*3.1. Characteristics of the Study Participants*

The subset of participants included in the present study was representative of the whole BIOSPHERE cohort in terms of age, sex distribution, clinical characteristics, and body composition and functional parameters [8]. The main characteristics of study participants are presented in Table 2. Sex distribution, BMI, number of comorbid conditions and medications, total serum protein concentrations, and albumin levels did not differ between older adults with and without PF&S. PF&S participants tended to be older than controls, but the difference did not reach statistical significance. As per the selection criteria, SPPB scores and aLM either crude or adjusted by BMI were lower in older adults with PF&S relative to non-PF&S participants.


**Table 2.** Participant characteristics according to the presence of physical frailty and sarcopenia.

Abbreviations: aLM, appendicular lean mass; aLMBMI, aLM adjusted by body mass index (BMI); non-PF&S, non-physically frail non-sarcopenic; PF&S: physical frailty & sarcopenia; SD: standard deviation; SPPB: short physical performance battery. ¥ includes hypertension, coronary artery disease, prior stroke, peripheral vascular disease, diabetes, chronic obstructive pulmonary disease, and osteoarthritis. # includes prescription and over-the-counter drugs

*3.2. Characterization of Small Extracellular Vesicles from the Serum of Participants with and without Physical Frailty and Sarcopenia*

### 3.2.1. Verification of the Purity of Serum Small Extracellular Vesicles

The purity of sEVs obtained by serum ultracentrifugation was ascertained according to the guidelines of the International Society of Extracellular Vesicles [29]. In particular, the presence of the cytosolic protein flotilin (positive control) and the absence of the non-sEV component HNRNPA1 (negative control) were verified (Figure 1A). The purified biospecimen was also analyzed by scanning electron microscopy (SEM) to confirm enrichment in sEVs. Small EVs appear in the scanning electron micrographs as objects of spherical shape and less than 100 nm in size (Figure 1B).

**Figure 1.** (**A**) Blots of the cytosolic protein flotilin and heterogeneous nuclear ribonucleoprotein A1 (HNRNPA1) as positive and negative markers respectively, in purified small extracellular vesicles (sEVs) obtained by serum ultracentrifugation from participants with physical frailty and sarcopenia (PF&S) and non-physically frail non-sarcopenic (non-PF&S) controls. The Michigan Cancer Foundation-7 (MCF-7) cell extract was used as the positive control for the anti-HNRNPA1 antibody. (**B**) Scanning electron micrographs of purified sEVs. The white-dashed box delimitates the area zoomed on the right. White arrows indicate some of the sEVs found in the observation field. Scale bar: 100 nm.

3.2.2. Quantification of the Amount of Circulating Small Extracellular Vesicles

The total amount of sEVs purified from the serum of PF&S participants was significantly greater than in non-PF&S controls (*p* < 0.0001, Figure 2).

**Figure 2.** Serum levels of small extracellular vesicles (sEVs) in non-physically frail non-sarcopenic (non-PF&S) controls (n = 10) and participants with physical frailty and sarcopenia (PF&S; n = 11). Data were normalized for the amount of total serum proteins and are shown as percentage of the control group set at 100%. Bars represent mean values (±standard error of the mean). \* *p* < 0.05 versus non-PF&S.

#### 3.2.3. Characterization of the Origin and Cargo of Small Extracellular Vesicles

Protein levels of the two tetraspanins, CD9 and CD63, were lower in participants with PF&S than in non-PF&S controls (Figure 3A,B), while CD81 content was unvaried between groups (Figure 3C).

**Figure 3.** Protein expression of (**A**) CD9, (**B**) CD63, and (**C**) CD81 in purified small extracellular vesicles (sEVs) from non-physically frail non-sarcopenic (non-PF&S) controls (n = 10) and participants with physical frailty and sarcopenia (PF&S; n = 11). Data were normalized for the amount of sEV total proteins and are shown as percentage of the control group set at 100%. Bars represent mean values (±standard error of the mean). \* *p* < 0.0001 versus non-PF&S.

As for sEV cargo characterization, protein levels of adenosine triphosphate 5A (ATP5A; complex V), nicotinamide adenine dinucleotide reduced form (NADH):ubiquinone oxidoreductase subunit S3 (NDUFS3; complex I), and succinate dehydrogenase complex iron sulfur subunit B (SDHB; complex II) were lower in participants with PF&S than in non-PF&S controls (Figure 4A–C). No signal was detected for mitochondrial cytochrome C oxidase subunit I (MTCOI, complex IV), NADH:ubiquinone oxidoreductase subunit B8 (NDUFB8; complex I), or ubiquinol-cytochrome C reductase core protein 2 (UQCRC2; complex III) in either participant group.

**Figure 4.** Protein expression of (**A**) adenosine triphosphate 5A (ATP5A), (**B**) nicotinamide adenine dinucleotide reduced form (NADH):ubiquinone oxidoreductase subunit S3 (NDUFS3), and (**C**) succinate dehydrogenase complex iron sulfur subunit (SDHB) in purified small extracellular vesicles (sEVs) from non-physically frail non-sarcopenic (non-PF&S) controls (n = 10) and participants with physical frailty and sarcopenia (PF&S; n = 11). Data were normalized for the amount of sEV total proteins and are shown as percentage of the control group set at 100%. Bars represent mean values (±standard error of the mean). \* *p* < 0.0001 versus non-PF&S.

#### **4. Discussion**

Among the factors involved in muscle degeneration associated with PF&S, mitochondrial dysfunction and the accrual of abnormal organelles have been indicated as relevant players [30]. However, the exact mechanisms underlying mitochondrial decay are not completely deciphered.

Derangements in MQC processes have been reported in older adults with PF&S [7,31,32]. Nevertheless, alterations in sEV trafficking, which might contribute to MQC dyshomeostasis in muscle [33], have remained largely unexplored. To start filling this gap in knowledge, we purified sEVs from the serum of older adults with and without PF&S and, after ascertaining purity of the preparation, we determined the overall quantity of the mixed sEV population. Our results show a greater amount of sEVs in serum of PF&S participants compared with non-PF&S controls (Figure 2). The verification of the three tetraspanins, CD9, CD63, and CD81, in purified sEVs allowed these vesicles to be identified as a fraction of endosome-derived vesicles, referred to as exosomes, originating from the fusion of multivesicular bodies with the plasma membrane [28]. A lower protein expression of CD9 and CD63 was found in the exosome fraction purified from participants with PF&S (Figure 3), while levels of CD81 were comparable between groups. These observations are in keeping with the heterogenous composition of exosomes themselves, likely reflecting a different vesicle trafficking regulation [34]. Indeed, RAB27A, a guanosine triphosphatase (GTPase) that modulates exosome secretion, has been shown to regulate the secretion of CD63-positive exosomes, but not of those positive for CD9 [35]. Notably, exosomes derived by B-cells are characterized by the tetraspanin markers CD9 and CD81, while CD63 is absent [36]. A previous report by our group showed that RAB7A, a small GTPase and a master regulator of the late endocytic pathway, was able to modulate secretion of CD9- and CD81-positive exosomes [37]. The decreased expression of tetraspanin CD63 found in the present study may therefore be indicative of an altered late endocytic pathway [38], possibly suggesting disarrangements in late endocytic trafficking in PF&S.

The identification of mitochondrial components within the purified material allowed for classification of MDVs among sEVs. In particular, lower levels of the mitochondrial components ATP5A (complex V), NDUFS3 (complex I), and SDHB (complex II) were found in participants with PF&S (Figure 4). With the intent of preserving mitochondrial homeostasis, mitochondrial hyper-fission segregates severely damaged or unnecessary organelles [39,40] that are subsequently disposed via mitophagy [41]. However, mitochondrial-lysosomal crosstalk may dispose mildly oxidized mitochondria via MDV release [42]. Such a mechanism may therefore restore mitochondrial

homeostasis before whole-sale organelle degradation is triggered [42]. Though, in the case of defective mitophagy or disruption of the mitochondrial-lysosomal axis, accrual of damaged mitochondria, misfolded proteins, and lipofuscin may occur as a result of inefficient cellular quality control [43]. Therefore, the increased sEV secretion in participants with PF&S (Figure 2) might reflect the cell's attempt to extrude dysfunctional mitochondria. However, the reduced secretion of MDVs in the same participant group (Figure 4) may indicate that the MQC flux is impaired or that the damage to mitochondria is too severe to be disposed via MDVs. This idea is in keeping with previous reports by our group showing derangements in the expression of key proteins of the MQC machinery in old hip-fractured patients with sarcopenia [7,31].

The retrieval of mitochondrial components within sEVs is particularly relevant as it provides novel insights into the mechanisms of sterile inflammation, an age-associated inflammatory response mounted in the absence of infections [44]. This process is framed within the innate immune response and has been included as part of the "danger theory" of inflammation [45]. According to this view, misplaced noxious material from injured cells (i.e., damage-associated molecular patterns (DAMPs)) triggers caspase-1 activation and the secretion of pro-inflammatory cytokines [46]. The release of MDV content (e.g., mitochondrial proteins, mtDNA) can activate inflammatory pathways by interacting with several receptors/systems including TLRs, family pyrin domain-containing 3 (NLRP3) inflammasome, and cGAS-STING DNA sensing system [47].

Recently, we described the existence of a frailty "cytokinome" in older adults with PF&S defined by higher levels of P-selectin, C-reactive protein, and interferon-γ-induced protein 10, and lower levels of myeloperoxidase, interleukin 8, monocyte chemoattractant protein-1, macrophage inflammatory protein 1-α, and platelet-derived growth factor BB [8]. Pro-sarcopenic/pro-disability effects have traditionally been attributed to inflammation [48,49] as much as to dysfunction of anti-inflammatory pathways [49,50]. Furthermore, circulating MDVs have been identified in serum of older adults with PD and associated with a specific inflammatory profile [14]. However, the liaison among failing mitochondrial fidelity pathways, MDV secretion, and systemic inflammation may not be exclusive of neurodegeneration. Indeed, other conditions, such as HIV infection, a model of accelerated and accentuated aging [51], are characterized by pyroptotic bystander cell death and release of DAMPs that may trigger the same pathways as those identified in PD and inflamm-aging [52]. In addition, a massive release of DAMPs is acknowledged as a factor in the development of multiorgan failure in patients with severe injuries or during hemorrhagic shock [53]. Although the pathophysiology of multiple organ failure syndrome, neurodegeneration, and PF&S is heterogeneous, the release of mitochondrial DAMPs might be a converging mechanism shared by all of them. Should this assumption hold true, the scavenging of circulating mitochondrial DAMPs might represent a yet unexplored therapeutic option for the management of age-associated disarrangements, including PF&S. From this perspective, our findings are in line with the geroscience hypothesis, according to which the roots of most chronic diseases may reside in perturbations of a set of basic mechanisms (i.e., hallmarks of aging), including mitochondrial dysfunction [54].

Albeit presenting novel and promising findings, our work has limitations that need to be discussed. First of all, the cross-sectional design of the study precludes establishing cause–effect or temporal relationships between the analyzed pathways and PF&S pathophysiology. Also, although participants were carefully selected and thoroughly characterized, we cannot rule out the possibility that unknown comorbidities may have affected our results. In addition, our study provides an initial characterization of the heterogeneous population of circulating sEVs. Indeed, the analysis of the MDV cargo was limited to selected components/subunits of the mitochondrial electron transport chain. Hence, we cannot exclude that the analysis of other biomolecules, including mtDNA, that may be transported along the same road could provide additional insights into the relationship between sEV trafficking and PF&S. Finally, a deeper characterization of sEVs for their structure and content by means of transmission electron microscopy analysis is needed to confirm and expand our findings as well as to gain further information into the dynamic regulation of vesicle trafficking in PF&S.

**Supplementary Materials:** The following is available online at http://www.mdpi.com/2073-4409/9/4/973/s1: Table S1: Technical specifications of the primary antibodies used for Western immunoblotting.

**Author Contributions:** Conceptualization, A.P., C.B., E.M., F.G., and R.C.; Data curation, A.P., F.G., and R.B. (Raffaella Beli); Methodology, A.P., F.G., H.J.C.-J., and R.B. (Raffaella Beli); Writing—original draft preparation, A.P., E.M., and R.C.; Writing—review and editing, C.B., F.G., F.L., and R.B. (Raffaella Beli); Supervision, F.L., and R.B. (Roberto Bernabei); Funding acquisition, C.B. and R.B. (Roberto Bernabei). All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by Innovative Medicine Initiative-Joint Undertaking (IMI-JU #115621), AIRC (Associazione Italiana per la Ricerca sul Cancro) Investigator grant 2016 #19068 to C.B., Ministero dell'Istruzione, dell'Università e della Ricerca (MIUR) to Consorzio Interuniversitario Biotecnologie (DM 587, 08/08/2018; CIB N. 112/19 to C.B.), 2HE-PONa3\_00334 grant for the Zeiss EVO HD 15 SEM, intramural research grants from the Università Cattolica del Sacro Cuore (D3.2 2013 and D3.2 2015), and the nonprofit research foundation "Centro Studi Achille e Linda Lorenzon".

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
