Individuals aged 45 to 85 years were eligible for participation upon signing an informed consent form covering all procedures outlined in the study protocol.

Participants were excluded if they presented with one or more of the following conditions or factors: acute or chronic diseases in the decompensation stage (including cardiovascular, respiratory, gastrointestinal, urinary, or other systems); severe musculoskeletal disorders; presence of electronic cardiac rhythm-regulating devices; any conditions in which physical performance testing might pose a risk to the participant’s health or life; pregnancy.

Medical history and physical examination were performed to confirm eligibility based on inclusion and exclusion criteria and to identify sarcopenia risk factors. These included medication histories, living conditions, harmful habits, presence of chronic diseases, previous trauma, prolonged immobilization, or surgical interventions requiring extended hospitalization.

Body weight was measured to the nearest 0.1 kg using the built-in scale of the InBody 770 body composition analyzer (InBody Co., Seoul, South Korea). Measurements were taken in the morning, in underwear, with an empty bladder and on an empty stomach. Height was measured to the nearest 0.1 cm using the MSK-233 stadiometer (MSK-233, Moscow, Moscow region, Russia), without shoes. Waist and hip circumferences were measured with a tape measure, accurate to 10 mm. The results were used to calculate body mass index (BMI) as weight (kg) / height² (m²) and waist-to-hip ratio.

Handgrip strength was assessed using a DK-140 hand dynamometer (JSC “Nizhny Tagil Medical Instrument Plant”, Nizhny Tagil, Russia). Participants performed three measurements with their dominant hand, and the final result was calculated as the arithmetic mean of the three attempts. All measurements were conducted in a seated position, with the shoulder adducted and neutrally rotated, elbow flexed at 90°, forearm in a neutral position, and wrist between 0° and 30° dorsiflexion, following recommendations from the American Society of Hand Therapists. The device was zeroed prior to each measurement session. Reduced grip strength was defined as 27 kg for men and 16 kg for women.

Muscle mass was measured using the InBody 770 analyzer (InBody Co., Seoul, South Korea) utilizing 8-point tactile tetrapolar multi-frequency bioelectrical impedance analysis. The device was calibrated according to manufacturer recommendations before each measurement session. Participants were instructed to refrain from physical activity, caffeine, and food intake for at least 2 hours prior to assessment. All measurements were performed in the morning, under standardized conditions. Appendicular lean mass (ALM) was recorded, and appendicular skeletal muscle index (ASMI) was calculated as ALM (kg) divided by height squared (m²). Low muscle mass was defined as ASMI 7.0 kg/m² for men and 5.5 kg/m² for women.

Functional performance was assessed using the SPPB. The 4-meter gait speed test was conducted on a flat, unobstructed surface, with participants starting from a static standing position. Timing started when the participant-initiated movement and stopped when one foot completely crossed the 4-meter mark. A trained evaluator provided verbal encouragement and ensured consistent measurement technique. An SPPB score of $\le$8 points indicated reduced muscle function. The battery included a 4-meter gait speed test, where a result of $\le$0.8 m/s was considered diagnostic for sarcopenia.

Sarcopenia screening was conducted using the SARC-F questionnaire, which includes five questions assessing signs of reduced muscle strength: difficulty in lifting weights, walking, standing up from a chair, climbing stairs, and falling [1]. Each question offered three response options. A total score of more than 4 was interpreted as probable sarcopenia. Quality of life was assessed using the sarcopenia quality of life (SarQoL) questionnaire, consisting of 22 questions grouped into seven domains: D1 “Physical and Mental Health”; D2 “Mobility”; D3 “Body Composition”; D4 “Functionality”; D5 “Activities of Daily Living”; D6 “Leisure”; and D7 “Fears”. Each domain was scored from 0 to 100, with higher scores indicating better quality of life. This evaluation was supplemented by the SF-36 questionnaire, assessing physical (SF-36 PH) and mental (SF-36 MH) well-being on a scale from 0 to 100, where higher scores reflect better quality of life in the respective domains. Physical activity levels were assessed using the international physical activity questionnaire (IPAQ) [2]; physical inactivity was verified at 14 points for individuals under 65 years and 7 points for those over 65 years.

A total of 55 participants were included in the study, stratified into control (n = 25), presarcopenia (n = 18), and sarcopenia (n = 12) groups. Gender distribution was unequal across groups, with some gender-specific subgroups comprising as few as 2 individuals. Participants were classified into three groups according to the EWGSOP2-2019 criteria based on the presence or absence of sarcopenia indicators [3]. All descriptive statistics are shown in Table 1.

Control group (25 participants: 18 men, 7 women): clinically healthy individuals without signs of reduced muscle strength, mass, or function. Presarcopenia group (18 participants: 14 men, 4 women): participants with isolated low muscle strength but preserved functional status and muscle mass. Sarcopenia group (12 participants: 10 men, 2 women): individuals meeting all three criteria—reduced muscle strength, decreased muscle mass, and impaired muscle function. The median age of sarcopenic patients (74.0 years) was significantly higher than that of controls (58.0 years), p = 0.0001.

Extracellular vesicles were obtained through a combination of ultrafiltration and ultracentrifugation techniques. Venous blood (27 mL) was drawn into K3EDTA-coated vacutainers. Initial cell separation was carried out by centrifugation at 1000 $\times$g for 20 minutes at 4 °C. The plasma supernatant was carefully transferred to a new tube and subjected to a second centrifugation step at 14,000 $\times$g for 15 minutes at 4 °C. The resulting supernatant was then diluted with PBS to a final volume of 33 mL and passed through a 220 nm pore-size filter (PES membrane, Wuxi NEST Biotechnology Co., Ltd., Wuxi, China).

The filtered plasma was subsequently ultracentrifuged using a bucket rotor (Optima XPN 80, Beckman Coulter, Brea, CA, USA) at 4 °C for 90 minutes. The pellet obtained after ultracentrifugation was resuspended in PBS and centrifuged again under the same conditions to further purify the extracellular vesicles (EVs). The resulting EV aliquots were frozen in liquid nitrogen and stored at –80 °C until further analysis.

The size distribution and concentration of isolated vesicles were evaluated by nanoparticle tracking analysis (NTA) on a NanoSight® LM10 system (Malvern Instruments, Worcestershire, UK). For optimal NTA measurements, the EV samples were diluted with PBS at ratios of 1:100, 1:1000, and 1:10,000. Each sample underwent a one-minute analysis to determine the mean vesicle size, size distribution profile, and particle concentration.

For transmission electron microscopy (TEM) imaging, the EVs were adsorbed onto copper grids pre-coated with a formvar-carbon film, followed by negative staining with 2% phosphotungstic acid. The prepared grids were then examined using a Talos L120C electron microscope (Thermo Fisher Scientific, Waltham, MA, USA).

4 µm aldehyde/sulfate latex beads (5 µL; Molecular Probes, Eugene, OR, USA) were washed twice with 100 µL 0.1 M MES buffer (pH 5.5) by centrifugation (3000 $\times$g, 15 min, room temperature), resuspended in 25 µL MES, and incubated with 3 µg anti-CD9 antibodies (1:1500; SAA0003, Antibody System, France) for 14 h at room temperature with agitation. EV aliquots ($\sim$30 µg protein) were then incubated with 3 $\times$ 10⁵ antibody-coated beads in 150 µL PBS at 4 °C for 14 h. Binding was blocked using 0.2 M glycine for 30 min at 4 °C. Complexes were washed twice with 2% EV-depleted bovine serum in PBS and treated with 5 µL Fc receptor binding inhibitor (14-9161-73; Invitrogen, CA, USA) for 10 min at room temperature.

For tetraspanin analysis, FITC-conjugated anti-CD63, anti-CD81, and anti-CD24 antibodies (1:100, 556019; 1:100, 555675; 1:100, 562439; BD Biosciences, Germany; 5 µL/test) were added to the complexes, followed by 20 min incubation and washing. Flow cytometry was conducted using a CytoFLEX cytometer (Beckman Coulter, Brea, CA, USA) and data were analyzed with CytExpert 2.4. median fluorescence intensity (MFI) was compared against isotype and negative controls (BD Biosciences, USA). In immunofluorescence staining, negative controls are essential for verifying the accuracy and reliability of results. They distinguish specific from nonspecific staining.

Typing of EVs was performed in accordance with the recommendations of the International Society for Extracellular Vesicles (ISEV) [33]. The characterization data, including TEM, NTA and profiling of major tetraspanins, are presented in Fig. 2.

Complexes of CD9 antibody-coated aldehyde/sulfate latex beads with EVs were prepared as previously described. The resulting bead–EV complexes were washed twice with washing buffer and incubated with 5 µL of an anti-human Fc receptor binding inhibitor at room temperature for 10 minutes, followed by an additional wash.

For the detection of matrix metalloproteinases, the complexes were stained with FITC-conjugated anti-MMP9 (1:150; FAA553Hu81, Cloud-Clone Corp., Wuhan, China; 2 µL/test), PE-conjugated anti-MMP2 (1:150; FAA100Hu41, Cloud-Clone Corp., China; 2 µL/test), and APC-conjugated anti-tissue inhibitor of metalloproteinase-1 (TIMP-1) (1:150; FAA552Hu51, Cloud-Clone Corp., China; 2 µL/test) antibodies. Staining was performed for 20 minutes at room temperature, after which the samples were washed again.

Flow cytometric analysis was carried out using a CytoFLEX cytometer, and data were processed with CytExpert 2.0 software (Beckman Coulter, Brea, CA, USA). During acquisition, bead populations carrying EVs (“Beads” gate) were first identified, and 10,000 events per sample were collected at a high flow rate. The results of MMPs profiling are shown in Fig. 3.

Data analysis was conducted using specialized software packages: Statistica 12.0 (TIBCO, Palo Alto, CA, USA), JASP 0.19.3 (Eric-Jan Wagenmakers, Department of Psychological Methods, University of Amsterdam, Netherlands), and SPSS 28.0.1 (IBM, New York, NY, USA).

The normality of data distributions was assessed using the Shapiro–Wilk test and, where necessary, the Lilliefors correction. Differences between multiple independent groups were analyzed using Kruskal-Wallis test and discriminant analysis, with statistical significance set at p 0.05. The results of descriptive statistics were presented as the median and interquartile range (Me; (Q1; Q3)).

Additional analyses included posterior probability estimation, correlation analysis, principal component analysis (PCA), and Random Forest modeling. Data visualization was performed using OriginLab and OriginPro 2025b (OriginLab Corporation, Northampton, MA, USA) and R software (version 4.X.X; R Core Team, Vienna, Austria).

The Kruskal-Wallis test, employed due to the non-normal distribution of the data, revealed several variables that significantly differed across study groups (p 0.05). Among clinical markers, age, abdominal and hip circumferences, BMI, body fat percentage, ALM, ASMI, total score on the SPPB, and a number of domains from the SarQoL questionnaire showed significant changes.

From the biochemical indicators, VTN, asprosin, and METRNL demonstrated significant variation. In contrast, vesicular markers, particularly those related to MMP expression, did not reach statistical significance in the KW analysis. Among immunological markers, only the proportion of CD14–CD163+206+ cells exhibited statistically meaningful differences.

In summary, the Kruskal-Wallis test primarily highlighted changes in general clinical status and selected biochemical parameters, while vesicle and immunological markers appeared less sensitive to group differentiation when considered individually.

3.3.2.1 Comparison Between Control and Presarcopenia

In individuals with presarcopenia compared to healthy controls, a statistically significant increase in age was observed (58 $\to$ 69.5 years; p = 0.00082). Muscle strength significantly decreased (25 $\to$ 20 kg; p = 0.00003), along with a reduction in ALM (18.07 $\to$ 13.68 kg; p = 0.01618) and ASMI (7.02 $\to$ 5.87 units; p = 0.02582). Quality of life in the physical domain significantly declined, as reflected by lower scores in the SF-36 Physical Health component (64 $\to$ 54.25 units; p = 0.0126) and SarQoL domain D1 (89.2 $\to$ 72.2; p = 0.00119). Additionally, vitronectin levels significantly decreased (173.8 $\to$ 148.0 ng/mL; p = 0.03598), as did trypsin-like protease activity (75.08 $\to$ 74.12 units; p = 0.00671).

3.3.2.2 Comparison Between Control and Sarcopenia

Compared to controls, participants with sarcopenia exhibited a significant increase in age (58 $\to$ 77 years; p 0.00001) (Table 4). Significant reductions were detected in abdominal circumference (94 $\to$ 87 cm; p = 0.00117), hip circumference (93 $\to$ 87 cm; p 0.00001), BMI (24.2 $\to$ 22.2 kg/m²; p = 0.00069), body fat content (34.2 $\to$ 27.3 units; p = 0.01013), muscle strength (25 $\to$ 12.5 kg; p 0.00001), ALM (18.07 $\to$ 12.94 kg; p = 0.00259), and ASMI (7.02 $\to$ 5.25 units; p = 0.00807). Functional capacity, measured by total SPPB score, also declined markedly (11 $\to$ 6 points; p = 0.00013).

Health-related quality of life was significantly lower in sarcopenic patients, including scores for SF-36 PH (64 $\to$ 24.67 units; p = 0.00001), SarQoL D1 (89.2 $\to$ 58.2; p = 0.00001), D2 (85.6 $\to$ 62.8; p = 0.00101), D3 (79.2 $\to$ 64.8; p = 0.00174), D4 (82.4 $\to$ 57.2; p = 0.01916), D5 (81.6 $\to$ 61.6; p = 0.0015), and overall SarQoL total score (88.2 $\to$ 65.7; p = 0.00019).

Among biochemical markers, vitronectin levels significantly decreased (173.8 $\to$ 111.6 ng/mL; p = 0.00416), as did asprosin (9.32 $\to$ 6.58 ng/mL; p = 0.04956) and METRNL (1376 $\to$ 692 pg/mL; p = 0.0578). Though the last two changes are borderline, they reflect a clear downward trend.

3.3.2.3 Comparison Between Presarcopenia and Sarcopenia

In the transition from presarcopenia to sarcopenia, age significantly increased (69.5 $\to$ 77 years; p = 0.0295). Anthropometric parameters also declined: abdominal circumference (93 $\to$ 87 cm; p = 0.023), hip circumference (90 $\to$ 87 cm; p = 0.00024), BMI (23.9 $\to$ 22.2; p = 0.01067), and body fat (30.1 $\to$ 27.3 units; p = 0.01047). ALM (13.68 $\to$ 12.94 kg; p = 0.00259) and ASMI (5.87 $\to$ 5.25; p = 0.00807) continued to decrease.

A further decline was seen in SarQoL D1 scores (72.2 $\to$ 58.2; p = 0.04286). Inflammatory/metabolic markers also showed deterioration: asprosin levels (9.32 $\to$ 6.58 ng/mL; p = 0.00606) and METRNL (853 $\to$ 692 pg/mL; p = 0.03432) both significantly decreased.

The direction and magnitude of changes were evaluated by comparing median values across the groups (see Table 3), while statistical significance for each comparison is reported in Table 4.

In parallel, DA was conducted to explore the classification potential of the study variables. Here, the list of significant predictors shifted slightly. DA indicated that waist-to-hip ratio, BMI, body fat content, ALM, ASMI, total SPPB score, IPAQ, the SF-36 Physical Health component, and several SarQoL domains (D4, D5) were influential in distinguishing between groups.

Notably, DA also brought to light immunological markers that the Kruskal–Wallis test had overlooked, including CD14+CD163+, CD14–CD163+, CD68–CD163+, and CD68+CD163–206+. Vesicular markers such as TIMP-1, as well as biochemical variables like triglycerides, elastase-like activity, trypsin-like activity, and $\alpha$1-antitrypsin, also contributed meaningfully to classification according to DA.

Discriminant analysis revealed that several markers, even those lacking strong univariate differences, exhibited meaningful classification ability. This indicates their potential use as predictive features for sarcopenia diagnosis and progression modeling, warranting further validation in larger cohorts.

PCA was performed on the two groups of following quantitative variables. First group included clinical variables and the second one — immunological, vesicular and metabolical data. The grouping variable was the disease status (control, presarcopenia, sarcopenia) (Fig. 5). Variables were standardized and missing data were removed. Feature space was orthogonalized into components with calculation of variance explained by each component. PCA was used to explore the structure of relationships among inflammation markers, extracellular matrix degradation markers, and metabolic indicators across different sarcopenia stages. Unlike univariate methods such as the Kruskal–Wallis test, PCA can capture complex interdependencies between variables and identify major axes of variation without prespecifying group memberships, thus better reflecting underlying biological processes. Methodologically, the use of PCA in this study was justified, as it enabled not only the identification of differences between groups based on individual markers but also the reconstruction of the complex biological landscape of sarcopenia’s markers [36].

The PCA performed on 24 clinical variables revealed key dimensions of variability in the cohort. Principal Component 1 (PC1) and Principal Component 2 (PC2) accounted for 32.03% and 19.26% of the total variance, respectively. The cumulative variance explained by the first five components reached 73.94%, indicating a high dimensional structure typical for sarcopenia-related clinical data. All variables were centered and scaled prior to analysis. A full breakdown of explained variance is provided in Supplementary Table 1. These components were selected for biplot visualization due to their maximal contribution to sample separation and interpretability (Supplementary Table 2).

The first principal component (PCA1) showed the highest factor loadings for variables reflecting the anthropometric characteristics of the participants (circumference of abdomen, circumference of hip, waist-to-hip ratio) and skeletal muscle status (appendicular lean mass, appendicular skeletal muscle index). These parameters decreased with sarcopenia progression and demonstrated a positive correlation with BMI, which also contributed significantly to PCA1.

The second principal component (PCA2) was primarily defined by age, patient-reported quality of life (as assessed by the Sarcopenia Quality of Life questionnaire domains D1, D2, D3, D4, D5, D7 and total score; SARC-F score; and SF-36 physical and mental health components), and functional capacity (maximal muscle strength, Short Physical Performance Battery). Disease severity showed a positive correlation with patient age and the SARC-F score, a trend most evident in the control and sarcopenia groups.

Body fat percentage did not contribute to either PCA1 or PCA2. However, it was strongly negatively correlated with physical activity level (International Physical Activity Questionnaire), gait speed, and Domain 6 of the SarQoL questionnaire, which reflects patients’ leisure activity.

The presented PCA biplot confirms a fundamental relationship between patients’ quality of life and sarcopenia severity and reveals bidirectional associations between these clinical characteristics and functional status (Fig. 6).

The PCA performed on 26 laboratory markers allowed for the identification of key biochemical, immunological, and metabolic features characterizing the transition from normal status to presarcopenia and sarcopenia. PC1 and PC2 explain 18.28% and 11.90% of the total variance, respectively. The total variance explained by the first two components is 30.18%. All variables were standardized prior to analysis (centered and scaled). Although this level of explained variance is moderate, it reflects the biological heterogeneity of the cohort and is typical for high-dimensional clinical datasets. A full breakdown of variance explained by each component is provided in Supplementary Table 1. The use of the complete set of variables and a careful analysis of their localization on the biplot and their contributions to PC1 and PC2 provided a more comprehensive understanding of the underlying pathogenic processes (Supplementary Table 2).

The first principal component (PC1) mainly reflected changes associated with enhanced inflammation, activation of matrix metalloproteinases, and ECM degradation. High contributions to PC1 were made by markers of ECM destruction (MMP9+, MMP2+, MMP2+9+TIMP+, MMP9+2–TIMP+), as well as by inflammatory macrophage subpopulations (CD68+163–206+, CD14–163+). These findings indicate the progressive remodeling of tissues and chronic inflammation correlated with the severity of sarcopenia.

The second principal component (PC2) emphasized differences associated with metabolic disturbances (Glucose, asprosin, triglycerides) and changes in macrophage activation profiles (e.g., CD68+163+, CD14+163+), reflecting compensatory or transitional processes that are particularly prominent at the presarcopenia stage.

Functional clustering of markers, based on their contribution to PC1/PC2 and spatial distribution on the biplot, allowed for the identification of four key groups.

3.4.2.1 Extracellular Matrix Degradation Cluster

This cluster includes MMP2+, MMP9+, MMP2+9+TIMP+, MMP2+9+TIMP–, MMP9+2–TIMP+, MMP9+2+TIMP–, and TIMP-1 as a metalloproteinase inhibitor. Their vectors were predominantly directed into the first and second quadrants, and a positive correlation was noted between metalloproteinase activity and sarcopenia progression. The increased expression of these markers accompanies the transition from healthy status to sarcopenia, reflecting ECM destruction and insufficient regulatory control.

3.4.2.2 Inflammation and Macrophage Activation Cluster

This group includes markers such as CD14–163+, CD14+163–, CD14–163+206+, CD68–163+, CD68–163+206+, and CD68+163–206+, representing inflammatory macrophage activation. Their contribution to PC1 is particularly significant, and their vectors are also directed mainly toward the first and second quadrants. This indicates that inflammatory response and tissue destructive processes are closely linked and progressively intensify in sarcopenia.

3.4.2.3 Homeostasis and Stabilization Cluster

This cluster includes CD14+163+, CD68+163+, CD14+163+206+, CD68+163+206+, and MMP9+2+TIMP+. These markers demonstrate smaller contributions to PC1 and moderate contributions to PC2, with vectors localized near the center of the biplot, corresponding to the control group. Their presence indicates preserved tissue homeostasis and a balance between degradation and remodeling, characteristic of healthy tissues.

3.4.2.4 Metabolic Disturbance Cluster

This group consists of glucose, asprosin, triglycerides, C-peptide, METRNL, and TP. Their contributions are more prominent along the PC2 axis, and their vector direction reflects energy metabolism disturbances characteristic of late-stage sarcopenia. Increased glucose and adipokine levels, along with disrupted lipid metabolism, indicate systemic metabolic changes associated with muscle mass and function loss.

Special attention should be given to $\alpha$1-antitrypsin, a marker of acute-phase inflammation, which is located within the region of active inflammatory response, confirming its involvement in sarcopenia pathogenesis. Elastase and VTN also contribute to tissue remodeling processes, emphasizing the importance of the imbalance between ECM destruction and regeneration.

Integration of PCA results demonstrated that the control group is characterized by the predominance of stabilizing macrophage profiles, moderate metalloproteinase activity, and preserved metabolic reserve capacity. In the early stages of sarcopenia, there is a noticeable rise in the activity of metalloproteinases, chronic inflammation, and significant metabolic disturbances (Table 5, Ref. [6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]).

Table 5. Results of principal component analysis (PCA) for laboratory indicators, including principal component coefficients, group categorization, biological interpretation, vector orientation, and type of correlation with principal components.

Laboratory indicator	PC1 coefficient	PC2 coefficient	Contribution to PC1 (%)	Contribution to PC2 (%)	Total contribution (%)	Quadrant	Correlation type	Interpretation	Link source
CD14+163+	–0.20367	–0.1296	5.9	2.32	8.22	III	positive	Anti-inflammatory macrophages, typical for control group.	[6]
CD14–163+	0.084272	0.119334	1.01	1.96	2.97	I	positive	Activation of inflammatory macrophages, increases with sarcopenia.	[7, 8]
CD14+163–	–0.1891	0.039169	5.08	0.21	5.3	II	negative	Transitional macrophage profile, initial inflammatory activation.	[9, 10, 11]
CD14+163+206–	0.033763	0.301404	0.16	12.53	12.7	I	positive	Moderately activated macrophages, early changes in sarcopenia.
CD14–163+206+	–0.01947	0.240272	0.05	7.96	8.02	II	negative	Inflammatory macrophages associated with tissue destruction.
CD14+163–206+	–0.00783	0.280749	0.01	10.87	10.88	II	negative	Transition to inflammation, onset of ECM remodeling.
MMP9+	0.105089	0.233195	1.57	7.5	9.07	I	positive	Active ECM degradation, sarcopenia.	[12]
TIMP-1	–0.09703	–0.21475	1.34	6.36	7.7	III	positive	Metalloproteinase inhibitor, imbalance in sarcopenia.	[13]
MMP2+	0.049124	0.257751	0.34	9.17	9.51	I	positive	Connective tissue destruction, activated in sarcopenia.	[14, 15, 16, 17, 18]
MMP2+9+TIMP+	–0.11119	–0.00665	1.76	0.01	1.76	III	positive	Complex ECM remodeling imbalance.
MMP2+9+TIMP–	0.099876	–0.03559	1.42	0.17	1.59	IV	negative	ECM destruction without inhibitory control.
MMP2+9–TIMP+	–0.209	–0.01623	6.21	0.04	6.25	III	positive	Controlled MMP activation, typical for presarcopenia.
MMP9+2+TIMP+	0.220789	–0.01712	6.93	0.04	6.97	IV	negative	Balanced ECM degradation, transitional phase.
MMP9+2+TIMP–	–0.20037	0.031233	5.71	0.13	5.84	II	negative	Enhanced ECM destruction, sarcopenia.
MMP9+2–TIMP+	0.109235	–0.05606	1.7	0.43	2.13	IV	negative	Active ECM destruction, late-stage sarcopenia.
VTN	0.130897	–0.12311	2.44	2.09	4.53	IV	negative	Vitronectin, modulation of cell adhesion, decreased in sarcopenia.	[19, 20]
Asprosin	0.202931	0.104386	5.85	1.5	7.36	I	positive	Regulator of glucose metabolism, increased in sarcopenia.	[21, 22]
METRNL	0.16293	0.109547	3.77	1.66	5.43	I	positive	Myokine response to inflammation, possible compensatory mechanism.	[23]
Glucose	0.19243	–0.0076	5.26	0.01	5.27	IV	negative	Glucose metabolism disturbance, associated with sarcopenia.	[24, 25]
Triglycerides	0.192231	–0.00294	5.25	0	5.25	IV	negative	Dyslipidemia, metabolic component of sarcopenia.	[26]
C-peptide	0.159222	0.077288	3.6	0.82	4.43	I	positive	Indicator of insulin secretion, impaired in sarcopenia.	[27]
Elastase-like activity	0.192222	–0.02785	5.25	0.11	5.36	IV	negative	Destruction of elastic fibers, tissue remodeling.	[28, 29]
Trypsin-like activity	0.208311	0.023308	6.17	0.07	6.24	I	positive	Total protein, general nutritional status, decreased in sarcopenia.	[30, 31]
$\alpha$1-antitrypsin	0.100607	–0.0813	1.44	0.91	2.35	IV	negative	Acute phase inflammation, activated in sarcopenia.	[32]

The authors acknowledge several limitations related to the study design and statistical methodology, which should be considered for a balanced interpretation of the findings. The cross-sectional nature of the study does not allow for definitive conclusions about causality. The patterns of biomarkers distribution were associated with sarcopenia severity, it remains unclear whether these changes are primary causes, compensatory mechanisms, or secondary effects. The prospective follow-up over 6–12 months is warranted to assess the dynamics and prognostic value of these indicators.

The patient stratification was performed according to the EWGSOP2 criteria, the practical implementation of these criteria (e.g., the handgrip strength measurement protocol, ASMI calculation using bioimpedance analysis, and threshold values for low muscle mass and strength) requires further standardization to enhance reproducibility and external validity.

One of the key limitations of this study is the relatively small and uneven sample size across defined subgroups. In particular, gender-stratified subgroups occasionally included as few as two individuals, which limits the statistical power for detecting subtle differences and increases the risk of type II error. The resulting imbalance should be considered when interpreting group-wise comparisons, especially in exploratory analyses involving immune or vesicular markers. Small sample size (n = 55) can limit the statistical power, particularly in multivariate analyses. PCA was used as an exploratory tool to identify biological gradients; however, clusters observed in biplots were interpreted visually and were not statistically validated using formal tests such as PERMANOVA. These facts may create a risk of overinterpreting visually apparent group separations and underscores the need for replication in larger cohorts.

Furthermore, no adjustments were made for potential confounding factors. Variables such as age, body mass index, and chronic conditions (e.g., diabetes mellitus and chronic kidney disease) can influence both metabolic and inflammatory markers. Their effects may overlap with or obscure sarcopenia-specific processes. Future analyses should incorporate multivariate approaches, such as generalized linear models (GLMs), to evaluate independent associations with sarcopenia status.

Although PCA enabled visualization of marker clusters and gradients, it has interpretational limitations. Variables that differed significantly between groups in univariate tests (e.g., METRNL) made only minor contributions to the first two principal components. Therefore, PCA results should be interpreted in conjunction with supervised methods and univariate statistics.

Finally, discriminant analysis and PCA helped to identify several potentially informative biomarkers, a complete diagnostic model was not developed in the current study. Receiver Operating Characteristic (ROC) analysis and Least Absolute Shrinkage and Selection Operator (LASSO) regression were not performed, and diagnostic cut-off values were not proposed. These steps are necessary to evaluate the clinical utility of the markers and should be included in future research.

In summary, the cross-sectional design, limited sample size, and lack of certain statistical adjustments call for cautious interpretation. Longitudinal, statistically robust studies with larger samples are required to validate the observed patterns and to develop reliable diagnostic and prognostic models.

Naturally, despite the many advantages of flow cytometry, both in the typing of mononuclear cells and in the subpopulation analysis of EVs protein markers, several limitations of the method must be acknowledged. Common issues include nonspecific staining, spectral overlap of fluorescent labels (necessitating compensation), and the instability of antibody-protein complexes, both on the surface of cells and on EVs surfaces [33]. For cell-based assays, there is also the problem of accurately representing subpopulation ratios due to cell disruption during sample preparation and pre-analytical handling. In the case of EVs analysis, staining procedures involve the additional challenge of unstable adsorption of EVs onto latex beads.

Other important limitations include background noise caused by antibody aggregates or unbound fluorophores, the potential for epitope masking due to conformational changes or competitive binding, and a general difficulty in detecting low-abundance markers owing to the small surface area of EVs. Furthermore, flow cytometry analysis of EVs is inherently limited by the instrument’s detection threshold, leading to the exclusion of small-sized vesicles or the misclassification of EVs subtypes.

As a result, truly accurate evaluation of immunological and vesicular markers requires further methodological refinement and the incorporation of complex correction strategies. Nevertheless, certain intrinsic limitations—such as physical detection limits, antibody specificity, and pre-analytical variability—will inevitably continue to distort the true results of the analysis, regardless of technical improvements.

The role of MMPs and their tissue inhibitor (TIMP-1) undoubtedly reflects ECM remodeling, a central component in the pathogenesis of sarcopenia. In our study, MMP2, MMP9, and TIMP-1 were assessed on the surface of CD9-positive EVs derived from the general circulation [17, 34]. Although individual comparisons between clinical groups did not reach statistical significance in univariate analyses (e.g., Kruskal–Wallis test), these vesicular markers contributed substantially to the PC1, which aligned with clinical severity and grouped inflammatory and matrix-remodeling variables. This suggests that, despite limited univariate discrimination, EV-associated MMPs carry multivariate value in the biological stratification of sarcopenia [17, 35].

However, the examination of these vesicular markers in the aggregate EV populations is limited by a lack of tissue-specificity. The observed differences may be attributed to the variations in clinical characteristics between the groups, such as age and body mass index. This is particularly true for markers that are linked to multiple age-related conditions, including inflammation, metabolic processes, mitochondrial function, and extracellular matrix remodeling. The detection of MMPs on CD9+ EVs may represent a generalized picture of ECM remodeling within the organism, including signals from diverse tissues such as vascular endothelium, adipose tissue, or smooth muscle. Therefore, while informative in a systems biology context, this approach cannot distinguish muscle-specific remodeling events from systemic matrix turnover [36].

To address this limitation, future approaches should focus on identifying a reliable skeletal muscle–specific marker detectable on circulating EVs. Such a marker should ideally be associated with structural or contractile proteins specific to muscle fibers or their resident cells — e.g., desmin, titin, myosin heavy chains, or caveolin-3 [37, 38, 39, 40]. The detection of EVs co-expressing MMPs and muscle-specific proteins would allow for more precise attribution of matrix remodeling to skeletal muscle, rather than other tissues undergoing parallel degeneration or remodeling processes.

Moreover, the need to differentiate skeletal muscle–derived EVs from those of smooth muscle and cardiac origin remains critical. Both vascular and myocardial remodeling can influence the EV cargo profile, particularly under conditions common in older adults, such as hypertension, atherosclerosis, or heart failure. Misattributing such vesicles to skeletal muscle could lead to erroneous conclusions regarding sarcopenia-specific processes. Incorporating muscle-specific EV markers, possibly in combination with transcriptomic or proteomic validation, will be essential to improving diagnostic and prognostic utility in sarcopenia research.

The interpretation of biochemical and immunological markers provides essential insights into the biological mechanisms underpinning sarcopenia progression. Our data support stage-specific trajectories of change, reflected both in median values and in statistically significant post hoc differences.

Among metabolic indicators, asprosin and METRNL demonstrated statistically significant increases across disease stages. Asprosin levels rose in serum from a median of 1.35 in controls to 1.74 in sarcopenia (p = 0.0036), while METRNL increased from 1.08 to 1.34 (p = 0.0067), indicating possible roles in energy mobilization and muscle regeneration. Vitronectin (VTN) levels decreased between presarcopenia and sarcopenia (medians 5.12 vs. 3.55; p = 0.0041), suggesting progressive disruption in extracellular matrix homeostasis.

Proteolytic enzyme activity markers showed consistent changes: elastase increased from control to presarcopenia (0.85 vs. 1.15; p = 0.0348), remaining elevated in sarcopenia (1.15). Similarly, trypsin-like proteolytic activity (TP) rose from 0.86 to 1.19 (p = 0.0139), indicating enhanced proteolytic stress during early and mid-stage muscle decline.

Among immunological variables, the CD14–CD163+CD206+ monocyte population exhibited a U-shaped dynamic: a significant decrease in presarcopenia compared to control (6.6 vs. 4.6; p = 0.03598), followed by a moderate increase in sarcopenia (6.2). This non-linear trend may reflect transient depletion of anti-inflammatory M2-like monocytes in early decline, with partial restoration in advanced disease. Additionally, significant differences were observed for CD68+CD163+, CD14+CD163+, and CD68–CD163+ populations, with decreased levels in sarcopenia relative to control or presarcopenia, supporting the hypothesis of altered monocyte/macrophage polarization along the sarcopenia trajectory.

These patterns are quantitatively supported by both median comparisons (Table 3) and post hoc significance (Table 4). Together, they indicate a progressive inflammatory and proteolytic shift during disease development, involving specific monocyte subsets and circulating biochemical regulators.

Inflammation plays a crucial role in the pathogenesis of sarcopenia, contributing to muscle loss and dysfunction. Chronic low-grade inflammation, often observed in aging, can lead to the release of pro-inflammatory cytokines, which in turn can impair muscle protein synthesis and promote muscle catabolism [3]. Additionally, inflammation can affect the ECM remodeling process. The ECM provides structural support and plays a key role in muscle repair and regeneration. Dysregulation of ECM remodeling, characterized by altered expression of MMPs and their inhibitors (TIMPs), can disrupt the balance between muscle degeneration and regeneration, further exacerbating muscle loss in sarcopenia [41].

Sarcopenia is characterized by progressive muscle mass, strength, and function loss, often linked to aging. Recent research emphasizes the roles of chronic inflammation and proteolytic dysregulation in its development.

Immunological markers, notably CD14–CD163+206+ cells, show significant deviations in sarcopenia, indicating a chronic inflammatory state. These cells suggest an ongoing inflammatory response contributing to muscle tissue degradation. Proteolysis markers also impact ECM remodelling. Elastase-like, trypsin-like activities, and $\alpha$1-antitrypsin are involved in pre-sarcopenia and sarcopenia, breaking down ECM proteins and leading to muscle tissue degradation and function loss.

Chronic inflammation, proteolytic dysregulation, and metabolic perturbations are interconnected. Inflammation releases pro-inflammatory cytokines, activating proteolytic pathways and disrupting metabolism. Proteolytic dysregulation breaks down muscle proteins, while metabolic perturbations affect energy production and nutrient utilization.

Understanding sarcopenia’s multifactorial nature is crucial for developing interventions. Targeting inflammation, proteolytic dysregulation, and metabolic issues may be effective. Anti-inflammatory therapies, protease inhibitors, and metabolic modulators could help preserve muscle mass and function, particularly in at-risk individuals.

Presarcopenia refers to the early stages of sarcopenia, where muscle mass and function begin to decline but have not yet reached clinically significant levels [3]. Early metabolic destabilization in presarcopenia involves alterations in energy metabolism, insulin sensitivity, and mitochondrial function. These metabolic changes can precede overt muscle loss and may contribute to the development of sarcopenia by impairing muscle energy production and utilization [42, 43]. Understanding the mechanisms underlying early metabolic destabilization is crucial for developing interventions that can slow or prevent the progression of sarcopenia.

The clinical markers exhibited significant intergroup disparities in age, abdominal and hip circumferences, BMI, body fat percentage, ALM, ASMI, SPPB total score, and SarQoL domains. These findings underscore the multifaceted and complex nature of sarcopenia’s impact on both physical performance and quality of life, highlighting the need for a comprehensive approach to its assessment and management.

To further elucidate the discriminatory power of these variables, a discriminant analysis was meticulously conducted to classify the groups. The results revealed a slight shift in the significance of the predictors. Specifically, the analysis identified waist-to-hip ratio (WHR), BMI, body fat content, lean mass, ASMI, SPPB score, physical activity levels (measured via the International Physical Activity Questionnaire — IPAQ), the physical health component of the SF-36, and certain SarQoL domains (D4, D5) as influential in distinguishing between the groups. These findings underscore the importance of a multifactorial approach in the diagnosis and classification of sarcopenia, emphasizing the integration of anthropometric, functional, and quality of life metrics to capture the full spectrum of the condition’s impact.

Macrophages are immune cells that play a key role in tissue repair, regeneration, and inflammation. In the context of sarcopenia, macrophages can be indicators of muscle tissue status due to their involvement in muscle repair and inflammatory processes [6]. The balance between pro-inflammatory (M1) and anti-inflammatory (M2) macrophages is crucial for muscle homeostasis. An imbalance favoring M1 macrophages can promote inflammation and muscle damage, while M2 macrophages are involved in tissue repair and regeneration. Assessing the polarization and activity of macrophages in muscle tissue can provide valuable insights into the underlying mechanisms of sarcopenia and potential therapeutic targets [9, 44]. For example, strategies aimed at modulating macrophage polarization towards an anti-inflammatory phenotype may help to reduce inflammation and support muscle repair in sarcopenia.

The distribution of CD14–CD163+206+ cells, exhibited significant disparities, underscoring the chronic inflammatory milieu associated with sarcopenia. The multivariate analysis further delineated immunological markers not discerned by the Kruskal–Wallis test, including CD14+CD163+ and CD14–CD163+ subpopulations, as well as CD68–CD163+ and CD68+CD163–206+ subsets. These findings provide a nuanced understanding of the immunological landscape in sarcopenia, shedding light on potential therapeutic targets and mechanisms underlying this geriatric syndrome.

EVs are membrane-bound particles released by cells into the extracellular environment. They play a crucial role in cell-to-cell communication and can carry various bioactive molecules, including proteins, lipids, and nucleic acids [44, 45, 46, 47]. In the context of sarcopenia, EVs can serve as integrative markers due to their ability to reflect the physiological and pathological status of muscle tissue. For example, changes in the composition and function of EVs released by muscle cells can indicate alterations in muscle metabolism, inflammation, and ECM remodeling [19, 48]. Furthermore, EVs can influence the behavior of recipient cells, such as macrophages and fibroblasts, thereby modulating muscle repair and regeneration processes. Investigating the role of EVs in sarcopenia may provide new insights into the pathogenesis of the disease and identify potential biomarkers for early detection and monitoring of sarcopenia [49].

Vesicular markers, such as TIMP-1, are molecules that can be found in vesicles — small membrane structures released by cells. They play a crucial role in intercellular communication and can transport various biological molecules, including proteins, lipids, and nucleic acids. The significance of vesicular markers as integrative indicators lies in their ability to reflect the condition of both tissues and blood. This makes them valuable tools for diagnosing and monitoring various diseases [50, 51].

TIMP-1 is one such vesicular marker that can be used to assess the state of tissues and blood. It is involved in regulating tissue remodeling, inflammation, and angiogenesis. The level of TIMP-1 in vesicles can change in various pathological conditions, such as cancer, cardiovascular diseases, fibrosis, and others.

As integrative indicators, vesicular markers, including TIMP-1, can provide information about the state of tissues and blood simultaneously. For example, changes in TIMP-1 levels in vesicles may indicate disturbances in tissue remodeling or inflammation processes, which may be associated with the development of certain diseases. Additionally, the analysis of vesicular markers can help in assessing the effectiveness of treatment and monitoring the patient’s condition. Therefore, vesicular markers, such as TIMP-1, are valuable tools for investigating the state of tissues and blood, as well as for diagnosing and monitoring various diseases.