Animal-origin-free medium enhances G2/M progression while maintaining identity and genomic stability in human adipose-derived MSCs

Abstract

Background

Mesenchymal stem cells (MSCs) are considered promising candidates for regenerative and immunomodulatory therapies. However, their clinical use requires xeno-free, good manufacturing practice (GMP)-compatible culture systems. Conventional serum-containing media (FCM) present variability and xenogeneic risk; as such, serum-free (SFM) and animal‑origin‑free (AOF) media have been produced to enhance reproducibility and safety. However, comparative analyses of MSC phenotype, function, and transcriptomic integrity across these media are scarce. Herein, we compared the features of MSC cultures in these different media.

Methods

Human adipose-derived MSCs from three donors were expanded in FCM, FDA-approved SFM, and AOF media at early, mid, and late passages. The growth kinetics, morphology, immunophenotype, and trilineage potential of these MSCs were assessed. Transcriptomic profiling (RNA-seq), pathway analysis, cell cycle inference, qRT-PCR confirmation, and copy number variation (CNV) analyses were further performed.

Results

AOF-supported MSCs exhibited the shortest population doubling time and stable morphology. AOF upregulated G2/M-phase regulators, including Cyclin B-CDK1 complexes and the anaphase-promoting complex/cyclosome, while downregulating negative proliferation and differentiation pathways. Cell cycle profiling confirmed G2/M enrichment. MSC identity (ISCT markers) and low inflammatory profiles, including reduced NGF and CSF1 levels, were maintained. CNV analysis showed no significant differences in genomic instability across media.

Conclusions

AOF medium promoted MSC proliferation through selective G2/M activation while maintaining MSC identity, immunomodulation-related marker profiles, and genomic stability under the tested conditions.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13287-026-05149-4.

Background

Mesenchymal stem cells (MSCs) are multipotent progenitor cells that can self-renew and differentiate into various mesodermal lineages, including osteoblasts, adipocytes, and chondrocytes [1]. The regenerative potential and immunomodulatory properties of MSCs make them promising candidates for cell-based therapies in regenerative medicine and for inflammatory diseases. However, the successful clinical translation of MSC-based therapies requires the establishment of standardised, reproducible, and xeno-free culture conditions that ensure cell identity, multipotency, and safety [2].

The reliance on foetal bovine serum (FBS) as a supplement to conventional media remains a critical challenge in MSC culture. While FBS provides essential nutrients and growth factors, it introduces batch-to-batch variability, undefined components, xenogeneic immune responses, and potential pathogen transmission risks, all of which limit its applicability for clinical use [3, 4]. To address these challenges, serum-free media (SFM) and animal origin-free media (AOF) have been developed [5]. These media both offer chemically-defined formulations that improve the reproducibility and reduce the immunogenicity of MSC cultures. Consequently, they are particularly crucial for ensuring good manufacturing practice (GMP)-compliant MSC production, as they enhance regulatory compliance and facilitate the transition to clinically applicable MSC therapies [6].

Given that MSCs are highly responsive to microenvironmental cues, such as extracellular matrix composition, soluble factors, and physical stimuli, alterations in culture conditions can significantly affect gene expression and functional potential. Indeed, prior reviews have highlighted how MSC behaviour is tightly regulated by their niche, and how environmental or epigenetic perturbations can modulate cell fate and therapeutic efficacy [7, 8]. Although prior studies have explored the proliferation and differentiation potential of MSCs cultured in FBS-containing media (FCM) and SFM, a comprehensive genetic and transcriptomic evaluation comparing MSCs expanded in FCM, SFM (e.g. FDA-cleared StemPro MSC SFM), and AOF remains lacking.

Beyond growth and differentiation, the composition of the MSC culture media influences cell secretory profiles, immunomodulatory potency, and genomic integrity, and thus directly affect benefit–risk assessments for clinical use [9, 10]. Regulatory agencies increasingly encourage the use of chemically defined, xeno-free conditions to minimise adventitious agents and lot variability, as well as to facilitate comparability across manufacturing sites [11, 12]. Nevertheless, evaluations that integrate the surface phenotype, proliferation kinetics, trilineage potential, cell-cycle behaviour, transcriptome-wide signatures, inflammatory marker expression, and genome-scale copy-number stability across the commonly used media remain scarce [13, 14].

To address this knowledge gap, in the present study, we systematically compared the features of adipose-derived MSCs from three donors maintained over early, mid, and late passages in conventional FCM, FDA-cleared SFM, and full AOF medium. By combining flow-cytometric immunophenotyping, growth and morphology metrics, differentiation assays, bulk RNA-seq with pathway/network analysis, cell-cycle inference, targeted qRT-PCR, and copy number variation (CNV) calling from RNA-seq, we delineated how media composition shapes MSC proliferation and differentiation programs, while preserving identity and safety-relevant attributes [15, 16]. These data will guide the selection of GMP-compatible media for scalable, reproducible, and clinically safer MSC manufacturing [9, 10].

In the present study, we evaluated the phenotypic, functional, and transcriptomic differences between MSCs cultured in FCM, SFM, and AOF media. This study focused on the reproducibility, potency, and safety under distinct culture conditions. To elucidate this effect, we employed a systematic set of methods, including RNA sequencing, flow cytometry-based immunophenotyping, proliferation assays, and multilineage differentiation assessments.

Materials and methods

Preparation of adipose-derived MSCs

Human adipose tissue samples were obtained from three healthy donors. Donor #1 was a female in her late 20s from the Republic of Korea, Donor #2 was a female in her early 30s from the Republic of Korea, and Donor #3 was a male in his late 20s from the Republic of Korea. The tissues were collected via surgical excision during orthognathic surgery at Dacapo Dental Hospital (Seoul, Republic of Korea). Approximately 50 to 100 g of discarded subcutaneous adipose tissue was harvested per donor. All procedures were performed with informed consent from the donors, and the study protocol was approved by the Institutional Review Board (IRB) of Xcell Therapeutics Inc. (IRB No. P01-201903-31-008).

The collected tissues were placed in sterile specimen bottles and transported to the laboratory on ice within 4 h. Upon arrival, the adipose tissue was finely minced, washed with Dulbecco’s phosphate-buffered saline (DPBS) supplemented with 1% penicillin/streptomycin (Cytiva, Marlborough, MA, USA), and centrifuged at 400 × g for 5 min at 20 °C to recover the tissue fragments.

For enzymatic dissociation, the minced tissue was digested with type I collagenase (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) at 3 mg/mL (approximately 840 U/mL) at 37 °C for 60–90 min with gentle agitation at 34 rpm using a multi-mixer with rolling and rocking motion in a CO₂ incubator until the tissue was sufficiently dissociated. The digestion reaction was terminated by the addition of 0.5 M EDTA solution (Lonza, Walkersville, MD, USA), followed by centrifugation at 400 × g for 5 min at 20 °C. The resulting pellet was resuspended, filtered through a 100-µm cell strainer to remove undigested debris, and centrifuged again under the same conditions to isolate the stromal vascular fraction. Erythrocyte lysis was not performed.

The isolated cells were then seeded in either a serum-containing medium or a chemically defined serum-free medium (CellCor™ SFD MSC, Xcell Therapeutics, Seoul, Republic of Korea). After 3 days of incubation, successful cell attachment to the culture dish was confirmed. Subsequently, the adipose-derived mesenchymal stem cells (AD-MSCs) were expanded and subcultured in their respective media.

Cell culture and expansion

MSCs from three donors were collected at passage 2 and seeded directly into different 25T flasks, each containing a distinct medium. FCM was composed of low-glucose DMEM (Cytiva, Uppsala, Sweden) supplemented with 10% FBS (HyClone, South Logan, UT, USA) and 1× penicillin/streptomycin (P/S; Sigma-Aldrich, St. Louis, MO, USA). The SFM was composed of StemPro^® MSC SFM XenoFree (Gibco), supplemented with 1× GlutaMax (Gibco) and P/S. AOF medium was composed of CellCor™ MSC CD AOF (Xcell Therapeutics) with P/S. The SFM medium cultured cells were coated with CTS CELLstart™ (diluted 1:50 in DPBS with Ca²⁺ and Mg²⁺) for 1 h on the flask surface prior to seeding to facilitate cell attachment, and the medium was subsequently washed before use. MSCs were subcultured every 3 days following detachment with TrypLE™ Express Enzyme (Gibco). Cells were seeded at a density of 2 × 10⁵ cells per T25 culture flask (Corning, Oneonta, NY, USA). And then incubated in a cell culture incubator under humid conditions at 37 °C in 5% CO₂.

Surface marker analysis via flow cytometry

The expression of surface markers on MSCs cultured in FCM, SFM, and AOF media at P5 and P10 was assessed using flow cytometry (Beckman Coulter, Brea, CA, USA). Briefly, 1 × 10⁵ cells were resuspended in 100 µL of staining buffer consisting of DPBS (Welene, Gyeongsan, Korea) supplemented with 2% FBS. The cells were then incubated with the respective phycoerythrin (PE)-conjugated monoclonal antibodies for 30–60 min at 4 °C in the dark. Following incubation, the cells were washed to remove unbound antibodies.

Cellular debris and nonviable cell populations were excluded from the analysis based on standard forward scatter (FSC) and side scatter (SSC) gating strategies. No additional viability dye was used in this assay. Fluorescence intensity was measured for 10,000 events per sample. Positive expression was defined as fluorescence intensity exceeding 99% of the corresponding isotype control or unstained control. Data were analyzed using FlowJo software (TreeStar Inc., Ashland, OR, USA). Detailed information regarding the antibody suppliers, clones, catalogue numbers, and working concentrations is summarized in Supplementary Table S1. The percentage expression and biological relevance of each surface marker are summarized in Supplementary Table S2.

Differentiation assays

For chondrogenic differentiation, MSCs were seeded at a density of 2.1 × 10⁴ cells per well in 0.02% gelatin (Sigma-Aldrich) coated 12-well plates in cultue medium. After 24 h, the medium was replaced with chondrogenic differentiation medium, and cells were cultured for 14 days. This medium consisted of DMEM containing 10% FBS, 1% P/S, 1% Insulin-Transferrin-Selenium-X supplement, 50 µM ascorbic acid, 1 µM dexamethasone, and 10 ng/mL transforming growth factor-β1 (TGF-β1; PeproTech, Thermo Fisher Scientific). After fixation, cells were stained with 0.5% Alcian Blue (Sigma-Aldrich).

For adipogenic differentiation, MSCs were seeded at a density of 2.1 × 10⁵ cells per well in 12-well plates (Corning) in AOF medium. After 2 days of culture, the medium was replaced with an adipogenic induction medium (StemPro Adipogenesis Differentiation Kit, Gibco), and cells were cultured for an additional 14 days. After fixation in 4% formaldehyde for 30 min, cells were stained in 0.6% Oil Red O solution (Sigma-Aldrich) at room temperature for 15 min, after which they were washed with 60% isopropanol and distilled water.

For osteogenic differentiation, cells (2.1 × 10⁴ cells per well in 12-well plates) were cultured in osteogenic differentiation medium for 21 days, using the StemPro Osteogenesis Differentiation Kit (Gibco). Following fixation, the cells were stained with 2% Alizarin Red S solution (Sigma-Aldrich) and washed with distilled water.

For the quantitative analysis of trilineage differentiation, all images were processed using ImageJ software. Alcian Blue staining images were quantified using Integrated Density (IntDen), which is calculated as the product of the stained area and the mean gray value of the corresponding color channels. Oil Red O and Alizarin Red S staining images were quantified to determine the percentage of the stained area relative to the total area (%Area).

Indoleamine 2,3-dioxygenase (IDO) ELISA

For the quantification of IDO levels, cells were initially seeded at 5.2 × 10⁴ cells/well in 6-well plates and allowed to adhere for 48 h. Immune priming was then induced by exposing the cells to 50 ng/mL IFN-γ (PeproTech) for an additional 24 h. Post-treatment, the cells were enzymatically harvested with TrypLE™ Express, centrifuged, and washed with cold 1X DPBS. Cell lysates were prepared to measure specific IDO levels using an IDO ELISA kit (Abcam, Cambridge, UK) as per the provided instructions. Optical density readings at 450 nm were acquired using a microplate absorbance reader (GloMax^® Discover System; Promega, Madison, WI, USA). Data analysis was performed with GraphPad Prism software (v. 10.6.0) by applying a four-parameter logistic (4-PL) regression model. Finally, the obtained IDO values were normalized to the total cellular protein concentration, which was evaluated using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific).

Cell growth kinetics analysis

MSCs at P2 were cultured until P15 in different media conditions. Cells were passaged and counted using a cell counter (NanoEnTek, Seoul, Korea). The calculations of the cell population doubling time (PDT) and accumulated cell number (ACN) were performed using the following equation:

where t is the duration of cell culture (h), Nh is the number of cells harvested after culture (t), Ns is the number of cells initially seeded, and ACN_prev is the accumulated cell number of caculated previous passage, ACN_prev of initial passage is equal to the the number of cells initially seeded.

Image analysis

Cell images were obtained 2 days after passage through flasks, and recorded using a Nikon Ti inverted microscope (Nikon Eclipse Ti, Tokyo, Japan) equipped with 4-20x magnification. The area and perimeter measurements were obtained using the Nikon Nis Element software (Nikon). The calculation of circularity was conducted using the following equation:

Cell cycle analysis via flow cytometry

For the purpose of the different experiments, MSCs were collected at P5, P10, and P15, and fixed in 80% ethanol (Sigma-Aldrich) at 4℃ for a least 1 h. After washing in DPBS, cells were stained in DPBS containing 50 ug/mL propidium iodide (Invitrogen, Carlsbad, CA, USA) and 100 ug/mL RNase A (Invitrogen) at 37℃ for 30 min. The samples were then diluted in DPBS, and analysed using a CytoFLEX flow cytometer (Beckman Coulter, Miami, FL, USA). Data were analysed using FlowJo software (v10.10.0) and the Dean-Jett-Fox data-fitting algorithm.

RNA sequencing

For these experiments, the cells were dissociated using a TrypLE solution. Total RNA was extracted from MSCs using Trizol™ reagent (Invitrogen), following the manufacturer’s instructions. RNA sequencing was performed using an Illumina platform.

Library preparation and RNA sequencing

Total RNA was extracted from MSCs cultured in three different media: FCM, SFM, and AOF, at P5, P10, and P15. The MSCs were obtained from three donors (A107, A108, and A109). RNA sequencing libraries were generated using NEBNext^® UltraTM RNA Library Prep Kit for Illumina^® (NEB, USA), according to the manufacturer’s recommendations. Paired-end reads were obtained by sequencing the pooled libraries on an Illumina platform. These reads were subsequently aligned to the human reference genome (NCBI GRCh38.p14) using spliced transcript alignment to a reference (STAR) version 2.7.10a [17]. The resulting binary alignment and map (BAM) files were processed using featureCounts in subread package version 2.0.3, a useful program employed for counting reads generated by RNA-seq [18].

Principal component analysis and clustering analysis

Principal component analysis (PCA) generates a plot of the distances among three media and three passages. PCA transformed the expression profiles of differentially expressed genes (DEGs) using RNA-Seq into a lower-dimensional space. In the PCA plot, the first principal component (PC1) represents the direction of the greatest variability while the second principal component (PC2) captures the next highest orthogonal variance. Based on these PCs, samples containing the three media and three passages are displayed in Fig. 3A. To analyse the relationships between the three media (FCM, SFM, and AOF), hierarchical clustering was conducted using Manhattan distances and complete linkage. A heatmap and dendrogram were subsequently generated to visualise the results of hierarchical clustering, and the expression values in the heatmap were scaled using the Z-score formula (Fig. 3B).

Identification of differentially expressed genes (DEGs)

Gene expression levels were quantified for each condition using counts per million (CPM) in EdgeR [19]. Differentially expressed genes (DEGs) were identified based on a fold-change cutoff > 2 within each cell line (A107, A108, and A109). We subsequently identified the common DEGs shared across all cell lines. Treating the three cell lines as biological replicates, we considered genes with false discovery rate (FDR) less than 0.05 to be statistically significant. To identify the DEGs under each condition, this analysis was repeated across comparisons among the three media (FCM, SFM, and AOF) and three passages (P5, P10, and P15). A complete list of DEGs is provided in Supplementary Table S3.

Biological pathway and network analysis

Over-representation analysis (ORA) was conducted to characterise the biological pathways related to DEGs using Metascape (https://metascape.org) and the Gene Ontology (GO) knowledgebase (https://geneontology.org/). Only biological pathways with a p-value of less than 0.01 which included more than five genes were selected. The biological pathway-pathway interaction network was visualised using Cytoscape (v3.11.0). In network analysis, the nodes represent each pathway that includes gene sets, while the edges describe the interactions between pathways using kappa similarity scores [20]. To highlight the significant biological pathways, only nodes without interactions and the top nodes with the highest p-values were selected.

Biological pathway analysis using heatmaps and boxplots

Heatmaps and boxplots were used to illustrate the expression levels of DEGs within specific biological pathways, and these visualisations were generated based on the DEGs. Heatmap images were generated using the R package (https://rdocumentation.org/packages/pheatmap/versions/1.0.12). Box plots were generated to visualise the expression patterns of gene sets within specific pathways, which were particularly related to the culture of stem cells. A boxplot from the ggplot R package was applied to compare gene expression levels between the two groups (https://ggplot2.tidyverse.org/index.html), and p-values were calculated using a paired t-test.

Comparison of stem cell surface and inflammatory markers

To evaluate whether MSCs cultured in FCM, SFM, and AOF media retain their MSC identity, we confirmed the expression levels of known MSC surface markers (eight positive and six negative markers), as defined by the International Society for Cellular Therapy (ISCT) [21]. In addition, the inflammatory response of MSCs was assessed using four anti-inflammatory markers, four multi-inflammatory markers, and 23 pro-inflammatory markers. We further examined the expression of seven housekeeping genes specific to adipose-derived MSCs to validate the stability of FCM, SFM, and AOF media. The expression levels of the markers in the MSCs were used as CPM in the three media after three passages. The statistical significance for all comparative analyses was determined using a paired t-test.

Quantitative real-time polymerase chain reaction (qRT-PCR)

cDNA synthesis was carried out using 1 µg of total RNA with the GoScript Reverse Transcription Mix kit (Promega, Madison, WI, USA). qRT-PCR was conducted using GoTaq qPCR Master Mix (Promega) and the StepOne™ Real-Time PCR System (Applied Biosystems, Foster City, CA, USA), in accordance with the manufacturer’s protocol. All data were normalised to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression and analysed using CT [22]. The primer sequences used in these assays are listed in Supplementary Table S4.

CNV analysis

To analyse CNVs, which serve as critical biomarkers for assessing the genomic integrity of ESCs, we performed computational analysis using the CaSpER package [23]. RNA-seq transcriptomic data were processed using the caSPER pipeline, which integrates allele-specific signals from bulk RNA-seq or single-cell RNA-seq data to identify large-scale chromosomal gains and losses. Processed BAM files produced using RNA-seq were used to generate the B-allele frequency (BAF) and log₂ ratio profiles through CaSpER with a 50 kb sliding window and a minimum of three single-nucleotide polymorphisms per window.

Cell cycle phase classification using gene expression

RNA-seq data were used to classify cells into different cell cycle phases using the cyclone algorithm in the Scran package [24]. The cyclone function was applied to the normalised expression matrix, generating probability scores for the G1, S, and G2/M phases in three MSCs of FCM, SFM, and AOF media. To ensure proportions across cell cycle phases, a distribution of proportions was constructed by shuffling the gene expression values within each MSC. The phase score quantifies the strength of evidence that a sample’s transcriptomic signature matches each cell cycle phase [24]. This outcome reflects the phase-specific expression patterns of cell cycle marker genes, which form the basis of the computational classification approach.

Statistical analyses

For the cell cycle assay and qRT-PCR, experimental data were analysed using GraphPad Prism v10.4.1 (GraphPad, San Diego, CA, USA) and Excel (Microsoft, Redmond, WA, USA) for statistical analysis and graph generation. The FlowJo software (v10.10.0) was used to fit the cell cycle proportions. All data are presented as the mean ± standard error of mean (SEM), with ‘n’ representing the number of data points.

Results

Characterization of isolated AD-MSCs

AD-MSCs were established from three independent donors and initially characterized by flow cytometry to confirm their MSC immunophenotype. Representative histograms and quantitative summaries of surface marker expression in the established AD-MSCs are provided in Supplementary Figure S1. The established AD-MSCs showed high expression of MSC-positive markers, including CD73, CD90, and CD105, and low expression of lineage-negative or immune-related markers, including CD14, CD34, CD45, and HLA-DR, supporting their MSC identity.

To further evaluate whether MSC identity was maintained under different culture media conditions, AD-MSCs were expanded in FCM, SFM, and AOF media at P5 and P10 and analyzed by flow cytometry. Representative flow cytometry histograms from Donor #2 are shown in Fig. 1A, and corresponding histograms from Donors #1 and #3 are provided in Supplementary Figure S2. Surface marker expression across Donors #1–3 is summarized as a heatmap in Fig. 1B, with donor-specific expression values provided in Supplementary Table 2. Across donors, media, and passages, AD-MSCs maintained high expression of CD73, CD90, and CD105 and low expression of CD14, CD34, and CD45. HLA-DR expression was generally low, although donor- and medium-dependent variability was observed. Overall, the immunophenotypic profile supported the maintenance of MSC identity under the tested culture conditions.

Fig. 1.

Immunophenotypic characterization and trilineage differentiation potential of AD-MSCs cultured in FCM, SFM, and AOF media. A Representative flow cytometry histograms of AD-MSCs from Donor #2 cultured in FCM, SFM, and AOF media at P5 and P10. Histograms show IgG isotype control, lineage-negative or immune-related markers CD14, CD34, CD45, and HLA-DR, and MSC-positive markers CD73, CD90, and CD105. Corresponding histograms from Donors #1 and #3 are provided in Supplementary Figure S2. B Heatmap showing surface marker expression in AD-MSCs from Donors #1–3 cultured in FCM, SFM, and AOF media at P5 and P10. Values indicate mean ± SEM percentages of marker-positive cells from three independent donors. C Representative trilineage differentiation images of Donor #2 AD-MSCs cultured in FCM, SFM, and AOF media at P5 and P10. Day 1 images indicate pre-induction baseline images, and post-induction images indicate day 14 for adipogenic and chondrogenic differentiation and day 21 for osteogenic differentiation. Corresponding differentiation images from Donors #1 and #3 are provided in Supplementary Figure S3. Scale bars: adipogenic images, 100 μm; chondrogenic and osteogenic images, 500 μm. D Quantitative analysis of Oil Red O-positive area, Alcian Blue Integrated Density (IntDen), and Alizarin Red S-positive area. Bars represent mean ± SEM from three independent donors, and symbols indicate individual donors. Statistical significance was determined by one-way ANOVA followed by Tukey’s multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001. FCM, fetal bovine serum-containing medium; SFM, serum-free medium; AOF, animal-origin-free medium

To determine whether culture media affected subsequent differentiation capacity, trilineage differentiation assays were performed using MSCs expanded in FCM, SFM, and AOF media at P5 and P10. Representative Day 1 and post-induction differentiation images from Donor #2 are shown in Fig. 1C, and corresponding images from Donors #1 and #3 are provided in Supplementary Figure S3. Quantitative image analyses from three independent donors are presented in Fig. 1D. Adipogenic, chondrogenic, and osteogenic differentiation were assessed by Oil Red O, Alcian Blue, and Alizarin Red S staining, respectively.

Quantitative analysis showed lineage- and passage-dependent differences among media. At P5, adipogenic differentiation was significantly decreased in SFM-expanded MSCs compared with FCM-expanded MSCs. Chondrogenic differentiation was significantly increased in SFM- and AOF-expanded MSCs compared with FCM-expanded MSCs at both P5 and P10, and AOF-expanded MSCs showed higher chondrogenic differentiation than SFM-expanded MSCs at P10. For osteogenic differentiation, AOF-expanded MSCs showed significantly higher staining intensity than SFM-expanded MSCs at P5 and significantly higher staining intensity than both FCM- and SFM-expanded MSCs at P10. Consistent with this trend, Donor #1 showed minimal osteogenic staining in the FCM and SFM conditions at P10, whereas staining was retained in the AOF condition in the representative images (Supplementary Figure S3B). These results indicate that AOF-expanded MSCs retained trilineage differentiation potential under the tested conditions.

Culture media conditions differentially influence MSC proliferation and morphological features

MSCs were expanded in FCM, SFM, and AOF media according to the optimized culture protocol recommended for each medium condition. Among the three culture systems, only SFM required flask coating with CTS CELLstart to support cell attachment, whereas FCM and AOF were maintained without additional coating (Fig. 2A). Thus, the following analyses compared each medium as a complete culture system under its recommended use condition.

Fig. 2.

Comparison of the growth kinetics and morphology of MSCs cultured in FCM, SFM, and AOF Media. A Representative phase-contrast images of human adipose-derived MSCs cultured in FCM, SFM, and AOF media at P5, P10, and P15. Scale bars, 300 μm. B Quantitative analysis of cell circularity at P5, P10, and P15. Data are presented as mean ± SEM (n = 120 cells). Statistical significance was determined by one-way ANOVA followed by Tukey’s multiple comparisons test. ^*p < 0.05, ^***p < 0.001. C, D Growth kinetics of AD-MSCs expanded in FCM, SFM, and AOF media up to P15, shown as population doubling time (PDT) (C) and accumulated cell number (ACN) (D). Data are presented as mean ± SEM (n = 6). Statistical significance was determined by one-way ANOVA followed by Tukey’s multiple comparisons test. For pairwise comparisons, * indicates AOF vs. FCM, # indicates AOF vs. SFM, and § indicates FCM vs. SFM. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001; ^§p < 0.05, ^§§p < 0.01, ^§§§p < 0.001. E Quantitative analysis of cell area at P5, P10, and P15. Each dot represents a single cell (n = 120 cells). Statistical significance was determined by one-way ANOVA followed by Tukey’s multiple comparisons test. ^**p < 0.01, ^***p < 0.001. FCM, fetal bovine serum-containing medium; SFM, serum-free medium; AOF, animal-origin-free medium

Morphometric analysis across early (P5), mid (P10), and late (P15) passages showed media-dependent differences in circularity and cell area. Circularity was lowest in FCM at P5 and highest in AOF at P15 (Fig. 2B). Growth kinetics were quantified up to P15. The population doubling time (PDT) was consistently shortest in AOF-cultured MSCs, approximately 21–31 h across passages, indicating faster expansion than in FCM- or SFM-cultured MSCs (Fig. 2C). The accumulated cell number (ACN) began to diverge from P5 and was highest in AOF by P15 (Fig. 2D). Cell area was smaller in AOF-cultured cells across passages compared with the other culture conditions (Fig. 2E).

Distinct gene expression patterns among MSCs cultured in different media conditions

To comprehensively evaluate the impact of culturing in FCM, SFM, and AOF media on MSCs, we conducted RNA-seq on MSCs cultured in the three media at three different passages (P5, P10, and P15). PCA of the RNA-seq data demonstrated a clear segregation of samples, primarily based on media conditions (Fig. 3A). The first principal component (PC1), which accounted for the greatest variability in gene expression, effectively separated MSCs cultured under different media conditions, thereby highlighting the substantial impact of the culture environment on the transcriptomic landscape. The second principal component (PC2) revealed further discrimination regarding AOF and two other types of media. Hierarchical clustering analysis of DEGs further confirmed the transcriptomic divergence among the three media conditions (Fig. 3B). The resulting heat map demonstrated distinct expression patterns for each condition. In Fig. 3B, the dendrogram structure revealed that the FCM and SFM clustered together on the same branch, whereas the AOF samples clustered separately on a different branch. The observed clustering indicated that the transcriptional differences induced by AOF medium were more substantial than those induced by FCM and SFM. This finding was consistent across all three passages.

Fig. 3.

Comprehensive analysis of transcriptomic changes across different culture conditions and passages. A Principal component analysis (PCA) of RNA-seq data from MSCs cultured in FCM, SFM, and AOF media at P5, P10, and P15. PCA plots are shown for each passage and for all samples combined. Each dot represents an individual sample, and colors indicate different media conditions. B Hierarchical clustering heatmap of differentially expressed genes (DEGs) across all experimental conditions. Each row represents an individual gene, and each column represents a sample grouped by media condition and passage. The dendrogram was constructed using Manhattan distance and complete linkage. Expression values were normalized and scaled as Z-scores. The color gradient from blue to red represents low to high expression levels. C MA plots showing DEGs at P5 in SFM versus FCM, AOF versus FCM, and AOF versus SFM comparisons. Red, blue, and gray dots indicate significantly upregulated genes, significantly downregulated genes, and genes with no significant differential expression, respectively. Horizontal dashed lines indicate the fold-change threshold. D Bar plot showing the number of upregulated and downregulated DEGs at P5 for each comparison. DEGs were defined by an absolute fold change ≥ 2 and FDR < 0.05. Upregulated and downregulated genes are shown in red and blue, respectively. FCM, fetal bovine serum-containing medium; SFM, serum-free medium; AOF, animal-origin-free medium; DEG, differentially expressed gene; FDR, false discovery rate

Analysis of DEGs revealed substantial transcriptomic differences among the three media conditions. In P5, a comparison between SFM and FCM media identified 713 DEGs (fold change ≥ 2 or ≤ -2, FDR < 0.05), with 179 genes significantly upregulated and 534 genes downregulated in SFM relative to FCM media (Fig. 33C, D). A comparison between AOF and FCM media identified 3,414 DEGs, of which 1,275 were upregulated and 2,139 were downregulated. Similarly, a comparison between AOF and SFM media revealed 3,234 DEGs, with 1,254 upregulated and 1,980 downregulated genes. In both comparisons, downregulated genes predominated, and the total number of DEGs in AOF was higher than that in SFM and FCM. A substantial number of DEGs observed between AOF and the other two media types aligned with the PCA and hierarchical clustering results, further confirming the distinctive transcriptomic signature of MSCs cultured under AOF conditions. The same outcome was observed at P10 and P15, although the number of DEGs decreased slightly across all the comparison groups at P15 (Supplementary Figure S3). Transcriptomic differences among the three media conditions were maintained throughout the extended cultivation period. These results indicated that the distinct gene expression signatures established between the three media conditions represent stable transcriptional adaptations rather than transient responses.

Media-dependent regulation of the cell cycle, proliferation, and differentiation process

To elucidate the biological pathways influenced by the three media conditions, we conducted ORA analysis and constructed a biological pathway-pathway network. Network analysis of the DEGs revealed interconnected clusters of biological pathways that were significantly altered across the different media conditions (Fig. 4A). The network highlighted the functionally related pathway clusters involved in MSC biology, including cell cycle regulation, cellular proliferation, differentiation, growth, and DNA-related processes. A complete list of the relevant biological pathways is provided in Supplementary Table S5. Notably, the mitotic cell cycle, negative control of proliferation, and mesenchymal cell differentiation, as the principal biological pathways, showed statistically significant differences (p < 0.01) in key MSC functional processes across the three media conditions (Fig. 4B).

Fig. 4.

Network analysis of DEGs and enriched biological pathways in MSCs cultured in different media conditions. A The biological pathway-pathway network based on DEGs at P5. The network was constructed where nodes represented individual biological pathways and edges indicated similarity among pathways based on kappa similarity scores. The node size indicates the statistical significance of each pathway and clusters of functionally related pathways were highlighted using colors and labeled according to their biological functions. B Bar plot showing the statistical significance for principal biological pathways. The X-axis indicates the significance level of each pathway

Figure 5A presents a heatmap analysis focusing on genes associated with the mitotic cell cycle, showing clear differences in expression, with higher levels observed specifically in AOF media than in FCM and SFM media. Among the 215 genes associated with the mitotic cell cycle, 185 genes showed the highest expression in AOF, 28 genes were highly expressed in FCM, and only 2 genes showed the highest expression in SFM. Further, boxplot analysis confirmed that the expression of mitotic cell cycle-related genes was significantly upregulated in MSCs cultured in AOF compared to both FCM and SFM (Fig. 5B, p < 0.001). Culturing MSCs in AOF media led to the upregulation of the mitotic cell cycle, as well as the downregulation of negative proliferation and differentiation compared with FCM and SFM conditions. Cell cycle phase classification based on transcriptomic signatures revealed distinct cell cycle distributions among the three different media conditions (Fig. 5C). This analysis demonstrates that AOF-cultured MSCs have a high probability of being in the G2/M phase. In contrast, MSCs cultured in SFM were predominantly in the S phase. FCM-cultured MSCs displayed an intermediate pattern between AOF and SFM conditions. Cell cycle–related transcriptomic signatures, quantified using S-phase and G2/M-phase scores derived from bulk RNA-seq, differed among the three media conditions (Fig. 5C). AOF-cultured MSCs localized predominantly to G2/M-dominant coordinates, whereas the SFM group was restricted to high S scores regions with negligible G2/M scores. FCM-cultured MSCs displayed an intermediate pattern between AOF and SFM. Consequently, we concluded that AOF media promoted the most active cell cycle progression and mitotic activity, while SFM conditions appear to maintain cells predominantly in the S phase, with reduced progression to G2/M compared to AOF. This differential expression pattern was consistent across all three passages. The findings of this study support the conclusion that AOF media promote the activation of cell cycle progression, whereas SFM conditions appear to suppress it relative to other media. The cell cycle pathway in the KEGG database showed specific molecular mechanisms underlying enhanced cell cycle progression in AOF-cultured MSCs (Fig. 5D). The molecular regulators of the cell cycle showed significant differences which emerged in S-G2 transition and G2/M regulation. The expression levels of G1 phase and G1-S transition regulators, including the Cyclin D-CDK4 complex and the Cyclin E-CDK2 complex, were comparable between FCM and AOF. Similarly, the expression levels of the primary S-phase regulator Cyclin A-CDK2 complex were comparable between FCM and AOF. These results suggested that MSCs cultured in either FCM or AOF maintain similar regulatory mechanisms for G1 phase progression, initial G1-S checkpoint regulation, and early DNA replication processes. However, notable differences were observed in the expression of S-G2 transition and mitotic phase markers. For example, the Cyclin A-CDK1 complex, which verifies the completion of DNA replication and initiates mitotic processes, was significantly upregulated in the AOF-cultured MSCs. This overexpression also extended to the mitotic Cyclin B-CDK1 complex, a key regulator of M phase entry and progression. Additionally, components of the anaphase-promoting complex/cyclosome (APC/C), including ANAPC5 and ANAPC10, were upregulated in AOF. Furthermore, several components of the mitotic spindle assembly checkpoint (SAC), including BUB1 Mitotic Checkpoint Serine/Threonine Kinase (BUB1), BUB1 Mitotic Checkpoint Serine/Threonine Kinase B (BUB1B), Mitotic Arrest Deficient 2 Like 1 (MAD2L1), and TTK Protein Kinase (TTK, also known as MSP1), exhibited significantly higher expression levels in AOF than in FCM. Collectively, these expression patterns indicate that AOF specifically enhances the expression of regulators governing the S-G2 transition, G2/M mitotic entry, and mitotic progression. This, in turn, results in accelerated cell cycle progression, as observed in cell cycle phase analysis. The upregulation of these cell cycle regulators provides a molecular basis for the enhanced proliferative capacity of AOF-cultured MSCs.

Fig. 5.

Cell cycle dynamics and pathway analysis of MSCs cultured in FCM, SFM, and AOF media. A The Heatmap was generated using the expression values of genes involved in the mitotic cell cycle across the three media conditions. The red, blue, and green bars on the left side of the heatmap represent the genes most strongly expressed in FCM, SFM, and AOF media, respectively. The color scale represents the normalised expression values scaled using the Z-score, with red and blue indicating high and low expression, respectively. B Boxplots were generated using expression values of genes involved in the mitotic cell cycle calculated per donor (n = 3 independent biological donors per condition). p values were calculated using a paired t-test. C The RNA-seq data were classified into G1, S, and G2/M phases based on the expression patterns of different cell cycle marker genes. The X-axis represents the probability of cells being in the S phase, and the Y-axis represents the probability of cells being in G2/M phase. G1 phase probabilities were not displayed as they were close to zero. Each point represents a sample, with different colors indicating the three media conditions. D KEGG pathway analysis of cell-related genes differentially expressed between FCM and AOF cultures. Genes overexpressed in AOF compared to FCM are depicted as red nodes. E–F Heatmap and boxplots of gene expression values involved in the negative regulation of cell population proliferation (n = 3 independent biological donors per condition). (G-H) Heatmap and boxplots of gene expression values involved in the mesenchymal cell differentiation (n = 3 independent biological donors per condition)

In accordance with the enhanced cell cycle activity observed in AOF media, a significant downregulation of genes involved in the negative regulation of cell population proliferation was detected in AOF compared to FCM and SFM (Fig. 5E, F). A heatmap visualising the expression of genes involved in the negative regulation of cell population proliferation generally indicated lower expression levels in AOF. Among the 240 genes related to the negative regulation of proliferation, 136 were most highly expressed in FCM, 92 in SFM, and only 12 in AOF. This was statistically confirmed by boxplot analysis, which demonstrated significantly lower expression of these proliferation inhibitors in AOF than in both FCM and SFM (p < 0. 01). Genes involved in suppressing cell proliferation, such as Cyclin-Dependent Kinase Inhibitor 1 A (CDKN1A), Tumor Protein P53 Inducible Protein 11 (TP53I11), BTG Anti-Proliferation Factor 1 (BTG1), and BTG Anti-Proliferation Factor 1 (BTG2), showed the most pronounced downregulation in the AOF medium. In addition, the mesenchymal cell differentiation pathway was significantly downregulated in AOF compared to in FCM and SFM (Fig. 5G, H). MSCs cultured in FCM and SFM media exhibited significantly higher expression of differentiation-related genes than AOF-cultured cells (p < 0.01). Among the 72 genes related to mesenchymal cell differentiation, 42 were highly expressed in FCM and 23 in SFM, whereas only seven showed the highest expression in AOF. This differential expression profile indicated that AOF effectively maintained a more undifferentiated state than FCM and SFM.

To validate these findings, qRT-PCR was conducted on MSCs from three donors at P5, P10, and P15. The expression levels of genes encoding the cyclin-CDK complex (CDK, CCNA2, CCNB1, and CCNB2) were significantly increased in AOF culture conditions at P5 compared to other culture conditions. However, a significant decrease in gene expression was observed in a study comparing SFM with other cultures (Fig. 6A). Specifically, APC5 and APC10, which belong to the APC/C gene family, demonstrated a significant increase under AOF culture conditions at P5, P10, and P15. Further, the SAC gene family members TTK, MAD2DL1, and BUB1 exhibited a significant increase under AOF culture conditions in P5 compared to other culture conditions. BUB1B expression, an additional gene belonging to SAC, exhibited a significant increase in AOF culture conditions at P5 and P10 compared with the alternative conditions (Fig. 6C). In addition, we examined the expression of BTG1, CDKN1A, and TP53I11, which are involved in the negative regulation of cell proliferation. BTG1 expression was significantly increased under AOF culture conditions at P10 and P15. CDKN1A expression significantly increased under FCM conditions at P5 and under AOF conditions at P15. TP53I11 expression exhibited differential trends between passages, with significant increases in SFM at P10 and in AOF at P15 (Fig. 6D).

Fig. 6.

qRT-PCR validation of cell cycle related markers in MSCs cultured in different media conditions. Gene expression of A Cyclin-CDK complex-related genes (CDK1, CCNA2,CCNB1, and CCNB2), B APC/C-related genes (APC5 and APC10), C spindle assembly checkpoint-related genes (TTK, MAD2L1, BUB1, and BUB1B), and D negative regulation of cell population proliferation-related genes (BTG1, CDKN1A, and TP53I11). Data are presented as the mean ± SEM. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001 by Dunnett’s test

Cell cycle distribution of MSCs is modulated by culture media composition

To determine whether the medium conditions of MSCs also affect the cell cycle, the distribution of cells in the G1, S, and G2/M phases was examined by FACS (Fig. 7A). For cell cycle analysis based on the culture medium conditions in the same passage, the fold-change values relative to FCM were analysed. Regardless of medium conditions, most cells were in the G1 phase during the cell cycle. Interestingly, the G2/M phase in AOF showed the highest proportion in P10 and P15 compared to SFM and FCM. More details, the proportions of cells in the cell cycle at P15 were determined to be 8.20 ± 0.65%, 3.59 ± 0.29%, and 10.14 ± 1.87% of the G2/M phase, respectively, in FCM, SFM, and AOF. The proportions of cells in the cell cycle at P10 were also determined to be 7.97 ± 0.42%, 5.57 ± 0.72%, and 10.18 ± 0.63% of the G2/M phase, respectively, in the FCM, SFM, and AOF (Fig. 7A). In addition, the S/G2/M phase showed the lowest percentage of cells under SFM culture conditions at P5, P10, and P15 (Fig. 7B).

Fig. 7.

Effects of culture media on MSC cell cycle dynamics over passages. A The percentages of cell cycle phase in FCM, SFM, and AOF condition at P5, P10, and P15 are presented. B A comparison of the proportion of cells in different cell cycle stages (G1, S, and G2/M) in MSCs cultures in FCM, SFM, and AOF at the same passage. The bar graphs were normalised to FCM and represented as mean ± SEM. A statistical analysis was conducted, employing Tukey’s multiple comparisons test. P-values less than 0.05 were designated as significant (ns, not significant; ^*p < 0.05, ^**p < 0.01, ^***p < 0.001). The results are derived from a total of three donors

Profiles of cell surface markers and inflammatory markers as MSC stability

To evaluate whether MSCs cultured in different media conditions retain their MSC identity, we analysed the expression levels of established MSC surface markers defined by the ISCT, including eight positive markers (NT5E, THY1, ENG, VCAM1, MCAM, ALCAM, CD44, and ITGB1) and six negative markers (CD34, PTPRC, ITGAM, CD14, CD19, and CD79A) (Supplementary Figure S5). A broad comparison considering the collective expression of all positive markers (Fig. 8A) and all negative markers (Fig. 8B) indicated no statistically significant differences among FCM, SFM, and AOF media (p > 0.05). These results indicate that, despite the substantial differences identified in the comprehensive transcriptomic analysis, MSCs maintained the expression of core surface markers that defined their identity, regardless of the media conditions.

Fig. 8.

Evaluation of MSC marker expression, inflammatory marker profiles, and immunomodulation-related IDO expression in AD-MSCs cultured in FCM, SFM, and AOF media. A Comparative analysis of expression levels of eight MSC-positive marker genes. B Comparative analysis of expression levels of six MSC-negative marker genes. C Heatmap showing the expression profiles of inflammation-related markers. Red and blue indicate relatively higher and lower expression levels, respectively. D–F Comparative analysis of anti-inflammatory (D), pro-inflammatory (E), and multi-inflammatory (F) marker expression levels. G, H Comparative analysis of NGF (G) and CSF1 (H) expression levels. I RT-qPCR validation of NGF and CSF1 expression in AD-MSCs cultured in FCM, SFM, and AOF media at P5, P10, and P15. J IDO protein levels measured by ELISA in AD-MSCs cultured in FCM, SFM, and AOF media at P5, P10, and P15 with or without IFN-γ stimulation. IFN-γ-untreated cells were used as negative controls. Data are presented as mean ± SEM from three independent biological donors per condition. For panels A–I, statistical significance was determined by one-way ANOVA followed by Dunnett’s test. For pane l J, statistical significance was determined by one-way ANOVA followed by Tukey’s multiple comparisons test. ns, not significant; ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. FCM, fetal bovine serum-containing medium; SFM, serum-free medium; AOF, animal-origin-free medium; IDO, indoleamine 2,3-dioxygenase; IFN-γ, interferon-γ

We further examined the expression of anti-, pro-, and multi-inflammatory markers to evaluate the stability of the MSC inflammatory profile in the three media conditions. Heatmap visualisation revealed that most anti-, pro-, and multi-inflammatory markers exhibited no distinct differences between the three media conditions (Fig. 8C). A comparative analysis of the expression levels of anti-inflammatory (Fig. 8D), pro-inflammatory (Fig. 8E), and multi-inflammatory markers (Fig. 8F) also showed low expression levels, with no statistically significant differences (p > 0.05). However, analysis of individual markers identified statistically significant differences in several genes, notably Nerve Growth Factor (NGF) and colony-stimulating factor 1 (CSF1) (Fig. 8G, H). Intriguingly, the expression levels of both NGF and CSF1 were significantly lower in MSCs cultured in AOF than in those cultured in FCM or SFM. The overall inflammatory profile remained stable across the media. To validate the expression of NGF and CSF1 under other culture conditions, we conducted qRT-PCR on MSCs cultured at P5, P10, and P15. In MSCs cultured in AOF, there was a significant decrease in the expression levels of NGF and CSF1 at all passages compared to other culture conditions (Fig. 8I).

To further support the immunomodulation-related findings at the protein level, IDO expression was measured by ELISA in MSCs cultured in FCM, SFM, and AOF media at P5, P10, and P15 with or without IFN-γ stimulation. IFN-γ stimulation markedly increased IDO protein levels across all media conditions, whereas unstimulated controls showed minimal IDO expression. At matched passages, IDO expression did not significantly differ among FCM-, SFM-, and AOF-cultured MSCs, although IDO levels tended to decrease with increasing passage number, particularly at P15 (Fig. 8J). These results suggest that AOF expansion did not markedly alter IDO expression as an immunomodulation-related marker under the tested conditions.

Evaluation of genomic stability via copy number variation analysis

To evaluate the genomic stability of MSCs cultured in FCM, SFM, and AOF, we conducted a comprehensive CNV analysis across all three passages. CNVs were identified using the BAF (Fig. 9A) and gene expression data derived from RNA-seq (Fig. 9B), enabling the detection of chromosomal gains and losses across the genome. Chromosome-wide visualisation of CNV events at P5 revealed distinct patterns of genomic alterations across all 22 autosomes in MSCs cultured in FCM, SFM, and AOF (Fig. 9C). The CNV profiles demonstrated regional variations in chromosomal gains (indicated in red) and losses (indicated in blue) distributed throughout the genome. The CNV data for P10 and P15 are shown in Supplementary Figure S6. MSCs cultured in AOF medium exhibited a slightly higher number of CNVs than those cultured in FCM and SFM at P5, P10, and P15 (Table 1). However, statistical analysis of the total number of CNV events and segment lengths across the three media conditions revealed no significant differences (Fig. 9D, E). These findings indicate that the genomic stability of MSCs is comparable across FCM, SFM, and AOF.

Fig. 9.

Copy number variation (CNV) analysis of MSCs cultured in different media. A Scale representation of B-allele frequency (BAF) values used in the CNV analysis. B Scale representation of gene expression values used in the analysis. C Chromosome-wide visualization of CNV events at P5 across different media conditions. Blue regions indicate chromosomal losses, while red regions represent gains. D Boxplots comparing the number of CNV events across the three media conditions. E Boxplots comparing the CNV segment length across the three media conditions

Table 1.

Number of CNV events

	Passage 5			Passage 10			Passage 15
	FCM	SFM	AOF	FCM	SFM	AOF	FCM	SFM	AOF
Loss	16	10	37	9	20	79	28	22	25
Gain	22	37	67	36	31	65	14	52	43
Total	38	47	104	45	51	144	42	74	68

FCM, fetal bovine serum-containing medium; SFM, serum-free medium; AOF, animal-origin-free medium

Discussion

In this study, we systematically compared the phenotypic, functional, and transcriptomic differences in human AD-MSCs maintained in FCM, SFM, and AOF media. The main finding of this study was the significant upregulation of mitotic cell cycle–related genes in MSCs cultured in AOF medium compared to those cultured in FCM and SFM, while largely maintaining inflammatory and safety-related marker profiles. Our RNA-seq analysis revealed that AOF specifically enhanced the expression of G2/M phase regulators, rather than broadly affecting all phases of the cell cycle, indicating that the AOF medium creates an optimal microenvironment for MSC expansion. The influence of media composition and growth factor combinations on MSC proliferation is well recognised [25]; however, our data newly identified a targeted regulatory mechanism that represents a distinct advantage of the AOF medium for efficient expansion. These findings thus provide critical insights into the optimisation of xeno-free and GMP-compatible culture systems for clinical MSC production.

To better understand the mechanisms underlying these observations, we analysed the expressions of molecular regulators of the cell cycle. This analysis showed that culturing in AOF medium significantly upregulated Cyclin A–CDK1 and Cyclin B–CDK1 complexes, which are crucial for the S–G2 transition and M-phase progression, verifying the completion of DNA replication and initiating mitosis, and regulating M-phase entry and progression, respectively [26]. This selective upregulation was consistent with our cell cycle phase analysis, which showed that AOF-cultured MSCs exhibited a higher probability of being in the G2/M phase than FCM-and SFM. We further observed an increased expression of anaphase-promoting complex/cyclosome (APC/C) components (ANAPC5, ANAPC10), suggesting enhanced control of sister chromatid separation and mitotic exit. APC/C is a critical E3 ubiquitin ligase that regulates mitosis and the G1 phase by targeting key proteins for degradation [26]. Additionally, SAC proteins (BUB1, BUB1B, MAD2L1, and TTK) were upregulated in the AOF-cultured MSCs. SAC proteins ensure proper chromosome alignment and prevent premature chromatid separation, thereby safeguarding genomic stability [27]. Collectively, these molecular signatures indicated that AOF provided a proliferative advantage while maintaining mitotic fidelity.

The enhanced proliferative capacity of AOF MSCs was further supported by the downregulation of genes involved in the negative regulation of proliferation, including CDKN1A (p21), TP53I11, BTG1, and BTG2. CDKN1A is a well-known inhibitor of p53-dependent G1 arrest [28], whereas BTG1 and BTG2 are known anti-proliferative regulators [29]. Their downregulation, combined with the upregulation of cyclin-CDK complexes, APC/C components, and SAC proteins, established an optimal molecular environment for accelerated and stable MSC expansion. In parallel, AOF-cultured MSCs exhibited a reduced expression of mesenchymal differentiation-related genes, consistent with their enhanced proliferative profiles. Despite reduced expression of mesenchymal differentiation-related genes, AOF-expanded MSCs retained functional trilineage differentiation potential under the tested conditions. This was supported by quantitative differentiation analysis and by donor-level representative images, in which Donor #1 at P10 showed retained osteogenic staining in AOF but minimal staining in FCM and SFM. This donor-level observation was interpreted as supportive of the overall trend rather than as an independent statistically significant finding. This inverse relationship between proliferation and differentiation has been well-documented in stem cell biology [30] and suggests that AOF medium favours expansion while maintaining stemness. In contrast, FCM and SFM culture conditions promoted stronger differentiation programs and imposed a tighter negative regulation of proliferation, making them potentially more suitable for lineage-specific applications. Despite these differences, MSC identity was preserved in all of the different media. We confirmed that ISCT-defined positive and negative surface markers remain stable [21] and inflammatory profiles show minimal expression changes, with only NGF and CSF1 being reduced in AOF cultures. These findings are consistent with reports that culture conditions can modulate immunomodulation-related features of MSCs without broadly altering their inflammatory marker profiles [28]. The preservation of marker expression and immunological stability underscores the translational potential of AOF-cultured MSCs for clinical use.

To further support the immunomodulation-related findings at the protein level, we assessed IDO expression by ELISA following IFN-γ stimulation. IFN-γ markedly induced IDO protein expression across all media conditions, whereas IDO expression did not significantly differ among FCM-, SFM-, and AOF-cultured MSCs at matched passages. These results suggest that AOF expansion did not markedly alter IDO expression as an immunomodulation-related marker under the tested conditions. However, because IDO expression alone does not directly demonstrate functional immunomodulatory potency, future immune-cell-based functional assays will be required to directly evaluate MSC immunomodulatory function.

Genomic stability is a critical consideration for the clinical application of expanded MSCs, as genomic aberrations can potentially compromise safety and therapeutic efficacy. Indeed, we found no statistically significant differences in CNVs across the three media. However, given that only three donors were included per condition, the statistical power of the analysis is limited, and the results should be interpreted with caution. This limitation underscores the necessity for confirmation in studies with larger sample sizes and greater statistical power. The slightly increased frequency of CNV events in AOF-cultured MSCs could be related to the upregulation of cyclin family members and CDKs, as rapid cell division can sometimes increase the risk of replication errors and genomic alterations [31]. However, the concurrent upregulation of spindle assembly checkpoint proteins (BUB1, BUB1B, MAD2L1, and TTK) in AOF-cultured MSCs likely contributed to the maintenance of genomic integrity by ensuring proper chromosomal segregation during mitosis. Previous studies have demonstrated that SAC proteins are essential for preventing chromosomal instability during cell division, particularly during rapid proliferation [32]. Thus, AOF medium appears to promote robust proliferation, without introducing additional genomic instability.

Senescence is another barrier to extending MSC cultures. Prolonged in vitro expansion typically results in longer population doubling times, increased cell size, and diminished proliferation [33–36]. In our study, AOF-cultured MSCs maintained the lowest and most consistent PDT values up to P15, exhibited stable morphology, and showed continuous increases in accumulated cell numbers. These observations suggest that AOF culture supported sustained proliferative capacity during long-term expansion under the tested conditions. However, MSC behavior can also be influenced by cellular heterogeneity, as previously reported [37]. Therefore, direct senescence assays will be required in future studies to determine whether AOF culture directly delays cellular senescence.

A limitation of this study is that the attachment conditions were not identical across all media. Each medium was used according to its optimized manufacturer-recommended culture protocol: SFM required CTS CELLstart coating, whereas FCM and AOF were maintained without additional coating. Because ECM coating can influence MSC attachment, morphology, and proliferation, the observed differences in circularity, cell area, and growth kinetics may reflect the combined effects of medium formulation and matrix-dependent attachment conditions. In addition, MSCs cultured in SFM without coating did not attach properly, indicating that SFM was coating-dependent under the tested conditions. Therefore, these results should be interpreted as comparisons of each complete culture system under its recommended use condition, rather than as isolated medium-composition effects.

Another limitation is that chondrogenic differentiation was evaluated using a 2D monolayer system rather than standard 3D pellet, spheroid, or multilayered culture conditions. Although TGF-β1-containing induction medium was used, MSC chondrogenic differentiation remains challenging and may be accompanied by incomplete cartilage-like maturation or hypertrophic features [38]. Moreover, 3D or multilayered culture systems can better support cell–cell interactions and matrix-rich microenvironments required for robust chondrogenesis [39]. Therefore, Alcian Blue staining should be interpreted as supportive evidence of chondrogenic induction under the tested conditions, and future studies using 3D or multilayered chondrogenic models will be needed to more rigorously assess chondrogenic potential.

Taken together, our findings build upon the results of prior work comparing growth and differentiation of MSCs in FBS-based and xeno-free systems. By integrating immunophenotyping, transcriptomics, differentiation assessment, immunomodulation-related marker analysis, and CNV analysis, we provided a more comprehensive framework for the evaluation of media suitability. Importantly, regulatory authorities and professional societies have emphasised the need for chemically defined, xeno-free media to improve reproducibility and reduce variability. By directly benchmarking the FCM, SFM, and AOF systems, this study addresses key translational gaps and highlights culture in AOF media as a promising solution for GMP-compliant MSC manufacturing.

Conclusion

In the present study, we demonstrated that AOF medium selectively accelerated G2/M cell cycle progression and enhanced the proliferative capacity of human adipose-derived MSCs, while preserving surface marker identity, immunomodulation-related marker profiles, differentiation capacity, and genomic integrity. Compared with foetal bovine serum-containing and serum-free systems, AOF provides a defined and reproducible culture environment that supports large-scale expansion while maintaining key MSC characteristics under the tested conditions. Overall, these findings suggest that AOF medium is a promising culture platform for GMP-compatible MSC manufacturing by reducing reliance on xenogeneic supplements and minimizing lot-to-lot variability. By integrating phenotypic, functional, transcriptomic, immunomodulation-related marker, and genomic analyses, this study provides a framework for evaluating AOF-based culture systems for standardized MSC production.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AD-MSCs

Adipose-derived mesenchymal stem cells

AOF

Animal‑origin‑free

BAF

B-allele frequency

CNV

Copy number variation

DEG

Differentially expressed genes

FBS

Foetal bovine serum

FCM

Serum-containing media

GMP

Good manufacturing practice

GO

Gene ontology

ORA

Over-representation analysis

ISCT

International society for cellular therapy

MSCs

Mesenchymal stem cells

P/S

Penicillin/streptomycin

PCA

Principal component analysis

PDT

Population doubling time

SFM

Serum-free medium

TGF-β

Transforming growth factor-β

Author contributions

Su-Jin Lee and Chunggeon Lee contributed equally to this work as co–first authors. Su-Jin Lee performed the majority of the experiments, and Chunggeon Lee contributed to RNA-seq data processing, bioinformatics analysis, and pathway enrichment. They drafted the initial manuscript. Eunji Kim assisted with cell culture, differentiation assays, and flow cytometry analysis. Min Hee Kang and Jae Woo Kang contributed to the provision of cells and AOF culture media, and participated in experiments. Joo Youn Lee provided overall comments, reviewed the manuscript, and assisted in data review to ensure consistency and clarity. Seyoung Ryu performed RNA-seq data processing. Daeui Park and Hyang-Ae Lee served as co–corresponding authors. They supervised the overall study design, guided data interpretation, critically revised the manuscript, and secured funding. All authors read and approved the final version of the manuscript.

Funding

This research was supported by the Technology Development Program of the Ministry of SMEs and Startups (MSS) (RS-2022-TI023709), the Research Program of the Korea Institute of Toxicology (KIT) (KS-2601), a grant from the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (MSIT) (RS-2023-00225239), and the Bio & Medical Technology Development Program of the NRF funded by MSIT (RS-2022-NR067336).

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. The RNA-seq datasets generated and analyzed during the current study are available in the NCBI Sequence Read Archive (SRA) under the accession number PRJNA1356668. The data can be accessed at the following link: https://dataview.ncbi.nlm.nih.gov/object/PRJNA1356668?reviewer=2ghkvo8npvq131f4ssr38n0pn0.

Declarations

Ethics approval and consent to participate

This study did not involve patients. Human adipose-derived mesenchymal stem cells were obtained from healthy donors with written informed consent under the approval of the Institutional Review Board of Xcell Therapeutics Inc. The study was conducted as part of the IRB-approved project titled “Research on the isolation of mesenchymal stem cells with excellent stemness from adipose tissue” (Ministry of Health and Welfare-designated Public Institutional Review Board, Approval No. P01-201903-31-008, approved on March 21, 2019). All procedures were performed in accordance with institutional and national ethical guidelines. The cells obtained under this approval were subsequently cultured in an animal-origin-free (AOF) medium to assess safety and G2/M cell cycle progression in the present study.

Consent for publication

Not applicable. This manuscript does not contain any person’s data in any form (including individual details, images, or videos).

Competing interests

Min Hee Kang, Jae Woo Kang, and Joo Youn Lee declare financial competing interests, as they are employees of Xcell Therapeutics, the manufacturer and supplier of the AOF medium used in this study. All other authors declare no competing interests.