The human DP1 promoter region spanning nucleotides −1080 to +1 was amplified by PCR using genomic DNA from normal human articular chondrocytes as a template and forward and reverse PCR primers carrying KpnI and XhoI recognition sequences, respectively (Table 1). The amplified DNA fragment was digested with KpnI and XhoI (Thermo Fisher Scientific, Mississauga, ON, Canada), purified, and cloned into the pGL-3-basic luciferase reporter vector (Promega, Madison, WI).

DNA fragments corresponding to the DP1 promoter region from −550/+1, −374/+1, −250/+1, −170/+1 and −120/+1 with a KpnI restriction site at the 5′end and an XhoI restriction site at the 3′end, were synthesized (GeneWiz, Cambridge, MA) and subcloned into the pGL-3-basic vector. A set of mutant promoter constructs bearing mutation in each, two or all three putative Sp1 binding sites, was also generated by direct DNA synthesis (GeneWiz). The mutated nucleotide sequences are shown in Figure 2A. Similarly, a DNA fragment corresponding to the DP1 promoter region from −374 to +1 was synthesized (GeneWiz) and inserted at the KpnI/XhoI sites of the pCpGL plasmid, a CpG-free luciferase reporter vector (Gebhard et al., 2010). All generated promoter constructs were verified by DNA sequencing.

Transient transfection experiments were performed using the Lipofectamine 3,000 transfection reagent (Thermo Fisher Scientific) following the manufacturer’s instructions. Briefly, the immortalized human chondrocytes, C28/I2 (kindly provided by Dr. Mary Goldring, Hospital for Special Surgery, NYC), were seeded at a 3.0 × 10⁴ cells/well density in 24-well plates, 1 day before transfection. Cells were transfected with 400 ng of the various DP1 reporter constructs and 10 ng of Renilla Luciferase Control Reporter plasmid (Promega, Madison, WI). In co-transfection assays, the expression vector for Sp1 (generously provided by Dr. Stephen Smale, UCLA) was used (10 ng/well). Forty hours post-transfection, cells were washed with PBS and lysed in 100 μL passive lysis buffer (Promega). Luciferase activity (20 μL of cell lysate) was measured using the Dual-Luciferase reporter assay system (Promega) following the manufacturer’s instructions. Each transfection was performed in triplicate, and each experiment was repeated at least three times.

Nuclear extracts were prepared and quantified as described previously (Cheng et al., 2004; Chabane et al., 2008). Oligonucleotides (Table 1) were synthesized and biotinylated on their 5′-end by Integrated DNA Technologies (Coralville, IA) and then annealed into double strands. Binding reactions were performed using a LightShift Chemiluminescent EMSA kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Briefly, 5 μg of nuclear extract was incubated with 20 fmol of the biotin-labeled probe in 1X binding buffer containing 50 ng/μL poly dI-dC, 2.5% glycerol, 5 mM MgCl2 and 0.05% NP40 in a final volume of 20 μL at room temperature for 20 min. In cold competition assays, 100-fold molar excess of cold wild-type or mutant oligonucleotide was used. In supershift assays, nuclear extracts were incubated with 1 μg of anti-Sp1 antibody (MilliporeSigma, Oakville, ON, Canada) for 15 min at room temperature before the addition of the probe. Binding complexes were resolved on a non-denaturing 6% polyacrylamide gel, transferred to a positively charged membrane (Invitrogen), and cross-linked with UV irradiation. Detection of biotin-labeled DNA was performed using the Chemiluminescent Nucleic Acid Detection Module (Thermo Fisher Scientific).

ChIP assays were performed using the Magna ChIP Assay kit (MilliporeSigma) in accordance with the manufacturer’s protocol. Briefly, cells were treated with formaldehyde to cross-link DNA and associated proteins. Cell lysates were sonicated to shear chromatin to fragments between 200–500 bp. The chromatin samples were diluted 10-fold in ChIP dilution buffer, and an aliquot was saved as the input DNA. The DNA-protein complexes were immunoprecipitated with an anti-Sp1 antibody (MilliporeSigma). Normal rabbit IgG was used as the negative control. DNA fragments were eluted from each immunoprecipitation and subjected to PCR analysis. Fold enrichment was assessed using qRT-PCR and the 2^−ΔΔCT method.

The primers used are listed in Table 1.

Macroscopically normal human knee cartilage was obtained at autopsy from 11 donors (5 females and 6 males) with no history of joint disease or trauma (mean age ±SD, 66 ± 10 years). The normality of the joint and integrity of cartilage were confirmed macroscopically at the time of sample collection. OA cartilage was obtained from 28 OA patients (17 females and 11 males) who underwent total knee replacement surgery (mean age ±SD, 70 ± 9 years). All patients had a Kellgren-Lawrence grade ≥3. Demographic characteristics of the donors are shown in Table 2. Informed consent was obtained from each donor, and the local ethics committee approved the study. In normal donors, cartilage samples were collected from the medial femoral condyle. In OA donors, cartilage was also collected from the medial femoral condyle, but only from macroscopically unaffected regions. Tissues were taken from the three (superficial, intermediate, and deep) layers of articular cartilage.

Cartilage samples were cut into small fragments, divided into portions for DNA and RNA extraction, flash-frozen in liquid nitrogen, and stored at −80°C until use. Samples were ground in liquid nitrogen and DNA, or RNA was isolated from each sample, using the DNeasy plant maxi kit or the RNAeasy Mini kits (Qiagen, Toronto, ON, Canada), respectively.

RNA was reverse transcribed and amplified using a QuantiTect RT-PCR kit (Qiagen) on a Rotor-Gene 3,000 real-time PCR system (Corbett Research), according to the manufacturer’s protocol. All values were normalized to the mean CT value of two reference genes, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and hypoxanthine-guanine phosphoribosyltransferase 1 (HPRT1). Data were expressed as relative gene expression using the 2^−ΔΔCT method or as fold change using the 2^−ΔΔCT method, as previously described (Livak and Schmittgen, 2001; Zayed et al., 2008a; Ouhaddi et al., 2017). Each PCR was performed in triplicate from 3 independent experiments. Primers are listed in Table 1.

Genomic DNA (500 ng) isolated directly from knee cartilage and from cultured chondrocytes was bisulfite converted using an EpiTect Bisulfite Kit (Qiagen) according to the manufacturer’s instructions. PCR was performed using the TaKaRa EpiTaq HS kit (Cedarlane, Burlington, ON, Canada) following the supplier’s instructions. The size and purity of the PCR products were confirmed by agarose gel electrophoresis. Pyrosequencing was carried out using a PyroMark Q24 system/instrument (Qiagen). PCR primers and sequencing primers (Table 1) were designed using PyroMark Assay Design 2.0 software (Qiagen). The methylation percentage at each CpG site of the DP1 promoter was determined using the PyroMark Q24 software version 2.0.6.

Primary human articular chondrocytes were released from cartilage by sequential enzymatic digestion with 2 mg/mL pronase (MilliporeSigma) for 1 h and 1 mg/mL collagenase type II (Thermo Fisher Scientific) for 6 h at 37°C. Cells were seeded at 1.2 × 10⁴ cells/cm² in 25 cm² flaks in DMEM supplemented with 10% (heat-inactivated) FCS and then cultured at 37°C for 72 h. Cells were treated either with the vehicle or with 5 μM 5-Aza-2′-deoxycytidine (5-Aza-dC) for 5 weeks, adding freshly prepared drug (5-Aza-dc) twice weekly. Finally, cells were washed with ice-cold PBS, and total RNA and DNA were isolated. Gene expression and DNA methylation were evaluated as described above.

The pCpGL-DP1-374/+1 luciferase reporter construct was methylated using the CpG methyltransferase, M. SssI (New England Biolabs, Whitby, ON, Canada). Briefly, plasmid DNA was incubated with 4 units of methyltransferase per μg DNA in the presence of 160 μM S-adenosylmethionine (SAM) for 4 h at 37°C. After the first 2 hours, the methylation reaction was supplemented with a further dose of 160 μM SAM. Mock methylation reactions in which SssI methyltransferase was omitted were performed in parallel. Successful methylation was verified by digestion with the restriction enzyme Hpa II (New England Biolabs). Methylated and mock-methylated plasmids were purified using a QIAquick PCR Purification Kit (Qiagen).

Data from transfection assays and real-time PCR analyses in cultured cells were assessed by one-way ANOVA, followed by either the Tukey or the Dunnett’s post hoc test for comparison of multiple groups. Comparison between two groups was performed using the Student’s t-test or chi-squared test, where appropriate. Comparison of the level of CpG methylation between groups was assessed using the Student’s t-test with Bonferroni correction for multiple testing. Relative DP1 expression in control and OA cartilage were compared using the Mann-Whitney U test. p-value <0.05 was considered significant. All analyses were performed using Prism 9.0 (GraphPad).

To characterize the transcriptional mechanisms underlying DP1 expression, a 1080-bp fragment containing the 5′-flanking region upstream of the translation start site (ATG) of the human DP1 gene (GenBank accession number NG_012118.1) was cloned into the pGL3-basic luciferase vector. As shown in Figure 1A, this region lacks a TATA box and contains GC-rich sequences. DNA sequence analysis revealed four putative Sp-1 binding motifs between positions −206 and −97 (Figure 1A).

To delineate the minimal DP1 promoter region required for basal DP1 promoter activity, a series of luciferase report constructs harboring various lengths of the DP1 promoter were generated. These constructs were then transiently transfected into the C28/I2 chondrocyte cell line, which expresses DP1 constitutively. As shown in Figure 1B, the fragments −1080/+1, −550/+1, −374/+1 and −250/+1 had virtually similar promoter activity, showing an increase in luciferase activity of 31-, 32-, 36- and 33-fold, respectively when compared to the empty pGL3-Basic vector. 5’ deletion of the −250/+1 construct from −250 to −170, and −120, gradually and drastically reduced luciferase activity, suggesting that the region between −250 and −120 contains essential element(s) for transcriptional activity of the human DP1 promoter. The observation that deletion of this region while maintaining the upstream and downstream portions, resulted in an 80% decrease in the promoter activity (Figure 1C) further supports that the primary promoter regulatory elements are located between −250 and −120 bp.

The −250/-120 fragment of the DP1 promoter includes three putative binding sites for Sp1, two of which overlap, and are defined as Sp1-1 (−206 to −196), Sp1-2a (−142 to −132) and Sp1-2b (−139 to −129) (Figure 2A). To determine whether the three putative Sp1 binding sites are essential for the transcriptional activation of DP1, we conducted transient transfection experiments with the WT DP1 promoter (−374/+1) construct or reporter constructs carrying a mutation in each, two, or all three Sp1 sites. As shown in Figure 2A, the single mutant constructs Sp1-1, Sp1-2a, and Sp1-2b displayed 43%, 36%, and 32% reduced activity, respectively, compared with the WT construct. The double mutants Sp1-1,2a (Sp1-1 and Sp1-2a mutated) and Sp1-1.2b (Sp1-1 and Sp1-2b mutated) displayed reduced activity by 68% and 64%, respectively, exhibiting an additive effect. The double mutant Sp1-2a, 2b (Sp1-2a and Sp1-2b mutated) displayed 38% reduced transcriptional activity. This value is virtually similar to that observed with individual mutants indicating, as expected with overlapping sites, that the effects of both mutations were not additive. Mutation of all sites reduced the activity by ∼80% (Figure 2A), suggesting that the three sites are essential for the regulation of the DP1 promoter.

Next, we tested the effect of Sp1 overexpression on the DP1 promoter activity. Cells were co-transfected with the WT DP1 promoter (−374/+1) construct, or the mutant constructs (Sp1-1, Sp1-2a, Sp1-2b, and Sp1-1,2a, 2b), and an expression vector encoding Sp1. As shown in Figure 2B, overexpression of Sp1 activated the wild-type DP1 promoter. The response of the single mutant constructs Sp1-1, Sp1-2a and Sp-1.2b to Sp1 overexpression was reduced by 41%, 32% and 26%, respectively, compared with the WT construct. The double mutant constructs Sp1-1,2a and Sp1-1.2b also displayed reduced response to Sp1 overexpression by 72% and 61%, respectively. The triple mutant Sp1-1,2a, 2b did not respond at all to Sp1 overexpression (Figure 2B).

To further assess the role of Sp1 in DP1 expression, cells were transfected with the DP1 promoter (−374/+1) construct and treated with increasing concentrations of mithramycin A (Mith), an Sp1 inhibitor. As shown in Figure 2C, treatment with Mith suppressed transactivation of the DP1 promoter by Sp1, in a dose dependent manner. Additionally, Mith treatment dose-dependently downregulated DP1 mRNA expression in primary OA chondrocytes (Figure 2D). These findings suggest that Sp1 is important for transcriptional regulation of DP1.

To determine whether Sp1 binds to the predicted binding sites at the DP1 promoter, we performed EMSA analysis using the biotin-labeled oligonucleotides encompassing the putative binding sites, and nuclear extracts from OA chondrocytes. As shown in Figure 3A, DNA-protein complexes were formed with the probe encompassing Sp1-1 (left panel, lane 2). These complexes disappeared in the presence of a 100-fold molar excess of unlabeled wild-type probe (lane 3) and a similar excess of a Sp1 consensus oligonucleotide (lane 4). Incubation with the unlabeled mutant oligonucleotide probe has no effect on the formation of these complexes (lane 5). The biotin-labelled probe encompassing Sp1-2a and Sp1-2b also formed DNA-protein complexes when incubated with nuclear extracts from chondrocytes (right panel, lane 2). Formation of these complexes was prevented in the presence of a 100-fold molar excess of an unlabeled WT probe (lane 3) or a Sp1 consensus oligonucleotide (lane 7). Since Sp1-2a and Sp1-2b overlap, we conducted competition experiments using oligonucleotides in which each Sp1 motif was individually mutated and one in which both motifs were simultaneously mutated. Competition with these oligonucleotides revealed that oligonucleotide mutated for Sp1-2a or Sp1-2b partially abolished the formation of the DNA-protein complex (lanes 4 and 5), whereas the double mutant oligonucleotide displayed no effect (lane 6). Supershift analyses revealed that DNA-complexes formed with both probes were shifted by anti-Sp1 antibody. No supershift was observed with the control IgG (Figure 3A).

We also performed ChIP assays to determine whether Sp1 binds to the DP1 promoter in vivo. Cross-linked chromatin from OA chondrocytes was immunoprecipitated with an anti-Sp1 antibody or a normal rabbit IgG (negative control). DNA isolated from the immunoprecipitates was analyzed by PCR using primers sets specific to each of the two regions containing the putative Sp1 binding sites. ChIP-qPCR assays showed that the promoter region encompassing the Sp1-1 site and the region encompassing both Sp1-2a and Sp1-2b sites were enriched in immunocomplexes using an anti-SP1 antibody by 15- and 12-fold, respectively, while no enrichment was detected in those using normal IgG. These results confirmed that Sp1 can bind to the endogenous DP1 promoter.

As mentioned above, the 5′-flanking region of the DP1 gene contains a CpG-rich area, suggesting that methylation of CpG dinucleotides might be involved in regulating DP1 expression. To test this possibility, we investigated the effect of the demethylating agent 5-Aza-dc on DP1 mRNA levels in human primary OA articular chondrocytes. DP1 mRNA levels in 5-Aza-dc treated cells were increased by 3.3-fold compared to control cultures (Figure 4A). Treatment with 5-Aza-dc had no effect on the mRNA and protein levels of Sp1 (data not shown), suggesting that the upregulation of DP1 expression by 5-Aza-dc was independent of an increase in Sp1 levels.

To evaluate whether the enhanced levels of DP1 mRNA in 5-Aza-dc-treated cells was associated with DNA demethylation, the methylation status of the DP1 promoter was quantified by bisulfite pyrosequencing. Methylation analysis performed with the corresponding samples indicates that the increased expression of DP1 was accompanied by reduced methylation levels of the CpG motifs contained in the minimal promoter region (−250 to −120 bp) (Figure 4B). Representative pyrograms of the 8 CpG motifs in region −249 to −177 are shown in Figure 4C.

These results suggest that demethylation of the DP1 promoter may contribute to the regulation of DP1 expression in chondrocytes.

To determine whether CpG methylation of the DP1 promoter alters its transcriptional activity, the 374-bp (−374/+1) promoter fragment was cloned into the CpG-free luciferase vector pCpGL (pCpGL-374/+1). The reporter construct was methylated in vitro using CpG methyltransferase (M.SssI). Successful methylation was confirmed by enzymatic digestion with HpaII, a methylation-specific restriction enzyme. C28/I2 cells were transfected with the methylated or unmethylated (control) construct and luciferase activity was quantified. The results showed that methylation of the DP1 promoter dramatically decreased basal DP1 promoter activity when compared with the unmethylated one (Figure 5A). Sp1-induced activation of the DP1 promoter was also suppressed by DNA methylation (Figure 5A). These results demonstrate that DNA methylation downregulates the DP1 promoter activity.

To investigate whether DNA methylation affects the binding of Sp1 to its binding sites at the DP1 promoter, we performed EMSAs using oligonucleotide probes in which the CpG sites within the Sp1 binding sites were unmethylated or methylated. As reported above, DNA–protein complexes formed when nuclear extracts from chondrocytes were incubated with unmethylated probes (Figure 5B, lane 2 in the left and right panels). Similar EMSA band patterns and intensities were observed in the presence of methylated probes (Figure 5B , lane 7 in the left and right panels). The addition of a 100-fold molar excess of unlabeled probe prevented the formation of the DNA-protein complex (lanes 3 and 8). Incubation with an anti-Sp1 antibody supershifted the protein-DNA complex (lanes 5 and 10) consistent with Sp1-specific binding to both probes. These results suggest that CpG methylation at the Sp1 binding site did not affect the binding of Sp1 to the DP1 promoter.

DNA methylation was reported to modulate the expression of many genes involved in the pathophysiology of OA (Cheung et al., 2009; Hashimoto et al., 2009; Loeser et al., 2009; Reynard et al., 2014; Takahashi et al., 2015). We therefore evaluated the methylation status of the −250/−120 bp proximal promoter region of DP1 in genomic DNA isolated directly from normal (n = 11) and OA (n = 14) cartilage. Pyrosequencing analysis revealed no difference in the methylation level at any CpG site analyzed between OA and normal cartilage (Figure 6A). Typical pyrosequencing results of the DP1 promoter region from −249 to −177 are shown in Figure 6B.

We also characterized the expression levels of DP1 in normal and OA cartilage. First, we compared histologic features in normal and OA cartilage. Safranin O staining showed that glycosaminoglycan (GAG) content was markedly reduced in OA cartilage than in normal cartilage. Moreover, OA cartilage displayed surface irregularities and small fibrillations, while normal cartilage appeared intact (Figure 7A). Next, we evaluated the expression level of OA related gene markers in normal and OA cartilage. As previously reported (Yang et al., 2016; Zhou et al., 2020), the mRNA expression of ACAN and COL2 were significantly decreased in OA cartilage compared with normal cartilage, whereas MMP-13 and ADAMTS5 mRNA expression were elevated in OA tissues (Figure 7B). As shown in Figure 7C, real-time qPCR analysis revealed no difference in DP1 mRNA expression level between OA and normal cartilage. These data suggest that the methylation status and the expression level of DP1 are not dysregulated in OA cartilage.