Hypoxia-conditioned extracellular vesicles from human amniotic membrane stem cells attenuate D-galactose-induced aging in mice

Abstract

BACKGROUND

Anti-aging for the preservation and recovery of physical and brain functions may be a major topic of regenerative medicine in the super-aging society. Stem cells and their secretory active molecules can play a central role of regenerative medicine.

AIM

To investigate whether extracellular vesicles (EVs) from amniotic membrane stem cells (AMSCs) enhance physical activity, including stamina, and cognitive function in a mouse model of facilitated brain-aging, and to elucidate underlying mechanisms.

METHODS

EVs were collected from conditioned media of AMSCs after hypoxic (2% O₂, 5% CO₂) cultivation for 3 days. The size and composition of EVs was analyzed via nanoparticle-tracking analysis and proteome/lipidome profiles, and functional molecules such as growth factors and neurotrophic factors were analyzed via enzyme linked immunosorbent assay. Male ICR mice were subcutaneously administered with D-galactose (300 mg/kg) for 6 weeks to induce facilitated aging, during which intravenously injected with EVs (6 × 10⁸ exosome particles/body) at weeks 0, 2, 4, and 6. Physical activity and cognitive function were assessed through Rota-rod, forced swimming and passive avoidance performances, respectively. To clarify underlying mechanisms, acetylcholine (ACh), brain-derived neurotrophic factor (BDNF), sirtuin 1 (SIRT1), glial fibrillary acid protein (GFAP), glycogen, and thiobarbituric acid-reactive substances (TBARS) were analyzed in the brain and muscles.

RESULTS

Six-week injection of D-galactose decreased physical activity and impaired learning and memory function, along with the reduced ACh, BDNF, SIRT1, and glycogen in the brain and muscles, whereas brain GFAP and muscular TBARS increased. However, EV treatment recovered the D-galactose-induced neurobehavioral deficits not only by increasing BDNF and SIRT1 (regulating neuro-muscular adaptation and function) and enhancing brain ACh (governing memory acquisition), but also by restoring muscular glycogen (an energy source) and attenuating brain GFAP and muscular TBARS (inflammatory and oxidative injury markers).

CONCLUSION

EVs from AMSCs restored cognitive function of mice with facilitated brain-aging by increasing ACh, BDNF, and SIRT1. EVs also enhanced stamina not only by attenuating tissue injury, but also by strengthening the muscles through the production of glycogen and BDNF.

Key Words: Aging; Physical activity; Cognitive function; Amniotic membrane stem cell; Extracellular vesicle; Growth factor; Neurotrophic factor

Core Tip: We obtained a large amount of extracellular vesicles containing high concentrations of growth factors and neurotrophic factors via hypoxic cultivation of amniotic membrane stem cells. The extracellular vesicles ameliorated physical and cognitive dysfunctions by recovering anti-aging factors and energy sources, in the brain and muscles, such as brain-derived neurotrophic factor, sirtuin 1, glycogen, acetylcholine, and antioxidation.

INTRODUCTION

Aging is a complex biological process that leads to a progressive decline in physiological functions, ultimately impacting physical activity, cognitive abilities, and overall quality of life. In addition to decreased physical activity, deterioration of learning and memory function is a characteristic feature in aging and Alzheimer’s disease (AD) patients[1], which has been linked to reduced neuroplasticity, increased oxidative stress, and a decline in regenerative capacity[2].

In AD patients, degeneration of presynaptic cholinergic system is one of the causes of cognitive dysfunction[3,4], in which loss of choline acetyltransferase activity and ensuing acetylcholine (ACh) depletion, has been reported[5-7]. As in AD, senescence-related brain aging leads to atrophy and degeneration of the cholinergic nervous system, resulting in profound neurobehavioral and cognitive dysfunction due to decreased ACh biosynthesis and reduced secretion of neurotrophins such as growth factors (GFs) and neurotrophic factors (NFs)[8,9]. Several hypotheses have been presented to explain the process of aging[10], in which oxidative stress was demonstrated to play a key role of senescence[11]. Increased oxidative stress causes mitochondrial dysfunction that produces high-level reactive oxygen species, accelerating neurodegeneration. Moreover, brain aging is often accompanied by astrocytic activation, reflected by increased glial fibrillary acidic protein (GFAP) levels, which signals neuroinflammation and brain injury[12,13].

D-galactose (D-Gal) is a nutrient derived from lactose in milk. Lactose is hydrolyzed to monosaccharides glucose and galactose. However, excessive D-Gal reacts with free amines of amino acids in proteins and peptides to form advanced glycation end products that cause facilitation of oxidative damage linked to the pathogenesis of degenerative diseases[14]. Especially, D-Gal makes cellular constituents vulnerable to oxidative reaction, and reduces mitochondrial energy production[15]. In addition, D-Gal not only accelerates inflammatory brain injury through activation and degeneration of astrocytes, but also causes cognitive deficits by activating acetylcholinesterase (AChE)[15,16]. Since D-Gal administration has been widely used to mimic aging-related biochemical and behavioral alterations, including reduced motor performance, cognitive impairment, and oxidative brain damage, we adopted D-Gal as a model compound for facilitated brain aging.

In our former studies, it was found that transplantation of neural stem cells encoding choline acetyltransferase gene, an ACh-synthesizing enzyme, recovered cognitive function and physical activity of animal models of AD[17] and aging[18]. Notably, amniotic membrane stem cells (AMSCs) and adipose-derived stem cells (ADSCs) were differentiated into neurons in the brain, recovered ACh and neurotrophins, improved physical and cognitive functions, and finally extended lifespan[19]. Interestingly, it was demonstrated that among GFs and NFs, brain-derived neurotrophic factor (BDNF) played a central role in the stamina-enhancing activity of ADSCs[20].

However, using stem cells has some drawbacks. The cells have to be directly transplanted into the brain for enough efficiency, because they cannot penetrate blood-brain barrier following intravenous (IV) administration. In addition, repeated injections of live cells can induce risks of immunogenicity and tumor transformation. It has been confirmed that bioactive molecules are released in a form of extracellular vesicles (EVs) from functional cells such as inflammatory, immune, and stem cells. EVs are lipid-membrane nanoparticles secreted by stem cells to encapsulate and deliver hundreds to thousands of substances such as intercellular signal transmitters, cytokines, and GFs and NFs. EVs containing these functional molecules have been shown to regulate bodily functions, inhibit inflammation, protect and regenerate cells, and repair tissues[21]. Recently, we collected a large amount of EVs by cultivating human AMSCs in a hypoxic condition (2% O₂, 5% CO₂). The EVs were confirmed to be 70-80 nm in diameter and contain high concentrations of GFs and NFs compared to the EVs obtained in a normoxic condition (20% O₂, 5% CO₂)[22,23]. It is interest to note that the EVs exerted beneficial effects on several diseases including wound healing[22], cerebral palsy[23], and AD (unpublished results).

The present study aimed to investigate the anti-aging potential of AMSC-derived EVs (AMSC-EVs) in mice with facilitated brain aging induced by D-Gal. We assessed the effects of EVs on motor coordination, swimming stamina, cognitive function, and key biochemical markers such as ACh, BDNF, sirtuin 1 (SIRT1), and glycogen in the brain and muscles. By exploring the multifaceted impacts of AMSC-EVs, we sought to shed light on their therapeutic potential for treating aging-related motor and cognitive impairments.

MATERIALS AND METHODS

Preparation of AMSCs

Human placenta were obtained from a healthy pregnant female donor during Caesarean section with informed consent. Amniotic membrane tissue was digested with collagenase I and neutralized with a medium containing 10% fetal bovine serum (Biowest, Kansas City, MO, United States). After removing erythrocytes, the remaining cells were suspended in Keratinocyte serum-free medium (Invitrogen, Carlsbad, CA, United States) supplemented with 5% fetal bovine serum and antibiotics[22]. Culture of AMSCs were maintained under 5% CO₂ at 37 °C in culture flask.

Preparation of EVs

The separated AMSCs were suspended in the defined serum-free medium in Hyper flask (Nunc, Rochester, NY, United States), and cultivated under normoxic oxygen (20% O₂, 5% CO₂) or hypoxic oxygen (2% O₂, 5% CO₂) tensions at 37 °C for 3 days[22,23]. The media were filtered through a bottle-top vacuum filter system (0.22 μm, PES membrane, Corning, Glendale, CA, United States). The conditioned media were 30-fold concentrated using Vivaflow-200 to obtain final EVs (Sartorius, Hannover, Germany).

Nanoparticle-tracking analysis of EVs

To analyze size distribution of EVs, nanoparticle-tracking analysis were performed using Nanosight NS300 equipped with v3.2.16 analytical software (Malvern Panalytical, Malvern, United Kingdom). Purified EVs were adjusted to 108 particles/mL according to the manufacturer’s recommendations in phosphate-buffered saline (PBS). The camera level was adjusted until the particle signal saturation did not exceed 20% and all particles were clearly visible, five images were captured in 60 seconds, and the size and particle concentration were analyzed[22,23].

Enzyme linked immunosorbent assay of GFs and NFs

Enzyme linked immunosorbent assay (ELISA) was conducted to analyze functional molecules, i.e., GFs and NFs related to cytoprotection and tissue regeneration in normoxic and hypoxic culture EVs[22,23]. ELISA kits for BDNF (ab212166; Abcam, United Kingdom), nerve GF (NGF, ab99986; Abcam, United Kingdom), platelet-derived growth factor (PDGF, ab100622; Abcam, United Kingdom), insulin-like growth factor (IGF, ab100695; Abcam, United Kingdom) or vascular endothelial growth factor (VEGF, ab222510; Abcam, United Kingdom) were used according to the manufacturer’s instructions. Briefly, EVs were incubated with the primary antibodies at room temperature, followed by secondary antibody, and color-developing substrate was applied for 10-30 minutes. After treatment with a stop solution, the absorbance was measured at 450 nm.

Proteome analysis of EVs

Proteome analysis, including CD markers of exosomes, was performed according to the standard protocol of KBio Health (Osong, Korea). For liquid chromatography-tandem mass spectrometry analysis of proteins, EVs were dissolved in 0.1% formic acid to make final 1 μg/μL. Protein concentration was determined using the BCA assay. Using a nano LC system, 3 μL (3 μg) EV sample was injected onto an LC analytical column (EASY-Spray 50 cm × 75 μm PepMap RSLC C18 2 μm) equipped with Q-Exactive™ Plus Hybrid Quadrupole-Orbitrap™ Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, United States), and separated with a linear gradient of solvent B (5%-95%) over 70 minutes at 300 nL/minute[23]. Data analysis was performed in Proteome Discoverer 2.4 (Thermo Fisher Scientific, Waltham, MA, United States), with searches run against the Homo sapiens database.

Lipidome analysis of EVs

Lipidome analysis was performed according to the standard protocol of KBio Health. Using a UPLC system for analysis of lipids, 5 μL of sample was injected onto an LC analytical column (Agilent Poroshell EC-120 C18 2.1 mm × 100 mm) equipped with Q-Exactive™ Plus Hybrid Quadrupole-Orbitrap™ Mass Spectrometer, and separated with a linear gradient of solvent B (1%-99%) over 30 minutes at 0.25 mL/minute. Lipidomic database searching was performed using LipidSearch 4.2.21 (Thermo Fisher Scientific, Waltham, MA, United States). Results were refined by removing peaks predicted to be false positives.

Animals

Eight-week-old male ICR mice were procured from the Daehan-Biolink (Eumseong, Korea). The animals (n = 6/group) were maintained at a constant temperature (23 ± 2 °C), relative humidity of 55% ± 10%, and 12-hour light/dark cycle. The animals were fed with standard rodent chow and purified water ad libitum.

Aging model and treatment

For the induction of facilitated aging, the animals were subcutaneously injected with D-Gal (300 mg/kg) for 6 weeks[14,16]. Treatment group mice received IV injections of EVs (6 × 10⁸ EV particles/body) at 2-week intervals, that is, at weeks 0, 2, 4, and 6. In the animal study, we used only hypoxic culture EVs, but not normoxic culture EVs, since the effects of normoxia-cultivated EVs were minimal in the preliminary and previous studies.

Measurement of physical activity

Rota-rod performance: Motor balance and coordination were evaluated using a Rota-rod test system (Panlab technology, Barcelona, Spain) on days 1, 7, 21, and 35. Mice were placed on a rotating rod at a constant speed of 4 rpm, and the time taken for the mice to fall off the rod was recorded[19,20]. The endpoint was set to 600 seconds. The average latency was calculated from 3 measurements. Forced swimming performance: To analyze physical stamina, all mice were allowed to swim for 30 minutes for adaptation to a swimming pool with a constant water temperature of 25 ± 0.5 °C for 3 days before the measurement. All mice were subjected to a forced maximum swimming performance on days 2, 8, 22, and 36. The mice were loaded with a lead ring weighing 5% of their body weight attached to the tail, and then placed in the swimming pool. The test was performed by forcing animals to swim until exhaustion, which was determined by observing loss of coordinated movements and failure to return to the water surface within 7 seconds[19,20]. This 7-second cut off time was used as the criterion for the maximum swimming capacity of the animals.

Measurement of cognitive function

For the evaluation of memory acquisition, the mice were subjected to a passive avoidance Shuttle box on day 43. Passive avoidance performance was assessed by 5 consecutive trials at 1-hour intervals to evaluate memory acquisition and retention. The Shuttle box apparatus consists of two compartments equipped with a lamp and a steel-grid floor for electric shock. On the trials, electric shock (0.5 mA for 2 seconds) was delivered when rats entered the dark compartment from the light room through a guillotine door[17-19]. The latency time of stay in the light room from light-on was recorded. The end-point was set at 300 seconds, denoting full acquisition of memory.

Analysis of brain ACh

The mice were sacrificed 24 hours (on day 44) after the final forced swimming tests. The brain and gastrocnemius muscles were weighed and homogenized in PBS containing a cocktail of protease inhibitors (Sigma-Aldrich, St. Louis, MO, United States). After centrifugation, the supernatant was measured for ACh concentration with the Amplex Red ACh/AChE assay kit (Molecular Probes, Eugene, OR, United States) according to the manufacturer’s instructions[17-19]. In this assay, ACh is hydrolyzed by AChE to release choline, which is then oxidized by choline oxidase to betaine and H₂O₂. H₂O₂ interacts with Amplex Red (7-dihydroxyphenoxazine) in the presence of horseradish peroxidase to generate the highly fluorescent resorufin. The resulting fluorescence was measured in a fluorescence microplate reader using excitation in the range of 530-560 nm and emission at 590 nm.

ELISA of brain and muscular functional proteins

Brain and gastrocnemius muscle extracts were analyzed using different mouse ELISA kits for the following proteins: BDNF (Antibodies, A1657; Cambridge, United Kingdom), SIRT1 (Abcam, ab206983, United Kingdom), and GFAP (Abcam, ab233621, United Kingdom). Brain tissues were homogenized in a lysis buffer containing protease and phosphatase inhibitors. Samples were centrifuged at 18000 × g for 20 minutes, and the supernatant was collected. All ELISA experiments followed the manufacturer’s protocols[24,25]. The optical densities of the samples were measured using a microplate reader at a wavelength of 450 nm, and the levels were calculated using the standard curves and normalized to the total protein content.

Analysis of muscle glycogen

Muscle tissue (30 mg) was hydrolyzed in 0.3 mL of 30% potassium hydroxide solution in a boiling water bath for 30 minutes under intermittent shaking. After cooling to room temperature, 0.1 mL of 1 M Na₂SO₄ and 0.8 mL of ethanol were added to the sample, and boiled again for 5 minutes. The glycogen pellet was dissolved in 0.2 mL of water, and additional ethanol precipitations were performed twice. The final pellet was dried, dissolved in 0.2 mL of 0.3 mg/mL amyloglucosidase in 0.2 M sodium acetate buffer (pH 4.8), and then incubated for 3 hours at 40 °C. The reaction mixture was added to 0.2 mL of the glucose assay solution containing 0.3 M triethanolamine-potassium hydroxide (pH 7.5), 1 mmol/L ATP, 0.9 mmol/L β-NADP, and 5 μg of G6PDH/mL[26]. The absorbance at 340 nm was determined before and after addition of 1 μg of hexokinase. Glycogen content was expressed as μmol of glucosyl units/g.

Measurement of muscular lipid peroxidation

Lipid peroxidation was measured by determining the formation of thiobarbituric acid-reactive substances (TBARS) in the gastrocnemius muscles. Briefly, the muscles were homogenized in cold PBS and centrifuged at 4 °C. SDS (500 μL of 8.1% solution) and acetic acid (1 mL of 20% solution; pH 3.5) were added to the muscle homogenate and centrifuged. Aliquots of the clear supernatant were mixed with an equal volume of TBA solution (0.8% w/v) and heated in a glass tube capped with aluminum foil at 95 °C for 30 minutes[20,26]. After cooling on ice, each sample was pipetted into 96-well plates, and the absorbance was read at 532 nm with a microplate reader.

Statistical analysis

The data were described as mean ± SEM. Statistical significance between the groups was analyzed by one-way analysis of variance with post-hoc tests (n = 6/group) using the SPSS statistical software 30.0.0. (SPSS Inc., Chicago, IL, United States). P-values of less than 0.05 were considered statistically significant.

RESULTS

Characteristics of AMSC-EVs

From nanoparticle-tracking analysis of the size distribution of AMSC-EVs, the major peak was observed at 71 nm (Figure 1A). It was found that the exosome particles in EVs were not aggregated or contaminated with larger EVs such as microvesicles and apoptotic bodies, since other considerable peaks were not observed. In ELISA, the concentrations of BDNF, NGF, PDGF, IGF, and VEGF which are neurotrophins related to neuroprotection and neuroregeneration, neuro-muscular adaptation, tissue regeneration, and angiogenesis were very low in normoxic (20% O₂) culture medium. By comparison, it was confirmed that the GFs and NFs were very high in hypoxic (2% O₂) culture EVs, reaching ten to a hundred times the normoxic culture EVs (Figure 1B).

Figure 1 Characteristics of extracellular vesicles obtained from amniotic membrane stem cells. A: Size distribution of extracellular vesicle particles from nanoparticle-tracking analysis; B: Concentrations of growth factors and neurotrophic factors in normoxic (white bars) and hypoxic (black bars) culture extracellular vesicles. ^aP < 0.05, compared with each normoxic culture extracellular vesicles. BDNF: Brain-derived neurotrophic factor; NGF: Nerve growth factor; PDGF: Platelet-derived growth factor; IGF: Insulin-like growth factor; VEGF: Vascular endothelial growth factor.

From proteomic analysis of EVs, total 751 proteins were detected. The top 30 proteins are listed in Table 1, showing rich in cysteine (SPARC), collagen α-1 (I) chain, isoform 1 of fibronectin, filamin-A, collagen α-2 (I) chain, vimentin, collagen α-3 (VI) chain, collagen α-1 (VI) chain, fibrillin, and collagen α-1 (III) chain. Collagens and fibronectins are essential components for restoring tissue integrity from injury, and SPARC plays a critical role for proper assembly and maturation of collagen fibrils. Interestingly, the composition of AMSC-EVs was largely different from that of umbilical cord blood stem cell-derived EVs (UCBSC-EVs), in which isoform 1 of fibronectin, collagen alpha-3 (VI) chain, and collagen alpha-1 (XII) chain were the highest. Notably, CD9, CD63, and CD81 were detected in both AMSC-EVs and UCBSC-EVs, confirming the nature of exosomes.

Table 1 Proteome analysis.

Accession	Protein name	Gene name	#PSM	Ratio	UCBSC (reference)	Remark
P09486	SPARC	SPARC	349	2.24%	0.13%
P02452	Collagen alpha-1 (I) chain	COL1A1	312	2.00%	0.51%	Collagen
P02751-1	Isoform 1 of fibronectin	FN1	297	1.90%	3.32%	Cytoskeleton
P21333	Filamin-A	FLNA	264	1.69%	0.58%	Cytoskeleton
P08123	Collagen alpha-2 (I) chain	COL1A2	214	1.37%	0.23%	Collagen
P08670	Vimentin	VIM	207	1.33%	0.74%
P12111	Collagen alpha-3 (VI) chain	COL6A3	204	1.31%	2.31%	Collagen
P12109	Collagen alpha-1 (VI) chain	COL6A1	187	1.20%	0.56%	Collagen
P35555	Fibrillin-1	FBN1	180	1.15%	0.72%
P02461	Collagen alpha-1 (III) chain	COL3A1	168	1.08%	0.14%	Collagen
P60709	Actin, cytoplasmic 1	ACTB	154	0.99%	0.92%
Q99715	Collagen alpha-1 (XII) chain	COL12A1	150	0.96%	1.04%
Q15582	Transforming growth factor-beta-induced protein ig-h3	TGFBI	150	0.96%	0.66%
A0A804HK76	Filamin B	FLNB	149	0.95%	0.14%
A0A804HII9	Actinin alpha 1	ACTN1	134	0.86%	0.49%
O00469-2	Isoform 2 of procollagen-lysine, 2-oxoglutarate 5-dioxygenase 2	PLOD2	131	0.84%	0.12%
P02787	Serotransferrin	TF	111	0.71%	ND	Iron transporter
O43707	Alpha-actinin-4	ACTN4	110	0.70%	0.50%
P04264	Keratin, type II cytoskeletal 1	KRT1	109	0.70%	0.25%
P68032	Actin, alpha cardiac muscle 1	ACTC1	107	0.69%	0.56%
P07585	Decorin	DCN	103	0.66%	0.32%
P01033	Metalloproteinase inhibitor 1	TIMP1	100	0.64%	0.03%	Collagenase inhibitor
P08253	72 kDa type IV collagenase	MMP2	99	0.63%	0.41%
A0A7P0TA35	Protein disulfide-isomerase	P4HB	97	0.62%	0.10%
P05121	Plasminogen activator inhibitor 1	SERPINE1	96	0.62%	0.36%
P26038	Moesin	MSN	94	0.60%	0.39%
P09871	Complement C1s subcomponent	C1S	91	0.58%	0.29%
Q12841	Follistatin-related protein 1	FSTL1	91	0.58%	0.06%
P35908	Keratin, type II cytoskeletal 2 epidermal	KRT2	90	0.58%	0.21%
A0A1B0GVI3	Keratin 10	KRT10	89	0.57%	0.34%
A0A7I2V3T0	CD9 molecule	CD9	16	0.02%	0.06%	Positive marker
F8VWK8	Tetraspanin (fragment)	CD63	11	0.03%	0.04%	Positive marker
H0YEE2	CD81 molecule	CD81	20	0.03%	0.07%	Positive marker

Top 30 proteins and exosome markers (CD9, CD63, and CD81) from total 751 proteins identified were listed. UCBSC: Umbilical cord blood stem cell; ND: Not detected.

From lipidomic analysis of EVs, 15 lipid species were identified (Table 2), wherein phosphatidylcholine (PC, 71.89%), ceramide (Cer, 14.31%), phosphatidylethanolamine (PE, 6.70%), and lysophosphatidylcholine (LPC, 2.37%) were the most abundant lipids, compared to the relatively-low levels of phosphatidylserine (PS), phosphatidylinositol (PI), and sphingomyelin (SM). Such high concentrations of PC, Cer, and PE may give the EV membranes fluidity and stability. In comparison, PC (51.70%), SM (21.44%), PI (10.36%), PS (3.92%), and PE (3.83%) were main lipid components in UCBSC-EVs, in which cell signaling- and apoptosis-related molecules were high.

Table 2 Lipidome analysis.

Lipid classes	Class	AMSC	UCBSC (reference)
Ceramide	Cer	14.31	1.52
Diacylglycerol	DG	0.90	0.53
Monohexosylceramide	HEX1 Cer	0.18	0.38
Dihexosylceramide	HEX2 Cer	ND	0.05
Lysophosphatidylcholine	LPC	2.37	1.77
Lysophosphatidylglycerol	LPG	0.00	ND
Lysophosphatidylinositol	LPI	0.00	1.30
Monoglyceride	MG	1.70	0.42
Phosphatidic acid	PA	0.00	ND
Phosphatidylcholine	PC	71.89	51.70
Phosphatidylethanolamine	PE	6.70	3.83
Phosphatidylglycerol	PG	0.00	0.87
Phosphatidylinositol	PI	0.00	10.36
Phosphatidylserine	PS	0.00	3.92
Sphingomyelin	SM	0.32	21.44
Triacylglycerol	TG	1.63	0.40
	Total	100.00	98.49

AMSC: Amniotic membrane stem cell; UCBSC: Umbilical cord blood stem cell; ND: Not detected.

Change in body weights

During 6-week treatment with D-Gal, the body weight gain was higher, to some extent, than that of normal animals (Figure 2). The effect of AMSC-EV treatment at 2-week intervals on D-Gal-induced increase was negligible, although the body weight gain was slightly attenuated.

Figure 2 Body weights of D-galactose-induced aging rats treated with extracellular vesicles. D-galactose (300 mg/kg) was subcutaneously administered every day for 6 weeks, and extracellular vesicles (6 × 10⁸ particles/body) were intravenously injected at 2-week intervals (on days 0, 13, 26, and 41). EVs: Extracellular vesicles.

Recovery of physical activity

To investigate the effect of EVs on motor balance and coordination, a Rota-rod test was performed. On day 1 post-D-Gal administration (Figure 3), the latency time of D-Gal-administered rats markedly decreased to 45% of normal animals, indicating impaired motor coordination. Such an impaired movement lasted for 35 days (Figure 3). In contrast, the EV treatment remarkably recovered the latency time throughout the experimental period. Similarly, the maximum swimming time of D-Gal-challenged rats significantly decreased, compared to the normal group (Figure 4). However, treatment with EVs led to a great improvement of swimming performance, especially increasing the latency time longer than that of normal animals.

Figure 3 Effects of extracellular vesicles on the Rota-rod performance of D-galactose-induced aging rats on days 1, 7, 21, and 35. D-galactose (300 mg/kg) was subcutaneously administered every day for 6 weeks, and extracellular vesicles (6 × 10⁸ particles/body) were intravenously injected at 2-week intervals (on days 0, 13, 26, and 41). ^aP < 0.05, compared with normal control; ^bP < 0.05, compared with D-galactose alone. EVs: Extracellular vesicles; D-Gal: D-galactose.

Figure 4 Effects of extracellular vesicles on the maximum swimming time of D-galactose-induced aging rats on days 2, 8, 22, and 36. D-galactose (300 mg/kg) was subcutaneously administered every day for 6 weeks, and extracellular vesicles (6 × 10⁸ particles/body) were intravenously injected at 2-week intervals (on days 0, 13, 26, and 41). ^aP < 0.05, compared with normal control; ^bP < 0.05, compared with D-galactose alone. EVs: Extracellular vesicles; D-Gal: D-galactose.

Recovery of cognitive function

The cognition-enhancing effect of EVs was investigated via a passive avoidance performance. Six-week injection of D-Gal resulted in profound impairment in cognitive tasks, displaying significantly-delayed acquisition of information (Figure 5). Notably, EV treatment improved the learning and memory function. To confirm the relationship between memory deficits and brain ACh level, we analyzed the concentration of ACh, a neurotransmitter governing memory acquisition. Repeated administration of D-Gal decreased ACh concentration in the brain by 43% (Figure 6A), which was significantly restored by EV treatment, that is, near-fully recovering to a normal level.

Figure 5 Effects of extracellular vesicles on the passive avoidance performance of D-galactose-induced aging rats on day 43. D-galactose (300 mg/kg) was subcutaneously administered every day for 6 weeks, and extracellular vesicles (6 × 10⁸ particles/body) were intravenously injected at 2-week intervals (on days 0, 13, 26, and 41). ^aP < 0.05, compared with normal control; ^bP < 0.05, compared with D-galactose alone. EVs: Extracellular vesicles.

Figure 6 Effects of extracellular vesicles on the brain acetylcholine, brain brain-derived neurotrophic factor, muscular brain-derived neurotrophic factor, and brain sirtuin 1 in D-galactose-induced aging rats on day 44. A: Effects of extracellular vesicles (EVs) on the brain acetylcholine in D-galactose (D-Gal)-induced aging rats on day 44; B: Effects of EVs on the brain brain-derived neurotrophic factor in D-Gal-induced aging rats on day 44; C: Effects of EVs on the muscular brain-derived neurotrophic factor in D-Gal-induced aging rats on day 44; D: Effects of EVs on the brain sirtuin 1 in D-Gal-induced aging rats on day 44. D-Gal (300 mg/kg) was subcutaneously administered every day for 6 weeks, and EVs (6 × 10⁸ particles/body) were intravenously injected at 2-week intervals (on days 0, 13, 26, and 41). ^aP < 0.05, compared with normal control; ^bP < 0.05, compared with D-galactose alone. ACh: Acetylcholine; BDNF: Brain-derived neurotrophic factor; SIRT1: Sirtuin 1; EVs: Extracellular vesicles; D-Gal: D-galactose.

Brain and muscular protection

Following D-Gal challenge, the concentrations of brain and muscular BDNF decreased by 31% and 40.4%, respectively (Figure 6B and C). However, EV treatment significantly restored both the brain and muscular BDNF, a key molecule for neuroprotection and neuro-muscular adaptation, up to near-normal levels. The brain level of SIRT1, an anti-aging molecule, was also significantly reduced in the D-Gal-challenged animals. The D-Gal-induced decrease in SIRT1 was attenuated by treatment with EVs to some extent (Figure 6D). Next, we examined astrocytic response by measuring GFAP level, a specific marker of brain injury and aging. The ELISA showed significant increase in the brain GFAP via activation of astrocytes, following 6-week injection of D-Gal (Figure 7A). Notably, the GFAP increase due to brain injury was fully recovered by EV treatment. D-Gal treatment caused marked depletion of glycogen, an energy source, in gastrocnemius muscles to a half level (Figure 7B). However, EV treatment led to a considerable recovery. To assess muscular damage, lipid peroxidation was evaluated using TBARS analysis. The TBARS content in the gastrocnemius muscles of the D-Gal-challenged rats showed a significant increase by 25.4% compared to the normal animals (Figure 7C). In contrast, the EV treatment lowered the TBARS concentration to a normal level, in parallel with the recovery of muscular glycogen (Figure 7B).

Figure 7 Effects of extracellular vesicles on the brain glial fibrillary acidic protein, muscular glycogen, and muscular thiobarbituric acid-reactive substances in D-galactose-induced aging rats on day 44. A: Effects of extracellular vesicles (EVs) on the brain glial fibrillary acidic protein in D-galactose (D-Gal)-induced aging rats on day 44; B: Effects of EVs on the muscular glycogen in D-Gal-induced aging rats on day 44; C: Effects of EVs on the muscular thiobarbituric acid-reactive substances in D-Gal-induced aging rats on day 44. D-Gal (300 mg/kg) was subcutaneously administered every day for 6 weeks, and EVs (6 × 10⁸ particles/body) were intravenously injected at 2-week intervals (on days 0, 13, 26, and 41). ^aP < 0.05, compared with normal control; ^bP < 0.05, compared with D-galactose alone. GFAP: Glial fibrillary acidic protein; TBARS: Thiobarbituric acid-reactive substances; EVs: Extracellular vesicles; D-Gal: D-galactose.

DISCUSSION

It was demonstrated that AMSCs and ADSCs exerted anti-aging effects, improving physical activity, recovering cognitive function, and extending health span and lifespan[19]. Furthermore, ADSCs enhanced physical stamina, which was mediated by neuroprotective and neuro-muscular adaptive activities of GFs and NFs including BDNF[20]. Notably, BDNF-loaded exosomes displayed neuroprotective effects in vitro oxidative stress and in vivo stroke animal models[27].

In the present study, the exosomal nature of EVs was confirmed with specific markers CD9, CD63, and CD81 in proteomic analysis (Table 1)[22]. ELISA revealed that the EVs contain various neurotrophins such as BDNF, NGF, PDGF, IGF, and VEGF (Figure 1). BDNF plays a critical role in synaptic development and plasticity via strengthening excitatory (glutamatergic) synapses[28,29]. Indeed, BDNF released from ADSCs increased physical activity in aged animals[19,20] and over-expression of BDNF led to an early improvement in locomotor function in spinal cord injury model[30]. BDNF acts as a muscle- and endothelial cell-derived NF for motoneurons[24], in which it rescued motoneurons from programmed or injury-induced cell death during development[31]. Furthermore, application of BDNF to the gastrocnemius muscles significant increased electrical excitability of motoneurons[32]. Notably, BDNF can induce fast-twitch (type II) to slow-twitch (type I) transformation of muscle fibers, thus enhancing muscle strength[19,33]. Furthermore, it promotes revascularization[34], and serves a key role in maintaining the population of muscular progenitor cells in adult muscles via regulation of satellite cell differentiation and skeletal muscle regeneration[35,36].

Interestingly, BDNF and NGF were confirmed to play a critical role in the improvement of learning and memory functions[37], as observed in the present study. IGF plays myogenic and angiogenic roles, increasing muscle mass[19,38,39], which may be supported by angiogenesis, muscular adaptation, and tissue protection mediated by VEGF, BDNF, and PDGF[38,40-42], leading to the enhanced stamina in forced-swimming in the present study. In addition, VEGF and IGF enhance antioxidative capacity by up-regulating antioxidant enzymes[38,40,41,43]. Similarly, the oxidative injury of gastrocnemius muscles was significantly attenuated by EV treatment (Figure 7). Therefore, it is believed that the neurotrophins from EVs and preserved brain and muscles contributed to the enhanced stamina and cognitive function.

Although total 751 proteins and 15 lipid species were detected in AMSC-EVs (Tables 1 and 2), it is not reasonable to interpret the roles of each component, since there are too many molecules and relationship among them. In brief, AMSC-EVs were rich in SPARC as well as collagen and fibronectin components. SPARC is a critical cellular matrix protein that regulates the proper assembly and maturation of collagen fibrils. The abundant collagens and fibronectins are essential components for restoring tissue integrity from injury. Interestingly, the composition of AMSC-EVs was largely different from that of UCBSC-EVs, in which several fibronectin and collagen components were the highest. However, it is not possible for us to compare the different natures between the two EVs due to too many factors affecting composition and function.

More interesting point is the lipidomic profile of EVs (Table 2). AMSC-EVs showed a composition of high PC, Cer, and PE concentrations, but relatively-low levels of PS, PI, and SM. The high concentrations of PC, Cer, and PE may play a role for membrane fluidity and stability of the EV membranes, whereas PS, PI, and SM could be signals for apoptosis-related scavenging by inflammatory cells. Notably, PC, SM, PI, and PS were main lipid components in UCBSC-EVs, in which cell signaling- and apoptosis-related molecules are high. It is interesting to note that AMSCs do not express major histocompatibility complex class-II mediating immune response[44], and thereby can avoid immune-related elimination or adverse effects during repeated administration.

However, there are clear limitations in comparing the composition and function of AMSC-EVs with other stem cell EVs, because we obtained data from hypoxia-cultivated AMSC-EVs, but not from hypoxic culture UCBSC-EVs or ADSC-EVs. Actually, hypoxic culture leads to different particle size, composition, and functional molecules in EVs. In addition, content of microRNAs (miRNAs) in EVs and the role of each miRNA remain to be analyzed in a follow-up study, since hundreds of miRNA are included in EVs, and they function as cell and tissue regulators.

Interestingly, the particle size of AMSC-EVs was found to be 71 nm in diameter (Figure 1) which is much smaller than other stem cell EVs[45]. Indeed, it was found that the AMSC exosomes readily penetrated both the normal and damaged keratinocytes, fibroblasts, osteoblasts, and retinal ganglionic cells[22]. Importantly, we demonstrated the fate (distribution and clearance) of fluorescence dye-labeled EVs in vivo: i.e., the EVs readily entered the brain tissue of cerebral palsy animals following IV injection[23]. Therefore, it is believed that the tiny AMSC-EVs can more efficiently deliver functional molecules to the injured tissues including the brain and muscles.

Aging may come from the loss of regenerative capacity, determining the gradual tissue injury, degeneration, and longevity[2]. D-Gal caused significant deficits of physical performances, consistent with previous studies highlighting the impact of aging on motor function and physical stamina[14-16]. As biochemical parameters of tissue injury following facilitated aging, increased brain GFAP and muscular TBARS were confirmed (Figure 7). It was found out that such D-Gal-induced tissue damages came from depletion of brain and muscular BDNF as well as muscular glycogen, an energy source[20,26] (Figures 6 and 7). As expected, AMSC-EVs restored BDNF and glycogen in the brain and muscles, and thereby lowered brain GFAP and attenuated muscular lipid peroxidation. SIRT1 participates in a variety of cellular metabolic processes, exerting anti-aging effects. Indeed, stem cell EVs delayed brain aging through neuroprotective, anti-apoptotic, and antioxidative activities via SIRT1-signaling pathway in senescence-accelerated mice[25]. In the present study, AMSC-EVs restored the brain SIRT1 depleted by D-Gal, and might contributed to the neurobehavioral recovery (Figure 6).

Brain aging leads to cognitive dysfunction, which are closely related to depletion of ACh, a neurotransmitter governing memory acquisition. D-Gal markedly reduced the brain ACh level (Figure 6), and eventually impaired the passive avoidance performance (Figure 5). BDNF and SIRT1 play a central role in the improvement of learning and memory functions, in addition to neuroprotective activity[25,37,46]. Therefore, it is believed that AMSC-EVs recovered ACh by enhancing BDNF and SIRT1 (Figure 6), and thereby improved learning and memory functions.

The enhanced physical and cognitive activities by AMSC-EVs were inferred from the high GFs and NFs, since the neurotrophins are well known to exert tissue-protective and -regenerative activities[19,22]. In addition, AMSC-EV treatment led to a marked increase in muscle glycogen levels, indicating improved metabolic function and energy storage in muscles (Figure 7). Lipid peroxidation was also fully blocked by EVs containing antioxidative BDNF and IGF[38,40,41,43]. Therefore, diverse active ingredients of AMSC-EVs are considered to exert positive effects in reducing muscle fatigue and injury. Indeed, in our previous study, BDNF released from ADSCs enhanced physical stamina and exerted multiple effects including antioxidative activity, glycogen metabolism, muscle type differentiation[20]. Interestingly, it was found that BDNF induces transformation of fast-twitch muscles to slow-twitch fibers, which are resistant to long-term fatigue[20,33]. These findings suggest that AMSC-EVs not only protect the brain function, but may also enhance peripheral tissue activity, contributing to overall physical performances.

CONCLUSION

Taken together, the analytical results of neurobehavioral functional and biochemical parameters provide compelling evidences that EVs have multi-faceted potentials, effectively mitigating the deleterious effects of D-Gal-induced aging. The observed improvements in motor coordination, physical stamina, and cognitive function, in addition to molecular markers of brain health, suggest that AMSC-EVs may offer a promising strategy to delay aging-related cognitive decline and motor impairments. Although further follow-up studies are necessary to elucidate the underlying mechanisms, it is clear that the AMSC-EVs exert beneficial effects, and could be a candidate for clinical applications to delay aging progress or overcome age-related diseases.