The present research was evaluated and approved by the Ethics Board of the Croatian Veterinary Institute, approval code Z-IV-4-2022/19, May 9, 2019, and Veterinary Ethics Committee at the Faculty of Veterinary Medicine, University of Zagreb, approval code 640-01/20-17/10, February 20, 2020, 640-01/20-17/55, September 28, 2020, and 640-01/22-02/07, April 20, 2022. Canine donor’s owners provided written informed consent before sampling.

The research involved obtaining abdominal adipose tissue samples from nine clinically healthy female dogs (Canis lupus familiaris) who underwent elective ovariohysterectomy or ovariectomy surgical procedures. We decided to collect only female donors due to the abundance of adipose tissue at elective surgeries compared to male donors. The latter would need laparoscopic sampling to obtain the adequate quantity, which we considered to be against the 3R approach (20). Nevertheless, sex was not recognized as influential in cell surface marker expression, viability, proliferation, or differentiation potential (21). The acquisition of adipose tissue, cAD-MSCs extraction, and propagation adhered to the established protocol previously detailed in Krešić et al. (18). Briefly, 1–7 g of adipose tissue (Table 1), depending on the availability, was obtained from biomedical waste of ovarian mesostructure, placed into a sterile Falcon tube devoid of the medium, and stored at 4°C until transportation to the laboratory. Table 1 details the donors’ age, breed, and adipose tissue mass. The extraction of cAD-MSCs was carried out within a 2-h window from sample collection. This process involved rinsing and cutting the adipose tissue with a scalpel, followed by enzymatic digestion using 0.5% collagenase type 1 (Gibco, Cat. 17100017). The obtained stromal vascular fraction was subsequently placed into T25 cell culture flasks (Nunc, Thermo Fisher Scientific) and cultured in basal media comprising 79% Dulbecco’s Modified Eagle Medium with Low Glucose (DMEM Low Glucose) (Gibco, Cat. 31885049), 20% Fetal Bovine Serum (FBS) (Gibco; Cat. 1027010), and 1% Penicillin/Streptomycin (Sigma-Aldrich, Cat. P4333-100ML) at 37°C. media was replaced after 24 h, and when adherent cells reached 90% confluence, subculturing was performed using a solution of 0.05% Trypsin and 0.02% EDTA (Sigma-Aldrich, Cat. 59417C-500ML) until proliferation arrest was observed. All experiments conducted in this study utilized freshly isolated sterility-tested cAD-MSCs.

Gene expression analysis using an RT-qPCR array was conducted following total RNA extraction at two time points during in vitro culture, P3 and P6. The experimental procedure followed the methodology outlined previously (17), incorporating certain modifications. The cAD-MSCs reached approximately 90% confluence in T75 cell culture flasks (Nunc, Thermo Fisher Scientific). Subsequently, the cells were detached using a cell scraper (Nunc, Thermo Fisher Scientific) and centrifuged at 235 × g for 10 min. Total RNA extraction was performed using the RNeasy Mini Kit (Qiagen, Cat. 74106) with 2 M dithiothreitol (Sigma-Aldrich, Cat. 43815-1G), following the manufacturer’s instructions for approximately 10⁷ cells. The extracted RNA was stored at −80°C until gene expression analysis.

Before commencing gene expression analysis, the RNA integrity score (RIS) and 28S:18S ratio were determined using the QIAxcel RNA QC Kit (Qiagen, Cat. 929104) on the QIAxcel Advanced capillary electrophoresis device (Qiagen) following manufacturer instructions. RIS number < 7 and 28S:18S <1 were considered low-quality RNA and excluded from further analysis.

Genomic DNA elimination and complementary DNA synthesis from 800 ng of RNA was accomplished using the RT2 First Strand kit (Qiagen, Cat. 330401) as instructed by the manufacturer. Subsequently, a commercially available RT² Profiler™ PCR Array for Dog Mesenchymal Stem Cells (PAFD-082ZR, Qiagen) with SYBR Green-optimized primer assays (Qiagen, Cat. 330603) was applied, following the manufacturer’s instructions for Rotor-Gene Q (Qiagen). The array featured primers targeting 84 genes, categorized as stemness markers, MSC-specific, MSC-associated, and MSC-differentiation genes associated with osteogenesis, adipogenesis, chondrogenesis, myogenesis, and tenogenesis. Detailed information, including gene names, symbols, and NCBI sequences, was thoroughly documented in the Supplementary material 1.

Following data acquisition, normalization, and in-depth analysis were conducted using the specialized RT2 Profiler PCR Array Data Analysis software, accessible online at https://dataanalysis2.qiagen.com/pcr (accessed November 10, 2023). The software analyzes the data and performs statistical analysis, which is fully explained and provided in a detailed report (Supplementary material 1). The statistical significance was calculated based on a Student’s t-test of the replicate 2^ (− Delta CT) values for each gene in the control and treatment groups. The software automatically appointed a fold change cutoff of 2.0, equal to log₂Fold Change ±1.0. The gene expression profiling data are publicly available on the NCBI Gene Expression Omnibus (GEO) database (Accession Number GSE255585). Data visualization was carried out using GraphPad Prism 10.1.2.

The alterations in the proteome composition of the cAD-MSCs secretome during in vitro culture were also assessed at two specific time points, P3 and P6, in randomly selected six donors (6/21, 9/21, 14/21, 1/22, 6/22, and 7/22). The cells were seeded in six replicates at 10⁵ cells per milliliter in 24-well plates (Nunc, Thermo Fisher Scientific). They were conditioned in basal media (with 20% FBS) at 37°C, 5% CO_2, and 95% humidity until they reached 90% confluence. Subsequently, the culture medium was aspirated, and the cells were rinsed with 100% DMEM Low Glucose before being incubated in 2 mL of 100% DMEM Low Glucose. After 48 h, the secretome was carefully collected, centrifuged at 2,100 × g for 10 min, filtered through a sterile 0.45 μm filter (Merck), and preserved at −80°C until proteomic analysis.

For sample preparation, the secretome proteins were first reduced with dithiothreitol (Sigma Aldrich, Cat. D0632-5G) at a final concentration of 10 mM at 57°C for 30 min and subsequently alkylated with iodoacetamide (Sigma, Cat. I6125-5G) at a final concentration of 25 mM for 30 min at room temperature. The extraction of secretome proteins from the culture medium was carried out using trichloroacetic acid (Merck, Cat. 1008070100) precipitation method as previously described (22). To summarize, 1,260 μL of the medium sample was combined with sodium deoxycholate (Sigma-Aldrich, Cat. D6750-100G) at a final concentration of 0.1%. Trichloroacetic acid was added to reach a final concentration of 7.5%, causing protein precipitation on ice for 2 h. Afterward, the samples were centrifuged for 10 min at 4°C and 10,000 × g. The supernatant was discarded, and the pellet underwent two washes with ice-cold tetrahydrofuran (Merck, Cat. 1081011000). Finally, the pellet was reconstituted in 50 μL of 50 mM triethylammonium bicarbonate buffer (Thermo Fisher Scientific, Cat. 90,114 1 M). Protein concentration was determined using the Bradford assay, and the concentrations of all samples were adjusted with 50 mM triethylammonium bicarbonate buffer.

The proteins were subjected to enzymatic digestion with trypsin (Promega, Cat. V5117) at a final 0.01 mg/mL concentration, carried out over 18 h at 37°C and 20 × g. Peptide separation was executed using the nanoLC EASY-nLC™ 1200 System (Thermo Fisher Scientific) on a 75 μm × 250 mm RPC column. The gradient length spanned 2 h with 0–80% acetonitrile (VWR Chemicals, Cat. 83639.320), and 0.1% formic acid (Merck, Cat. 1002641000) with a flow rate of 1 μL/min, and the injection volume was 2 μL. The nanoLC system was connected to the Q Exactive™ Plus Hybrid Quadrupole-Orbitrap™ Mass Spectrometer (Thermo Fisher Scientific). The mass spectrometer operated in a data-dependent mode, where MS1 spectra were recorded within the range of 350–1,800 m/z at a resolution of 70,000. The top 12 ions were selected for fragmentation (MS2) and recorded at a resolution of 17,500.

Raw data were analyzed using Scaffold Quant Q + S 5.3.0 utilizing protein sequence data of Canis lupus familiaris reference proteome obtained from the UniProt database Proteome ID UP000805418 (accessed on October 30, 2023, with a total of 20,991 entries). Search parameters included the allowance of a missed trypsin cleavage, carbamidomethylation as a fixed modification, and a precursor and fragment ion mass tolerance of 10 ppm. Peptide search results were subsequently analyzed with Scaffold Quant version 5.0.3, employing untargeted label-free quantification and statistical analysis based on the spectral counting method, utilizing a t-test to verify the results of secretome composition. The proteins were filtered to include only the ones with a minimum of two identified peptide sequences. Statistical probability was defined at the p < 0.05 unless otherwise stated. A fold change cutoff of 1.3 was appointed, equal to log₂Fold Change ±0.3785. Gene Ontology (GO) Panther 18.0 was utilized in the research of cellular components, protein class, molecular function, biological process, and pathways analysis, using Fisher’s exact test type and false discovery rate correction; data presented as raw p value and false discovery rate (FDR). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (23) partner repository with the dataset identifier PXD049324 and 10.6019/PXD049324. Data visualization was carried out using GraphPad Prism 10.1.2.

The cAD-MSCs cultures from all donors were successfully established. The spindle-shaped isolated cells underwent in vitro cultivation, where morphological changes were observed with each passage, exhibiting an increase in both size and granularity, coupled with a decrease in cell population density attributable to senescence, i.e., in vitro aging. Proliferation arrest, on average, occurred at P8. Microscopic images of three representative donors in P3 and P6 were presented in Supplementary material 2.

The trilineage differentiation capacity (adipogenic, osteogenic, and chondrogenic) of all isolated cells was preserved at P3, and the cAD-MSCs were confirmed to possess differentiation potential following the International Society for Cellular Therapy (ISCT) criteria (24). Microscopic images in Figure 1 depict undifferentiated cells (A), adipogenic (B), osteogenic (C), and chondrogenic (D) differentiation from a representative 6/21 donor experiment.

Isolated cAD-MSCs at P3 underwent flow cytometry analysis, revealing the presence of markers CD90 (87.64 ± 5.3%), CD105 (41.44 ± 11.64%), CD44 (74.32 ± 6.33%), and CD29 (97.93 ± 0.65%), confirming the cells’ stem cell status according to ISCT guidelines. Notably, the expression of CD73, for which a polyclonal antibody was utilized, was not observed. The analysis also showed the absence of expression for markers CD271 (1.61 ± 1.21%), CD45 (0%), and CD34 (4.01 ± 1.65%). Figure 1E represents CD surface marker expression histograms for donor 13/21 in comparison to isotype and unstained cells. Additionally, Whisker-box plots depict the MFI fold changes for all CD markers, with donor-specific data points represented by different-colored dots (Figure 1F).

RNA quality analysis confirmed that the RNA in every sample was of high quality (Table 3). Gene expression analysis revealed that the overall change in expression between P3 and P6 was significantly downregulated for MSC-associated genes IL10 and PTPRC. Nonsignificant downregulations and upregulations were observed in stemness, MCS-specific, MSC-associated, and MSC-differentiation genes, as indicated in Figures 2A–D. Figure 2E represents fold-regulation values for downregulated and upregulated genes. Fold-regulation represents fold-change results in a biologically meaningful way. The complete RT2 Profiler PCR Array Data Analysis software report is attached as a Supplementary material 1.

During the proteomic analysis of cAD-MSCs secretome, 1,187 proteins were detected in P3 and P6, with 90% proteins in common. A complete list of detected proteins, their gene symbols, t-test significance, and fold change results, is provided in the Supplementary material 3.

Proteins detected in P3 and P6 were compared by functional annotation to cellular components, protein class, molecular function, and biological process on GO complete analysis (Figure 3). Similar involvement in all detected functions was observed. Cellular component analysis revealed that ≈75% of protein belonged to cellular anatomical entities, 22.7% belonged to protein-containing complexes, and the rest were unassigned. More than 50% of cAD-MSCs secretome proteins comprised metabolite interconversion enzyme translational proteins, protein modifying enzymes, cytoskeletal proteins, translational proteins, and protein-binding activity modulators (Figure 3A). Their molecular functions were mainly binding and catalytic activity (Figure 3B), and they were predominantly involved in cellular and metabolic processes, biological regulation, response to stimulus, and localization (Figure 3C).

Secretome analysis of commonly produced proteins showed downregulation in 648 proteins, while 302 proteins were upregulated (Figure 4). Correlation of statistically significant (p < 0.05) and fold change cutoff >1.3 proteins indicated nine biologically significant downregulated proteins—PCBP1, RPS8, HSP70, EEF1G, SRSF1, NAGA, FASN, SERPINB1, and COQ10B and 12 biologically significant upregulated proteins—DKK3a (isomer 1), PRSS23, HPRT1, LOC102152698, LOXL2, AP1B1, ME1, EIF3F, THBS3, HNRNPU, RPS19, and DKK3b (isomer 2) (Figure 4).

The 10% of explored proteins were distinctive for each passage; 63 and 52 proteins were expressed only in P3 and P6, respectively. Distinctive proteins expressed in P3 were grouped with biologically significant downregulated proteins (Group P3), while those expressed in P6 were grouped with biologically significant upregulated proteins (Group P6). Moreover, these two groups were analyzed through GO Panther Pathways analysis to see if any critical protein pathways were influenced by the in vitro aging of cAD-MSCs. Group P3 proteins showed significant involvement in cytoskeletal regulation by Rho GTPase (p = 7.45E⁻¹², FDR = 1.2E⁻⁹), nicotinic acetylcholine receptor signaling pathway (p = 1.2E⁻⁷, FDR = 9.8E⁻⁶), inflammation mediated by chemokine and cytokine signaling pathway (p = 6.7E⁻⁷, FDR = 3.6E⁻⁵), Wnt signaling pathway (p = 7.3E⁻⁵, FDR = 0.00235), and CCKR signaling map (p = 0.000601, FDR = 0.0161). In group P6, the expressed proteins were significantly related to the xanthine and guanine salvage pathway (p = 0.000193, FDR = 0.031), adenine and hypoxanthine salvage pathway (p = 0.000377, FDR = 0.0202), and blood coagulation pathway (p = 0.00032, FDR = 0.0258).