Preparation and Characterization of ECM-mimicking RSF small pore size scaffolds and RSF Hydrogel.
In our previous work [13], we showed successful preparation of RSF scaffolds through a combination of n-butanol addition and freezing treatment, which together induced the conformational transition of RSF from a random and/or helical structure to β-sheets. Scheme 1 showed a schematic fabrication process for obtaining the RSF/CS/COL Ⅰ/HA/β-TCP scaffolds. Fig. 1a was the gross morphology of scaffolds after the freeze-drying process.
The morphology is of the ECM-added scaffolds a little varied according to the mass ratio of RSF and HA, COL Ⅰ and β-TCP. Fig. 1b-d showed SEM images of the RSF scaffold prepared by freezing n-butanal at −20 °C for 24 h. It revealed that the surface of the RSF scaffold was rough and the pore size of the 5%HA group was 41.5±8.8 μm, the 10% COL group was 70.6±15.5 μm, 1:1 β-TCP group was 20.5±9.5μm (Fig. S table1, Supporting Information). The compressive modulus of 10%COL was 1100 kPa (Fig. 1e), 5%HA was 175kPa (Fig. 1f) and 1:1 β-TCP was 900kPa (Fig. 1g). When the compressive strain was 20%, the compressive stress of the 10%COL, 1:1 β-TCP group was about 200 kPa (Fig. 1h-j) while 5% HA is 30 kPa. The porosity of HA, COL Ⅰ and β-TCP scaffolds was approximately 90% (Fig. 1km).Thus HA is used as the cartilage surface layer and COL as the middle layer from the mechanical strength analysis.
At room temperature, high concentration RSF solution (12-14 wt%) was in a sol state and very sensitive to shear. After shearing, the RSF solution quickly turned into a gel state, which had the potential to be a glue to adhere substances. Fig. 1n and 1o showed that 14 wt% RSF solution had excellent bonding performance.
In vitro chondrogenesis effects for the ECM-inspired small aperture scaffolds
In Vitro Screening Ingredients at Different Concentrations for ECM-inspired Small Aperture Scaffolds
As shown in Fig. S2, Supporting Information, natural cartilage was roughly divided into 2 layers -cartilage, and subchondral bone, so we designed a bionic gradient structure, in which 5% HA loaded with chondrocytes and 10% COL carried with BMSCs or mixed cells (number ratio, chondrocytes: BMSC=2:1) were the cartilage layer, and 1:1 β-TCP co-cultivated with BMSCs was the subchondral bone layer (Fig. S2a). GAG and COLⅡare the main ECM components of knee cartilage, thus the content of dsDNA, total GAG and collagen were detected separately after chondrocytes and BMSCs co-culturing with scaffolds for 7 d in vitro. Compared with the RSF, HA and COL groups both increased dsDNA content, among which 10% COL, 5%, 15% HA and 1:1 β-TCP groups were the most obvious (Fig. S2b-d). 10% COL, 15% HA and 1:1 β-TCP group also had an up-regulated expression in GAG content (Fig. S2e-g). At the same time, the most significant up-regulation of total collagen expression were 10% COL, 5%, 10% HA and 1:1 β-TCP groups (Fig. S2h-j).
COLⅡ, a biomarker of knee cartilage, was regarded as a detection indicator as it should be. Chondrocytes and BMSCs were seeded on scaffolds and then tested with COLⅡimmunohistochemical (IHC) staining after 1 week. From IHC results, 10% COL group was the best in aspects of the green fluorescent probe (GFP)-labeled chondrocytes number and COLⅡ(red spots) production (Fig. 2a-b). In addition, it can be seen that the different collagen groups were covered with a large number of nucleis from 3D view and magnification dates, especially 10% COL (Fig. 2a-b). Not only had 5%, 15% HA a higher COLⅡlevel, but the GFP-labeled chondrocytes reached a larger quantity (Fig. 2c-d). However, the mixed cells aggregated spherically on HA scaffolds (Fig. 2c-d), which might be due to a high viscosity of HA itself. The more HA added, the poorer formability, and which was proved by the phenomenon that 15% HA exhibited an extremely fragile feature in vitro. For porous scaffolds, the cells' nutrient supply on the scaffolds has always attracted much attention. In order to solve this issue inside the cell sphere, a HA scaffold was put as the upper layer of the cartilage gradient structure so that it received nutrition from subcutaneous and surrounding cartilage defect areas at an early date.
For clinical applications, favorable biocompatibility is an essential requirement for scaffolds [14]. The biocompatibility of the scaffolds in vitro was assessed by live/dead staining and the Cell Counting Kit (CCK-8) test. First, rabbit-derived chondrocytes and BMSCs were cultured on different scaffolds, respectively. A live/dead staining assay was performed after 7d of culture. As shown in Fig. 2e, most chondrocytes and BMSCs were alive (green fluorescence) in the HA, COL and β-TCP groups, with only a few dead cells present (red fluorescence), indicating that the ECM-inspired small aperture scaffolds could support cell adhesion and growth. Besides, the number of live cells in 1:1 β-TCP group was significantly higher than that in other groups (Fig. 3e). The natural presence of ECM components in RSF was likely to enhance cell adhesion to the scaffolds. Then, the CCK-8 assay demonstrated that cell viability compared to the control group (RSF group) was undecreased when cells co-culturing with the scaffolds for one, three, five or seven days, demonstrating the lack of cytotoxic effects (Fig. 2f). After one and three days of culture, the cell viability increased in all groups, indicating that the formulated scaffolds could promote cell proliferation.
Scanning electron microscope (SEM) results also directly clarified that chondrocytes and BMSCs not only adhered well in the pores, but also could be seen to secrete a large number of spun collagen fibers covering scaffolds' surface (Fig. S1, Supporting Information). Chondrocytes co-cultured in RSF hydrogel for 2 weeks still stretched and grew well (Fig. 2g). Taken together, these results implied that the RSF/CS/COL Ⅰ/HA, RSF/β-TCP hybrid scaffolds and RSF hydrogel were biocompatible, non-hemolytic, and supported cell adhesion, growth, and proliferation.
In vitro chondrogenesis effects for the ECM-inspired small aperture scaffolds
As shown in Fig. 3, for the construction of biomimetic cartilage, 5% HA was used as the upper layer, 10% COL was applied as the middle layer (cartilage layer), and 1:1 β-TCP and BMSCs were regarded as the subchondral bone layer [15]. Because of insufficient nutrient supply, the 10% COL middle layer was co-cultured with chondrocytes and BMSCs and the ratio was approximately 2:1[16] in groups C, D, and E. Chondrocytes created a differentiation microenvironment for BMSCs into chondrocytes, meanwhile, BMSCs provided a variety of growth factors for chondrocytes [17].
Then, it could be seen from the results that dsDNA, total GAG and collagen content, group D had the strongest expression after 4 weeks in vitro (Fig. 3b, c, d). HE staining showed that there were a certain number of cells in groups A, B, C and D when cells and scaffolds were co-cultured for 4 weeks in vitro (Fig. S3, Supporting Information). The expression of COL II was the highest in groups B and D from COL II IHC staining (Fig. S3, Supporting Information). In addition, the cell-scaffold complex, secreting flocculent substances similar to ECM attached to their edge, still maintained a good morphology after 4, 8 weeks. And mixed cells stretched well around the complex at the 4th week (Fig. S4, S5, Supporting Information).
Effects of ECM-inspired small aperture scaffolds and (or) RSF hydrogel in vivo
degradation of ECM-inspired small aperture scaffolds in vivo
The degradation time and products of materials in vivo have always attracted much attention. It has reached a consensus that the degradation products of RSF were mainly amino acids absorbed by our body without causing rejection [18], and the ideal degradation time needs to match the time of nascent tissue formation [19]. It is too quick to maintain a scaffold's microstructure, and too slow to lead an inflammatory response [20]. Therefore, we added rhodamine to the RSF solution, stirred it thoroughly and obtained a hybrid scaffold under freezing and thawing conditions. Next, the scaffolds (1 cm in diameter, 3 mm in thickness) were implanted subcutaneously in rats, and then were collected after 3, 4 months (mths) (Fig. 4a). At the 3rd mth, the three groups' scaffolds implanted in rats were all wrapped by subcutaneous tissue, and their volume were significantly reduced (Fig. 4b). The 5% HA, 1:1 β-TCP group degraded three-quarters of their total volume, and the 10% COL group arrived at two-thirds (Fig. 4c,d); at the 4th mth, the scaffolds in first 2 groups could not be seen, however, the remaining part in the latter group approximately accounted for one-fifths after analysis (Fig. 4c,d). Even if there was a small amount of 10% COL left in the 4th month, RSF had extremely excellent biocompatibility, and the added HA, CS, COL components have also been proven to have high biosafety [21], thus the degradation rates were probably acceptable.
Cell fluorescence tracer of ECM-inspired small aperture scaffolds in vivo
Cell survival time in vivo has always been a concern. These cells may undergo apoptosis because of a host's immune rejection and insufficient nutrient supply. Rabbit chondrocytes labeled with different fluorescence (DIL and DIO, Fig. S6, Supporting Information) were loaded on scaffolds (Fig. 4e) , and then transferred and tracked in nude mice for 1 week (Fig. 4f). There were a large number of chondrocytes in 5% HA group, and they aggregated in clumps (Fig. 4g), which was consistent with the results in vitro. In addition, green and red fluorescent cells were seen in both the stem cell groups and the mixed culture cell groups (Fig. 4g, h).
Effects of ECM-inspired small aperture scaffolds and RSF hydrogel on osteochondral defect in vivo
The in vitro studies demonstrated that the ECM-inspired small aperture scaffold sealed by RSF hydrogel are a promising candidate as osteochondral defects. Therefore, a cartilage defect rabbit model was used to assess the healing-acceleration performance of the 2 combinations (Fig. 5a, b). New Zealand white rabbits were injected with ketamine, a full-thickness cartilage defect (5 mm diameter, 4-5 mm depth) was created, and cartilage identification in the defect area was positive for safranin-fast green staining and COLⅡ staining (Fig. 5c). Subsequent treatments consisting of A, B, C, D and E groups were assessed through observation and tissue collection for a long time on mths 4 and 4.5 (Fig. 5d, e, f). Macroscopically (ICRS scores), the remaining 4 groups (B, C, D and E) displayed faster cartilage formation compared to the A group (Fig. 5g). The marginal sealing hydrogel-treated E group has the highest healing scores on mths 4 and 4.5, whereas only ≈20% closure was achieved in the A groups on month 4 (Fig. 5d). This result was consistent with the chondrogenic induction test in vitro, which confirmed that cartilage had proliferation-limited capability. Micro-CT datas also showed that group E had more osteogenesis compared with others (Fig. 5d, e, f). Consequently, edge-sealed and BMSCs-mixed RSF might be critical to enable biomedical applications.
Results from histological analysis by Masson-Goldner staining, Saf-O/Fast Green staining and collagen II immunohistological staining were consistent with the wound healing rate (Fig. 5d, g). The statistical measurement of cartilage and subchondral bone tissue showed that superficial cartilage formation of the E group was faster than the other groups at a later stage of healing (4, 4.5 mths). Specifically, defects from group E after 4.5 mths treatment were almost healed and exhibited a typical histological architecture similar to the rabbit knee joint. It appeared mostly scarless and had the greatest well-defined anatomy of cartilage and subchondral bone (Fig. 5h, i). Taken together, the results demonstrated that group E had superior cartilage healing ability compared to the other four groups.
Mechanism of ECM-inspired small aperture scaffolds and RSF hydrogel on chondrogenesis regulation in vivo
RNA sequencing (RNA-Seq) was used to assess changes in messenger RNA (mRNA) levels in chondrogenesis regulation cultured on the A, B, C, D or E groups in vivo for 4.5 mths. Compared with control groups, the E groups have 968 differentially expressed genes including 340 highly expressed genes and 628 low expressed genes. (Fig. 6a,b). GO analysis demonstrated that among the differentially expressed genes, 64 were associated with the immune system (Fig. 6c,d), and the implants caused the downregulation of genes involved in immunity and inflammation. Notably, KEGG pathway analysis of the differentially expressed genes associated with the chondrogenesis regulation were cytokine--cytokine receptor interaction, cell cycle and proliferation, ECM-receptor interaction and hematopoietic cell lineage. Environmental information processing demonstrated that PI3K-Akt and MAPK signaling pathway may be involved. In vitro, chondrocytes (C28/I2) (Fig. S7, Supporting Information) and (or) human mesenchymal stem cells (hBMSCs), which grew on small aperture scaffolds, secreted cartilage matrix (collagen Ⅱ, aggrecan) possibly through activating Akt, MAPK molecules (Fig. 6g, h) verifying results in vivo. Additionally, the differentially expressed genes associated with the immune system revealed that host defense mechanisms, such as NOD-like receptor and IL-17. E treatment group caused a negative regulation of the “NOD-like receptor signaling pathway”, including the signal transducer and activator of genes, such as NLRP12 and IL1B. NLRP12 acts as a negative regulator of the NFκB and MAPK signaling pathways. Canonical and non-canonical signaling in T cells is negatively regulated by NLRP12 causing exacerbated autoimmune diseases. NLRP12 also functions as an inflammasome or as a negative regulator is context-dependent [22]. IL1B genes were low expressed in the immune system in this research, especially, IL1B was producted by a major T cell effector function in human newborns, which has the potential to activate antimicrobial neutrophils and γδ T cells [22]. In addition, our study also found RSF/BMSCs/hydrogel (group E) treatment did not affect the expression of the chemokine as CXCL3 in “IL-17” signaling pathways. CXCL3 is a well-known inflammatory and structural cells chemoattractant that plays a pivotal role in the migration of keratinocytes during epithelial repair, and in angiogenesis [23].
Thus, analyses of the mRNA-seq suggested that the RSF/BMSCs/hydrogel (group E) decreased the NLRP12, IL1B activation and downstream immune response and inflammatory response, but didn't inhibit cell proliferation and replication (Fig. 6e, f).