With Institutional Review Board approval, discarded cartilage samples were obtained from the knee joints of 13 OA patients aged 58 to 77 years (mean age, 66 ± 5.2 years; 11 female and two male Caucasians) who underwent joint replacement surgery at Palmetto-Richland Hospital, Columbia, SC, USA. The macroscopic cartilage degeneration was determined by staining of femoral head samples with India ink 36, and the cartilage with smooth articular surface (unaffected cartilage) was resected and used to prepare chondrocytes by enzymatic digestion as previously described 3132333437. Histological analysis of some of the unaffected cartilage samples was performed on 5-μm-thick sections stained with H & E and Safronin O and graded using the Mankin score 38. Grading of the histology slides (data not shown) revealed that all of the cartilage pieces taken from the unaffected area had a Mankin score <2 for structure and a Mankin score of 1 for cellularity. Isolated chondrocytes were plated at a density of 1 × 10⁶/ml in 35-mm tissue culture dishes (Corning, NY, USA) in complete Ham's F-12 medium as previously described 31. In some cases, OA chondrocytes passaged once were used.

AGE-BSA was prepared by reacting BSA (Sigma Chemical Co., St Louis, MO, USA) with glycoaldehyde (Sigma), according to the method described by Valencia and colleagues 21 with slight modifications. Briefly, endotoxin-free BSA (2 mg/ml) was incubated under nonreducing conditions with 70 mMfresh glycoaldehyde in PBS (pH 7.4) without calcium chloride and magnesium chloride for 3 days at 37°C. The reaction was terminated by removing nonreacted glycoaldehyde by dialyzing extensively against PBS.

Characterization of AGE-BSA was performed as previously described in detail 212223. Briefly, the AGE-specific absorbance was read at 340 nm 21 and AGE-specific fluorescence was detected at excitation/emission wavelengths of 360/430 nm 22, 330/395 nm, 365/440 nm, 485/530 nm, 280/350 nm and band widths set at ex40/em40 23. All fluorescence data are given normalized to the corresponding control. Absorbance and fluorescence were read using the Synergy HT microplate reader (Biotek Instruction, Winooski, VT, USA). The UV-visible absorption spectra of native BSA and of AGE-BSA were recorded on a Shimadzu UV-1800 Spectrophotometer. The electrophoretic migration of native and modified BSA samples was analyzed by reducing SDS-PAGE on 10% polyacrylamide gel with 4% stacking gel 39.

We first determined whether AGE-BSA and EGCG (purity ≥ 95%, catalogue number E4143; Sigma) affect the viability of OA chondrocytes in vitro. Human OA chondrocytes (1 × 10⁶/ml) were plated in 35-mm culture dishes (Becton-Dickinson, Franklin Lakes, NJ, USA) in complete Ham's F-12 medium and serum-starved for 12 hours/overnight. Chondrocytes were treated with various doses of AGE-BSA (20 to 600 μg/ml) and EGCG (25 to 200 μM), and after 24 hours the incubation cytotoxicity of AGE-BSA and EGCG was examined using the CytoTox-Glo™ Cytotoxicity Assay Kit (Promega, Madison, WI, USA). Primary chondrocytes were pretreated with different doses of EGCG for 1 or 2 hours prior to stimulation with AGE-BSA. Chondrocytes cultured without AGE-BSA or EGCG served as controls. All experiments were performed within 4 days of the primary culture to avoid dedifferentiation of OA chondrocytes.

Real-time quantitative PCR was used to quantify the expression of mRNA for TNFα and MMP-13 with expression of GAPDH as endogenous control. Total RNA was separated from OA chondrocytes by the Quick Gene automated system according to the manufacturer's instruction (Quick Gene, Holliston, MA, USA). First-strand cDNA was synthesized using 1 μg total RNA and the SuperScript First Strand cDNA synthesis kit (Invitrogen, Carlsbad, CA, USA). Primers used for PCR assisted amplification were: TNFα (NM_000595: forward, 5'-AGG ACG AAC ATC CAA CCT TCC CAA-3'; reverse, 5'-TTT GAG CCA GAA GAG GTT GAG GGT-3'), MMP-13 (NM_002427: forward, 5'-CGC CAG AAG AAT CTG TCT TTA AA-3'; reverse, 5'-CCA AAT TAT GGA GGA GAT GC-3'), and GAPDH (NM_002046.3: forward, 5'-TCG ACA GTC AGC CGC ATC TTC TTT-3'; reverse, 5'-ACC AAA TCC GTT GAC TCC GAC CTT-3'). PCR amplification was carried out using the core kit for SYBR Green (Applied Biosystems, Foster City, CA, USA) and the Step One Real Time PCR System (Applied Biosystems). Typical profile times used were an initial step of 95°C for 10 minutes, followed by a second step at 95°C for 15 seconds and 60°C for 60 seconds for 40 cycles with melting curve analysis. The level of target mRNA was normalized to the level of GAPDH and compared with control (untreated sample). Data were analyzed using the ΔΔCT method 40.

To analyze the transcription of type-2 collagen (COL2A1), type-10 collagen (COL10A1), aggrecan (ACAN), and SRY-box containing gene 9 (SOX-9), total cellular RNA was prepared using a commercially available kit (Qiagen, Valencia, CA, USA). First-strand cDNA was synthesized using 1 μg total RNA and the SuperScript First Strand cDNA synthesis kit (Invitrogen). cDNAs of COL2A1, COL10A1, ACAN, and SOX-9 were used as positive controls in the PCR reaction and were generously provided by Dr Thomas Herring (Department of Orthopaedics, Case Western Reverse University School of Medicine, Cleveland, OH, USA). RT-PCR was carried out utilizing the PTC-100™ Peltier Thermal Cycler (MJ Research, Ramsey, MN, USA). Primers used for PCR-assisted amplification were: COL2A1 (NM_001844: forward, 5'-ACG TGA AAG ACT GCC TCA GC-3'; reverse, 5'-TTT CAT CAA ATC CTC CAG CC-3'; expected size of the DNA fragment, 352 bp), COL10A1 (NM_000493: forward, 5'-TGA TCC TGG AGT TGG AGG AC-3'; reverse, 5'-GAG ATC GAT GAT GGC ACT CC-3'; expected size of the DNA fragment, 703 bp); ACAN (NM_013227: forward, 5'-GAC CTG CAA GGA GAC AGA GG-3'; reverse, 5'-CCA CTG GTA GTC CTT GGG CAT-3'; expected size of the DNA fragment, 256 bp), and SOX-9 (NM_011448: forward, 5'-GAT TTT TCA CGC AGC CCT AA-3'; reverse, 5'-ATA CAG TCC AGG CAG ACC CA-3'; expected size of the DNA fragment, 637 bp). The reaction was cycled according to the following scheme: 40 minutes at 72°C, followed by 40 cycles of 1 minute at 95°C, 1 minute at 60°C, and 2 minutes at 72°C, followed by a final 15-minute extension. The amplified products were electrophoresed on 1.2% agarose gels in TAE buffer and were visualized by ethidium bromide staining.

Briefly, OA chondrocytes were stimulated with AGE-BSA (600 μg/ml) for 24 hours with or without pretreatment with EGCG. TNFα present in the culture medium was quantified using the TNFα-specific ELISA according to the manufacturer's instructions (R&D Systems, Minneapolis, MN, USA). Plates were read at 450 nm using a Synergy HT microplate reader (Biotek Instrument, Winooski, VT, USA).

Stimulated and control chondrocytes were washed with cold PBS and lysed using the previously described cell lysis buffer (50 mM Tris-HCl, pH 7.4; 150 mM NaCl; 1% Triton X-100; 0.1% SDS; 0.5% sodium deoxycholate; 1 mM EDTA; 1 mM EGTA; Complete^® protease and phosphatase inhibitors) 41. Cytoplasmic and nuclear fractions were prepared as previously described 42. Total lysate or nuclear/cytoplasmic fraction protein (45 μg/lane) or concentrated cell culture supernatant was resolved by SDS-PAGE (10% resolving gel with 4% stacking) and was transferred to nitrocellulose membranes (Bio-Rad, Hercules, CA, USA). Membranes were blocked with blocking buffer containing nonfat dry milk powder in Tris-buffered saline and 0.1% Tween-20. Blots were probed with 1:200 to 1:1,000 diluted primary antibodies specific for the target protein. Immunoreactive proteins were visualized using 1:1,000 diluted horseradish peroxidase-linked secondary antibodies and enhanced chemiluminescence (GE Healthcare, Milwaukee, WI, USA) 43. Images were captured using AFP-Imaging System (Minimedical Series, Elms Ford, NY, USA). Anti-MMP-13 antibody (sc-30073) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA), anti-NF-κB p65 antibody (IMG-150) was from Imgenex (San Diego, CA, USA), and anti-IκBα antibody (#9242), anti-phospho-p38 MAPK antibody (#9215S), anti-p38 MAPK antibody (#9212), anti-phospho-SAPK/JNK antibody (#9251S), anti-SAPK/JNK antibody (#9252), anti-phospho ERK P44/42 MAPK antibody (#9101S) and anti-ERK P44/42 MAPK antibody (#9102) were from Cell Signaling Technology (Amherst, Beverly, MA, USA).

Gelatin zymography was performed essentially as previously described 41. Briefly, an equal volume of cell culture supernatant was mixed with nonreducing sample buffer (4% SDS, 0.15 M Tris (pH 6.8), and 20% (volume/volume) glycerol containing 0.05% (weight/volume) bromophenol blue) and resolved on a 10% polyacrylamide gel containing copolymerized 0.2% gelatin (Bio-Rad). Commercially available MMP-13 (catalogue number PF094; EMD Chemicals, Gibbstown, NJ, USA) was used as a positive control. The product used was supplied as concentrated conditioned medium, and 2.5 μl was used. After electrophoresis, gels were washed twice, for 15 minutes each time, with 2.5% Triton X-100. Digestion was carried out by incubating the gel in the gelatinase buffer (50 mM Tris-HCl (pH 7.6), 10 mM CaCl₂, 50 mM NaCl, and 0.05% Brij-35) at 37°C for 16 hours. The gel was stained with 0.1% Coomassie brilliant blue R350 (GE Healthcare, Piscataway, NJ, USA), and the locations of gelatinolytic activity were revealed as clear bands on a background of uniform light blue staining. Electrophoretic migration of MMP-13 in the supernatant was compared with known molecular weight standards and also with clear bands of MMP-13 activity produced by the positive control. After development, gel images were captured and the clear bands were analyzed using Un-Scan-It software (Silk Scientific Corporation, Orem, UT, USA).

Activation of NF-κB p65 in human OA chondrocytes pretreated with EGCG and then stimulated with AGEs was determined using a highly sensitive and specific Transcription Factor ELISA Kit according to the instructions of the manufacturer (catalogue number EK1121; Panomics, Fremont, CA, USA). The color intensity was read at 450 nm using the Synergy HT ELISA plate reader (Biotek).

The effect of EGCG on purified IκB kinase (IKK) activity was determined with a highly sensitive HTScan^® IKKβ Kinase Assay Kit according to the instructions of the manufacturer (catalogue number 7549; Cell Signaling Technology). The kit provides a means of performing kinase activity assays with recombinant human IKKβ kinase. It includes active IKKβ kinase (supplied as GST fusion protein), a biotinylated peptide substrate, and a phospho-serine antibody for detection of the phosphorylated form of the substrate peptide. Purified IKKβ kinase was pretreated with different doses of EGCG (5 to 200 μM) 5 minutes prior to the treatment of substrate peptide. The assay was performed on a 96-well high-binding streptavidin-coated plate, and the absorbance (A) of each well was read at 450 nm using the Synergy HT ELISA plate reader (Biotek). Each kinase assay was performed in triplicate. The percentage inhibition of IKKβ kinase activity was calculated using the formula:

Measurements of the scanned bands were performed using Un-Scan-It software. Each band was scanned five times, and the mean band intensity (pixels per band) was obtained. Data were normalized to suitable loading controls and expressed as the mean ± standard deviation followed by appropriate statistical analysis.

All measurements were performed in duplicate and were repeated at least three times using different age-matched and sex-matched OA samples. All statistical analyses were performed using Origin 6.1 software package (one-paired two-tailed t test with one-way analysis of variance and Tukey's post-hoc analysis) and P < 0.05 was considered significant. Values shown are the mean ± standard error of the mean unless stated otherwise.

Glycoaldehyde-induced modifications of BSA were studied by UV-visible absorption spectroscopy, gel electrophoresis, and AGE-specific fluorescence. The absorption spectra of AGE-BSA (AGEs) revealed a marked UV hyperchromicity (74.4%) at 280 nm (Figure 1a). For further characterization of in vitro-produced AGEs, samples were analyzed by SDS-PAGE under reducing conditions. An accompanying decrease in the electrophoretic mobility and smear towards higher molecular size ranges from 65 to 98 kDa was noted, indicating a higher level of glycoaldehyde-mediated modifications (Figure 1b). AGE-specific absorbance was detected at 340 nm 21 and was found to be significantly high in glycoaldehyde-treated BSA as compared with native BSA (P < 0.001). AGE-specific fluorescence was detected at the excitation/emission wavelengths described elsewhere 2223 and were found to be very high for glycoaldehyde-treated BSA compared with native BSA, signifying the presence of a higher level of modifications. Characterization of AGE-specific modifications on BSA resulting from incubation with glycoaldehyde is summarized in Table 1.

We determined whether primary human OA chondrocytes and chondrocytes passaged once used in these studies maintained their phenotype – by analyzing the expression of type-2 collagen, aggrecan, and SOX-9 mRNA, which are considered to be signatures of the chondrogenic phenotype 44. Our results show that primary or passage-1 human OA chondrocytes in monolayer culture maintained their phenotype, when they were plated (1 × 10⁶/ml) in 35-mm culture dishes in complete Ham's F-12 medium with 10% FCS and were allowed to grow at 37°C and 5% CO₂ in a tissue culture incubator, as judged by the continued expression of COL2A1 (Figure 2a), aggrecan, and SOX-9 mRNAs, whereas COL10A1 mRNA was not expressed (Figure 2b). Based on these data, OA chondrocytes were used within 72 hours after plating to avoid possible dedifferentiation.

OA chondrocytes were treated with increasing doses of AGE-BSA (20 to 600 μg/ml) and the mRNA expression of TNFα and MMP-13 was determined using a highly sensitive and specific quantitative RT-PCR method. We found AGE-BSA dose-dependently induced TNFα and MMP-13 mRNA expressions, and the maximum stimulation of chondrocytes was found to be at 600 μg/ml AGE-BSA (data not shown). Based on these data, we used a concentration of 600 μg/ml AGE-BSA for stimulation of chondrocytes in all of our experiments.

OA chondrocytes (80% confluent) were pretreated with EGCG (25 to 150 μM) for 2 hours, and were then stimulated with in vitro-generated AGE-BSA (600 μg/ml) for 8 hours. No cytotoxic effect of EGCG was noted at the dose used (data not shown). The level of TNFα mRNA was quantified by a highly sensitive and specific quantitative RT-PCR method, and values were compared with control. Our results showed that OA chondrocytes treated with AGE-BSA had a higher level of TNFα mRNA compared with unstimulated OA chondrocytes. TNFα mRNA levels showed a marked decline, however, in the samples pretreated with EGCG and then stimulated with AGE-BSA (Figure 3a). To determine whether inhibition of gene expression also affected the protein level, culture supernatants were assayed for TNFα protein using TNFα-specific ELISA. As shown in Figure 3b, pretreatment with 25 to 150 μM EGCG significantly decreased the AGE-BSA-induced TNFα production in the culture supernatant of AGE-BSA-stimulated OA chondrocytes. It is also to be noted that maximum suppression was observed in cultures treated with 75 μM EGCG, after which no further decline was found (Figure 3b).

Pretreatment of OA chondrocytes with EGCG inhibited AGE-BSA-induced gene expression of MMP-13 in a dose-dependent manner as determined by quantitative RT-PCR (Figure 4a). Again the maximum inhibitory effect of EGCG was found at 75 μM concentration. We also determined the effect of EGCG on the production of MMP-13 in the culture supernatant by western immunoblotting using anti-MMP-13 antibody (Figure 4b). Analysis of the immunoblot revealed that the levels of MMP-13 were high in the supernatant of chondrocytes treated with AGE-BSA when compared with the levels detected in untreated chondrocytes culture, where MMP-13 was barely detectable. Importantly, the AGE-BSA-stimulated increase in MMP-13 expression was inhibited by EGCG to less than the basal level (Figure 4b). These results were further verified using gelatin zymography (Figure 4c). Zymographic analysis showed that OA chondrocytes stimulated with AGE-BSA had enhanced levels of active MMP-13 compared with the levels detected in control OA chondrocytes. Importantly, chondrocytes pretreated with different doses of EGCG and then stimulated with AGE-BSA showed reduced levels of active MMP-13 in the culture supernatant. Taken together these data indicate that the AGE-BSA-induced production of active MMP-13 was inhibited by EGCG in human OA chondrocytes (Figure 4c).

Activation of MAPKs is intimately associated with the expression of proinflammatory mediators 26. To determine whether the inhibition of TNFα and MMP-13 expressions was due to EGCG-mediated inhibition of the MAPK pathway, we examined the effect of EGCG on the activation of MAPKs in AGE-BSA-stimulated human OA chondrocytes. OA chondrocytes were pretreated with EGCG (25 to 150 μM) for 1 hour and then stimulated with AGE-BSA for 45 minutes, and the cell lysate was analyzed by western immunoblotting. Pretreatment of chondrocytes with EGCG attenuated the AGE-BSA-induced phosphorylation of p38-MAPK, JNK-MAPK and to a lesser extent of ERK-MAPK (Figure 5a,b). To further strengthen the relation of p38, JNK and ERK inhibition by EGCG and proinflammatory cytokine TNFα and MMP-13 expressions in AGE-BSA-stimulated human OA chondrocytes, we used pharmacological agents that inhibit p38-, JNK- and ERK-MAPKs.

Treatment of OA chondrocytes with the selective p38 inhibitor SB202190 (100 μM), the JNK inhibitor SP600125 (10 μM), and the ERK inhibitor PD98059 (50 μM) blocked the AGE-BSA-induced TNFα mRNA expression as determined by quantitative RT-PCR, but the effect was more pronounced in the case of p38-MAPK and JNK (Figure 3a). AGE-BSA-induced MMP-13 mRNA expression was significantly inhibited by SB202190 and SP600125 (P < 0.05), but the affect was more pronounced when SB202190 was used. Pretreatment of OA chondrocytes with PD98059 had no effect on MMP-13 mRNA expression in AGE-BSA-stimulated OA chondrocytes (Figure 4a). These data support the contention that inhibition of AGE-BSA-induced TNFα (Figure 3a) and MMP-13 (Figure 4a) expressions by EGCG in OA chondrocytes was mediated, at least in part, by the inhibition of AGE-induced activation of the p38-MAPK and JNK-MAPK pathways.

NF-κB is an important transcriptional regulator of inflammatory cytokine gene expression and plays a crucial role in immune and inflammatory response. After the ubiquitination and phosphorylation of IκBα, the inhibitor is degraded and the NF-κB is translocated to the nucleus, where it binds and activates the promoter of target genes. To investigate the mechanism responsible for the inhibitory effect of EGCG on proinflammatory cytokine gene expression, we examined the effect of EGCG on NF-κB activation and translocation to the nucleus using western blotting. Stimulation of OA chondrocytes with AGE-BSA induced the degradation of IκBα and nuclear translocation of NF-κB p65 (Figure 6a,b). Pretreatment with EGCG (25 and 75 μM) inhibited the AGE-BSA-induced degradation of IκBα and nuclear translocation of NF-κB p65 (Figure 6a,b). To determine whether EGCG also inhibits AGE-BSA-induced DNA binding activity of NF-κB, we used the Transcription Factor ELISA kit. Exposure of OA chondrocytes to AGE-BSA significantly enhanced the DNA binding activity of NF-κB p65 compared with controls (P < 0.0001), and increasing doses of EGCG (25 to 75 μM) significantly reduced the AGE-BSA-induced DNA binding activity of NF-κB p65 (P < 0.05) (Figure 6c).

To further strengthen the relation of inhibition of the NF-κB pathway and the expression of TNFα in our studies, we next investigated the effect of a pharmacological agent (MG-132, a known inhibitor of NF-κB) on the expression of TNFα and MMP-13. Treatment of chondrocytes with the proteasome inhibitor MG-132 (100 μM) significantly blocked the AGE-BSA-induced TNFα expression (P < 0.05) (Figure 3a), whereas MMP-13 mRNA expression was also inhibited by MG-132 but the inhibition was not statistically significant (P > 0.05) (Figure 4a). Together these results suggest that EGCG exerts its inhibitory effect on TNFα expression in AGE-BSA-stimulated OA chondrocytes via modulation of the activation and DNA binding activity of NF-κB.

The effect of EGCG on the phosphorylating activity of IKKβ kinase was determined using an HTScan^® IKKβ Kinase Assay Kit (Cell Signaling Technology). Purified IKKβ kinase was pretreated with different doses of EGCG (5 to 200 μM) 5 minutes prior to incubation with the substrate peptide. Figure 6d shows that IKKβ kinase activity was significantly inhibited by EGCG treatment (P < 0.001). A maximum of 85% inhibition of enzymatic activity was observed with 5 μM EGCG, after which no significant inhibition (P > 0.05) of enzyme activity was observed (Figure 6d). These data suggest that EGCG suppresses the activation of NF-κB by inhibiting the enzyme activity of IKK complex.