibody at room temperature for 1 h. Analysis of GAPDH (Novus Biologicals, Cat# NB600-502FR or Cat# NB600-5021) was performed as an internal control and for sample normalization. Bands were imaged with the Odyssey CLX Imaging System (LICOR) and signal quantification was performed with Image Studio 4.0 software (LICOR).

Seahorse XF analyzer
Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured on a Seahorse XF24 Analyzer (Agilent Technologies, Inc.) using the Seahorse Mito Stress Test kit (Agilent Technologies, Inc., Cat# 103015-100). ReN cells were seeded at a density of 40,000 cells per well in a 24-well Seahorse tissue culture microplate (Agilent Technologies, Inc., Cat# 100777-004) coated with purified mouse laminin (Millipore, Cat# CC095-5MG). Once cells reached ~80% confluence, they were treated with EV-PBS, EV-Aβ 40 , or Aβ 40 alone for 24 h. Three basal respiratory measurements were obtained before injections of oligomycin (1.5 mM), FCCP (0.5 mM), and rotenone/antimycin A (0.5 mM). OCR and ECAR measurement results were then recorded, and all data were normalized to total protein levels (mg/mL) per well as measured by a BCA assay. Each data point represents the average measurement from five separate wells.

Confocal microscopy
NPCs were cultured on laminin-coated eight-well chamber slides (Ibidi, Cat# 80841) and treated with EV-PBS, EV-Aβ 40 , or Aβ 40 for 24 h. The cells were subsequently fixed with 4% PFA for 10 min, washed with PBS, and permeabilized with PBS containing 0.1% Triton X-100 (Millipore Sigma, Cat# 93443). The fixed and permeabilized cells were incubated with blocking buffer (20% FBS and 0.5% Tween-20 in PBS) for 1 h. Then, samples were incubated overnight at 4 °C with primary antibodies [Supplementary Table  3]. The next day, primary antibody was aspirated, and slides were washed with PBS and incubated with Alexa Fluor-conjugated secondary antibodies for 2 h at room temperature. Slides were mounted with a glass coverslip in Fluoromount-G Mounting Medium with 4',6-diamidino-2-phenylindole (DAPI; Invitrogen, Cat# 00-4959-52) to visualize the nuclei. Immunofluorescent images were captured with a Fluoview 1200 confocal microscope (Olympus, 2× dry lens, or 40× or 100× oil immersion lens).
For the nuclear factor-κB (NFκB) measurements, mean fluorescence intensity was measured in the nuclear areas outlined by DAPI staining using FIJI software (v.2.3.0, NIH) and the Isodata algorithm. It was then used to convert the NFκB image to a binary mask that included all fluorescence data above the background. Nuclear and whole-cell histogram data were calculated and normalized for the total number of nuclei included in each image. A comparison was then made from the difference in staining intensities. NFκB nuclear translocation is represented by an increase in nuclear:cytoplasmic ratio of signal intensity. For TOM20 measurements, the total area of fluorescence intensity on the acquired images was normalized to the number of nuclei. Confocal microscopy images were analyzed using FIJI software (v.2.3.0, NIH). Images were pre-processed using the Bio-Formats (Available from: https://github.com/ome/bioformats) plug-in splitting fluorescent channels: DAPI (408 nm) and TOM20 (488 nm). Mitochondrial network parameters including mitochondrial footprint, mean branch length, total branch length, and mitochondrial network were analyzed using the Stuart Lab Mitochondrial Network Analysis (MiNA) (Available from: https://github.com/StuartLab) plug-in [30] .

Statistical analysis
Data are presented as the mean ± standard error of the mean (SEM). All statistical tests were performed and all graphs were constructed using GraphPad Prism software 9.3.1. Statistical comparisons were conducted using one-way ANOVA with Tukey's multiple comparisons test, or an unpaired t-test where indicated. Levels of significance were set to *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Characterization of brain endothelium-derived EVs loaded with Aβ 40
The BBB has been shown to act as a route of entry for Aβ 40 to transfer into the brain at concentrations in the nanomolar range. Therefore, our studies examined these physiological concentrations using Aβ 40 -loaded EVs. In the initial series of experiments, EVs were isolated from the cell culture medium collected from brain endothelial cells (hCMEC) that had been treated for 48 h with exogenous Aβ 40 or PBS as a vehicle control. The isolated EVs were assessed by nanoparticle tracking analysis (NTA) and transmission electron microscopy (TEM) to characterize EV size, morphology, and distribution. These analyses showed that the isolation procedure yielded a heterogeneous range of sizes of EVs isolated from both control [ Figure 1A] and Aβ 40 -treated [ Figure 1B] endothelial cells. Most EVs fell between 100 and 400 nm in size for each group. EVs isolated from control hCMEC had a mean size of 182.2 nm (±7.3 nm) and a total mean EV concentration of 3.67 × 10 8 particles/mL. EVs isolated from hCMEC treated with Aβ 40 had a mean size of 186.4 (±7.8 nm) and a total mean EV concentration of 2.73 × 10 8 particles/mL [ Figure 1A and B]. We did not observe any significant change in particle size after treatment with Aβ 40 . TEM analysis of the isolated EVs confirmed the size and morphology of control EVs and EV-Aβ 40 ( Figure 1A and B, respectively). EVs isolated from hCMEC treated with Aβ 40 exhibited effective packaging of Aβ 40 comparable to that of those isolated from PBS-treated cells, as measured by ELISA [ Figure 1C]. Once the size distribution was determined, we analyzed EVs for the presence of specific markers, as well as the presence or absence of Aβ 40 in whole-vesicle lysates. Ubiquitous vesicular markers -CD9, CD63, and CD81 -are tetraspanin proteins present in lipid membrane vesicles of various sizes. All three markers were expressed in isolated EVs, along with Aβ 40 in EVs isolated from Aβ 40 -treated cells [ Figure 1D]. Additionally, we probed for CD31 to confirm that our EVs were derived from endothelial cells making up the BBB. Finally, hCMEC were transfected with CD9-RFP, followed by treatment with green fluorescent Aβ 40 , to visualize co-localization of the EV marker CD9. Uptake of the CD9-RFP and co-localization with the green fluorescence signal from Aβ 40 HiLyte were assessed by fluorescence microscopy. A surface intensity plot was then created to visualize the intensity peaks of both green (Aβ 40 ) and red (CD9) signals and their co-localization. Partial overlap of the two signals demonstrated the diversity of EV species and the efficacy of Aβ 40 uptake into these vesicles [ Figure 1E]. In addition, the EVs stained positive for DAPI, suggesting the uptake of genetic material into them. It is expected that there is a heterogeneous mixture of free Aβ 40 (partially due to EV rupture during the preparation and the efficacy of Aβ 40 uptake) as well as EV-associated Aβ 40 in these isolates.

NPCs are vulnerable to aberrant neurogenesis when treated with brain endotheliumderived EVs containing Aβ 40
In the next series of experiments, we assessed the impact of EVs on NPC differentiation using ReN cells. The ReN-derived NPCs were allowed to differentiate into neurons through the removal of growth factors from the medium, and images were taken every hour over 14 days. Treatment with EV-Aβ 40 resulted in lower ratios of neurite extension (traced in pink) to cell body (traced in yellow), compared with those of NPCs exposed to control EVs, which displayed more healthy dendritic extensions and synapse formation between cells [ Figure 2A]. Cell-body clusters, outlined in yellow, tended to form dense clusters in the EV-Aβ 40 and Aβ 40 groups compared with those in the untreated and EV-PBS NPCs [ Figure 2B]. The total number of neurite branch points was calculated as an indicator of proper development and synapse formation. In addition to a higher tendency for cell bodies to cluster together, both Aβ 40 groups (EV-Aβ 40 and Aβ 40 alone) showed reduced numbers of branch points (indicated by pink tracing of the neurites). The number of branch points/ cell-body area tended to be higher in control and EV-PBS groups than in EV-Aβ 40 -and Aβ 40 -treated ReN-derived NPCs [ Figure 2C]. A decrease in the neurite branch points to cell-body ratio was observed during this differentiation time course following treatment with Aβ 40 . Meanwhile, the control groups displayed a tendency for an elevated growth rate [ Figure 2D]. Upon treatment with EVs and fresh medium, it was common to observe elevated rates of growth upon IncuCyte analysis (spikes in plots B-D), with cells returning to a basal state of growth in the time following replacement of the medium.
We also confirmed the impact of EV-Aβ 40

Brain endothelium-derived EVs containing Aβ 40 alter mitochondrial crista morphology
To investigate the cause of NPC disruption and aberrant neurogenesis upon exposure to EV-Aβ 40 , we sought to understand whether metabolic regulation of NPCs was being affected through alterations in mitochondrial functionality. Mitochondria are essential organelles that mediate cellular function and survival. Organs and tissues that have high energetic requirements, such as the brain, have higher overall levels of mitochondria. The morphology of mitochondria in NPCs was analyzed using TEM following 24 h of treatment with medium alone (control), EV-A 40 , control EVs (EV-PBS), or A 40 alone. Untreated NPCs and NPCs exposed to EV-PBS exhibited healthy mitochondria with dense, regular, inner mitochondrial membrane formation [ Figure 3A and B]. Meanwhile, NPCs exposed to EV-Aβ 40 and Aβ 40 alone exhibited morphological changes in crista structures [ Figure 3C and D]. In addition, the presence of white patches, or crista "voids" (indicated by the arrows), was consistently seen in both EV-Aβ 40 -and Aβ 40 -exposed groups, but not in the control groups (control and EV-PBS).
The mitochondrial area and crista structures were analyzed in the perinuclear region of NPCs, as previously described [29] . Limiting the analysis to this region is crucial to control for the high rate of mitochondrial dynamics normally found in the growth extension regions of differentiating cells. The total perinuclear mitochondrial area was not significantly changed between groups (data not shown). Using TEM analysis, we observed a significant decrease in crista number in both EV-Aβ 40 -and Aβ 40 -exposed NPCs compared with the level in untreated or EV-PBS-exposed cells [ Figure 3E]. The total crista area was determined from individual mitochondria and the summation of outlined crista surface area was measured. The surface area of perinuclear mitochondrial cristae was significantly decreased between control NPCs compared with that of EV-Aβ 40 -exposed cells [ Figure 3F]. In addition, we measured the density of cristae, which reflects the ratio of crista surface area to the associated mitochondrion area. The relative density of cristae exhibited a significant decrease in both Aβ 40 treatment groups (EV-Aβ 40 and Aβ 40 alone) compared with the level in the control groups [ Figure 3G]. Similarly, mitochondria in each group were assigned morphometric crista scores to visually assess crista definition. This scale ranges from 0, corresponding to no sharply defined cristae, to 4, corresponding to many regular cristae. Treatment with EV-Aβ 40 or Aβ 40 alone resulted in a significantly lower crista score than in the control groups. This suggests that exposure to Aβ 40 results in the adoption of irregular shapes [ Figure 3H]. Circularity index is a relative measure of mitochondrial shape, with 1.0 representing a complete sphere and 0.0 representing an extreme elliptical or elongated shape. NPCs treated with EV-Aβ 40 tended to display a greater fluctuation between spherical and elongated extremes than in the control treatments [ Figure 3I].

Mitochondrial footprint alterations in NPCs exposed to EVs containing Aβ 40
To further explore the mitochondrial dynamic changes in NPCs exposed to EV-Aβ 40 or Aβ 40 alone, we next focused on morphological features of mitochondrial networks. TOM20, a translocase protein located on the outer mitochondrial membrane, was used to stain mitochondrial networks within treated NPCs [ Figure 4A]. Using MiNA, we quantified the networks using a skeletonized overlay approach. Mitochondrial footprint, which is the total area of the image with a signal, was reduced in both vesicle groups (EV-PBS and EV-Aβ 40 ) compared with that in untreated cells; however, this impact was more pronounced in NPCs exposed to EV-Aβ 40 [ Figure 4B]. Interestingly, the overall mitochondrial network number, a measure of networks comprising more than one branch, was unchanged among all groups [ Figure 4C]. While networks remained unchanged, network fusion, or the restricted ability of mitochondria to divide, was apparent in Aβ 40 -treated NPCs, indicating possible mitochondrial disruption. This was evident by enhanced fragmentation of larger networks in the Aβ 40 group, resulting in longer branch lengths, compared with fragmentation of slightly smaller relative networks in control NPCs [ Figure 4D]. Additionally, the presence of long branches [ Figure 4E] and a larger network size may indicate events leading to the occurrence of hyperfusion and the potential for many of the mitochondria to exhibit an elongated branched network. Hyperfusion can also be induced by stress or mitophagy and can decrease the ability of the cell to divide and differentiate properly. An increase in total overall branch length [ Figure 4E] may be related to this phenomenon when fission machinery is disrupted or inhibited as a compensatory mechanism. Overall, our mitochondrial tracking using TOM20 demonstrated that the exposure of NPCs to Aβ 40 (either alone or in the form of EV-Aβ 40 ) results in morphological changes to the mitochondria, as well as alterations in the organelle structure that lead to possible dynamic shifts within the cell.

EV-Aβ 40 contribute to the dysregulation of mitochondria by impairing metabolic respiration
After confirming differential regulation in mitochondrial footprint and branch length dynamics, we sought to evaluate whether these effects, combined with changes in the morphology of cristae, altered the respiratory capacity of NPCs. Healthy differentiating NPCs rely mainly on mitochondrial respiration and maintain glycolysis at low levels. Therefore, we analyzed whether EV-Aβ 40 treatment could alter mitochondrial respiration and subsequent ATP production using Seahorse technology. First, we assessed oxygen consumption rate (OCR) as an indicator of mitochondrial respiration [ Figure 5A and B]. Treatment with Aβ 40 , either alone or in the form of EV-Aβ 40 , resulted in a decrease in the compensatory rate of respiration compared with that of untreated NPCs or treatment with control EVs. In the subsequent analyses, OCR was measured following injections with oligomycin, to inhibit complex V (ATP synthase) and force the cell into compensatory mechanisms for ATP production; FCCP, to disrupt membrane potential and uncouple the proton gradient across the inner membrane; and rotenone/antimycin A, to block complexes I and III, which eliminates electron transport chain (ETC) mitochondrial respiration. NPCs treated with EV-Aβ 40 had significantly lower basal respiration [ Figure 5C], maximal respiration [ Figure 5D], and spare respiratory capacity [ Figure 5E] than untreated NPCs or EV-PBS-treated cells. These parameters measure the capacity of the cells to respond to energy demands and their ability to increase respiration. Proton leakage is indicated by the remaining basal respiration not coupled to ATP production, which showed no statistically significant differences among any of the treatment groups [ Figure 5F]. In contrast, both vesicle-treated groups (EV-PBS and EV-Aβ 40 ) appeared to exhibit an increased ability to respire by non-mitochondrial cellular enzymes that consume oxygen [ Figure 5G]. ATP production after treatment with EV-Aβ 40 exhibited a tendency to decrease compared with that in the controls [ Figure 5H]. We also assessed the extracellular acidification rate (ECAR) as a measure of cellular glycolysis. ECAR values demonstrated a tendency to decrease in NPCs exposed to EV-Aβ 40 or Aβ 40 alone relative to the levels in control groups [ Figure 5I].

EV-Aβ 40 dysregulate mitochondrial fusion and fission machinery
Mitochondrial morphology is crucial for cellular health and normal metabolic function. To better understand the main functions that regulate mitochondrial morphology, mitochondrial fusion and fission processes were assessed after treatment with EVs or Aβ 40 . Mitochondrial fusion is mainly regulated by proteins of the outer mitochondrial membrane, mitofusin 1 (Mfn1) and mitofusin 2 (Mfn2), and of the inner mitochondrial membrane, optic atrophy protein (OPA1). We assessed the expression of Mfn1, Mfn2, and OPA1 through mRNA and total protein expression. Although no significant change in total mRNA expression was found among the studied fusion-related genes [ Figure 6A], differential protein expression was observed. Specifically, there was a significant decrease in the expression of OPA1 and Mfn2 proteins in NPCs exposed to EV-Aβ 40 , while Mfn1 showed a modest but not statistically significant increase in the cells treated with Aβ 40 alone [ Figure 6B-D].
In addition to mitochondrial fusion, fission proteins are essential to mitochondrial maintenance. They include mitochondrial fission factor (Mff), dynamin-related protein 1 (Drp1), and mitochondrial fission protein 1 (Fis1). Similar to the results of the fusion genes, the expression of the fission genes showed no significant differences across the experimental groups [ Figure 6E]. Nevertheless, fission proteins were differentially expressed between the treatment groups. Fis1, a common receptor for Drp1 to promote mitochondrial division, was significantly decreased in both Aβ 40 groups (EV-Aβ 40 and Aβ 40 alone) [ Figure 6F]. In addition, Drp1 protein expression was found to be decreased in EV-Aβ 40 -treated NPCs [ Figure 6G]. Expression of Mff protein, which is an important regulator in the recruitment of Drp1 to the mitochondria, was also significantly reduced after treatment with Aβ 40 and tended to show lower levels in EV-Aβ 40 [ Figure 6H]. This overall decrease in fission regulators supports our findings in mitochondrial morphology that indicated a possible hyperfused state in which mitochondrial division is impaired, leading to a shift in the dynamic balance that results in cell stress and metabolic dysfunction.
Mitochondria are also key regulators of the apoptotic pathway. This involves Bcl-2, a modulator suppressing apoptosis and regulating cell death by mediating mitochondrial membrane permeability. To assess Bcl-2 levels, total mRNA levels were assessed by qPCR and showed a significant decrease in NPCs exposed to all of the treatments (EV-PBS, EV-Aβ 40 , or Aβ 40 alone) compared with that of the untreated control [ Figure 6I]. Consistent with this, total protein expression for Bcl-2 was significantly lower in all experimental groups than in untreated NPCs [ Figure 6J]. Taken together, our data suggest that an imbalance in normal fusion and fission processes occurs in NPCs exposed to EV-Aβ 40 or Aβ 40 alone. This imbalance may lead to the remodeling of cristae, as shown in Figure 3, and further apoptosis-associated responses in the mitochondria.

Diminished SIRT3 expression and activation of NFκB in NPCs exposed to EV-Aβ 40
To assess the downstream pathways that may alter mitochondrial dynamics in NPCs, we probed for sirtuin 3 (SIRT3) expression, a mitochondrially specific NAD-dependent deacetylase. One of the prominent targets of SIRT3-mediated deacetylation is NFκB [ Figure 7A]. SIRT3 gene expression was assessed by qPCR, which showed no significant differences across any of the treatment groups [ Figure 7B]. However, there was a marked decrease in SIRT3 protein expression in NPCs exposed to EV-Aβ 40 and Aβ 40 alone [ Figure  7C and D].
We then stained NPCs to examine the expression levels of p65, a subunit of the NFκB heterodimer that is needed for nuclear translocation [ Figure 7E]. We assessed the levels of NFκB in the nucleus as an indicator of activation and demonstrated that the group treated with EV-Aβ 40 had increased expression of NFκB in the nucleus compared with the levels in other treated groups, as quantified by total fluorescence intensity [ Figure 7F] and the relative intensity of nuclear staining [ Figure 7G]. We also assessed the ratio of p65 fluorescence in the nucleus compared to that in the cytoplasm [ Figure 7H]. This analysis indicated that NPCs treated with Aβ 40 alone exhibited the most significant translocation of NFκB into the nucleus, while the EV groups showed a trend towards higher nuclear-to-cytoplasmic translocation ratios. It is important to consider that the relative nuclear-to-cytoplasmic ratio in the EV-Aβ 40 group might be affected by aberrant neurogenesis that also occurs in this group.

DISCUSSION
Unbalanced oxidation and oxidative stress have been widely postulated to be common factors that contribute to pathological aging and aberrant neurogenesis of NPCs. Aβ 40 has also been shown to possibly impair the viability and differentiation of NPCs by disrupting mitochondrial signaling [31,32] . The significance of these events is related to the fact that dysfunctional mitochondria are the main inducers of oxidative stress and imbalance of the redox status, which are the critical factors in the stimulation of immune activation, neurotoxicity, and aberrant NPC proliferation [33][34][35][36][37] . Additionally, EVs have recently been postulated to be significantly involved in various neurodegenerative diseases, including Aβ pathology [16,17] . One study has demonstrated that Aβ 40 can be packaged into EVs and shed into the brain microenvironment, and that human Aβ 40 plaques are enriched with EV proteins [9] . The implications of endogenous Aβ 40 and Aβ 40 trafficked into the brain may translate into clinical manifestations of AD, such as dementia, and neurodegeneration. Because neurogenesis has been accepted to occur in the adult brain, it is also a crucial modulator of brain plasticity, repair, as well as learning and memory functions [24,38] . Neurogenesis in the adult brain occurs in distinct zones from NPCs that reside in several parts of the brain, including the subgranular zone (SGZ) and the dentate gyrus (DG) in the hippocampus [39][40][41] . Furthermore, the integrity of the brain endothelium is vital for normal self-renewal, proliferation, and differentiation of NPCs [42] . In this study, we investigated the role of brain endothelium-derived EV biology in the context of Aβ 40 pathology to understand the mechanistic dysfunction that may induce aberrant neurogenesis in NPCs.
The level of Aβ 40 in the CNS is dependent on the balance achieved between endogenous production in the brain, influx from the plasma, and efflux via BBB efflux transporters [43] . The relative contributions of these sources to amyloid levels in the brain are difficult to estimate, but they all appear to be critically important. Neurotoxic Aβ 40 is generated by proteolytic cleavage of the ubiquitously expressed amyloid precursor protein (APP) commonly expressed by neurons [44] . Amyloidogenic Aβ peptides range from 30 to 42 amino acids in length, with Aβ 40 and Aβ 42 , being the two main toxic Aβ species. Aβ 40 is the main amyloidogenic species, accounting for 90% of all amyloid protein produced. The doses of Aβ 40 used in this study have pathological relevance as the cerebrospinal fluid Aβ 40 levels in AD brains are in the nanomolar range [28] .
The first goal of the present study was to evaluate whether differentiation of NPCs could be altered after exposure to EVs containing Aβ 40 . We demonstrated that EV-Aβ 40 had the ability to alter neurite outgrowth and branching in developing NPCs, prompting us to hypothesize that mitochondrial metabolism was being affected. We also observed a similar decrease in neurite length in a human-induced stem cell-derived neuronal model [45] . Generally, mitochondria have a tightly regulated distribution within the cytoplasm of cells; however, when NPCs are rapidly differentiating and require high energetic turnover, the mitochondria are more commonly found in specific growth cones and extensions of these cells. This has broad implications in aiding ATP and calcium release, presynaptic terminal formation, and overall growth of developing neurite extensions. Therefore, we decided to evaluate the impact of EV-associated Aβ 40 on the mitochondrial morphology and function of NPCs. While we quantitatively confirmed a decrease in the number and density of cristae, we also noticed structural changes to the integrity of individual mitochondria, with areas showing a lack of crista structure or "voids." These areas with voids may directly correspond to losses in the number, density, and even score of cristae. This observation is significant because it has recently been demonstrated in many pathological studies revolving around organs with a high-energy demand, such as the brain and heart [46][47][48][49] . This has implications for cell function because it directly corresponds to decreases in mitochondrial membrane permeability, electron transport chain function, and ATP synthesis.
Next, we used mitochondrial network analysis as an alternative method to assess mitochondrial morphology. Interestingly, the mitochondrial footprint was decreased in EV-Aβ 40 -treated NPCs, while branch length of the networks increased, suggesting a tendency towards a hyperfused state with elongated branches. The presence of a hyperfused state in NPCs is controversial; however, our data suggest that the restricted ability of the organelle to properly divide supports the theory that diseased NPCs have abnormal mitochondria and may exhibit lower ATP production, affecting their potential for self-renewal or differentiation [50] .
The mitochondrial membrane is largely impermeable, and the transfer of molecules across it is mediated by channels, pumps, and transporters. Proper assembly of the inner mitochondrial membrane is crucial for maintaining the electron transport chain and ATP bioenergetic production. While NPCs normally maintain glycolysis at low levels, environmental stressors may cause a switch in cell machinery when metabolic respiration is compromised. Indeed, treatment of NPCs with EV-Aβ 40 and Aβ 40 alone resulted in a decrease in metabolic respiration. This led us to hypothesize that the metabolism of NPCs is affected by EVs containing Aβ 40 . With a decrease in metabolic respiration and the inability of Aβ 40 -exposed NPCs to switch to compensatory glycolysis, as measured by ECAR levels, we confirmed that overall metabolic regulation in the mitochondria was significantly hindered by EV-Aβ 40 .
The circularity index of a mitochondrion is a relative measure of its shape. Typically, more spherical mitochondria are associated with swelling, while elongated ones result from the dynamic fission and fusion of mitochondria that occur in healthy cells. In our study, NPCs treated with EV-Aβ 40 showed a tendency for greater fluctuation between spherical and elongated morphology, as well as a decreased respiratory rate. Mitochondrial maintenance is regulated by a balance between fusion and fission [51,52] and a shift in this balance may lead the mitochondria to induce mitophagy or cellular apoptosis [53] . The dysregulation of this process can have pathological consequences. This may be particularly important in cells that have high energy demands, such as NPCs, and may have driven the deficits in neurogenesis observed when these cells were exposed to Aβ 40 [54] . In general, mitochondrial fusion acts to ameliorate stressful conditions in cells by substituting damaged mitochondria with healthy ones. Meanwhile, mitochondrial fission is necessary to promote mitochondrial division and growth, but can also be used by the cell as a quality control mechanism under pathological conditions to remove damaged mitochondria [55] . Fission in the mitochondria is mainly mediated by the GTPase Drp1, which binds to its receptor Fis1 near the endoplasmic reticulum [56] . In contrast, mitochondrial fusion is mediated by mitofusin proteins, while the integrity of mitochondrial cristae is maintained by OPA1 [57][58][59] . Our data suggest that mitochondrial fusion machinery is dysregulated in NPCs upon exposure to EV-Aβ 40 and/or Aβ 40 alone. Indeed, these treatments decreased the expression of OPA1 and Mfn1, while Mfn2 expression was significantly increased. In contrast, the elements of the fission machinery Drp1 and its receptor, Fis1, were significantly downregulated upon treatment with EV-Aβ 40 , supporting our observation of a hyperfused state. Nevertheless, this could also be a compensatory mechanism achieved by cells due to the loss of the integrity of cristae observed in the TEM and metabolic analyses.
Given the irregular mitochondrial dynamics observed in the Aβ 40 -treated cells, we assessed the NPCs for potential engagement in the mitochondrial apoptotic pathway with a focus on Bcl-2 levels. The loss in OPA1 protein expression [ Figure 6A] suggests a loss in inner mitochondrial membrane (IMM) fusion and crista structure [ Figure 3]. OPA1 is a critical protein for IMM fusion and for preserving crista structure and sequestering cytochrome c in the intermembrane space. A loss of this protein may lead to release of cytochrome c and reduced density of mitochondrial cristae, resulting in white patches or voids in TEM [60] . Our results also reveal that a significant decrease in OPA1 levels was associated with a decreased level of Bcl-2, a mitochondrially specific pro-apoptotic inhibitor [ Figure 6I-K], an event that can stimulate the release of cytochrome c, which further triggers apoptotic cascades [61] .
SIRT3 is a mitochondrially specific NAD-dependent deacetylase that is an essential part of mitochondrial maintenance. Targets of SIRT3 include the electron transport chain, enzymes of intermediary metabolism, and antioxidant defense [62][63][64] . Although functions of SIRT3 in peripheral metabolic regulation have been studied, its action in the brain is not fully understood. One important function of SIRT3 is the deacetylation of NFκB. Thus, decreased activity of SIRT3, as seen in Figure 7A-C, may cause activation of NFκB (acetylated NFκB is active) and induction of downstream inflammatory reactions. In addition, SIRT3 expression is transcriptionally regulated by NFκB [65] . Indeed, our data indicate that a decrease in SIRT3 protein expression was associated with the translocation of NFκB into the nucleus in the groups with EV-Aβ 40 and Aβ 40 alone. This has implications for future studies on inflammatory and signaling pathways that mediate aberrant neurogenesis.
In summary, the results of the present study indicate that EVs derived from brain endothelial cells can effectively transfer Aβ 40 to NPCs and affect their neurogenesis via alterations of mitochondrial functions. These events have the potential to modulate the dynamics and severity of AD. Therefore, a better understanding of the interactions between Aβ 40 and the BBB, and the underlying mechanisms that mediate amyloid accumulation in the brain parenchyma may provide targets for pharmacological intervention. Such interventions may facilitate Aβ removal or protect against entry across the BBB and the development of neurological, neurodevelopmental, and neurocognitive alterations.   Quantification of (E) crista number, (F) crista surface area, (G) crista density, (H) crista score, and (I) circularity index. Scale bar for representative images is 2 μm and zoom inserts are 800 nm. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. n = 30 images per Analysis of the mitochondrial network skeletons in NPCs using MiNA. (A) Confocal images of NPCs stained with DAPI (blue) and TOM20 (green) for tracking the mitochondrial membrane. MiNA plug-in on ImageJ was used to skeletonize the mitochondria. Box and whisker plots quantify the (B) mitochondrial footprint, (C) total mitochondrial networks, (D) mean branch length, and (E) total branch length. Box plots show the mean (horizontal line), first to third quartile (box), and the extreme values within the 95% confidence interval of the interquartile range (whiskers). Scale bar for representative images is 20 μm. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, using one-way ANOVA. n = 6-9 images per group, each image containing 3-12 NPCs. EV: Extracellular vesicles; NPCs: neural progenitor cells; MiNA: mitochondrial network analysis; DAPI: 4′,6-diamidino-2phenylindole. Bcl-2 was probed for (I) total mRNA gene expression and (J) total protein expression by western blotting. Bar graphs indicate mean ± SEM. Statistical analysis was performed using one-way ANOVA, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, n = 3-6 independent experiments per group. EV: Extracellular vesicles; NPC: neural progenitor cell; SEM: standard error of the mean. Expression of SIRT3 and NFκB is affected upon exposure to Aβ 40 . (A) Schematic of mitochondrially dependent SIRT3 function affecting NFκB activation in NPCs. Reduced SIRT3 expression inhibits the deacetylation of NFκB, which allows the acetylated form to translocate into the nucleus and induce downstream inflammatory responses. SIRT3 expression was probed for both (B) total mRNA expression and (C) total protein expression.