The increasing number of aged individuals in the global population will likely lead to an increase of costs accounting on health care system. Thus, it is of crucial importance improving the comprehension of the mechanisms underneath ageing processes and developing new therapeutic strategies in order to reduce the effects of age-related morbidities. The goal standard will promote an improvement in health lifespan and a reduction of age-related co-morbidities conditions.

Aging is defined as a loss of physiological function during the time and is regulated by specific molecular pathways (Xu and Sun, 2015). Aging process leads to an increased risk of several chronic diseases such as cancer, cardiovascular disease, and autoimmune disease, but also dementia. It is associated with the body’s altered capacity to face up stress caused by metabolism, infection, and damage to cellular macromolecules. The comprehension of molecular mechanisms driving aging will help the scientific community to figure out why aged individuals are more vulnerable to those diseases and why they may be less stress-resistant (Panagiotou et al., 2018).

Aging is considered conserved across taxia and it is characterized by nine hallmarks comprising: genomic instability, telomere shortening, epigenetic alterations, loss of proteostasis, deregulated nutrient sensing, mitochondrial dysfunction, senescence, stem cell exhaustion and alteration in intercellular communication (Shiels et al., 2017). Moreover, the so-called “inflammaging,” a chronic inflammatory status, represents the main feature of aging process (Salvioli et al., 2013).

While aging involves the entire organism, not all cell types age at the same rate and it is conceivable that senescent cells may contribute to spread senescence to young cells (Olivieri et al., 2015).

Senescence is a particular phenotype of eukaryotic cells leading to a loss of replication ability in response to several stimuli that induce DNA damage (Campisi and d’Adda Di Fagagna, 2007).

The major component in the signal transmission from senescent cells to the surrounding tissue is the SASP that can facilitate the removal of senescent and remodeling of tissue by attraction of phagocytic immune cells (Urbanelli et al., 2016). Beside previously known SASP components, many Wnt ligands have been counted. Wnt is a secreted signaling molecule extremely conserved playing critical roles in many processes including stem cell proliferation and maintenance of homeostasis in the canonical pathway (β-catenin dependent) and transcriptional and non-transcriptional cellular responses in the non-canonical ones (β-catenin independent) triggered by calcium or other Wnt ligands, as Frizzled receptors (Nusse, 2005). It should be noted that these two different pathways interact with each other and other multiple pathways, including the NF-κB, MAPK, and JNK pathways making Wnt signaling extremely complex and articulated (Zhang et al., 2014; Ma and Hottiger, 2016). It is not surprising that Wnt is involved in aging too (Nusse, 2005). Indeed some sources suggest that Wnt signaling decays with aging in brain impairing adult neurogenesis (Okamoto et al., 2011) and lung (Hofmann et al., 2014 but at the same time it may increase in an age-dependent manner (Brack et al., 2007; Liu et al., 2014). Furthermore, the key members of Wnt pathways involved in SASP are secreted in the extracellular space by exosomes. These vesicles carrying Wnt proteins on their surface have been reported to active Wnt signaling in target cells. These findings highlighted a new role of exosomes in mediating the cell-to-cell transmission of senescence signals, suggesting that exosomes represent a new SASP component (Urbanelli et al., 2016). For the first time in 2008, Lehmann et al. (2008) described an increase of exosomes secretion by senescent cells. This increase seems to be a general feature of cellular senescence and has been observed in fibroblasts, epithelial cells, and cancer cells (Takasugi, 2018). On the contrary, Eitan et al. (2017) in cross-longitudinal study showed that plasma exosomes concentration decreased with human age, at least from the early 30s to late 60s. Monocytes and B cells internalized more exosomes rather than T cells, even if these ones are the most representative kind of PBMC in bloodstream. In addition, exosomes were more incorporated in B cells and monocytes of aged donors suggesting that exosomes internalization not only is cell-specific but age-dependent. Moreover, aging can alter RNA and protein composition of exosomes. For example Galectin-3, which plays a role in osteoblast maturation, was reduced in the plasma exosomes of elderly people, presumably as consequence of the stem cell functionality loss in the skeleton, classical of aging process (Weilner et al., 2016). Plasma exosomes isolated from young but not elderly donors promoted the osteogenic differentiation of mesenchymal stem cells in a galectin-3-dependent manner (Weilner et al., 2016). In particular, this protein, belonging to the lectin family, consists of carbohydrate recognition and collagen α-like domains. This chimeric structure allows Galectin-3 to interact with a multitude of intra-and extracellular proteins, in the nucleus as well as in the cytoplasm or on the membrane and in the extracellular space, after its secretion from different types of cells and tissues. Interacting with a myriad of proteins, Galectin-3 is involved in multiple biological processes, physiological and pathological, such as development, neuronal functions, immune reactions, endocytosis, neoplastic transformation and metastasis, and osteoblastogenesis, which impairing seems to contribute to age-related bone frailty (Dumic et al., 2006).

Exosomal miRNAs are also involved in brain aging (Pusic and Kraig, 2014). Peripheral exosomes isolated from young Wistar rats promoted differentiation in primary oligodendrocyte precursor cell (OPC) differentiation and remyelination in slice cultures. Moreover, nasal administration of EVs from young rats increased myelination in aged rat brain due to the presence of high levels of miR-219, which reduced the expression of inhibitory regulators of OPC differentiation (Pusic and Kraig, 2014). Recently, it has been described that the activity of acetylcholinesterase protein (AChE) was increased in young as well as old Wistar rats. An age-related increase was observed in CD63 levels in CSF exosomes but a decrease was observed in plasma vesicles of the older group. The authors showed that the young adult rats had significantly higher circulating IL-1β levels in the exosomes compared to the aged ones, without any effect on central content. These data suggest that the normal aging process caused different changes in the profiles of central and circulating exosomes. Altered IL-1β levels in circulating EVs could be linked, at least partly, to age-related inflammatory conditions, and a disruption of the CSF exosomes in aged rats, evaluated by CD63 levels, could be related to susceptibility to neurodegenerative disorders (Gomes de Andrade et al., 2018).

Cellular senescence triggered by specific conditions as irradiation, DNA-damaging reagents, and oncogenic RAS expression, all enhance exosomes secretion. This increase is mediated by p53 (Lehmann et al., 2008) and one of its targets, Tumor suppression-activated pathway 6 (TSAP6), but the mechanism whereby TSAP6 regulates exosomes secretion is not well understood. Nevertheless, it has been demonstrated that exosomes creating pro-inflammatory environment accelerate the aging process (Biran et al., 2017). Interestingly, exosomes contain various lengths of genomic DNA fragments and seem to be one of the major routes of DNA secretion (Fernando et al., 2017) and DNA secretion from exosomes increases upon cellular senescence (Takasugi, 2018). Intriguingly, cH2AX positive cytoplasmic chromatin fragments appear in senescent cultured cells (IMR90 and HEMa-LP); suggesting that damaged DNA may be the major source of exosomes associated DNA in senescent cells (Ivanov et al., 2013). Exosomes are involved also in promoting genomic instability, another aging hallmark, thought the transfer of retrotransposons that are DNA elements able to create and insert multiple copies of themselves into host genomes. It is interesting to note that retrotransposons expression has been found to increase in senescent mouse cells (strain C57BL/6; De Cecco et al., 2013).

The deposition of toxic proteins is another event correlated with aging. To date, it is demonstrated that exosomes are involved in the transport of pathogenic proteins in the brain and in the progression of neurodegenerative diseases (Bellingham et al., 2012). Recently, the NDE levels of six neuronal proteins have been quantified in cognitively intact older subjects. Except for Phosphorylated tau-S396, the exosomal levels of Phosphorylated tau P-181, Beta Amyloid 42 (Aβ1–42), chatepsin D, repressor element 1-silencing transcription factor (REST) and neurogranin are significantly modified with aging (Goetzl et al., 2016).

MicroRNAs are short non-coding RNAs that regulate negatively gene expression at post-transcriptional level. It has been reported that several miRNAs are involved in aging and cellular senescence (Urbanelli et al., 2016). The importance of exosomal miRNAs analysis lies in the fact that these molecules could potentially transmit signals to surrounding tissues with a good impact, but also with a detrimental role. Moreover, exosomal miRNAs are interesting in the context of aging biomarker search (Sprott, 2010). MiRNAs released from senescence cells in the extracellular environment by exosomes have been reported that are able to spread senescence in surrounding cells.

For example, miR-433 promoted the induction of senescence in ovarian cancer cells (A2780) and when overexpressed, miR-433 was released in association with exosomes (Weiner-Gorzel et al., 2015). MiR-34a and miR-29 can induce cell cycle arrest in colon carcinoma cell line (HCT116 cells) contributing to the stabilization of p53/p21 by targeting proteins relevant for its regulation such as Sirtuin 1 (SIRT1; Yamakuchi and Lowenstein, 2009). The miR17–92 cluster is down regulated in several cell aging models as endothelial cells, replicated CD8+ T cells, renal proximal tubular epithelial cells, and skin fibroblasts and they can target p53/p21 (Weilner et al., 2013). Another miRNA involved in cellular senescence is miR-146, whose expression was increased in senescent human fibroblasts (HCA2) when compared with proliferating quiescent ones. It is important to underlie that miRNA-146 targets IL-6 and IL-8, SASP components with pro-inflammatory function, suggesting a role for miR-146 as senescence-associated inflammation modulator (Bhaumik et al., 2009). On the other hand, miRNAs encapsulated in exosomes are able to suppress cellular senescence; this is the case of miR-214, involved in angiogenesis, that plays a role in vesicle-mediated signaling between endothelial cells. Exosomes derived from human microvascular endothelial cell line (HMEC-1) stimulated migration and angiogenesis in recipient cells, whereas exosomes from miR-214-depleted endothelial cells failed to stimulate these processes preventing senescence and allowing blood vessel formation (van Balkom et al., 2013). A recent microarray study performed on salivary exosomes miRNAs from young and old healthy subjects has identified mir-24-3p as a possible peripheral aging biomarker (Machida et al., 2015). Exosomes isolated from the bone marrow of young and aged C57BL/6 mice showed a similar concentration and size distribution. However, bioanalyzer data indicated that exosomes from young and aged mice were differently enriched in miRNAs. The amount of miR-183-5p was increased in aged bone marrow exosomes, and its overexpression detected also in bone marrow stromal cells, mimicked the effects of aged bone marrow exosomes (Davis et al., 2017).

AD is considered the most frequent cause of dementia. It is characterized, clinically, by cognitive and behavioral disorders and, pathologically, by the extracellular deposit of insoluble Aβ-amyloid and intracellular neurofibrillary tangles (NFTs), consisting of tau fibrils. The amyloid plaques derived from impaired processing of the APP leading to the formation of the toxic Aβ-amyloid. APP is translocated into the endoplasmic reticulum (ER) and matures through Golgi apparatus. The mature form of APP is transported to the cell membrane where it undergoes to further proteolytic cleavage from β- and γ-secretases acting together to produce fibrils of the toxic Aβ-amyloid that accumulates with age in human AD brains (Busciglio et al., 1993; Takahashi et al., 2002). The dosage of Aβ-amyloid, total tau and phosphorylated tau in the CSF is recognized as the “core” of AD biomarkers in the clinical practice (Molinuevo et al., 2018). Interestingly, the first link between exosomes and AD proposed that Aβ-amyloid was released in association with exosomes. Moreover, the presence of other specific exosomal proteins as Alix and Flotillin-1 were also found accumulating into the AD brain (Rajendran et al., 2006; Sharples et al., 2008). A prion-like mechanism to explain how aggregates of Aβ-amyloid seem to self-propagate and spread to cells out of CNS is confirmed in AD mouse models. Seeding of Aβ was observed when extracts of AD human brain were injected in healthy mice that express the human wild-type APP gene causing the formation of the plaques in the site of injection and adjacent brain region (Morales et al., 2011). In addition, tauopathy was inducted in ALZ17 transgenic mice injecting aggregated of tau protein (Clavaguera et al., 2009). In the light of these findings, the hypothesis that exosomes could use a prion-like mechanism to disseminate toxic proteins associated with AD is taking hold (Coleman and Hill, 2015; Thompson et al., 2016). Exosomes could be involved in the trafficking of amyloid aggregates because Tg2576 mouse brain mice and post-mortem human AD brains were enriched in exosome markers within amyloid plaques (Kokubo et al., 2004; Rajendran et al., 2006). Phosphorylated tau was also detected in the exosomes from CSF of early-onset AD patients (Saman et al., 2012). ADAM10, Beta-secretase 1 (BACE1), nicastrin, and presenilin 1 and 2 (PSEN1 and 2) are other examples of AD pathogenic proteins found inside exosomes of transgenic mouse brain (Tg2576) and cell culture APP models, as CHO cell line (Sharples et al., 2008; Perez-Gonzalez et al., 2012). More recently, researchers have found that exosomes could stimulate aggregation of Aβ-amyloid and tau in vivo models, 5XFAD and rTg4510 transgenic mice (Dinkins et al., 2014; Polanco et al., 2016). In other words, exosomes, removing the excess of intracellular Aβ, shuttled it outside the cells concurring to plaque formation (Joshi et al., 2015). On the other hand, a neuroprotective role is also proposed. Neural exosomes could uptake Aβ-amyloid reducing the Aβ load in the brain as seen in the brains of mouse models (C57BL/6, KM670/671NL and V717F) where after the injection of exosomes, a decrease of Aβ and amyloid deposition was observed (Yuyama et al., 2015). Furthermore, extracellular tau could arise by secretion through exosomes in SH-SY5Y and COS-7 cell lines (Simón et al., 2012). Even if there is a body of literature arguing the role of exosomes in aggregate transmission, the fact remains that this theory assumes the presence of pathogenic proteins within exosomes about which the functional evidences are few or controversial (Lim and Lee, 2017). Nevertheless, the exosome hypothesis is appealing and partly explains the intercellular transmission of proteinopathies. Worth mentioning also, a research field that proposes exosomes as source of biomarkers for CNS disorders (Figure 3) due to their interesting characteristics suitable to the clinic (e.g., presence in many biological fluids, crossing the BBB, protection of the biomolecules inside them, etc.). Goetzl et al. (2015, 2018) measured the levels of different pathogenic proteins, Aβ-amyloid, total tau and p-tau isoforms inside NDEs immunoprecipitated with L1CAM to isolate specifically neuronal exosomes from blood of AD, Mild Cognitive Impairment (MCI) and controls (Kapogiannis et al., 2015). They found higher levels of these proteins vs. controls able to predict the development of AD 10 years before clinical onset (Fiandaca et al., 2015) or the progression from MCI to dementia (Winston et al., 2016). Instead, a contrary study showed no difference in NDEs total tau levels for AD patients (Shi et al., 2017). It is known that type-2 diabetes is an AD risk factor thus AD brains have markers of insulin resistance as Insulin Receptor Substrate-1 (IRS-1). Altered forms of IRS-1 were detected in NDEs of AD plasma patients and at lower levels compared to controls and to patients with type 2 diabetes with intermediate levels. This is interesting because NDEs IRS-1 protein levels could contribute to discriminate MCI/AD to controls and patients with type-2 diabetes at the same time (Kapogiannis et al., 2015). Pathological proteins were also found in exosomes extracted from CSF. In the work of Saman et al. (2012), the tau phosphorylated at threonine 181 (pT181) was more concentrated in CSF exosomes than in the total CSF and in early stage of AD, while it was absent in other dementing conditions as vascular or Lewy body diseases. This exosomal tau detected so early in AD suggests that CSF tau could be secreted, not shed from dead neurons (Saman et al., 2012).

FTD is the most common form of dementia in the presenium accounting for up to 20% of patients with an onset before 65 years. Three clinically different syndromes characterize FTD: behavioral variant (bv) FTD, Progressive Non Fluent Aphasia (PNFA), and Semantic Dementia (SD). These different subtypes are related to different clinical features but mostly patients present a profound alteration in the behavior and personality, often associated with cognitive and executive impairment, except for PNFA and SD where the language impairment is prevalent (Snowden et al., 2007). All of these syndromes at pathological level are characterized by Frontotemporal lobar degeneration (FTLD). Histopathologically FTLD is defined on the type of protein depositing into FTLD-Tau, FTLD-TAR DNA Binding protein (TDP)-43, and FTLD-Fused in Sarcoma (FUS; Fenoglio et al., 2018).

Up to 40% of patients have a history of familial transmission with nearly 10% of patients showing an autosomal dominant inheritance pattern. The majority of familial FTLD account mutations in the microtubule associated protein tau (MAPT) and progranulin (GRN) genes, and the pathologic expansion of the hexanucleotide GGGGCC repeat in the first intron of C9ORF72 gene (Rademakers and Hutton, 2007).

As discussed above for AD, also for FTD, the involving of exosomes in the pathology has been investigated although the current knowledge is still limited. FTD is characterized by TDP-43 aggregates accumulation throughout the nervous system. As for AD pathogenic proteins, also TDP-43 protein can be exchanged via exosomes between neuronal cells (Neuro2a cells and primary neurons) leading to propagation of TDP-43 proteinopathy in a “prion-like” manner (Iguchi et al., 2016). Indeed the uptake of exosomal TDP-43 oligomers from recipient cells induces higher toxicity than free TDP-43 in murine primary cortical neuron cell culture (C57Bl76J; Feiler et al., 2015). Furthermore, exosomes derived from ALS-FTD-CSF cell model showed a high concentration of full length and TDP-43 C-terminal fragments (CTFs). The latter lead to the formation of cytoplasmic inclusions within cells, so authors suggest that aberrant cleavage of TDP-43 in these exosomes acting as “seed” induces the formation of TDP-43 aggregates in the ALS-FTD-CSF-cultured cells (Ding et al., 2015). Not only TDP-43, but also dipeptide repeat proteins (DPRs) produced by aberrant translation of C9ORF72 FTD patients throughout CNS seem to spread between cells via exosome-dependent pathways (Westergard et al., 2016).

As was the case for AD patients, levels of Aβ and pT181 were increased in FTD as well (Fiandaca et al., 2015). Interestingly, the levels of synaptophysin, synaptopodin, synaptotagmin-2, and neurogranin dosed in NDEs were decreased in patients with FTD compared to controls, probably because of reduced functionality of synaptic proteins in senile dementias. These levels were low years before dementia making synaptic NDEs proteins useful for preclinical diagnosis of dementia (Goetzl et al., 2016). Instead, the Repressor Element 1 Silencing Transcriptor factor (REST) was significantly high in FTD over controls and AD representing a potential marker to discriminate FTD patients from AD (Goetzl et al., 2015). The IRS-1 phosphorylated in serine 312 was able to distinguish at 84% of accuracy between FTD patients and controls (Kapogiannis et al., 2015).

Lastly, Benussi et al. (2016) studied human primary fibroblasts without GRN null mutations. They conclude that the glycosylated form of PGRN was released with exosomes and in the presence of mutation, the secretion of exosomes was extremely reduced and their composition changed enriching in Lamp1 protein. Overall, the GRN null mutations cause an alteration in the intercellular communication.

PD is a chronic neurodegenerative disease characterized by motor impairments due to the selective death of dopaminergic neurons. Cognitive impairments can arise in the course of the disease at any time. The most of PD cases are sporadic but there are rare familial forms linked to mutations in several genes; SNCA, parkin, DJ-1, PTEN-induced kinase 1 (PINK-1) and Leucine-rich repeat kinase 2 (LRRK2; Thomas and Beal, 2007). Even if the molecular pathogenesis of PD is not fully understood, it’s now universally accepted that α-synuclein plays a predominant role in PD accumulating in Lewy Bodies, a pathological hallmark of PD (Spillantini and Goedert, 2018). Indeed α-synuclein aggregates are responsible for synaptic pathology and neurodegeneration (Kramer and Schulz-Schaeffer, 2007). In addition, mutations that involve duplication or triplication of the wild-type SNCA are associated to autosomal dominant PD with a severity proportional to the degree of α-synuclein over-expression whereas missense mutations in SNCA (e.g., A53T) are linked to dominantly inherited forms of PD (Thomas and Beal, 2007). Therefore, it is not surprising that the spreading of pathology, already proposed for AD and demonstrated for PD involved α-synuclein. Several studies on murine primary cortical neurons and SH-SY5Y cell lines reported that α-synuclein was secreted from exosomes (Emmanouilidou et al., 2010; Danzer et al., 2012). Moreover, exosomes, providing environments for α-synuclein nucleation, catalyzed its aggregation in N2a cells and cultured hippocampal neurons (Olanow and Brundin, 2013; Grey et al., 2015). Another study in human H4 cell line demonstrated that the loss of function of P-type ATPase ion pump PARK9/ATP13A2 led to a decrease in secretion of α-synuclein into extracellular space, indeed the overexpression of PARK9/ATP13A2 caused the opposite effect, suggesting that PARK9/ATP13A2 was involved in the α-synuclein secretion at least in part via exosomes (Tsunemi et al., 2014). This consequence could have a neuroprotective effect. Probably because the increased release of exosomes containing α-synuclein reducing the intracellular levels of that protein, it could explain the surviving of neurons of substantia nigra in sporadic PD patients that overexpress PARK9/ATP13A2.

Furthermore, the biogenesis of α-synuclein exosomes seems to be modulated by zinc levels regulated from PARK9/ATP13A2 in SHSY5Y cells (Kong et al., 2014). Stuendl et al. (2016) measured the levels of CSF exosomal α-synuclein, and found differences among patients with PD and Lewy bodies. In accordance with previous studies in glioblastoma cell lines, the same group demonstrated that CSF exosomes derived from patients with PD and dementia with Lewy bodies induced the oligomerization of soluble α-synuclein in target cells in a dose-dependent manner (Stuendl et al., 2016). In this regard, Shi et al. (2014) were able to demonstrate in mouse models that CSF α-synuclein was promptly transported to blood, with a small portion within exosomes but CNS specific. An increased releasing of this protein to the blood of PD patients was explained discovering that in a large cohort of clinical samples (267 PD and 215 controls); the levels of plasma exosomal α-synuclein were significantly higher in PD patients. Another protein secreted from exosomes is LRRK2. Mutations in the LRKK2 gene cause late-onset PD. LRKK2 secretion was regulated by 14-3-3 protein. Indeed, using 14-3-3 inhibitor, the LRRK2 secretion from exosomes was interrupted in mouse primary neurons and macrophages (Fraser et al., 2013).