Normal human osteoblasts were purchased from PromoCell GmbH (Heidelberg, Germany) and were cultured according to the manufacturer's specifications. OA osteoblasts were isolated from trabecular bone specimens from patients suffering from advanced OA and undergoing primary total knee replacement. The experimental protocol was approved by the Research Ethics Board at Sacre-Cæur Hospital of Montreal. The osteoblast cell cultures were prepared as already described[11]. Briefly, trabecular bone samples were cut into small pieces of 2 mm² prior to their sequential digestion in the presence of 1 mg/ml collagenase type I (Sigma-Aldrich Canada Ltd., Oakville, ON, Canada) in BGJb media (Invitrogen Canada Inc., Burlington, ON, Canada) without serum at 37°C for 30, 30, and 240 minutes. After being washed with the same media, the digested bone pieces were cultured in 25 cm² plastic cell culture flasks (Corning Incorporated, Corning, NY, USA) with BGJb media containing 20% foetal bovine serum (FBS) (Invitrogen Life Technologies). This medium was replaced every 2 days until cell outgrowths appeared around the explants. At confluence, cells were split once and plated at 50,000 cells per cm² in culture plates (Falcon, Lincoln Park, NJ, USA) with Ham's F-12/Dulbecco's modified Eagle's medium (HAMF-12/DMEM) (Sigma-Aldrich Canada Ltd.) containing 10% FBS and 50 mg/ml ascorbic acid and grown to confluence again. Only first-passage cells were used in our experiments.

Normal and OA osteoblasts were incubated for 24 hours with or without increasing concentrations of H₂O₂ (1 to 100 μM). Total cellular levels of HNE/protein adducts were assessed in cellular extracts of osteoblasts using an in-house enzyme-linked immunosorbent assay (ELISA) as previously described [5].

Osteoblasts were incubated for 24 hours in HAMF-12/DMEM containing 2% charcoal-stripped FBS, which yields maximal stimulation of ALPase activity and OC secretion. Cells were then incubated for 48 hours in the same medium in the presence of increasing concentrations of HNE (0 to 20 μM). The medium was collected at the end of the incubation and frozen at -80°C prior to assay. Cells were washed twice with phosphate-buffered saline, pH 7.4, and solubilised in ALPase buffer (100 mM glycine, 1 mM MgCl₂, 1 mM ZnCl₂, 1% Triton X-100; pH 10.5) for 60 minutes with agitation at 4°C. Cellular ALPase activity was determined as the release of p-nitrophenol hydrolysed from p-nitrophenyl phosphate (12.5 mM final concentration) at 37°C for 30 minutes after cell solubilisation in ALPase buffer as described above. Protein determination was performed by the bicinchoninic acid method[14]. Nascent OC was determined by a specific enzyme immunoassay (Biomedical Technologies, Inc., Stoughton, MA, USA). The detection limit of this assay is 0.5 ng/ml, and 2% charcoal-treated FBS contains less than 0.1 ng/ml of OC.

For the HNE dose-response curves, osteoblasts were incubated in 0.5% FBS/HAMF-12/DMEM for 48 hours with increasing concentrations of HNE (0 to 20 μM). After incubation, the culture medium was collected and the IL-6 and PGE₂ levels were determined using specific commercial kits from R&D Systems, Inc. (Minneapolis, MN, USA) and Cayman Chemical Company (Ann Arbor, MI, USA), respectively, according to the manufacturers' specifications. The sensitivities of the assays were 3 and 9 pg/ml, respectively.

Osteoblasts were incubated in fresh medium containing 0.5% FBS/HAMF12/DMEM in the presence of increasing concentrations of HNE (0 to 20 μM) for 24 hours or in the presence of 20 μM of HNE for increasing periods of incubation. Twenty to 50 μg of cellular protein extract was subjected to discontinuous 4% to 12% SDS-PAGE under reducing conditions and transferred onto nitrocellulose membrane (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The membranes were immersed overnight at 4°C in a blocking solution consisting of TTBS (20 mM Tris, pH7.4, 150 mM NaCl, 0.1% Tween 20) and 5% skim milk and incubated again overnight in blocking buffer containing the polyclonal rabbit anti-COX-2 or anti-Col I (1:1,000 dilution; Oncogene Research Products, San Diego, CA, USA). The membranes were then washed three times with TTBS and incubated for 1 hour at 22°C with the second antibody (anti-rabbit immunoglobulin G-horse radish peroxidase; New England Biolabs Ltd., Mississauga, ON, Canada) and washed again. Detection was carried out using Supersignal west dura extended duration substrate (Pierce Biotechnology, Inc., Rockford, IL, USA). Membranes were prepared for autoradiography and exposed to clear-blue x-ray film (Pierce) and then subjected to a digital imaging system (Bio-Rad Laboratories, Inc.). For the total and phosphorylated level of mitogen-activated protein kinases (MAPKs) (p38, c-Jun NH₂-terminal kinase [JNK] 1/2, and extracellular signal-regulated kinase [ERK] 1/2) as well as transcription factors (activating transcription factor-2 [ATF-2], CRE-binding factor-1 [CREB-1], and IκBα), we used specific PhosphoPlus kits (New England Biolabs Ltd.).

Total RNA was extracted from OA osteoblasts using TRIzol^® reagent (Invitrogen Life Technologies) according to the manufacturer's recommendations. The RNA was quantitated using the RiboGreen RNA quantitation kit (Molecular Probes, now part of Invitrogen, Carlsbad, CA, USA), dissolved in RNase-free H₂O, and stored at -80°C until use. One microgram of total RNA was reverse-transcribed using Moloney murine leukaemia virus reverse transcriptase (Fermentas Canada Inc., Burlington, ON, Canada) as detailed in the manufacturer's guidelines. One fiftieth of the reverse transcriptase reaction was analysed by real-time quantitative polymerase chain reaction (PCR). The nucleotide sequence of primers are shown below:

ALPase [15]: 5'-CCCAAAGGCTTCTTCTTG-3' (sense)

5'-CTGGTAGTTGTTGTGAGCAT-3' (anti-sense),

OC [15]: 5'-ATGAGAGCCCTCACACTCCTC-3' (sense)

5'-GCCGTAGAAGCGCCGATAGGC-3' (anti-sense),

Col I α 1 [16]: 5' CATCCTCGACGGCATCTCAGC-3' (sense)

5'-TTGGGTCAGGGGTGGTTATTG-3' (anti-sense),

IL-6 [17]: 5'-TTCAAATGAGATTGTGGGAAAATTGCT-3' (sense)

5'-AGTTCATCTCTGCCTGAGTATCTT-3' (anti-sense),

COX-2 [18]: 5'-TTCAAATGAGATTGTGGGAAAATTGCT-3' (sense)

5'-AGTTCATCTCTGCCTGAGTATCTT-3' (anti-sense), and

glyceraldehyde-3-phosphate dehydrogenase (GAPDH) [19]:

5'-CAG AAC ATC ATC CCT GCC TCT-3' (sense)

5'-GCT TGA CAAAGT GGT CGT TGA G-3' (anti-sense).

Quantitative PCR analysis was performed in a total volume of 50 μl containing template DNA, 200 nM of sense and antisense primers, 25 μl of SYBR^® Green master mix (Qiagen Inc., Mississauga, ON, Canada), and uracil-N-glycosylase (UNG) (0.5 Units; Epicentre Biotechnologies, Madison, WI, USA). After incubation at 50°C for 2 minutes (UNG reaction) and at 95°C for 10 minutes (UNG inactivation and activation of the AmpliTaq Gold enzyme), the mixtures were subjected to 40 amplification cycles (15 seconds at 95°C for denaturation and 1 minute for annealing and extension at 60°C). Incorporation of SYBR^® Green dye into PCR products was monitored in real time using a GeneAmp 5700 Sequence detection system (Applied Biosystems, Foster City, CA, USA) allowing determination of the threshold cycle (C_T) at which exponential amplification of PCR products begins. After PCR, dissociation curves were generated with one peak, indicating the specificity of the amplification. The C_T value was obtained from each amplification curve using the software provided by the manufacturer (Applied Biosystems). Preliminary experiments showed that the amplification efficiency of COX-2, ALPase, OC, Col α1, IL-6, and GAPDH was similar.

Relative amounts of mRNA in normal and OA cartilage were determined using the standard curve method. Serial dilutions of internal standards (plasmids containing cDNA of target genes) were included in each PCR run, and standard curves for the target gene and for GAPDH were generated by linear regression using log (C_T) versus log (cDNA relative dilution). The C_T values were then converted to number of molecules. Relative mRNA expression in cultured chondrocytes was determined using the ΔΔC_T method, as detailed in the manufacturer's guidelines (Applied Biosystems). A ΔC_T value was first calculated by subtracting the C_T value for the housekeeping gene GAPDH from the C_T value for each sample. A ΔΔC_T value was then calculated by subtracting the ΔC_T value of the control (unstimulated cells) from the ΔC_T value of each treatment. Fold changes compared with the control were then determined by raising 2 to the ΔΔC_T power. Each PCR reaction generated only the expected specific amplicon as shown by the melting-temperature profiles of the final product and by gel electrophoresis of test PCR reactions. Each PCR was performed in triplicate on two separate occasions for each independent experiment.

OA osteoblasts were incubated with HNE alone or in combination with 1 ng/ml tumour necrosis factor-α (TNF-α) for 1 hour. Nuclear extracts were prepared and electrophoretic mobility shift assay (EMSA) was performed as previously described[20]. Double-stranded oligonucleotide probes for CRE (5'-AGAGATTGCCTGACGTCAGAGAGCTAG-3') and NF-κB (5'-AGTTGAGGGGACTT TCCCAGGC-3') were end-labeled with [γ-³²P]-ATP using a kit (Promega Corporation, Madison, WI, USA). The binding reactions were conducted with 5 μg of nuclear extract and of 2 × 10⁵ cpm of [γ-³²P]-labeled oligonucleotide probe at 22°C for 20 minutes in a final volume of 10 μl, and complexes were resolved on non-denaturing 6% polyacrylamide gels. Then, gels were fixed, dried, and exposed to clear-blue x-ray film (Pierce).

Supershift assays were performed as described above with nuclear extracts from cells treated with HNE (20 μM) or TNF-α (1 ng/ml) for 1 hour. Two micrograms of the antibodies were added to the shift reaction mixture 20 minutes after the incubation period, followed by another incubation at 4°C overnight. The antibodies were specific for the transcription factors ATF-2, p65, and p50 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA).

The human COX-2 promoter constructs used included a wild-type (WT) (-415)-Luciferase (Luc) COX-2 promoter plasmid, mutated ATF/CRE (-58/-53) (-415)-Luc COX-2 promoter plasmid, and mutated NF-κB (-223/-214) (-415)-Luc COX-2 promoter plasmid, as previously described [21]. Expression vectors for WT (pCMV-Flag-p38) and dominant negative (DN) (pCMV-Flag-p38) p38 MAPK were a kind gift from Dr. R.J. Davies (University of Massachusetts). Expression vector for IKKα was generously given by Dr. M. Karin (University of California). A pCMV-β-galactosidase (pCMV-β-gal) reporter vector was purchased from Promega Corporation.

Human MG-63 osteoblast-like line cells (American Type Culture Collection, Manassas, VA, USA) (approximately 50% confluence) were transiently transfected in 12-well cluster plates using lipofectamine 2000™ reagent methods (Invitrogen Life Technologies) according to the manufacturer's protocol. Briefly, transfections were conducted for 6 hours with DNA lipofectamine complexes containing 10 μl of lipofectamine reagent, 2 μg DNA plasmid, and 0.5 μg of pCMV-β-gal (as a control of transfection efficiency). After washing, medium was replaced by a fresh medium containing 1% FBS and experiments were performed in this medium supplemented with the factors under study. For promoter study, Luciferase activity was determined in cellular extracts by a kit (Luciferase Assay System; Promega Corporation) using a microplate luminometer (Applied Biosystems) and normalised to β-gal level, which was quantified by a specific ELISA (Roche Diagnostics Canada, Laval, QC, Canada). To study the effect of p38 MAPK and IKKα overexpression on PGE₂ and IL-6 production, cells were transfected with the appropriated WT p38 MAPK, DN p38 MAPK, or IKKα expression vector as described above and then culture medium was collected for PGE₂ and IL-6 assay as described above.

The data are expressed as the mean ± standard error of the mean. Statistical significance was assessed by unpaired Student t test, and P < 0.05 was considered significant.

To provide evidence that HNE production was increased during OA development, the level of this aldehyde was determined in cellular extract of normal and OA osteoblasts. As shown in Figure 1a, HNE/protein adduct levels were 1.4-fold higher in OA cells compared with the normal (p ≤ 0.05). To confirm that OA osteoblasts are able to produce HNE under oxidative stress, cells were incubated for 24 hours with increasing concentrations of H₂O₂ and levels of HNE/protein adducts were quantified in cellular extracts. The data showed that H₂O₂ at different concentrations induces the formation of HNE/proteins adducts in OA osteoblasts in a dose-dependent manner (Figure 1b).

Because ALPase, OC, and Col I are the principal biomarkers of osteoblasts and considered to be good indicators for bone formation and metabolic activity, we tested the ability of HNE to alter their expression and, in turn, the phenotype of the osteoblasts. Figure 2 depicts the variation in osteoblast production of ALPase (Figure 2a,b), OC (Figure2c,d), and Col I (Figure 2e,f) after HNE incubation. Compared with control, HNE dose-dependently inhibited significantly osteoblast ALPase activity by 19.6%, 25.4%, and 32.1% (p < 0.01) in the presence of 5, 10, and 20 μM of HNE, respectively (Figure 2a). However, ALPase mRNA expression was significantly inhibited only at 20 μM HNE (20%; p < 0.05) (Figure 2b).

In contrast to the inhibition of ALPase, OC protein level was increased significantly in the presence of HNE (15%, 25%, and 20% at 5, 10, and 20 μM, respectively; p < 0.05) (Figure 2c). OC mRNA levels were also increased at different concentrations of HNE, with a maximum stimulation of 155% at 5 μM HNE (p < 0.01) (Figure 2d).

Finally, we explored the effect of HNE on Col I, which constitutes 90% of the total organic ECM in mature bone. Our data showed that HNE increased Col I protein expression by factors of 2.4, 2.1, 4.6, and 8.4 at concentrations of 1, 5, 10, and 20 μM, respectively (Figure 2e), although this inductive effect of HNE was not manifested at the mRNA level (Figure 2f).

To determine whether HNE is a modulator of IL-6 production, osteoblasts were incubated with 0 to 20 μM of HNE for 48 or 4 hours for IL-6 protein and mRNA determination, respectively. As shown in Figure 3, there was a significant dose-dependent inhibition of IL-6 protein release (Figure 3a) and mRNA expression (Figure 3b) after incubation of osteoblasts with increasing concentrations of HNE. To test the combined effect of HNE with TNF-α, osteoblasts were preincubated with HNE (20 μM) for 30 minutes and then stimulated with TNF-α (1 ng/ml). Compared with untreated cells, TNF-α significantly induced IL-6 release by 549% (Figure 3c). This induction was completely inhibited in the presence of 20 μM HNE (47% of untreated cells). At mRNA level, HNE also showed a significant (approximately 70%) decrease of TNF-α-induced IL-6 mRNA expression (Figure 3d).

To better characterise the properties of HNE cell signaling in OA osteoblasts, we evaluated COX-2 gene expression and PGE₂ production in response to HNE stimulation. Compared with untreated cells, PGE₂ level was increased significantly by 209%, 240%, 551%, and 2,434% at concentrations of 1, 5, 10, and 20 μM HNE, respectively (Figure 4a). The increase of PGE₂ production related directly to an increase in COX-2 protein and mRNA OA osteoblasts. The protein and mRNA levels were increased in a dose-dependent manner by incubation of cells with HNE (0 to 20 μM), with a maximal stimulation at 20 μM HNE (8.5- and 4.6-fold, respectively) (Figure 4b,c).

To delineate the signaling pathways involved in HNE-induced COX-2 expression in pilot experiments, we used cell-permeable chemical inhibitor of p38 MAPK, SB202190. This inhibitor had no effect on the basal PGE₂ release (data not shown). As shown in Figure 4d, HNE significantly induced PGE₂ release by 340% in comparison with untreated cells. However, the p38 MAPK inhibitor significantly reduced HNE-stimulated PGE₂ production. Identical results were obtained with COX-2 protein level (data not shown).

Next, we examined whether HNE can also modulate the activity of ALPase as well as the production of OC, PGE₂, and IL-6 in normal osteoblasts. In a similar pattern, our data showed that HNE at 20 μM inhibits ALPase activity (Figure 5a) and IL-6 production (Figure 5c) but, in contrast, induced OC (Figure 5b) and PGE₂ (Figure 5d) release.

To gain insight into the signaling pathway activated by HNE in human OA osteoblasts, we first examined the HNE-induced phosphorylation patterns of MAPKs over increasing periods of time. Our data indicated that HNE stimulated p38 MAPK phosphorylation within 5 minutes and remained in a phosphorylated state for 120 minutes (Figure 6a). JNK2 (p46) was phosphorylated in a time-dependent manner, reaching a maximum between 5 and 30 minutes and returning to the basal level at 60 minutes. HNE had no effect on the ERK1/2 phosphorylation levels. No change in the total protein level of MAPKs was noted (data not shown).

Next, we investigated the effect of HNE on p38 MAPK downstream transcription factors CREB-1 and ATF-2. Our data showed that exposure of 20 μM HNE resulted in an early phosphorylation of ATF-2 and CREB-1 after 5 minutes of incubation (Figure 6a). We also examined the effect of HNE on the total NF-κB/p65 and phosphorylated IκBα. Our data showed that HNE had no significant effect on the basal level of the phosphorylated IκBα and cytosolic and nuclear NF-κB/p65 in osteoblasts (data not shown). However, combined with TNF-α, HNE inhibited strongly NF-κB/p65 protein translocation in the nucleus in a dose-dependent manner (Figure 6b).

To explore the effect of HNE on DNA-binding activity of ATF/CRE and NF-κB, OA osteoblasts were incubated for 60 minutes with 20 μM HNE or 1 ng/ml TNF-α. The latter was used as a positive control of NF-κB activation. EMSA data showed that HNE increased the DNA-binding activity of ATF/CRE to 170% compared with unstimulated cells (Figure 6c). However, TNF-α (but not HNE) induced the DNA-binding activity of NF-κB by 160% (Figure 6d). Basal and induced binding was displaced by adding 50-fold excess cold ATF/CRE and NF-κB oligonucleotide (competition).

To further identify the ATF/CRE and NF-κB protein complexes that bind on these motifs, specific anti-ATF-2, anti-p65, and anti-p55 antibodies were added to the shift reaction mixture. As illustrated in Figure 6c,d, the ATF-2 was supershifted in HNE-treated osteoblasts, and the p65 and p50 proteins were supershifted in TNF-α-treated osteoblasts.

To examine for elements of transcriptional control of the COX-2 gene stimulated by HNE, we conducted transient transfection analyses with a WT (-415)-Luc COX-2 promoter construct harbouring enhancer elements [21-23] for critical transcription factors, including an ATF/CRE site (-58 to -53). Data showed that HNE (20 μM) as well as TNF-α (1 ng/ml) upregulated the COX-2 promoter activity by 7.8- and 8.4-fold, respectively, compared with control (Figure 7). The mutation of the ATF/CRE site decreased the basal COX-2 promoter activity. In addition, their inducibility by either HNE or TNF-α was completely abrogated. However, the mutation of proximal NF-κB site (-223/-214) in the human COX-2 promoter construct was without effect in terms of basal and HNE-stimulated luciferase activity.

Finally, for a better understanding of the role of IKKα in NF-κB-mediated IL-6 production, constitutive activated IKKα was overexpressed in MG-63 osteoblast-like cells and then cells were incubated with TNF-α, HNE, or TNF-α combined with HNE. Our data showed that IKKα overexpression stimulated IL-6 production in the presence of 1 ng/ml TNF-α, an effect completely abrogated by HNE (Figure 8a). These results indicated that the inhibition of the IKKα pathway is the major regulator of the IL-6 response to HNE.

To further confirm that p38 MAPK plays a principal role in mediating HNE-induced COX-2 in osteoblasts, we transfected expression vectors of WT p38 MAPK and DN p38 MAPK followed by HNE stimulation. Overexpression of WT p38 plasmid markedly increased PGE₂ production, and HNE treatment further enhanced PGE₂ release compared with control cells (Figure 8b). However, the overexpression of DN p38 MAPK abrogated this effect.