Immunity-and-matrix-regulatory cells promote hyaline-like cartilage repair in osteoarthritis

Abstract

Articular cartilage repair in osteoarthritis (OA) remains a major clinical challenge. Although approaches using mesenchymal stem cells (MSCs) have demonstrated partial efficacy in OA, the regenerated tissue is often unable to form stable hyaline-like cartilage phenotype. Here, we introduce immunity-and-matrix-regulatory cells (IMRCs), derived from human embryonic stem cells (ESCs), as an alternative cell source for OA therapy. Using a multiple-level platform including 2D chondrocytes, 3D OA cartilage organoids (OCOs), rabbit cartilage explants, and a rat OA model, we demonstrate that IMRCs significantly enhance hyaline-like matrix formation while suppressing inflammation, angiogenesis, and fibrotic remodeling. Mechanistically, IMRC-mediated cartilage repair is partially associated with MMP1-dependent matrix remodeling. This study establishes a translational evaluation strategy for IMRCs and points to their potential as an alternative cell source for inducing cartilage repair in OA. Notably, the OCOs established in this study provide a pathologically relevant in vitro platform for evaluating cartilage regeneration in OA.

Graphical abstract

Highlights

• •ESC-derived IMRCs promote hyaline-like cartilage regeneration and exhibit superior anti-fibrotic effects compared with UC-MSCs.
• •IMRC-driven hyaline-like cartilage repair is partially associated with MMP1-dependent matrix remodeling.
• •OA cartilage organoids (OCOs) provide a promising platform for evaluating the efficacy of cell-based therapies.

1. Introduction

Osteoarthritis (OA) is a prevalent joint disease characterized by articular cartilage degeneration, subchondral bone remodeling, and synovial inflammation. With population aging and rising obesity rates, OA affects more than 600 million individuals globally and is projected to increase substantially in the coming decades [1]. The intractable nature of OA and need for chronic management impose a substantial burden on both individuals and societies. A key pathological feature of OA is irreversible cartilage degeneration, driven by the limited regenerative capacity of avascular cartilage and the low proliferative potential of mature chondrocytes [2,3]. Current pharmacological treatments mainly focus on symptomatic relief, but have minimal effects on structural disease progression or hyaline-like cartilage regeneration [4,5]. For patients with advanced OA, joint replacement remains the most effective option for restoring function, but it is highly invasive and associated with perioperative risks, tissue damage, and prolonged rehabilitation [6].

Recently, cell-based therapies have attracted considerable interest, particularly intra-articular injections of mesenchymal stem cells (MSCs). Preclinical and clinical investigations have reported improvements in pain and joint function, and sometimes partial alleviation of cartilage degeneration following MSC injection in OA joints [[7], [8], [9], [10]]. However, their clinical outcomes have been inconsistent due to heterogeneous cell populations, variable chondrogenic potential, and limited long-term cartilage restoration in the inflammatory OA environment. MSC-derived products such as small extracellular vesicles (sEVs) and mitochondria have also been explored for OA treatment, but their therapeutic efficacy remains variable and requires further validation [[11], [12], [13]]. In addition, MSC-mediated repair may result in fibrocartilaginous tissue rather than durable hyaline-like cartilage, with uncertain long-term efficacy and safety [[14], [15], [16], [17], [18]]. These factors restrict the clinical application of MSC-based therapies and underscore the importance of identifying alternative cell sources with more defined functional properties and enhanced capacity to regulate joint inflammation and cartilage matrix homeostasis.

Immunity-and-matrix-regulatory cells (IMRCs), derived from embryonic stem cells (ESCs), represent a novel cell population with enhanced immunomodulatory and matrix-regulatory functions compared with primary tissue-derived MSCs [19]. Through the secretion of diverse bioactive factors, IMRCs can more effectively modulate inflammatory signaling, regulate extracellular matrix (ECM) turnover, and improve the tissue microenvironment to favor repair. To date, IMRCs have demonstrated notable therapeutic potential in several preclinical disease models, including lung injury, cognitive impairment, membranous nephropathy (MN), intrauterine adhesion (IUA), and meniscus injury [[20], [21], [22], [23]]. As a novel cell source, IMRCs may offer advantages for joint repair in OA. Their immunoregulatory capacity may enable more effective suppression of chronic inflammation in the OA joint, while their matrix-regulatory properties may help restore the balance between cartilage matrix synthesis and degradation. Compared with MSC-based approaches, these features may enable IMRCs to provide a more robust and controllable strategy for modulating the joint microenvironment and promoting stable hyaline-like cartilage regeneration, thereby offering new promise for cell-based therapies targeting OA progression.

In recent years, organoid technology has emerged as a powerful platform for regenerative medicine and disease modeling, enabling more controlled in vitro systems for studying disease progression and therapeutic responses [[24], [25], [26]]. In the context of studying articular cartilage, organoids better preserve 3D cartilage architecture and microenvironmental features compared with conventional 2D chondrocytes, allowing for their wide application in cartilage regeneration and relative mechanistic studies [[27], [28], [29]]. By constructing appropriate 3D cartilage organoids, the key pathological features of OA cartilage can be recapitulated in a controlled in vitro setting, such as matrix degradation, chondrocyte hypertrophy, fibrotic remodeling, and inflammatory responses, hence facilitating detailed investigation of cell-cell interactions and tissue-level responses during disease progression [[30], [31], [32]]. The development of a robust OA cartilage organoid system enables a unique set of features compared to 2D cell cultures, ex vivo tissue samples, or in vivo models for the systematic assessment of regenerative interventions including cell-based treatments. Importantly, they offer a disease-relevant and controllable in vitro system for mechanistic investigations of how candidate cell populations influence cartilage homeostasis and inflammatory signaling.

This study aims to systematically evaluate the therapeutic potential of IMRCs as a novel cell source for hyaline-like cartilage repair in OA, compared with umbilical cord-derived mesenchymal stem cells (UC-MSCs, hereafter referred to as MSCs). We first characterized their key functional properties relevant to joint repair, including resistance to oxidative stress, migration and colonization abilities, and chondrogenic potential. Their effects on OA cartilage were then assessed using a multi-level experimental platform, from 2D chondrocyte cultures to a newly established 3D OA cartilage organoid (OCO) model, followed by rabbit cartilage explants and an anterior cruciate ligament transection (ACLT)-induced rat OA model (Fig. 1). By integrating biomimetic in vitro, ex vivo, and in vivo systems, this multilevel framework enables a comprehensive assessment of IMRC-mediated hyaline-like cartilage repair, as well as mechanistic investigations into their immunomodulatory and matrix-regulatory functions. Moreover, we showed that the OCO model we established offers an in vitro platform for evaluating cell-tissue interactions during cartilage repair under OA-associated pathophysiological conditions. Together, this study demonstrates a preclinical evaluation framework for IMRCs and highlights their potential as an alternative therapeutic cell source for OA cartilage repair.

Fig. 1.

Schematic overview of the experimental design. Human ESC-derived IMRCs were systematically evaluated for OA treatment using a multi-level platform comprising ATDC5 cells, OCOs, cartilage explants, and an ACLT-induced rat OA model. Compared with MSCs, IMRCs exhibited superior regenerative properties, promoting anabolic cartilage remodeling while suppressing inflammation, angiogenesis, hypertrophy, and fibrosis. Mechanistically, IMRC-derived MMP1 contributed to ECM remodeling by reducing COL-I deposition and increasing the COL-II/COL-I ratio, thereby facilitating hyaline-like cartilage repair. Created with BioRender.com.

2. Results

2.1. Characterization of the key properties of IMRCs for cartilage repair

Analysis of cell surface markers by flow cytometry indicated a mesenchymal profile [33] in IMRCs, which was similar to that of MSCs (CD45⁻/CD73⁺/CD90⁺/CD105+) (Fig. S1). We first evaluated the key cellular properties of IMRCs compared to conventional, human umbilical cord-derived MSCs that might be relevant for OA cartilage repair. Under H₂O₂-induced oxidative stress, IMRCs exhibited superior resistance compared to MSCs, evidenced by reduced senescence as indicated through senescence-associated β-galactosidase (SA-β-Gal) staining (Fig. 2A) and higher cell viability in cell counting kit-8 (CCK-8) assays (Fig. 2B). RT-qPCR analysis also showed higher expression of LMNB1 and lower expression of P21 and P16 in IMRCs than MSCs (Fig. 2C), reflective of enhanced resistance and survival under OA-relevant oxidative stress conditions.

Fig. 2.

Preliminary comparative evaluation of IMRCs and MSCs relevant to OA cartilage repair. (A) SA-β-Gal staining following H₂O₂ treatment and quantification of senescent cells, n = 5 per group. (B) Cell viability measured by CCK-8 assay after exposure to oxidative stress, n = 5 per group. (C) RT-qPCR analysis of senescence-associated genes LMNB1, P21, and P16, n = 4 per group. (D) Schematic illustration of migration and colonization assays. (E) Representative images and quantification of Transwell migration, n = 5 per group. (F) Fluorescence images and quantification of cell attachment to cartilage defect surfaces, n = 5 per group. (G) RT-qPCR analysis of adhesion-related genes, n = 4 per group. (H) AB and SO-FG staining of cartilage pellets differentiated from both cell types. (I) Immunofluorescence (IF) staining of COL-II and COL-X in cartilage pellets. (J) Bern score quantification of cartilage pellets, n = 5 per group. (K) Quantitative analysis of COL-II and COL-X expression, n = 5 per group. (L) RT-qPCR assessment of cartilage metabolism-related genes in pellets, n = 4 per group. Data are presented as mean ± SD, ns: no significance, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Scale bar = 200 μm.

Next, we assessed the migration and colonization capacity of both cell types (Fig. 2D). In Transwell assays, IMRCs displayed superior migratory ability compared to MSCs (Fig. 2E). Moreover, cartilage defect colonization assays using Dil-labeled cells revealed significantly higher colonization efficiency of IMRCs (Fig. 2F), which was potentially related to their elevated expression of CXCL12 (Fig. 2G) [[34], [35], [36]]. We also compared the chondrogenic potential of IMRCs and MSCs by differentiating them into 3D chondrogenic cartilage pellets. Alcian Blue (AB) and Safranin O-Fast Green (SO-FG) staining indicated more robust glycosaminoglycan (GAG) deposition in IMRC-derived pellets (Fig. 2H), which was accompanied by higher COL-II and lower COL-X protein expression (Fig. 2I), suggesting greater resemblance of hyaline-like cartilage. These observations were consistent with quantifications of Bern score (Fig. 2J) [37] and fluorescence area of COL-II and COL-X (Fig. 2K). RT-qPCR analysis of cartilage metabolism-related genes revealed higher expression of anabolic genes (SOX9, COL2A1, ACAN) and reduced expression of hypertrophic genes (IHH, COL10A1) in IMRC-derived cartilage pellets (Fig. 2L). Collectively, compared with MSCs, IMRCs displayed greater resistance to oxidative stress, enhanced migration and colonization capacity, and better potential to differentiate into tissue resembling hyaline-like cartilage. These characteristics point to the possibility that IMRCs may achieve improved therapeutic outcomes over MSCs when used for OA cartilage repair.

2.2. Effects of IMRCs on OA chondrocytes in 2D culture

To understand the potential therapeutic effects of IMRCs on OA cartilage repair, we tested them in a range of models replicating the key features of OA disease. The first and simplest model consisted of OA-like chondrocytes in 2D culture, whereby ATDC5 cells were subjected to inflammatory stimulation (INFL) followed by co-culture with MSCs or IMRCs in a Transwell system (Fig. 3A). Flow cytometry (FCM) analysis demonstrated markedly elevated apoptosis in ATDC5 cells after inflammatory stimulation, which was greatly reduced following co-culture with both MSCs and IMRCs (Fig. 3B). There were no significant differences between the MSC and IMRC groups, suggesting that these two cell types had similar ability to rescue OA chondrocytes from apoptosis under inflammatory conditions. RT-qPCR analysis of ATDC5 cells following inflammatory stimulation (Fig. 3C) indicated significant downregulation of chondrogenic genes (Sox9 and Col2a1), as well as upregulation of genes associated with matrix degradation and hypertrophy (Mmp13 and Col10a1) or OA-related inflammation (Il1b and Il6). Groups subsequently co-cultured with MSCs or IMRCs showed non-significant changes in Sox9 and Col2a1 but greatly attenuated expression of Col10a1 and Il1b, whereby the two cell types exerted similar effects. Meanwhile, compared to inflammatory ATDC5 cells co-cultured with MSCs, where the expression levels of Mmp13 and Il6 did not change significantly, those co-cultured with IMRCs showed markedly reduced expression of both genes approaching normal cells not subjected to inflammatory stimulation. Additionally, GAG deposition in inflammatory ATDC5 cells was markedly reduced compared to the normal unstimulated group (Fig. 3D). Interestingly, while co-culture with IMRCs partly restored GAG content, co-culture with MSCs showed no improvements or even lower staining intensity than the control inflammatory group. These results collectively suggest that in the 2D-cultured OA chondrocyte model, IMRCs can exert better chondroprotective effects than MSCs, including by promoting cartilage matrix synthesis, inhibiting hypertrophy, and providing anti-inflammatory effects.

Fig. 3.

Comparative effects of IMRCs and MSCs on 2D-cultured OA chondrocytes. (A) Schematic illustration of the experimental design showing that ATDC5 cells were subjected to INFL for 48 h and subsequently co-cultured with MSCs or IMRCs for an additional 48 h. (B) Apoptosis of ATDC5 cells assessed by Annexin V-FITC/PI FCM, with quantitative analysis of apoptotic cells (Q2 + Q3), n = 3 per group. (C) RT-qPCR analysis of genes associated with chondrogenesis (Sox9, Col2a1), matrix degradation and hypertrophy (Mmp13, Col10a1), and pro-inflammatory cytokines (Il1b, Il6) in ATDC5 cells under different treatment conditions, n = 4 per group. (D) AB and SO staining showing GAG deposition in ATDC5 cells under different treatment conditions, with quantitative analysis of relative staining intensity, n = 4 per group. Data are presented as mean ± SD, ns: no significance, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Scale bar = 1 mm.

However, a 2D chondrocyte model is fundamentally limited by non-physiological cell growth and possible loss of phenotype. This model does not adequately recapitulate the complex cell-ECM interactions present in native cartilage tissue that exist in 3D, particularly under the complex pathological conditions of OA. The limitations of the 2D system made it difficult to discern the therapeutic effects of IMRCs compared to MSC on short-term matrix accumulation, and more comprehensive histological evaluation was not feasible. These observations highlight the necessity of establishing a physiologically and pathologically relevant 3D cartilage model to more comprehensively assess possible therapeutic differences between the two stem cell types.

2.3. Construction and validation of 3D OCOs

We developed an optimized 3D OCO model based on previously reported protocols [30,[38], [39], [40], [41]]. Briefly, human MSCs were used as the source cells to first generate hyaline-like cartilage organoids (HCOs) through chondrogenic induction for 14 days, followed by an additional 7 days of hypertrophic and inflammatory stimulation to obtain OCOs (Fig. 4A). Macroscopic observation of organoid morphology showed that HCOs at 14 days and OCOs at 21 days both had relatively regular spherical shapes. However, OCOs had significantly smaller diameter than HCOs (Fig. 4B), which could be a result of increased matrix catabolism under inflammatory conditions.

Fig. 4.

OCOs recapitulated key features of OA cartilage. (A) Schematic illustration of the induction strategy for constructing OCOs. (B) Representative macroscopic morphology of HCOs (d14) and OCOs (d21) with quantitative analysis of organoid diameter, n = 8 per group. (C) Histological evaluation of organoids by AB and SO-FG staining with Bern score quantification, n = 5 per group. (D) Representative IF images showing COL-II expression in cartilage organoids. (E) IHC staining of COL-X, MMP13, and COL-I in organoid sections. (F) Quantitative analysis of COL-II, COL-X, MMP13, and COL-I expression from IHC, n = 5 per group. (G) ELISA analysis of IL-6, MMP13, and VEGF levels in organoid culture supernatants, n = 4 per group. (H) Western blot analysis of proteins involved in OA-related signaling pathways from cartilage organoids. (I) RT-qPCR analysis of cartilage metabolism-related genes, n = 4 per group. (J) RT-qPCR analysis of osteogenesis-related genes, n = 4 per group. (K) RT-qPCR analysis of inflammatory cytokines and angiogenic factors, n = 4 per group. Data are presented as mean ± SD, ns: no significance, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Scale bar = 200 μm.

To understand the matrix composition and structure of OCOs and their resemblance to OA cartilage, we performed comprehensive histological assessments. AB and SO-FG staining showed that HCOs were characterized by abundant GAG deposition, well-organized matrix architecture, and clearly defined lacunar structures (Fig. 4C). Meanwhile, OCOs exhibited markedly reduced GAG content and disorganized matrix arrangement, accompanied by a significant decrease in Bern scores compared to HCOs, indicative of OA-like cartilage degeneration. In addition, OCOs showed a significant reduction in COL-II immunofluorescence staining compared to HCOs (Fig. 4D), as well as greatly increased COL-X, MMP13, and COL-I expression as indicated by immunohistochemical (IHC) staining (Fig. 4E–F). These results suggest that OCOs can recapitulate some of the key histological features of OA cartilage, including a loss of ECM components that define hyaline-like cartilage alongside increases in hypertrophic and fibrotic characteristics.

In addition to their structural features, the molecular expression of OA-related markers was characterized in OCOs. ELISA results showed that OCOs secreted significantly higher levels of IL-6, MMP13, and VEGF compared to HCOs, respectively suggesting increased inflammatory activity, matrix degradation, and pro-angiogenic signaling under OA-like conditions (Fig. 4G). As aberrant angiogenesis is a pathological feature of OA, we tested the responses of human umbilical vein endothelial cells (HUVECs) to the conditioned medium (CM) derived from HCOs and OCOs, and found that HUVECs cultured in OCO-CM exhibited increased tube formation ability compared to those cultured in HCO-CM (Fig. S2). In addition, Western blot analysis revealed the activation of OA-related signaling pathways in OCOs, including the NF-κB pathway (indicated by enhanced phosphorylation of P65) and SMAD1/5/9-dependent BMP/TGF-β signaling (indicated by enhanced phosphorylation of SMAD1/5/9) (Fig. 4H). The activation of these molecular pathways is reflective of inflammation and terminal chondrocyte differentiation characteristically observed in OA [[42], [43], [44], [45]]. At the transcriptional level, OCOs exhibited significant downregulation of anabolic genes (SOX9, COL2A1, COMP), accompanied by upregulation of genes associated with cartilage hypertrophy or degradation (IHH, COL10A1, MMP13) and osteogenic differentiation (RUNX2, BMP2, ALP, OCN, OPN, IBSP) (Fig. 4I–J). In parallel, OCOs also showed greatly elevated expression of genes encoding inflammatory cytokines (IL1B, IL6, IFNG) and angiogenic factors (VEGF, HIF1A, FGF2) compared to HCOs (Fig. 4K). Collectively, these molecular-level alterations complement the structural observations, confirming that OCOs can recapitulate the key pathological features of OA cartilage including reduced anabolic activity, enhanced matrix degradation, terminal chondrocyte differentiation, inflammatory activation, and increased angiogenic propensity. Hence, the OCOs we developed in this study represent a biomimetic platform and useful tool for studying OA pathophysiology and evaluating potential therapeutics.

2.4. Short-term effects of IMRCs on 3D OCOs

To evaluate the short-term effects of stem cell treatment on a 3D model of OA cartilage, IMRCs or MSCs were co-cultured with OCOs for 7 days, with a non-treated group included as a control (Fig. 5A). Following co-culture, macroscopic observation showed no obvious differences in organoid size among the three groups (Fig. 5B). Histological analysis indicated higher GAG deposition and COL-II protein expression in the IMRC-treated group compared to the MSC-treated and non-treated groups (Fig. 5C–D), suggesting greater effectiveness of IMRCs in promoting ECM synthesis resembling hyaline-like cartilage. IHC analysis showed an insignificant difference in COL-X expression among the three groups, while interestingly, MMP13 and COL-I were markedly elevated only in the MSC-treated group but similar to the control in the IMRC-treated group (Fig. 5E–F). These results pointed to an interesting phenomenon whereby short-term stem cell intervention tended to induce a certain degree of cartilage hypertrophy, but the hypertrophic signal was more pronounced in the MSC-treated group and accompanied by an obvious fibrotic tendency, while this adverse effect was largely absent in the IMRC-treated group.

Fig. 5.

Short-term modulatory effects of IMRCs and MSCs on OCOs. (A) Schematic illustration of short-term (7-day) co-culture of IMRCs or MSCs with OCOs. (B) Macroscopic morphology of OCOs after co-culture and quantitative analysis of organoid diameter. (C) SO-FG staining of organoid sections. (D) IF staining of COL-II in organoid sections. (E) IHC staining of COL-X, MMP13, and COL-I in organoid sections. (F) Quantitative analysis of COL-II, COL-X, MMP13, and COL-I protein expression. (G) RT-qPCR analysis of inflammatory gene expression. (H) RT-qPCR analysis of angiogenic gene expression. (I) RT-qPCR analysis of chondrogenic gene expression. (J) RT-qPCR analysis of hypertrophic gene expression. (K) RT-qPCR analysis of osteogenic gene expression. (L) ELISA analysis of IL-6, MMP13, and VEGF secretion by organoids. (M) Western blot analysis of proteins involved in OA-related signaling pathways from organoids. Data are presented as mean ± SD, ns: no significance, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, n = 4 per group, except for panel B (n = 8). Scale bar = 200 μm.

The short-term effects of stem cell treatment on OCOs were further characterized by molecular-level analyses. RT-qPCR results showed that both cell interventions significantly reduced gene expression of pro-inflammatory cytokines (IL1B, IL6, IFNG) and angiogenic factors (VEGF, HIF1A, FGF2) in OCOs, with more pronounced reductions observed in the IMRC-treated group (Fig. 5G–H). For chondrogenic genes, both cell-treated groups showed markedly upregulated expression of SOX9, COL2A1, and ACAN compared to the control, while the IMRC-treated group also achieved much higher expression of COL2A1 than the MSC-treated group (Fig. 5I). While both cell-treated groups also showed increased expression of genes associated with hypertrophy (IHH, COL10A1, MMP13) and osteogenesis (RUNX2, COL1A1), the overall expression levels in OCOs treated with IMRCs were much lower than those treated with MSCs (Fig. 5J–K). ELISA analyses for the secretion levels of IL-6, MMP13, and VEGF between groups followed the same trends as the RT-qPCR results (Fig. 5L). Western blot analysis further demonstrated that IMRCs markedly suppressed the activation of NF-κB and SMAD-1/5/9-mediated BMP/TGF-β signaling in OCOs, whereas MSC treatment did not show a comparable inhibitory effect on these OA-related pathways (Fig. 5M).

Collectively, the results of short-term (7 days) co-culture using a 3D model of OCOs suggested several regulatory advantages of IMRCs compared to conventional MSCs in OA cartilage repair. First, IMRCs could more effectively suppress inflammatory responses and angiogenesis-related signaling while promoting the synthesis of hyaline-like cartilage matrix. Second, unlike MSCs, IMRCs did not significantly induce adverse activation of hypertrophic and fibrotic signaling, which could prevent the premature transition of chondrocytes towards terminal differentiation. These findings suggest a more balanced paracrine regulatory profile of IMRCs than MSCs that could be helpful in providing a favorable microenvironment for OA cartilage repair. However, short-term observations do not capture the sustained impact of IMRCs on OCOs, such as long-term regulatory capacity and tissue remodeling effects, which would be relevant for their clinical application. Hence, it remains necessary to characterize the in vitro effects of IMRCs over a prolonged culture period.

2.5. Long-term effects of IMRCs on 3D OCOs

The long-term modulatory effects of IMRCs and MSCs on OCOs were evaluated by co-culturing for 21 days (Fig. 6A). After this prolonged in vitro co-culture, RT-qPCR analysis indicated that both cell types reduced gene expression associated with hypertrophy or matrix degradation (IHH, COL10A1, MMP13) and osteogenesis (RUNX2, BMP2, ALP, OSX, OPN) in OCOs, with more pronounced reductions seen in the IMRC-treated group (Fig. 6B–C). Remarkably, significant upregulation of chondrogenic genes (SOX9, COL2A1) was only noted in the IMRC-treated group, which reached levels matching or exceeding the control group, while expression levels in the MSC-treated group were considerably lower than the control. Meanwhile, COL1A1 expression was upregulated in OCOs treated with both cell types, suggesting an unavoidable tendency towards fibrotic repair. However, the level of COL1A1 upregulation in the IMRC-treated group was relatively mild compared to the control, while that in the MSC-treated group was dramatically elevated and far exceeded the other two groups. These results may reflect a better ability of IMRCs to protect against a fibrotic repair response in an OA-like environment compared to MSCs. For genes associated with inflammation (IL1B, IL6, TNFA) or angiogenesis (VEGF, HIF1A, FGF2), IMRCs and MSCs generally had similar capacity to downregulate their expression compared to the control group, with the IMRCs showing more pronounced downregulation for selected genes such as IL6 and FGF2 (Fig. 6D). Western blot analysis further indicated that after prolonged co-culture for 21 days, both cell types exerted similar inhibitory effects on NF-κB signaling in OCOs, with IMRCs showing more effective suppression of SMAD1/5/9-dependent TGF-β signaling (Fig. 6E).

Fig. 6.

Long-term modulatory effects of IMRCs and MSCs on OCOs. (A) Schematic illustration of long-term (21-day) co-culture of IMRCs or MSCs with OCOs. (B) RT-qPCR analysis of cartilage metabolism-related genes. (C) RT-qPCR analysis of bone metabolism-related genes. (D) RT-qPCR analysis of inflammatory and angiogenic genes. (E) Western blot analysis of proteins involved in OA-related signaling pathways. (F) Representative images of SO-FG staining, as well as IF and IHC staining for COL-II, COL-X, MMP13, and COL-I expression in OCOs after 21 days. (G) GO enrichment analysis of OCOs after 21 days of co-culture with MSCs or IMRCs. Data are presented as mean ± SD, ns: no significance, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, n = 4 per group. Scale bar = 200 μm.

Histological analysis at 21 days indicated markedly enhanced cartilage-like matrix structure in OCOs treated with either IMRCs or MSCs compared to the control, as seen through SO-FG staining, while COL-II immunofluorescence staining was also greatly elevated in both cell-treated groups but strongest in the IMRC-treated group (Fig. 6F, Fig. S3). Immunohistochemical staining for COL-X showed similar staining area among groups, while for MMP13 and COL-I, only the IMRC-treated group showed significant reduction in staining area compared to the control, with the MSC-treated group having minimal effects or even slightly exceeding the control group. These staining results were consistent with RT-qPCR data, suggesting an improved ability of IMRCs to induce repair in OA cartilage with a hyaline-like cartilage phenotype while suppressing fibrosis compared to MSCs.

In parallel, transcriptomic sequencing of OCOs treated with IMRCs compared to MSCs at 21 days detected a total of 128 differentially expressed genes (DEGs), including 61 upregulated and 67 downregulated genes (Fig. S4). GO analysis revealed that these genes were predominantly enriched in ECM regulation, microtubule-related processes, receptor-ligand interactions, and immune-related pathways (Fig. 6G). These findings suggest that IMRCs may contribute to cartilage repair through multiple mechanisms such as remodeling the ECM, enhancing secretory function through the regulation of cytoskeletal and intracellular trafficking, and reshaping the inflammatory microenvironment through receptor-ligand signaling.

2.6. Comparative secretome analysis of IMRCs and MSCs in ECM regulation

Since transcriptomic sequencing of OCOs treated with IMRCs and MSCs suggested that the secretory function of stem cells could be a key mechanism in modulating cartilage repair, we compared the secretome profiles of IMRCs and MSCs by liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis of CM collected from both cell types after 24 h (Fig. 7A). Differential protein analysis identified a total of 29 proteins with significantly altered secretion in IMRCs compared to MSCs, of which 21 were upregulated and 8 were downregulated (Fig. 7B–C). Notably, matrix metalloproteinase-1 (MMP1) was markedly upregulated in IMRCs as a key regulator of ECM metabolism [46], highlighting enhanced matrix-regulatory function of the IMRC secretome. Protein-protein interaction (PPI) analysis revealed a tightly connected network indicative of strong functional associations among the differentially secreted proteins (Fig. 7D). Notably, multiple collagens including COL3A1, COL5A1, COL4A2, COL6A1, and COL6A2 occupied central nodes in this network, highlighting significant differences between IMRCs and MSCs in the secretion patterns of collagens and hence their potential to regulate collagen metabolism. Consistent with this observation, gene set enrichment analysis (GSEA) showed that the IMRC secretome was significantly enriched for the collagen-containing extracellular matrix (GO:0062023) gene set (Fig. 7E, Fig. S5). These findings collectively suggest a unique secretory signature of IMRCs that is involved in ECM regulation, which may in turn contribute to regulating cartilage matrix homeostasis and improving repair outcomes.

Fig. 7.

IMRCs exhibit an ECM-remodeling-enriched secretome compared to MSCs. (A) Schematic illustration of the experimental workflow for secretome analysis, where CM was collected from IMRCs and MSCs after 24 h culture in serum-free medium. (B) Volcano plot showing differentially secreted proteins between IMRCs and MSCs. (C) Heatmap showing the distinct protein expression profiles between IMRCs and MSCs. (D) PPI network of differentially secreted proteins generated using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database. (E) GSEA of the secretome revealed significant enrichment of collagen-containing extracellular matrix (GO:0062023) in IMRCs compared to MSCs.

2.7. IMRC-mediated cartilage repair is partially associated with MMP1-dependent matrix remodeling

We proceeded to further investigate the role of MMP1 in mediating the effects of IMRCs on OCOs, informed by the results of secretome analysis whereby IMRCs were found to express significantly higher levels of MMP1 than MSCs. Moreover, previous studies have shown that IMRC-derived MMP1 exerts significant therapeutic effects in pulmonary fibrosis [19,22]. Mechanistically, these therapeutic effects were associated with the ability of MMP1 to efficiently degrade COL-I and remodel the ECM, thereby attenuating fibrosis. Hence, MMP1 may serve as a contributing factor through which IMRCs regulate tissue matrix homeostasis. Considering that abnormal accumulation of COL-I is a hallmark of fibrosis and chondrocyte dedifferentiation during OA cartilage degeneration, we hypothesized that IMRCs may act partly through high expression of MMP1, which degrades COL-I and improves cartilage matrix metabolism, thereby delaying cartilage degeneration. To test this hypothesis, we performed MMP1-related intervention experiments in organoid models and evaluated outcomes associated with maintaining the hyaline-like cartilage phenotype using molecular and histological analyses.

First, RT-qPCR analysis revealed that MMP1 gene expression in IMRCs was elevated by approximately 600-fold compared to MSCs (Fig. 8A). Considering that multiple members of the MMP family have been recognized as markers of OA and are closely associated with cartilage degeneration, we first evaluated the potential toxicity of MMP1 on healthy cartilage by directly adding it to the HCO model (Fig. 8B). SO-FG staining showed no obvious change in staining intensity after MMP1 treatment (Fig. 8C). RT-qPCR analysis indicated generally beneficial effects of MMP1 on cartilage metabolism in HCOs, including dramatically upregulated expression of COL2A1 and ACAN accompanied by downregulation of the hypertrophic and fibrotic markers IHH, COL10A1, MMP13, RUNX2, and COL1A1 (Fig. 8D). Meanwhile, similar effects were observed in ATDC5 cells following MMP1 treatment (Fig. S6). These results suggest that MMP1 is safe for healthy cartilage and may even be beneficial for maintaining normal cartilage ECM composition and metabolism.

Fig. 8.

IMRC-mediated cartilage protection is partially associated with MMP1-dependent matrix remodeling in OA. (A) RT-qPCR analysis of MMP1 expression in MSCs and IMRCs. (B) Schematic illustration of MMP1 treatment (10 ng/mL) in HCOs for 7 days. (C) Representative SO-FG staining images of HCOs after MMP1 treatment. (D) RT-qPCR analysis of cartilage metabolism-related genes in HCOs, together with COL2A1/COL1A1 ratio. (E) Schematic illustration of MMP1 treatment (10 ng/mL) in OCOs for 7 days. (F) Representative images of SO-FG staining, COL-II IF staining, and COL-I IHC staining in OCOs after MMP1 treatment. (G) RT-qPCR analysis of cartilage metabolism-related genes in OCOs, together with COL2A1/COL1A1 ratio. (H) Schematic illustration of co-culturing IMRCs after siRNA-mediated MMP1 silencing with OCOs for 7 days. (I) RT-qPCR analysis confirming MMP1 knockdown efficiency in IMRCs at 48 h after siRNA transfection. (J) Western blot analysis of MMP1 protein expression in IMRCs at 3 and 7 days after siRNA-mediated silencing. (K) RT-qPCR analysis of cartilage metabolism-related genes in OCOs after co-culture with siMMP1-treated IMRCs, together with COL2A1/COL1A1 ratio. (L) Representative SO-FG staining and IHC/IF images showing COL-II, COL-I, MMP13, and COL-X expression in OCOs after co-culture with siMMP1-treated IMRCs. (M) Quantitative analysis of COL-II and COL-I staining in OCOs. Data are presented as mean ± SD, ns: no significance, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, n = 4 per group. Scale bar = 200 μm.

Next, we evaluated the effects of MMP1 on OA-like cartilage by directly adding it to OCOs (Fig. 8E). Similar to HCOs, histological analysis of OCOs treated with MMP1 showed no significant changes in SO-FG staining intensity compared to the control group (Fig. 8F). However, there was marked elevation of COL-II protein levels and marked reduction of COL-I in the MMP1-treated OCOs, suggestive of decreased fibrosis and preservation of hyaline-like cartilage characteristics. RT-qPCR analysis further confirmed overall improvements in the expression of cartilage metabolism-related genes following MMP1 treatment, particularly noting significant changes in SOX9, COL10A1, MMP13, and COL1A1 as well as a dramatic increase in COL2A1 expression and the COL2A1/COL1A1 ratio (Fig. 8G). Western blot analysis of MMP1-treated OCOs showed significantly reduced SMAD1/5/9 phosphorylation, whereas no obvious effect on p65 phosphorylation was detected (Fig. S7), indicating that MMP1 may partially mediate the regulation of SMAD1/5/9 signaling by IMRCs and contribute to their ECM-remodeling effects. These findings support a beneficial regulatory effect of MMP1 on cartilage degeneration and ECM composition in an OA-like environment.

To validate the mechanistic contribution of IMRC-derived MMP1 in OA cartilage repair, we silenced MMP1 expression in IMRCs before co-culturing them with OCOs (Fig. 8H). After siRNA-mediated silencing, MMP1 expression in IMRCs was effectively suppressed at both the mRNA and protein levels (Fig. 8I–J). When siMMP1-IMRCs were co-cultured with OCOs, the previously observed beneficial effects of IMRCs were abolished, evidenced by RT-qPCR analysis showing overall deterioration of cartilage metabolism-related gene expression, particularly in SOX9, COL2A1, IHH, COL10A1, and MMP13 accompanied by significant reduction in the COL2A1/COL1A1 ratio (Fig. 8K). Similar results were shown by staining images for SO-FG and the expression of COL-II and COL-I proteins (Fig. 8L–M). Taken together, these findings suggest that MMP1 in IMRCs may contribute, at least in part, to their observed effects on OA-like cartilage, including suppression of cartilage fibrosis and maintenance of a hyaline-like cartilage phenotype.

2.8. Therapeutic effects of IMRCs in ex vivo rabbit OA cartilage explants

While our established cartilage organoid models recapitulate key features of healthy and OA-like human cartilage, we acknowledge that they may not substitute for native tissues and that the above-observed effects of IMRCs would benefit from systematic verification using ex vivo tissue samples. Here, rabbit cartilage explants were used as the ex vivo model, where an OA-like state was induced by treatment with an inflammatory cocktail (INFL) comprising IL-1β, IL-6, and TNF-α (Fig. 9A). Following co-culture of IMRCs or MSCs with OA-like cartilage explants for 10 days, their therapeutic effects were assessed by histological and immunohistochemical analyses (Fig. 9B–D). Notably, across all stains, the IMRC group exhibited the closest resemblance to the appearance of normal cartilage explants not subjected to inflammatory stimulation, while the MSC group largely showed qualitative and quantitative staining results similar to the blank group (subjected to inflammatory stimulation and left untreated). In particular, the IMRC-treated group showed significant elevation in Saf-O positive area (Fig. 9C) and COL-II positive area (Fig. 9D) compared to the MSC-treated group, suggesting an improved ability to maintain a cartilage matrix with hyaline-like characteristics that is rich in GAG and COL-II. Collectively, the results obtained using an ex vivo OA cartilage model verify the therapeutic effects of IMRCs seen in 2D chondrocytes and 3D cartilage organoid models, demonstrating that IMRCs can outperform MSCs in an OA-like environment by promoting cartilage matrix preservation and regeneration.

Fig. 9.

IMRCs promoted matrix preservation and regeneration in rabbit OA cartilage explants. (A) Schematic of experimental design. (B) Representative images of histological staining of rabbit OA cartilage explants following different treatments, including AB and SO-FG, as well as IHC for COL-II, COL-X, and MMP13. (C) Quantitative analysis of SO-FG staining. (D) Quantification of IHC staining for COL-II, COL-X, and MMP13. Data are presented as mean ± SD, ns: no significance, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, n = 4 per group. Scale bar = 200 μm.

2.9. Therapeutic effects of IMRCs in an in vivo rat ACLT-induced OA model

The final step in our systematic evaluation of IMRC therapeutic efficacy involved in vivo cartilage repair in a rat ACLT-induced OA model, whereby intra-articular injections of PBS, MSCs, or IMRCs were administered every two weeks from the second week after ACLT followed by final evaluation at ten weeks post-surgery (Fig. 10A). To assess the retention of transplanted cells in the OA joint, DiR-labeled MSCs and IMRCs were subjected to in vivo fluorescence imaging 14 days after intra-articular injection. Quantitative analysis revealed significantly higher fluorescence intensity in the IMRC group than in the MSC group, indicating greater retention of fluorescent signal at the injection site (Fig. S8). At the study endpoint, OARSI and Mankin histological scoring both indicated attenuation of cartilage degeneration in the medial tibial plateau of animals in the cell-treated groups compared to the PBS control, with the IMRC group exhibiting a more pronounced protective effect compared to the MSC group (Fig. 10B). HE staining of knee joint sections and quantitative analysis of the hyaline cartilage to calcified cartilage (HC/CC) ratio indicated significant structural damage to cartilage in the PBS group, characterized by disorganized chondrocyte arrangement and impaired tissue layering, while these structural alterations were partially improved in the MSC group and largely eliminated in the IMRC group (Fig. 10C).

Fig. 10.

IMRC treatment attenuated OA-associated cartilage degeneration in a rat ACLT-induced model. (A) Schematic of experimental design. (B) OARSI and Mankin scores of knee joint sections from the three groups at 10 weeks post-ACLT. (C) Representative HE staining of articular cartilage from the three groups, with quantitative analysis of the hyaline cartilage (HC)/calcified cartilage (CC) ratio. (D) Representative AB and SO-FG staining images of articular cartilage from the three groups. (E) Representative IHC staining images of articular cartilage for COL-II, MMP13, COL-X, and COL-I. (F) Quantitative analysis of IHC staining results. Data are presented as mean ± SD, ns: no significance, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, n = 4 per group. Scale bar = 200 μm.

Consistent with these observations, AB and SO-FG staining demonstrated superior matrix preservation and structural resemblance to normal hyaline-like cartilage in the IMRC group, with rich GAG content and a highly organized ECM appearance (Fig. 10D). Meanwhile, the PBS group showed extensive structural damage in the articular cartilage accompanied by GAG loss, which was only partly alleviated in the MSC group. Immunohistochemical staining indicated that both MSC and IMRC treatments promoted COL-II deposition and suppressed MMP13 expression compared to the PBS group, with more pronounced effects observed in the IMRC group (Fig. 10E–F). Meanwhile, only the IMRC group showed significant efficacy in suppressing the aberrant expression of COL-X and COL-I, while the MSC group had similar expression levels as the PBS control for these markers of cartilage hypertrophy and fibrosis. The expression levels of the pro-inflammatory cytokines IL-1β and TNF-α in synovial tissues were further evaluated (Fig. S9). Quantitative analysis demonstrated that both MSC and IMRC treatments reduced synovial inflammatory marker expression compared with the PBS group, with the lowest staining intensity observed in the IMRC group. These results suggest that IMRCs more effectively attenuate synovial inflammation within the OA joint microenvironment. Collectively, the repair outcomes obtained using an in vivo rat OA model demonstrate superior therapeutic efficacy of IMRCs compared to MSCs, including marked attenuation of OA-associated cartilage degeneration, preservation of a hyaline-like cartilage phenotype with reduced hypertrophic and fibrotic features, and alleviation of synovial inflammation.

3. Discussion

As an underexplored cell source for therapeutic intervention in OA, IMRCs hold potential for addressing some of the key drawbacks experienced with conventional UC-MSCs, presenting advantages such as cellular homogeneity and matrix-regulatory functions. However, progressing their practical application requires robust preclinical evidence that demonstrates the safety and therapeutic potential of IMRCs compared to UC-MSCs as a more commonly tested cell source. This study presents the first systematic, multi-level evaluation of IMRCs in OA cartilage repair across in vitro, ex vivo, and in vivo models. Uniquely, it also enhances the robustness and physiological relevance of the in vitro evidence through short- and long-term evaluations of IMRCs in 3D OCOs. The results collectively suggest significant advantages of IMRCs compared to UC-MSCs in promoting anabolic metabolism, maintaining a hyaline-like phenotype, and suppressing fibrosis in OA cartilage repair. Importantly, the multi-level preclinical evidence obtained through four types of models provide multi-angle back-to-back comparison of IMRCs with UC-MSCs in safety and efficacy, thereby comprehensively validating the therapeutic potential of IMRCs that could inform their future clinical use as an alternative form of OA intervention.

Unlike previous studies that mainly relied on molecular-level analyses or drug screening approaches [27,47,48], this study is the first to employ a cartilage organoid model as a key component in systematically evaluating the therapeutic effects of cell-based intervention for OA, providing physiologically relevant information at the level of 3D human tissues. As a 3D microenvironment, organoids recapitulate the major structural and functional features of cartilage, including cell-cell interactions, ECM composition, and long-term metabolic dynamics, enabling a more physiologically and pathologically relevant assessment of stem cell effects. Notably, this platform captures hallmark OA processes such as proteoglycan loss, matrix degradation, and fibrotic remodeling, which are not adequately represented in conventional 2D systems. In addition, organoids may partially substitute for animal models, reducing ethical concerns and the costs associated with in vivo studies [49,50]. By integrating molecular and histopathological analyses of 3D cartilage organoids, this study delineates the multi-dimensional effects of IMRCs on cartilage matrix homeostasis, fibrosis suppression, and ECM remodeling in an OA-like environment, while enabling parallel investigations of the underlying mechanisms. This strategy overcomes the key limitations of conventional 2D cultures and reductionist approaches, establishing a robust methodological framework for screening stem cell candidates with therapeutic potential in cartilage repair, and highlighting the unique value of organoids in tissue engineering and regenerative medicine research. Notably, although a scaffold-free organoid strategy was employed in this study, this platform could be further integrated in the future with emerging biomaterials, such as hydrogels, decellularized extracellular matrix (dECM) scaffolds, or smart drug delivery systems, to achieve synergistic integration of cells, biomaterials, and bioactive factors, thereby establishing a more complex and biomimetic OA microenvironment. Such an integrated strategy may more faithfully recapitulate the pathological and regenerative processes of cartilage tissue. Moreover, our organoid system may also serve as an advanced in vitro platform for evaluating biomaterial-based delivery systems under disease-relevant OA conditions. In particular, lesion-targeted biomaterial delivery strategies designed to modulate the local microenvironment, such as promoting the clearance of apoptotic chondrocytes and enhancing cartilage tissue activity [51], can be further investigated and optimized using this organoid platform. In summary, our organoid platform holds great promise for integration with biomaterials in future applications.

Furthermore, it should be noted that OA is a highly heterogeneous and complex disease that can exhibit distinct clinical phenotypes and molecular characteristics across different patients, disease stages, and pathogenic backgrounds [52]. Therefore, there is currently no single in vitro model capable of universally representing all pathological states of OA cartilage. The OCO model established in this study was not intended to fully recapitulate the entire complexity of clinical OA, but rather to specifically simulate key pathological states during OA progression. Notably, the OCO model in the present study demonstrated ECM degradation, chondrocyte hypertrophy, and upregulation of inflammatory response and angiogenesis, which are highly consistent with the core pathological features reported in publicly available human OA cartilage transcriptomic datasets (such as GSE114007 and GSE169077) and previous studies [[53], [54], [55]]. Therefore, we believe that this model reasonably reflects the OA-associated cartilage pathological microenvironment to a certain extent.

Members of the matrix metalloproteinase (MMP) family, such as MMP-1, -3, -9, and -13, are traditionally regarded as key mediators of cartilage degradation in OA [[56], [57], [58]]. In particular, MMP1 has been widely implicated in ECM breakdown and impaired tissue regeneration. However, our findings challenge this conventional paradigm. Our results suggest that IMRC-mediated cartilage repair is partially associated with MMP1-dependent matrix remodeling. This beneficial effect is mediated through the efficient degradation of COL-I, a principal component of fibrocartilaginous tissue, as previously reported [19]. Importantly, this degradative activity appears selective and does not noticeably impact key hyaline-like cartilage components such as COL-II and ACAN. Our findings suggest that IMRC-derived MMP1 may partly contribute to ECM remodeling associated with a hyaline-like cartilage phenotype rather than indiscriminate matrix degradation. This mechanistic insight provides a compelling explanation for the enhanced anti-fibrotic and regenerative capacity of IMRCs compared to UC-MSCs observed across our in vitro and in vivo models. In addition, MMP1 should be considered a contributing mechanism rather than the sole mediator of IMRC function. We speculate that IMRC-derived MMP1 may synergistically interact with other secreted factors to suppress inflammatory signaling, maintain cartilage matrix homeostasis, promote cartilage repair, and reduce fibrotic matrix deposition. Such coordinated regulation of extracellular matrix remodeling and inflammation may collectively contribute to the superior regenerative effects of IMRCs observed in this study. Nevertheless, the interplay between MMP1 and other IMRC-derived factors warrants further investigation.

Despite establishing key evidence on the modulatory potential of IMRCs across multiple dimensions in OA disease models, this study has several limitations that could influence the interpretation of results. First, the established OCO model predominantly comprises a single chondrocyte population and lacks other key joint cell types such as macrophages, synovial cells, and subchondral bone cells. Consequently, it does not fully recapitulate the complex cellular and biochemical microenvironment of the OA joint, particularly the inflammatory crosstalk and multi-cellular interactions that constitute key processes of disease progression. In addition, the absence of mechanical loading limits the ability of the model in mimicking the native biomechanical and physiological conditions of articular cartilage, which play essential roles in maintaining tissue homeostasis and regulating degeneration. Future studies should focus on developing more advanced multicellular and mechano-responsive cartilage organoid systems by incorporating synovial, immune, and subchondral bone components, as well as physiologically relevant mechanical stimulation. Such models would better recapitulate the in vivo joint microenvironment and further enhance the translational relevance of organoid-based platforms for OA research and therapeutic evaluation. Second, the MSCs used in this study were derived from human umbilical cord (UC-MSCs), which were selected due to their well-documented advantages, including high proliferative capacity, relatively low immunogenicity, and favorable chondrogenic potential. In addition, UC-MSCs have been extensively investigated in preclinical and clinical studies for OA, making them a representative and clinically relevant MSC source [[59], [60], [61], [62], [63], [64]]. However, it should be noted that MSCs derived from different tissue sources may exhibit distinct biological properties and therapeutic potentials. Therefore, although our results demonstrate that IMRCs exhibit superior regenerative and anti-fibrotic capacities compared with UC-MSCs, caution should be exercised when extrapolating these findings to MSCs from other sources. Future studies are warranted to systematically compare IMRCs with MSCs derived from multiple tissue origins to further validate the generalizability of the observed effects. Third, there are some limitations in the in vivo validation. The absence of a sham-operated control group limits comparison with normal cartilage, and the in vivo role of MMP1 was not directly validated. In addition, DiR fluorescence reflects local signal retention within the joint but does not definitively indicate long-term survival or viability of IMRCs. Further studies are warranted to validate IMRC efficacy in more comprehensive OA models, including larger animals and longer-term evaluations for assessing regenerative outcomes.

It is also worth noting that although stem cells such as IMRCs and UC-MSCs were observed in this study to promote cartilage regeneration, enhance COL-II expression, and stimulate anabolic metabolism, they also led to simultaneous upregulation of COL-X and COL-I following treatment, indicating an inevitable tendency toward hypertrophy and fibrotic remodeling in OA cartilage. This may be related to certain pro-fibrotic factors in the stem cell secretome or excessive ECM synthesis signaling that drives aberrant matrix deposition. Therefore, while stem cell therapies have potential to promote regeneration and suppress inflammation and pain in OA, their long-term effects need to be carefully modulated to preserve a stable hyaline-like cartilage phenotype. In this context, IMRCs may offer improved capacity for sustaining long-term repair in OA towards a native-like tissue state. Future strategies could involve comprehensive screening of stem cell secreted factors or combinational approaches with relevant agents, such as PTHrP [65,66], to further mitigate the risk of fibrosis.

Although our findings demonstrate the therapeutic potential of IMRCs in multiple OA models, several important preclinical studies are still required before clinical translation. In particular, validation in large animal OA models with joint structures and biomechanical properties closer to humans will be essential to further evaluate therapeutic efficacy and translational feasibility. In addition, long-term studies assessing cell retention, biodistribution, immune compatibility, tumorigenic risk, and repeated-dose safety, as well as dose-response optimization, are necessary to comprehensively establish the safety profile of IMRC-based therapy. Future investigations should also focus on standardized large-scale manufacturing and GMP-compatible production strategies to facilitate potential clinical application. Moreover, additional considerations including potential tumorigenicity concerns, defined release criteria for clinical-grade manufacturing, and requirements for long-term genomic stability assessment should be systematically addressed to ensure regulatory compliance and safe clinical translation of ESC-derived IMRCs.

4. Conclusion

In conclusion, this study demonstrates the modulatory effects of IMRCs in OA hyaline-like cartilage repair across cellular, organoid, explant, and in vivo models. IMRCs enhance cartilage matrix synthesis, suppress inflammation and angiogenesis, and effectively limit fibrotic remodeling. Compared to UC-MSCs, IMRCs show a greater ability to maintain cartilage homeostasis and inhibit fibrosis, and these effects are partially associated with MMP1. Collectively, these findings support the potential of IMRCs as an alternative therapeutic cell source for hyaline-like cartilage repair in OA and establish a multi-level preclinical evaluation framework for evaluating and optimizing future stem cell-based regenerative strategies.

5. Experimental section

5.1. Cell culture and characterization

UC-MSCs, IMRCs, ATDC5 cells, and HUVECs were used in this study. All cell types were cultured in α-MEM (TBD, China) supplemented with 10% fetal bovine serum (FBS; Gibco, USA) and 1% penicillin-streptomycin (P-S; Hyclone, USA), and cultures were maintained at 37 °C in an incubator with 5% CO₂. For surface marker characterization of MSCs and IMRCs, cells were harvested and incubated with fluorophore-conjugated antibodies for 30 min at 4 °C in the dark, including FITC anti-human CD45, FITC anti-human CD73, APC anti-human CD90, and PE anti-human CD105 (BioLegend, USA). After this, cells were washed with phosphate-buffered saline (PBS; Hyclone, USA) and analyzed using a flow cytometer (BD Biosciences, USA), and data were processed using FlowJo software.

5.2. Oxidative stress and senescence assay

MSCs and IMRCs were treated with 100 μM H₂O₂ for 4 h, after which the medium was replaced with fresh culture medium, and cells were cultured for a further 24 h period. Cellular senescence was assessed using a SA-β-Gal staining kit (Beyotime, China) according to the manufacturer's protocol. Cell viability was measured using the CCK-8 kit (Dojindo, Japan). The expression levels of senescence-related genes were analyzed by RT-qPCR.

5.3. Migration and colonization assays

Cartilage explants were obtained from the femoral condyles of the knee joints of euthanized New Zealand rabbits using a sterile blade. For each experiment, explants from two rabbits were pooled and randomly assigned to two groups: the MSC group and the IMRC group (n = 5 per group). For the migration assay, cartilage explants were placed in 24-well plates, and 2 × 10⁴ MSCs or IMRCs were seeded in the upper chamber of Transwell inserts (8 μm pore size, polycarbonate membrane; Corning, USA). After 24 h, cells were fixed with 4% paraformaldehyde (PFA) and stained with crystal violet for 5 min. Migrated cells on the underside of the Transwell membrane were observed under a light microscope. For the colonization assay, rabbit cartilage explants were placed in 96-well plates with the injured surface facing upward, then added with a suspension of 1 × 10⁴ MSCs or IMRCs labeled with Dil dye (Yeason, China). After incubating for 6 h in the dark, unattached cells were gently washed away using PBS, and attached cells were visualized using a fluorescence microscope (Olympus, Japan).

5.4. Chondrogenic differentiation

MSCs or IMRCs were subjected to chondrogenic induction using the pellet culture method. Briefly, each sample comprised 3 × 10⁵ cells centrifuged in a 15 mL tube to form a cell pellet, which was cultured in chondrogenic medium consisting of α-MEM supplemented with 100 nM dexamethasone (DEX; Selleck, USA), 10 μM L-ascorbic acid (Sigma, USA), and 10 ng/mL TGF-β3 (Solarbio, China). The medium was refreshed every 2-3 days. On day 21, the cartilage pellets were fixed with 4% PFA before being subjected to paraffin embedding and sectioning for histological evaluation. In parallel, RT-qPCR analysis was conducted to analyze the expression of chondrogenic genes.

5.5. OA chondrocyte model

To generate OA-like chondrocytes, ATDC5 cells grown to confluence were stimulated with a cytokine mix comprising 10 ng/mL IL-1β (Beyotime China), IL-6 (Navoprotein, China), and TNF-α (GenScript, USA) for 48 h to induce an inflammatory state [[67], [68], [69]]. After inflammatory induction, the medium was replaced with fresh culture medium and cells were used for subsequent experiments.

5.6. Construction of OCO model

OCOs were constructed using the micromass method. First, 1 × 10⁵ UC-MSCs (P4-P7) per well were seeded in round-bottom 96-well plates. The cells formed pellets after 24 h, which were transferred to suspension culture and commenced induction in chondrogenic medium. HCOs were obtained after 14 days of chondrogenic differentiation. Then, HCOs were cultured for an additional 7 days in α-MEM containing 100 nM DEX, 50 μM L-thyroxine (T4; Solarbio, China), 10 mM β-glycerophosphate (β-GP; Sigma, USA), and a pro-inflammatory cytokine mix comprising IL-1β, IL-6, and TNF-α each at 100 pg/mL, to induce hypertrophy and an inflammatory phenotype, resulting in OCOs. The concentrations of the induction components were optimized based on our previous studies and relevant published literature [30,[38], [39], [40], [41],[70], [71], [72]]. A half-volume medium change was performed every 2 days during the whole culture period. HCOs and OCOs were subsequently subjected to downstream experiments and analyses.

5.7. RT-qPCR analysis

Total RNA was extracted from cultured cells or organoids using TRIzol reagent (Invitrogen, USA) according to the manufacturer's instructions. RNA concentration and purity were determined by ultraviolet spectrophotometry using a Microdrop micro-spectrophotometer (Bio-DL, China). For each sample, 1 μg of total RNA was reverse transcribed into cDNA using the PrimeScript RT reagent kit (Takara, Japan). The primers used for RT-qPCR were synthesized by Sangon Biotech (China) (Tables S1–S2). GAPDH/Gapdh was used as the internal reference gene of human/mouse samples. Quantitative real-time PCR was performed using SYBR Green reagent (Takara, Japan) on a Bio-Rad CFX Connect Real-Time PCR Detection System (Bio-Rad, USA). Relative mRNA expression levels were calculated using the 2^−ΔΔCt method.

5.8. Western blot analysis

Cultured cells or organoids were lysed using radioimmunoprecipitation assay (RIPA) buffer (Beyotime, China) supplemented with protease inhibitor (Beyotime, China) and phosphatase inhibitor (Beyotime, China). The lysates were incubated on ice for 30 min and centrifuged at 12,000 g for 10 min to collect the supernatants. Protein concentrations were determined using a BCA protein assay kit (Beyotime, China). Protein samples were mixed with 5× SDS loading buffer and denatured at 95 °C for 10 min. Equal amounts of protein were separated by SDS-PAGE and transferred onto PVDF membranes (Millipore, USA). The membranes were blocked with 5% non-fat milk or bovine serum albumin (BSA) at room temperature for 1 h, followed by incubation with primary antibodies at 4 °C overnight. After washing with Tris-buffered saline with Tween 20 (TBST), the membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (Beyotime, China) at room temperature for 2 h. Protein bands were visualized using an enhanced chemiluminescence (ECL) detection reagent (Millipore, USA) through a gel imaging system. GAPDH was used as the internal loading control. The primary antibodies used in this study included: anti-GAPDH (Solarbio, China), anti-P65 (Affinity, China), anti-phospho-P65 (Affinity, China), anti-SMAD1/5/9 (Abmart, China), anti-phospho-SMAD1/5/9 (Abmart, China), and anti-MMP1 (Proteintech, USA).

5.9. ELISA assay

Organoids subjected to ELISA were thoroughly washed with PBS to remove residual factors from the induction medium and cultured in fresh α-MEM for an additional 48 h. The culture supernatants were then collected for analysis. The secretion levels of IL-6, MMP13, and VEGF were quantified using corresponding ELISA kits (Proteintech, USA) according to the manufacturer's protocols. All assays were performed based on the double-antibody sandwich method.

5.10. HUVEC tube formation assay

Matrigel (Corning, USA) was evenly coated onto 96-well plates (50 μL per well) and incubated at 37 °C for 30 min. HUVECs were resuspended and seeded at a density of 1 × 10⁴ cells per well in 100 μL of culture medium. The cells were subsequently grown in CM collected from HCOs or OCOs. After 8 h of incubation, the medium was carefully removed, and the cells were washed with PBS. Calcein-AM (Beyotime, China) was then used to stain the cells for 10 min, followed by washing with PBS, and fluorescence images were captured using a fluorescence microscope. Quantitative analysis was performed using ImageJ software.

5.11. Histological assessment

For cell samples (ATDC5 cells), they were fixed with 4% PFA and washed with PBS. GAG deposition was evaluated by AB (Psaitong, China) and SO (Solarbio, China) staining according to the manufacturers’ instructions. For tissue samples (pellets/organoids, cartilage explants, and knee joint tissues), they were collected and fixed in 4% PFA. The samples were decalcified if necessary, followed by paraffin-embedding and sectioning. Sections were subjected to hematoxylin & eosin (HE), AB, and SO-FG staining to evaluate tissue morphology and GAG deposition. For immunofluorescence (IF) staining, sections were incubated with primary antibodies and fluorescence-labeled secondary antibodies, and cell nuclei were counterstained with DAPI staining solution (Beyotime, China). For IHC staining, sections were deparaffinized, rehydrated, and subjected to antigen retrieval, followed by blocking and incubation with primary antibodies at 4 °C overnight. After incubation with HRP-conjugated secondary antibodies, immunoreactive signals were visualized using DAB substrate. All stained sections were imaged using light or fluorescence microscopy. Primary antibodies used included anti-COL-II (Novus, USA), anti-COL-X (Bioss, China), anti-COL-I (Novus, USA), and anti-MMP13 (Proteintech, USA).

5.12. Transcriptomic sequencing analysis (RNA-seq)

Organoids were collected and homogenized after 21 days in culture. Total RNA was extracted using TRIzol reagent according to the manufacturer's instructions. RNA quality was assessed, and samples meeting sequencing requirements were sent to Novogene (China) for library preparation and high-throughput sequencing. The raw sequencing data were subjected to initial quality control and cleaning, and DEGs were identified using the DESeq2 package. The DEGs were subjected to downstream analyses including Gene Ontology (GO) enrichment.

5.13. Secretome proteomic analysis

IMRCs and MSCs were cultured to 80% confluence, after which the medium was replaced with serum-free medium and the cells were cultured for a further 24 h period to collect CM. The CM was subsequently filtered through a 0.45 μm syringe filter to remove cellular debris. The filtered supernatant was concentrated using ultrafiltration units (Vivaspin Turbo 15 PES, 3 kDa MWCO) to prepare the concentrated CM. Protein concentrations were determined using a BCA assay kit, and equal amounts of protein were used for each sample. Proteins were reduced with dithiothreitol, alkylated with iodoacetamide, and digested with trypsin overnight at 37 °C. The resulting peptides were desalted and analyzed by LC-MS/MS on an Orbitrap mass spectrometer (Thermo Fisher, USA). Proteins were identified and quantified against the UniProt human protein database. Differentially secreted proteins between IMRCs and MSCs were determined using a fold-change cutoff of 2 and a statistical significance threshold of p < 0.05. Differential expression of secreted proteins was visualized using volcano plots and heatmaps. PPI networks were constructed using the STRING database. GSEA was performed using GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) gene sets to determine the biological significance of the proteins.

5.14. siRNA transfection

Well-grown IMRCs were seeded in 6-well plates and cultured until 50–70% confluence before transfection. Cells were transfected with MMP1-specific siRNA (ApexBio, USA) using Lipofectamine 3000 (Thermo Fisher, USA) according to the manufacturer's instructions. Total mRNA was extracted 2 days post-transfection for RT-qPCR analysis, and protein was collected at 3 and 7 days post-transfection for Western blot analysis to verify the efficiency of MMP1 knockdown. In the co-culture experiments with OCOs, IMRCs were pretreated with siMMP1 for 2 days before being used for subsequent co-culture.

5.15. Ex vivo analysis using rabbit articular cartilage explants

Full-thickness cartilage explants were obtained from the knees of euthanized New Zealand rabbits using a sterile 4-mm-diameter drill. The explants were pretreated for 5 days in α-MEM (TBD, China) supplemented with 10 ng/mL IL-1β, IL-6, and TNF-α to induce an inflammatory state (INFL) [69,73,74]. The medium was then replaced with fresh α-MEM, followed by co-culturing the explants with MSCs or IMRCs in a Transwell system for 10 days. Cells in the upper chamber were refreshed every 3 days to maintain their viability and function. At the end of the culture period, explants were fixed in 4% PFA, paraffin-embedded, and sectioned for subsequent histological analysis.

5.16. In vivo analysis using ACLT-induced rat OA model

A total of 12 male Sprague–Dawley (SD) rats (8 weeks old, weighing 250-300 g) were randomly assigned to three groups: PBS control, MSC treatment, and IMRC treatment (n = 4 per group). Right knee joints of rats were subjected to ACLT surgery under anesthesia to establish the OA model [75]. Prophylactic penicillin was administered once daily for 3 consecutive days after surgery to prevent postoperative infection. Starting from 2 weeks post-surgery, an 100 μL suspension containing PBS only (control group) or 1 × 10⁶ MSCs or IMRCs (treatment groups) was administered by intra-articular injection every 2 weeks, with the final injection administered at 8 weeks post-surgery. To evaluate the in vivo retention of transplanted cells, MSCs and IMRCs were labeled with the fluorescent dye DiR (Yeason, China) prior to injection, and in vivo fluorescence imaging was performed using an IVIS imaging system 14 days after intra-articular administration. Rats were euthanized at 10 weeks post-surgery and knee joints were harvested for histological and IHC analyses.

5.17. Statistical analysis

All data are presented as mean ± SD. n refers to the number of biological replicates. Statistical analyses were performed using GraphPad Prism software. Comparisons between two groups were conducted using two-tailed Student's t-test, while comparisons among multiple groups were performed using one-way ANOVA with Tukey's multiple comparison test. p < 0.05 was considered a statistically significant difference between groups.

Ethics approval and consent to participate

All animal-related experiments were conducted under the approval of the PKUPH Animal Use Protocol & Ethic Review Committee (2023PHE099). This study did not involve human participants, and informed consent was not applicable.

Funding information

This work was supported by National Key Research and Development Program of China (2024YFA1108600, 2023YFC3605100), the Beijing Natural Science Foundation (L232094, L244019, Z240018), the National Natural Science Foundation of China (82272538, 82302776, 82472482, 82370851), Peking University People's Hospital Scientific Research Development Funds (RS2024-04, RDXL2025-02, RDXL2025-04), and Rebecca L Cooper Medical Research Foundation (Australia).

CRediT authorship contribution statement

Yiqi Su: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Writing – original draft. Zihao He: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Writing – original draft. Kai Wang: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization. Tingting Gao: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization. Jiao Jiao Li: Funding acquisition, Methodology, Resources, Visualization, Writing – review & editing. Qianqian Chen: Formal analysis, Investigation, Methodology, Software, Validation. Long Chen: Formal analysis, Investigation, Methodology, Software, Visualization. Shasha Liu: Formal analysis, Investigation, Methodology, Software, Visualization. Fanhao Ye: Conceptualization, Data curation, Investigation, Validation. Jingzhou Wang: Conceptualization, Data curation, Investigation, Validation. Shichun Xu: Conceptualization, Data curation, Investigation, Validation. Du Wang: Data curation, Formal analysis, Investigation, Methodology. Zhen Yang: Data curation, Formal analysis, Investigation, Methodology. Zijin Zhou: Data curation, Formal analysis, Investigation, Methodology. Caiyi Wei: Data curation, Formal analysis, Investigation, Methodology. Jun Wu: Conceptualization, Funding acquisition, Methodology, Project administration, Resources. Jianhao Lin: Conceptualization, Funding acquisition, Methodology, Project administration, Resources. Hui Li: Conceptualization, Funding acquisition, Project administration, Resources, Validation, Writing – review & editing. Dan Xing: Conceptualization, Investigation, Methodology, Project administration, Resources, Validation, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article.