It is known that intracellular neurofibrillary tangles formed by hyperphosphorylated tau protein are one of the typical pathological features of AD, and the pathological proliferation of tau is also a hallmark of AD. Although the toxic effects of tau occur primarily inside the cell, tau can be released into the extracellular space as a “seed”. The tau “seeds” can be taken up by other healthy recipient cells and transferred from one cell to another (Figure 2C). Hence, tau pathology is spread through different regions of the brain in a manner similar to prions. This proliferative effect significantly promotes the pathological process of tau. However, in earlier times, investigators believe that tau is passively released into the extracellular space only through cell death, and then these membrane-free tau “seeds” spread to neighboring nerve cells (Gómez-Ramos et al., 2006). On the contrary, in recent years, with the discovery of EVs, people gradually realize that exosomes can also serve as an important vehicle to actively secrete tau “seeds” into the extracellular environment. Recently, the role of exosomes in the spread of tau pathology has gradually become one of the hot issues concerned by many researchers.

Tau oligomers first pass through the endoplasmic reticulum and Golgi apparatus, then become part of vesicles and subsequently enter exosomes. Under physiological conditions, exosomes containing tau are degraded by lysosomes. However, once the lysosome is damaged, it contributes to tau accumulation in exosomes. Earlier studies have reported that some forms of tau exist in the EVs (Dujardin et al., 2014), and these EVs isolated from neuronal culture can induce abnormal tau phosphorylation in the brain of wild-type mice, showing significant neurotoxic effects (Aulston et al., 2019). Moreover, another study revealed that most of the intracellular tau is full-length, while the bulk of the extracellular tau is free-floating, unaggregated and C-terminal truncated, lacking the microtubule binding regions necessary for aggregation. Only a small number of tau is encapsulated in exosomes (Kanmert et al., 2015). These tau proteins packed in exosomes contain the microtubule binding regions and have the ability to aggregate. Furthermore, it has been confirmed that there are small amounts of full-length tau in neuronal exosomes (Guix et al., 2018). In summary, the detailed mechanisms of tau encapsulated in exosomes still need further study.

Evidence from some studies illustrates that tau aggregates can spread and replicate in the brain in a prion-like manner via exosomes. Tau oligomers delivered by exosomes may act as “seeds”, and the uptake of pathological tau “seeds” leads to tau misfolding into a toxic conformation in recipient cells (Jackson et al., 2022). Exosomes isolated from the brains of tau transgenic rTg4510 mice carry a large number of tau “seeds”, thereby inducing endogenous tau aggregation in the recipient cells and increasing tau phosphorylation and oligomer formation. This evidence confirms that EVs are the carriers of tau pathology (Baker et al., 2016; Polanco et al., 2016). Many in vivo experiments have demonstrated that tau-containing neuronal exosomes from AD patients can induce the formation of neurofibrillary tangles (NFT) in rodents, exhibit tau neurotoxic effects, and thus reproduce the disease characteristics triggered by the original exosome contents. Winston et al. found that NDEs extracted from the plasma of patients with mild cognitive impairment and AD are able to seed p-tau pathology and induce AD-like neuropathology in the brains of normal mice (Winston et al., 2016). In addition, another study has shown that neuronal exosomes derived from human tau have toxic effects on recipient mouse neurons in vivo and cause long-distance propagation of tau pathology and neurodegeneration (Winston et al., 2019). A similar phenomenon has also been observed in the study of Reilly et al. (2017). Surprisingly, further research by Polanco et al. showed that a simple neuronal circuit model was built to observe the diffusion of exosomes in the interconnected neurons using a microfluidic culture system. The findings suggest that exosomes can spread the tau protein over long distances through a novel hijacking mechanism of endosomes, thereby increasing the pathogenic potential and radius of action of the exosomes (Polanco et al., 2018). In a recent study, to reveal how tau seeds contained within internalized exosomes exploit mechanisms of lysosomal degradation to escape the endosome and induce tau aggregation in the cytoplasm, their research team also found that the enzymatic activities of lysosomes can penetrate the exosomal and endosomal membranes, thereby promoting the entry of exosomal tau seeds into the cytoplasm and inducing its aggregation (Polanco et al., 2021). In earlier studies on the relationship between exosomes and tau, several studies described that the spread of tau pathology is achieved by the direct transmission of exosomes across synapses between interconnected neurons, and the depolarization of neurons faciliates the release of tau-containing exosomes (Wang et al., 2017). Recent literatures have also reported that only exosomes and vesicle-free tau seeds are shown to propagate across synapses along neural networks (Polanco and Götz, 2022). Most of the exosomal tau released from AD synapses is C-terminal-truncated and oligomeric, and with seeding activity that is enhanced by Aβ (Miyoshi et al., 2021). These results also support the ideas that Aβ-induced pathology may directly or indirectly drive tau-mediated neurotoxicity and NFT formation (Blurton-Jones and Laferla, 2006).

It is becoming increasingly clear that microglia may play a critical role in the interneuronal transmission of tau pathology promoted by exosomes. It has been shown that microglia can spread tauopathy by internalization and secretion of exosomes. Microglia propagates tau pathology via secretion of tau-containing exosomes. The depletion of microglia or the inhibition of exosome synthesis significantly reduces tau propagation in vitro and in vivo. Surprisingly, tau propagation mediated by microglia can occur not only through trans-synaptic transmission, but also in non-trans-synaptic pathways (Asai et al., 2015). Most importantly, several reports have illustrated that some genes closely related to the risk of AD morbidity can interfere with the spread of tau pathology by influencing the production of exosomes in microglia. Bridging integrator 1 is a gene associated with late-onset AD. Overexpression of bridging integrator 1 promotes the release of tau-enriched EVs from microglia, and contributes to exacerbate tau pathology in P301S tau transgenic PS19 mice, thus promoting the spread of AD-related tau pathology (Crotti et al., 2019). Triggering receptor expressed on myeloid cells 2 (TREM2) is a transmembrane protein produced by microglia in the brain. Variations in TREM2 have been shown to increase the risk of late-onset AD (Ulland and Colonna, 2018). It has been demonstrated that Trem2 deletion can enhance the transport, distribution and seeding of tau by microglial exosomes (Zhu et al., 2022).

Increased research has focused on exosomes as important communication mediators between neurons and glial cells. Exosomes induce neuroinflammation by acting as inflammatory mediators via carrying inflammation-causing cargoes (Weng et al., 2022). The accumulation of Aβ and hyperphosphorylation of tau have been shown to continuously activate microglia and astrocytes, and promote inflammatory responses (Novoa et al., 2022). Aβ or tau can be effectively encapsulated in exosomes, and activated glial cells or neuronal cells can release exosomes into the extracellular space, thus amplifying the neuroinflammatory effects caused by toxic proteins (Gupta and Pulliam, 2014). Additionally, microRNAs, as an important nucleic acid cargo in exosomes, have been found to be involved in the induction of neuroinflammation. It has been reported that NDEs containing miR-21-5p can be phagocytosed by microglia and induce microglia M1 polarization, which leads to increased release of neuroinflammatory factors, inhibition of neurite outgrowth, increased accumulation of P-tau, and increased the apoptosis of PC12 cells (Yin et al., 2020). Another study from Wei et al. has shown that miR-182-5p delivered by plasma exosomes can target brain-derived neurotrophic factor (BDNF) and activate the nuclear transcription factor-κB pathway, thereby promoting sevoflurane-induced neuroinflammation and cognitive dysfunction in aged postoperative cognitive dysfunction rats (Wei et al., 2022). In addition to enhancing and spreading inflammation in the extracellular microenvironment, exosomes can also play an anti-inflammatory role. Recent studies have reported the anti-inflammatory therapeutic value of exosomes due to their excellent properties, such as carrier transport, ease of modification and encapsulation of therapeutic agents. Related studies have also demonstrated that the activation of nucleotide-binding oligomerization domain-like receptor pyrin domain-containing 3 (NLRP3) inflammasome is closely associated with the occurrence of AD (Liang et al., 2022). By blocking the assembly of NLRP3 inflammasome, the exosome-like nanoparticles from ginger rhizomes strongly inhibit the downstream pathways of NLRP3 inflammasome activation, including caspase1 autocleavage, interleukin 1β and interleukin 18 secretion, and pyroptotic cell death (Chen et al., 2019). Under some specific conditions, exosomes released from microglia can also exert an anti-inflammatory role. For instance, the increased miR-124-3p in microglial exosomes promotes the anti-inflammatory M2 polarization of microglia, inhibits neuronal inflammation, and transfers into neurons to promote neurite outgrowth during traumatic brain injury (Huang et al., 2018). Taken together, whether exosomes promote or inhibit the inflammatory response may be related to the different stimulants and the cell types that produce exosomes, and further research is needed to explore the potential mechanism.

In contrast, while accumulated evidence has demonstrated that NDEs can be involved in the formation and spreading of the pathological features of AD, some studies have confirmed that NDEs also exert a protective role in the progression of AD. Therefore, NDEs play a dual role. Growing evidence indicates that NDEs and glial cell-derived exosomes may participate in Aβ clearance through independent or synergistic action and transfer neuroprotective substances between cells to alleviate the nervous system injury (Figure 2D). Earlier studies have shown that exosomes released from neuroblastoma N2a can capture and bind to Aβ via GSLs on their surface and be transported into microglia for degradation. Moreover, in vivo studies further provide evidence that the infusion of NDEs into the brains of APP mice can reduce Aβ levels and ameliorate Aβ-related pathologies (Yuyama et al., 2014). The next year, further research demonstrated that neuronal exosomes, but not glial exosomes, have abundant GSLs and can capture Aβ. Infusion of neuronal exosomes into the brains of APP transgenic mice can reduce Aβ and amyloid deposition, and these findings highlight the role of neuronal exosomes capable of capturing Aβ as messengers for Aβ clearance (Yuyama et al., 2015). All the above evidence have clearly demonstrated that intracerebral NDEs administration can improve the Aβ-related pathogenesis, which may provide a new direction for the treatment of AD. Furthermore, Yuyama et al. investigated that NDEs can drive the conformational changes of Aβ through GSLs on their surface. The changes in Aβ conformation can serve as seed on which other Aβ combines to form Aβ oligomers, and then form amyloid fibrinates. The Aβ-bound exosomes are endocytosed into endosomal-lysosomal system by microglia in a phosphatidylserine-dependent manner and promote the degradation of Aβ. Moreover, Yuyama et al. further demonstrated that the secretion of NDEs is modulated by the activity of sphingomyelin metabolic enzymes, including nSMase2 and sphingomyelin synthase 2. Inhibition of nSMase2 prevents exosome production, while inhibition of sphingomyelin synthase 2 promotes exosome release (Yuyama and Igarashi, 2022). In the co-culture experiment of N2a cells transfected with human APP and microglia cells, inhibition of sphingomyelin synthase 2 activity to improve the secretion of neurons’ exosomes can enhance the uptake of Aβ by microglia cells, and significantly decrease the level of extracellular Aβ (Yuyama et al., 2012). However, contrary to the results of this study, Dinkins et al. revealed that ADEs can also promote Aβ accumulation, but interfere with the uptake of Aβ by glial cells, thus promoting the formation of plaques in vivo. Inhibition of SMase2 activity by GW4869 or 5 × FAD mouse model targeted at nSMase2 gene deletion blocks exosome synthesis in vivo and in vitro, decreases the levels of Aβ1-42, alleviates amyloid plaque load in the brain of mice, and improves the cognitive deficits (Dinkins et al., 2014, 2016). These results suggest that neurons and astrocyte-derived exosomes may differ in glial Aβ clearance, but the underlying mechanisms remain unclear.

The evidence suggests that microglia and microglia-derived exosomes also play active roles in the mechanisms of Aβ clearance or tau pathologic protection. Earlier researches have shown that microglia exosomes contribute to the secretion of insulin degrading enzyme, which is an important enzyme for the extracellular degradation of Aβ (Tamboli et al., 2010). According to a recent study, triggering receptor expressed on myeloid cells-2 (TREM2), expressed on the membrane of microglia exosomes, can mediate the secretion of exosomes. Free microglia exosomes can bind to Aβ through TREM2, releasing chemokines to alter the inflammatory levels surrounding Aβ and promote the recognition and phagocytosis of Aβ by microglia cells (Huang et al., 2022). In addition to its roles in Aβ clearance, TREM2 can also affect tau pathology. Trem2 deletion, as previously described, can enhance tau trafficking, distribution, and pathological spread through microglia exosomes (Jain and Ulrich, 2022; Zhu et al., 2022). Conversely, upregulation of TREM2 in microglia inhibits the inflammatory response and ameliorates the pathological effects of activated microglia on neuronal tau hyperphosphorylation (Jiang et al., 2018). These evidence confirm that TREM2 can be involved in protecting tau pathology via microglia exosomes. P2X purinoceptor 7 (P2RX7) is an ATP-gated cationic channel, and its enrichment in microglia promotes exosome secretion. Administration of the P2RX7-specific inhibitor GSK1482160 to P301S tau mice significantly reduces the number of exosomes and accumulation of tau in the hippocampus, thereby improving working and contextual memory of model mice (Ruan et al., 2020). Inhibition of P2RX7 to down-regulate exosome expression can also be regarded as a novel pathway to improve tau pathology (Figure 2E).

It is necessary to take into consideration that recent studies have focused on the therapeutic effects of exosomes in AD, which further demonstrates the beneficial effects of exosomes. In this review, the therapeutic values of exosomes from different sources in AD are summarized in Table 1. However, it should be noted that exosomes that play protective roles in these studies are mainly from healthy cells. At present, numerous reports have described that stem cell-derived exosomes may play a positive role in alleviating the pathological features of AD (Figure 2F). Research by Lee et al. showed that exosomes secreted by adipose-derived stem cells reduce β-amyloidosis and neuronal apoptosis in AD transgenic mouse models and enhance the axon growth in the brain of AD patients (Lee et al., 2018). Further study indicates that mesenchymal stem cell (MSC)-derived exosomes can also significantly improve the pathological and cognitive deficits of AD (Chen et al., 2021). Exosomes derived from bone-MSCs alleviate cognitive decline in AD-like mice by improving BDNF-associated neuropathology (Liu et al., 2022). In addition, double transgenic APP/PS1 mice injected with bone-MSCs can activate the sphingosine kinase-1/sphingosine-1-phosphate signaling pathway to reduce Aβ deposition and promote cognitive function recovery in AD mice (Wang and Yang, 2021). Exosomes isolated from human umbilical cord MSCs (hucMSCs) have been shown to have therapeutic effects in many inflammation-related diseases. Ding et al. observed that hucMSC-exosomes injection contributes to the removal of Aβ deposition, alleviates cognitive dysfunction in AD model mice, and regulates the activation of microglia cells, thus alleviating neuroinflammation (Ding et al., 2018). Katsuda et al. also found that human adipose tissue-derived MSCs (ADSCs) secrete exosomes carrying the enzyme activity neprilysin (NEP). NEP is an enzyme that is important for the extracellular degradation of Aβ. The ADSCs-derived exosomes are transferred into N2a cells and reduced intracellular Aβ levels in N2a cells (Katsuda et al., 2013, 2015). In addition to stem cells, some other cell-derived exosomes or bioengineered exosomes also play beneficial roles in AD. Pan et al. used human brain microvascular endothelial cells derived exosomes inheriting p-glycoprotein as an Aβ cleansing system to remove Aβ peptides from the brain by specific capture between p-glycoprotein and Aβ, which can facilitate the clearance of Aβ and effectively ameliorate the cognitive dysfunction of AD mice (Pan et al., 2020). Moreover, M2 microglia-derived exosomes may also play a protective role in the pathogenesis of AD by improving PINK1/Parkin-mediated mitophagy (Li et al., 2022a). Based on bioengineering technology, microglia-targeting exosomes and targeted drug delivery can be designed. Mannose-modified exosomes containing gemfibrozil can not only bind with Aβ, but also specifically target microglia, thereby promoting the entry of Aβ into microglia, activating lysosome activity by exosome gemfibrozil, and accelerating lysosome-mediated clearance of Aβ in microglia (Hao et al., 2022).