A Behavior‐Integrated Potency Assay for Quantitative Evaluation of Extracellular Matrix Remodeling by Mesenchymal Stem/Stromal Cells

ABSTRACT

Reliable potency assays are critical for ensuring the therapeutic consistency and manufacturing quality of mesenchymal stem/stromal cells (MSCs). Conventional assays based on soluble factor quantification often fail to capture extracellular matrix (ECM) remodeling capacity and exhibit high variability. We developed a biologically relevant, image‐based potency assay that directly quantifies MSCs‐mediated ECM remodeling under defined biophysical conditions. A collagen‐coated (CL) surface that provides a stable, proliferation‐suppressive microenvironment, enabling single‐cell quantification of remodeling activity using collagen hybridizing peptide staining. The ECM remodeling index (I _CHP+) defined an optimal working range (3.0 × 10³ ≤ cells/cm² ≤ 4.5 × 10³), where spatially separated MSCs achieved high analytical precision (CV < 10%). The CL‐based assay demonstrated robust repeatability across MSCs sources and biologically relevant quantitation limits. Integration of I _CHP+ with γH2AX, enabled simultaneous assessment of functional potency and cell health. Response surface methodology modeling of culture duration, seeding density, and passage number revealed a dome‐shaped relationship, with maximal remodeling at early passages (N _p1–N _p3) and 3–5 days of culture, whereas extended culture (> 6 days) decreased potency alongside elevated γH2AX expression. The model exhibited strong predictive performance (R² > 0.95), supporting its utility for process optimization. Collectively, this mechanism‐based potency assay enables standardized platform for potency evaluation and manufacturing control for consistent, high‐quality MSCs therapeutics.

1. Introduction

Human mesenchymal stem/stromal cells (MSCs) have been extensively expanded in vitro for clinical applications due to their broad spectrum of functional attributes (Song et al. 2023; Liu et al. 2024). These include immunosuppressive and immunomodulatory effects, as well as the capacity for extracellular matrix (ECM) remodeling, which together contribute to the multifunctional phenotype of MSCs and underpin their potential for diverse therapeutic applications. However, despite this broad therapeutic potential, the clinical translation of MSCs‐based therapies remains limited by substantial functional heterogeneity (Palamà et al. 2025; Olmedo‐Moreno et al. 2022). This heterogeneity results from both donor‐to‐donor variability and differences in in vitro culture conditions, which often give rise to marked inconsistencies in cellular behavior. These variations pose a significant challenge to the reproducibility and standardization of MSC‐based products, frequently leading to divergent phenotypic profiles and unpredictable therapeutic outcomes. The extent of this variability reveals critical gaps in current quality control practices and emphasizes the necessity for robust, mechanism‐informed potency assays capable of quantitatively assessing the influence of specific culture parameters on MSC functionality (Rebelatto et al. 2023). Incorporating such assays into manufacturing workflows would enable early identification of population instability, promote therapeutic consistency, and enhance the reliability of clinical performance. Therefore, the development of a reliable, cell‐based in vitro functional assay is essential to ensure consistent MSCs production and to establish potency metrics aligned with therapeutic efficacy.

To enable practical implementation and ensure reliable assessment of MSCs potency, the assay's analytical performance should be validated in accordance with the International Council for Harmonisation (ICH) Q2 guidelines (European Medicines Agency EMA 2023; Gimona et al. 2021). The validation should address key parameters, including working range, quantitation limit, precision, and robustness, thereby meeting regulatory expectations and supporting the development and manufacturing of cell‐based therapies. To meet these analytical requirements, the assay must quantify a biological readout that is both measurable and mechanistically linked to the therapeutic function of MSCs. Among the various biological activities attributed to MSCs, their anti‐fibrotic function is particularly relevant, as it plays a key role in tissue repair and fibrosis resolution. MSCs mediate these effects through the secretion and activation of ECM remodeling enzymes, including matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs) (Burk et al. 2022). Critically, this function is cell‐intrinsic and matrix‐contact‐dependent and cannot be reliably captured by bulk cytokine secretion assays. However, most conventional potency assays rely on bulk measurements of soluble factors—such as cytokines, growth factors, or MMPs—released into the culture medium. While these readouts reflect the overall paracrine activity of MSC populations, they fail to capture functional heterogeneity within these inherently diverse cell populations (Nataliya et al. 2023; Sagaradze et al. 2023). This limitation is particularly significant, as MSC subpopulations can differ markedly in their capacity for ECM remodeling. Moreover, the inability of conventional assays to evaluate ECM remodeling activity at the single‐cell level restricts their capacity to accurately define the therapeutic potential of individual MSCs. In addition, many existing assay formats permit cell proliferation during the assay period. Such proliferation can alter the initial population composition and obscure the intrinsic functional properties of the input cells, resulting in readouts that reflect proliferation‐driven changes rather than the true potency of the starting MSC population. Therefore, establishing a more robust and biologically meaningful assessment of MSC potency requires assay designs that explicitly address these limitations by incorporating both single‐cell resolution and mechanistic relevance. Therefore, ECM remodeling capacity―particularly, anti‐fibrotic collagenolytic activity mediated by MMP‐1―was selected as the primary mechanism underlying this assay, with anti‐fibrotic therapy as the intended clinical application context.

In our previous study, we investigated the mechanism by which MSCs remodel the ECM on a collagen‐coated (CL) surface, with a focus on cell migratory behavior and collagen fiber reorganization (Kim et al. 2023). This remodeling process involves dynamic cell migration accompanied by focal adhesion turnover, as indicated by paxillin localization at leading edges, and the generation of traction forces that physically reorganize collagen fibrils. Concurrently, localized collagen degradation occurs through the activity of MMPs, particularly MMP‐1 and MMP‐9, which facilitate matrix remodeling and cell motility (Kim et al. 2023; Zhao et al. 2023). The CL surface mimics aspects of the early in vivo microenvironment and provides a physiologically relevant context for assessing MSCs remodeling activity (Seo et al. 2020).

Building on this foundation, the present study aimed to develop and analytically validate a biologically relevant, mechanism‐informed, and highly reproducible potency assay that directly quantifies ECM remodeling by MSCs. The assay employs a CL surface with a densely structured collagen fibril layer, designed to suppress cell proliferation and preserve population composition while enabling single‐cell assessment of ECM remodeling capacity. A quantitative ECM remodeling potency index is derived using collagen hybridizing peptide (CHP) staining, and mechanistic validation is supported by the analysis of migration‐dependent remodeling behaviors. These features collectively establish a biologically meaningful and analytically robust working range. The assay demonstrates high analytical precision (coefficient of variation (CV < 10%), strong repeatability, and defined quantitation limits. Additionally, incorporation of γH2AX as a biomarker of DNA damage and cellular senescence enables simultaneous evaluation of MSCs functional capacity and health status. Together, these elements constitute a standardized, mechanism‐based assay platform suitable for potency testing, process optimization, and manufacturing control in MSCs‐based therapeutic development.

2. Materials and Methods

2.1. Cells and Culture Conditions

Human adipose tissue‐derived mesenchymal stem/stromal cells (AT‐MSCs; PromoCell, Germany) and bone marrow‐derived mesenchymal stem/stromal cells (BM‐MSCs; Lonza, USA) were purchased from commercial suppliers. Human amniotic tissue‐derived mesenchymal stem/stromal cells (AM‐MSCs) were kindly provided by Prof. K. Yamahara (Hyogo Medical University, Japan). All MSCs were expanded on tissue culture‐treated polystyrene (PS) surfaces (Corning Inc., USA) in MSCs growth medium (Lonza, USA) maintaining a temperature of 37°C, with a humidified atmosphere of 5% CO₂. Cells were detached by enzymatic treatment with 0.1% trypsin/0.02% ethylenediaminetetraacetic acid (EDTA) solution (Sigma‐Aldrich, USA) when cultures reached 70% confluency. Cells were reseeded onto culture surface at a standard seeding density of 3.0 × 10³ cells/cm². For assay evaluation experiments, passage 1 AT‐MSCs were used, and cultures were maintained for 4 days unless otherwise specified. To evaluate donor variability, three independent AT‐MSC donors were employed, comprising two donors obtained from PromoCell (Germany) and one donor from CellSource (Japan). To assess the effects of culture parameters, AT‐MSCs at passages 3, 5, 7, and 9 were additionally employed.

2.2. Establishment of CL ‐Integrated ECM Remodeling Potency Assay

Figure 1 outlines the systematic workflow used to establish a biologically relevant and reproducible potency assay. The process begins by defining the intended purpose of the assay and explicitly linking the underlying biological mechanism to that purpose. After selecting an appropriate assay format, feasibility studies are performed to confirm experimental practicality and to identify quantitative index. Initial assay development is then undertaken to establish baseline analytical conditions. Subsequently, assay parameters are refined by key variables such as coating conditions, culture environment, and analytical endpoints. The resulting datasets are evaluated for analytical working range, repeatability, and robustness to ensure assay validity and reproducibility. Specifically, we (i) determined the working range of cell seeding densities that preserves single‐cell resolvability and yields acceptable precision; (ii) established quantitation limits (LLOQ/ULOQ) at the boundaries of this range based on variance behavior and loss of single‐cell signal; (iii) verified precision by targeting a coefficient of variation (CV < 10%) across technical replicate measurements; (iv) demonstrated repeatability (intra‐assay: same operator/day; inter‐assay: different days/aliquots) with CV < 10%; and (v) confirmed robustness across independent donor‐derived MSCs lots under a standardized protocol. Finally, the validated assay procedures are documented and implemented as a standardized method, providing a robust framework for consistent potency evaluation and manufacturing control.

Figure 1.

Systematic workflow for the development and validation of a biologically relevant and reproducible ECM remodeling potency assay. The flowchart outlines the stepwise process to establish a biologically relevant and reproducible potency assay. Key steps include: defining assay purpose and mechanistic linkage, confirming feasibility and index quantifiability, and developing baseline assay conditions. Assay parameters (e.g., coating density, seeding density, analytical endpoints) are optimized and validated through determination of analytical working range, quantitation limits (LLOQ/ULOQ), precision (CV < 10%), repeatability (intra‐/inter‐assay), and robustness across donor‐derived MSCs. The validated assay is implemented as a standardized method for potency assessment and process control.

The detailed procedure for establishing the CL‐integrated ECM remodeling potency assay is described in the following steps.

2.2.1. Step 1. Preparation of CL Surface

To generate a surface with high‐density collagen fibrils suitable for ECM remodeling analysis, a CL surface was prepared as described previously (Kino‐oka et al. 2005). An acid‐solubilized bovine collagen type I (0.5% collagen I‐AC, Koken Co., Ltd., Japan) was coated onto a conventional 3.5 cm² PS culture dishes (μ‐Dish 35 mm; Ibidi) under aseptic conditions. For experiment to determine the standardized density, dishes were prepared at coating densities of 0.131, 0.263, 0.525, 0.700, 0.875, and 1.05 mg/cm². The coated dishes were then dried in a vacuum chamber for 3 days to promote fibril formation. Before cell seeding, all dishes were rinsed twice with sterile phosphate‐buffered saline (PBS; Sigma‐Aldrich).

2.2.2. Step 2. Staining and Image Acquisition for ECM Remodeling

To detect ECM remodeling potency at the single‐cell level, cells were fixed with 4% paraformaldehyde in PBS (Fujifilm Wako Chemical Corporation, Japan) for 10 min at room temperature and rinsed twice with PBS. Cells were then incubated overnight at 4°C with heat‐treated collagen hybridizing peptide (CHP; R‐CHP, 3Helix Inc., USA). Then, cells were rinsed with PBS twice and soaked in PBS of 0.5% polyoxyethylene octyl phenyl ether (Triton X‐100; Fujifilm Wako Chemical Corporation) for 10 min. After blocking non‐specific proteins with Block Ace (Dainippon Sumitomo Pharma Co., Ltd., Japan) for 90 min at room temperature, the cells were incubated with primary antibodies against phospho‐histone H2AX (γH2AX; Cell Signaling Technology, USA) in PBS containing 10% Block Ace at 4°C overnight. The cells were first washed with Tris‐buffered saline (TBS; Dako, USA). Then, they were incubated with secondary antibodies, Alexa Fluor 647‐conjugated goat anti‐rabbit IgG (Thermo Fisher, USA) for 1 h at room temperature. γH2AX staining was performed when evaluating the effects of culture parameters on cellular condition. F‐actin was stained using Alexa Fluor 488 phalloidin (Life Technologies, USA), while DAPI was used to stain the nuclei. Finally, images were acquired using a fluorescence microscope (Axio Observer; Zeiss, Germany) with a 10× objective lens.

2.2.3. Step 3. Quantification of ECM Remodeling Potency Index

To evaluate the ECM remodeling capacity of cells, we defined the ECM remodeling potency index (I _CHP+) as the ratio of collagen hybridizing peptide‐positive (CHP + ) cells to the total number of nuclei stained with DAPI (DAPI + ). For each experimental condition, at least four randomly selected microscopic fields were analyzed, ensuring that a minimum of 100 cells were evaluated to provide statistical robustness.

The ECM remodeling potency index was calculated as follows:

$$\text{ECM}\text{remodeling}\text{potency}\text{index}:I_{\text{CHP}+}(-)=\frac{\text{Number of CHP}+\text{cells}}{\text{Number of DAPI}+\text{cells}}$$ (1)

To further assess the remodeling potential of MSCs in relation to DNA damage incurred during culture, we incorporated the DNA damage marker γH2AX. This allowed us to stratify the MSCs population into two subgroups: “healthy cells” (γH2AX‐) and “damaged cells” (γH2AX‐). Subpopulation‐specific remodeling indices were calculated as follows:

For healthy cells (γH2AX‐):

$$I_{\text{CHP}+/\text{γH}2\text{AX}-}(-)=\frac{\text{Number of CHP}+/\text{γH}2\text{AX}-\text{cells}}{\text{Number of γH}2\text{AX}-\text{cells}}$$ (2)

For damaged cells (γH2AX+):

$$I_{\text{CHP}+/\text{γH}2\text{AX}+}(-)=\frac{\text{Number of CHP}+/\text{γH}2\text{AX}+\text{cells}}{\text{Number of γH}2\text{AX}+\text{cells}}$$ (3)

To enable quantitative comparison of ECM remodeling potency across different culture conditions, remodeling index values obtained after culture were normalized to baseline values measured immediately after thawing (pre‐culture). This normalization yielded a relative cell potency level, reflecting culture‐induced changes in functional remodeling capacity. The relative potency was defined as the base‐10 logarithm of the ratio between post‐culture and pre‐culture remodeling indices:

$$\text{Relative cell potency level}(-)={\text{Log}}_{10}\frac{\text{ECM remodeling index of cells after culture}}{\text{ECM remodeling index of cells before culture}}$$ (4)

The resulting dataset provided a quantitative framework for evaluating how specific culture variables influence MSC functional performance. For multi‐parameter optimization, the dataset was further analyzed using response surface methodology (RSM) based on a third‐order polynomial regression model, as previously reported (Roeb 2018; Prihadi et al. 2018; Patel et al. 2017; Chen et al. 2022; Kusena et al. 2021).

2.3. Validation of the CL‐Integrated ECM Remodeling Potency Assay

2.3.1. Working Range

The analytical working range was defined as the range of cell seeding densities over which the assay exhibited consistent and precise performance. Specifically, the working range corresponded to the interval in which the ECM remodeling index remained constant, and the coefficient of variation (CV) was < 10%. These criteria ensured reliable single‐cell level quantification and confirmed the assay's reproducibility and robustness under these conditions.

2.3.2. Quantitation Limit

The lower and upper limits of quantitation (LLOQ and ULOQ) were established based on the boundaries of the validated working range. Cell densities below the LLOQ demonstrated increased variability and inconsistent ECM remodeling index values, indicating insufficient assay reliability. Conversely, cell densities above the ULOQ led to cell overcrowding, which impaired single‐cell resolution and attenuated remodeling signals. These observations defined the functional quantitation limits of the assay.

2.3.3. Repeatability

Repeatability was assessed at both intra‐assay and inter‐assay levels to determine the assay's consistency within and across experimental runs. Intra‐assay repeatability was assessed using ten technical replicates prepared from a single MSCs batch seeded at a standardized density of 3.0 × 10³ cells/cm². All replicates were processed and analyzed on the same day by a single operator to minimize operator‐induced variability. The resulting remodeling index yielded CV < 10%, which was considered acceptable.

Inter‐assay repeatability was determined through three independent experiments performed on separate days using distinct cryovials from the same MSCs lot. The remodeling index remained consistent across experiments, with a CV < 10%, confirming inter‐assay reliability.

2.3.4. Robustness

Assay robustness was assessed by performing the complete performance using MSCs obtained from three independent donor lots. All samples were performed according to the standardized protocol. The remodeling index was calculated for each donor‐derived sample, and a CV < 10% across the three lots was considered indicative of acceptable biological robustness.

2.4. Mechanistic Analysis

2.4.1. Quantitative Evaluation of Cell Migratory Behaviors by Time‐Lapse Imaging

Time‐lapse imaging was conducted using a 4 × phase‐contrast objective lens (BioStudio; Nikon Engineering Co., Japan). Images were acquired every 30 min for up to 72 h. Cell migratory behavior was quantitatively analyzed by calculating the migration rate and turning angle for individual cells. For each seeding condition, a total of 45 cells were manually tracked for statistical analysis.

The cell migration rate, V (μm/h) was calculated based on the total displacement of a cell's centroid during the tracking period, using ImageJ software (National Institutes of Health, USA) with the Manual Tracking plugin, according to the following equation:

$$V\left(\text{μm}/\text{h}\right)=\frac{d_{\text{total}}}{t_{\text{n}}-t_0}$$ (5)

where d _total is the cumulative path length obtained by summing stepwise frame‐to‐frame displacements across the tracking, and t _n and t ₀ denote the final and initial time points of tracking, respectively.

The turning angle, θ (degrees) was calculated to evaluate changes in migration direction between two consecutive displacement vectors using the following equation:

$$\theta (°)={\cos }^{-1}\left(\frac{\overrightarrow{r_1}·\overrightarrow{r_2}}{|\overrightarrow{r_1}||\overrightarrow{r_2}|}\right)\times \frac{180}{\text{π}}$$ (6)

where $\overrightarrow{r_1}$ is the displacement vector between the first two frames (from t ₀ to t ₁) and $\overrightarrow{r_2}$ is the subsequent displacement vector (from t ₁ to t _n). The turning angle θ represents the change in migration direction between consecutive steps in the cell trajectory. The highest frequency of movement for each experimental condition was displayed within the 330°–30° range.

2.4.2. Scanning Electron Microscopic Observation

Scanning electron microscopy (SEM) samples were prepared as previously described (Matsumoto et al. 2022). Samples were fixed in 2% glutaraldehyde (Electron Microscopy Sciences, USA) in 0.1 M potassium dihydrogen phosphate/disodium hydrogen phosphate buffer at 4°C and rinsed in the same buffer for overnight at 4°C. The specimens were then dehydrated through a graded ethanol series (Nacalai Tesque Inc., Japan), followed by critical‐point drying. The dried samples were coated with a thin osmium layer using a plasma ion coater and observed with a field‐emission SEM (S‐4800; Hitachi High‐Tech America Inc., USA) operated at 7–15 kV.

2.4.3. Quantification of Cytokine Production via Enzyme‐Linked Immunosorbent Assay

Quantification of cytokine production was performed using an enzyme‐linked immunosorbent assay (ELISA). Conditioned media were collected and stored at −20°C until use. MMP‐1 and TIMP‐1 levels in culture medium were measured using a commercial ELISA kit (R&D Systems Inc., USA) in accordance with the manufacturer's protocol. Briefly, 100 μL of standard dilutions of anti‐human MMP‐1 and TIMP‐1 antibodies and collected medium were dispensed in a 96‐well plate coated with MMP‐1 or TIMP‐1 antibodies, respectively, to each well. The plate was sealed and incubated at room temperature for 2 h, followed by washing with PBS containing 0.05% Tween 20. Then, 100 μL of diluted antibodies were added to each well with sealed and incubated at room temperature for 2 h. After the plate was washed, 100 μL of diluted streptavidin conjugated to horseradish‐peroxidase was added. The plate was placed in the dark for 20 min at room temperature. A 1:1 mixture of H₂O₂ and tetramethylbenzidine of 100 µL was applied for 20 min. After the addition of 50 μL H₂SO₄ of stop solution, optical density values of each well were calculated with comparable standard curves by microplate spectrophotometer (TriStar 2; Berthold Technologies, Germany) at 450 nm. Concentration of secretomes (pg/ml) were determined from the standard curve.

2.5. Statistical Analysis

Data is presented as the mean ± standard deviation (SD) from at least three independent experiments. Statistical comparisons between groups were performed using one‐way or two‐way analysis of variance (ANOVA), followed by Tukey's post hoc test to identify significant main effects. Differences were considered statistically significant at p < 0.05, p < 0.005, or p < 0.0001. The CV was calculated to evaluate assay repeatability. A third‐order polynomial regression model was used to simulate surface trends and optimize culture parameters using OriginPro 2025 (OriginLab Corporation, USA). All statistical analyses were performed using GraphPad Prism version 10.3 (GraphPad Software Inc., USA).

3. Results

3.1. Identification of Assay Conditions for the CL‐Integrated ECM Remodeling Potency Assay

To establish appropriate assay conditions, we first characterized the physical properties of the CL surface. As shown in Figure 2A, SEM revealed that the CL surface exhibited a dense collagen fibrillar network with nanoscale architecture (≈100 nm band spacing), closely resembling native ECM organization and early fibrotic matrix formation. This structure closely mimics the in vivo collagen matrix architecture (Matsumoto et al. 2022; Tjin et al. 2017).

Figure 2.

Identification of assay conditions for CL‐integrated ECM remodeling potency assay. (A) Scanning electron microscopy (SEM) of the CL surface showing dense collagen fibrillar structure. Inset shows magnified view (scale bars: 5 μm, 100 nm). (B) Schematic workflow of MSCs culture on CL and PS surfaces over 120 h to evaluate time‐dependent remodeling activity. (C) Phase‐contrast images of MSCs morphology on CL and PS surfaces at 24, 72, and 120 h (scale bar: 50 μm). (D) Fluorescence images of collagen hybridizing peptide (CHP), F‐actin, and DAPI staining at each timepoint (scale bar: 50 μm). (E) Quantification of cell density, X _t (cells/cm²) and I _CHP+ during early (24–72 h) and late (72–120 h) phases. Data represents mean ± standard deviation (SD) (n = 3). Statistical analysis by two‐way analysis of variance (ANOVA), followed by Tukey's post hoc test (**p < 0.005, **** p < 0.0001). (F) Illustration summarizing the design criteria for the CL‐integrated assay platform, highlighting the 72 h timepoint on the CL surface as a remodeling‐competent phase characterized by maximal ECM remodeling activity and limited MSCs proliferation—defining an optimal window for functional potency assessment.

Subsequently, we examined the time‐dependent changes in AT‐MSCs morphology and ECM remodeling capacity on CL and PS surfaces over a 120‐h culture period to determine the appropriate incubation time for the ECM remodeling potency assay (Figure 2B). As shown in Figure 2C, phase‐contrast imaging revealed MSCs on the CL surface retained a spindle‐shaped morphology whereas on the PS surface, progressive cell spreading and increased confluency was observed throughout the culture period. To further characterize remodeling behavior, fluorescence imaging was performed using collagen hybridizing peptide (CHP), F‐actin, and DAPI staining (Figure 2D). MSCs on the CL surface exhibited detectable CHP‐positive (CHP + ) signals as early as 24 h, which markedly increased by 72 h and persisted through 120 h, although single‐cell evaluation became difficult due to increased cell density. In contrast, on the PS surface, CHP+ signals were scarcely detected until 120 h, likely reflecting remodeling of cell‐derived ECM rather than direct interaction with the substrate. To objectively assess these qualitative observations, ECM remodeling index (I _CHP+) was subsequently quantified.

Quantitative analysis of cell density and I _CHP+ over time (Figure 2E) further supported these morphological observations. On the CL surface, cells exhibited minimal proliferation during early phase (24‐72 h), followed by a gradual increase at late phase (72‐120 h; p < 0.0001). In contrast, cells on the PS surface proliferated continuously from 24 h to 120 h (p < 0.0001), with density reaching significantly higher levels than the CL surface by 72 h (p < 0.0001). A similar trend was observed for I _CHP+ analysis. On the CL surface, I _CHP+ remained constant during early phase before increasing significantly after 72 h (p < 0.005). Conversely, I _CHP+ was zero on the PS surface until 120 h mark. At 72 h, I _CHP+ levels were significantly higher on the CL surface than on PS surface (p < 0.0001), confirming that the observed remodeling phenomena were specific to the CL surface and could not be captured on PS surface. Taken together, the growth profile and ECM remodeling index collectively indicate that MSCs on the CL surface remain in a remodeling‐competent state at 72 h. This phase is characterized by collagen turnover (I _CHP+ signal) under proliferation‐suppressed conditions, consistent with previous studies reporting active MMP secretion (Kim et al. 2023; Roeb 2018; Prihadi et al. 2018). This 72‐h time point thus represents an optimal window for assessing the ECM remodeling potency of individual MSCs (Figure 2F).

3.2. Determination of Collagen Type I Coating Density for CL‐Integrated ECM Remodeling Potency Assay

To identify the optimal collagen type I coating density for the standardized ECM surface, the I _CHP+ was quantified across a defined range of coating densities (0.131‐1.05 mg/cm²) (Figure 3A). As shown in Figure 3B, I _CHP+ values increased progressively from 0.131 to 0.525 mg/cm², then plateaued and remained stable up to 1.05 mg/cm². Notably, measurement variability decreased with increasing collagen density, with coefficients of variation (CVs) falling below 10% at densities above 0.525 mg/cm². The lowest variability was observed at 1.05 mg/cm² (CV = 5.12%), supporting its selection as the standardized coating condition for minimizing technical variability and enhancing assay reproducibility. These results directly address within‐study substrate variability: by fixing the coating density at 1.05 mg/cm² with a single commercially standardized bovine collagen product, the protocol ensures a consistent and reproducible ECM substrate across assay runs.

Figure 3.

Determination of coating density for Type I collagen for ECM remodeling potency assay. (A) Schematic workflow for assessment of coating density. MSCs were cultured on CL surfaces coated with increasing collagen density (0.131–1.05 mg/cm²), and I _CHP+ was evaluated at 72 h. (B) Scatter plots of I _CHP+ values at each coating density (n = 6 per density) with coefficient of variation (CV) shown above each plot to indicate variability and identify the coating density for reproducible evaluation. The red line represents a four‐parameter logistic sigmoidal dose‐response fit.

3.3. Determination and Validation of Optimal MSCs Seeding Density for CL‐Integrated ECM Remodeling Potency Assay

To determine the optimal MSCs seeding density on the standardized CL surface (collagen type I, 1.05 mg/cm²), cells were plated at densities ranging from 0.5 to 12.0 × 10³ cells/cm², and the ECM remodeling index was quantified to define the assay's analytical working range (Figure 4A). As shown in Figure 4B, I _CHP+ values increased progressively from 0.5 to 3.0 × 10³ cells/cm², followed by a plateau that was maintained up to 12.0 × 10³ cells/cm². Variability was higher at lower seeding densities (0.5–2.5 × 10³ cells/cm²; CV > 10%) and decreased substantially, remaining at or below 10% at densities above 3.0 × 10³ cells/cm².

Figure 4.

Determination and validation of the optimal MSC seeding density for defining the analytical working range of the CL‐integrated ECM remodeling potency assay. (A) Schematic workflow for determining the optimal cell seeding density. MSCs were seeded onto standardized CL surfaces (collagen type I, 1.05 mg/cm²) at densities ranging from 0.5 to 12.0 × 10³ cells/cm², followed by assessment for I _CHP+, proteases secretion and migratory behavior. (B) Scatter plots show ECM remodeling index, I _CHP+ values across seeding densities. ECM remodeling potency was quantified by analyzing 100 individual cells per well (n = 6 wells per condition). Coefficient of variation (CVs) are indicated above each data cluster. Red line represents a four‐parameter sigmoidal logistic fit. Based on I _CHP+ trends and CV distribution, seeding densities were categorized into three analytically defined ranges: Range I (0.5–2.5 × 10³ cells/cm²), Range II (3.0–4.5 × 10³ cells/cm²), and Range III (5.0–12.0 × 10³ cells/cm²). Green dotted lines denote the lower and upper limits of quantitation. (C) Concentration of MMP‐1 and TIMP‐1 (pg/ml) in culture supernatant across seeding densities, plotted on a logarithmic scale, and quantified by enzyme‐linked immunosorbent assay (ELISA) for a total of 6 wells per condition. Red line represents a third‐order polynomial fit for MMP‐1 and a four‐parameter variable slope sigmoidal fit for TIMP‐1. (D) Scatter plot of cell migration rates, V measured over 72 h. Each dot represents a single cell (n = 45 cells per condition). Rose plots of turning angles for the representative seeding density condition in each range. Red line represents a four‐parameter sigmoidal logistic fit. Dotted line indicates representative density in each range: 0.5 × 10³ cells/cm² (purple, Range I), 3.0 and 4.5 × 10³ cells/cm² (green, Range II), and 12.0 × 10³ cells/cm² (blue, Range III). (E) Conceptual summary illustrating how cell migratory behavior transitions across different seeding densities and its implications for ECM remodeling assay accuracy and precision.

To determine whether ECM remodeling was attributable to changes in matrix remodeling enzyme secretion, levels of MMP‐1 and TIMP‐1 were quantified across the same range of seeding densities using ELISA (Figure 4C). MMP‐1 levels showed an increasing trend as seeding density increased from 0.5 to 7.0 × 10³ cells/cm², followed by a constant level maintained up to 12.0 × 10³ cells/cm². In contrast, TIMP‐1 levels mirrored the trend of the remodeling index, increasing significantly from low densities and reaching a stable plateau beyond 3.0 × 10³ cells/cm²—a level that was maintained up to 12.0 × 10³ cells/cm². This parallel pattern suggests that ECM remodeling potency, as reflected by I _CHP+, is closely associated with the secretion levels of matrix remodeling enzymes under varying cell densities.

To investigate whether cell behavior within the assay microenvironment influences assay accuracy, quantitative single‐cell tracking was performed across the same range of seeding densities. For each condition, both the cell migration rate and directional variability—quantified as turning angle—were assessed by tracking 45 individual cells per well, ensuring consistent single‐cell resolution. The cell migration rate, V remained relatively high and stable from 0.5 to 3.0 × 10³ cells/cm², followed by a gradual decline between 3.0 and 5.0 × 10³ cells/cm² (Figure 4D). Beyond this range, migration rates remained relatively constant up to 12.0 × 10³ cells/cm². Based on these data, we defined the 0.5 × 10³ cells/cm² (Range I, low density), 3.0–4.5 × 10³ cells/cm² (Range II, intermediate density), and 12.0 × 10³ cells/cm² (Range III, high density).

Distinct migratory patterns were observed across these density ranges (Figure 4D, Movie S1). In Range I, turning angles were widely distributed (0–360°), reflecting highly dynamic and random migration. However, sparse seeding resulted in long, overlapping trajectories, which hindered the accurate attribution of ECM remodeling activity to individual cells due to limited spatial resolution. In Range II, although turning angles remained randomly oriented, cell trajectories became spatially confined near their points of origin. This confinement reduced trajectory overlap and enabled accurate and reproducible tracking of ECM remodeling at the single‐cell level. In Range III, turning angles were narrowly distributed (330–30°), indicating directionally biased and restricted movement. Excessive cell crowding under these conditions led to short, aligned, and overlapping trajectories, which severely compromised the resolution necessary for single‐cell‐level analysis of ECM remodeling.

As summarized in Figure 4E, combined analysis of ECM remodeling potency and migration dynamics revealed a clear density‐dependent transition in assay performance. At low densities, excessive motility and trajectory overlap limited analytical precision, whereas at high densities, crowding restricted migration and reintroduced path overlap. Only within the intermediate‐density range (Range II: 3.0–4.5 × 10³ cells/cm²) were high motility, spatial separation, and low variability (CV < 10%) achieved simultaneously. Taken together, these findings define Range II as the optimal analytical working range for the CL‐integrated ECM remodeling potency assay, supporting consistent and reproducible single‐cell–level functional assessment.

3.4. Assay Validation: Intra‐/Inter‐Assay Repeatability and Robustness

To evaluate the precision of the CL‐integrated ECM remodeling potency assay, both repeatability and robustness were assessed by analyzing the CV and SD of the I _CHP+. Intra‐assay repeatability was assessed using ten technical replicates (n = 10) derived from a single MSC batch. The I _CHP+ value demonstrated low variability, with CVs ranging from 3.46% to 5.40%, well within the generally accepted threshold of less than 10% for bioassay precision (Figure 5A). Inter‐assay repeatability was evaluated through three independent experiments conducted on separate days using distinct vials of cryopreserved MSCs from the same lot. Across these experiments, I _CHP+ values remained consistent, with a standard deviation of 0.024 (Figure 5B), confirming that I _CHP+ is a stable and reproducible metric under varying experimental conditions.

Figure 5.

Assessment of intra‐assay and inter‐assay repeatability and robustness of I _CHP+ across different MSCs sources. (A) Intra‐assay repeatability evaluated across n = 10 cultures within a single MSCs batch. CV were calculated for each batch. (B) Inter‐assay repeatability assessed in three independent experiments (n = 3 batches) of MSCs derived from the same tissue source. SD were calculated for each condition. (C) Robustness of the assay was evaluated using MSCs derived from adipose tissue (AT‐MSCs, black), amniotic tissue (AM‐MSCs, red), and bone marrow (BM‐MSCs, blue), all tested under identical assay conditions.

To evaluate assay robustness across different MSCs sources, cells derived from adipose tissue (AT‐MSCs), amniotic tissue (AM‐MSCs), and bone marrow (BM‐MSCs) were tested under standardized assay conditions. Each MSCs source was analyzed in biological triplicates (n = 10 per source), and the CV of the ECM remodeling index was calculated to determine robustness (Figure 5C). The mean I _CHP+ values were consistent across MSCs sources, 0.82 for AT‐MSCs, 0.71 for AM‐MSCs, and 0.79 for BM‐MSCs. Importantly, I _CHP+ demonstrated CV values below 10% across all three MSCs types (5.40%, 6.93%, and 3.95%, respectively), indicating high reproducibility.

Collectively, these results confirm that the ECM remodeling potency assay, as defined by the I _CHP+ index, demonstrates high repeatability and robustness. The assay yields consistent and reproducible results across both technical and biological replicates and is applicable to MSCs derived from diverse tissue origins under standardized conditions. To further evaluate assay performance across donors within the same tissue source — a critical consideration for manufacturing lot screening — I _CHP+ was assessed across 3 independent healthy AT‐MSC donors under identical standardized assay conditions (Figure S1). I _CHP+ mean values showed differences across donors, 0.82 for donor 1, 0.53 for donor 2, and 0.67 for donor 3, while CV remained below 10% within each individual donor lot (5.40%, 5.52%, and 8.04%, respectively). These findings confirm that the assay possesses sufficient biological dynamic range to discriminate between donors of differing potency, supporting its utility as a starting material quality screening confirming robust reproducibility at the donor level.

3.5. Optimization of MSCs Culture Parameters Using the CL‐Integrated ECM Remodeling Potency Assay and Response Surface Modeling

To demonstrate the applicability of the validated CL‐integrated ECM remodeling potency assay for culture process optimization, we systematically applied the assay to evaluate how various culture conditions influence the ECM remodeling capacity of AT‐MSCs. MSCs were first thawed and then subjected to a series of culture conditions varying in culture time, seeding density, and passage number (Figure 6A). Following culture, cells were harvested and assessed using the CL‐integrated potency assay, which quantitatively measures ECM remodeling activity at the single‐cell level. To further characterize functional status, γH2AX—a DNA damage marker associated with cellular senescence—was employed to identify damaged or senescent cells within the cultured population. Remodeling index values were normalized to pre‐culture baselines to calculate relative potency, enabling direct comparison across conditions. The resulting dataset served as a basis for multi‐variable optimization using response surface modeling (RSM).

Figure 6.

Optimization of MSC culture parameters using the CL‐integrated ECM remodeling potency assay and response surface modeling (RSM). (A) Schematic overview of the experimental design. MSCs were thawed and cultured under systematically varied conditions, including seeding density, culture duration, and passage number. Post‐culture, cells were evaluated using the CL‐integrated ECM remodeling potency assay to assess single‐cell functional remodeling capacity, stratified by γH2AX expression status. (B) Relative potency levels for total (I _CHP+), healthy (I _{CHP+/γH2AX‐}), and damaged (I _CHP+/γH2AX+) cell populations across a range of culture durations (1–20 days). (C) Relative potency levels across a range of seeding densities (1.0–27.0 × 10³ cells/cm²) at a fixed culture duration. (D) Relative potency levels across different passage numbers (N _p1–N _p9), showing progressive functional decline with increasing passage. (E) Response surface models illustrating the combined effect of culture duration and passage number on ECM remodeling potency. Dome‐shaped distributions were observed for both I _CHP+ and I _{CHP+/γH2AX‐} indices, with peak values at early passages (N _p1–N _p3) and intermediate culture durations (3–5 days). Data represents mean values from three independent experiments. All statistical analyses were performed using third‐order polynomial regression and validated by ANOVA (p < 0.05 for accepted models).

To evaluate the effects of each parameter, we systematically varied culture time (2–20 days), then seeding density (1.0–27.0 × 10³ cells/cm²) at the selected optimal culture time and finally passage number (N _p1–N _p9). As shown in Figure S2, all three ECM remodeling indices (I _CHP+, I _{CHP+/γH2AX‐}, and I _CHP+/γH2AX+) exhibited a gradual decline with increasing culture time, seeding density, and passage number. When normalized to baseline (post‐thaw) values, the relative potency levels both the total (I _CHP+) and healthy (I _{CHP+/γH2AX‐}) cell populations followed consistent patterns across all parameters. Increases were observed when culture time was kept within 8 days, seeding density was maintained at or below 9.0 × 10³ cells/cm², and cells were used within the early passages (up to N _p3). Once these parameters extended beyond those ranges—specifically, when culture time exceeded 8 days, seeding density surpassed 9.0 × 10³ cells/cm², or cells were used at later passages—a consistent reduction in potency was observed (Figure 6B,C). In contrast, the ECM remodeling potency of damaged cell population (I _CHP+/γH2AX+) exhibited a non‐linear trend. This population showed reduced activity under short culture time (less than 3 days) and low seeding densities (below 2.0 × 10³ cells/cm²). The ECM remodeling potency then increased when culture time reached 3 days or more and seeding density was raised to 2.0 × 10³ cells/cm² or higher, peaking within an intermediate range (approximately 4 to 8 days of culture and 3.0 to 9.0 × 10³ cells/cm² seeding density). However, further increases in either parameter once again led to downregulation. Regarding passage number, this subpopulation demonstrated a gradual decrease in remodeling potency as passage increased, mirroring the trend observed in the total and healthy cell populations (Figure 6D).

To systematically assess the combined influence of culture parameters on ECM remodeling potency, RSM was applied using a third‐order polynomial regression model, enabling prediction and identification of optimal culture parameters. Among the three evaluated parameters—seeding density, culture time, and passage number—seeding density exhibited the most prominent and consistent impact. Both I _CHP+ and I _{CHP+/γH2AX‐} indices reached maximum values at a seeding density of 3.0 × 10³ cells/cm², followed by a rapid decline at higher densities. Based on this result, seeding density was fixed at this optimal value for subsequent modeling, allowing focused analysis of the interactive effects of culture time and passage number. The resulting RSM models revealed a dome‐shaped surface for both I _CHP+ and I _{CHP+/γH2AX‐}, with peak values consistently observed at early passage numbers (N _p1–N _p3) and short culture time (3–5 days) (Figure 6E). Deviation from these ranges—either through prolonged culture or extended passaging—was associated with a decrease in remodeling potency. The statistical performance of the predictive models was high for I _CHP+ (R² = 0.95, p < 0.05) and I _{CHP+/γH2AX‐} (R² = 0.96, p < 0.05), supporting their validity in modeling ECM remodeling behavior across culture conditions. In contrast, the model for I _CHP+/γH2AX+ demonstrated poor fit (R² = 0.72, p > 0.05) and was excluded from subsequent analysis.

Taken together, these results identify I _{CHP+/γH2AX‐} as the most reliable indicator of ECM remodeling potency under varying culture conditions. While I _CHP+ reflects the total remodeling activity within the MSC population, I _{CHP+/γH2AX‐} provides enhanced resolution by capturing the contribution of non‐damaged, functionally competent cells. This distinction allows for improved assessment of cellular performance and is applicable to defining potency‐based design of experiment and quality control parameters in MSCs culture processes.

4. Discussion

This study demonstrates that the rational design of biomaterial‐based microenvironments is essential for developing high‐precision and reliable cell potency assays. By engineering a structurally stable, high‐density collagen fibril substrate that recapitulates key features of the physiological ECM, we established a reproducible matrix context that ensures uniform cell–matrix engagement and supports robust single‐cell analysis of MSCs behavior. This biomimetic microenvironment serves as a functional platform for capturing dynamic, mechanistically relevant cellular activities that are indicative of therapeutic potential (Kalukula et al. 2025). Within this context, we developed a CL‐integrated potency assay that directly quantifies ECM remodeling by MSCs at the single‐cell level. In contrast to conventional secretome‐based assays (such as IDO induction or cytokine profiling) that infer function from population‐level soluble factor release, our behavior‐driven approach assesses how individual MSCs physically engage with and remodel their surrounding matrix (Chinnadurai et al. 2018). While previous studies have relied on bulk metabolic markers to predict efficacy, these methods often overlook the physical engraftment potential required for tissue repair. This mechanistic strategy, centered on direct matrix interaction and remodeling, enables precise and reproducible potency assessment through the I _CHP+ index. Importantly, it overcomes major limitations of traditional methods—such as high variability, limited reproducibility (Pissas and Eleftheriadis 2019), and the inability to resolve functional heterogeneity at the single‐cell level (Chinnadurai et al. 2018)—thus offering a robust and scalable platform for MSCs quality control and manufacturing optimization.

The CL surface contributes to assay robustness through three core features. First, the CL surface with high‐density collagen fibrils establishes a stable and physiologically relevant microenvironment for MSCs adhesion and interaction with the ECM. The vacuum‐dried collagen coating forms a dense fibrillar network that remains structurally stable throughout culture (Kim et al. 2023), in contrast to conventional collagen‐coated surfaces or low‐density hydrogels, which often undergo partial dissolution or cell‐mediated contraction (Velankar et al. 2012). This stable matrix architecture supports consistent cell behavior, contributing to the assay's reproducibility. Moreover, the CL surface recapitulates the structural characteristics of an early fibrotic matrix and inherently suppresses cell proliferation. This allows potency evaluation within a maintained MSCs population without chemical proliferation inhibitors like Mitomycin C. As Navarrete et al. previously noted, such inhibitors can alter cell metabolism and thus introduce possible batch‐to‐batch variability (Figure 2) (Navarrete et al. 2014). By reducing proliferative noise and matrix instability, this assay enables accurate quantification of ECM remodeling activity—addressing limitations of soluble factor–based assays. Moreover, evidence from our previous work demonstrated that MSCs cultured on conventionally coated collagen surfaces lacking native fibril architecture did not exhibit remodeling activity or CHP+ signal, underscoring that substrate microstructure is a critical determinant of the potency readout. Alternative substrates such as fibronectin, laminin, or Matrigel would engage different MMP isoforms and alter remodeling kinetics in ways not reflective of physiologically relevant fibrillar collagen turnover in fibrotic tissue (Kanta 2015; Khalili‐Tanha et al. 2025). Although in vivo MSC function involves coordinated proliferation, paracrine signaling, and dynamic niche adaptation, the proliferation‐suppressive microenvironment is a deliberate design choice to preserve the input population composition during the assay window, ensuring that the readout reflects the intrinsic functional capacity of the starting cells. Critically, cells retain the functional behaviors relevant to anti‐fibrotic activity, including motility, MMP secretion, and ECM engagement. By regulatory definition, a potency assay measures a defined functional attribute under controlled and reproducible conditions rather than recapitulate full in vivo niche complexity.

Second, the established CL‐integrated ECM remodeling potency assay provides a reproducible approach by capturing mechanism‐driven behavior. This assay's sensitivity and reproducibility were then quantitatively validated by analyzing the relationship between cell seeding density, the image‐based I _CHP+ index, and the CV. The analysis of cell seeding density highlights a fundamental trade‐off between cell behavior and analytical precision that is often overlooked in conventional assays (Figure 4). At low seeding densities (Range I), the observed density‐dependent response of ECM remodeling activity, suggests that under sparse conditions, individual MSCs predominantly engaged in cell–matrix adhesion with minimal cell–cell contact, leading to enhanced migratory potential and highly random motility (Köhler et al. 2025). While conventional assays often utilize low densities to isolate single‐cell effects, our findings demonstrate that the stochastic nature of cell migration leads to path convergence and overlapping trajectories. This overlapping behavior, combined with variability in cell‐matrix interactions, contributed to elevated CV and heterogeneity in ECM remodeling activity, thereby reducing the precision of potency measurements (Bertolo et al. 2015). Unlike previous studies that overlook spatial noise in sparse cultures, our findings suggest that low‐density conditions make it difficult to reliably assess single‐cell potency. In contrast, the saturation of remodeling capacity observed at high seeding densities (Range III), underscores the limitations of prioritizing population‐level output. While increased cell–cell interactions under these conditions promote slower (Kyburz and Anseth 2013) but more coordinated and directional migration (Friedl and Gilmour 2009)—synergizing with aligned collagen fibers to guide persistent movement (Fraley et al. 2015)—the resulting cell crowding compromises spatial resolution. This represents a significant limitation for conventional bulk‐well assays (such as ELISA or secretome‐based PCR). While these methods successfully capture population‐level averages, they cannot resolve the loss of individual cell behavior and signal overlap that occurs when cells are crowded. The most robust and reproducible results were obtained at intermediate seeding densities (Range II), specifically between 3.0 and 4.5 × 10³ cells/cm². Within this range, the assay facilitates balanced and coordinated cell‐matrix‐cell interactions (Saraswathibhatla et al. 2023), in agreement with previous mechanobiological studies that have demonstrated the critical role of mechanical coupling in regulating matrix remodeling (Ozguldez et al. 2018; Zheng et al. 2019; Monteiro et al. 2018). By providing sufficient cell density to stabilize migration without compromising spatial resolution, this optimized microenvironment enables precise quantification of ECM remodeling activity at the single‐cell level. Importantly, this allows the I _CHP+ to maintain a CV below 10% across various donor origins and donor biological variability (Figure 5, Figure S1), satisfying key analytical validation parameters defined in the ICH Q2 guidelines for repeatability and robustness, and further support the assay's utility for starting material screening (European Medicines Agency EMA 2023). This level of analytical precision is a critical benchmark, as conventional secretome‐based assays often suffer from inter‐donor variability and high CVs that exceed these regulatory thresholds. These findings underscore the critical role of cell behavior in determining assay precision, indicating that the design of robust potency assays must account for the dynamic and spatial aspects of cell behavior within their microenvironment.

Third, the CL‐integrated ECM remodeling potency assay provides a robust framework for optimizing culture process parameters in MSCs manufacturing. By incorporating γH2AX as a marker of cellular health status and senescence, we developed stratified indices to evaluate how the initial MSCs condition dictates ECM remodeling outcomes across varying culture process parameters (Figure 6). This approach reveals that remodeling potency is not an isolated metric but is intrinsically tied to the physiological state of cell population, providing a more comprehensive readout than assays that measure function without accounting for senescence. Response surface methodology (RSM) analysis demonstrated a strong model fit (R² > 0.95), supporting its predictive utility for process control. The analysis revealed a dome‐shaped relationship between relative remodeling potency and culture conditions. This functional window peaks at early passage (N _p1‐3) and intermediate culture time (3‐5 days), which aligns with prior observations that early‐passage MSCs retain superior MMP activity and dynamic cell‐matrix interactions (Burk et al. 2022; Dias et al. 2019; Samsonraj et al. 2015). We found that cultures maintained for less than 2 days likely suffer from insufficient matrix engagement, while prolonged cultures beyond 6 days lead to diminished potency driven by senescence‐associated dysfunction (Pustovaloca et al. 2016). This decline, as indicated by elevated γH2AX expression, is consistent with senescence‐driven impairments in cytoskeletal structure, adhesion turnover, and protease secretion reported in advanced‐passage cells (Kumari and Jat 2021; Samsonraj et al. 2015). Taken together, these results demonstrate that ECM remodeling potency is tightly linked to cellular health status and responsiveness to the microenvironment. This further highlights the importance of defining a culture window that preserves functional integrity to ensure consistent and effective MSCs‐based therapeutic applications.

The current assay is well‐suited for integration at critical quality control checkpoints within MSC manufacturing workflows, including cell bank qualification, raw material screening, and pre‐release potency testing, where throughput demands are moderate and functional characterization is prioritized over speed. Advances in automated imaging platforms and high‐content analysis systems further offer a practical pathway to streamline image acquisition and quantification, reducing hands‐on time and enhancing scalability. These adaptations would support the incorporation of this assay as a functional release criterion without imposing prohibitive operational burdens on manufacturing pipelines.

Despite these promising attributes, the assay's current validation is limited to in vitro conditions, and several limitations define clear priorities for future manufacturing‐focused validation. For longitudinal use spanning multiple collagen lots, a brief lot‐qualification step confirming I _CHP+ values of a reference MSC batch fall within a pre‐established acceptance window at 1.05 mg/cm² is recommended. Formal process deviation sensitivity testing, including controlled temperature shifts and media lot changes, will further be necessary to complete regulatory‐grade robustness assessment. Ultimately, correlating I _CHP+ cell frequency with in vivo efficacy readouts in relevant fibrosis animal models will be essential to establish a statistically grounded batch release threshold, linking ECM remodeling activity to therapeutic outcomes and positioning this assay as a robust functional release criterion.

5. Conclusion

This study established a robust platform for MSCs potency assessment, centered on the strategic design of a dense collagen fibril (CL) surface. By providing a proliferation‐suppressive and biophysically stable microenvironment, the CL surface serves as a critical biological benchmark that overcomes the variability inherent in conventional culture substrates, a factor currently overlooked by regulatory guidelines. This CL‐based potency assay quantifies MSC‐mediated ECM remodeling with high precision. By coupling the functional output (CHP+ remodeling) with health status (γH2AX), the platform offers a dual‐layer assessment of cell quality at a single‐cell resolution. Ultimately, this material‐centric assay provides a reproducible and scalable solution for manufacturing systems, making it an ideal candidate for standardizing MSC potency assessment, process control, and regulatory‐compliant quality assurance in cell therapy manufacturing.

Author Contributions

Shao Ying Tan: conceptualization, data curation, formal analysis, investigation, methodology, validation, visualization, writing – original draft. Mee‐Hae Kim: conceptualization; formal analysis, visualization, supervision, writing – review and editing. Masahiro Kino‐oka: funding acquisition, project administration, resources, supervision. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work.

Ethics Statement

The permission to use human mesenchymal stem/stromal cells in this study was granted by Hyogo College of Medicine (approval number: 0325) and Osaka University (approval numbers: R‐53).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Assessment of donor variability in I _CHP+ across three independent adipose tissue‐derived MSC donors (n = 10 per donor batch). CV were calculated for each donor batch.

Figure S2: Evaluation of individual culture parameters based on ECM remodeling indices of (A) I _CHP+, (B) I _{CHP+/γH2AX‐} and (C) I _CHP+/γH2AX+.

Movie S1: Time‐lapse observation of MSC culture on collagen‐coated (CL) surface seeded at density of 0.5 × 10³ cells/cm², 3.0 × 10³ cells/cm², 4.5 × 10³ cells/cm², 12.0 × 10³ cells/cm². These conditions represent the representative densities within the analytically defined ranges (Range I, II and III) in this assay. Cross marks indicate the trajectories and intersection paths of individually tracked cells. Scale bar: 200 μm.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request. Data will be made available on request.