Neuron-derived extracellular vesicles in blood reveal effects of exercise in Alzheimer’s disease

Background Neuron-derived extracellular vesicles (NDEVs) in blood may be used to derive biomarkers for the effects of exercise in Alzheimer’s disease (AD). For this purpose, we studied changes in neuroprotective proteins proBDNF, BDNF, and humanin in plasma NDEVs from patients with mild to moderate AD participating in the randomized controlled trial (RCT) of exercise ADEX. Methods proBDNF, BDNF, and humanin were quantified in NDEVs immunocaptured from the plasma of 95 ADEX participants, randomized into exercise and control groups, and collected at baseline and 16 weeks. Exploratorily, we also quantified NDEV levels of putative exerkines known to respond to exercise in peripheral tissues. Results NDEV levels of proBDNF, BDNF, and humanin increased in the exercise group, especially in APOE ε4 carriers, but remained unchanged in the control group. Inter-correlations between NDEV biomarkers observed at baseline were maintained after exercise. NDEV levels of putative exerkines remained unchanged. Conclusions Findings suggest that the cognitive benefits of exercise could be mediated by the upregulation of neuroprotective factors in NDEVs. Additionally, our results indicate that AD subjects carrying APOE ε4 are more responsive to the neuroprotective effects of physical activity. Unchanged NDEV levels of putative exerkines after physical activity imply that exercise engages different pathways in neurons and peripheral tissues. Future studies should aim to expand upon the effects of exercise duration, intensity, and type in NDEVs from patients with early AD and additional neurodegenerative disorders. Trial registration The Effect of Physical Exercise in Alzheimer Patients (ADEX) was registered in ClinicalTrials.gov on April 30, 2012 with the identifier NCT01681602. Graphical abstract Supplementary Information The online version contains supplementary material available at 10.1186/s13195-023-01303-9.


Background
Epidemiological studies indicate that lifestyle factors, such as a healthy diet and physical activity, may reduce the incidence of Alzheimer's disease (AD) [1].Multiple clinical trials assessing exercise as a non-pharmacological intervention for AD have found improvements in cognition, neuropsychiatric symptoms, and physical function.In elderly people with increased familial and genetic risk of AD, physical activity by itself or in conjunction with diet and cognitive training can improve or maintain cognitive and physical functions [2,3].The randomized controlled trial (RCT) ADEX [short for "Preserving Cognition, Quality of Life, Physical Health and Functional Ability in Alzheimer's Disease (AD): The Effect of Physical Exercise"] evaluated the cognitive [4,5] and functional [6] effects of exercise in community-dwelling individuals with mild to moderate AD [7].In ADEX, 16 weeks of moderate to high-intensity aerobic exercise had a positive effect on neuropsychiatric symptoms [4,8].Moreover, the primary outcome, the symbol digit modalities test (SDMT), showed a lesser cognitive decline in participants achieving high exercise attendance and training intensity compared to control participants [4], suggesting a positive effect of relatively intense exercise on cognition.
To unravel the molecular mechanisms underlying ameliorative effects of exercise in AD, studies utilizing the ADEX cohort assessed biomarkers for classic AD pathologies [9], neurodegeneration [10,11], and neuroinflammation [12] in cerebrospinal fluid (CSF), plasma, and serum.Results did not reveal any exercise effects on CSF Aβ, tau, synaptic proteins, or serum neurofilament light chain (NfL), suggesting that the positive functional and cognitive effects observed were perhaps attributable to activation of neuroprotective and neuroregenerative mechanisms rather than the main pathogenic cascades.
Exercise can modulate brain function both directly, through increased blood flow and cellular respiration, and indirectly, through the release of exercise-induced cytokines, termed "exerkines", that operate in an endocrine-like manner [13].Among neuronal proteins known to be induced by exercise are brain-derived neurotrophic factor (BDNF) and humanin, neuroprotective agents with recognized therapeutic potential in AD.Both molecules regulate synapse development and plasticity and promote neuronal survival in the face of noxious stimuli, and their depletion is part of AD pathogenesis [14].
To gain a window into the molecular changes induced by exercise in neurons, we leveraged circulating neuronderived extracellular vesicles (NDEVs).Extracellular vesicles (EVs) are lipid membrane-enclosed nanoparticles secreted by all cells and present in plasma.These nanoparticles carry variable cargo and play critical roles in inter-cellular communication [13].EVs secreted by brain neurons can cross the blood-brain barrier (BBB) and are detected in blood [15], thus representing a window into the brain.We and others have isolated NDEVs via immunoprecipitation targeting the neuronal cell-adhesion molecule L1CAM and have shown that NDEVs can be leveraged as a diagnostic tool for AD and other neurological disorders [16][17][18][19] and as surrogates of target engagement in clinical trials [20], including trials in AD [21].Therefore, NDEVs reflect the state of signaling pathways in neurons and can help interrogate mechanisms engaged by pharmacologic and non-pharmacological interventions.
In this study, we hypothesized that neuronal effects of exercise may be reflected by increased NDEV levels of proBDNF, BDNF, and humanin and tested this hypothesis utilizing plasma samples from ADEX.To further explore whether putative exerkines mediate neuronal effects of exercise, we quantified NDEV levels of several candidates, especially irisin, and studied their associations with proBDNF, BDNF, and humanin.

Methods
A description of the ADEX cohort, EV characterization methodology, and statistical analyses employed are included in Supporting Information.

Plasma samples from the ADEX cohort
We studied all remaining 162 plasma samples from the ADEX cohort, sourced from 47 participants in the control group (n = 38 at baseline, n = 43 at 16 weeks), and 48 participants in the exercise group (n = 44 at baseline, n = 37 at 16 weeks).

Results
Cohort demographics, clinical information, and cognitive performance measures at baseline are summarized in Table 1.Randomization successfully balanced groups for important parameters that could have acted as confounders of exercise effects.

Exercise does not alter NDEV concentration and size
Previous findings have shown that plasma contains a high concentration of soluble L1CAM that could significantly interfere with the enrichment of NDEVs via L1CAM immunoprecipitation [26].Hence, we first sought to validate the EV composition of our NDEV preparations by NTA and quantify the enrichment of L1CAM + EVs via flow cytometry analysis (FCA), following established guidelines [27].NTA showed that NDEVs had sizes between 50 and 450 nm, which is typical for a mixed population of exosomes and smaller microvesicles (Fig. S1C).NTA results were consistent with FCA of total EVs and NDEVs, with a violet side scatter (vSSC) signal range mainly within that of nanobeads under 500 nm (Fig. S1D).The achieved enrichment was determined by quantifying L1CAM + nanoparticles before and after L1CAM immunoprecipitation using FCA (Fig. S1D-H).Among EVs gated based on the detection of blue succinimidyl ester (BSE) staining all EVs (Fig. S1D), ~ 0.5% were positive for L1CAM (Fig. S1E); this portion drastically increased to over 40% after L1CAM immunoprecipitation (Fig. S1F).A vast percentage of L1CAM + EVs in NDEV preparations were double-positive for the canonical EV markers CD9, CD63, and CD81 (Fig. S1G), confirming their EV identity.The specificity of the anti-L1CAM antibody was confirmed by FCA of NDEVs labeled with an isotype control antibody showing the absence of EVs within the L1CAM gate (Fig. S1H).The abolition of detected events after treatment with NP40 detergent further demonstrated that FCA signals originated from membrane-enclosed nanoparticles (Fig. S1D).
Given variable reports regarding the acute effects of exercise on the concentration and size of total circulating EVs [28,29], we examined the effect of 16-week exercise on NDEVs.There were no group differences or changes over time in the concentration (Fig. S1A) and particle size (Fig. S1B and S1C) of NDEVs based on NTA.To account for any effects of differential NDEV yield on NDEV protein biomarkers, the concentration of NDEVs was used as a covariate in all models, as previously done [30].

NDEV biomarkers of exercise are inter-related
As physical activity can profoundly alter the neuronal proteome [32], we explored whether exercise modulates NDEV cargo for exerkines expressed by neurons (as confirmed by https:// www.prote inatl as.org/) (Table S2).We were particularly interested in irisin, an exerciseinduced hormone shown to regulate cognitive function by promoting BDNF expression [33].Although no changes were observed in NDEV irisin with exercise, as was also the case for all exerkines (Table S2), we observed a positive correlation with BDNF at baseline that was preserved at 16 weeks for both groups (Fig. S2C; Tables S3 and S4).Similar positive correlations were observed between levels of NDEV irisin and humanin and between humanin and proBDNF (Fig. S2A and B; Tables S3 and S4).We also examined the relationships between exercise-induced changes in NDEV biomarkers and the only significant association observed was between the change in NDEV irisin and that of BDNF in the exercise group (Fig. S2D; Table S5).

Discussion
In this study, we sought to shed light on the molecular mechanisms underlying the beneficial effects of exercise in AD by examining a priori hypothesized molecular effectors in plasma NDEVs from AD patients participating in a RCT of exercise.We found that in patients with mild to moderate AD, 16 weeks of aerobic exercise increased NDEV levels of BDNF, proBDNF, and humanin (Fig. 1).These findings strengthen the notion that NDEV cargo reflects the molecular state of brain neurons and any dynamic changes to it, a thesis supported by previous evidence showing that NDEV biomarkers can track AD progression [34] and demonstrate target engagement in clinical trials [20].
BDNF is a neurotrophin produced upon proteolytic cleavage of its precursor, proBDNF, which is also bioactive.Physical activity increases BDNF concentrations in brain [35] and plasma [36].Previous studies have shown that proBDNF and mature BDNF mRNA and protein, are decreased in early and end-stage AD brain in correlation with cognitive measures [14].The upregulation of BDNF by exercise has been observed in both animal [37] and human studies [38] supporting its development as a non-pharmacological intervention for AD.Proposed mechanisms on how exercise may increase BDNF levels in neurons include: (i) effects through the PGC1α-dependent myokine, irisin [33], (ii) the myokine Cathepsin B, which when increased peripherally by exercise can cross the BBB and enhance BDNF production and hence neurogenesis [39] and (iii) β-hydroxybutyrate, which also increases during aerobic exercise and increases BDNF expression in the brain [37].As previously shown, both proBDNF and BDNF are present in NDEVs, at higher levels compared to plasma [40].Moreover, NDEV proBDNF, but not BDNF, was associated with physical activity in a large longitudinal cohort of aging [41].Future studies, unrestrained by the amount of plasma available, may Fig. 2 NDEV exercise effects stratified by APOE ε4 genotype.Bar graphs show the concentrations of proBDNF (A), BDNF (B), and humanin (C) in NDEVs from ε4 carriers and non-carriers from the exercise group at baseline (light blue) and 16 weeks (dark blue).Graphs show the estimated marginal mean ± standard error of the mean from repeated measures mixed-effects models.* indicates p < 0.05; ns: not significant use NDEVs as a tool to further dissect the downstream effects of proBDNF and BDNF, such as by measuring their receptor levels (p75 NTR and TrkB), downstream effectors and functional outcomes, such as levels of synaptic proteins [34].
Humanin is a mitochondria-derived peptide that suppresses neuronal apoptosis, preserves synapses, reduces inflammation, and supports glucose and oxidative metabolism [42].Plasma humanin decreases with age in humans and mice [43], and upon replenishment, cognition in aged mice is improved [44].Humanin mRNA in plasma EVs has been found decreased in AD compared to control individuals [45], whereas protein levels of humanin in NDEVs have been found decreased in multiple neuropsychiatric disorders [46,47].Interestingly, physical activity increases humanin in plasma [48].Humanin can be destabilized via ubiquitination assisted by TRIM11 [49], a process that could be interrupted by aerobic exercise, which inhibits the ubiquitin-proteosome pathway [50].In turn, increased humanin levels could inhibit the pro-apoptotic factors BAX and BID, thus preventing mitochondrial-outer membrane permeabilization and enhancing neuronal survival and cognitive performance in AD [51].These observations suggesting that humanin depletion is associated with mitochondrial function deficits in AD further support humanin augmentation via physical activity as a non-pharmacological intervention for AD.
Exercise-induced factors released from muscle and other organs, collectively coined exerkines, can mediate systemic responses to physical activity [13].The extracellular milieu is a harsh environment for soluble exerkines, which may have driven the evolution of a parallel lipidenclosed delivery mode via EVs, which are known to exert autocrine, paracrine, and endocrine signaling functions [13].Previous studies have shown that endurance exercise increases circulating EVs, including a musclederived subpopulation capable of being biodistributed to the brain after peripheral injection [13,52].Thus, it has been widely hypothesized that physical activity modulates neuronal function indirectly, through the activity of soluble or EV-associated peripheral exerkines capable of crossing the BBB.Although many of these peripheral exerkines are also expressed in neurons, they remained unchanged in NDEVs after exercise (Table S2), suggesting that physical activity engages different pathways in neurons and peripheral organs.
Our study also indicates that neuronal exercise effects are coordinated across multiple inter-connected pathways; this is suggested by the correlation between changes in NDEV BDNF and irisin (Fig. S2), proteins whose expression is known to be co-regulated [53], as well as correlations between humanin with irisin and proBDNF, both at baseline and after exercise, (Fig. S2) which have not been previously reported.
We explored whether exercise effects may differ by APOE genotype, as previous studies have shown that, in both healthy elderly and AD individuals, beneficial effects of physical activity on cognition and hippocampal volume are stronger in ε4 carriers [54].These studies suggest that the presence of ε4 increases the responsiveness of neurons to exercise.In favor of such a hypothesis, we found more prominent increases in NDEV proBDNF and humanin with exercise in ε4 carriers (Fig. 2).

Limitations
The randomized controlled design of ADEX offers the highest level of clinical evidence, especially given a relatively large N, rigorous inclusion criteria including PET-PiB brain imaging and good adherence [4].Sixty-two percent of the patients adhered to the exercise intervention both in terms of attendance (attended more than 80% of planned visits) and exercise intensity (exercised at more than 70% of maximal heart rate during sessions) [4].Nevertheless, intrinsic generalizability constraints limit the relevance of our results to patients with mild to moderate AD.Also, the present study was limited in terms of enrolling a racially and ethnically homogeneous Danish population; future studies should aim to increase cohort diversity to assess the potential effects of racial and ethnic backgrounds and relevant disparities.Moreover, results may not be extrapolated to interventions of different exercise duration, intensity, and type.Our study was also constrained by the amount of available plasma, which limited the number of biomarkers we were able to quantify.A further limitation is that most of the participants in this study had a normal body mass index (BMI) (Table 1).Therefore, it would be very interesting to assess the effects of exercise in NDEV biomarkers in overweight and obese individuals.Increased BMI is associated with neuroinflammation [55] and increased risk of dementia in older age [56].Our results motivate the hypothesis that both the beneficial cognitive and neuropsychological effects of physical activity and the corresponding increase in NDEV proBDNF, BDNF, and humanin may be enhanced in individuals with AD and increased BMI.This hypothesis should be addressed in future studies leveraging study cohorts that enroll individuals with high BMI and/or related comorbidities, such as insulin resistance or metabolic syndrome.

Conclusions
The present study offers a mechanistic basis for the beneficial effects of aerobic exercise in early AD by implicating the upregulation of proBDNF/BDNF and humanin.To further advance our understanding of the effects of non-pharmacological treatments for AD and related dementias, future studies should leverage samples from RCTs of exercise alone and/or multidomain lifestyle interventions and examine the relationships of EV biomarkers with cognitive and functional outcomes.

Fig. 1
Fig. 1 Neurotrophic and neuroprotective factors carried by NDEVs increase after exercise.Bar graphs show the concentrations of proBDNF (A), BDNF (B), and humanin (C) in NDEVs of control and exercise group participants at baseline (light blue) and 16 weeks (dark blue).D Humanin in EVs of all cellular origins immunocaptured by targeting panTET-EVs.Graphs show the estimated marginal mean ± standard error of the mean from repeated measures mixed-effects models.* indicates p < 0.05; ns: not significant

Table 1
Cohort demographics For the rest of the comparisons, no significant differences were found between the control and exercise groups using t-test (continuous variables) or Pearson's chi-squared test (categorical variables).MMSE, Mini-Mental State Examination.SDMT-120, correct responses in 120 s on Symbol Digit-Modalities Test.VO2max, the maximal oxygen uptake estimated by the submaximal Astrand test * P = 0.011, t-test.