Small-Molecule Loaded Biomimetic Biphasic Scaffold for Osteochondral Regeneration: An In Vitro and In Vivo Study

Osteoarthritis is a prevalent musculoskeletal disorder in the elderly, which leads to high rates of morbidity. Mesenchymal stem cells (MSCs) are a promising approach to promote tissue regeneration in the absence of effective long-term treatments. Small molecules are relatively inexpensive and can selectively alter stem cell behavior during their differentiation, making them an attractive option for clinical applications. In this study, we developed an extracellular matrix (ECM)-based biphasic scaffold (BPS) loaded with two small-molecule drugs, kartogenin (KGN) and metformin (MET). This cell-free biomimetic biphasic scaffold consists of a bone (gelatin/hydroxyapatite scaffold embedded with metformin [GHSM]) and cartilage (nano-gelatin fiber embedded with kartogenin [NGFK]) layer designed to stimulate osteochondral regeneration. Extracellular matrix (ECM)-based biomimetic scaffolds can promote native cell recruitment, infiltration, and differentiation even in the absence of additional growth factors. The biphasic scaffold (BPS) showed excellent biocompatibility in vitro, with mesenchymal stem cells (MSCs) adhering, proliferating, and differentiated on the biomimetic biphasic scaffolds (GHSM and NGFK layers). The biphasic scaffolds upregulated both osteogenic and chondrogenic gene expression, sulfated glycosaminoglycan (sGAG), osteo- and chondrogenic biomarker, and relative mRNA gene expression. In an in vivo rat model, histo-morphological staining showed effective regeneration of osteochondral defects. This novel BPS has the potential to enhance both subchondral bone repair and cartilage regeneration, demonstrating excellent effects on cell homing and the recruitment of endogenous stem cells.


Introduction
Due to the lack of blood vessels and perichondrium, the articular surface has limited self-healing ability [1] and a poor capacity to repair [2]. Consequently, full-thickness articular defects common in young and active individuals can lead to cartilage degeneration, osteochondral damage, and eventually to osteoarthritis (OA) [3]. OA is a significant cause of mobility problems and reduced quality of life among the elderly [4].
Recent studies have focused on reconstructing hyaline cartilage in osteoarthritis. Surgical strategies for osteochondral reconstruction have been extensively studied to restore joint architecture and activity [5]. Several methods for treating full-thickness articular damage, such as microfracture [6], mosaicplasty [7], and autologous chondrocytes implantation [8] have been attempted. However, these procedures have limitations, such

Materials and Methods
The experimental design and flow chart of this study was presented as Figure S1.

Preparation of Nano-Gelatin Fiber Embedded with Kartogenin (NGFK)
A homogeneous solution of 10% gelatin and 10 M kartogenin (KGN) was prepared at 50 • C through a magnetic stirring for 30 min. The temperature was then maintained between 45 and 70 • C using an internal heater (AREX-6; VELP Scientifica, Deer Park, NY, USA) as the solution was transferred to a 10 mL syringe (with an inner diameter of needle outlet of 1.5 mm). The solution was injected through a needle with a 20 kV potential and was applied to droplets after being connected to a high-voltage power supply (SC-PME50; Cosmi Global Co. Ltd., New Taipei City, Taiwan). A distance of 7 cm was set between the collector and capillary tip, and aluminum foil was used to cover the collector. To ensure complete solvent removal, the mats were left for at least 6 h. After 24 h of glutaraldehyde vapor crosslinking (Sigma-Aldrich), the fiber samples were rinsed in ethanol (20 times) before being dried.

Preparation of Metformin Embedded Gelatin/Hydroxyapatite Scaffold (GHSM)
In situ for mineralization of hydroxyapatite (Hap) on gelatin fibrils was achieved by co-precipitating orthophosphoric acid and calcium hydroxide. To obtain a 1:1 weight ratio of gelatin/Hap, 3.665 g of Ca(OH) 2 was added to 95 mL 3.7 wt% gelatin solution and allowed to completely dissolve. A 0.15 M solution of H 3 PO 4 (1.95 mL, at 85 • C) was added drop-wise at a rate of 0.3 mL/min, followed by continuous magnetic stirring for 24 h. The pH was adjusted to 9 to facilitate pure nano-HAp precipitation, and then further adjusted to a neutral equivalent. Metformin (MET) was added to achieve a final concentration of 50 µM. Finally, 0.1% microbial transglutaminase (mTGase; Activa, Ajinomoto, Japan) was used as a cross-linking agent to crosslink the GHSM sponge. The specimens were moved to 4 • C for 24 h, frozen at −20 • C for 24 h, −80 • C 24 h, and then lyophilized for 72 h.

Synthesis of Biphasic Scaffold and Study of Morphology
The biphasic scaffold used for osteochondral regeneration was synthesized by coupling the GHSM sponge (lower layer) and NGFK scaffold (upper layer) with 0.1% microbial transglutaminase. The samples were dehydrated by graded ethanol solutions (50% for 5 min, 75% for 5 min, 85% for 5 min, 95% for 5 min, and 100% for 10 min, repeated three times). A uniform thickness of gold was sputtered and coated on the surface of all samples, which were then observed under an electron microscope (Hitachi, Tokyo, Japan).

Identification of Crystal Phase and Functional Group
Powders were mounted onto a Ni filter of the sample holder, and an X-ray diffractometer (Rigaku Corporation, Tokyo, Japan) was scanned from 10 • to 70 • (at a speed of 2 • /min, potential: 30 kV, current: 15 mA) to determine the structures of crystal under Cu KI (* = 0.15406 nm) radiation. The X-ray diffraction (XRD) patterns were analyzed using an auto-matching model in Jade 6.0 software (Materials Data, Livermore, CA, USA), with the direction database of the international center for diffraction data (ICDD).
Fourier transform infrared (FTIR) spectroscopy was used to analyze the functional groups of the different scaffolds by detecting the dipole moment change for the specific target molecule. FTIR spectra of the scaffolds were measured using a spectrophotome-ter (Perkin Elmer, Waltham, MA, USA) at a 450-4000 cm −1 wavelength range. Scaffold characterization was compared with natural bone as a control.

In Vitro Experiment
The Institutional Review Board of National Taiwan University Hospital (NTUH IRB no. 201704005 RINA) approved the experimental protocol for use of human mesenchymal stem cells (hMSCs). During total hip-and knee-joint arthroplasty surgery, hMSCs were collected from bone marrow aspirates and mononuclear cells were isolated by Ficoll-Paque PLUS (GE Healthcare, Amersham, UK) solution. All hMSCs used in the following experiments were from passages 3 to 4.

Cell Viability
The scaffold was added to high-glucose DMEM (Sigma-Aldrich) at a concentration of 0.2 g/mL, and the extract medium was prepared after incubating at 37 • C for 24 h. Following the ISO 10993-5 standard, a water-soluble tetrazolium (WST-1) assay was performed to evaluate the cell viability of L929 cells (BCRC Strain Collection, Hsinchu, Taiwan). A density of 5 × 10 3 cells/well was seeded in 96-well plates and incubated at 37 • C for 1 d. The culture medium was removed and replaced with the prepared extract medium, and the samples and cells were incubated for 1 to 3 d. For the WST-1 assay, each well was pretreated with 10 µL reagent for 4 h. The amount of formazan formed during the incubation was determined by using a spectrophotometric reader (Sunrise ELISA plate reader, Tecan, Männedorf, Switzerland), with absorbance measured at 450 nm (reference filter at 600 nm). Cell viability was calculated using the following equation:

Cytotoxicity
The CytoTox 96 assay kit (Promega, Madison, WI, USA), which measures extracellular lactate dehydrogenase (LDH), was used to assess the cytotoxicity effect. The suspension medium was transferred to new enzymatic assay plates and the LDH substrate solution was added. After a 30 min incubation, the stop solution was added, and the absorbance of each well at 490 nm was measured by using the ELISA reader (Sunrise ELISA plate reader, Tecan, Männedorf, Switzerland). Cytotoxicity was calculated using the following equation: Cytotoxicity (%) = [(Experimental value − Negative control) × 100]/(Positive control − Negative control). (2)

Live/Dead Assay
To evaluate the live/dead status of cells within constructs, staining was performed using 4 µM calcein AM (Life Technologies, Thermo Fisher Scientific Inc., Carlsbad, CA, USA) and 4 µM propidium Iodide (PI; Life Technologies, Thermo Fisher Scientific Inc.) for 30 min. The live cells were identified by calcein AM, which emits green fluorescence upon excitation (~495 nm/~515 nm); while dead cells were stained by PI, which emits red fluorescence upon excitation (~540 nm/~615 nm).

Biochemical Analysis
The content of sulfated glycosaminoglycan (sGAG) was determined using the 9-dimethyl methylene blue chloride (DMMB) method with a Blyscan assay kit (Biocolor, Carrickfergus, UK). The amounts of cell numbers were determined by extraction of total DNA using a genomic geno plus DNA extraction miniprep system (Viogen, Taipei, Taiwan) and the Quant-iT PicoGreen dsDNA assay kit (Thermo Fisher Scientific, Waltham, MA, USA). Boneand cartilage-specific proteins were determined using a commercial ELISA kit (FineTest, Wuhan Fine Biotechnology, Wuhan, China), and the ratio of total sGAG to dsDNA was averaged across samples.

In Vivo Experiment
The animal experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals. Adult male Wistar rats (weighing 190-210 g) were provided with standard chow and water ad libitum and housed in a temperature-controlled room (at 23 • C) with a fixed 12-h light/dark cycle. The Institutional Animal Care and Use Committee (IACUC) of the National Taiwan University, School of Medicine, Taipei, Taiwan (No. 20180364), pre-approved the protocol for animal experiments.

Generation of Osteochondral Defect
To generate an osteochondral defect, a transverse medial parapatellar incision was made, and the patella was laterally dislocated. A circular hole (1 × 1 mm) was drilled in both the medial and lateral femoral condyle until bleeding from the subchondral bone was observed. The scaffolds were then implanted into one of the defective sites. After three months of implantation, the joints were harvested for histological evaluation. Buffered 4% paraformaldehyde was used to fix the joints and the specimens were demineralized, dehydrated, defatted, cleared with xylene, and then embedded in wax.

Histological Staining
All specimens were fixed using formaldehyde and decalcified with Plank-Rychlo solutions. The specimens were then divided into anterior and posterior sections, embedded in paraffin, cross-sectioned and stained on glass slides with a thickness of 5 µm. Hematoxylin and eosin (H&E) and Alcian blue (AB)/periodic acid-Schiff (PAS; Polysciences Inc., Warrington, PA, USA) staining of rat articular sections was performed for histological analysis. Examination of the regeneration of bone and cartilage in the defects was observed under a light microscope (IX71; Olympus, Tokyo, Japan). Images were visualized and captured at 40× and 400× magnification.

Statistical Analysis
Quantitative data are presented as the mean ± standard deviation (SD). One-way ANOVA was used to assess the difference between groups and the level of statistical significance was set at p < 0.05. The statistical analysis was performed using SPSS Statistics 29 (IBM, Armonk, NY, USA). The surface morphology of the biphasic scaffold (BPS) was examined using scanning electron microscopy (SEM). The images showed that the mean pore size of the gelatin/hydroxyapatite scaffold embedded with metformin (GHSM) was 80.93 ± 16.7 µm (Figure 1), and the diameter of the nano-gelatin fiber embedded with kartogenin (NGFK) was 183 ± 57 nm ( Figure 1). The biphasic scaffold (BPS) had a highly interconnected porous structure in the GHSM and multilayer structure in the NGFK, allowing distribution of optimal oxygen and nutrients throughout the scaffold.

Morphology of the Biphasic Scaffold (BPS)
The surface morphology of the biphasic scaffold (BPS) was examined using scann electron microscopy (SEM). The images showed that the mean pore size of the gelatin/ droxyapatite scaffold embedded with metformin (GHSM) was 80.93 ± 16.7 μm (Figure and the diameter of the nano-gelatin fiber embedded with kartogenin (NGFK) was 18 57 nm ( Figure 1). The biphasic scaffold (BPS) had a highly interconnected porous struct in the GHSM and multilayer structure in the NGFK, allowing distribution of optimal ygen and nutrients throughout the scaffold.

Functional Group Identification
The infrared spectrum of the gelatin/hydroxyapatite scaffold embedded with metformin (GHSM) and nano-gelatin fiber embedded with kartogenin (NGFK) was analyzed to identify their functional groups ( Figure 1). Peaks located in the 500-1100 cm −1 region were characteristic of hydroxyapatite (Hap), with the peak found at 1062 cm −1 corresponding to the asymmetric bending and stretching band of the (PO 4 ) 3− group. The peaks observed at 1650 cm −1 , 2800-2950 cm −1 , and 3430 cm −1 were characteristic peaks for the C=O group, C-H stretching, and O-H stretching of gelatin, respectively. In the FTIR spectrum for the GHSM, characteristic bands were observed at 1625 cm −1 and 1569 cm −1 (corresponding to C-N stretching of MET), a band at 3150 cm −1 (due to N-H secondary stretching of MET), and two typical bands at 3290 cm −1 and 3370 cm −1 (relative to N-H primary stretching of MET).

Biocompatibility of Biphasic Scaffold (BPS)
To assess the biocompatibility of the biphasic scaffold (BPS), L929 fibroblasts were cultured with a scaffold, and the cell viability and cytotoxicity were evaluated using WST-1 and LDH assays on days 1, 2, and 3 of culturing ( Figure 2). The viability of cells cultured with a biphasic scaffold (BPS) was similar to that of cells cultured under control conditions, indicating that the biphasic scaffold (BPS) is not toxic to L929 fibroblasts. Live/dead staining revealed no difference in the ratio of red (dead cells) to green signals (live cells) among the experimental and control groups, regardless of the concentration of the BPS extract. Therefore, we concluded that the biphasic scaffold (BPS) did not induce cytotoxicity.
GHSM, characteristic bands were observed at 1625 cm −1 and 1569 cm −1 (corresponding to C-N stretching of MET), a band at 3150 cm −1 (due to N-H secondary stretching of MET), and two typical bands at 3290 cm −1 and 3370 cm −1 (relative to N-H primary stretching of MET).

Biocompatibility of Biphasic Scaffold (BPS)
To assess the biocompatibility of the biphasic scaffold (BPS), L929 fibroblasts were cultured with a scaffold, and the cell viability and cytotoxicity were evaluated using WST-1 and LDH assays on days 1, 2, and 3 of culturing ( Figure 2). The viability of cells cultured with a biphasic scaffold (BPS) was similar to that of cells cultured under control conditions, indicating that the biphasic scaffold (BPS) is not toxic to L929 fibroblasts. Live/dead staining revealed no difference in the ratio of red (dead cells) to green signals (live cells) among the experimental and control groups, regardless of the concentration of the BPS extract. Therefore, we concluded that the biphasic scaffold (BPS) did not induce cytotoxicity.

Osteo-and Chondrogenic Differentiation Potential of the Biomimetic Biphasic Scaffold
To evaluate the osteogenic differentiation potential of hMSCs on the gelatin/hydroxyapatite scaffold embedded with metformin (GHSM), we examined the expression of osteogenic genes using qPCR after 4 weeks of culture ( Figure 3). The results showed a significant upregulation of ALP, RUNX2, SP7, SPARC, BGLAP, and COL1A1 genes in hMSCs cultured on the GHSM, indicating their differentiation into pre-osteoblasts. To assess chondrogenic differentiation potential, we examined the expression of chondrogenic genes COMP, ACAN, COL2A1, and SOX9 in hMSCs cultured on the nano-gelatin fiber embedded with kartogenin (NGFK) using qPCR (Figure 3). The results revealed a significant upregulation of these genes, indicating the differentiation of hMSCs into pre-chondroblasts. teogenic genes using qPCR after 4 weeks of culture ( Figure 3). The results showed a significant upregulation of ALP, RUNX2, SP7, SPARC, BGLAP, and COL1A1 genes in hMSCs cultured on the GHSM, indicating their differentiation into pre-osteoblasts. To assess chondrogenic differentiation potential, we examined the expression of chondrogenic genes COMP, ACAN, COL2A1, and SOX9 in hMSCs cultured on the nano-gelatin fiber embedded with kartogenin (NGFK) using qPCR (Figure 3). The results revealed a significant upregulation of these genes, indicating the differentiation of hMSCs into pre-chondroblasts.  Western blotting analysis confirmed the increased expression of osteogenic markers ON, OC, and COL1A1 and chondrogenic markers ACAN, COMP, and COL2A1 in hMSCs cultured on GHSM and NGFK, respectively ( Figure 3). Moreover, the 1,9-dimethylmethylene blue (DMMB) assay showed a significant increase in the sGAG/dsDNA ratio over 4 weeks of culture on both the GHSM and NGFK (Figure 3), indicating chondrogenic differentiation and extracellular matrix production.

Biochemical Analysis
ELISA analysis revealed that the expression levels of ON, OC, and COL1A1 were significantly upregulated compared to the control group and were also increased with the increasing culture time (Figures 4 and 5). Similarly, the expression levels of ACAN, COMP, and COL2A1 were upregulated when compared to the control group.

Biochemical Analysis
ELISA analysis revealed that the expression levels of ON, OC, and COL1A1 we significantly upregulated compared to the control group and were also increased with t increasing culture time (Figures 4 and 5). Similarly, the expression levels of ACA COMP, and COL2A1 were upregulated when compared to the control group.

Immunofluorescence Analysis
After 4 weeks of culture with the gelatin/hydroxyapatite scaffold embedded with metformin (GHSM) and nano-gelatin fiber embedded with kartogenin (NGFK), hMSCs were subjected to immunofluorescence staining. On day 28, positive staining for osteogenic and chondrogenic proteins was observed ( Figure 6). These results suggest that the biphasic scaffold (BPS) facilitates the attachment of hMSCs and the production of an osteochondral tissue construct.

Immunofluorescence Analysis
After 4 weeks of culture with the gelatin/hydroxyapatite scaffold embedded with metformin (GHSM) and nano-gelatin fiber embedded with kartogenin (NGFK), hMSCs were subjected to immunofluorescence staining. On day 28, positive staining for osteogenic and chondrogenic proteins was observed ( Figure 6). These results suggest that the biphasic scaffold (BPS) facilitates the attachment of hMSCs and the production of an osteochondral tissue construct.

In Vivo Histological Evaluation
After 12 weeks of post-implantation, the formation of osteochondral tissue in def tive rat joints was evaluated by histological staining in the control and BPS groups. In t experimental groups, new bone regeneration at the defect site was confirmed by hem toxylin and eosin (H&E) staining (Figure 7). On the other hand, the control group show incomplete trabecular bone formation, with fibrous connective tissue filling most of t defect sites. The sections from the experimental group, after BPS implantation, show positive double staining with Alcian blue (AB) and Periodic acid-Schiff (PAS) stains (F ure 7). The positive AB staining indicated glycosaminoglycans (GAGs) accumulation, i parting a blue color to acidic mucins and other carbonylated or weakly sulfated a muco-substances. This finding suggests that the healing of articular cartilage and su chondral bone was significantly better in the biphasic scaffold (BPS) group than those the control group.

In Vivo Histological Evaluation
After 12 weeks of post-implantation, the formation of osteochondral tissue in defective rat joints was evaluated by histological staining in the control and BPS groups. In the experimental groups, new bone regeneration at the defect site was confirmed by hematoxylin and eosin (H&E) staining (Figure 7). On the other hand, the control group showed incomplete trabecular bone formation, with fibrous connective tissue filling most of the defect sites. The sections from the experimental group, after BPS implantation, showed positive double staining with Alcian blue (AB) and Periodic acid-Schiff (PAS) stains (Figure 7). The positive AB staining indicated glycosaminoglycans (GAGs) accumulation, imparting a blue color to acidic mucins and other carbonylated or weakly sulfated acid muco-substances. This finding suggests that the healing of articular cartilage and subchondral bone was significantly better in the biphasic scaffold (BPS) group than those in the control group.

Discussion
The osteochondral unit experiences high pressure, frequent motion, and complex interactions with the subchondral bone, which are further compounded by the limited healing potential of articular cartilage [28]. Conservative management is the current treatment for articular defects or osteoarthritis, but no single solution has demonstrated complete and durable functional repair of osteochondral lesions [29]. The development of improved and innovative therapeutic approaches to promote osteochondral tissue regeneration is imperative due to the immense socioeconomic concern of osteochondral-related problems without suitable long-term treatment options [30]. Because cells from different sources may have different differentiation potentials, the selection of appropriate cell sources for osteochondral tissue engineering is a critical issue [31]. Although osteochondral tissue can be engineered from various cell types, mesenchymal stem cells (MSCs) are still the preferred cell type [32].
Electrospinning has emerged as a promising approach in osteochondral tissue engineering due to its ability to provide mechanical properties to the scaffolds while supporting cell growth [33]. By mimicking the fibrils embedded in the extracellular matrix of natural tissues, fibrous structures in scaffolds can replicate the morphology and scale of native tissue. The use of microfibers and/or nanofibers in scaffolds has shown promise in enhancing cell recruitment, infiltration, and differentiation, even in the absence of growth factors [34]. The small diameter of fibers produced by electrospinning provides a higher surface area to volume ratio, better tunable porosity, and enables the integration of more bioactive molecules to enhance cellular response [35].

Discussion
The osteochondral unit experiences high pressure, frequent motion, and complex interactions with the subchondral bone, which are further compounded by the limited healing potential of articular cartilage [28]. Conservative management is the current treatment for articular defects or osteoarthritis, but no single solution has demonstrated complete and durable functional repair of osteochondral lesions [29]. The development of improved and innovative therapeutic approaches to promote osteochondral tissue regeneration is imperative due to the immense socioeconomic concern of osteochondral-related problems without suitable long-term treatment options [30]. Because cells from different sources may have different differentiation potentials, the selection of appropriate cell sources for osteochondral tissue engineering is a critical issue [31]. Although osteochondral tissue can be engineered from various cell types, mesenchymal stem cells (MSCs) are still the preferred cell type [32].
Electrospinning has emerged as a promising approach in osteochondral tissue engineering due to its ability to provide mechanical properties to the scaffolds while supporting cell growth [33]. By mimicking the fibrils embedded in the extracellular matrix of natural tissues, fibrous structures in scaffolds can replicate the morphology and scale of native tissue. The use of microfibers and/or nanofibers in scaffolds has shown promise in enhancing cell recruitment, infiltration, and differentiation, even in the absence of growth factors [34]. The small diameter of fibers produced by electrospinning provides a higher surface area to volume ratio, better tunable porosity, and enables the integration of more bioactive molecules to enhance cellular response [35].
In addition to growth factors, small molecules have also emerged as potential modulators of MSC behavior. These molecules are cost-effective and can selectively alter stem cell behavior during differentiation, making them attractive options for clinical use. In this study, we aimed to promote regeneration of osteochondral defects using the small molecules kartogenin (KGN) and metformin (MET) embedded in a biphasic scaffold (BPS). The BPS is composed of a bone layer scaffold, gelatin/hydroxyapatite scaffold embedded with metformin (GHSM), and a cartilage layer, nano-gelatin fiber embedded with kartogenin (NGFK).
Johnson's first report showed that kartogenin (KGN) could induce in vitro differentiation of human mesenchymal stem cells (hMSCs) into hyaline cartilage nodules containing proteoglycans and collagen II [36]. In osteoarthritis (OA), KGN can prevent cartilage degeneration and subchondral bone changes [37]. Furthermore, KGN-incorporated scaffolds have shown excellent effects on cartilage regeneration by recruiting endogenous cells from the host without the need for cell transplantation [38]. KGN selectively upregulates the expression of chondrogenic markers such as ACAN, COL2A1, SOX9, and lubricin to promote chondrogenic differentiation [39]. Our study found similar results in our biphasic scaffold.
Metformin (MET) exerts direct osteogenic effects by activating AMP-activated protein kinase (AMPK) and promoting osteoblast differentiation, which can antagonize bone loss, maintain bone mineral density, and protect bone microarchitecture [40]. Metformin's osteogenic potential is attributed to the upregulation of RUNX2 expression via the AMPK signaling pathway in MSCs, pre-osteoblasts, and osteoblasts [41,42]. In vitro, metformin can upregulate Runx2 expression in bone marrow mesenchymal cells, increase their ALP activity, collagen synthesis, osteocalcin production, and extracellular calcium deposition [43]. Runx2 phosphorylation is critical for osteogenesis as it can stimulate the secretion of bone sialoprotein (BSP) and differentiation MSCs or pre-osteoblasts to osteoblasts [44]. Conversely, the loss of Runx2 may activate adipogenesis, which is directly related to the AMPK-mediated phosphorylation of Runx2 serine [45]. In vivo, metformin can reduce bone loss by inducing osteoblast genes, such as Lrp5 and Runx2 [46], stimulate osteoprotegerin (OPG) expression, and reduce receptor activator RANKL levels [47]. Our study also demonstrated similar osteogenic effects in vitro.
To ensure clinical safety, autologous human serum is preferable to fetal bovine serum (FBS) for use in musculoskeletal tissue engineering. The most favorable cell source for cartilage regeneration are endogenous mesenchymal stem cells (MSCs). In this study, a novel biphasic scaffold (BPS) showed promising results in enhancing both subchondral bone repair and cartilage regeneration. In vitro, MSCs adhered, proliferated, and differentiated on the gelatin/hydroxyapatite scaffold embedded with metformin (GHSM) and nano-gelatin fiber embedded with kartogenin (NGFK) layers. In vivo, after implantation of the biphasic scaffold (BPS) [which contains both kartogenin (KGN) and metformin (MET)] into the sites of osteochondral defects without exogenous cell seeding, the biphasic scaffold (BPS) promoted the complete regeneration of osteochondral defects in rats. This study effectively enabled the recruitment of endogenous bone marrow MSCs for osteogenesis and chondrogenesis, facilitating osteochondral regeneration. However, the in vitro drug release and degradation profile was not evaluated in this study and the relative time-related osteogenic and chondrogenic gene expression effects were not analyzed; the possible timerelated benefit from different components may have been overlooked. In this study, we used an empty osteochondral defect (without control scaffold) as a control; another sham control with a control scaffold can be used to validate the role of small molecules in osteochondral regeneration. Although there have been great advances in musculoskeletal tissue engineering, current clinical treatments still have drawbacks in reproducing hyaline cartilage with a full function [48]. The gene-activated cell-free strategy offers a better scaffold for osteochondral reconstruction. In this study, the metformin/kartogenin-based cell-free osteochondral scaffold has been demonstrated to completely restore osteochondral defects. By developing an electrospun biphasic scaffold capable of mimicking the extracellular matrix, better tissue maturation and clinical outcomes can be achieved.

Conclusions
In conclusion, the composite biomimetic biphasic scaffold (BPS) demonstrated excellent performance in repairing osteochondral defects in rats. This scaffold, based on smallmolecule drugs and an extra-cellular matrix (ECM), provides a micro-physiological joint organoid model system for studying spatiotemporal responses across the osteochondral interface. By using small molecules, this study not only demonstrated in vitro expansion and differentiation of endogenous stem cells but also direct in vivo modulation of stem cells. The biphasic osteochondral scaffold was spatially fabricated with biochemical cues to guide specific functional tissue regeneration. Site-specific delivery of inducible and tunable gene signals provides spatial control over both the phase of osteochondral composition and the remodeling. The biphasic scaffold (BPS) showed excellent biocompatibility and enabled rapid osteo-and chondro-neogenesis, further supporting its potential for clinical application. Based on our findings, the regeneration of osteochondral defects may be more feasible when small molecules rather than growth factors are used. Despite the promising results of this study, small animal models may not be the most suitable models for assessing clinical outcomes, and investigating the biphasic scaffold (BPS) in large animal studies is imperative in the near future.