Introduction

Articular cartilage is an avascular tissue with poor spontaneous healing potential once injured, ultimately progressing to osteoarthritis (OA)(Lou et al. 2021). The repair and restoration of articular cartilage have presented a complex challenge for researchers and clinicians. Although small defects can spontaneously be filled over time without interventions, the repaired tissue is formed with fibrocartilage-like consisting predominantly of collagen fibers, making it mechanically inferior to native hyaline cartilage (Armiento et al. 2019). A wide range of cell-based techniques are utilized for cartilage restoration in clinical practice; however, the optimal cell source accounting for donor variability and the assessment of ideal timing and delivery techniques to the intended site of action remain unclear because they are also dependent on the host microenvironment, not just the graft (Kean et al. 2013; Markov et al. 2021). Consequently, no consensus has been established about the most appropriate clinical application that should be used to repair articular cartilage. In addition, securing appropriate raw materials and establishing shipment conditions including preservation are very essential for manufacturing because manufacturing of homologous products is labor intensive, particularly cell-based products with donor variations.

Mesenchymal stem cells (MSCs) have become an attractive treatment strategy for various diseases (Jovic et al. 2022). MSCs have been proposed for cartilage tissue engineering and regenerative medicine applications, and a wealth of knowledge has been generated regarding their properties and possible implementation. The translation of cell therapies into the clinic offers considerable promise by providing safer and more effective treatments for life-threatening intractable diseases, turning from merely treating symptoms to curing them completely. MSCs have the potential to differentiate into multiple tissues, exert immunomodulatory effects, and secrete soluble factors with multiple behaviors that act on their own or surrounding cells in their microenvironment in an autocrine/paracrine manner (Aggarwal and Pittenger 2005; Sid-Otmane et al. 2020). Synovial MSCs (sMSCs) have a greater potential for proliferation and chondrogenic differentiation than do those derived from other sources; therefore, they have been given greater attention in this field (Sakaguchi et al. 2005).

A previous study developed a scaffold-free three-dimensional tissue-engineered construct (TEC) comprised of sMSCs and the self-synthesized extracellular matrix (ECM) (Ando et al. 2007). Clinical studies have verified that the autograft implantation of TEC in five patients did not have serious adverse events and improved the clinical scores up to 5 years postoperatively, suggesting amelioration of knee cartilage defects (Registration: 000008266; UMIN Clinical Trials Registry number) (Shimomura et al. 2021, 2023). However, even with this promising candidate, by elucidating the early clinical proof of concept (POC), clinical applications remain a challenging task, requiring the use of a growth medium containing fetal bovine serum (FBS), which has potential risks concerning contamination by viruses and other pathogens or susceptibility to high lot-to-lot variability (Van Raalte and Dalton 2022). Reducing variability to the maximum extent possible is indispensable because FBS has significant effects on quality control for finished products more than just safety, particularly in large-scale production. In addition, autologous cell implantation, in which the patient’s own cells are isolated, processed, and later returned to the same patient, requires processing cells separately for each patient and is unable to take advantage of efficient productivity of scale in providing treatment for numerous patients. Therefore, further improvements as next-generation product of economic benefits are needed.

So far, serum-free media have been limited to research use only; however, recently, novel serum-free media are being developed for clinical practice as well (Gottipamula et al. 2013; Schepici et al. 2022), which allows the standardization of safer and higher-performance culture conditions by avoiding the uncertainly risk, including contamination of unknown factors, uncontrollable variation of serum-used products, and inhibition of the decrease in cell expansion capacity in a few passages (Gottipamula et al. 2013). Previously, we developed two types of serum-free media (STK1 for primary and STK2 for expansion) in which human MSCs showed greater potency for proliferation and differentiation into trilineage than FBS-supplemented medium (Ishikawa et al. 2009). However, no study has yet investigated the application of three-dimensional constructs such as TECs using STK1 and STK2 media.

Long-term preservation techniques are essential for supply chain to support the progress in cell-based products (Li et al. 2019). A previous study demonstrated that the viability of umbilical cord-derived MSC sheets after preservation at 4 °C for 24 h was 80%–90% (Wang et al. 2022). However, refrigeration, simplified storage, and transport conditions cannot be fully exploited for therapeutic potential because of its inadequacy for large-scale manufacturing due to concerns about immediate loss of quality. Effective preservation methods permit the evaluation of quality, potency, and safety. Therefore, we attempted to extend the expiration date by freezing at ultralow temperatures; however, a thick three-dimensional structure was estimated to make homologous freezing difficult because of insufficient permeation of the medium. Moreover, it was necessary to consider freezing methods, such as cryopreservation, which are more labor-intensive and complex owing to the specific freezing time and temperature setting and the type of cryoprotective agent (CPA) to be utilized compared with the conditions required for refrigeration. The cryopreservation of chondrocyte sheets was achieved by vitrification, resulting in comparable to non-preserved efficacy (Tani et al. 2017). In addition, three-dimensional clumps of MSCs and ECM complex have been investigated for cryopreservation; however, they were used in rat bone marrow-derived MSCs and have distinctive shapes and sizes (Motoike et al. 2018). Despite some emerging technologies, these technological elements continue to be complicated, and thus, the manufacturing process is incomplete unless all the elements are incorporated and verified.

In this study, we appropriately combined TEC technology with serum-free media for a renewal model to develop allografts as a viable off-the-shelf construct, termed “gMSC1”, which is a guaranteed product to meet the quality standard established in preliminary studies. In addition, we examined whether long-term preservation is possible by freezing to establish technical feasibility for any patient, safety, and cost-effectiveness. Finally, this study investigated the effect of gMSC1 on cartilage repair, elucidating the mechanism of action underlying this outcome and simultaneously affecting cryopreservation.

Methods

Human synovial MSC isolation and gMSC1 preparation

All experiments were approved by the Ethics Committee of Osaka University and TWOCELLS Co. Ltd., and all patients provided written informed consent. MSCs were isolated from the synovium obtained from three human donors. After providing informed consent, all donors who required knee joint surgery and did not have an affected normal synovium were enrolled in this study. The study protocol was approved by our ethics committee on human research. Specifically, the synovium minced into small sections with scissors was seeded on a culture dish in STK1 (Kohjin Bio, Saitama, Japan) and cultured at 37 °C in a humidified 5% CO2 atmosphere. In each dish, the media were changed on days 5 and 8. After heterogeneous cell colonies were formed on day 11, adherent and outspreading cells from the synovium sections were harvested with TrypLE Select-CTS enzyme (Thermo Fisher Scientific, MA, USA) and then reseeded at 5000 cells/cm2 in STK2 (Kohjin Bio). After reaching subconfluence, sMSCs expanded as adherent cells were passaged for a further time. For gMSC1 preparation, sMSCs expanded up to passage 5 were seeded at high density (4.0 × 105 cells/cm2) in STK2, which was changed on days 3 and 5. sMSCs were developed as a sheet-like construct composed of sMSCs and ECM during high-density culture. After day 7, these constructs were carefully detached from the culture dish with pipette tips, scratching from the bottom rim; subsequently, they spontaneously formed three-dimensional tissues with active contraction. This scaffold-free construct was termed “gMSC1,” which can be prepared for various sizes by changing the seeding cell number, not the cell density, depending on the area of the culture vessel.

gMSC1 preservation

After culture, gMSC1s were divided into three groups according to the preservation conditions: Control group with fresh (nonpreservation) gMSC1s, Ref-gMSC1 group with gMSC1s stored in a refrigerator, and Fro-gMSC1 group with gMSC1s stored in an ultralow freezer. The gMSC1s were transferred to cryovial tubes containing Solulact (Lactated Ringer; Terumo, Tokyo, Japan) and stored at 10 °C in a refrigerator for 2 or 3 days (Ref-gMSC1), whereas others were transferred to cryovial tubes in custom-made, serum-free cell cryopreservation medium based on STK2, added with 10% dimethyl sulfoxide (DMSO), and stored using a control rate freezer (Strex, Osaka, Japan) set to − 1 °C/min until − 80 °C, followed by storage in a deep freezer (Nihon freezer, Tokyo, Japan) at − 80 °C within 1 month (Fro-gMSC1). Others were examined immediately after culture (fresh gMSC1). If Fro-gMSC1s were investigated in later experiments, they were thawed at room temperature (RT) by just leaving them in place without any special processing.

Weight measurement and cell viability test

Wet weight of gMSC1s prepared in 24-well culture plates was measured after draining preservation solution using a precision balance (GX-200; A&D, Tokyo, Japan). Then, gMSC1s were dissociated into single cells using 560 U/mL collagenase (Worthington Biochemical, NJ, USA) at 37 °C for 90 min with occasional vortexing. After the incubation and checking for no visible clumps, aliquots of the cell suspension were mixed with 0.4% trypan blue (Thermo Fisher Scientific, MA, USA), counted with a hematocytometer, and viability was calculated by dividing live cells by dead cells.

Flow cytometry analysis

After the cell viability test, a part of the cell suspension was washed in phosphate-buffered saline (PBS) and then resuspended in 0.5% human serum albumin in PBS. The cells were stained for 60 min at RT with antibodies against CD13, CD44, CD73, and CD90 for positive MSC markers and CD11b, CD34, CD45, and CD90 for negative MSC markers (BioLegend, CA, USA). All flow cytometric analyses were performed using a CytoFLEX Flow Cytometer (Beckman Coulter, CA, USA). Positive cells were determined as the proportion of the population with fluorescence > 99% that of the isotype control. Histograms are provided in the supplementary data (Supplementary file1).

Chondrogenic differentiation

gMSC1s prepared in 24-well culture plates were centrifuged at 1500 rpm for 5 min in 15-mL polypropylene conical tubes and then subjected to pellet culture in mesenchymal stem cell chondrogenic differentiation medium (Promocell, Heidelberg, Germany) at 37 °C in 5% CO2. The culture medium was changed every 2–3 days. The cell pellets derived from gMSC1 aggregates were harvested on days 1, 7, 14, and 21 for quantitative real-time polymerase chain reaction (qRT-PCR) analysis and only harvested on day 21 for glycosaminoglycan (GAG) assay.

qRT-PCR

After chondrogenic induction, total RNA was isolated using NucleoSpin RNA (Takara Bio, Shiga, Japan). Complementary DNA (cDNA) was then synthesized using a PrimeScript RT Reagent Kit (Perfect Real Time; Takara Bio) according to the manufacturer’s instructions. qRT-PCR was performed with cDNA using TaqMan Fast Advanced Master Mix (Thermo Fisher Scientific), primers, TaqMan probe, and QuantStudio 12 K Flex system (Thermo Fisher Scientific). Primers COL1A2 (Hs01028956_m1), COL2A1 (Hs00264051_m1), COL10A1 (Hs00166657_m1), ACAN (Hs00153936_m1), and SOX9 (Hs00165814_m1) were used. GAPDH was included as an endogenous normalization control to adjust for unequal amounts of RNA. Each sample was run in duplicate. The cycling parameters were 50 °C for 2 min and then 95 °C for 20 s, followed by 40 cycles of 95 °C for 1 s and 60 °C for 20 s × 40 cycles. Relative gene expression was analyzed using a standard curve-based method with GAPDH as the endogenous control. The results are presented as target gene expression normalized to that of GAPDH RNA. The results were subjected to statistical analysis.

Determination of GAG accumulation

Following chondrogenic differentiation, gMSC1s were digested with a papain extraction reagent for 60 min at 60 °C with occasional vortexing. The GAG content in the extraction reagent from gMSC1 was determined using a Blyscan Glycosaminoglycan Assay Kit (Biocolor, Northern Ireland, UK) as described in the manufacturer’s instructions, and the GAG (mg) content was quantified by measuring the absorbance at 650 nm. The GAG fold change showed the GAG content volume after differentiation divided by that of undifferentiation.

Analysis of the concentration of soluble factors by enzyme-linked immunosorbent assay (ELISA)

Five gMSC1s prepared under each preservation condition in 24-well culture plates were further cultured in 1.2 mL of STK2 at 37 °C and 5% CO2 in the same 15-mL polypropylene conical tubes. On day 2 after incubation, supernatants were centrifuged to remove cell debris and were collected to measure protein expression. Lysates from gMSC1 were extracted using Minute Total Protein Extraction Kit for Animal Cultured Cells and Tissues (Invent Biotechnologies Inc., MN, USA). Briefly, gMSC1s were placed in a filter cartridge and ground with a plastic rod, and a cell lysis buffer was added to the filter and continuously ground. Then, the cell lysis buffer was centrifuged, and the lysates were collected. ELISA was performed using ELISA Kits [TGF-β1 (DB100C), fibronectin (DFBN10), TSP-2 (DTSP20), TIMP-1 (DTM100), and FGF-2 (DFB50), R&D Systems. MN, USA] to determine the culture supernatant or the gMSC1 lysate. The STK2 after incubation and the lysate solution were measured as negative controls. The obtained results were subjected to statistical analysis.

gMSC1 implantation in a rat cartilage defect model

All animal experiments were performed with the approval of the Animal Care and Use Committee of TWOCELLS Co. Ltd.. Ten male nude rats (F344/NJcl-rnu/rnu) were supplied by CLEA Japan (Tokyo, Japan). Animals were maintained under a 12-h light–dark cycle and had free access to chow and water. These 12-week-old rats were anesthetized using isoflurane, and an articular cartilage defect (3 ~ 4 mm × 1.4 mm, depth 1 mm) was created in the femoral trochlear groove in both knees using a drill. Tissue fragments created by cartilage defect were cleaned up with saline. Slight bleeding from the subchondral bone was inhibited using gauze containing 10% adrenaline (Bosmin solution) by pressing against the defect to avoid spontaneous repair or insufficient engraftment. Thereafter, the gMSC1s prepared in 24-well culture plates were divided into three groups: non-application (defect only; n = 4), Ref-gMSC1 (n = 8), and Fro-gMSC1 application (n = 8). Both types of gMSC1 were implanted into cartilage defects after returning to RT from each storage device. The surgical protocol was completed by repairing the muscle attachment and closing the skin in layers after patella restoration. After the operation, all rats were returned to their cages and allowed to move freely. At 56 days post-operation, all rats were sacrificed, and specimens containing the defect sites were harvested. The femurs were excised and fixed in 4% paraformaldehyde phosphate-buffered saline and then decalcified in 20% EDTA/2Na (pH 7.0–7.4) until all minerals were removed.

Histological analysis

Following decalcification, the specimens were embedded with paraffin, cut into 3-µm sections along the axial plane, and divided into three parts (proximal, center, and distal) per defect. All sections were then stained with hematoxylin–eosin (HE) and toluidine blue (TB) according to general protocols.

Immunohistochemical evaluation

For immunohistochemical staining for SOX9, deparaffinized sections were washed in distilled water and treated with primary antibodies against SOX9 (ABE2868, Merck Millipore) at a dilution of 1:500 at 4 °C overnight. The stained sections were washed with PBS, treated with polyclonal goat anti-rabbit immunoglobulins/biotinylated (E0432, Agilent (Dako), CA, USA) for 30 min at RT and Ultra-Sensitive ABC Peroxidase Standard Staining Kit (32050, Thermo Fisher), immersed in buffer, and then counterstained with hematoxylin.

For multiplex immunohistochemistry, staining for human vimentin and type II collagen (COL 2), the tissues were conducted with an ImmPRESS Duet Double staining polymer Kit (MP-7724, Vector laboratories, CA, USA) according to the manufacturer’s instructions. Briefly, deparaffinized sections were washed, followed by antigen retrieval with 2% bovine testicular hyaluronidase (H3506, Sigma‒Aldrich, MA, USA) at RT for 60 min. After washing, the sections were incubated with BLOXALL Blocking Solution at RT for 10 min and blocked with normal horse serum to avoid nonspecific staining. Then, a primary antibody mixture was prepared with type II collagen (Kyowa Pharma Chemical, F-57) (1:100) with human vimentin prediluted solution (413541, Nichirei). The primary antibody mixture was applied to each section and incubated at RT for 60 min. Detection was then performed in a step-by-step manner using ImmPRESS Duet Reagent, ImmPACT DAB EqV Substrate and ImmPACT Vector Red Substrate.

Evaluation of repair tissue

The original semiquantitative scoring systems that modified the previously reported double staining methods for distribution and localization in the tissue was defined, and the chondrogenic differentiation capacity of gMSC1 was evaluated using this system according to the degree of intensity and range in the marginal region (human vimentin × COL2 double staining) (Table 1)(Sun et al. 2010; Mifune et al. 2013). Scoring was performed blindly, with the scorer being unaware of any processing details used in the examined specimens to prevent bias. Eventually, the highest score within three parts (proximal, center, and distal) of the defect was adopted as the degree of repair in each individual defect.

Table 1 Historical scoring system of human vimentin and type II collagen (COL2) double immunostaining
Full size table

Immunohistochemical staining of gMSC1

Fresh and Fro-gMSC1 prepared in 150-mm culture dishes were fixed in 4% paraformaldehyde, dehydrated in a gradual ethanol series, and then embedded in paraffin. The tissues were cut into 5-µm sections, deparaffinized, and rehydrated using a graded ethanol series. The sections were stained with H&E to evaluate the general pathological changes. For immunostaining of type I collagen (COL1), type II collagen (COL2), fibronectin, and hyaluronic acid-binding protein (HABP), the sections for heat-induced antigen retrieval were placed in citrate buffer (pH 6) and heated in a microwave (for COL1 and HABP). Then, 0.3% H2O2 in methanol was added to deplete endogenous peroxidase, blocked with G-Block (GenoStaff, Tokyo, Japan) and an Avidin/Biotin Blocking Kit (Vector, CA, USA), and incubated with antibodies against collagen I (ab34710, Abcam, Cambridge, UK) at 0.4 µg/mL, collagen II (NOVUS, NBP1-05169, MN, USA) at 2 µg/mL, fibronectin (ab2413, Abcam) at 0.4 µg/mL, and HABP (BC41, Hokudo, Hokkaido, Japan) at a dilution of 1:200 at 4 °C overnight, followed by incubation with anti-rabbit Ig Biotin (E0432, Dako) for collagen I and fibronectin or anti-mouse IgG biotin (Vector BA-9200) for collagen II for 30 min. Finally, peroxidase-labeled streptavidin (426062, Nichirei, Tokyo, Japan) and diaminobenzidine were added. After counterstaining with hematoxylin, all images were acquired using a microscope.

Ultrastructural observation of gMSC1

Ultrastructural analysis of fresh and Fro-gMSC1 was conducted to investigate the effect of freezing on the morphological integrity of cells and the ECM. The specimens were fixed with 1% glutaraldehyde and post-fixed with 1% OsO4. After dehydration in a graded ethanol series, they were prepared by freeze-drying with t-butanol. Finally, osmium coating was performed, and the specimens were imaged under a scanning electron microscope (JSM-7800F, JEOL, Tokyo, Japan).

Statistical analysis

For qRT-PCR and ELISA, the results were expressed as means ± standard deviations (SDs). Multiple comparisons with the control were performed, and p < 0.05 was deemed significant using Dunnett’s test utilizing JMP ver. 17.0.0 (SAS Institute Inc., NC, USA). For animal experiments, interval estimation using the Welch–Satterthwaite approximation was used to calculate 95% confidence intervals for the difference between the Ref-gMSC1 group and the non-application group and between the Fro-gMSC1 group and the non-application group, considering the different variances between implanted and non-application groups. The 95% confidence intervals were calculated utilizing SAS 9.4 software (SAS Institute Inc.). Detailed method and data were shown in the supplementary file.

Results

Characterization of gMSC1s among preservation conditions

Initially, we examined the quality of gMSC1s by comparing the preservation conditions. On day 7 after harvest, gMSC1s were characterized as elastic and sticky and contracted slightly with thick sheet-like constructs (Fig. 1a). Cell suspensions obtained from three donors (MB02-009, SYN098, and SYN107) after collagenase dissociation were counted using trypan blue, and wet weight, cell number, and viability under preservation were nearly comparable, except for donor variation (Fig. 1b–d). Cell surface markers dissociated from gMSC1 were identified as many as common MSC markers by mainly referring to the ISCT criteria (Dominici et al. 2006), suggesting that the quality of gMSC1 after preservation could be sustained stability by considering the procedure despite cryopreservation (Fig. 1e).

Fig. 1
figure 1

Characterization of gMSC1 by preservation. Gross appearance of gMSC1 after 7 days of high-density culture (a). Wet weight (b), cell number (c), and viability with enzymatic dissociation of gMSC1 (d). Values are presents as means ± SDs (n = 5). Cell surface marker expression after the enzymatic dissociation of gMSC1 (e) (n = 1). Fresh: non-preservated gMSC1, Ref: refrigerated gMSC1; Fro: frozen gMSC1

Full size image

Chondrogenic potency of gMSC1

Chondrogenic differentiation was induced by a chondrogenic medium in the pellet culture, and the response over time was measured by qRT-PCR and GAG assay. As shown in Fig. 2, the mRNA expression levels of chondrogenesis-related genes, such as COL2A1 (Fig. 2a), COL10A1 (Fig. 2b), and ACAN (Fig. 2c), gradually displayed strong induction in all donors and preservation conditions. By contrast, no expression change was found in COL1A1 (Fig. 2d), a marker of the main type of fibrocartilage, at any time point. The gene expression of SOX9, a transcription factor of chondrogenesis (Song and Park 2020), stably increased from the early developmental stage until day 7 and lasted until day 21 (Fig. 2e). The GAG content within gMSC1s after differentiation on day 21 tended to increase regardless of the preservation conditions but varied among donors (Fig. 2f), indicating the continuous differentiation of gMSC1s during chondrogenesis until day 21, and the change resembled the trend of general MSCs during chondrogenesis (Yang et al. 2012). Certain studies have firmly established the importance of soluble factors, including TGF-β1, fibronectin, TIMP-1, TSP-2, and FGF-2, in the development, differentiation, maturation, expansion, and activation of chondrogenic progenitor cells or mature chondrocytes (Watt et al. 2020; Chen et al. 2021). These are candidates for the critical quality attribute and elucidation of the mechanism of action of gMSC1. Consistent with these findings, TGF-β1 and fibronectin were expressed at extremely high levels in both the supernatants and lysates (Fig. 3a, 3b). The lysates of TIMP-1 and supernatant of FGF-2 collected from Fro-MSC1 increased compared with the other conditions; however, the other factors did not show any significant differences, although some variations were noted among donors (Fig. 3c-3f). The concentrates of each control group were both low and nearly negligible in all factors. These results indicate that gMSC1s produced effective soluble factors internally and externally from their own constructs, which may be relevant to gMSC1 treatment without modulation by preservation.

Fig. 2
figure 2

Gene expression levels of chondrogenic markers and the increase rate of glycosaminoglycan (GAG) contents of gMSC1. COL2A1 (a), COL10A1 (b), ACAN (c), COL1A1 (d), SOX9 (e), and GAG fold change (f) after differentiation relative to that of undifferentiation. Values are expressed as means ± SDs of the replicates and reported as fold change with respect to day 1 after normalization of GAPHD mRNA (n = 3). Dunnett’s multiple comparisons test was used to distinguish the level of significance based on *p < 0.05 compared with day 1. The GAG content within the construct was measured quantitatively. Fresh: non-preserved gMSC1, Ref: refrigerated gMSC1; Fro: frozen gMSC1

Full size image
Fig. 3
figure 3

Measurement of protein concentration in the supernatant (left) and lysate (right) of gMSC1 cells. The presence of TGF-β1 (a), fibronectin (b), TIMP-1 (c), TSP-2 (d), and FGF-2 (e) was revealed. Values are expressed as means ± SDs of the replicates (n = 3). Control of the supernatant indicates STK2 after incubation, and that of the lysate does the extraction buffer. Statistically significant differences were determined by analysis based on Dunnett’s test compared with the control group. * P < 0.05 Fresh: non-preserved gMSC1, Ref: refrigerated gMSC1; Fro: frozen gMSC1

Full size image

Effect of gMSC1 implantation on rat models

At 56 days following the implantation, samples containing the defect area were harvested from the three groups. Upon gross observation, no abnormalities were found in any of the animals, including normal clinical signs and increasing body weight until sacrifice, indicating that gMSC1s have no adverse systemic or localized effects. On H&E staining, defects in the non-application group were commonly filled with newly formed fibrous tissue scarring with dense cells, which represented the activation of fibroblasts underlying the subchondral bone (Fig. 4a). By contrast, Ref-gMSC1 and Fro-gMSC1 were filled with white to slightly pink sparse cell structures. Furthermore, gMSC1s integrated with the adjacent tissues completely, namely, human cells and host, that is, rat cells were mixed with each other. TB histological and SOX9 immunohistochemical staining for regenerated cartilage showed that the Ref-gMSC1 group had a positive region of cartilage-like structure that accumulated pericellular GAG-rich ECM from the middle to bottom parts of the defect, whereas the Fro-gMSC1 group was distributed in a small region of the middle defect (Fig. 4b, 4c). Notably, gMSC1s integrated with adjacent tissue completely and mixed with human cells (donor) and rat cells (host) in both implantation groups; however, this could not be confirmed in the non-application group. To investigate whether gMSC1s per se contributed to cartilage formation, human-specific vimentin (red) was used to detect gMSC1s and collagen type II (COL2) (brown) was used as a cartilage marker with double immunostaining. Expectedly, Ref-gMSC1s were detectable in large merged red and brown regions in the defect, and Fro-gMSC1s were found in similarly detectable merged regions. By contrast, non-application did not result in a reddish human cell region; thus, human cells did not exist (Fig. 4d). Semiquantitative scoring of the merged regions of human vimentin and COL2 double staining was blinded. The mean of Ref-gMSC1 (1.0 ± 0.8) was slightly higher than that of Fro-gMSC1 (0.8 ± 0.5); however, the 95% CI for the difference between the two gMSC1 groups was comparable to that of the non-application group (Table 2, supplementary file1), elucidating that gMSC1 can differentiate into chondrocytes in vivo and that its quality was not changed by preservation.

Fig. 4
figure 4

Histological examination of the regenerated tissue in all groups 56 days after the operation. H&E (a), toluidine blue (b), SOX9 (c), and human vimentin (red)/COL2 (brown) (d). Scale bars: 1 mm

Full size image
Table 2 Overlap degree of human vimentin and COL2 immunostaining
Full size table

Characterization of the gMSC1 structure

The expression of MSC and chondrogenic markers within Fro-gMSC1s was examined by immunostaining and SEM observations and compared with that in fresh gMSC1s. COL1 expression was slightly detectable (Fig. 5a), and COL2 expression was weakly positive overall (Fig. 5b). In particular, fibronectin and HABP were noticeably positive in both types of gMSC1s (Fig. 5c, 5d). The distribution and expression levels of Fro-gMSC1s in MSCs and cartilage marker proteins were comparable to those of fresh gMSC1s. These results confirmed that the gMSC1s maintained in undifferentiated conditions were comparable to those of sMSCs, which have a high potential for chondrogenesis even under high-density culture in STK2.

Fig. 5
figure 5

Morphological analysis of gMSC1. Immunohistochemical and ultrastructural images of fresh gMSC1 and frozen gMSC1 (Fro-gMSC1). Sections were stained with COL1 (a), COL2 (b), fibronectin (c), hyaluronic acid-binding protein (d), and scanning electron microscopy (SEM) images (e). Blue circles indicate the cell. Scale bars: 100 µm (a)–(d); 10 µm (e)

Full size image

Ultrastructural analysis utilizing SEM revealed that fresh gMSC1s and Fro-gMSC1s had similar distribution architectures, with distinctive elongated cells (Fig. 5e). The shapes of both gMSC1 types showed spindle-shaped nuclei that were extended to form distinctive ECM and a three-dimensional reticular and collagen fibrous appearance surrounding their cells resembling MSCs. They did not closely attach, i.e., cell–cell adhesion like in epithelial cells, but were largely embedded into the abundant ECM. Cavities created by lost cell traces were observed partly in both conditions; therefore, Fro-gMSC1s were not affected by cryopreservation, and precision structures affecting performance were not damaged.

Discussion

Based on the previous study on the preparation of TEC technology, we developed a novel, advanced scaffold-free TEC and explored the long-term preservation method. We identified the benefits of gMSC1 treatment from some quality tests and potency as chondrogenic differentiation potential, both in vitro assay using chondrogenic media and in vivo assay in the rat cartilage defect model. Long-term preservation methods of gMSC1s were resolved using a custom-made cryopreservation medium and a control rate freezer controlled by − 1 °C/min. In clinical practice, the unique characteristic of gMSC1 has meaning for practicality where implantation could be performed via knee arthroscopy owing to the plasticity fitting of cartilage defects of different shapes and sizes. In addition, because of the distinctive sticky constructs, periosteal flaps, artificial/animal-origin membrane suturing, and fibrin glue sealing, which lead to the loss of cartilage quality, are not required for retaining gMSC1(Hunziker and Stähli 2008). Moreover, it enables implantation of more than two gMSC1s by adjusting the defect size because of high adhesion to adjacent tissues.

We explored the potential mechanism of action of gMSC1 against cartilage defects, and as shown in the double staining for human vimentin/COL2, it was mainly distributed in the middle to bottom of the defects in a rat cartilage defect model (Fig. 4). This implies that gMSC1 influences the injured cartilage directly by differentiation into chondrocytes. This method allows simultaneous localized evaluation of both specific markers, namely, vimentin in human mesenchymal cells and COL2 in a xenograft cartilage model. This original scoring system, defined concerning the chondrogenic differentiation based on the double staining in this study (Tables 1 and 2), could help predict efficacy; furthermore, it can be applied for other xenograft models by changing the target marker antigen or different species of animal.

A previous study described that MSC sheets facilitated integration with the host tissue within a short time (Chang et al. 2015). Considering adherence to adjacent cartilage and bone because of abundant collagen, fibronectin, and HA-related factors, such as HABP (Fig. 5), we propose the following mechanism of action of gMSC1 as a cell-based regenerative agent: immediately adheres to the adjacent cartilage tissue after implantation into the cartilage defects, differentiates into chondrocytes by secreting soluble factors while mixing with each other, accumulates cartilaginous ECM internally and externally, and eventually fills the defect with an equivalent level of normal cartilage.

Unfortunately, the fate of exogenously administered MSCs is highly controversial and has received marked attention from researchers owing to their multipotentiality (Pittenger et al. 2019; Carneiro et al. 2023). Few studies have reported that implants directly contribute to tissue repair by replacing lost tissue and cells, as were presumed (Goodman-Bacon and Nikpay 2017). As previously shown, allogeneic iPS cell-derived cartilage organoids integrated with host native cartilage and remodeled cartilage with survival in a primate model (Abe et al. 2023). This is supportive of our study; since the cell source used in the report was iPS cell-derived cartilage organoids, the process until regeneration may be slightly different from that of MSCs concerning the differentiation behavior or response to stimuli. While affected by the cell conditions, cell dose, choice of MSC donor, and particularly the cell-passage number, MSCs ultimately triggered blood responses and were then eliminated upon entering the bloodstream (Moll et al. 2012). Meanwhile, the intravenous administration of MSCs causes insufficient retention because of their rapid trapping mainly in the lung, not in the injured site, and finally disappearing in the host’s body (Schrepfer et al. 2007). Administered MSCs may work as endocrine reservoirs and confer trans-organ communication through the bloodstream by protective effects even if not homing to the injured site where the spatial gap exists with the lung (Huang et al. 2022). Instead, improving the dosage form, for example, constructs such as cell sheets, spheroids, or encapsulates, is expected to extend the duration of cell engraftment at the injured site even if transient. As normal articular cartilage is a type of hyaline featuring GAG abundance, we have identified the GAGs contents as a critical quality attribute of gMSC1 and verified the high level of GAG contents and gene expression (Fig. 2). On the other hand, other studies have focused on the therapeutic effect, which is mainly due to trophic factors from the implant materials that prompt the bioactivity process of the host tissue (Teo et al. 2023), which means that MSCs do not differentiate when they are associated with damaged tissues but participate in activating intrinsic progenitor cells responsible for regeneration by immunomodulating and secreting trophic factors (Caplan and Correa 2011). Instead of using live cells, active medical ingredients such as exosomes, particularly microRNAs or plate-rich plasma (PRP), have been reported to trigger the host’s capacity for cartilage regeneration (Mianehsaz et al. 2019; Moretti et al. 2022). We are aware that the specific factors contributing to cartilage repair have not been fully identified, as predicted by the fact that many possible candidates for multiple factors interact dynamically and complexly (Liu et al. 2022). This variance of the condition, suspension versus construct, allograft versus xenograft, and so on, makes it more difficult to understand how administered MSCs affect the cartilage. In addition, the ambiguity of mechanism effects causes more confusion because the identity of MSCs in vivo is not well defined (Kfoury and Scadden 2015). In this study, the trophic factors measured in this study, relative to cartilage development or repair, were secreted abundantly inside and outside gMSC1s (Fig. 3), indicating the possibility that these soluble factors supportively provoke intrinsic and chondrogenic progenitor cells for differentiation. TGF-β1, TIMP-1, and TSP-2 are reported to promote chondrogenesis during development in mesenchymal lineages or progenitor cells (Liao et al. 2010; Jeong et al. 2013; Ozeki et al. 2016). Growth factors FGF-2 and fibroblasts facilitate MSC expansion and act as adhesive proteins constituting the majority of the ECM of MSCs (Ando et al. 2007; Kabiri et al. 2012). gMSC1s are expected to exert the long-lasting release of the beneficial trophic factor on the defect via these signaling pathways. Defining functionalities underlying gMSC1 treatment is a key to predicting therapeutic outcomes based on clear logical evidence and definition of the manufacturing process.

The use of undefined animal serum products in MSC culturing imposes numerous safety issues and product variety; therefore it is desired to apply not only serum-free media but also animal-derived component-free media as little as possible. Our previous work showed that serum-free media, STK1 and STK2, are suitable for the proliferation and differentiation of synovial MSCs into chondrocytes, adipocytes, and osteoblasts. These advantages can be applied to gMSC1, in which low immunogenicity of residual host cells occurs for allogeneic implantation and more improving the safety (Barry et al. 2005).

Long-term preservation techniques have attracted increasing attention recently as clinical approach in three-dimensional biomaterials, including cell sheets with complex structures (Li et al. 2019; Miyamoto 2023). They are sensitive to their environment and display intrinsic variability; as a result, heterogeneity causes fluctuations in the manufacturing process, preventing the stable preservation for the long term. Refrigeration enables easy operation in terms of slowing down metabolic activity; however, it cannot be stopped fully, leading to cell death. With freezing, although > 90% viability of MSCs at a month has been shown using protective reagents (Ginis et al. 2012), the formation of ice crystals below the equilibrium freezing point is hampered by destroying cells (Rall et al. 1978). We considered the slow freezing method to be superior to vitrification considering the results of ovarian tissue cryopreservation (Lee et al. 2019; Miyamoto 2023), whose medium for vitrification appears cytotoxic because high CPA concentrations are necessary to prevent the formation of ice crystals (Lawson et al. 2011). CPAs, such as DMSO and glycerol to protect against cell damage, are widely used today (POLGE et al. 1949; Siebzehnruebl et al. 1989). In this study, the STK2-based medium maintained its quality, which may have caused gMSC1s to create an ideal environment for cell activity support systems, that is, acclimatization, expansion, and fabrication. HA, secreted from MSC abundantly, was a candidate for effective PCA (Gurruchaga et al. 2019). Unlike an additive high-molecular-weight CPAs like HA, which reported to hamper penetration into cells (Pilbauerova et al. 2022), the ECM secreted from and accumulated in gMSC1s could protect endogenous CPA. In this study, 10% DMSO was added as a CPA; however, considering safer development, a lower DMSO concentration may be available, particularly in gMSC1s.

gMSC1s represent a therapeutic strategy for operating immediately when needed and decreasing the burden on patients and offer novel pharmacological agents, whereas existing ones can only delay OA progression, except for temporary pain relievers (Peng et al. 2023). They will assist and offer a meaningful solution for various issues, i.e., how to supply the final products immediately at the time of need that are ready to use or how to stock cells derived from donors and easily maintain suitable cells as high-quality, widespread, ideally universal master cell banks in advance, that could be integrated into an innovative ecosystem in activities for bringing synergic value from the clinic to market (Banda et al. 2018). gMSC1s have crucial implications by becoming a critical driver of robust and reproducible methods for cell-based therapy. In future, the technological issues in large industrial process on a per-unit cost of production, such as reducing the variability quality, remain to be addressed, whose resolution leads to improving the therapeutic value of gMSC1 with clinical implications.

Conclusions

This study might have significant implications for clinical practice and manufacture of cartilage defects of the knee because gMSC1, composed of sMSCs prepared from serum-free media and applicable as an allograft and cryopreservation for the long term, presents unique opportunities in the absence of fundamental treatment. This unprecedented approach, considering safety and production efficiency, will lay the foundation for its wide usage.