Fabrication of Injectable Chitosan-Chondroitin Sulfate Hydrogel Embedding Kartogenin-Loaded Microspheres as an Ultrasound-Triggered Drug Delivery System for Cartilage Tissue Engineering

Ultrasound-responsive microspheres (MPs) derived from natural polysaccharides and injectable hydrogels have been widely investigated as a biocompatible, biodegradable, and controllable drug delivery system and cell scaffolds for tissue engineering. In this study, kartogenin (KGN) loaded poly (lactide-co-glycolic acid) (PLGA) MPs (MPs@KGN) were fabricated by premix membrane emulsification (PME) method which were sonicated by an ultrasound transducer. Furthermore, carboxymethyl chitosan-oxidized chondroitin sulfate (CMC-OCS) hydrogel were prepared via the Schiff’ base reaction-embedded MPs to produce a CMC-OCS/MPs scaffold. In the current work, morphology, mechanical property, porosity determination, swelling property, in vitro degradation, KGN release from scaffolds, cytotoxicity, and cell bioactivity were investigated. The results showed that MPs presented an obvious collapse after ultrasound treatment. The embedded PLGA MPs could enhance the compressive elastic modulus of soft CMC-OCS hydrogel. The cumulative release KGN from MPs exhibited a slow rate which would display an appropriate collapse after ultrasound, allowing KGN to maintain a continuous concentration for at least 28 days. Moreover, the composite CMC-OCS@MPs scaffolds exhibited faster gelation, lower swelling ratio, and lower in vitro degradation. CCK-8 and LIVE/DEAD staining showed these scaffolds did not influence rabbit bone marrow mesenchymal stem cells (rBMMSCs) proliferation. Then these scaffolds were cultured with rBMMSCs for 2 weeks, and the immunofluorescent staining of collagen II (COL-2) showed that CMC-OCS hydrogel embedded with MPs@KGN (CMC-OCS@MPs@KGN) with ultrasound had the ability to increase the COL-2 synthesis. Overall, due to the improved mechanical property and the ability of sustained KGN release, this injectable hydrogel with ultrasound-responsive property is a promising system for cartilage tissue engineering.


Introduction
Injuries to articular cartilage are a common health problem, causing pain and disability in all age groups [1]. Due to the lack of vasculature, neural, and lymphatic network and low chondrocyte density, the regenerative capacity of articular cartilage after damage or disease is limited [2]. The field of tissue engineering holds great promise for the cartilage repair and sustained release and evaluate the systems for promoting directed chondrogenic differentiation of bone marrow mesenchymal stem cells in vitro.

Characterizations of PLGA MPs@KGN
The PLGA MPs with and without KGN loading were prepared by the premix membrane emulsification method. They were consistently spherical in shape and exhibited a discrete and smooth surface without any pinholes or cracks. Both MPs with and without KGN displayed an obvious loss after ultrasound treatment for 2 and 5 min in structure showing an evident collapse. After soaking in phosphate buffered saline (PBS) for one month, obvious shrinkage could be seen on the surface of the MPs without cracks ( Figure 1A,B). In order to effectively control the KGN release rate, we used ultrasonic treatment for 2 min each time. As shown in Figure 1C, histograms represented the size distribution of MPs. The diameter of MPs was found to be in the range of 9.2-11.8 μm. Additionally, the medium sizes of MPs with and without KGN were 10.1 μm. In general, owing to their high surface As shown in Figure 1C, histograms represented the size distribution of MPs. The diameter of MPs was found to be in the range of 9.2-11.8 µm. Additionally, the medium sizes of MPs with and without KGN were 10.1 µm. In general, owing to their high surface area to volume ratio of the larger size of MPs, the drug loading rate would be increased [35]. The KGN loading content was measured through UV spectrophotometry at 279 nm and Pharmaceutics 2021, 13,1487 4 of 15 the concentration was calculated by the standard curve. The encapsulation efficiency and drug loading capacity of PLGA MPs were (36.02 ± 0.78)% and (0.16 ± 0.05)%, respectively. Figure 1D showed the KGN release behavior from MPs, which exhibited a low burstrelease within the initial 24 h. Once exposed to ultrasound, the KGN could be quickly released with a clear burst-release of KGN each time. Figure 2A showed a schematic illustration of the cross-linked reaction in the CMC-OCS hydrogel. CMC could react with oxidized chondroitin sulfate (OC) through a Schiff base reaction to form a stable structure. KGN was encapsulated into PLGA MPs which were mixed in hydrogels. FTIR was used to get further insight into the molecular organization of the polysaccharide derivatives and cross-linked hydrogels ( Figure 2B). Compared with chondroitin sulfate (CS), the spectrum of OCS showed an absorption of 1715.3 cm −1 which was assigned to the aldehyde groups. The spectrum of CMC-OCS hydrogel was similar to that of CMC in which the peak at 1241.9 cm −1 represented the C-O stretching vibrations. In addition, the disappearance of peak at 1715.3 cm −1 indicated that OCS was involved in the gelation through Schiff base reaction. The chemical structure of CS, OCS, and CMC are shown in Figure 2C-E. The hydroxyl groups of CS were oxidized into aldehyde groups of OCS using the sodium periodate.

Synthesis of Hydrogel Scaffolds
Pharmaceutics 2021, 13, x 4 of 15 area to volume ratio of the larger size of MPs, the drug loading rate would be increased [35]. The KGN loading content was measured through UV spectrophotometry at 279 nm and the concentration was calculated by the standard curve. The encapsulation efficiency and drug loading capacity of PLGA MPs were (36.02 ± 0.78)% and (0.16 ± 0.05)%, respectively. Figure 1D showed the KGN release behavior from MPs, which exhibited a low burstrelease within the initial 24 h. Once exposed to ultrasound, the KGN could be quickly released with a clear burst-release of KGN each time. Figure 2A showed a schematic illustration of the cross-linked reaction in the CMC-OCS hydrogel. CMC could react with oxidized chondroitin sulfate (OC) through a Schiff base reaction to form a stable structure. KGN was encapsulated into PLGA MPs which were mixed in hydrogels. FTIR was used to get further insight into the molecular organization of the polysaccharide derivatives and cross-linked hydrogels ( Figure 2B). Compared with chondroitin sulfate (CS), the spectrum of OCS showed an absorption of 1715.3 cm −1 which was assigned to the aldehyde groups. The spectrum of CMC-OCS hydrogel was similar to that of CMC in which the peak at 1241.9 cm −1 represented the C-O stretching vibrations. In addition, the disappearance of peak at 1715.3 cm −1 indicated that OCS was involved in the gelation through Schiff base reaction. The chemical structure of CS, OCS, and CMC are shown in Figure 2C-E. The hydroxyl groups of CS were oxidized into aldehyde groups of OCS using the sodium periodate.

Characterization of Hydrogel Scaffolds and MPs-Incorporated Scaffolds
The hydrogel and hydrogel@MPs were easy to be extruded from a medical syringe ( Figure 3A). After mixing with MPs, colorless and transparence hydrogel scaffolds

Characterization of Hydrogel Scaffolds and MPs-Incorporated Scaffolds
The hydrogel and hydrogel@MPs were easy to be extruded from a medical syringe ( Figure 3A). After mixing with MPs, colorless and transparence hydrogel scaffolds became white and opaque ( Figure 3B,C). According to the cross-sectional SEM photographs, both hydrogel and hydrogel@MPs exhibited an interconnected porous 3D framework which enabled 3D culture for cell activity, and the MPs were distributed evenly in the hydrogel to form a composite scaffold with slightly rougher internal structure than the hydrogel scaffold without MPs which might facilitate cell adhesion ( Figure 3D). Furthermore, slicing two full hydrogels into halves and reassembling them with each other after exchange could intuitively demonstrate the healing capability of broken hydrogel ( Figure 3E). The high adhesiveness of the materials enables instant reconnection. After 15 min of healing, an ambiguous boundary at the healed interface was exhibited by the assembly material, which could maintain the integrity of the healing interface despite subjected to external tensile strength. As a widely used hydrogel material, diverse benefit for a scaffold could be brought by the stable self-healing effect (e.g., injectability, flexible delivery, and adaptability) [36][37][38].
Pharmaceutics 2021, 13, x 5 of 15 became white and opaque ( Figure 3B,C). According to the cross-sectional SEM photographs, both hydrogel and hydrogel@MPs exhibited an interconnected porous 3D framework which enabled 3D culture for cell activity, and the MPs were distributed evenly in the hydrogel to form a composite scaffold with slightly rougher internal structure than the hydrogel scaffold without MPs which might facilitate cell adhesion ( Figure 3D). Furthermore, slicing two full hydrogels into halves and reassembling them with each other after exchange could intuitively demonstrate the healing capability of broken hydrogel ( Figure 3E). The high adhesiveness of the materials enables instant reconnection. After 15 min of healing, an ambiguous boundary at the healed interface was exhibited by the assembly material, which could maintain the integrity of the healing interface despite subjected to external tensile strength. As a widely used hydrogel material, diverse benefit for a scaffold could be brought by the stable self-healing effect (e.g., injectability, flexible delivery, and adaptability) [36][37][38].  The pore sizes of CMC-OCS, KGN-incorporated CMC-OCS (CMC-OCS@KGN), PLGA MPs@KGN-embedded CMC-OCS (CMC-OCS@MPs@KGN) were 33.03 ± 1.19 µm, 32.71 ± 1.20 µm, and 24.80 ± 1.33 µm, respectively ( Figure 4A). Figure 4B shows the porosities of CMC-OCS, CMC-OCS@KGN, CMC-OCS@MPs@KGN as 84.60 ± 5.285%, 82.20 ± 3.43%, and 72.2 ± 3.49%, respectively. The interconnected porous structures can transport nutrients and metabolic waste. The gelation times of CMC-OCS, CMC-OCS@KGN, and CMC-OCS@MPs@KGN were 221.80 ± 12.07 s, 233.60 ± 12.60 s, and 264.60 ± 25.18 s, respectively ( Figure 4C). By introducing solid MPs, gelation time was increased significantly. According to the stress-strain curve, the compressive elastic modulus of CMC-OCS and CMC-OCS@KGN hydrogels were 4.90 ± 0.68 kPa and 5.78 ± 0.62 kPa, while hydro-gel@MPs showed a significantly larger compressive modulus (23.40 ± 1.34 kPa), which indicated that mixing with MPs could greatly improve the mechanical properties for the cartilage repair ( Figure 4D). The pore sizes of CMC-OCS, KGN-incorporated CMC-OCS (CMC-OCS@KGN), PLGA MPs@KGN-embedded CMC-OCS (CMC-OCS@MPs@KGN) were 33.03 ± 1.19 μm, 32.71 ± 1.20 μm, and 24.80 ± 1.33 μm, respectively ( Figure 4A). Figure 4B shows the porosities of CMC-OCS, CMC-OCS@KGN, CMC-OCS@MPs@KGN as 84.60 ± 5.285%, 82.20 ± 3.43%, and 72.2 ± 3.49%, respectively. The interconnected porous structures can transport nutrients and metabolic waste. The gelation times of CMC-OCS, CMC-OCS@KGN, and CMC-OCS@MPs@KGN were 221.80 ± 12.07 s, 233.60 ± 12.60 s, and 264.60 ± 25.18 s, respectively ( Figure 4C). By introducing solid MPs, gelation time was increased significantly. According to the stress-strain curve, the compressive elastic modulus of CMC-OCS and CMC-OCS@KGN hydrogels were 4.90 ± 0.68 kPa and 5.78 ± 0.62 kPa, while hydrogel@MPs showed a significantly larger compressive modulus (23.40 ± 1.34 kPa), which indicated that mixing with MPs could greatly improve the mechanical properties for the cartilage repair ( Figure 4D).  The rheological results further demonstrated that the storage modulus (G ) and loss modulus (G ) of CMC-OCS and CMC-OCS@KGN hydrogel scaffolds remained constant within the strain range of 0.01% to 5%, and then it was detected that the value of G was lower than G , indicating that the structure was damaged under large deformations ( Figure 4E). Moreover, the structures of CMC-OCS@MPs@KGN hydrogel scaffolds were destroyed under smaller deformations (3%). Furthermore, all the hydrogel scaffolds exhibited similar shear thinning behaviors ( Figure 4F) confirming that the hydrogels can be smoothly squeezed out of the syringe. The incorporation of MPs improved the modulus and viscosity of hydrogel scaffolds, resulting in better mechanical properties. Figure 4G showed that equilibrium swelling ratio of all hydrogel and hydrogel@MPs could be reached in PBS at 37 • C in 96 h. It was noticeable that compared with CMC-OCS and CMC-OCS@KGN hydrogels, hydrogel@MPs exhibited significantly lower swelling ratio, which was consistent with another study (21), since MPs may take up a portion of space of porous hydrogel. The in vitro degradation was monitored as a function of incubation time in PBS at 37 • C. Hydrogel@MPs had an influence on weight loss as shown in Figure 4H. After incubation for 14 days, the degradation ratio of CMC-OCS and CMC-OCS@KGN hydrogels (43.73 ± 6.52% and 44.76 ± 5.91%) was higher than that of hydrogel@MPs (35.42 ± 3.65%). The values of the three groups were 43.73 ± 6.52%, 44.76 ± 5.91%, and 35.42 ± 3.65% after 21 days of culture, which was consistent with swelling and compressive modulus properties. The KGN was loaded into the hydrogel by physical absorption or MPs encapsulation. The cumulative release behavior of KGN in CMC-OCS@KGN was significantly different from that in CMC-OCS@MPs@KGN. In the CMC-OCS@KGN group, KGN showed obvious burst release behavior at the initial release stage, and the burst release rate was 41.83 ± 3.70% and the release rate was 94.35 ± 3.76% at 28 days. In the hydrogel@MPs group, KGN showed lower burst release behavior at the initial release stage. The cumulative release rate of 12 h and 30 days was 20.94 ± 2.64% and 44.51 ± 2.77%, respectively. Upon exposure to ultrasound, KGN showed a burst release from MPs@KGN and CMC-OCS@MPs@KGN. The cumulative release rate of 28 days increased to 74.71 ± 4.03% and 72.52 ± 1.75%, respectively ( Figure 4I).

The Cytocompatibility of Hydrogel Scaffolds and Scaffolds Incorporated with MPs In Vitro
Cell compatibility of scaffolds was evaluated by cell viability and cell proliferation by CCK-8 and LIVE/DEAD staining as shown in Figure 5. The results showed that the cell viability ratio of all groups was higher than 90%, which indicated that all kinds of scaffolds and extra ultrasound had good cell compatibility ( Figure 5A). No statistical differences were observed among the four groups. The rabbit BMMSCs (rBMMSCs) proliferation on scaffolds was analyzed by CCK-8 ( Figure 5B). There was no statistical differences among four groups at 1, 3, 5, and 7 days. When it came to the 10th day, cell proliferation slowed in the KGN release group, and there were statistical differences among the four groups. Since the proliferation rate of chondrocytes was slower than that of rBMMSCs, we speculated that the KGN-released groups promoted the chondrogenic differentiation of rBMMSCs. LIVE/DEAD assay ( Figure 5C) was performed after seeding cells onto the scaffolds for 7 days. Live/dead assay showed a large area of living cells (green) and few dead cells (red) in these four groups, indicating good biocompatibility of the scaffold.

Bioactivity of Scaffolds In Vitro
To evaluate the effect of the scaffolds or scaffolds incorporating with ultrasound on differentiation of rBMMSCs in vitro, they were stained with collagen II (COL-2) immunohistochemistry after 14 days. As shown in Figure 6, after 14 days of incubation in vitro, COL-2 gradually increased and COL-2 in CMC-OCS@MPs@KGN/ultrasound group was significantly higher than that of CMC-OCS@MPs@KGN and CMC-OCS@KGN groups ( Figure 6B). The CMC-OCS group barely showed any staining for COL-2. The results showed that controlled KGN-released through ultrasound significantly increases the expression of COL-2, which might be helpful for the repair of cartilage.

Bioactivity of Scaffolds In Vitro
To evaluate the effect of the scaffolds or scaffolds incorporating with ultrasound on differentiation of rBMMSCs in vitro, they were stained with collagen II (COL-2) immunohistochemistry after 14 days. As shown in Figure 6, after 14 days of incubation in vitro, COL-2 gradually increased and COL-2 in CMC-OCS@MPs@KGN/ultrasound group was significantly higher than that of CMC-OCS@MPs@KGN and CMC-OCS@KGN groups ( Figure 6B). The CMC-OCS group barely showed any staining for COL-2. The results showed that controlled KGN-released through ultrasound significantly increases the expression of COL-2, which might be helpful for the repair of cartilage.

Preparation of PLGA MPs@KGN
The PLGA MPs@KGN were prepared via an oil-in-water (O/W) emulsion method combined with premix membrane emulsification. Briefly, PLGA and KGN were dissolved in component solvent of dichloromethane (DCM) and dimethyl sulfoxide (DMSO) (50:1 for volume ratio). Under mechanical stirring, the coarse double emulsion was prepared by pouring the solvent into an aqueous PVA solution immediately. Then, under a certain

Preparation of PLGA MPs@KGN
The PLGA MPs@KGN were prepared via an oil-in-water (O/W) emulsion method combined with premix membrane emulsification. Briefly, PLGA and KGN were dissolved in component solvent of dichloromethane (DCM) and dimethyl sulfoxide (DMSO) (50:1 for volume ratio). Under mechanical stirring, the coarse double emulsion was prepared by pouring the solvent into an aqueous PVA solution immediately. Then, under a certain nitrogen pressure, the crude double emulsion was extruded several times through the SPG membrane. Then, MPs were collected by centrifugation, washed three times with distilled water, and lyophilized.

Surface Morphology and Size Distribution Measurements of PLGA MPs
The MPs and MPs@KGN dispersion were sonicated for 2 or 5 min using an ultrasound transducer (Chattanooga Co., Chattanooga, TN, USA, Thermal index/Ti = 0.9 and Mechanical index/Mi = 1.6) which had been applied in a previous study [39]. Meanwhile, they were soaked in PBS solution (pH = 7.4) at room temperature for one month. The surface morphology and shape of PLGA MPs with or without ultrasound were specified by scanning electron microscope (SEM, FEI Quanta 200, Thermo Fisher Scientific, Waltham, MA, USA). The particle size distribution of MPs used in the experiment was characterized using a laser particle size analyzer (Zetasizer Nano ZS90, Malvern, Malvern, UK).

Measurement of KGN Loading Efficiency and Encapsulation Efficiency
The prepared 5 mg of PLGA MPs@KGN (n = 5) were dissolved in component solvent of dichloromethane (DCM) and ethanol (4:1 for volume ratio) of 5.0 mL. The KGN concentration in the solution was spectrophotometrically measured at 279 nm. The drug loading efficiency (LE) and encapsulation efficiency (EE) of the MPs were calculated using the following Equations (1) and (2)

KGN Release from Micropheres
For the release study, 5 mg of lyophilized PLGA MPs@KGN (n = 5) were placed in PBS solution (pH = 7.4, 3 mL) at 37 • C under gentle agitation. At each predetermined time point (1, 4, 7, 10, 14, 21 and 30 days), for 3 samples, the solution was sonicated for 2 min using an ultrasound transducer. Then we withdraw the supernatants of the 3 samples after being centrifuged for 2 min for further analysis. Subsequently, the MPs were redispersed in the same volume of fresh PBS by vortexing and then incubated until the next release time point.

Synthesis of OCS
OCS was synthesized according to a slightly modified reporting procedure [40]. After dissolving 5 g of CS in 100 mL of distilled water, it was reacted with 3.5 g of NaIO 4 in the dark at room temperature, and stirring was continued for 6 h. Total of 2.5 mL of ethylene glycol was added to the flask and stirred for 1 h. Distilled water was dialyzed for 3 days to remove residual periodate and ethylene glycol. A white foam material was obtained after the purified product was lyophilized. The actual aldehyde content of OCS was measured by elemental analyzer (Thermo Flash Smart, Milan, Italy), and the oxidation degree was found to be 53.6%.

Fabrication of Hydrogel Scaffolds
To prepare the CMC-OCS hydrogel, CMC (30 mg/mL) and OCS (100 mg/mL) were dissolved in PBS (pH = 7.4) and mixed at a volume ratio of 4 at room temperature. To form MPs-embedded hydrogel, PLGA MPs@KGN was added to OCS solution and OCS@MPs were mixed with CMC solution. MPs were embedded within hydrogel at a concentration of 0.6 mg/mL. The gelation time of hydrogel scaffolds were obtained by vial tilting methods [41].
Fourier transform infrared spectroscopy (FTIR, TENSOR II, Bruker, Billerica, MA, USA) was employed to confirm the chemical structure of CS, OCS, CMC as well as CMC-OCS hydrogels.

Surface Morphology of Hydrogel Scaffolds
The surface morphologies of the PLGA MPs, hydrogel, and hydrogel@MPs were investigated with SEM. Briefly, the frozen MPs, sliced hydrogel, and hydrogel@MPs were gold-coated and observed using SEM at a 3 kV accelerating voltage.

Gelation Time and Self-Healing Performance
The gelation time of different formulations of hydrogels was measured by the vial inversion method [42]. Briefly, 2 mL of hydrogel was added to a glass bottle at 37 • C for different periods of time. When the hydrogel solution no longer flowed after tilting the glass bottle, the gelation time was determined. Two adhesives were fully set in tablet shape and dyed with different colors into halves and reconnected after exchange to characterize the self-healing behavior of the hydrogel. After 15 min of healing, tensile stress was manually applied by reconnecting samples with tweezers.

Porosity Determination
The porosity of hydrogel (n = 5) and hydrogel@MPs (n = 5) was calculated through ethanol displacement method. Samples were cut into regular cylinders, and then the initial volume and weight (V 0 and W 0 ) were calculated and weighted. Then samples were saturated in ethanol at room temperature for 5 min and reweighted (W 1 ) [43]. The porosity of hydrogel could be obtained as follows: ρ represents the density of ethanol.

Rheological Test of Hydrogel Scaffolds
Different rheological characteristics of CMC-OCS, CMC-OCS@KGN, and CMC-OCS@ MPs@KGN hydrogel scaffolds were carried out on a rheometer (Thermo Haake Rheometer, Newington, NH, USA). For cyclic alternating shear strain sweep tests, with T = 30 • C, f = 0.02 Hz, the shear strain was changed successively from strain = 0.01% to strain = 6% (60 min), and the cycle was repeated for three times, and was tested under the action of 6% strain. After each phase of the test, the next phase was immediately switched without retention. At least three effective samples were tested for each hydrogel sample. For shear rate sweep tests, shear viscosity was recorded over a shear rate range of 0.004 to 0.4 s −1 .

Swelling Properties
The dry hydrogel (n = 5) and hydrogel@MPs (n = 5) of the same size weighted W d were incubated in phosphate-buffer saline (PBS, pH = 7.4) at 37 • C for certain times (0. 25, 0.5, 1, 2, 4, 6, 12, 24, 48, 72, 96, 120, 144, and 168 h). The samples were taken out from swelling medium and placed between two sheets of paper to remove the excess water on surface and weighted W s . The swelling ratio was calculated as below: (4)

Compressive Testing
The compressive testing of hydrogel (n = 5) and hydrogel@MPs (n = 5) were calculated by a universal material testing machine of Instron 3365 (Instron Co., Norwood, MA, USA). Samples were cut into cylinders (diameter 15 mm and height 7.5 mm) and the compressive testing was carried out with a beam velocity of 3 mm/min. The compressive elastic modulus was calculated according to the stress-strain curve [44].

KGN Release Profile from Scaffolds
To examine the KGN release profile, hydrogel@MPs (5.0 mg, n = 5) were placed in PBS solution (pH = 7.4, 1.0 mL) and agitated continuously at 37 • C. At specific time intervals, PBS solution was obtained and an equivalent amount of fresh PBS was added. The concentration of KGN was measured by UV spectrophotometry at 279 nm.

In Vitro Degradation
The degradation of hydrogel and hydrogel@MPs (n = 5) in vitro was examined through weight loss in PBS solution during 21 days. Initial weight of lyophilized samples (W 0 ) were recorded and immersed in PBS solution at 37 • C for certain time (1, 4, 7, 14, and 21 days). Then the samples were taken out, lyophilized and weighted again (W 1 ). The degradation ratio was defined as follows: 3. 16. In Vitro Studies of Scaffolds 3.16.1. In Vitro Cytotoxicity Cytotoxicity of hydrogel was measured using Cell Counting Kit-8 (CCK-8, Dojindo, Kumamoto, Japan) assay system at 12, 24, and 48 h post-treatment by contacting the extracts of hydrogel. The rBMMSCs (P2) (2 × 10 3 /100 µL/well) were seeded in 96-well microplates and then incubated at 37 • C in 5% CO 2 for 12 h to obtain a monolayer of cells. Then hydrogel extracts (n = 3) or PBS (10 µL) were added to each well and incubated for predetermined time. After 12, 24, and 48 h of incubation, the cell culture medium was removed and then 100 µL of fresh culture medium and 10 µL of CCK-8 were added to the 96 wells for 2 h. Subsequently, 110 µL was transferred to a 96-well plate, and the absorbance was read at 450 nm on a microplate reader (Thermo, Waltham, MA, USA). Cell viability (%) was calculated using the following Equation: The data represented the mean of six independent experiments and were expressed as means ± SD. Cell viability of <70% was considered cytotoxic [41].

Cell Proliferation on Scaffolds
Cell proliferation was measured with the CCK-8 assay kit. The rBMMSCs (P2) were cultured with hydrogel and hydrogel@MPs (n = 3) for 1, 3, 5, 7, and 10 days. When treatments were completed, the cell culture medium was substituted with 100 µL fresh cell culture medium and 10 µL of CCK-8, and then the cultures were incubated for 2 h at 37 • C. The absorbance (OD) value at 450 nm for each well was recorded by using a microplate reader. Cell proliferation curves were depicted according to the OD value at indicated time points.

Cell Seeding on Scaffolds
Total of 50 µL concentrated rBMMSCs solution (1.0 × 10 7 cells/mL) was seeded in lyophilized hydrogel and hydrogel@MPs (n = 3) (2 mm in thickness and 6 mm in diameter) by centrifugation [45]. After cell implantation, the scaffold was incubated for 1 h to attach cells, and then 2 mL of fresh culture medium was added.

Cell Compatibility and Distribution on Scaffolds
Cell culture medium was replaced every 3 days and subsequent to being cultured for 7 days, each scaffold was washed in PBS twice, and immersed in 500 µL of PBS with 2 mM of calcein AM and 4 mM of PI reagents (Invitrogen, Carlsbad, CA, USA) before incubation for 1 h at 37 • C. Confocal laser-scanning microscopy (Leica, Wetzlar, Germany) with excitation wavelength of 488 and 568 nm was used to detect live (green) and dead (red) cells.

Cell Bioactivity Assessment of KGN Released from Scaffolds
The cell-seeded scaffolds were fed with chondrogenic differentiation solution (Cyagen Biosciences Inc., Suzhou, China) and the solution was changed every other day. Immunofluorescent staining was used to analyze the formation of COL-2 in cartilage scaffold. After being cultured for 14 days, the scaffolds were washed with PBS and fixed with 4% paraformaldehyde at room temperature for 10 min. Then, scaffolds were blocked with 5% bovine serum albumin (BSA) for 30 min before exposure to rabbit anti-COL II primary antibody (1:200, Abcam, Cambridge, UK) overnight at 4 • C followed by secondary antibodies (anti-rabbit IgG, green, Abcam, Cambridge, UK) at 37 • C for 40 min. Then the immunofluorescence images were obtained by using confocal microscopy (Leica TCS SP8, Leica Microsystems Inc., Heidelberg, Germany). By calculating and normalizing the fluorescence intensity of COL-2 to the number of the cells where the COL-2 fluorescence was unevenly distributed in the immunofluorescence images, the amount of COL-2 was quantitatively evaluated by Image-Pro Plus software (6.0; Media Cybernetics Inc., San Diego, CA, USA) (n = 3).

Statistical Analysis
All data are presented as the mean ± standard deviation (mean ± SD). Statistical analysis among groups were calculated using a one-way analysis of variance ANOVA after testing for homogeneity of variance (test Levene's) with post hoc contrasts by least significant difference test (LSD-t). Significance thresholds were set at p < 0.05 (significant) and p < 0.01 (highly significant). Data analysis was performed using SPSS 25.0 for Windows (SPSS Inc., Chicago, IL, USA).

Conclusions
In this study, an injectable CMC-OCS hydrogel containing MPs was developed via the Schiff base cross-linking reaction and PLGA MPs@KGN were prepared through the PME method. The obtained CMC-OCS@MPs@KGN exhibited a shorter gelation time and slower weight loss than control hydrogels in vitro. Compressive elastic modulus of the hydrogel@MPs was significantly improved and beneficial to the cartilage repair. Meanwhile, the PLGA MPs@KGN could respond to ultrasound with controlled burst-release of KGN. In vitro rBMMSCs seeded in the scaffolds actually showed significant higher viability during culture time of 7 to 10 days. These results indicated the advanced strategy on fabrication of controlled KGN release which could be used in tissue-engineered cartilage.