Abstract
A promising therapeutic intervention for preventing the onset and progression of Alzheimer's disease is to protect and improve synaptic resilience, a well-established early vulnerability associated with the toxic effects of oligomers of amyloid β (AβO) and Tau (TauO). We have previously reported that exosomes from hippocampal neural stem cells (NSCs) protect synapses against AβO. Here, we demonstrate how exosomes can also shield against TauO toxicity in adult mice synapses, potentially benefiting primary and secondary tauopathies. Exosomes from hippocampal NSCs (NSCexo) or mature neurons (MNexo) were delivered intracerebroventricularly to adult wild-type male mice (C57Bl6/J). After 24 h, TauO were administered to suppress long-term potentiation (LTP) and memory, measured by electrophysiology and contextual memory deficits measured using novel object recognition test. We also assessed TauO binding to synapses using isolated synaptosomes and cultured hippocampal neurons. Furthermore, mimics of select miRNAs present in NSCexo were delivered intracerebroventricularly to mice prior to assessment of TauO-induced suppression of hippocampal LTP. Our results showed that NSC-, not MN-, derived exosomes, prevented TauO-induced memory impairment, LTP suppression, and reduced Tau accumulation and TauO internalization in synaptosomes. These findings suggest that NSC-derived exosomes can protect against synaptic dysfunction and memory deficits induced by both AβO and TauO, offering a novel therapeutic strategy for multiple neurodegenerative states.
Significance Statement
Exosomes from hippocampal NSCs provide an unprecedented therapeutic strategy targeting an early vulnerability driven by amyloidogenic toxic oligomers associated with multiple neurodegenerative states.
Introduction
Alzheimer's disease (AD) stands as the most prevalent form of age-related neurodegenerative dementia in the United States, with projections indicating that the number of affected individuals will more than double by 2050. Despite ongoing efforts, effective treatments for AD remain elusive, prompting a consensus that successful therapies should target prevention and/or minimization of symptoms prior to the onset of irreversible neuronal loss that characterizes the disease's advanced stages (https://www.alz.org/media/documents/alzheimers-facts-and-figures.pdf).
Over the last two decades, AD research has been significantly shaped by the identification of individuals who maintained cognitive function despite advanced AD pathology, termed non-demented with AD neuropathology (NDAN; Zolochevska and Taglialatela, 2016). This discovery highlights the brain's potential to resist or delay the neurotoxic processes that typically lead to cognitive decline in AD. Notably, our previous research uncovered distinctive proteomic profiles in hippocampal synapses of NDAN subjects (Zolochevska et al., 2018), along with resistance to harmful amyloid β oligomers (AβO; Bjorklund et al., 2012). As synaptic dysfunction, induced by AβO and Tau oligomers (TauO), emerges as an early event in AD progression (Klein, 2013; Spires-Jones and Hyman, 2014; Tu et al., 2014; Fa et al., 2016; Selkoe and Hardy, 2016), preserving synaptic integrity, as observed in NDAN individuals, holds promise as a therapeutic approach, albeit challenging to achieve.
New neurons are generated from neural stem cells (NSCs) via a process known as neurogenesis (Altman and Das, 1965). Neurogenesis occurs in the mammalian brain, in the hippocampus, throughout life and is known to contribute to brain plasticity and repair (Altman and Das, 1965). Extensive evidence indicates that hippocampal neurogenesis decreases significantly more in individuals with AD than in those experiencing normal aging (deToledo-Morrell et al., 2007; Kuhn et al., 2007; Ohm, 2007; Appel et al., 2009; Barnes et al., 2009; Mu and Gage, 2011; Hamilton et al., 2013; Horgusluoglu et al., 2017). Our research revealed that the brains of NDAN individuals exhibit greater numbers of hippocampal NSCs compared with the brains of healthy age-matched controls (Briley et al., 2016). This suggests that maintaining robust neurogenesis could mitigate cognitive decline in the presence of AD pathology.
NSCs release exosomes (Marzesco et al., 2005; Stevanato et al., 2016), small vesicles containing various molecular components, including microRNAs (miRNAs; Raposo and Stoorvogel, 2013; Sato-Kuwabara et al., 2015), which are increasingly recognized for their role in mediating the functional effects of NSCs (Baulch et al., 2016; Han et al., 2016; Zhang et al., 2017). These exosomes have been implicated in controlling aging processes (Zhang et al., 2017) and ameliorating synaptic dysfunction and cognitive decline in AD models (Cui et al., 2018). Moreover, a meta-analysis reveals a connection between exosomes, synaptic plasticity, and neurodegenerative diseases, including AD (Wang et al., 2017).
Our previous research indicated that hippocampal NSC-derived exosomes rendered central nervous system synapses resistant to AβO by delivering specific miRNA cargo (Micci et al., 2019). Additionally, we found that compared with individuals with AD, the brains of NDAN subjects exhibited increased numbers of NSCs in the hippocampus. Since the number of NSCs, but not neurons, positively correlated with cognitive function, we hypothesized that NSCs may support cognitive resilience through mechanisms other than neurogenesis (Briley et al., 2016).
Building on these findings, here we investigate whether NSC-secreted exosomes (NSCexo) prevent functional synaptic deficits and memory impairments driven by TauO and whether this protection involves enhanced synaptic resistance to TauO binding. Moreover, given that miRNAs are a predominant cargo of exosomes, we aimed to identify specific miRNAs found exclusively or highly enriched in exosomes derived from NSCs that may mediate their protective effects on target neuronal synapses.
Materials and Methods
Animals
Male mice (C57BL/6J; 6–8 weeks old) were purchased from The Jackson Laboratory. The Institutional Animal Care and Use Committee of the University of Texas Medical Branch approved all animal experiments, which were conducted according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The mice were housed four per cage on a 12/12 h light/dark cycle with ad libitum access to food and water.
Randomization and blinding procedures
Animals were randomized to the experimental groups using a simple randomization procedure. To maximize scientific rigor, we achieved a simple blinding procedure by coding exosome preparations (NSCexo or MNexo), vehicle, or the miRNAs to be employed in the various experiments. The correspondence of the codes was known only to Guilio Taglialatela, PhD, and Maria-Adelaide Micci, PhD, and was revealed only after final statistical analyses were performed.
NSC and mature neuron (MN) cultures
Adult rat hippocampal NSCs were purchased from MilliporeSigma. NSCs were cultured as neurospheres on low-attachment plates in expansion media consisting of EmbryoMax DMEM/F12 containing l-glutamine without HEPES, B27 with retinoic acid, GlutaMAX (all from Invitrogen, Thermo Fisher Scientific), FGF-b (20 ng/ml, MilliporeSigma), and antibiotics (PSF, Invitrogen, Thermo Fisher Scientific). Cells were passaged every 5–7 d using NeuroPapain (Genlantis).
For the generation of MNs, NSC neurospheres were dissociated using NeuroPapain (Genlantis), 2 mg/ml in basal media, plated out onto poly-ornithine/laminin-coated plates at a density of 8.0 × 105 cells/cm2 and cultured in differentiation media consisting of expansion media without FGF-b and with the addition of 1 µM retinoic acid (MilliporeSigma) and 5 µM forskolin (MilliporeSigma) for 5 d. Immunocytochemistry and Western blot analyses confirmed the purity of the cultures by identifying the presence or absence of specific NSCs (Sox2 and nestin) and neuronal markers (βIII-tubulin and NeuN) as previously reported (Micci et al., 2019).
TauO preparation
Prepared recombinant TauO were provided by the laboratory of Rakez Kayed, PhD. They were produced and characterized following established and published protocols (Gerson et al., 2017; Sengupta et al., 2018) and labeled TauO488 using the Microscale Protein Labeling Kit from Thermo Fisher Scientific (catalog #A30006; lot #2413458) following the manufacturer’s instructions. Briefly, 100 µl of TauO 1 mg/ml was mixed with 1/10 volume of 1 M sodium bicarbonate. Then, 11.3 µl of the Alexa Fluor 488 TFP ester was added following the kit molar ratio recommendations and incubated for 15 min at room temperature (RT). The labeled TauO were purified by centrifugation at 16,000 × g in spin filters with yields between 60 and 90% and quantified spectrophotometrically using a NanoDrop 2000C (Thermo Fisher Scientific) to be used for flow cytometry analyses. The determined degree of labeling was higher than two.
Isolation and characterization of exosomes
Exosomes were isolated from conditioned culture media using the ultracentrifugation method (Greening et al., 2015; Lobb et al., 2015). Comprehensive characterization of isolated exosomes was performed using electron microscopy (EM), dot blotting, and nanoparticle tracking analyses (Fig. 1) according to the guidelines of the International Society of Extracellular Vesicles (Lotvall et al., 2014; Witwer et al., 2021) and as previously reported (Micci et al., 2019). Briefly, 225 ml of conditioned media collected from ∼60 million cultured cells (NSC or MNs) were centrifuged at 2,000 × g, at 4°C for 10 min to remove cells and debris. The resulting supernatant was transferred to new tubes and centrifuged at 10,000 × g, at 4°C for 10 min. The supernatant was transferred to Beckman 60 Ti ultracentrifuge tubes and centrifuged at 100,000 × g for 3 h at 4°C. The resulting pellet was resuspended in 2 ml PBS containing a protease inhibitor cocktail (MilliporeSigma) and centrifuged at 182,000 × g for 1 h at 4°C in a Beckman TLA110 centrifuge. The resulting pellet, containing exosomes, was resuspended in 1× PBS containing protease inhibitors to a concentration of ∼109 exosomes per microliter.
Characterization of isolated exosomes. Schematic representation of the methods used to characterize the exosome preparations. Left, Representative high-resolution transmission EM image of NSCexo and MNexo, isolated using the ultracentrifugation method (scale bar, 0.1 μm). Middle, Representative NTA profile for NSCexo and MNexo showing the size and concentration of particles. Right, Representative dot blot comparing exosome markers expression in NSCexo, MNexo, and total homogenate from mouse parietal cortex (mPC-H). NSCexo, neural stem cells exosomes; MNexo, mature neurons exosomes; NTA, nano tracking analysis.
Ultrastructure analysis of exosomes was performed with 5 μl drops adsorbed on a 200 mesh coated resin grid (FCF 200–CU Formvar/Carbon, Electron Microscopy Sciences) for 10 min at RT. Grids were blotted with filter paper and stained with 2% aqueous uranyl acetate (catalog #541-09-3, Electron Microscopy Sciences) for negative staining for 1 min at RT. The uranyl acetate was then removed using filter paper, and the grids were dried with warm regular light for 2 min. Images were acquired with a Philips CM-100 transmission electron microscope at 60 kV with an Orius SC2001 digital camera (Gatan). For the dot blot, NSCexo, MNexo, and mouse parietal cortex were lysed with RIPA buffer containing protease inhibitors, and 0.5 µg of each sample was dotted on an S nitrocellulose membrane. The membrane was let dry at RT for 30 min, blocked with Intercept (TBS) Blocking Buffer (LI-COR Biosciences) for 1 h at RT and probed overnight at 4°C with primary antibodies against CD9, CD63, CD81, Hsp70, and GM130 (1:1,000; EXOABSBI Biotechnology), followed by IR Dye secondary antibodies (1:1,000; LI-COR Biosciences). Nanoparticle tracking analysis was performed using the NanoSight N300 system and NTA 2.1 operating system (Malvern Panalytical) according to manufacturer's instructions. Briefly, the exosome preparation was diluted in PBS and injected into the NanoSight for analysis of both particle size and concentration.
Intracerebroventricular injections
Adult (6–8-week-old) male C57BL/6J mice were anesthetized with isoflurane and subjected to intracerebroventricular injections using the freehand injection method (Clark et al., 1968; Dineley et al., 2010; Krishnan et al., 2018). One day after intracerebroventricular injection of exosomes (prepared from NSC or MN conditioned media) or PBS (vehicle), mice were killed for the electrophysiology study or injected intracerebroventricularly with 3 μl of 0.55 mM TauO (prepared as described above) or ACSF and killed 24 h later (for biochemistry studies) or 48 h later (after behavioral testing) by deep usoflurane anesthesia followed by decapitation. Brains were removed and divided into two groups: (1) fresh hippocampal slices were prepared for assessment of LTP, or (2) brains were further dissected into the hippocampus, frontal cortex, parieto-occipital cortex, and midbrain, snap frozen on dry ice, and stored at −80°C until ready to use for the synaptosome preparation.
Brain slice preparation and electrophysiological assessment of long-term potentiation (LTP)
Brain slices were prepared and electrophysiological assessment of LTP was performed according to the methods reported by Ting et al. (2014). Mice were deeply anesthetized with isoflurane and transcardially perfused with 25−30 ml of RT carbogenated gas mixture (95% O2 and 5% CO2) and NMDG ACSF consisting of the following (in mM): 93 NMDG, 2.5 KCl, 1.25 NaH2PO4, 30 NaHCO3, 20 HEPES, 25 glucose, 2 thiourea, 5 Na-ascorbate, 3 Na-pyruvate, 0.5 CaCl2·2H2O, and 10 MgSO4·7H2O, titrated to pH 7.3–7.4 with HCl.
Following perfusion, brains were carefully extracted within 1 min. Each brain was mounted using Super Glue, on the mounting cylinder, at a transverse slicing angle, to generate 350 µM sections using the Compresstome VF-300 (Precisionary Instruments) protocols. The sliced hippocampi were allowed to recover from the procedure at 32–34°C for <12 min. The slices were then transferred to a new holding chamber with carbogen-saturated HEPES ACSF recovery solution (in mM: 92 NaCl, 2.5 KCl, 1.2 NaH2PO4, 30 NaHCO3, 20 HEPES, 25 glucose, 2 thiourea, 5 Na-ascorbate, 3 Na-pyruvate, 2 CaCl2·2H2O, and 2 MgSO4·7H2O) at pH 7.3–7.4, RT.
Recording was done in constantly flowing oxygenated (95% O2/5% CO2) ice-cold normal ACSF consisting of the following (in mM): 124 NaCl, 2.5 KCl, 1.2 NaHPO4, 24 NaHCO3, 5 HEPES, 13 glucose, 2 CaCl2·2H2O, and 1.2 Mg SO4·7H2O, at pH 7.3-7.4. Slices were perfused at a rate of ∼3 mL/min. Field excitatory postsynaptic potentials (fEPSPs) were collected using an Axon MultiClamp 700B amplifier connected to a Windows computer-running Clampex 8.2 software (Molecular Devices). All electrodes were placed under the visual guidance of an upright microscope (Olympus BX51WI). The slope from a single fEPSP trace was calculated from the initial slope of the fEPSP relative to the slope of the 10 ms interval immediately preceding afferent stimulation. The current magnitude was delivered through a digital stimulus isolation amplifier (AMPI) and set to elicit an fEPSP of ∼30% of the maximum for synaptic potentiation experiments using platinum–iridium-tipped concentric bipolar electrodes (SUK 30200, FHC). Using a horizontal P-97 Flaming/Brown Micropipette puller (Sutter Instruments), borosilicate glass capillaries (catalog #BF150-110-10, Sutter Instruments) were used to pull electrodes and filled with ACSF to get a resistance of 1–2 MΩ.
A stable baseline (for a minimum of 10 m) was obtained by delivering a single pulse stimulation at 20 s interstimulus intervals. fEPSPs in the CA1 were evoked by stimulating the Schaffer collaterals (SC; CA3→CA1) using a conditioning stimulus (CS) consisting of three trains of 100 pulses at 100 Hz, 20 s apart (high-frequency stimulation; HFS). Input–output experiments were conducted to measure basal dendritic excitation in response to increasing applied current in ACSF. Evoked fEPSP responses were digitized via Digidata 1550B, and the initial slope of the fEPSP was analyzed using pClamp 10.6 software (Molecular Devices). All data are represented as percentage change from the initial average fEPSP slope, defined as the average slope obtained for the 10 min prior to CS application.
Intracerebroventricular injection technique
Intracerebroventricular injections were performed on deeply anesthetized mice using a 29 gauge needle, firmly held in place using hemostatic forceps to leave 4.5 mm of the needle tip exposed and connected to a 25 µl Hamilton syringe via 0.38 mm polyethylene tubing. Infusions were performed at the rate of 2 µl/min for a total volume of 3 μl, using an electronic programmable microinfuser (Harvard Apparatus). After intracerebroventricular injection, the needle was left in place for 2 min, and the mouse was allowed to recover while lying on a heated pad under warm light.
Novel object recognition testing
The novel object recognition (NOR) test was performed as described previously (Taglialatela et al., 2009; Comerota et al., 2017; Krishnan et al., 2018). Each mouse was habituated to an empty NOR open-field box that served to test each animal for normal locomotion. The sessions commenced with two 10 min habituation sessions spaced 24 h apart, during which the TopScan (Clever Sys) video-tracking software quantified various locomotor parameters. Twenty-four hours after the last habituation session, mice were subjected to 10 min training sessions consisting of exposure to two identical, nontoxic objects (metal or hard plastic items) in the same open-field box to which the mouse had previously been acclimated, so that the mice would be more likely to maximize time exploring the objects as opposed to the environment.
The time spent exploring each object was recorded using ObjectScan (Clever Sys). The time spent in each quadrant zone, and object zone surrounding each object, was also recorded. After the training session, the animal was returned to its home cage. Mice were returned once more to the same box after retention intervals of either 2 or 24 h, for the test session. One object identical to the familiar one but previously unused (to prevent olfactory cues) and one completely unfamiliar, hence, novel, object would be in the box. The animal was allowed to explore for 10 min, during which the amount of time spent exploring each object was recorded.
Objects were randomized and counterbalanced across animals. The ratio of the time spent exploring the novel object versus the time spent exploring the familiar object was reported as the discrimination index (Buenz et al., 2009; Bussian et al., 2018). An index above 1 is indicative of recognition that an object is novel. Each mouse was tested at 2 and 24 h intervals, with the intention of assessing the shorter and longer time frames in memory recall. To avoid the experience in the 2 h test affecting the performance in the 24 h test, we used different novel objects for the two memory recall tests.
Tissue collection and processing
Animals were deeply anesthetized using isoflurane and transcardially perfused with PBS. Brains were carefully removed and then cut in two halves along the sagittal line. One hemisphere was collected for biochemistry assays. The other hemisphere was postfixed in 4% paraformaldehyde in 0.1 M PBS, pH 7.4, for 48 h at 4°C and cryoprotected by suspension in 30% sucrose solution for 48 h at 4°C. Brains were then embedded in OCT compound (Tissue-Tek) and frozen on dry ice prior to storage at −80°C. For immunofluorescence experiments, mouse brain blocks were removed from storage at −80°C, equilibrated at −20°C, and sectioned at 12 μm onto Superfrost Plus slides (catalog #12-550-15, Fisherbrand, Thermo Fisher Scientific). Prepared slides were stored at −80°C.
Immunofluorescence staining
Slides were processed as previously described (Fracassi et al., 2023). Briefly, slides were fixed in 4% paraformaldehyde in 0.1 M PBS, pH 7.4, for 30 min at RT. Nonspecific binding sites were blocked with 5% bovine serum albumin (catalog #A4503-100G, Sigma-Aldrich)/10% normal goat serum (NGS; catalog #S26-100ml Sigma-Aldrich), and sections were permeabilized with 0.5% Triton X-100/0.05% Tween-20 for 1 h at RT. Slides were incubated overnight at 4°C with the following primary antibodies diluted in PBS containing 1.5% NGS: chicken anti-Tau (1:100, #ab75714, Abcam–Cambridge Science Park); mouse anti-phospho–Tau (Ser202/Thr205-AT8) (1:100, MN1020, Thermo Fisher Scientific); and rabbit anti-NeuN (1:200, #12943S Cell Signaling Technology). Slides were washed in PBS before incubation with the appropriate Alexa Fluor-conjugated secondary antibodies (goat anti-chicken Alexa Fluor 594, 1:400, #A-11042; goat anti-mouse Alexa Fluor 488, #A-10680, goat anti-rabbit Alexa Fluor 647, 1:400, #A32733, Thermo Fisher Scientific) in PBS containing 1.5% NGS/0.25% Triton X-100 for 1 h at RT. Finally, slides were washed in PBS, treated with 0.3% Sudan Black B prepared in 70% EtOH for 10 min to block lipofuscin autofluorescence, washed again with deionized water, and coverslipped using Fluoromount-G containing 4′,6-diamidino-2-phenylindole (DAPI; catalog #0100-20, SouthernBiotech) and sealed.
Quantitative analysis of immunofluorescence
All immunoreacted sections were acquired with a Keyence BZ-X800 (Keyence) microscope by using immersion oil 60×. Three sections were analyzed for each animal, and at least two images per section were taken at 1,920 × 1,440 pixel resolution, with the z-step size of 2 μm at 12 μm thickness. We analyzed five different regions: dentate gyrus (DG), CA1, CA3, and frontal and parietal cortex. For the feasibility of the quantification, all layers from a single image stack were projected on a single slice (stack/Z projection). Quantitative analyses were performed using the ImageJ software (https://imagej.nih.gov/ij, NIH). We analyzed the intensity of fluorescence for each marker (Tau and posphoTau AT8) “per” area (integrated density) when the overall distribution of a specific marker was studied. The colocalization between Tau and either NeuN or Iba1 was evaluated and quantified using Pearson's correlation coefficient. Representative images were composed in an Adobe Photoshop CC2020 format.
Synaptosome isolation
We isolated synaptosomes from the hippocampi of mice previously injected intracerebroventricularly with PBS, NSCexo, or MNexo. The synaptosomal fraction containing both pre- and postsynaptic components was isolated using a well-established method developed in our laboratory (Franklin and Taglialatela, 2016; Comerota et al., 2017; Franklin et al., 2019; Marcatti et al., 2022). The snap-frozen hippocampus from mouse brains was lysed using Syn-PER lysis buffer (Thermo Fisher Scientific) with 1% protease and phosphatase cocktail inhibitors, and the homogenates were centrifuged at 1,200 × g for 10 min at 4°C. The supernatants (containing the synaptosomes) were collected and centrifuged at 15,000 × g for 20 min at 4°C. The synaptosomal pellets were resuspended in HEPES-buffered Krebs-like (HBK) buffer (in mM: 143.3 NaCl, 4.75 KCl, 1.2 MgSO4·7 H2O, 1.2 CaCl2, 20.1 HEPES, 0.1 NaH2PO4, and 10.3 d-glucose), pH 7.4. Finally, 0.5% of Pluronic F-68 nonionic surfactant (catalog #24040-032, lot #2275337; Thermo Fisher Scientific) was added to prevent synaptosome aggregation.
The quality and concentration (synaptosomes/µl) of isolated synaptosomes were routinely verified by flow cytometry, EM, and Western blot as previously reported (Franklin and Taglialatela, 2016; Micci et al., 2019; Marcatti et al., 2022).
Enzyme-linked immunosorbent assay of total Tau
Synaptosomes isolated from the hippocampi of mice injected with PBS, NSCexo, or MNexo (intracerebroventricularly) were challenged with TauO (5 nM) and then digested with proteinase K (PK). Total Tau content was measured by commercial Tau enzyme-linked immunosorbent assay (ELISA) kits according to the manufacturer's instructions (catalog #KHB0041; Thermo Fisher Scientific). Ten million synaptosomes were incubated for 1 h at RT with 5 nM of TauO. Half of synaptosomes were incubated with 1 mg/ml of PK (catalog #70663-4, lot #3018798; EMD Millipore) for 30 min at 37°C [1 mg of PK is the equivalent of 30 mAU, where AU is an Anson unit that represents the amount of enzyme that liberates 1.0 μmol (181 μg) of tyrosine from casein per minute at pH 7.5 at 37°C]. The other half was subjected to the same conditions to serve as control. Synaptosomes were then washed three times with HBK buffer and lysate in 10 µl of Cell Extraction Buffer (catalog #FNN0011; Thermo Fisher Scientific) containing 0.3 M of PMSF and 1% protease cocktail inhibitors. Finally, standards and samples were applied to the ELISA plate. After washing, a biotin-conjugated detection antibody was applied. The positive reaction was enhanced with streptavidin–horseradish peroxidase and colored by 3,3′,5,5′-tetramethylbenzidine. The absorbance at 450 nm was then measured, and the concentration of Tau protein was calculated from the standard curve.
TauO binding challenge to synaptosomes and PK digestion
Synaptosomes were treated with TauO for binding challenges, and the binding percentages and median fluorescence intensity (MFI) were evaluated with flow cytometry. The same number of synaptosomes isolated from each mouse of the three experimental groups was pooled together. We incubated two million synaptosomes for 1 h at RT without oligomers (control) as well as with TauO tagged as described above at concentrations of 0, 0.025, 0.05, 0.25, 0.5, 1 μM. Following the challenge experiments, synaptosomes were digested with 1 mg/ml of PK (catalog #70663-4, lot #3018798; EMD Millipore) for 30 min at 37°C [1 mg of PK is the equivalent of 30 mAU, where AU is an Anson unit that represents the amount of enzyme that liberates 1.0 μmol (181 μg) of tyrosine from casein per minute at pH 7.5 and 37°C]. Synaptosomes were then pelleted, washed three times with HBK buffer, and resuspended in HBK. Oligomer fluorescence positivity was acquired by a Guava EasyCyte 8 flow cytometer (EMD Millipore) and analyzed using the InCyte software (EMD Millipore).
TauO binding to hippocampal neurons in vitro
Hippocampal neurons were prepared by differentiating adult hippocampal NSC as described above and plated on poly-ornithine/laminin-coated plates. Cultures were treated with exosomes (NSCexo or MNexo; 1 × 106) or an equivalent volume of PBS for 24 h at 37°C. After washing, cultures were exposed to 2.5 μM TauO for 60 min at 37°C. One separate set of cells was treated with PBS for 24 h, followed by a vehicle to control for nonspecific staining of the anti-hTau antibody. At the end of incubation, the cells were washed and fixed in 4% paraformaldehyde for 15 min. After two washes in PBS, the cells were blocked and permeabilized in PBS containing 5% NGS for 30 min at RT. The cells were incubated with mouse anti-βIII–tubulin primary antibody (1:1,000, Promega) and Tau5 antibody (1:100, Thermo Fisher Scientific), diluted in 1.5% NGS in PBS overnight at 4°C in a humid chamber. Following three washes in PBS, the cells were incubated with Alexa Fluor 488-conjugated anti-mouse and Alexa Fluor 594-conjugated anti-rabbit secondary antibodies (1:400; Invitrogen) in 1.5% NGS in PBS for 1 h at RT in a humidified chamber. The cells were washed in PBS and coverslipped with Prolong Gold AntiFade Reagent with DAPI (Thermo Fisher Scientific). Images were acquired with an Olympus confocal microscope (FV1200, Olympus Life Science), and quantification of TauO was performed by an investigator blinded to the experimental groups by counting the number of fluorescent puncta in dendrites using the ImageJ software.
Intracerebroventricular injection of miRNA mimics and target engagement validation
Mice were injected (intracerebroventricularly) with miRNA mimics (scrambled, miR-17, miR-322, miR-485; Thermo Fisher Scientific) dissolved in ACSF. Scrambled miRNA was delivered at 1 nmole per mouse, while 0.33 nmole of miR-17, miR-322, and miR-485 were mixed, and a final concentration of 1 nmole of miRNA per mouse was administered. Four animals per group were used. Twenty-four hours after intracerebroventricular injection, RNA was extracted from the hippocampi of intracerebroventricularly injected mice using the RNeasy Mini Kit (Qiagen), and cDNA was generated with the amfiRivert cDNA Synthesis kit (GenDEPOT). Quantitative RT-qPCR was performed using specific sense and antisense primers for known targets (STAT3, SYN5X, HIFa3) in a 20 μl reaction volume containing 10 μl of KAPA SYBR FAST qPCR Master mix (Kapa Biosystems), 0.5 μl of 10 µmol/L primer stock, 1 μl cDNA, and 8 μl double-distilled H2O. Data were normalized to β-actin.
Statistical Analysis
The sample sizes were based on a two-sided power analysis performed on the experimental groups to allow us to detect a 10-unit difference in the most variable measurement at α = 0.05 and β = 0.8 based on the typical variance we have in our data. Data is expressed as mean ± SEM or SD. Analysis of variance (ANOVA) followed by multiple-comparison post hoc tests were performed using the GraphPad Prism software. Behavioral data (NOR) was analyzed using repeated two-way ANOVA with Dunnett correction for multiple-comparison correction. Immunofluorescence data were analyzed using ordinary one-way ANOVA with Tukey’s post hoc test and two-way ANOVA with Šídák's multiple-comparison tests or Tukey's multiple-comparison test where appropriate. For all statistical analyses, significance was defined at *p < 0.05.
Results
NSCexo prevent TauO-induced short–term memory deficits
Wild-type mice were injected (intracerebroventricularly) with ACSF, NSCexo, or MNexo (1 × 109 exosomes) 24 h prior to receiving intracerebroventricular injections of 3 μL of 0.55 mM TauO or PBS. The mice underwent NOR training 4 h after treatment, followed by memory recall tests at 2 h and 24 h after that. The treatment schematic is shown in Figure 2A. Throughout the training phase, animals across all groups exhibited equal exploration of the two identical objects, suggesting that the different treatments had no discernible impact on their inherent exploratory tendencies (Fig. 2B).
TauO-induced memory deficit is rescued by NSCexo. A, Schematic of the experimental design. Mice were injected (intracerebroventricularly) with either 1 × 109 exosomes (NSCexo or MNexo) or ACSF 24 h before injection (intracerebroventricularly) of 3 μl of 0.55 μM TauO or PBS. Four hours later, they were subjected to the NOR memory test. B, Discrimination index during the training phase when the mice are exposed to two identical objects. One-way ANOVA (F = 1.580) and Tukey’s multiple-comparison test. n = 7–10 mice/group. Data are mean ± SD. C, Box plot representation of the discrimination index calculated as the ratio between time spent with the novel object and the time spent with the familiar object during the 2 and 24 h memory recall tests. Two-way ANOVA of repeated-measure data with Dunnett multiple-comparison test versus training. *p < 0.05; **p < 0.01; ***p < 0.005. n = 7–10 mice/group. Data are min–max. NSCexo, neural stem cell exosomes; MNexo, mature neuronal exosomes; PBS, phosphate-buffered saline; TauO, Tau oligomers; ACSF, artificial cerebrospinal fluid; SD, standard deviation.
The intracerebroventricular injection of ACSF, NSCexo, or MNexo alone, akin to the PBS treatment, demonstrated no influence on the mice's capability to differentiate between the familiar and novel object after the 2 and 24 h testing intervals (Fig. 2C). However, mice administered with ACSF or MNexo, followed by intracerebroventricular injection of TauO, exhibited equivalent exploration time for both the familiar and novel objects during the 2 h memory recall test, indicating memory impairment. On the other hand, mice treated with NSCexo prior to receiving TauO spent significantly more time in the novel object area, revealing preserved memory function despite the administration of the toxic TauO. During the memory recall test at 24 h, mice in all four experimental groups spent significantly more time exploring the novel object, demonstrating integrity in a longer-term expression of the memory (Fig. 2C). Representative movement traces of the mice during the NOR testing are shown in Figure 3.
Representative movement traces of mice during the NOR testing. Representative movement traces of mice intracerebroventricularly injected with either PBS, NSCexo, or MNexo prior treatment with TauO or ACSF. Left panels show data from the training phase of the test; middle panels show data from the 2 h memory recall test; right panels show data from the 24 h memory recall test. N, novel object; PBS, phosphate-buffered saline; NSCexo, neural stem cells exosomes; MNexo, mature neurons exosomes; TauO, Tau oligomers; ACSF, artificial cerebrospinal fluid.
NSCexo protect the hippocampus from TauO-induced suppression of LTP expression via select miRNAs
To investigate the cellular underpinnings of NSCexo protective effects against TauO-induced memory impairments, we assessed LTP expression in hippocampal slices from mice injected with or without exosomes, then exposed to TauO using a well-documented protocol for TauO-mediated LTP inhibition (Dineley et al., 2010). Wild-type mice were injected (intracerebroventricularly) with NSCexo or MNexo (1 × 109 exosomes in 3 μl of PBS) 24 h before killing. Hippocampal slices were prepared and incubated in HEPES solution with or without preformed TauO (50 nM) for 1 h (Fig. 4A).
TauO-induced suppression of LTP expression in the hippocampus is abolished by intracerebroventricular injection of NSCexo. A, Schematic of the experimental design: Mice were administered intracerebroventricular injections of either NSCexo, MNexo, or PBS. After 24 h, they were killed. Brain sections from these mice were subsequently treated with or without TauO. Field recordings of LTP were then conducted in the SC region of the hippocampus. B,D, Field potential recording of LTP, compared with the percentage of the baseline, from brain slices pretreated with NSCexo but not with MNexo, in the presence or absence of TauO, revealed differences in the slope of EPSPs. C, For each experimental condition, the amplitude of fEPSP during the final 10 min (from the 50th to 60th minute post-HFS) was averaged. TauO significantly reduced LTP in the brain slice taken from mice injected with either PBS or MNexo. Conversely, this suppressive effect was not seen in slices from mice that received the NSCexo injection. PBS, n = 7; NSCexo, n = 7; and MNexo, n = 6 (1–4 slices per mouse). *p < 0.05. Two-way ANOVA followed by Tukey's multiple comparisons test. NSCexo, neural stem cells exosomes; MNexo, mature neurons exosomes; PBS, phosphate-buffered saline; TauO, Tau oligomers; LTP, long-term potentiation; fEPSP, functional excitatory postsynaptic potential.
Following a stable 10 min baseline, slices underwent HFS for 1 min, and LTP was recorded for 60 min. When comparing LTP expression after administering NSCexo or MNexo to the control treatment with PBS, there were no significant differences, with LTP expression being ∼161% of the baseline. When brain slices were treated with TauO, a clear reduction in LTP expression (to 131%) was observed in slices from animals previously treated with either PBS (131%) or MNexo (134%). However, slices from NSCexo-treated animals showed less reduction in LTP (155%; Fig. 4B). A quantitative analysis of LTP in the final 10 min of recording (Fig. 2C) confirmed a significant decline in HFS–LTP triggered by TauO in slices from animals treated with PBS and MNexo, but not in slices from animals treated with NSCexo. Thus, given their potential role in memory protection, NSCexo appears to diminish the vulnerability of hippocampal synapses to TauO-induced LTP disruption. We also observed a trend of reduction without significance in the synaptic strength attributed to the toxic effects of TauO on the slope and on the FV as seen with MNexo (Fig. 5).
Application of TauO reduces the slope, while application of MNexo affects the fiber volley. The input–output (IO) plotted as slope (mV/ms) on the Y-axis as a function of fiber volley (FV in mV) on the X-axis is provided prior to HFS (pre-, clear circles) and after (post-, filled circles) for (A) PBS treated, (B) NSCexo treated, (C) MNexo treated. The panel on the left is untreated, while the panel on the right is from TauO treated. Slopes for (D) PBS treated, (E) NSCexo treated, (F) MNexo treated and FV as a function of stimulation (μA) for (G) PBS treated, (H) NSCExo treated, (I) MNexo treated provide additional insight into the role of TauO and MNexo treatment on synaptic strength.
We have previously reported that specific miRNAs enriched in NSCexo as compared with MNexo (miR-17, miR-322, miR-485) render synapses physically resistant to the detrimental binding of AβO and functionally resilient to their disruptive action (Micci et al., 2019). To test whether these miRNAs can also protect the hippocampus against TauO-suppression of LTP, wild-type mice received intracerebroventricular injections of mimics for three miRNAs (miR-485, miR-17, and miR-322) or an equal amount of scrambled RNA as control. Mice were killed 24 h later, and hippocampal slices were prepared and treated for 1 h with preformed TauO (50 nM) or PBS (Fig. 6A). After a stable baseline of 10 min, slices were subjected to HFS for 1 min, and LTP was measured for 1 h after that. We found that TauO-driven reduction of LTP was absent in mice treated with miRNA mimics (Fig. 6B,C). Proper target engagement for the miRNAs was confirmed by RT-qPCR in a separate set of experiments, as previously reported (Micci et al., 2019).
TauO-induced LTP suppression in the hippocampus is abolished by mimics of miRNAs enriched in NSCexo. A, Schematic of the experimental design. MiRNAs (scrambled or mimics of miR-322, miR17, and miR-485) or ACSF (vehicle) were injected intracerebroventricularly into adult mice 24 h before killing. B, SC field recording of LTP (indicated as the percentage of the baseline in the slope of fEPSPs) was performed on brain slices in the presence of TauO (50 nM) or ACSF. TauO abolished LTP in ACSF-injected mice and in scrambled-treated mice but not in miRNAs combo-treated mice. C, Average for each condition of the fEPSP amplitude for the final 10 min (time points 50–60 min post-HFS). TauO significantly reduced LTP in brain slices from mice injected with vehicle or with scrambled miRNA, but not in brain slices from mice treated with combo miRNAs mimics. N = 4 mice/group (3 slices per mouse). **p < 0.01; ****p < 0.0001; one-way ANOVA followed by Dunnett multiple-comparison test. ACSF, artificial cerebrospinal fluid; ICV, intracerebroventricular; LTP, long-term potentiation; fEPSPs, functional excitatory postsynaptic potentials; TauO, Tau oligomers.
Intracerebroventricular delivery of NSC exosomes prevent neuronal accumulation of TauO in the hippocampus and cortex
At the end of NOR testing, mice were killed, perfused with PBS, and fixed in 4% paraformaldehyde. The brains were processed for Tau and pTau immunofluorescence analyses (refer to Fig. 1A for the experimental design). We found that significantly less Tau and pTau were present in the hippocampus (DG, CA1, and CA3) of mice that received NSCexo intracerebroventricularly as compared with PBS and MNexo-injected mice (Fig. 7). Moreover, we observed reduced levels of Tau and pTau in the parietal and frontal cortex of NSCexo-treated mice compared with those treated with MNexo and PBS (Fig. 8).
NSCexo reduce Tau and pTau accumulation in the hippocampus. Representative images and quantitative analyses of triple immunofluorescence staining for NeuN (magenta) in combination with total Tau (red) and pTau (AT8-green) revealed a significant decrease in Tau and pTau levels after treatment with NSCexo in DG, CA1, and CA3. Scale bar, 20 μm. Statistical analyses were made using one-way analysis of variance, following Tukey's multiple-comparison test. Values are expressed as the mean ± SD. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. NeuN, neuronal nuclei; NSCexo, neural stem cells exosomes; DG, dentate gyrus.
NSCexo reduce Tau and pTau accumulation in the cortex. Representative images and quantitative analyses of triple immunofluorescence staining for NeuN (magenta) in combination with total Tau (red) and pTau (AT8-green) revealed a significant decrease in Tau and pTau levels after treatment with NSCexo in both the parietal and frontal cortex. Scale bar, 20 μm. Statistical analyses were made using one-way analysis of variance, following Tukey's multiple-comparison test. Values are expressed as the mean ± SD. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. NeuN, neuronal nuclei; NSCexo, neural stem cells exosomes; SD, standard deviation.
NSC exosomes reduce the internalization of TauO into synaptosomes
Because of the observed protective effects of NSCexo on TauO-induced memory disruption and inhibition of LTP expression, we explored whether exosomes released by hippocampal NSCs could mitigate synaptic susceptibility to TauO. Synaptosomes were isolated from the hippocampus of mice previously injected (intracerebroventricularly) with PBS or exosomes (NSCexo or MNexo) and incubated for 1 h with preformed TauO (5 nM). After washing away excess Tau, synaptosomes were incubated for 30 min with PK, to remove surface-bound Tau, or vehicle. Internalized and total content of Tau was determined by ELISA (Fig. 9A, schematic). We found that the amount of total Tau (internalized and at the surface) in hippocampal synaptosomes was not significantly different between the experimental groups (Fig. 9B). However, after PK treatment, the amount of TauO measured in hippocampal synaptosomes isolated from mice treated with NSCexo was significantly reduced compared with synaptosomes from mice treated with PBS or MNexo (p < 0.01 NSCexo vs PBS, p < 0.05 NSCexo vs MNexo; Fig. 9C). This suggests that NSCexo reduces TauO internalization but not binding at the synaptosome surface.
NSCexo reduce hippocampal synaptic vulnerability to TauO. Exosomes (NSCexo or MNexo) or PBS were injected intracerebroventricularly 24 h before sacrifice. Brain synaptosomes were prepared from the hippocampus and challenged with preformed 5 nM TauO for 1 h followed by treatment with PK or vehicle for 30 min. After washing, the amount of TauO bound to synaptosomes was quantified by ELISA. A, Schematic of the experimental design; (B) quantification of total Tau; (C) quantification of internalized Tau after PK treatment. n = 8 mice/group (average of two independent experiments). *p < 0.05; **p < 0.01 versus PBS unpaired T test. Data are median. NSCexo, neural stem cells exosomes; MNexo, mature neurons exosomes; PBS, phosphate-buffered saline; TauO, Tau oligomers; PK, proteinase K.
To further determine the effect of NSCexo on TauO binding and internalization in the synapses, we incubated the synaptosomes with increasing concentrations of labeled TauO488 (25, 50, 250, 500, 1,000 nM) in the presence or absence of PK and analyzed by flow cytometry (Fig. 10A, schematic). Our results revealed a dose-dependent increase in both the percentage of TauO-positive synaptosomes (Fig. 10B) and the amount of TauO bound to synaptosomes expressed as MFI (Fig. 10D) across all experimental groups (PBS, NSCexo, and MNexo). However following PK treatment, a greater reduction in TauO positivity was observed in synaptosomes isolated from NSCexo-treated mice at the TauO 50 nM concentration compared with synaptosomes from MNexo-treated mice (Fig. 10C). Additionally, we observed a greater reduction in TauO MFI in NSCexo-treated mouse synaptosomes compared with those from PBS-treated mice at 25 and 250 nM TauO, as well as MNexo-treated mouse synaptosomes at 25 nM TauO (Fig. 10E). When we compared the outcomes of PK-digested samples with their respective control conditions (without PK) for each TauO concentration (Fig. 11), we found that synaptosomes from NSCexo-treated mice, exposed to 25 nM TauO concentration, exhibited a significantly lower percentage of TauO-positive synaptosomes after PK digestion compared with synaptosomes treated with PBS or MNexo, relative to their respective control groups (Fig. 11B). No differences between the groups were noted in the percentage of synaptosomes positive for TauO before and after PK treatment at higher TauO concentrations (50, 250, 500, 1,000 nM; Fig. 11C–F).
Effect of NSCexo on synaptic TauO binding and internalization. Exosomes (NSCexo or MNexo) or PBS were injected intracerebroventricularly 24 h before sacrifice. Synaptosomes were isolated from the hippocampus and challenged with preformed labeled TauO488 at increasing concentrations or PBS for 1 h. After washing, synaptosomes were treated with PK or vehicle for 30 min, washed and analyzed by flow cytometry. A, Schematic of the experimental design: B,C, The percentage of synaptosomes positive for TauO488 with and without PK treatment. D–E, MFI of TauO488-positive synaptosomes with and without PK. Comparisons were performed within each group between data in the presence or absence of PK. *p < 0.05 (multiple unpaired T test). Synaptosomes were pooled from eight individual mice per group. N = 3 flow cytometry runs. Data are mean ± SD. NSCexo, neural stem cells exosomes; MNexo, mature neurons exosomes; PBS, phosphate-buffered saline; TauO, Tau oligomers; PK, proteinase K; SD, standard deviation.
NSCexo reduce the internalization of TauO into synaptosomes. A, Schematic of the experimental design. B–F, Flow cytometric analysis of the percentage of TauO488-positive synaptosomes (after being challenged with increasing concentration of TauO) derived from NSCexo-, MNexo-, or PBS-treated mice before and after PK digestion. G–K, Flow cytometric analysis of the MFI of TauO488-positive synaptosomes derived from NSCexo-, MNexo-, or PBS-treated mice before and after PK digestion. *p < 0.05; two-way ANOVA followed by Tukey's multiple-comparison tests. Synaptosomes were pooled from eight individual mice per group. n = 3 flow cytometry runs. Data are mean ± SD. NSCexo, neural stem cells exosomes; MNexo, mature neurons exosomes; PBS, phosphate-buffered saline; TauO, Tau oligomers; PK, proteinase K.
Interestingly, when we compared the amount of TauO (MFI) before and after PK treatment, we found significantly lower levels of TauO after PK in synaptosomes from NSCexo-treated mice but not in those from PBS- and MNexo-treated mice at all Tau concentrations tested (Fig. 11G–K). These data suggest that synaptosomes isolated from PBS or MNexo-treated mice internalize most, if not all, TauO, while synaptosomes from NSCexo-treated mice do not internalize TauO, so most remains bound to the outer surface. Taken together, the ELISA and flow cytometry data suggest that NSCexo reduces TauO internalization at the synapse without affecting TauO binding at the surface.
NSCexo reduce TauO internalization in hippocampal neurons in vitro
To test whether NSCexo-mediated reduction of TauO internalization at the synapse resulted from a direct effect on neurons, we exposed cultures of mature hippocampal neurons, generated from NSC differentiation, to NSCexo or MNexo for 24 h. After removing excess exosomes, TauO (2.5 μM) was added to the cultures for 60 min, and the cells were washed, fixed, and stained with anti-βIII–tubulin and anti-hTau (Tau5) antibodies. Confocal images revealed the presence of Tau bound to the dendritic processes and within the cell bodies (as previously reported; Puangmalai et al., 2020; Fig. 12A–C). Quantitative analysis showed that, while TauO did not reduce the number of neurons (Fig. 13A), NSCexo treatment significantly reduced the amount of Tau found inside the neurons, specifically in the nucleus (Fig. 12D), and increased the number of neurites without TauO puncta (Fig. 6E). The number of Tau puncta per neurites was reduced in both NSCexo- and MNexo-treated neurons as compared with those treated with PBS (Fig. 12F). No puncta were visible in cultures not treated with TauO (Ctrl) and stained with anti-hTau antibody (Fig. 13B). These results suggest that NSCexo reduced the internalization of Tau into neurons (in accordance with the ELISA and flow cytometry data and the in vivo immunofluorescence data).
NSCexo reduce TauO internalization into hippocampus neurons in vitro. Hippocampal neurons generated by differentiation of adult rat hippocampus NSC were treated with NSCexo, MNexo, or PBS for 24 h before being challenged with TauO (2.5 μM) for 1 h. A, Representative confocal images (60× with 2 zoom) of neurons pretreated with PBS, NSCexo, and MNexo (hTau in red, βIII-tubulin in green, and the nuclei in blue; scale bar, 20 µm). B, Representative confocal images (60× w/7.5 zoom) showing the presence of hTau inside the neurites (scale bar is 2 µm). C, Representative confocal images (60× with 7.5 zoom) showing the presence of hTau in the neuronal cell bodies (scale bar, 5 µm). D, Quantification of the number of hTau puncta inside neurons (5 images acquired from 8 independent experiments). *p < 0.05; **p < 0.01; one-way ANOVA (F = 6.701) followed by Tukey’s multiple-comparison test. E, Quantification of the number of neurites without hTau puncta (5 images acquired from 5 independent experiments). *p < 0.05; ****p < 0.0001; one-way ANOVA (F = 22.82) followed by Tukey's multiple-comparison test. F, Quantification of the number of hTau puncta on neurites (5 images acquired from 5 independent experiments). *p < 0.05; **p < 0.01; one-way ANOVA (F = 7.975) followed by Tukey's multiple-comparison test. Data are mean ± SEM. NSCexo, neural stem cells exosomes; MNexo, mature neurons exosomes; PBS, phosphate-buffered saline; TauO, Tau oligomers; SEM, standard error of the mean.
A, Hippocampal neurons generated by differentiation of adult rat hippocampus NSC were treated with PBS for 24 h before being incubated with PBS for 1 h. Neurons were fixed and processed for immunofluorescence analysis using anti-hTau antibody and anti-βIII–tubulin. Left, Representative confocal images (460× with 2 zoom) of neurons pretreated with PBS (hTau in red, βIII-tubulin in green, and the nuclei in blue; scale bar, 20 µm). Middle, Representative confocal images (60× with 7.5 zoom) showing the absence of hTau inside the neurites (scale bar, 2 µm). Right, Representative confocal images (60× with 7.5 zoom) showing the absence of hTau in the neuronal cell bodies (scale bar, 5 µm). NSC, neural stem cells; PBS, phosphate-buffered saline. B, Quantification of the number of DAPI+ cells across the experimental groups (5 images acquired from 8 independent experiments). One-way ANOVA (F = 1.035) followed by Tukey’s multiple-comparison test.
Discussion
NSCs are critical for neurogenesis in the hippocampus, a process necessary for synaptic plasticity, learning, and memory (Kempermann et al., 2018). Previously, we showed that aged individuals who remained cognitively intact despite the presence of Aβ plaques and Tau tangles in the brain had high numbers of NSCs in the hippocampus (Briley et al., 2016). Consistent with this discovery, we showed that exosomes released by hippocampal NSCs reduced the binding of AβO at synapses and prevented AβO-induced suppression of LTP and memory recall deficits (Micci et al., 2019).
Here, we report, in wild-type mice, that NSCexo reduced internalization of TauO into hippocampal neurons, protected synapses from TauO-induced suppression of LTP expression, and prevented memory deficits driven by intracerebroventricular injection of TauO and that these effects may occur via delivery of specific miRNA exosomal cargo. Most importantly, we show that this mechanism is specific to NSCexo because exosomes secreted by MN, generated from the differentiation of hippocampal NSCs, were unable to provide similar protections. Using exosomes from MNs derived from differentiation of NSCs enables the comparison of exosomes specifically associated with both a “stem” state and a MN phenotype within the same cell lineage. Consequently, this approach excludes the potential of nonspecific effects of exosome vesicles, such as scavenging of exogenously added TauO (An et al., 2013; Yuyama et al., 2015).
Our data show that both synaptic plasticity (LTP) and memory (NOR) can be protected against TauO-induced deficits in mice treated with NSCexo prior to challenging with TauO. These results indicate that exosomes released by hippocampal NSCs play a key role in promoting or maintaining synaptic resistance to the damaging effects of TauO that ultimately result in memory deficits. Comparable trend of fEPSP reduction was observed following TauO application in PBS or NSCexo-treated animals, which was greater in the MNexo group where a significant effect is observed with the fiber volley that will be explored in the future. These data suggest other mechanisms than a direct effect on synaptic strength via action of NSCexo are involved with improvement of cognitive response observed.
Our data show that intracerebroventricular injection of TauO impair short-term memory but not long-term memory in the NOR test. This may be explained by their preferential impact on hippocampal synaptic plasticity and function, which are critical for short-term memory consolidation.
TauO are known to disrupt synaptic integrity, impair LTP, and alter neuronal communication in the hippocampus. These disruptions can selectively affect the encoding and retrieval of short-term memory, which relies heavily on intact hippocampal circuits. Conversely, long-term memory may involve broader cortical networks and compensatory mechanisms that are less immediately impacted by acute toxic effects of TauO intracerebroventricular injection.
Our data show that intracerebroventricular injection of NSCexo, but not MNexo, prior to TauO delivery, reduces Tau and pTau accumulation in the hippocampus and cortex. Our previously published data showed that exosomes are taken up by neurons. Here we show that NSCexo can reduce TauO internalization when administered to neuronal cultures. The possibility that NSCexo are also taken up by glial cells when injected intracerebroventricularly cannot be excluded. Further studies are necessary to determine whether this is the case and what the functional significance would be.
TauO are known to accumulate in synapses resulting in disrupted synaptic transmission, ultimately leading to dendritic spine retraction and cognitive impairments (Lacor et al., 2004, Tu et al., 2014). Our data show that NSCexo reduce TauO internalization into isolated synaptosomes and cultured hippocampal neurons. The mechanisms remain unclear; they may involve the low-density lipoprotein receptor-related protein 1 (LRP1). LRP1 is abundantly expressed at synapses and has been recently shown to function as a major transporter for TauO allowing rapid internalization of Tau and its subsequent spreading (Rauch et al., 2020; Cooper et al., 2021). Our data are consistent with the idea that an action of the cargo of NSCexo on the LRP1–Tau complex could be responsible for reduced Tau internalization but unaffected Tau surface binding in synaptosomes of NSCexo-treated mice.
MiRNAs are abundant in exosomes and, in addition to proteins and lipids, represent their main bioactive cargo (van den Boorn et al., 2013; Bayraktar et al., 2017). Indeed, compelling evidence indicates that many of the effects elicited by exosomes (including NSCexo) can be ascribed to the action of their specific miRNA cargoes, including modulation of aging (Zhang et al., 2017), cognition and synaptic function (Hebert et al., 2009; Konecna et al., 2009; Cohen et al., 2011), and neuroprotection (Properzi et al., 2015). Importantly, miRNAs have been shown to regulate transcriptional modulation of key synaptic proteins within the synaptic compartment itself (Schratt, 2009; Garza-Manero et al., 2014). Furthermore, our own previous findings show that resilient hippocampal synapses in NDAN individuals with high numbers of NSCs (Briley et al., 2016) have a unique proteomic signature (Zolochevska et al., 2018), suggesting that selective changes in synaptic protein expression mark synaptic resistance to toxic AβO and TauO in these humans. This is consistent with upstream regulation by specific miRNAs, as revealed by IPA analysis of our proteomic data. We have previously shown that, compared with MNexo, NSCexo contain a unique set of miRNAs that are known to modulate the expression of proteins involved in synaptic function (Micci et al., 2019). Most importantly, we also found that when injected intracerebroventricularly in mice, mimics of such unique NSCexo-derived miRNAs render synapses physically resistant to the detrimental binding of AβO and functionally resilient to their disruptive action. Here we demonstrated that these unique NSCexo-derived miRNAs protect mice from TauO-induced LTP and memory dysfunctions. These data further support the notion that selected miRNAs uniquely present in NSCexo mediate their protective action on synapses of target neurons and further suggest the exciting possibility of new drug discovery to promote cognitive resilience in patients with AD.
Our data point to NSC-secreted exosomes, and their miRNA cargo, as critical elements of the yet poorly understood, neurobiological basis of the relationship between brain reserve, cognitive resilience, and resistance to AD neuropathology. While the relative quantity of exosomes released by NSC in the hippocampus is not known, we know that exosomes injected in the DG can spread to other regions of the hippocampus and to the cortex (Zheng et al., 2017). An intriguing possibility is that a decrease in the number of NSCs in the hippocampal DG, such as that which occurs during aging (Kuhn et al., 2007; Ben Abdallah et al., 2010; Lugert et al., 2010), could lead to increased synaptic vulnerability to the toxic binding of AβO and TauO, because of reduced NSCexo in the hippocampus. Such a mechanism could be one factor that links aging to increased risk of AD (Querfurth and LaFerla, 2010).
The future pursuit of this novel concept has the potential to open the door to new therapeutic strategies for AD centered on NSC-secreted exosomes and their bioactive cargo. Indeed, our previous reports and the results of this work, demonstrating the effect of NSC-secreted exosomes in increasing synaptic resistance to AβOs and TauO, support this possibility. NSC-secreted exosomes could delay the onset and/or mitigate the severity of AβO- and TauO-dependent synaptic and cognitive dysfunctions associated with age-dependent decreases in neurogenesis.
Data Availability
Raw data are available from the corresponding authors upon reasonable request.
Footnotes
This work was supported by NIA/NIH 5R01AG042890 (to M.-A.M. and G.T.). We thank Dr. Vsevolod L. Popov and Zhixia Ding for their helpful suggestions and for the use of the Department of Pathology EM facility and Stacy L. Sell, PhD, for editing.
↵*G.T. and M.-A.M. contributed equally to this work.
The authors declare no competing financial interests.
- Correspondence should be addressed to Maria-Adelaide Micci at mmicci{at}utmb.edu or Giulio Taglialatela at gtaglial{at}utmb.edu.
