Introduction

Osteoarthritis (OA) is a prevalent degenerative joint illness which can cause disability and severely harm human health, affecting approximately 3.6% of the global population [1]. Cartilage degeneration, characterized by gradual loss of articular cartilage, reconfiguration of subchondral bones, as well as osteophyte formation at joint borders, is main pathological hallmark of OA joints [2, 3]. However, the exact mechanism of occurrence and development of OA has not yet been fully elucidated. OA has become widely accepted as a complex degenerative disease affecting the entire joint rather than merely a mechanical cartilage damage condition [3, 4].

There are two components of articular cartilage: chondrocytes and extracellular matrix (ECM) [5]. Among transcription factors, SOX9, with other transcription factors such as runt-related transcription factor 2 (RUNX-2), hypoxia and hypoxia-inducible factor (HIF-1), as well as forkhead box O3A (Foxo3A), is involved in cartilage phenotype regulation and has an essential function in cartilage ECM homeostasis and chondrocyte survival [6,7,8,9,10]. In chondrocytes, if balance between catabolism and anabolism of SOX9 is disrupted, it increases matrix metalloproteinase-3 (MMP3) expression and decreases COL2A1 expression, leading to the destruction of cartilage ECM homeostasis; therefore, the SOX9 stability is crucial for articular cartilage [5]. In addition to MMPs, another matrix-degrading enzyme of a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTSs) is highly expressed in OA joint, suggesting their involvement in matrix degradation during OA development [11, 12].

In vertebrates, homeobox genes (HOX genes) have an essential involvement in cell differentiation regulation, body patterning, as well as cell migration by regulating target gene expression [13]. The HOXD gene family’s significance in vertebrate limb development has been widely researched. In vertebrate limb development, 5′-HOXD gene (HOXD9-13) is required for establishing patterning and is expressed in a nested manner along the A–P axis, it can be divided into three sections patterned along the proximal–distal (P–D) axis: the stylopod (humerus and femur), the zeugopod (radius/ulna and tibia/fibula), and the autopod (the wrist/forepaw, ankle/hindpaw) [14, 15]. HOXD9 and HOXD10 function in the stylopod region [16, 17], HOXD11 in the zeugopod region [16], and HOXD12 and HOXD13 in the autopod region [18]. An abnormally expressed gene can result in dramatic, region-specific limb deformities. For example, a polyalanine expansion in the N-terminus of the HOXD13 gene results in polydactyly in humans [19], and idiopathic congenital clubfoot is associated with mutations in the HOXD12 and HOXD13 [20]. Since many 5′-HOXD genes have extensive functional overlap, it is difficult to completely detect the role of a single member [21]. One of the 5′-HOXD genes, HOXD9, is elevated in synovial cells of rheumatoid arthritis patients. HOXD9 may play a significant involvement in retinoic acid (RA) because of its association with synovial cell proliferation [22, 23]. HOXD9 has been implicated in chondrogenesis through regulating SOX9 and COL2A1 directly, according to our previous studies [24]. However, no direct genetic evidence links 5′-HOXD genes to OA development.

As mentioned above, we hypothesized that one or more 5′-HOXD genes may be associated with cell proliferation, cell apoptosis, and expression of cartilage-specific factors involved in cartilage degradation and inducing osteoarthritis development. In this study, we aimed to investigate if 5′-HOXD genes were involved in cartilage degradation and OA development. First, we collected and separated cartilage samples from normal and OA patients and cultured the primary chondrocytes. We then examined cell proliferation by CCK-8 analysis, cell apoptosis by flow cytometric analysis, and cartilage-specific factors (SOX9-COL2A1) expression by quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) and Western blot (WB) in normal and OA chondrocytes. We identified 5′-HOXD genes (HOXD9-13) that were differentially expressed (DE) in OA chondrocytes through bioinformatic analysis using Gene Expression Omnibus (GEO) database. Next, we investigated differentially expressed genes (DEGs) expression by qRT-PCR and WB analysis. Finally, we explored whether overexpression of HOXD11 using lentivirus could prevent cartilage degradation in OA.

Materials and methods

Human articular cartilage samples and chondrocyte culture

Normal human cartilage specimens were extracted and separated from patients’ knee joints (n = 6) who had suffered above-knee amputations after severe lower extremity injuries. Human cartilage samples were taken from knee joints of OA patients (n = 6) who had underwent total knee arthroplasty (TKA). Southern Medical University’s Institute Research Ethics Committee reviewed and approved this study. Patients/participants gave their explicit written consent to take part in our research.

Primary chondrocytes were separated from the cartilage of the human tibial plateau and rinsed with Dulbecco’s modified Eagle’s media (DMEM; Sigma, St., USA) according protocol as previously described. Furthermore, cartilage tissue specimens were isolated and divided into fragments, before being treated with 0.25% trypsin for 30 min and 0.2% collagenase type II for 8 h at 37 °C (Sigma-Aldrich, USA). The digest was diluted in DMEM with 10% (v/v) fetal bovine serum (FBS) (Hyclone, USA), 1% penicillin/streptomycin (v/v) (Invitrogen, USA), 2 mM glutamine (Sigma-Aldrich, USA), as well as 50 g/mL ascorbic acid (Sigma-Aldrich, St. USA). Cells were plated in a Petri dish at a density of 1 × 105 cells/mL as well as cultivated at 37 °C in an incubator of 95% O2 and 5% CO2. Cultivated chondrocytes were used in subsequent experiments when cells were at 80% confluence.

Analysis of cell proliferation

A cell counting kit-8 assay (CCK-8; Beyotime, China) was utilized to evaluate cell proliferation based on manufacturer’s instruction described in the previous study. Cells were plated in 6-well plates at a density of 1 × 105 cells/well as well as incubated for confluence, then measured at 0, 1, 3, 5, and 7 d. A microplate reader (Bio-Rad, USA) was utilized to measure optical density (OD) at 450 nm.

Flow cytometric analysis

An Annexin V-fluorescein isothiocyanate (FITC) apoptosis detection kit (BD Biosciences, USA) was utilized to evaluate percentage of apoptotic cells following manufacturer’s instructions. Briefly, chondrocytes were seeded in 6-well plates at a density of 1 × 105 cells/well as well as incubated until confluence, after which cells were washed with 4 °C phosphate buffer saline (PBS) thrice and fixed at −20 °C for 1 h at least in ethanol. After thrice washes, cells were incubated in dark for 30 min after being treated with 10 μL of Annexin V-FITC and 5 μL of propidium iodide (PI). Then, 0.6 mg/mL RNase in PBS with 0.5% (v/v) Tween 20 as well as 2% FBS was added to cells. A FACSCalibur flow cytometer (BD Bioscience, USA) was used to detect stained cells using CellQuest software. Approximately 1 × 104 cells of each sample were counted. Utilizing WinMDI software, data were analyzed (version 2.9, Bio-Soft Net).

qRT-PCR

Total RNA was separated utilizing a TRIzol reagent (Invitrogen, USA). Utilizing a PrimeScript RT Reagent Kit (Takara Bio, China), first-strand cDNA synthesis was accomplished. qRT-PCR was conducted utilizing SYBR Premix Ex Taq (Takara Bio, Beijing, China) based on manufacturer’s instructions on a LightCycler 480 SYBR Green I Master (Roche, Indianapolis, USA) at 95 °C for 10 min, 40 cycles at 95 °C for 15 s and 60 °C for 1 min. Performed genes and primer sequences are registered in Table 1. Gene expression was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

Table 1 qRT-PCR primer sequence and cycling conditions
Full size table

Western blot

Cells were rinsed with 4 °C PBS thrice as well as added with radioimmunoprecipitation assay (RIPA) buffer (Beyotime, China) with 1% phenylmethylsulfonyl fluoride (Sigma, St. USA). Moreover, 10% SDS–polyacrylamide gel electrophoresis (SDS-PAGE) was utilized to extract total protein, and then, proteins were translocated to polyvinylidene difluoride membranes (Thermo-Fisher, Hampton, NH). Primary antibodies against ADAMTS4 (1:500 dilution, Proteintech, 11865-1-AP), ADAMTS5 (1:1000 dilution, Abcam, ab41037)), MMP3 (1:500 dilution, Proteintech, 17873-1-AP), MMP13 (1:500 dilution, Proteintech, 18165-1-AP), SOX9 (1:1000 dilution, Cell Signaling Technology, 82630), COL2A1 (1:1000 dilution, Abcam, ab34712), HOXD11 (1:500 dilution, Proteintech, 18734-1-AP), and GAPDH (1:1000 dilution, Proteintech, 60004-1-Ig) were diluted at appropriate concentration and incubated with membranes at 4 °C overnight. Horseradish peroxidase (HRP)-conjugated anti-rabbit immunoglobulin G (Santa Cruz Biotechnology, sc-2004) or anti-mouse IgG (Santa Cruz Biotechnology, sc-2005) was diluted at 1:3000 and incubated with the membranes at RT for one hour. LI-COR Imaging System (Biosciences, USA) was utilized to visualize protein bands. The band intensities were assessed utilizing ImageJ analysis (National Institutes of Health, USA). Each gene’s band intensity was normalized to GAPDH.

Cell transfection

GenePharma (Shanghai, China) supplied HOXD11 overexpressing lentivirus (OE-HOXD11) and control vector. Lentivirus was transfected into human chondrocytes based on manufacturer’s instruction. Cells were transfected for 48 h, and then, following experiments were performed.

Immunohistochemistry and histomorphometry analysis

The OA and normal human articular cartilage were treated with 10% buffered formalin for fixation and decalcified with 10% (w/v) EDTA; pH 7.4 for three weeks and then embedded with paraffin. Sections were cut into 6 μm thickness as well as further dyed with anti-SOX9 (1:500 dilution, Cell Signaling Technology, 82630), anti-HOXD11 (1:300 dilution, Proteintech, 18734-1-AP) at 4 °C overnight, then stained with second antibodies conjugated with HRP (Santa Cruz Biotechnology, CA, USA). Microscope BX51 (Olympus, Japan) was utilized to obtain images. ImageJ analysis software was utilized to quantify the intensity of HOXD11 and SOX9 expression (National Institutes of Health, USA). For histomorphometry, six-micrometer-thick sections of knee cartilage were dyed with H&E, Safranin O and fast green (SOFG) to evaluate cartilage damage. ImageJ was used to quantify safranin O loss in relation to total cartilage. Histological assessment was evaluated based on Osteoarthritis Research Society International (OARSI) histological scoring system and performed by two blinded observers.

Analysis of DEGs

We searched for data sets related to OA in the GEO and obtained the GSE169077 data set, which contained knee cartilage samples from five normal humans as well as six OA patients. Then, we analyzed DEGs using the GEO2R tool attached to the GEO. As the criteria of |log2(FC)|> 1 and P < 0.05, DEGs were screened from the GSE169077 data set.

Statistical analysis

Data were presented as the means ± standard deviations. All the statistical analyses were performed utilizing SPSS 13.0 (SPSS, Chicago, USA). Unpaired Student’s t test was utilized for analyzing the differences between two groups. We used one-way analysis for variance and Tukey’s test for multiple comparisons. Probability (P) values > 0.05 were considered statistically significant.

Results

Inhibited proliferation and enhanced apoptosis in human OA chondrocytes

We isolated primary chondrocytes from OA and normal knee articular cartilage, then compared cell absorbance at 450 nm, to evaluate proliferation effects in OA chondrocytes (Fig. 1a, b). The results depicted inhibited proliferation in OA chondrocytes as relative to control group, a 45.75% reduction at 5 d and a 44.69% reduction at 7 d, they were both reduced significantly (P < 0.05). We further investigated the effects of cell apoptosis in OA chondrocytes using flow cytometry analysis (Fig. 1c, d). Results revealed that the apoptotic percentage was 15.32% in OA and 7.35% in the controls (P < 0.05). These results suggested that cell proliferation was suppressed, and apoptosis was improved in OA chondrocytes.

Fig. 1
figure 1

Cell proliferation and apoptosis in OA and normal chondrocytes. a OA and normal chondrocyte cells were cultivated in 60 mm plates (magnification, × 100). Scale bars: 200 μM. b Cell absorbances (450 nm) were estimated at 0, 1, 3, 5, and 7 d by CCK8 assay. c Flow cytometry with Annexin V-FITC/PI dual dying was utilized to quantify OA cells and normal chondrocytes apoptosis. d Apoptosis percentage between OA cells and normal chondrocytes. Data are presented as mean ± SD of at least three independent experiments. *P < 0.05 and **P < 0.01 as relative with control

Full size image

Reduced expression of specific factors in human OA chondrocytes

Cartilage-specific markers: SOX9 and COL2A1, have been identified as dominant transcription factors necessary for chondrogenesis [25]. In qRT-PCR assays, SOX9 and COL2A1 mRNA levels in OA were reduced than controls (P < 0.05 vs controls) (Fig. 2b). WB assays displayed that SOX9 and COL2A1 were reduced in OA compared to controls (P < 0.05) (Fig. 2a). These findings indicated that SOX9 and COL2A1 expressions were inhibited in OA chondrocytes.

Fig. 2
figure 2

SOX9 and COL2A1 expression in OA cells and normal chondrocytes. a WB was carried out to quantify SOX9 and COL2A1 protein levels in OA cells and normal chondrocytes. b qRT-PCR was performed to quantify SOX9 and COL2A1 mRNA level. *P < 0.05 and **P < 0.01 as relative with control

Full size image

For the expressions of ADAMTSs and MMPs, qRT-PCR and WB were utilized to study mRNA and protein levels. The findings exposed that mRNA levels of ADAMTSs (ADAMTS4, ADAMTS5) and MMPs (MMP3, MMP13) were upregulated in OA than in control (Fig. 3b). Moreover, WB assays exposed upregulated protein levels of ADAMTSs (ADAMTS4, ADAMTS5) and MMPs (MMP3, MMP13) in OA compared with control (P < 0.05) (Fig. 3a).

Fig. 3
figure 3

Expression of ADAMTSs (ADAMTS4 and ADAMTS5) and MMPs (MMP3 and MMP13) in OA cells and normal chondrocytes. a WB was carried out to evaluate ADAMTSs and MMPs in OA cells and normal chondrocytes. b qRT-PCR was performed to quantify mRNA level of ADAMTSs and MMPs. *P < 0.05 and **P < 0.01 as relative with control

Full size image

Downregulated HOXD11 in human OA chondrocytes

Given the importance of 5′-HOXD genes in chondrogenesis, we analyzed the DEGs in HOXD genes using the GEO data set. As shown in the Uniform Manifold Approximation and Projection (UMAP) diagram (Fig. 4a), the differences in gene expression levels between the knee cartilage samples of normal people and the OA patients can be distinguished well. Through the analysis of GEO2R, we obtained 720 genes upregulated and 810 genes downregulated in OA patients, shown in the volcano diagram (Fig. 4b). Given master involvement of HOXD genes in vertebrate limb development, we concentrated HOXD genes expression. In HOXD genes, the chip included HOXD4, HOXD9, HOXD11, and HOXD13 (Table 2). Among them, HOXD11 met the screening criteria for DEGs, the P value was 0.0161, and ǀlog2FCǀ was 1.31709. Therefore, HOXD11 was selected as an important DEG in our further analysis.

Fig. 4
figure 4

Identification of DEGs analysis. a UMAP plot of knee cartilage samples in GSE169077 data set. The purple dot represents knee cartilage samples of five normal humans, and the green dot represents knee cartilage samples of six OA patients. b Volcano plot of DEGs in GSE169077 data set. The red dot indicates elevated genes (n = 720), whereas blue dot indicates downregulated genes (n = 810)

Full size image
Table 2 The DEGs
Full size table

To verify the DEGs, we examined the effects of 5′-HOXD genes in OA chondrocytes, and qRT-PCR assays demonstrated that HOXD11 reduced significantly in OA as compared with control (P < 0.05) (Fig. 5a). mRNA levels of the other 5′-HOXD genes were not significantly different. To test the hypothesis that HOXD11 expression modulates cartilage degradation in OA, we also measured HOXD11 protein levels (Fig. 5b). WB showed a reduction in OA as compared with control (P < 0.05).

Fig. 5
figure 5

Expression of HOXD11 in OA cells and normal chondrocytes. a qRT-PCR was conducted to quantify 5′-HOXD mRNA levels (HOXD9, HOXD10, HOXD11, HOXD12, and HOXD13). b WB was conducted to quantify HOXD11 protein levels in OA cells and normal chondrocytes. *P < 0.05 and **P < 0.01 as relative with control

Full size image

HOXD11 overexpression reversed effect of human OA chondrocytes

To determine if HOXD11 exerts its role via regulating SOX9 expression, HOXD11 overexpression experiments were conducted. HOXD11 overexpression was achieved by lentivirus transfection according to qRT-PCR analysis (Fig. 6b). Proliferation impact in OA chondrocytes was reversed by HOXD11, as determined by CCK-8 analysis (Fig. 6c). Moreover, HOXD11 overexpression inhibited cell apoptosis in OA chondrocytes, as shown by flow cytometry (Fig. 6d). SOX9 and COL2A1 expressions were also detected. qRT-PCR and WB results suggested that HOXD11 enhanced SOX9 and COL2A1 expressions (Fig. 6e, f), which were initially reduced. Furthermore, qRT-PCR and WB were conducted to detect ADAMTS5 and MMP13 expressions. The findings exhibited that HOXD11 inhibited ADAMTS5 and MMP13 expression (Fig. 6g, h), which were initially upregulated.

Fig. 6
figure 6

Overexpression of HOXD11 in OA cells reverses the effects. a The cultivated chondrocyte cells in 60 mm plates (magnification, × 100). Scale bars: 200 μM. b qRT-PCR was performed to evaluate HOXD11 mRNA expression. c CCK-8 was conducted to evaluate proliferation. d Flow cytometry was performed to evaluate cell apoptosis. e WB was performed to evaluate expression of SOX9 and COL2A1. f qRT-PCR was conducted to evaluate SOX9 and COL2A1 mRNA levels. g WB was used to quantify MMP13 and ADAMTS5 expression. h qRT-PCR was utilized to quantify MMP13 and ADAMTS5 mRNA levels. *P < 0.05, **P < 0.01 versus control or OA + vector

Full size image

Formed aberrant subchondral bone and downregulated expression of SOX9 and HOXD11 in OA cartilage

To measure aberrant subchondral bone formation in OA cartilage, we performed the experiments of H&E staining and SOFG staining (Fig. 7a, b). The results showed that the OARSI histological score was 0.1667 in controls and 4.833 in the OA group (Fig. 7e). For proteoglycan loss (% relative to total) experiments, the results showed that 6.843 percentage of proteoglycan loss in the controls, and 37.9 percentage in OA cartilage (Fig. 7f) (Fig. 8).

Fig. 7
figure 7

Cartilage staining and immunohistochemistry assay of SOX9 and HOXD11 expression. a and b H&E and SOFG were used to stain the cartilage and assess the cartilage damage (original magnification × 40). Scale bars: 500 μM. c and d Immunohistochemistry of SOX9 and HOXD11 in human normal and OA knee joint cartilage (original magnification × 400). Scale bars: 50 μM. e OARSI histological score of OA severity in normal and OA knee joint cartilage. f Quantification of proteoglycan loss in normal and OA knee joint cartilage. g and h SOX9, HOXD11-positive cells were quantified in normal and OA knee joint cartilage. *P < 0.05, **P < 0.01 versus control

Full size image
Fig. 8
figure 8

Model for inhibition of HOXD11 promotes cartilage degradation and induces osteoarthritis development (red arrow). In OA chondrocytes, the decreased expression of HOXD11 inhibits cell proliferation and enhances apoptosis. HOXD11 downregulates SOX9 via inducing catabolism and restraining anabolism, leading to the upregulation of MMP13 and ADAMTS5, which downregulates COL2A1, ultimately promoting cartilage degradation and inducing osteoarthritis development. Lentivirus-mediated HOXD11 overexpression reverses these effects (green arrow)

Full size image

We further detected the expression of SOX9 and HOXD11 in knee joint cartilage utilizing immunohistochemistry. Findings displayed that SOX9 expression was significantly reduced in OA cartilage as relative with controls (P < 0.01) (Fig. 7c, g). Similarly, HOXD11 expression was reduced in OA cartilage (Fig. 7d, h).

Discussion

During OA development, chondrocyte is characterized by elevated apoptosis, cytokine production, and matrix degeneration [26,27,28,29,30]. Chondrocyte apoptosis in OA cartilage has been recognized as one of the most crucial factors in pathophysiology of OA illness process [26, 31]. This study illustrated significantly inhibited cell proliferation while markedly increased cell apoptosis in human OA chondrocytes (Fig. 1), indicating the participation of cell apoptosis in OA development, consistent with previous studies [32, 33].

SOX9 is a master transcription regulator required for chondrogenesis, which controls condensation as well as differentiation of mesenchymal cells by cooperating with other cartilage-specific factors, such as COL2A1 [25], aggrecan [34], and cartilage link protein [35]. COL2A1, regulated by SOX9, serves as the master component in ECM; thus, it is recognized as a target gene of SOX9 in vivo [36]. This study found that SOX9 and COL2A1 expression was noticeably decreased in human OA chondrocytes, demonstrating that the signaling pathway of SOX9-COL2A1 might have an essential role in cartilage degradation and OA development [24].

5′-HOXD genes roles in many diseases have been reported, especially in cancer and congenital malformations, and they are therefore emerging as novel pharmacological targets [37,38,39,40]. There is increasing evidence that HOXD9, one of the 5′-HOXD genes, has a role in development of normal joints in early stages and the pathological process of arthritis [22, 23, 41, 42]. Our previous study has found that HOXD9 has a close relationship with chondrogenesis by regulating SOX9 and its down-target, COL2A1 [24]. In this study, four DEGs in the GSE169077 dataset were screened in OA and normal cartilage samples based on the differential analysis in the GEO database. Based on the analysis (Fig. 4 and Table 2), HOXD11 was downregulated and identified as DE in OA. To verify whether HOXD11 was downregulated in OA chondrocytes, we detected the mRNA levels of 5′-HOXD genes utilizing qRT-PCR analysis. We found that HOXD11 mRNA levels were reduced in OA, which supported the findings of bioinformatic analysis (Fig. 5). WB and immunohistochemistry results also showed that HOXD11 protein levels decreased in OA chondrocytes, which corresponded with qRT-PCR results. Conclusively, HOXD11 was markedly decreased in the human OA chondrocytes in vitro, consistent with the cartilage-specific factors of SOX9 and COL2A1. To further understand the relationship between HOXD11 and SOX9-COL2A1, we performed the HOXD11 overexpression experiments and found an increased expression of SOX9 and COL2A1, which reversed downregulating effects. These results strongly show that HOXD11 may have a key involvement in cartilage degradation and the development of OA via the SOX9-COL2A1 signaling pathway. Interestingly, this inference differs from our previous conclusion about HOXD9 and cartilage formation [24]. We attribute the reason to the difference in study species and cell type. Besides, this inference just showed the overlapping function of the 5′-HOXD genes [21].

In the cartilage degradation of OA, MMPs and ADAMTSs are examples of matrix-degrading enzymes responsible for degrading ECM. Among the MMPs, MMP13 is primary collagenase in OA cartilage [43, 44]. Recently, research showed that ADAMTS5 is superior to ADAMTS4 at aggrecan breakdown of the experimental models and human OA cartilage [12]. We found the upregulated expression of MMPs (MMP3 and MMP13) and ADAMTSs (ADAMTS4 and ADAMTS5) in OA-cultured chondrocytes, similar to the reports before [31] (Fig. 3). In the HOXD11 overexpression experiments, we found ADAMTS5 and MMP13 expression were reduced, which reversed upregulating effects. These findings strongly show that HOXD11 may have a key involvement in cartilage degradation and HOXD11 overexpression may play an anti-inflammatory and anabolic effect in cultivated chondrocytes.

Nonetheless, this study had a number of limitations. Firstly, HOXD11 knockdown effect in OA chondrocytes was not identified, despite HOXD11 overexpression results were displayed in our study, which necessitates further research. Second, the molecular mechanism is relatively simple to explore in the present study. Whether the HOXD11 has a related signaling pathway and whether the pathway is related to osteoarthritis, it needs further research on the relationship between the HOXD11 and the signaling pathway.

To conclude, this study demonstrated inhibited expression of HOXD11 in human OA chondrocytes and cartilages, parallel with observed expression levels of SOX9 and COL2A1. Lentivirus-mediated overexpression of HOXD11 reversed OA effects on chondrocyte proliferation and apoptosis and reduced cartilage damage. Our findings proposed that HOXD11 has a critical role in OA development as well as HOXD11 overexpression could be effective therapeutic strategy for preventing OA cartilage degradation.