Regeneration of full-thickness cartilage defects using small umbilical cord–derived fast proliferating cells combined with magnesium in a rat model

Abstract

The effective application of allogeneic mesenchymal stem cells (MSCs) has the potential to enhance cartilage regeneration. This study aimed to evaluate the therapeutic efficacy of intra-articular (IA) injections of small umbilical cord–derived fast proliferating cells (smumf cells) combined with magnesium (Mg²⁺) in a rat model of full-thickness cartilage defects (FTDs). Adhesion of smumf cells was assessed on type I collagen-coated surfaces in vitro, in an uncontained ex vivo model, and via cell tracking in a rat FTD model. Therapeutic efficacy was evaluated using histological analyses of macroscopic and microscopic in vivo (at 4 and 8 weeks, n = 6). Mg²⁺ improved the adhesion of smumf cells by up to 1.89-fold in the ex vivo model, and by 2.80-fold in cell tracking. A single injection of smumf cells alone improved histological scores by 2.33-fold at 4 weeks, whereas the combination with Mg²⁺ resulted in further improvements in both macroscopic (1.30-fold) and microscopic (1.26-fold) scores at 8 weeks. Moreover, the smumf cells + Mg²⁺ group showed significant increases in tissue thickness (1.40-fold), safranin O-positive area (2.88-fold), and type II collagen synthesis (1.22-fold) in rat model. This study demonstrates that a single IA injection of smumf cells combined with Mg² enhances cell homing and functional adhesion at the defect site, thereby promoting superior cartilage regeneration compared with smumf cells alone.

Graphical Abstract.

Introduction

Knee osteoarthritis (OA) has the highest global prevalence among all joint diseases and is characterized by the presence of cartilage defects, making early treatment crucial to prevent or delay disease progression^1,2. Cartilage defects can generally be classified into three types based on the depth of the damage: osteochondral defect (OCD), full-thickness defect (FTD), and partial-thickness defect, each associated with a distinct microenvironment^3,4. In FTD, an intact subchondral bone plate blocks the migration of endogenous bone marrow–derived mesenchymal stem cells (BM MSCs) into the defect area ⁵ . Mesenchymal stem cells (MSCs), whether autologous or allogeneic, have shown great potential in the treatment of musculoskeletal disorders due to their immunomodulatory, anti-inflammatory, and multipotent nature^6–9. Recent evidence also suggests that stem cells exert paracrine effects, facilitating the endogenous reparative process by secreting numerous cytokines and growth factors ¹⁰ . The selection of the source of MSCs should meet the requirements for clinical applications. However, it is even more critical to isolate cells with both high quality and sufficient quantity and to culture them in a manner that preserves robust proliferative capacity while preventing cellular senescence ¹¹ .

The use of autologous MSCs has been criticized for several disadvantages including invasive harvesting techniques, low collection efficiency, and decreased quality with donor age and comorbidities^12–15. Umbilical cord–derived MSCs (UC MSCs) are fetal MSCs isolated from umbilical cord tissue, which can overcome the limitations associated with MSCs derived from other tissues ¹⁶ . Over the past decade, UC MSCs have emerged as a dominant source of MSCs for clinical trials^17,18. MSCs isolated from fetal tissues have been shown to secrete higher levels of growth factors than obtained from adult tissues^19,20. Previous studies have established a suitable method for clinical scale production of MSCs from human umbilical cord tissue—namely, small umbilical cord–derived fast proliferating cells (smumf cells), which demonstrate excellent proliferative capacity without transformation ¹¹ . Moreover, studies have shown regenerative effects in tendon injury following injection of smumf cells, even without the use of carrier materials^21,22. For cartilage regeneration, a few studies have shown regenerative potential using hyaluronic acid (HA) in an OCD animal model^23,24. However, the regenerative efficacy of UC MSCs for cartilage defects via direct intra-articular (IA) injection has not yet been reported.

For clinical trials, key challenges extend beyond selecting an appropriate MSC type to include choosing suitable carrier for delivery into injured joints ²⁵ . A strategy that facilitates the attachment of allogeneic MSCs to the lesion site through a clinically translatable method is crucial ²⁶ . The direct cell-to-cell communication and paracrine signaling are considered key mechanisms in MSC-mediated regeneration ²⁷ . Studies have shown that direct cell-to-cell contact between MSCs and chondrocytes may facilitate the deposition of extracellular matrix (ECM) in cartilage tissue^28,29. Moreover, reduced cell survival and engraftment rates lead to a diminished paracrine effect of the stem cells ³⁰ . Therefore, rapid and sufficient adhesion of MSCs to the defect site is important after injection. In addition, IA injection of MSCs is a simple, minimally invasive, and efficient delivery procedure ³¹ . Research focusing on the homing and adhesion of UC MSCs is limited, particularly studies investigating the effects of rapid adhesion on cartilage regeneration. Among the candidate agents that may enhance adhesion potential of MSCs, an appropriate concentration of magnesium (Mg²⁺) has been shown to have potential under local adhesion ³² . Mg²⁺ is an important factor in both the growth and maintenance of live cells, which also enhances the adhesion of cells by integrin pathway ³³ . Moreover, studies have confirmed Mg²⁺ injection promotes cartilage matrix synthesis and exhibits anti-inflammatory properties ³⁴ . However, the combined effects of UC MSCs and Mg²⁺ on cartilage repair have not yet been investigated. The ultimate goal of MSC-based therapies is clinical translation; therefore, the translational relevance of the selected carrier must be considered. Mg²⁺-containing isotonic solutions are balanced electrolyte intravenous fluids that are used to provide volume replacement and correct electrolyte imbalances, and they may serve as an ideal carrier given their widespread clinical use and well-established safety profile ³⁵ .

Thus, the purpose of this study is to evaluate the efficacy of smumf cells combined with Mg²⁺ in regenerating FTDs in a rat model. We hypothesized that the addition of Mg²⁺ would enhance cartilage regeneration through cell adhesion, resulting in superior macroscopic and histological outcomes compared to smumf cells treatment alone.

Materials and methods

The smumf cells isolation and cryopreservation

The institutional ethical review board approved tissue collection for this research. Human umbilical cords were obtained from healthy full-term deliveries by cesarean section after informed consent at SMG-SNU Boramae Medical Center (IRB No. 06–2011–211), and smumf cells were isolated and cultured using the minimal cube explant method ¹¹ . The smumf cells were continuously cultured up to passage 8, and smumf cells were cryopreserved with a STEM-CELLBANKER^® (ZENOGEN PHARMA CO., LTD., Fukushima, Japan) ³⁶ . Cryovials were stored at −80°C in a deep-freezer and transferred to a −196°C liquid nitrogen tank for preservation for at least 1 month. For use in experiments, cryopreserved smumf cells were thawed immediately in the 37°C water bath within 1 min and cultured in a 5% CO₂ incubator with humidified air at 37°C. The subculture protocol was the same as the smumf cells isolation protocol ¹¹ . The smumf cells at passage 10 were used for all experiments, and characterization including morphology, growth kinetics, colony-forming-unit fibroblasts (CFU-F) assays, flow cytometry, and trilineage differentiation was performed as previously reported ³⁷ .

The adhesion of smumf cells on type I collagen-coated surfaces

In vitro cell adhesion assay was performed on type I collagen-coated 96 well plates. The details are described in the Supplemental Material (S4). The detached smumf cells were resuspended with saline containing different doses of Mg²⁺ and Mg²⁺-containing isotonic solution (MgIS, Plasma solution-A, HK inno.N Co., Ltd., Cheongju, Korea). A total of 32,000 cells with carrier were injected per well, and the procedure was repeated in triplicate. The plate was incubated in a 5% CO₂, 37°C incubator for 30 min, and then washed twice with Dulbecco’s Phosphate-Buffered Saline(DPBS). Cell viability was assessd using the EZ-CyTox cell viability assay kit (Daeil Lab Service, Seoul, Korea) with a 3-h incubation. The absorbance at 450 nm was measured using a microplate reader to quantify the number of adhered cells. A standard curve was generated using the same procedure, and the number of adhered cells was calculated based on OD values. The cell morphologies were observed under a microscope (CKX53 Olympus culture microscope; Olympus, Tokyo, Japan).

The adhesion test of smumf cells in human osteochondral explant

Tissue collection was approved by the Institutional Review Board of Seoul Metropolitan Government-Seoul National University Boramae Medical Center (IRB No. 30-2019-124). Human osteochondral samples were obtained from patients undergoing knee replacement surgery. Cylindrical osteochondral explants (4 mm in diameter, 4 mm in height) were extracted using a bur, and the cartilage layer was completely removed with a surgical blade to mimic a cartilage defect. A 2% agarose gel (Agarose MP, Roche, Basel, Switzerland) was loaded into 48-well plates to a height of 4 mm. After solidification, a 4-mm diameter hole was created using a biopsy punch, and an osteochondral explant without the cartilage layer was inserted into the agarose hole (Fig. 1c). Depending on the group, smumf cells were suspended in different carriers—including saline, MgIS, HA (1%; LG Life Sciences, Seoul, Korea), and fibrin glue (FG; Tisseel, Baxter korea, Seoul, Korea) and injected into the wells. After the 30 min of incubation to cell adhesion, the explants were carefully removed and washed twice thoroughly with DPBS to eliminate unattached cells. The explants were then immersed in the EZ-CyTox cell viability assay kit and incubated for 1 h. The absorbance at 450 nm was measured using a microplate reader to quantify the number of adhered cells. A standard curve was generated using the same procedure, and the number of adhered cells was calculated based on optical density (OD) values.

Figure 1.

Cell adhesion test of smumf cells in vitro and ex vivo. (a) Morphology of smumf cells adhered to type I collagen condition; Scale bar:100 µm. (b) Quantitative analysis of smumf cells adhesion on type I collagen surface under varying concentrations of magnesium ions and Mg²⁺-containing isotonic solution (MgIS). (c) Schematic representation of bone explant preparation to the ex vivo model setup. (d) Comparative analysis of early adhesion in the ex vivo model. (e) Comparison of adhesion test using other carriers. Data represent mean ± standard deviation (SD); n > 5. *P < 0.05; **P < 0.01; ***P < 0.001.

Animal model of FTD

All procedures were conducted following the ARRIVE guidelines and approved by the Institutional Animal Studies Committee (IACUC No. 2022-0016). Forty-eight healthy male Sprague-Dawley rats (12 weeks old, weight 360–380 g) were using for modeling surgery. Animals were anesthetized with zoletil and rompun (30 mg/kg + 10 mg/kg), and surgery was performed on the right hind limb only, following a unilateral design. A parapatellar medial surgical approach was used with lateral dislocation of the patella to expose the trochlear groove of the femur. The FTD model (defect: 1.75 mm × 1.5 mm, oval) was created by using a 1.5-mm biopsy punch. The cartilage defect was debrided using a biopsy curette to remove the calcified cartilage layer without damaging the subchondral bone and bleeding. The patella was relocated, and the joint capsule, subcutaneous tissue was closed layer by layer. All animals were allowed free cage activity before being sacrificed.

The cell tracking of PKH26 labeled-smumf cells in vivo

The smumf cells were labeled with the fluorescent PKH26 (Sigma-Aldrich, St Louis, MO, USA) according to the manufacturer’s protocol to facilitate in vivo cell trafficking ³⁸ . The details are described in Supplemental Material (S4). After labeling, the cells were counted by hemocytometer, and confirmation of red fluorescence was obtained using a fluorescence microscope (Leica DMI 4000B, Leica, Wetzlar, Germany) before injection ¹² . For a cell-tracking study, six FTD model rats were divided into two groups (smumf cells or smumf cells + Mg²⁺) for tracking at 1 and 3 days after injection. For IA injection, 1 × 10⁶ PKH26-labeled smumf cells were suspended in either saline or MgIS to a final volume of 100 μL. At 1 and 3 days after injection, the entire right knee joint was harvested (Supplemental Material S4). To evaluate the adhesion of PKH26-labeled smumf cells at defect sites, specimens were sectioned at 6 μm thickness using a freezing microtome (Leica CM3050S, Leica, Wetzlar, Germany) and carefully trimmed. Slides were mounted with VECTASHIELD® mounting medium containing 4’,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories Inc., Burlingame, CA, USA). High-magnification images (×100) of the defect sites were obtained using fluorescence microscopy, and PKH26-positive cells were counted per image.

In vivo study for IA injection of smumf cells

For in vivo study, 36 Sprague-Dawley rats were randomly divided into three groups (n = 6 in each group for two time points) as follows: (1) FTD + MgIS (control group); (2) smumf cells + saline (smumf cells group); (3) smumf cells + MgIS (smumf cells + Mg²⁺ group). The cryo-preserved and cultured smumf cells were suspended in either saline or MgIS to a final volume of 1 × 10⁶/100 μL. A 31-gauge needle was used to inject the cells through the patellar ligament into the IA space of the right knee. The control group was injected to 100-μL MgIS by the same protocol. Exclusion criteria included unexpected death, surgical or injection failure, severe local or systemic complications, inability to complete the follow-up period, or technical inadequacy of tissue samples for analysis.

Histological evaluation and immunohistochemical evaluation

Four and 8 weeks after the surgery, the animals were sacrificed. Tissues surrounding the trochlear groove were harvested, and macroscopic evaluations were independently performed in a blind manner by two researchers using the International Cartilage Repair Society (ICRS) macroscopic evaluation scoring system ³⁹ . The distal femurs were fixed in 4% (w/v) paraformaldehyde (PFA; Merck, Darmstadt, Germany) for 48 h at 4°C, followed by decalcification in 5% formic acid (Biosesang, Korea) for 3–4 days. After decalcification, tissues were embedded in paraffin, and serial sections of 4 µm thickness were prepared. For microscopic evaluation, slides were stained with safranin-O and fast green (Saf’O), observed under light microscopy, and assessed independently by two blinded researchers according to the O’Driscoll histologic grading system ⁴⁰ . For glycosaminoglycans (GAGs) content analysis, the initial cartilage defect area, the total area of regenerated tissue filling the defect, and the Saf’O-positive stained areas within the regenerated tissue were measured using ImageJ software (National Institutes of Health, Behesda, MD, USA). Values for total regenerated tissue were expressed as percentages relative to the original cartilage defect area (set as 100%). Values for Saf’O-positive areas were expressed as percentages of the total regenerated tissue. For measurement of thickness, the initial cartilage thickness, the total thickness of regenerated tissue was measured using ImageJ software (National Institutes of Health, Behesda, MD, USA). Thickness values of regenerated tissue were expressed as percentages of initial cartilage thickness (set as 100%).

The same measurement method was used to quantify collagen-positive areas within the regenerated tissue ³⁸ . For immunohistochemistry (IHC), slides were blocked with Superblock (Scytek, Logan, UT, USA; NC944219), incubated overnight at 4°C with primary antibodies against Collagen I (LSBio, Seattle, WA, USA; LS-C343921-100) or Collagen II (Abcam, Cambridge, UK; ab34712), and subsequently incubated with secondary antibody (anti-rabbit IgG, HRP(horseradish peroxidase)-linked antibody, Cell Signaling, 7074) at room temperature for 30 min. Finally, sections were incubated with Liquid DAB+ substrate (GBI Labs, Rockville, MD, USA; C09-100) for detection of HRP enzyme activity, dehydrated, and mounted.

Analysis of short-term repeat

To identify the human-derived gene of the repair tissue at 8 weeks, genomic DNA was isolated from rat cartilage tissue from the smumf cells + Mg²⁺ group. DNA amplification and genotyping were performed using the AmpFlSTR® Identifiler® PCR Amplification Kit (Applied Biosystems, Foster City, CA, USA). The prepared samples were loaded into the appropriate wells of a genetic analyzer plate and analyzed using a 3730xl DNA Analyzer (Applied Biosystems). The resulting data were processed using GeneMapper ID-X Software (Applied Biosystems).

Statistics

Results are shown as mean ± standard deviation (SD). The t test was used for statistical analysis of in vitro and ex vivo studies, results of histological evaluations were statistically analyzed using the one-way ANOVA test followed by Bonferroni corrections. A P-value less than 0.05 was considered statistically significant. All statistical analyses were conducted using IBM SPSS Statistics version 20.0 (IBM Corp., Armonk, NY, USA).

Results

Adhesion of smumf cells on type I collagen-coated surface in vitro

After 30 min of initial adhesion, smumf cells exhibited a rounded morphology while adhering to the type I collagen-coated plate surfaces. A greater number of round-adhered cells were observed in the Mg²⁺ and MgIS groups compared to the control group (Fig. 1a). Quantitative analysis revealed that Mg²⁺ enhanced smumf cells adhesion by 3.26- to 4.49-fold (P < 0.001). However, no significant differences were observed between the different Mg²⁺ ion concentrations and the MgIS group (Fig. 1b).

Adhesion of smumf cells to human osteochondral explants ex vivo

In an ex vivo model using human bone explants, the MgIS group showed a 1.46-fold and 1.87-fold increase in cell adhesion at 0.5 h and 1 h, respectively, compared to the control group (P < 0.05) (Fig. 1d). In comparison of cell adhesion among different carrier types, the MgIS group demonstrated a 1.89-fold increase compared to the HA group and a 1.39-fold increase compared to the FG group (Fig. 1e). The differences in adhesion between MgIS and both HA and FG groups were statistically significant.

In vivo tracking image of smumf cells after IA injection

PKH26-labeled smumf cells were clearly observed on the defect surface and remained detectable at the site on day 3 (Fig. 2b). One day after injection, in the smumf cells group, the average number of PKH26-positive cells was 20.33 ± 5.68/0.05 mm², whereas in the smumf cells + Mg²⁺ group, it increased to 28.0 ± 6.55/0.05 mm². Although the number of PKH26-positive cells in the smumf cells + Mg²⁺ group showed a 1.38-fold increase compared to that in the smumf cells group, the difference was not statistically significant. Three days after injection, the number of PKH26-positive cells in the smumf cells + Mg²⁺ group increased to 50.3 ± 17.01/0.05 mm², showing a more than 2.80-fold increase compared to the smumf cells group (P < 0.01) (Fig. 2c).

Figure 2.

Cell tracking of labeled smumf cells after IA injection in a rat model. (a) Gross morphology of in vivo samples and microscopic images; scale bar: 2 mm. (b) Tracking of smumf cells at early time points after intra-articular injection; scale bar: 50 µm. (c) Quantification of smumf cells exhibiting PKH26-positive cells at the defect site. Data represent mean ± SD; n = 3, **P < 0.01.

Histological outcomes at 4 weeks after the IA injection

At 4 weeks after the injection, the defects in the control group were still depressed, although reparative tissue filled the defect site partially in best case. However, the filled tissue appeared opaque or showed irregularity of the articular surface, and the borders of the defect remained visible in all cases (Fig. 3a). In contrast, the defects in both smumf cells–treated groups were nearly completely filled with repaired tissue, and the neo-cartilage-like tissue was observed in all cases. Unlike the control groups, the tissues exhibited a transparent, shiny, and smooth articular surface and were integrated with the adjacent tissue. The ICRS macroscopic score showed significantly worse repair in the control group than in the other two smumf cells groups at 4 weeks. The macroscopic scores in the two smumf cells–treated groups were higher than those in the control group (2.83 ± 1.47 vs 6.6 ± 1.14, P < 0.01; vs 6.33 ± 1.37, P < 0.05) at 4 weeks (Fig. 3b). In addition, the heatmap of ICRS scores provided the numerical distribution of reparative levels the best, moderate, or worst outcomes in each group. Treatment with smumf cells improves the overall regenerative effects compared to the control group (Fig. 3c). For microscopic evaluation, the control group showed irregular thickness and cracks in the defect, along with hypocellularity at the integration sites. In contrast, the smumf cells group exhibited more GAGs content, which integrated well with the surrounding cartilage and subchondral bone. Surprisingly, the smumf cells + Mg²⁺ group exhibited nearly normal hyaline cartilage structures with rich GAGs content and a smooth surface in both the best and moderate cases. Even in the worst case, GAGs content was observed compared to that in the other groups. In addition, the repaired tissue remained intact in the subchondral bone region and integrated with the normal cartilage. In total O’Driscoll scores, both the smumf cells and smumf cells + Mg²⁺ groups showed significant differences compared to the control group (10.33 + 1.21 vs 15.00 + 2.35, P < 0.05; vs 15.83 + 3.19, P < 0.01) (Fig. 3d). However, no statistically difference was observed between the two smumf cells–treated groups at 4 weeks. The heatmap of O’Driscoll scores provided the numerical distribution of reparative levels the best, moderate, and worst outcomes in each group (Fig. 3e). Treatment with smumf cells improved the overall regenerative effects, similar to the macroscopic outcomes in each case. In particular, the smumf cells + Mg²⁺ group demonstrated markedly higher scores in the Saf’O and cellular parameters compared to the control group (1.67 + 0.82 vs 0.00 + 0.00, P < 0.01), while the smumf cells group showed significantly higher thickness than the control group (1.60 + 0.55 vs 0.33 + 0.52, P < 0.05) (Supplemental Table S1).

Figure 3.

Histological evaluation of articular cartilage defects regenerated using smumf cells at 4 weeks after injection. (a) Macroscopic images representing the best, moderate, and worst outcomes in each group; scale bar: 2 mm. (b) Total International Cartilage Repair Society (ICRS) scores; n = 6. (c) Heatmap visualization of ICRS scores corresponding to the best, moderate, and worst outcomes. (d) Total O’Driscoll scores; n = 6. (e) Heatmap visualization of O’Driscoll scores corresponding to the best, moderate, and worst outcomes. (f) Representative microscopic images of best, moderate, and worst outcomes in each group; scale bar: 500 µm. (g) Immunohistochemical staining for type II collagen expression; scale bar: 500 µm. n = 3. Data are represented as mean ± SD. *P < 0.05; **P < 0.01.

Histological outcomes at 8 weeks after IA injection

At 8 weeks after injection, the defects were macroscopically filled to the original cartilage height in the control group, but defects or cracks were observed in the filled tissue (Fig. 4a). In contrast, the defects in both smumf cells–treated groups remained nearly completely filled, and the tissues also exhibited a transparent, shiny, and smooth articular surface in all cases. The ICRS macroscopic score showed the best repair remaining in the smumf cells + Mg²⁺ group, which was higher than that in the control group at 8 weeks (9.00 ± 1.26 vs 4.83 ± 1.17, P < 0.01) (Fig. 4b). In addition, the heatmap of ICRS scores demonstrated that the smumf cells + Mg²⁺ group showed improved overall regenerative effects compared to the control group (Fig. 4c). For microscopic evaluation, the control group exhibited poor GAGs content and obvious cracks in the repaired tissue at the center or at integration sites (Fig. 4f). In addition, abnormal subchondral bone bonding and hypocellularity were also observed in the worst case. Although the smumf cells group exhibited less GAGs content and cracks with an irregular thickness, it remained well integrated with the surrounding tissue and subchondral bone. In contrast, the smumf cells + Mg²⁺ group exhibited nearly normal thickness, maintaining a smooth surface in both the best and moderate cases. In the worst case, there was irregular thickness in the center, but hypocellularity or abnormal integration with the normal cartilage was not observed. The smumf cells + Mg²⁺ group showed significant differences in total O’Driscoll scores compared to the control group (15.14 ± 1.26 vs 9.33 ± 2.42, P < 0.01) (Fig. 4d). In the sub-parameters, the smumf cells + Mg²⁺ group exhibited significantly higher scores in cellularity (P < 0.05), surface, and thickness (P < 0.01) than the control group (P < 0.01) (Supplemental Table S1). The heatmap of O’Driscoll scores demonstrated that the smumf cells + Mg²⁺ group showed improved overall regenerative effects compared to the other groups (Fig. 4e).

Figure 4.

Histological evaluation of articular cartilage defects regenerated using smumf cells at 8 weeks after injection. (a) Macroscopic images representing the best, moderate, and worst outcomes in each group; scale bar: 2 mm. (b) Total International Cartilage Repair Society (ICRS) scores; n = 6. (c) Heatmap visualization of ICRS scores corresponding to the best, moderate, and worst outcomes. (d) Total O’Driscoll scores; n = 6. (e) Heatmap visualization of O’Driscoll scores corresponding to the best, moderate, and worst outcomes. (f) Representative microscopic images of best, moderate, and worst outcomes in each group; scale bar: 500 µm. (g) Immunohistochemical staining for type II collagen expression; scale bar: 500 µm. Data are represented as mean ± SD. **P < 0.01; ***P < 0.001.

Quantitative analysis of regenerated tissue at 4 and 8 weeks

For the quantitative quality of the repaired tissue, the smumf cells + Mg²⁺ group exhibited superior hyaline-like cartilage formation (Fig. 5a), which was significantly thicker (64.21% vs 38.74%, P < 0.05) and had richer Saf’O positive area than the control group (42.83% vs 3.05%, P < 0.01) and the smumf cells group at 4 weeks (vs 14.52%, P < 0.01) (Fig. 5b, c). Type II collagen formation showed no significant difference among all groups (Fig. 5d), and the same was observed for type I collagen at 4 weeks (Supplemental Fig. S2). However, the smumf cells + Mg²⁺ group exhibited significantly superior type II collagen formation (69.52% vs 42.84%, P < 0.05) and tissue thickness (68.78% vs 35.29%, P < 0.01) than the control group at 8 weeks (Fig. 5f, h).

Figure 5.

Quantitative analysis of articular cartilage defects regenerated using smumf cells at 4 and 8 weeks after injection. (a) Microscopic images showing regenerated tissue thickness and GAGs content in the best outcomes of each group; scale bar: 250 µm. (b) Percentage of normal cartilage thickness repaired in regenerated tissue. (c) Percentage of Safranin O-positive area in regenerated tissue. (d) Immunohistochemical analysis showing areas positive for type II collagen; n = 3. (e) Microscopic images showing regenerated tissue thickness and type II collagen expression of each group; scale bar: 250 µm. (f) Percentage of normal cartilage thickness repaired in regenerated tissue. (g) Percentage of Safranin O-positive area in regenerated tissue. (h) Immunohistochemical analysis showing areas positive for type II collagen; n = 3. Data are represented as mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001.

Short Tandem Repeat analysis

At the final time point, no human-specific short tandem repeat (STR) profiles were detected in the smumf cells + Mg²⁺ group. Detailed information can be found in the accompanying report (Supplemental data S3).

Discussion

The most important findings of this study were that Mg²⁺ increased the adhesion of smumf cells in an ex vivo model and enhanced their local homing and adhesion to the defect site in a rat model. In the in vivo study, a single injection of smumf cells improved histological outcomes, and the combination with Mg²⁺ further enhanced regenerative effects, showing significant increases in tissue thickness, Safranin O–positive area, and type II collagen–positive area compared with the smumf cell group. Taken together, these findings suggest that Mg²⁺ enhances the homing and initial adhesion of smumf cells within cartilage defects. Furthermore, while a single injection of smumf cells can promote cartilage regeneration through endogenous regenerative process in the rat FTD model, the combination with Mg²⁺ further improves both the quantity (tissue thickness) and quality (GAGs synthesis and type II collagen deposition) of the regenerated tissue.

Magnesium is an important factor in both the growth and maintenance of live cells, and it also enhances the adhesion of cells by integrin pathway ³³ . Mg²⁺ concentrations of 1.25–10 mM were shown to promote rabbit chondrocyte proliferation in culture, and 10 mM Mg²⁺ in an OCD ex vivo model increased the number of adherent synovial-derived MSCs approximately 2-fold compared to Phosphate-buffered saline (PBS)^32,41. To mimic the homing process within a joint cavity, we developed an uncontained ex vivo model that simulates cell homing and targeted attachment to the defect surface rather than local adhesion (Fig. 3c). The smumf cells suspended in MgIS—a clinically applicable carrier containing 3.15 mM Mg²⁺ and without Ca²⁺—were injected into the ex vivo model to compare their initial attachment ability. We first compared the early adhesive rates of smumf cells on type I collagen-coated plates, which simulated FTD surface condition, and observed no significant differences among the different concentrations of Mg²⁺, also including MgIS. Compared with the other carriers, HA and FG, the MgIS group also showed a significant increase in cell adhesion. HA serves as a scaffold with anti-inflammatory and analgesic effects, and FG enhances MSC viability and tissue adhesion but is technically challenging to deliver via minimally invasive IA injection^42,43. Increasing evidence supports that mitochondrial dysfunction plays an essential role in MSC senescence, also impacting the quantity and quality of MSCs and limiting their clinical use^44,45. Meanwhile, mitochondrial loss in MSCs can decrease their secretory capacity and interfere with cytokine secretion^46,47. In our study, tracking of PKH26-labeled smumf cells showed increased retention when combined with Mg²⁺, with signals detected at the lesion site for up to 3 days in the rat FTD model. Using a mitochondrial activity assay to confirm MSC adhesion, rather than cell counting, demonstrates that the cells are not just physically attached but are also viable and functionally adhering to the surface. Although most of the injected cells may adhere to the synovial tissue rather than the cartilage, the regenerative effect is likely to depend more directly on the direct adhesion of MSCs to the defect site ⁴⁸ . Such adhesion is of particular importance because, only after MSCs adhere to a target site, they release various paracrine factors with immunomodulatory, angiogenic, and antiapoptotic properties, all of which exert therapeutic effects on cartilage tissue^49,50. Collectively, these data confirm that smumf cells successfully homed to and persisted at the defect site even in vivo, and we further demonstrate that Mg²⁺ enhanced the initial attachment capacity of the injected smumf cells rather than merely causing them to be physically attached.

For cartilage defects, increased thickness or volume of regenerated tissue can provide additional mechanical support to the joint, which will result in improved function and reduced pain^51,52. Clinical studies have shown that patients with thinner fibrocartilage repair tissue are more susceptible to subsequent cartilage degeneration ⁵² . Several studies have demonstrated that, without scaffolds, achieving sufficient tissue thickness to adequately fill the defect remains challenging^53,54. However, the use of scaffolds may often lead to the invasion of bony structures into the repaired cartilage, causing the cartilage to become thinner than the surrounding normal articular cartilage ⁵⁴ . In contrast, the use of scaffold-free carriers for MSC administration may minimize the potential impact on the surrounding microenvironment or the implanted MSCs ⁵⁵ . In our animal experiments, smumf cells—either alone or in combination with an MgIS—were injected into rat FTDs, and cartilage regeneration was first observed at 4 weeks after single injection. As observed in the best and moderate histological images, the defects in both smumf cells–treated groups were uniformly filled and resembled the surrounding normal articular cartilage. Notably, by 8 weeks, the smumf cells + Mg²⁺ group exhibited a significantly greater thickness in the regenerated tissue—up to 1.95-fold—compared to the control group. When human MSCs are used for cell therapy, extensive preclinical culture expansion inevitably drives the cells toward replicative “aging,” leading a consequent decline in quality ⁵⁶ . These cells display senescence signatures, including reduced proliferative and immunosuppressive capacity, weakened regenerative potential, and proinflammatory features^56,57. Our previous study reported that when smumf cells were cultured up to passage 10, no difference was observed in the expression of p16, p21, or p53, whereas the expression levels of these genes were significantly increased in BM MSCs ¹¹ . Research also demonstrates that aged BM MSCs produce significantly less cartilage matrix than younger cells ⁵⁸ . Unlike aged MSCs, which tend to produce fibrocartilage rather than hyaline cartilage, the smumf cells used in this study may not have this limitation ⁵⁸ . These findings suggest that injection of smumf cells not only promotes regeneration and restores the thickness of defects, when combined with Mg²⁺, it also prevents further degradation of the newly formed tissue.

Articular cartilage is a type of hyaline cartilage that is rich in aggrecan and type II collagen, which provide compressive, tensile, and shear strength ³⁸ . GAGs contained within aggrecan also play a vital role in ECM assembly, influencing the mechanical properties of cartilage ⁵⁹ . At later stages, insufficient type II collagen formation may lead to the formation of fibrocartilage instead of hyaline cartilage, thereby compromising long-term functional integrity ⁶⁰ . It remains unclear whether there is a specific temporal relationship between GAGs deposition and collagen formation; however, both are closely associated with the functional properties and quality of the regenerated tissue. In canine OCD model, the combined treatment of microfracture (MFx) and ultra-purified alginate gel resulted in an Saf-O positive area that was nearly 3-fold more than that of the control group at 27 weeks ⁶¹ . In addition, enriched type II collagen was observed; however, the majority of it was localized in the subchondral bone region rather than the cartilage layer. In another study using a rabbit OCD model, synovial MSCs were delivered either via IA injection or through a local adherent method ⁶² . Although more cartilage matrix was observed in the IA group at an early time point, the defects were filled with fibrous tissue and showed minimal GAGs formation by a late timepoint. Magnesium has been shown to promote not only cell adhesion but also the anabolic synthesis of GAGs ⁶³ . Previous studies have reported that Mg²⁺ concentrations within the range of 2–5 mM exhibit significant effects on GAGs production and cartilage regeneration, suggesting this range as an optimal concentration for such biological activities ⁶⁴ . In an additional in vivo study, the combination of MFx with magnesium injection demonstrated that a Mg²⁺ injection dose of 0.5 mol/L was the most effective in enhancing MFx-mediated cartilage repair ⁶⁵ . However, the study employed a multi-injection protocol (three times) and showed a type II collagen–positive rate of approximately 50% at the late time point. Although differences in animal species and defect models may have influenced the outcomes, the OCD model is generally expected to produce superior regenerative effects ⁶⁴ . In our study of the rat FTD model, the smumf cells group showed a 4.76-fold increase in GAGs synesis compared to the control group. Notably, the smumf cells + Mg²⁺ group showed a significant 14.0-fold increase at 4 weeks, indicating a pronounced early enhancement in cartilage matrix synthesis. In this study, a relatively small number of cells (1 million smumf cells delivered via a single injection) was sufficient to achieve effective adhesion at the defect site, leading to increased GAG synthesis at early time points and a marked improvement in type II collagen formation (69.52%) within the cartilage region. These findings indicate that a single IA injection of smumf cells, when combined with Mg²⁺ as a carrier, further improves the quality of the regenerated tissue.

The STR analysis revealed that the origin of the cells from regenerated tissue was not human-derived, consistent with the findings reported in previous references ⁶⁶ . Studies have demonstrated that positive effects of MSCs come from the paracrine mechanism by secretion of cytokines and growth factors ⁶⁶ . Consistent with these findings, our study and several others observed that although the injected smumf cells gradually disappeared over time, the local environment modulated by the smumf cells may continue to exert tissue regenerative effects. On the other hand, the anti-inflammatory effects are also involved in tissue regeneration. Previous studies have shown that the conditioned medium from smumf cells contains elevated levels of bioactive factors related to inflammation, such as interleukin-4 (IL-4, 28.58-fold increase) and hepatocyte growth factor (HGF, 16.78-fold increase), compared to the control ¹¹ . Moreover, MSCs possess the ability to modulate the local inflammatory environment by inducing macrophage recruitment and polarization ⁶⁷ . This is particularly relevant, as inflammation and cartilage damage are accompanied by the degradation of type II collagen, with these localized lesions being primarily mediated by macrophages ⁶⁸ . Although the mechanisms by which smumf cells modulate IA macrophages to improve the inflammatory microenvironment remain unclear, further investigation is required. Taken together, the exact mechanism by which the injected and adhered smumf cells form articular cartilage would be through endogenous regeneration, not by direct differentiation of the injected cells.

There are several limitations to this study. First, we primarily focused on early cell adhesion and did not perform long-term tracking of the fate of the injected MSCs. Although STR analysis was performed at the final regeneration time point to detect the presence of human-derived gene, this approach was insufficient to fully elucidate the contribution of the injected smumf cells to the regenerated tissue, and long-term tracking and more sensitive fate mapping approaches will be needed for further studies. Second, as mentioned earlier that inflammation modulation also plays a crucial role in tissue regeneration, our study did not directly assess the anti-inflammatory effects of smumf cells. Comparative analyses involving synovial tissue or synovial fluid were not performed. These concerns need to be addressed through further studies to comprehensively evaluate both the regenerative mechanisms and the immunomodulatory effects of MSCs prior to considering their clinical application. Third, future studies should incorporate functional evaluations to provide a more comprehensive understanding of therapeutic efficacy. Finally, because the study used a small-animal preclinical model, validation in larger animal models will be required before translation to clinical application.

Conclusion

In conclusion, this study demonstrates that a single IA injection of smumf cells combined with Mg²⁺ improves cell homing and functional adhesion to defect site, leading to superior macroscopic and histological outcomes compared to smumf cells treatment alone.