All cell culture reagents were acquired from Gibco^® by Life Technologies, unless otherwise stated. The ATDC5 cell line was purchased from Xiangf Bio (Shanghai, China). The cells were cultured and passaged in the normal medium of 1:1 mixture (v/v) of Dulbecco’s modified Eagle’s medium and Ham’s F12 medium (DMEM/F12) with 10% (v/v) fetal bovine serum (FBS). The eighth passage (P8) was used. For chondrogenic differentiation, the macromass method was applied as reported previously (Yao et al., 2017). A 10 μL ATDC5 cell suspension with 2×10⁷ cells/mL was added dropwise into the center of each well of 24-well plates (Corning, United States), incubated for 2 h, and then cultured using a chondrogenesis medium with 2, 8, 35, and 70 mM galactose (Aladdin, Shanghai, China) for 21 days. The cells in the negative control group (marked as ATDC5) were maintained using a normal medium, and the cells in the positive control group (marked as 0 mM) were maintained using a chondrogenic medium, that is, DMEM/F12 was supplemented with 10% FBS, 1 mM sodium pyruvate, 40 μg/ml L-proline (Sigma, United States), 1% (v/v) ITS-G, 50 μg/mL l-ascorbic acid (Sigma), 100 nM dexamethasone (MCE, United States), and 1% (v/v) penicillin/streptomycin. The medium was changed every 2 days.

Rabbit articular chondrocytes were isolated as described in the literature with slight modification (Pei et al., 2009; Gong et al., 2010). In brief, rabbit chondrocytes were harvested from the articular cartilage of five 3-week-old white rabbits (Huadong Xinhua Laboratory Animal Farm, Guangzhou, China). Cartilage was extracted from the femoral condyles and the posterior surface of the patella by excision with a scalpel. Then, the cartilage tissue was minced into small fragments under aseptic conditions. Subsequently, the cartilage slices were digested in 0.1% (w/v) collagenase type II (Sigma-Aldrich, United States) for 14 h in an incubator at 37°C and 5% CO₂. The cell suspensions were filtrated through a Falcon^® Cell Strainer (mesh size 40 μm). All isolated chondrocytes were suspended in DMEM with 10% (v/v) FBS, 10 mM HEPES, 0.1 mM MEM NEAA, 0.4 mM L-proline (Sigma-Aldrich), 50 mg/L L-ascorbic acid (Sigma-Aldrich), 100 U/ml penicillin, and 100 mg/L streptomycin. In this study, chondrocytes at the second passage (P2) were used.

For cartilage matrix formation, the macromass method was used was used as described previously. Briefly, 10 μl of a cell suspension of 2×10⁷ cells/mL was dropped into the center of each well of 24-well plates and incubated in the incubator for 2 h. Then, the cultures were maintained using a medium with 2, 8, 35, and 70 mM galactose (Aladdin, Shanghai, China) for 21 days. The control group was marked as 0 mM, and the medium was replaced every 2 days.

The cytotoxicity of various concentrations of galactose toward ATDC5 cells and chondrocytes was characterized on days 1, 3, 5, and 7 using Cell Counting Kit-8 (CCK-8, Dojindo, Japan) assays according to the manufacturer’s instructions. The cells were seeded in 96-well plates (Corning, United States) at a density of 2000 cells/well, ATDC5 cells were maintained in a normal medium, and both types of cells were supplemented with galactose at a range of concentrations. At the indicated time point, the cells were incubated for 4 h in a CCK-8 working solution at 37°C and 5% CO₂. After incubation, the absorbance of the solution was measured at 450 nm using a Thermo 3,001 microplate reader (Thermo Scientific, United States). The IC₅₀ curves were generated by adding different concentrations of galactose to the medium, and the cells were cultured for 72 h, followed by cytotoxicity measurements using the CCK-8 assay.

The cells on days 3, 7, 14, and 21 were fixed in 4% paraformaldehyde (Macklin, China) for 12 h at 4°C and stained with 1% (w/v) Alcian Blue 8GX (Solarbio, China) aqueous solution (Challa et al., 2010). The stained cells were assembled on plates, washed three times with deionized water to ensure that only specific staining remained, and photographed using a stereomicroscope (Jiangnan, China), and the relative staining level was processed using ImageJ software.

At the specified time points, the cells were washed with PBS three times and then were frozen in liquid nitrogen and lyophilized for 12 h. The remaining cells were collected and digested with the GAG digestion solution (papain 0.0156 g, NaOH 0.2 g, EDTA 0.0731 g, sodium dihydrogen phosphate 0.39 g, and cysteine 0.0304 g in 25 ml ultrapure water) for 16 h at 60°C. The GAG content was measured using the Blyscan sGAG Assay (Biocolor, United Kingdom), and the standard curve was generated by chondroitin sulfate standard using the kit. The DNA content was quantified using Hoechst 33,258 (Biosharp, China). Calf thymus DNA (Sigma) was used to generate a standard curve (Farndale et al., 1986).

At the specified time points, total RNA was isolated using the HiPure Total RNA Kit (Magentec., China) and the concentration of RNA was determined using a Nano drop 2000 spectrophotometer (Thermo Scientific). Then, the cDNA was reverse-transcribed using a PrimeScript RT reagent kit with gDNA Eraser (TaKaRa Biotechnology, Japan) according to the manufacturer’s protocol. Finally, quantitative real-time PCR was performed using a SYBR Green System (GeneCopoeia, United States) on an RT-PCR instrument (LightCycler 96, Roche). The relative quantification of target genes was performed with normalization to glyceraldehyde 3-phosphate dehydrogenase (GAPDH), and the 2^−ΔΔCt method was used to calculate the gene expression. The PCR primers used in this study are listed in Supplementary Tables S1, S2.

Induced cells were fixed in 4% paraformaldehyde for 12 h at 4°C. The cell aggregates were carefully collected from the plates and immersed in a gradient concentration of sucrose solution (10, 20, and 30% w/v) to dehydrate the frozen sections. Then, the cell aggregate samples were embedded at optimal cutting temperature (O.C.T) and cut into 6-μm-thick sections using a freezing microtome (Leica). For immunofluorescence staining, the samples in the sectioned slides were permeabilized and blocked with 0.3% Triton X-100 and 10% goat serum for 2 h at room temperature. The samples were incubated with primary antibodies against aggrecan (1:200, Affinity), collagen II (1:200, Bioss), collagen I (1:200, Bioss), and collagen X (1:200, Bioss) overnight at 4°C. The secondary antibody was diluted 1:500 (goat antirabbit IgG FITC-conjugated, goat antirabbit IgG Fluor 594-conjugated, Affinity), and the sections were incubated with the secondary antibody for 2 h at room temperature. Then, the chondrocyte mass sections were stained with H&E, safranin-O, and Masson. The images were taken using a digital pathology scanning system (3DHISTECH P250 FLASH).

All data are presented as mean ± standard deviation (STDEV) for at least three repeated individual experiments. Significant differences were confirmed by analysis of variance (two-way ANOVA) for independent samples. Statistical significance was defined as p < 0.05.

After treatment for 72 h, cell behavior exhibited a significant dose-dependent relationship with galactose. The IC₅₀ represents the concentration of a drug that is required for 50% inhibition (Sebaugh, 2011; Swinney, 2011). The IC₅₀ value of galactose for ATDC5 was approximately 99.6 mM (Supplementary Figure S1A), which is considerably higher than that of glucosamine, which was 13 mM (Yao et al., 2017). Generally, an IC₅₀ value greater than 1 mM is considered to be free of any toxicity, demonstrating that galactose can be used for more promising applications.

To further elucidate the cytotoxic effect of galactose on ATDC5 cells, CCK-8 assays were carried out at various time points during 7 days of culture. During the culture, ATDC5 showed an increasing trend of proliferation with excellent growth. For ATDC5, on days 5 and 7, cell proliferation was slightly lower in the experimental groups than that in the control group (Supplementary Figure S1B), but there was no significant difference among all ATDC5 groups, indicating that no evident cytotoxicity was observed below the galactose concentration of 32 mM. Based on these results, 2, 8, 35, and 70 mM were selected for the following tests.

The ECM of hyaline cartilage in knee joints, which mainly includes proteoglycans and type II collagen, has a unique physiological function and plays an important role in daily life. This entitles glycosaminoglycan (GAG) deposition as an important phenotypic parameter for chondrogenesis and chondrogenic differentiation. Alcian blue, a cationic dye, chemically binds to the acidic functional groups in GAG, so the accumulation of GAG in the deposited cartilage ECM can be directly observed by Alcian blue staining. In Figure 1A, staining in the 0 mM group (chondrogenic medium) was stronger than that in the undifferentiated group of ATDC5 (normal medium), demonstrating successful chondrogenic differentiation of ATDC5 cells. In addition, the galactose-treated groups showed more pronounced staining as the galactose concentration increased with time, suggesting increased GAG deposition in the ECM. The staining intensity of Alcian blue was quantified using ImageJ to measure the tiny difference and the quantification of the GAG/DNA content. The 8 mM group reached a peak of 1.85 ± 0.072 μg per μg DNA on day 21, significantly higher than that of the control group (Supplementary Figure S2). Therefore, galactose can enhance GAG accumulation during ATDC5 chondrogenic differentiation.

ATDC5 cells were assessed on days 3, 7, 14, and 21 for cartilage-related genes, such as aggrecan, type II collagen, type X, and type I collagen, to further define the ability of galactose to regulate the cell phenotype and promote chondrogenic differentiation. As depicted in Figure 1, the 0 mM group (positive control) showed elevated expression levels of all four genes compared to those of the undifferentiated ATDC5 (negative control), indicating that the applied differentiation method was effective. The expression levels of the hyaline cartilage–related genes aggrecan (Figure 1B) and type II collagen (Figure 1C) increased over time during the first 14 days and significantly upregulated with increasing galactose concentrations. However, the results of the type X collagen gene (a marker of cellular hypertrophy, Figure 1D) showed that the high concentrations of galactose after 14 days could lead to cell hypertrophy. In contrast, the expression of the type I collagen gene (a typical fibrocartilage marker, Figure 1E) was significantly upregulated at 2 mM for the first 3 days and peaked at 8 mM on days 14 and 21, but the expression level was much lower than that of hyaline cartilage–related genes, suggesting that galactose could effectively promote chondrogenic differentiation of ATDC5 cells.

After assessment of gene expression levels, immunofluorescence staining was performed under the same conditions to detect the protein levels of aggrecan, type II, type X, and type I collagen in ATDC5 cells. Overall, the positive staining of these proteins with different intensities was observed in all experimental groups on days 14 and 21, but the cartilage-related protein expression was largely consistent with gene expression levels. For ATDC5, staining was slightly stronger on day 21 than on day 14, and the expression levels of aggrecan (Figure 2A) and type II collagen (Figure 2B) were upregulated with increasing galactose concentrations, demonstrating that galactose can successfully promote chondrogenesis. In addition, as investigated in other articles, the estimated half-life of proteoglycans is about 26 days, while that of collagen is over 100 years (Lai et al., 2013; Linka et al., 2016; Christiansen-Weber et al., 2018). Both are longer than the experiment period in this work. Although the gene expression levels of aggrecan and type II collagen were lower on day 21 than those on day 14, the experimental groups still had higher gene expression levels than those of the control group. Therefore, it is possible to infer that aggrecan and type II collagen proteins were being synthesized, and the accumulation was observed higher on day 21 than that on day 14 in the immunofluorescence staining in Figure 2. However, the expression of type X collagen (Supplementary Figure S3A) was not significantly different from that in the other experimental groups, suggesting that gene expression levels may not directly contribute to differences in protein levels. The differences among the groups treated with galactose concentrations under 8 mM and the control groups were not evident for type I collagen (Supplementary Figure S3B). Collectively, these results demonstrated that galactose at concentrations under 8 mM had a better effect on chondrogenesis.

Having shown that galactose successfully promoted chondrogenic differentiation of ATDC5 cells, we then attempted to verify the direct effect of galactose on the chondrocyte phenotype. The IC₅₀ value of galactose on chondrocytes was approximately 498.5 mM (Supplementary Figure S4A), which may indicate that chondrocytes are better adapted to a high-sugar environment than ATDC5. During the 7-day culture (Supplementary Figure S4B), chondrocytes presented a similar increasing trend of proliferation as ATDC5. On days 5 and 7, cell proliferation was slightly lower in the experimental groups than that in the control group, but still higher than 80%. Galactose concentrations of 2, 8, 35, and 70 mM were selected for the following tests, consistent with those for the ATDC5.

To evaluate the capacity of cartilage matrix formation by chondrocytes, cell masses were cultured at different concentrations of galactose for 21 days, followed by Alcian blue staining. All cell masses gradually became larger and thicker and proliferated outward with the extension of time. As shown in Figure 3A, the chondrocyte mass exhibited stronger staining as the galactose concentration increased with time, directly indicating that the ability of chondrocytes to generate GAG was robustly promoted by galactose stimulation. Quantitative measurements also indicated that GAG accumulation was significantly promoted with the addition of galactose, peaking on day 21 at 8 mM (35.63 ± 1.82 μg per μg DNA, Supplementary Figure S5). Interestingly, small matrix clumps were observed on day 3 and day 21, and the cell masses in the high–galactose concentration groups (35 and 70 mM) appeared partially sparse.

Then, chondrocytes were assessed on days 3, 7, 14, and 21 for cartilage-related genes, such as type II collagen, aggrecan, type X, and type I collagen, to evaluate the influence on the chondrogenic phenotype. As depicted in Figure 3, the expression of the aggrecan gene (Figure 3B) was upregulated gradually during incubation and was significantly different from that of the control group on days 14 and 21. On day 14, there was a significant upregulation of aggrecan gene expression at galactose concentrations of 2 and 70 mM and a slight decrease at galactose concentrations of 8 and 35 mM. The aggrecan gene expression peaked on day 21, showing a significant increase with increasing galactose concentration. For type II collagen (Figure 3C), the gene expression levels were similar to those of aggrecan. Type II collagen gene expression levels were progressively upregulated with increasing galactose concentrations on days 3 and 7, while galactose concentrations resulted in increased expression at 2 and 70 mM. The type X collagen gene and type I collagen were also examined. There was a modest elevation in collagen type X (Figure 3D) and type I gene (Figure 3E) expression compared to that of the control group, which is likely to be a general phenomenon of drug treatment. In addition, the elevation of type II collagen and aggrecan was superior to that of type X and type I collagen after galactose treatment. In particular, for the type X collagen gene, there was no significant difference between the experimental groups before 14 days of culture, which indicated that galactose treatment may be more effective in promoting cartilage expression within 14 days.

After the assessment of gene expression levels, immunofluorescence staining was performed under the same conditions to detect the protein levels of aggrecan, type II, type X, and type I collagen in chondrocytes. Consistent with the ATDC5 results, the protein expression was largely consistent with gene expression levels. For chondrocytes, staining was slightly stronger on day 21 than on day 14 and showed a significant increase in aggrecan (Figure 4A) and type II collagen (Figure 4B) expression with increasing galactose concentrations, demonstrating that galactose can successfully promote cartilage ECM secretion and accumulation. In addition, the protein expression levels of type X (Supplementary Figure S6A) and type I collagen (Supplementary Figure S6B) were visually different compared to those of the control group (0 mM). A moderate increase was observed in type I collagen levels at concentrations of 2 and 8 mM and was significantly weaker than that of aggrecan and type II collagen. Therefore, although there was an increase in the gene expression of type X and type I collagen in chondrocytes, associated with cell hypertrophy and fibrocartilage, no significant expression of type X and type I collagen was detected by immunofluorescence staining, suggesting that galactose also triggered anabolism and enhanced cartilage matrix formation by chondrocytes.

In particular, we also performed histological staining of the chondrocyte masses (H&E, safranin-O, and Masson) (Supplementary Figure S7). H&E staining was used to assess the overall structure of the in vitro cultured cartilage matrix, which showed that the chondrocytes were round and evenly distributed in the ECM. Safranin-O could detect the anionic GAG expression in the cartilage matrix. In Masson’s staining, the cytoplasm was stained red and the collagen fibers were stained blue. As the culture time was extended from 14 to 21 days, the matrix staining gradually deepened, indicating that the matrix secreted by the chondrocytes gradually increased. In addition, the formation of distinct cartilage lacunae in the cell masses was observed in the staining results, indicating successful cartilage matrix formation.