All procedures in this study, including in vitro and in vivo experiments, were approved by the Ethics Committee of the Second Hospital of Shanxi Medical University (NO. SYDL2019001). Human chondrocytes used in this study were derived from the articular cartilage library of Shanxi Key Laboratory of Bone and Soft Tissue Injury Repair. All patients provided informed consent, and all procedures were approved by the Ethics Committee of the Second Hospital of Shanxi Medical University (NO. 2019YX260). Based on the purpose of this study, under the guidance of statistical experts, the number of samples numbersin this study was six. For the in vitro experiments, a sample size of six was used. For the in vivo experiments, the number of animals in each group was determined as six. Patient baseline characteristics were as follows: age, 64 ± 3.38 (years); sex, female (n = 4)/male (n = 2); height, 163 ± 3.53 (cm); weight: 65 ± 3.85 (kg). Mini pigs were purchased from the Beijing Shichuang Century Mini pig Breeding Base (Certificate number: SCXX (jing)2013-0008). All animals were housed at the China Institute for Radiation Protection (Certificate number: SYXK (Jin)2016-0002).

The following antibodies were used in this study: collagen-2 (Col-2, ab34712, Abcam), matrix metalloproteinase -3 (MMP-3, bs-0413R, Bioss), MMP-9 (bs-4593R, Bioss), MMP-13 (K009743P, Solarbio), Col-10 (bs-0554R, Bioss), and Runt-related transcription factor 2 (Runx-2, ab76956, Abcam).

Whole blood (13 ml) was collected in a coagulation tube (BD Vacutainer SSTTM II,United States ) and centrifuged at 2,000 × g for 20 min to obtain 6 ml of serum, which was then added to the upper filter of the ultrafiltration tube (Cytonics Corporation, West Palm Beach, Florida, United States ). The upper concentrate was obtained under different conditions of centrifugal force (3,000, 4,000, and 5,000 × g) and time (20, 30, and 40 min). Finally, the best concentration conditions were determined based on the concentration of α2M in the upper concentrate.

Human chondrocytes were isolated as previously described (Guo et al., 2019) and plated in 6-well culture plates at a density of 1 × 10⁶ cells/plate. Chondrocytes culture medium was Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F-12) containing 10% fetal bovine serum (Hyclone) At 90% confluence, the cells were cultured overnight under serum-free conditions and then treated with 10 ng/ml recombinant human interleukin (IL)-1β for 2 h before treatment with α2MRS. It was ensured that the concentration of α2M was 0.25 mg/ml in the culture medium. The culture medium and chondrocytes were collected and analyzed.

α2M concentrations in the upper concentrate under different centrifugal conditions were determined using ELISA (EK1118, Boster Bio, China). The human chondrocyte culture medium was collected 24 h after α2MRS treatment and analyzed for the presence of MMP-13, tumor necrosis factor-α (TNF-α), and IL-6 using ELISA.

Human chondrocyte proliferation was detected at 0, 24, 36, and 48 h using the Cell Counting Kit-8 (CCK-8) cell viability assay kit (Boster Biological Technology, China). Human chondrocytes were collected 24 h after α2MRS treatment and apoptosis was detected using a Terminal transferase dUTP end Labeling (TUNEL) assay kit (Key GEN Bio TECH, China). The percentage of positive cells was determined. The detailed procedure was in accordance with the manufacturer’s protocol.

mRNA levels of col-2, aggrecan, MMP-3, and MMP-13 in human chondrocyte samples and those plus collagen-10 a1 (col-10 a1) and Runx-2 in minipig cartilage weight-bearing sites of the medial tibial plateau were measured by real-time PCR. Primer pairs are listed in Supplementary Table S1. Levels of gene expression were normalized to 18S rRNA expression. The data were analyzed using the comparison Ct (2^−ΔΔCt) method and expressed as the fold-change relative to the respective control. The detailed PCR procedure was described previously (Gu et al., 2019).

Eighteen mature female mini pigs (age, 18 ± 1.55 months; weight, 43.3 ± 3.67 kg) were randomized into three groups based on animal ear numbers: sham (n = 6), IACL-R (n = 6), and IACL-R+α2MRS (n = 6). All surgeries were performed under anesthesia via an intramuscular injection of 25 mg/ml tiletamine and 25 mg/ml zolazepam (Zoletil 50, 1 ml/15 kg; Virbac Group, Carros, France). Unilateral surgery was performed on the right hind limbs of all mini pigs. Mini pigs in the IACL-R and IACL-R+α2MRS groups were subjected to surgery based on methods described previously (Figure 1; Bryan et al., 2011; Han et al., 2018). The minipigs in the sham group underwent arthrotomy, temporary patellar dislocation, and coring of one-third of the length of the lateral femoral condyle. Specific information on animal care can be found in the Supplementary Text S1.

With the help of a veterinarian, 120 ml of whole blood was collected from the anterior vena cava of each mini pig into coagulation tubes before IACL-R. According to the best concentration conditions (centrifugal force: 5,000 × g; time: 30 min), 12–15 ml of α2MRS was obtained per mini pig (marked according to the mini pig ear number), and these samples were frozen at −80°C.

Synovial fluid from the right hind limbs of all animals was collected preoperatively (day 0) and postoperatively on days 3, 6, 14, 29, and 90. The detailed procedure was described previously (Wei et al., 2010).

Under general anesthesia, intra-articular injections were administered 2, 6, 14, and 29 days post-surgery using a sterile syringe. Under aseptic conditions, 2.5 ml autologous α2MRS was injected into the right hind limbs of mini pigs in the IACL-R+α2MRS group on the indicated days. Animals in the sham and IACL-R groups were administered an equivalent volume of saline.

The Millipore Porcine Cytokine Magnetic Bead Panel (EMD Millipore, No. PCYTMAG-23K) was used to measure the levels of IL-1α, IL-1β, IL-2, IL-6, IL-8, IL-18, TNF-α, and granulocyte-macrophage colony-stimulating factor (GM-CSF). Luminex assays were performed as previously described (Han et al., 2018).

Six gait indicators related to biomechanics—maximum force, contact area, peak force, impulse, stance time, and swing time—were determined using the Tekscan Walkway system (Tekscan Inc., United States ) (Rashid et al., 2013; Shah et al., 2020). To rule out individual differences in learning skills and walking states, each animal was subjected to over ten successful training sessions per day for ten consecutive days before the surgery, and all indicators were expressed as the ratio of the average values for the left hind limb divided by the average values for the right hind limb (Ruan et al., 2013). Gait data were collected preoperatively (day 0) and postoperatively on days 7, 15, 30, 45, 60, 75, and 90. All results obtained were the average of five successful repeated walkway trials performed at each time point for each animal.

Three months after the surgery, the mini pigs were euthanized with a pentobarbital overdose, and their right hind limbs were severed from the hip joint. Each right hind limb semi flexed was immediately subjected to X-ray examination, computed tomography (CT), three-dimensional CT reconstruction (3D CT), and magnetic resonance imaging (MRI). The specific imaging parameters are listed in Supplementary Text S2.

We determined the Kellgren-Lawrence grade of the right hind limb of each animal by examining the X-ray image (Kellgren and Lawrence, 1957; Misir et al., 2020). The CT values and thickness of the subchondral bone plate were determined (Hu et al., 2020). To avoid interference by metal artifacts, we obtained CT scans of only the middle sagittal plane from the medial femoral condyle and medial tibial plateau of the right hind limb. We also determined the whole-organ MRI score (WORMS) of the medial femoral condyle and medial tibial plateau of the right hind limb (Pozzi et al., 2015; Xue et al., 2021).

Macroscopic damage to the articular cartilage surfaces and osteophyte formation on the medial femoral condyle, medial tibial plateau, lateral femoral condyle, lateral tibial plateau, and trochlea were scored according to the Osteoarthritis Research Society International (OARSI) recommendations for sheep and goats (Little et al., 2010)

Cartilage samples were obtained by drilling (φ8 mm; MOC Medizinische Geräte Gmbh, Fedderingen, Germany) the weight bearing site of the medial femoral condyle. The cartilage tissue sections (6 µm) were stained with safranin O and fast green as previously described and scored according to the recommendations of OARSI (Little et al., 2010). Furthermore, we collected synovium samples from inside the joint capsule. Synovium sections (4 µm) were stained with hematoxylin and eosin (H&E) as previously described and scored according to the OARSI recommendations (Little et al., 2010). Vertical meniscus slices from the middle region of the medial meniscus were processed and stained using H&E and scored according to the protocol detailed by Pauli et al. (2011).

Immunohistochemical analysis was conducted as reported previously (Li et al., 2020). Briefly, to detect the distribution of the target protein, 6-μm thick cartilage tissue sections from the medial femoral condyle were collected on positively-charged glass slides. Endogenous peroxidase was blocked by treating the sections with 3% hydrogen peroxide in methanol. The sections were incubated with specific antibodies against Col-2 (1:500), MMP-3 (1:200), MMP-9 (1:100), MMP-13 (1:100), Col-10 (1:50), and Runx-2 (1:50) at 4 °C overnight. Thereafter, the sections were treated sequentially with a biotinylated secondary antibody and streptavidin–peroxidase conjugate and then developed with 3,3′-diaminobenzidine chromogen. Quantitative immunohistochemical analysis was performed using an imaging analyzer.

The SPSS statistical software (version 13.0) was used to analyze the collected data. Differences in gaits and inflammatory factor levels between the preoperative (day 0) and postoperative time points in the same group were analyzed using multiple comparisons of repeated measurement data. Differences in human chondrocytes, minipig gait, inflammatory factor levels, CT values, thicknesses of the subchondral bone plate T2 values, and quantitative immunohistochemical analysis at the same time point among the different groups were estimated using one-way analysis of variance. Differences in macroscopic cartilage and osteophyte scores; microscopic cartilage, synovium, and meniscus scores; and WORMS were estimated using nonparametric tests (Wilcoxon rank-sum test). Differences in Kellgren-Lawrence grades were estimated using Fisher probabilities. Statistical significance was set at p < 0.05.

As the centrifugal force increased and centrifugation time was prolonged, human α2M concentrations in the upper concentrate increased correspondingly (Table 1). Considering different factors, such as protein biological activity and concentrate volume, the concentration effect of α2M under the conditions of centrifugation at 5,000 g for 30 min was ideal. For human α2MRS, the concentration of α2M was 11.13 mg/ml, which was 4.88-fold higher than that in normal human serum. In mini pigs, the concentration of α2M was 12.32 mg/ml, which was 6.48-fold higher than that in normal pig serum (Figure 2A).

ELISA results demonstrated that exogenous α2MRS significantly inhibited the induction of MMP-13 (p = 0.001), TNF-α (p = 0.005), and IL-6 (p < 0.001) activity in IL-1β-induced human primary osteoarthritis chondrocytes (Figures 2B–D).

α2MRS promoted the proliferation and reduced the apoptosis of human chondrocytes in vitro. The results of the CCK-8 assay showed that the viability of chondrocytes was higher in the IL-1β+α2MRS group than in the IL-1β group (p < 0.001), and the viability gradually increased with treatment time (Figure 2E). The results of the TUNEL assay showed that apoptosis was significantly reduced in the IL-1β+α2MRS group (18.33% ± 5.71%) relative to that in the IL-1β group (32.33% ± 7.23%; p < 0.001; Figures 2F,G).

Real-time PCR data indicated that supplementation with α2MRS reduced cartilage matrix catabolism and enhanced anabolic metabolism in vitro (Figures 3A–D) and in vivo (Figures 3E–J). mRNA levels of MMP-3 (p < 0.001), MMP-13 (p < 0.001), Col-10 a1 (p = 0.001), and Runx-2 (p < 0.001) were lower in the IL-1β+α2MRS and IACL-R+α2MRS groups than in the IL-1β and IACL-R groups. In contrast, mRNA levels of Col-2 (p < 0.001) and aggrecan (p < 0.001) showed the opposite pattern. Both were increased in the IL-1β+α2MRS and IACL-R+α2MRS groups as compared to the levels in the IL-1β and IACL-R groups, respectively.

Postoperatively, changes in the levels of inflammatory factors in different groups, except for IL-18 in the IACL-R group, showed similar trends; the levels markedly increased in the early stage and then decreased significantly. The IL-2 concentration in the IACL-R group subsequently showed an increasing trend from day 30 to day 90. The concentrations of inflammatory factors, including IL-1β, IL-6, IL-18, TNF-α, and GM-CSF, in the IACL-R and IACL-R+α2MRS groups after surgery were significantly higher than those before surgery (p < 0.001). The concentration of all tested inflammatory factors other than IL-1α after surgery was significantly lower in the IACL-R+α2MRS group than in the IACL-R group, and a significant difference in peak concentrations was observed in all factors (p < 0.001). Moreover, the peak concentrations of all detected inflammatory factors, other than IL-18 in the IACL-R group, appeared within 3–14 days after surgery (Figure 4).

Across all groups, the ratios of all gait parameters of the left hind limb to those of the right hind limb initially showed an increasing trend, followed by a decreasing trend, except in the IACL-R group, which showed another increasing trend toward the end. In the IACL-R group, no gait parameters, from day 45 until day 75, differed significantly from those on day 0 (p > 0.05). In the IACL-R+α2MRS group, no gait parameters, from day 30 until euthanasia, differed significantly from those on day 0 (p > 0.05) (Figure 5).

The ratios of the left hind limb to the right hind limb of all gait parameters were similar and close to one for symmetry and did not differ significantly in all groups before surgery (p > 0.05). Meanwhile, the postoperative ratios of the left hind limb to the right hind limb of the gait parameters were significantly greater than one in all groups, and the values in the IACL-R group were significantly greater than those in the other two groups other than swing time, especially on days 7 and 15 (p < 0.001). The ratios of the left hind limb to the right hind limb of the gait parameters were close to one on days 45 and 60, indicating that there were no significant differences among the groups (p > 0.05). On day 75, this ratio increased only in the IACL-R group whereas it remained constant in the sham and IACL-R+α2MRS groups. On day 90, all gait parameters other than swing time (p = 0.345) significantly differed between the IACL-R group and the other two groups (p < 0.001; Figure 5).

The X-ray examinations showed that joint degeneration in the IACL-R group was relatively noticeable. The joints had a blurred border and mild osteophyte formation compared with those in the IACL-R+α2MRS group (Figure 6A). The Kellgren-Lawrence grades did not significantly differ among the three groups (p > 0.05; Table 2). Three-dimensional CT reconstruction showed that all joint surfaces were relatively smooth and flat in the IACL-R+α2MRS group compared to the IACL-R group. Osteophyte was seen on both sides of the patellofemoral joint in the IACL-R group (Figure 6B). Both the medial femoral condyle (p < 0.001) and medial tibial plateau (p < 0.001) showed significant differences in CT values in the subchondral bone plate between the sham and IACL-R groups. In addition, significant differences in CT values were found between the IACL-R+α2MRS and IACL-R groups in the medial femoral condyle (p = 0.028; Figure 6D). Significant differences in the thickness of the subchondral bone plate were found only between the sham and IACL-R groups in the medial tibial plateau (p = 0.020; Figure 6E). MRI OSag-fs PD showed that cartilage continuity was better without obvious local defects in the IACL-R+α2MRS group than in the IACL-R group (Figure 6C, left). WORMS of the medial femoral condyle (p = 0.006), medial tibial plateau (p = 0.014), and sum (p = 0.004) were significantly lower in the IACL-R+α2MRS group than in the IACL-R group (Figures 6G–I; Supplementary Table S2). Osag T2MAP showed that regular orange-red layers were more obvious in the IACL-R+α2MRS group than in the IACL-R group (Figure 6C, right). The T2 values of the medial femoral condyle (p < 0.001) and medial tibial plateau (p < 0.001) were significantly lower in the IACL-R+α2MRS group than in the IACL-R group (Figure 6F).

Compared to that in the IACL-R group, cartilage degeneration was relatively low and no obvious cartilage defects or large erosions were found in the IACL-R+α2MRS group (Figure 7A). OARSI scores of macroscopic cartilage were significantly lower in the IACL-R+α2MRS group than in the IACL-R group (p = 0.031; Figure 7B; Supplementary Figures S1A–E; Supplementary Table S2). Mild irregular protrusions were found only on sides of the trochlea in the IACL-R group (Figure 7A). No differences were found in the OARSI sum scores of the osteophyte among the three groups (p = 0.438; Figure 7C; Supplementary Figures S1F–J; Supplementary Table S2).

Less decreases in safranin O staining and surface fibrillation were observed in the IACL-R+α2MRS group as compared to the IACLR-group (Figure 8A). The microscopic OARSI cartilage scores were lower in the IACL-R+α2MRS group than in the IACL-R group (p = 0.015; Figure 8B; Supplementary Figures 2A–E; Supplementary Table S2). Based on H&E staining of the synovium, we found mild intimal thickening, low inflammatory cell infiltration, and slight sub-intimal fibrosis and vascularity in the IACL-R+α2MRS group, relative to those in the IACL-R group (Figure 8C). Thus, the microscopic OARSI synovium scores, both total scores (p = 0.002) and single indicator scores, showed that synovial damage was lower in the IACL-R+α2MRS group than in the IACL-R group (Figure 8D; Supplementary Figures S2F–I; Supplementary Table S2). H&E staining of the meniscus also revealed mild surface fibrillation, normal cell distribution, and a normal collagen fiber organization (Figure 8E), and the meniscus score was lower in the IACL-R+α2MRS group than in the IACL-R group (p < 0.001; Figure 8F; Supplementary Figures 2J–L; Supplementary Table S2).

Both articular cartilage (Figures 9A,B) and synovium (Figures 9C,D) immunostaining showed that MMP-3 (p < 0.001), MMP-9 (p < 0.001), MMP-13 (p < 0.001), Col -10 (p < 0.001), and Runx-2 (p = 0.001) staining significantly increased in the IACL-R group compared with the IACL-R+α2MRS group. In contrast, Col-2 (p < 0.001) expression in articular cartilage was higher in the IACL-R+α2MRS group than in the IACL-R group.