Chronic Obstructive Pulmonary Disease Airway Epithelial Cell–derived Extracellular Vesicles Spread Cellular Senescence via MicroRNA-34a

Abstract

Chronic obstructive pulmonary disease (COPD) is associated with the acceleration of lung aging and the accumulation of senescent cells in lung tissue. MicroRNA-34a (miR-34a) induces senescence by suppressing the antiaging molecule sirtuin-1 (SIRT1). Senescent cells spread senescence to neighboring and distant cells, favoring COPD progression and its comorbidities. Mechanisms for spreading senescence remain undetermined but may be mediated by the transfer of microRNAs in extracellular vesicles (EVs). We analyzed the microRNA content of EVs in COPD and explored their effect on cellular senescence of healthy cells. EVs were isolated from small airway epithelial cells (SAECs) from healthy donors or patients with COPD. Recipient healthy SAECs were cultured with EVs, and the expression of miR-34a and markers of cellular senescence p21^CIP1 (cyclin-dependent kinase inhibitor-1) and SIRT1 was measured. We have shown that EVs from COPD cells induce senescence in healthy recipient cells via the selective transfer of miR-34a. COPD SAECs produce increased numbers of EVs enriched with miR-34a. EVs are taken up by healthy cells, resulting in reduced expression of the antiaging molecule SIRT1 and increased expression of markers of senescence, such as p21^CIP1 and positive staining for senescence-associated β-galactosidase, which were blocked by a specific miR-34a antagomir. Our findings provide evidence of the mechanism by which EVs spread cellular senescence in human primary cells via miR-34a rather than via soluble mediators. EVs enriched with miR-34a may spread senescence locally, accounting for disease progression, but also provide a mechanism for distant spread to account for comorbidities and multimorbidity in elderly individuals.

Clinical Relevance

Extracellular vesicles produced by chronic obstructive pulmonary disease small airway epithelial cells contain miR-34a, which is transferred and induces cellular senescence of recipient healthy cells. Thus, extracellular vesicles spread aging in chronic obstructive pulmonary disease lungs, accounting for disease progression.

Chronic obstructive pulmonary disease (COPD) is associated with accelerated lung aging (1). Vulnerabilities to chronic diseases and multimorbidity increases with aging (2, 3) and has major consequences on the burden of health care (2). Aging is a major risk factor for COPD, and its prevalence is two to three times higher in patients older than 60 years than in younger age groups (4). Cellular senescence contributes to the aging process (5) and is defined by the cessation of cell division and distinctive phenotypic alterations, leading to irreversible proliferation arrest (6, 7). Senescent cells show activation of several characteristic markers, including p53 (cellular tumor antigen p53), p21^Cip1 (cyclin-dependent kinase inhibitor-1), and p16^INK4a (cyclin-dependent kinase inhibitor-2A); increased activity of senescence-associated β-galactosidase (SA-βGal) (8); and secretion of various bioactive molecules, collectively known as the senescence-associated secretory phenotype (SASP) (6). After chronic exposure to hazardous substances that leads to chronic oxidative stress, COPD lungs show signs of accelerated lung aging with the accumulation of senescent cells, together with the loss of endogenous antiaging molecules and a persistent state of chronic inflammation that has a similar profile to the SASP (9).

MicroRNAs (miRNAs) are small endogenous noncoding RNAs, between 18 and 23 nt, that play important roles in gene and protein regulation (10). Changes in miRNA expression have been described in plasma, sputum, BAL fluid (BALF), and lung tissue of patients with COPD (11–13). Specifically, miRNA-34a (miR-34a) and miR-570 concentrations are elevated in lung tissues and small airway epithelial cells (SAECs) from patients with COPD and are induced by oxidative stress in healthy cells. MiR-34a and miR-570 promote cellular senescence by targeting and suppressing the expression of the antiaging molecules SIRT1 (sirtuin-1) and SIRT6 (14).

Senescent cells may damage their local environment and induce paracrine senescence in bystander cells because of secretion of the SASP (15). However, the mechanism for spreading senescence is not fully understood. Extracellular vesicles (EVs) encapsulate specific cargo composed of RNA, miRNA, DNA, proteins, or metabolites (16) and have the capacity to trigger molecular and phenotypical changes in recipient cells (17). Previous studies have proposed that EVs induce cellular senescence in recipient cells, but these studies used mouse cells or transformed cell lines and did not address the mechanism (18–21). Modification of EV production and composition has been described in sputum and plasma of subjects with COPD, but the function of EVs in the acceleration of lung aging has not been explored (22, 23). Therefore, we hypothesized that EVs are capable of inducing cellular senescence in healthy SAECs through a specific mechanism involving the transport and transfer of the crucial miRNAs miR-34a and miR-570.

Some of the results of these studies have been previously reported in preprint form (https://doi.org/10.1101/2023.12.18.572220).

Methods

Cell Culture and Transfections

Primary SAECs were cultured as monolayers in LHC-9 medium (Invitrogen) on collagen-coated (1% wt/vol) plates. Cells were extracted from lung tissue from subjects undergoing lung resection surgery at the Royal Brompton Hospital in London. Subjects were matched for age and smokers and subjects with COPD for smoking history (see Table E1 in the data supplement). All subjects gave informed written consent, and the study was approved by the London-Chelsea Research Ethics committee (study 15/SC/0101). Recipient SAECs or BEAS-2B cells were transfected with mirVana miRNA inhibitor (mirVana miRNA inhibitor negative control 1, hsa-miR-34a MH11030) (Ambion, Life Technologies) or miRNA mimic (mirVana miRNA mimic negative control 1, hsa-miR-34a MC11030) using Lipofectamine RNAiMAX (ThermoFisher) for 24 hours before being stimulated.

Isolation of EVs

EV isolation and experiments comply with the latest (2024) Minimal Information for Studies of Extracellular Vesicles guidelines (24). Two million SAECs were seeded in T75 flasks coated with 1% wt/vol collagen in 10 ml LHC-9 depleted of EVs by ultracentrifugation (100,000 × g for 2 h). Five milliliters of LHC-9 was added to the flasks after 2 and 4 days of culture. After 7 days, cell viability was measured using annexin V/propidium iodide staining and was above 95% (see Figure E1). Conditioned media was centrifuged at 300 × g for 10 minutes to eliminate debris, then centrifuged at 20,000 × g for 30 minutes to isolate large EVs and at 100,000 × g for 2 hours for small EVs. The resultant pellets were resuspended in PBS or media. Expression of the EV markers CD9 (EPR2949, ab92726; Abcam), CD81 (EPR4244, ab232390; Abcam), CD63 (ab68418; Abcam), Alix (ab88388; Abcam), and calreticulin (MAB38981; Bio-Techne) was analyzed using western blot. EV concentration was measured using nanoflow cytometry (NanoFCM) and flow cytometry with Dynabeads coated with a primary monoclonal antibody specific for CD9 (Invitrogen) (see the data supplement). Large and small EVs treated for 20 minutes with 0.01% vol/vol Triton X-100 (Sigma-Aldrich) were used as an internal control.

Stimulation of Recipient Cells with EVs

EVs isolated from SAECs were resuspended in 300 μl media. EVs were quantified using nanoflow cytometry, and the amount of EVs proteins was analyzed using Pierce Bradford Plus protein assay kits (Life Technologies). Large EV concentrations varied from 1.86 × 10⁸ to 2.66 × 10¹¹ and small EVs from 9.30 × 10⁸ to 1.46 × 10¹³. The volume of EVs was normalized to the number of donor cells, and 150 μl EV solution (5 μg proteins equivalent to ∼2 × 10¹¹ particles) was applied to 250,000 recipient SAECs plated in six-well plates. To control the effect of soluble factors, recipient cells were treated with 150 μl conditioned media depleted of EVs, called “no EVs.” Expression of miR-34a was analyzed after 3 hours, and expression of SIRT1, p21^CIP1 (CDKN1a) and p16^INK4a (CDKN2A) was analyzed after 12 hours using quantitative PCR analysis. Protein expression of SIRT1, p21^CIP1, and p16^INK4a was analyzed after 48 hours using western blot. The activity of SA-βGal was analyzed by colorimetric analysis after 48 hours (see the data supplement).

Analysis of Cell Uptake of EVs

EVs were labeled with the fluorescent lipid membrane label PKH67 according to the manufacturer’s instructions (Sigma-Aldrich). Briefly, 2× dye solution was prepared in diluent C by adding 2 μl pkh67 solution in 1 ml Diluent C. EVs (1:1 dilution EVs:dye) or PBS (as a negative control) were incubated at room temperature for 10 minutes and centrifuged as previously described. EV pellet was washed in PBS to avoid free dye in suspension and, after ultracentrifugation, resuspend in 1 ml LHC-9. SAECs were stimulated with 100 μl EV stained solution for 3, 12, 24, or 48 hours at 37°C with 5% (vol/vol) CO₂. Cells were fixed with paraformaldehyde 4% and stained with Red Cell trace (Life Technologies) and DAPI (Abcam) and imaged using confocal microscopy. Cells were trypsinized, and PKH67 fluorescence was detected in cells using a Canto II flow cytometer (Beckton Dickinson Biosciences).

Statistical Analysis

Data are expressed as mean ± SEM. Results were analyzed using two-way ANOVA with Bonferroni posttest correction and Mann-Whitney, Wilcoxon, Kruskal-Wallis, or Friedman tests as appropriate. Prism 9.2 (GraphPad Software) was used for analysis. P values ≤0.05 were considered to indicate statistical significance.

Results

SAEC from Subjects with COPD Produce EVs Enriched with MiR-34a

The characterization of EV concentration and the size of EVs produced by SAECs showed that large EVs (99.7 ± 3.6 nm) are significantly larger than small EVs (85.8 ± 2.8 nm) (Figure 1). EV markers were detected using western blot, and the vesicular structure of EVs was confirmed using flow cytometry (see Figure E2). SAECs from subjects with COPD produced 10-fold more large vesicles than EVs derived from healthy SAECs (Figure 1A) (healthy, 1.13 ± 0.35 × 10¹⁰ particles/ml, n = 9; COPD, 1.35 ± 0.99 × 10¹¹ particles/ml, n = 9), with no difference in their average size (Figure 1B). Similarly, COPD SAECs produced more small EVs than SAECs from healthy donors (Figure 1C) (healthy, 2.2 ± 0.21 × 10¹² particles/ml, n = 9; COPD: 1.37 ± 0.39 × 10¹² particles/ml, n = 9), but not significantly (P = 0.055), and there was no difference in their size (Figure 1D). SAECs from COPD subjects expressed more miR-34a compared with SAECs from healthy donors at baseline and after 7 days of culture (Figures 2A and E3), and this was also reflected in EVs with both large and small EVs derived from COPD SAECs, showing enrichment for miR-34a compared with EVs derived from healthy cells (Figures 2B and 2C). The expression of miR-570 was increased in COPD SAECs compared with nonsmoker cells (see Figure E4A), but no difference was detected in large or small EVs (see Figures E4B and E4C). Thus, EV content is not a snapshot of miRNA expression from SAECs that produce them, and miR-34a packaging in EVs is an active process.

Figure 1.

Chronic obstructive pulmonary disease (COPD) SAECs produce increased numbers of extracellular vesicles (EVs). Large and small EVs produced by healthy and COPD SAECs were isolated from 7-day conditioned media. (A–D) The concentration (A and C), and size (B and D) of EVs were analyzed using NanoFCM. Each point represents a different patient. Data are mean ± SEM, analyzed using the Mann-Whitney test. *P < 0.05 (n = 9). NS = nonsmoker; SAEC = small airway epithelial cell.

Figure 2.

COPD SAEC EVs are enriched with microRNA-34a (miR-34a). Large and small EVs produced by healthy and COPD SAECs were isolated from 7-day conditioned media. (A–C) Total RNA was extracted, and miR-34a expression was determined in cells (A; n = 12), large EVs (B; n = 9), and small EVs (C; n = 9) using qRT-PCR. Each point represents a different patient. Data are mean ± SEM, analyzed using the Mann-Whitney test. *P < 0.05.

EVs Produced by SAEC Are Taken Up by Healthy Recipient SAEC

To determine whether EVs can be taken up by epithelial cells, EVs were isolated and labeled with the lipophilic fluorescent dye PKH67. Recipient healthy SAECs were treated with PKH67-labeled EVs for up to 16 hours and analyzed using flow cytometry. Internalization of EVs was time dependent, as after 3 and 16 hours of exposure to large EVs, the proportion of recipient SAECs positive for PKH67 compared with untreated cells increased from 5.7 ± 0.59% to 53.6 ± 0.83% (Figure 3A), with no difference between the EVs derived from COPD and healthy SAECs. A similar pattern was seen for the uptake of small EVs (3 h, 38.7 ± 10.2% positive cells; 16 h, 92.7 ± 0.08% positive cells), such that the proportion of recipient cells containing EVs was greater than that for larger EVs (Figure 3B). The proportion of recipient cells positive for PKH67 was not different when exposed to small EVs derived from COPD SAECs compared with healthy SAEC–derived EVs (Figure 3B). Moreover, the percentage of cells positive for EVs was not modified between recipient healthy and COPD SAECs at any time point (data not shown). Uptake is an active process, as EVs were not detectable in recipient cells when incubated at 4°C. Internalization was confirmed after exposure of cells to PKH67-labeled EVs for 6 hours using fluorescence microscopy and was lost when EVs were treated with the detergent Triton X-100 (Figure 3C). These data suggest that within 3 hours of incubation, EVs are taken up by healthy recipient SAECs.

Figure 3.

EVs are taken up by recipient SAECs. Large EVs (lEVs) and small EVs (sEVs) were isolated from healthy or COPD SAEC media, stained with PKH67, and resuspended in 300 μl media. (A and B) Healthy SAECs were stimulated with PKH67-labeled lEVs (A) or sEVs (B), and the uptake of EVs was analyzed using flow cytometry at 1, 3, 6, and 16 hours. (C and D) PKH67-labeled lEVs (C) or PKH67-labeled EVs treated with Triton X-100 (D) were incubated with 125,000 recipient healthy SAECs for 6 hours. Cells were fixed and nuclei stained with DAPI (blue) and cytoplasm with CellTrace (red) and imaged using fluorescence microscopy. Images are representative of four experiments. Data are expressed as mean ± SEM and were analyzed using two-way ANOVA with the post hoc Šidák test. Scale bars, 10 μm. *P < 0.05 and ****P < 0.0001.

COPD SAEC–derived EVs Transfer Functional MiRNA-34a into Healthy SAECs

Having shown that EVs can enter recipient cells, we assessed the transfer of vesicular miR-34a in recipient cells using EVs isolated from BEAS-2B cells transfected with a miR-34a mimic or scrambled sequence. In response to miR-34a–loaded large and small EVs, we determined that recipient BEAS-2B cells exhibit increased expression of miR-34a after 3 hours (see Figures E5A and E5C) and that the increased expression of miR-34a decreased SIRT1 mRNA after 12 hours (see Figures E5B and E5D). Moreover, expression of the protein SIRT1 decreased in recipient cells treated with miR-34a–laden large EVs but not small EVs (see Figures E5C and E5F). We next examined the functional effect of EVs on recipient cells by incubating 2.5 × 10⁵ healthy SAECs with 150 μl EVs (∼2 × 10¹¹ vesicles) derived from either healthy or COPD SAECs. Transfer of large, but not small, EVs from COPD SAECs significantly increased the expression of miR-34a (Figures 4A and E6A) and decreased the expression of SIRT1 mRNA (Figures 4B and E6B) in recipient cells compared with EVs from healthy SAECs. This effect was due to large EVs, as there was no modification in miR-34a and SIRT1 mRNA expression in recipient cells treated with EVs previously exposed to Triton X-100 or with media that had been depleted of EVs (Figures 4C and 4D). In the same way, transfer of large, but not small, EVs diminished the protein concentration of SIRT1 in recipient SAECs, and this reduction in SIRT1 was lost when EVs were dissolved by Triton X-100 or with media depleted of EVs (Figures 4E and E6C). It remained a possibility that the presence of large EVs could induce endogenous production of miR-34a in recipient cells. To address this, recipient cells were pretreated with the transcriptional inhibitor actinomycin D before exposure to EVs (see Figure E7). This did not affect the increased miR-34a expression observed in response to EVs from subjects with COPD and supported the concept that miR-34a is transferred from EVs and does not result from endogenous synthesis by recipient cells. To confirm that the effect of COPD EVs was due to a change in their cargo, recipient cells were treated with 1 × 10¹² large EVs isolated from nonsmoker SAECs or COPD SAECs. Transfer of large EVs from COPD but not healthy SAECs significantly increased the expression of miR-34a and decreased the RNA and protein expression of SIRT1 (see Figure E8). Furthermore, recipient healthy SAECs were treated with a concentration range of COPD EVs, showing that recipient SAECs treated with 5 μg EVs exhibited increased expression of miR-34a and decreased expression of SIRT1, but these effects disappeared when EVs were diluted 10 times (see Figure E9). These results further confirm that the modification of miR-34a and SIRT1 expression in recipient SAECs is driven by the modification of EV content rather than a modification of their quantity.

Figure 4.

Large EVs from COPD SAECs transfer functional miR-34a in recipient SAECs. Large EVs were isolated from 7-day conditioned media from COPD or nonsmoker SAECs. A total of 250,000 recipient healthy SAECs were stimulated with 150 μl large EVs, with EVs treated with Triton X-100 or with no EVs. (A–D) Expression of miR-34a (A and C), and SIRT1 (B and D) were measured in recipient cells after 3 hours (for miR-34a) and 12 hours (SIRT1) by qRT-PCR. (E) After 48-hour stimulation, SIRT1 protein was analyzed using western blotting. Data are expressed as mean ± SEM and were analyzed using the Kruskal-Wallis test with the post hoc Dunn test. *P < 0.05 and **P < 0.01. no EV = conditioned media depleted of EVs; NT = untreated.

Taken together, these data suggest that large EVs produced by COPD SAECs transfer miR-34a into recipient epithelial cells, which leads to the inhibition of its target, SIRT1.

COPD SAEC–derived EVs Induce Senescence Markers into Healthy SAECs

To assess whether EV-mediated reduction in SIRT1 led to an expected increase in markers of senescence, expression of p21^CIP1 and p16^INK4a, IL-6 production, and the activity of SA-βGal were measured in recipient cells after incubation for 48 hours with EVs from COPD SAECs. Transfer of large EVs from subjects with COPD significantly increased mRNA and protein expression of p21^CIP1 in recipient cells (Figures 5A and 5B). Moreover, recipient cells treated with COPD large EVs expressed significantly higher concentrations of p16^INK4a and produced significantly more IL-6 than untreated cells (Figures 5C and 5D). CXCL8 concentration was analyzed in supernatants, but no modifications in response to EVs were demonstrated (data not shown). Recipient cells treated with large EVs from COPD SAECs also exhibited increased SA-βGal compared with untreated cells, which was lost when EVs were treated with Triton X-100 (Figure 5E).

Figure 5.

Large EVs from COPD SAECs induce senescence markers in recipient SAECs. EVs were isolated from 7-day conditioned media from COPD or nonsmoker SAECs. A total of 250,000 recipient healthy SAECs were stimulated with 150 μl large EVs from NS or COPD donors. (A and B) Expression of p21^CIP1 was measured in recipient cells using (A) qRT-PCR after 12 hours and (B) western blotting after 48 hours. (C) p16^INK4a expression was measured in recipient cells using western blotting. (D) IL-6 secretion was measured in recipient cell media using ELISA after 48 hours. (E) Recipient SAECs were stained for senescence-associated β-galactosidase and imaged using light microscopy. Images are representative of four experiments. Each point represents EVs isolated from different SAEC donors (n = 6). Data are expressed as mean ± SEM and were analyzed using the Kruskal-Wallis test with the post hoc Dunn or Mann-Whitney test as appropriate. Scale bars, 100 μm. *P < 0.05 and **P < 0.01. p^16INK4a = cyclin-dependent kinase inhibitor-2A; p21^CIP1 = cyclin-dependent kinase inhibitor-1; SA = senescence-associated; SN-EVs = conditioned media depleted of EVs.

p21^CIP1 mRNA expression was also significantly increased in recipient cells exposed to COPD small EVs compared with nonsmoker small EVs (see Figure E10A), but no effect of COPD small EVs was detected on protein expression of p21^CIP1 in recipient cells after 48 hours (see Figure E10B). Similarly, IL-6 production was not modified in recipient cells treated with small EVs (see Figure E10C).

Taken together, these data suggest that large EVs produced by COPD SAECs induce expression of p21^CIP1, production of the SASP factor IL-6, and increased activity of SA-βGal, confirming the potential role of EVs in the induction of cellular senescence.

COPD SAEC–derived EVs Induce Senescence Markers in Recipient SAECs through MiR-34a Transfer

Having identified EVs as a potential vector of cellular senescence, we aimed to confirm the involvement of the miR-34a pathway. SAECs from healthy donors were transfected overnight with an antagomir against miR-34a or a control scrambled sequence before being incubated with large EVs derived from healthy or COPD SAECs (Figure 6A). Expression of miR-34a induced by large EVs from COPD cells was significantly abrogated in recipient cells transfected with the antagomir compared with cells transfected with a control sequence (Figure 6B). Consequently, recipient cell transfection with the antagomir prevented reduction of SIRT1 mRNA and protein expression (Figures 6C and 6D) and blocked the augmentation of p21^CIP1 mRNA and protein expression (Figures 6E and 6F). To further confirm the role of miR-34a contained in EVs in the induction of cellular senescence, recipient cells were treated with EVs derived from donor healthy and COPD cells transfected with an antagomir against miR-34a (see Figure E11A). Large EVs from control COPD SAECs increased miR-34a expression in the recipient cells, which was blocked when donor cells were transfected with the antagomir (see Figure E11B). Consequently, treatment of recipient cells treated with EVs isolated from COPD SAEC–antimiR-34a prevented reduction of SIRT1 mRNA and protein and augmentation of p21^CIP1 mRNA and protein expression (see Figures E11C–E11F). Moreover, p16^INK4a expression increased in cells treated with COPD EVs, and its expression diminished in cells treated with EVs isolated from COPD SAEC–antimiR-34a but nonsignificantly (see Figure E11G). Taken together, these data suggest that transfer of miR-34a via large EVs from COPD SAECs induces cellular senescence in recipient healthy cells by inhibiting the expression of SIRT1 and subsequently increasing expression of p21^CIP1. Furthermore, pretreatment of SAECs with an antagomir against miR-34a is sufficient to inhibit these effects.

Figure 6.

Large EVs from COPD SAECs induce senescence through miR-34a transfer. (A) Recipient healthy SAECs were transfected with Anta or with a Ctrl sequence overnight before treatment with large EVs isolated from NS or COPD SAEC media. (B, C, and E) Changes in (B) miR-34a, (C) SIRT1, and (E) p21 mRNA were measured using qRT-PCR after 3 and 12 hours of stimulation. (D and F) SIRT1 (D) and p21 (F) protein expression was analyzed 48 hours after recipient SAEC stimulation using western blotting. The band density of each blot is represented as a histogram. Each point represents treatment with EVs isolated from different SAEC donors. Data are expressed as mean ± SEM and were analyzed using the Wilcoxon or Friedman test with the post hoc Dunn test as appropriate. *P < 0.05. Anta = antagomiR against miR-34a; Ctrl = control sequence.

Discussion

Recent studies suggest that senescent cells are capable of inducing neighboring cells to undergo senescence through cell-to-cell contact or in a paracrine fashion (25, 26). The possible role of EVs in the induction of paracrine senescence is suggested by some studies, but the mechanism has not been fully determined (27, 28). Our experimental results have highlighted a mechanism by which EVs produced by SAECs transfer miR-34a and induce cellular senescence in recipient epithelial cells. Thereby, EVs and associated miR-34a may contribute to the spread of senescence and hence aging. As such, this mechanism may contribute to the pathophysiology of COPD but more broadly to age-related diseases and to the development of multimorbidity in elderly individuals.

The function of EVs in aging and in COPD is currently uncertain. Contradictory results have been described depending on the methods used to isolate EVs (29, 30), confirming that the analysis methods used to characterize and describe EVs are of utmost importance. Thus, and in accordance with current guidelines, we have tested several speeds and durations of ultracentrifugation and carefully selected the method that allowed the isolation of EVs (31). Nomenclature and definition of EVs are an ongoing topic of discussion, as the present markers and size may not accurately reflect the functional differences between EV populations (32). According to their biogenesis, three main biotypes of EVs are currently described: apoptotic bodies, ectosomes, and exosomes. Ectosomes originate from the outward budding at the plasma membrane, whereas exosomes are generated from the endosomal system as intraluminal vesicles that are secreted by fusion with the plasma membrane (17). Ectosomes are usually described as a heterogeneous population of vesicles whose sizes range from 100 to 1,000 μM, whereas exosomes are believed to be 30–100 μM in diameter. Markers of EVs include tetraspanins (CD9, CD81, and CD63), proteins involved in EVs biogenesis such as Alix or Tsg101, and markers of the cells from which EVs originate, such as EpCam (33). However, these size ranges do not represent pure populations and are difficult to verify because of the lack of unique and specific markers of EVs (32). As no consensus exists to precisely characterize EV subtypes, the International Society for Extracellular Vesicles has published guidelines to characterize EV-associated functions. Particularly, the society advised to refer to EV subtypes by their physical characteristics, such as their size, biochemical composition, or descriptions of conditions or cells of origin (31). In our study, we isolated two subtypes of EVs: large EVs isolated after 30 minutes of centrifugation at 20,000 × g and small EVs isolated after 2 hours at 100,000 × g. We have confirmed the lipid nature of EVs by analyzing the expression of specific markers and using a detergent to disrupt the vesicles (see Figure E3). We have analyzed the expression of common EV markers and have shown that large vesicles were predominantly positive for CD9 and EpCam and small EVs for CD63. Thus, SAECs seem to produce two subtypes of EVs that are heterogeneous in size and markers. This was also observed in a study by Willms and colleagues in which two subtypes of EVs were isolated from melanoma cells by density gradient. Even if both populations were enriched with CD63, CD81, and CD9, analysis by mass spectrometry showed that the protein composition of subtypes of EVs were different (34). Proteomic comparisons of EVs produced by dendritic cells also showed that small EVs were enriched with CD63 compared with large EVs (35). In this study, we isolated two subpopulations of EVs produced by SAECs that seem to express specific markers, but we cannot rule out that there is an overlap between our EV populations. As they are both produced by COPD SAECs, we aimed to elucidate their biological impact particularly in aging mechanisms. We then showed that SAECs from subjects with COPD produced more EVs than SAECs from age-matched healthy subjects on the basis of nanoflow cytometry (Figure 1). In agreement with these results, changes in EV production have been demonstrated in age-related diseases. The concentration of CD9-positive EVs is significantly elevated in the plasma of subjects with COPD and is correlated with proinflammatory factors, including C-reactive protein and IL-6 (23).

Vesicular miRNAs are mediators of genetic exchange between cells. After delivery, miRNAs regulate the translation of target genes and ultimately the function of the recipient cells (36). Our data showed that miR-34a associated with EVs can be transferred into recipient SAECs to suppress expression of SIRT1, leading to increased expression of senescence markers. Our results are supported by a study showing that EVs produced by mouse myoblasts in response to oxidative stress contain miR-34a, which can be transferred into bone marrow mesenchymal (stromal) cells with an increase in the senescence biomarker SA-βGal (18). MiR-34a is ubiquitously expressed and is involved in numerous cellular functions, including proliferation, cell cycle, and cellular senescence (37), and is thus implicated in numerous age-related diseases, including cancer (37) and cardiovascular diseases (38). Although the composition of EVs reflects the pathophysiological state of the source cell, miRNA content is often markedly different from the miRNA content of the parent cells (39, 40). These findings are consistent with our data showing increased expression of miR-34a and miR-570 in source SAECs from subjects with COPD but increased expression of only miR-34a in EVs derived from COPD cells. This implies that miRNAs are not passively sorted in EVs but are selectively and actively packaged. Dysregulation of miRNAs associated with EVs have been described in COPD pathology. In plasma-derived EVs, upregulation of miR-22-3p, miR-99a-5p, miR-151a-5p, miR-320b, and miR-320d and downregulation of miR-335-5p, miR-628-3p, miR-887-5p, and miR-937-3p have been observed in patients with COPD compared with smokers and nonsmoker donors (41). Another study showed that miR-23a, miR-1, miR-574, miR-152, and miR-221 were upregulated and miR-3158, miR-7706, miR-685, and miR-144 were downregulated in patients with COPD compared with healthy donors. The expression of vesicular miR-23a, miR-221, and miR-574 was negatively correlated with lung function in patients with COPD (42). Analysis of miRNAs contained in EVs isolated from BALF and lung tissues from patients with COPD has shown that miR-122-5p is downregulated in EVs derived from COPD lung tissues, whereas miR-320b and miR-22-3p are upregulated and miR-423-5p is downregulated in EVs derived from COPD BALF (43). However, these studies aimed to use miRNAs associated with EVs as biomarkers of COPD pathology and severity. Therefore, very little has been done on the biological function of vesicular miRNAs, and their role in COPD pathophysiology is yet to be elucidated. Our study was focused on miR-34a, as its biological role in the induction of cellular senescence in COPD has been well characterized. Therefore, we aim to analyze and show that this miRNA could be packed in EVs and is an essential mechanism by which EVs participate to the transmission of cellular senescence.

The alteration of EV composition in COPD and aging may be due to oxidative damage of donor cells. Proteomic analysis of EVs produced by BEAS-2B cells revealed that 33% of proteins were differentially expressed after exposure to cigarette smoke extract, with upregulation of proteins involved in cell-to-cell communication and immune responses (44). Moreover, stimulation of BEAS-2B cells in vitro with cigarette smoke extract leads to the production of EVs containing miR-210 (45). The comparison of miRNA expression in EVs isolated from BALF obtained from nonsmokers and smokers showed that smoking alters the miRNA profile, with increased the expression of miR-21 and miR-27a (46). Without any additional stimulation, EVs derived from COPD SAECs are enriched with miR-34a. Once the disease is established, oxidative stress persists in cells even after smoking cessation because of persistent inflammation or impaired endogenous antioxidant defenses (47). Thus, oxidative stress may be essential in the expression and packaging of miR-34a in EVs.

Our data have collectively shown that even if large and small EVs derived from COPD SAECs expressed higher concentrations of miR-34a than EVs derived from healthy age-matched SAECs, only large EVs triggered increased expression of miR-34a and the expression of senescence markers in recipient healthy SAECs. These results were also seen in BEAS-2B cells (see Figure E5), in which large EVs isolated from BEAS-2B cells transfected with miR-34a mimic induced increased expression of miR-34a and decreased expression of SIRT1 for up to 48 hours, whereas small EVs induced a very small increase in the expression of miR-34a after 3 hours and decreased expression of SIRT1 only after 12 hours. It has been showed that subtypes of EVs isolated from the same cells exhibit different proteins and RNA compositions and elicit different response in recipient cells (34). However, miRNA composition has not been assessed. Another hypothesis is that large and small EVs enter cells by different mechanisms and are sequestered in different compartments of recipient cells. The cellular uptake of EVs is believed to happen mainly by endocytosis pathways, including phagocytosis, pinocytosis, and clathrin-dependent and clathrin-independent endocytosis (48). There is no consensus on the cell specificity of EV uptake, which could be dependent on the interaction with a membrane receptor on target cells (49). We have shown that large and small EVs are detected and are equally able to enter in recipient healthy SAEC from 3 hours. No difference has been demonstrated whether EVs were isolated from healthy or COPD SAECs, but the entry of EVs into cells was an active process, as it was inhibited when cells were treated at 4°C. Finally, the fate of EVs within the receptor cells is not fully characterized. Once in the cells, EV content can be release in the cytosol, degraded within lysosomes, or resecreted in the extracellular space through recycled EVs. Tracking of internalized EVs has shown that 20–30% escaped the endosomes and released their content, mRNA and proteins, within the cytosol (50, 51). However, these studies have focused mainly on small EVs (or exosomes) and not on different EV subtypes. Thus, the fate of EV subtype and the release of EV contents within the cells need to be better characterized and could explain our results on large and small EVs.

Previously, the spread of cellular senescence has been ascribed to components of the SASP, which in vitro may induce senescence in healthy cells (26, 52). However, our study does not support this hypothesis, as EV-depleted cell media and the disruption of vesicles with a detergent failed to reduce SIRT1 or induce an increase in p21^CIP1 in recipient cells. This suggests that EVs are a more effective means of propagating senescence in SAECs.

Our results suggests that EVs produced by COPD SAECs induced cellular senescence of healthy SAECs, suggesting that EVs spread cellular senescence locally and participate in COPD progression. Pulmonary epithelial cells are prominent producers of EVs, together with macrophages and endothelial cells (53, 54). EVs produced by one type of cells are believed to be able to enter into other cell types (55), but it is unclear whether EVs would induce the same response into these cell types. Recent studies have shown that EVs could be engineered, for example, with glycan ligands, to trigger a response specifically from endothelial and dendritic cells (56). Also, cross-talk between two different types of cells through EVs has been demonstrated. BEAS-2B cells treated with cigarette smoke produced EVs containing miR-210 that can enter into fibroblasts and induce their differentiation into myofibroblasts by inhibiting the autophagy process (45). An elegant study by Tian and colleagues revealed that aging is a multisystem process whereby the aging of one organ can influence the aging of multiple other systems. Notably, they showed that aging of the pulmonary systems leads to faster cardiovascular aging, which in turn results in faster aging of the musculoskeletal and renal systems (57). Thus, EVs may spread senescence locally but also within the lungs and within other organs, which would account for comorbidities and multimorbidity in elderly individuals.

Conclusions

Our data demonstrate that EVs act as vehicles that are central in the induction of cellular senescence and participate in the development of COPD and age-related diseases. We showed that miR-34a could be packed into EVs produced by SAECs and be actively transferred into healthy SAECs, consequently promoting the downregulation of SIRT1 and an increase in cellular senescence. As EVs are detectable in every body fluid and have a longer half-life than free RNA, they can be measured in biological fluids as a biomarker and may also be used as a delivery system for miRNA or antagomirs and could be engineered to target specific types of cells. Thus, targeting vesicular miR-34a to reverse cellular senescence could be a novel therapeutic target in the treatment of COPD and age-related diseases.