β-Endorphin Regulation of MAPKs in Cultured Human Articular Chondrocytes: MAPK Inhibitors Prevent the Increase of IL-1β Protein Levels During β-Endorphin Stimulation

Abstract

We investigated the effect of β-endorphin on the activities of mitogen-activated protein kinases in cultured human articular chondrocytes in order to elucidate its effect on cartilage. Monolayer cultures of chondrocytes obtained from patients undergoing total knee arthroplasty were treated with 60, 600, or 6000 ng/ml β-endorphin, or 100 ng/ml naltrexone combined with 600 ng/ml β-endorphin. The regulation of three major mitogen-activated protein kinases phosphorylation, ERKp44/p42, p38, and JNK, was determined by Western blotting. We also examined the influence of specific mitogen-activated protein kinase inhibitors on IL-1β protein levels during β-endorphin stimulation. The results demonstrate that β-endorphin, dependent on concentration and duration of stimulation, significantly affected the activation of the three mitogen-activated protein kinases in cultured human articular chondrocytes. Naltrexone in some cases significantly regulated the mitogen-activated protein kinases in different ways when added to β-endorphin 600 ng/ml. Furthermore, specific mitogen-activated protein kinase inhibitors hindered the increase of IL-1β during β-endorphin incubation. The effect of β-endorphin seen in this study is considered critical for the production of several mediators of cartilage damage in an arthritic joint.

Keywords:

• chondrocytes
• monolayer
• μ-opioid receptor
• β-endorphin
• MAPKs
• osteoarthritis

INTRODUCTION

Mitogen-activated protein kinase (MAPK) activation is reportedly a critical event in the pro-inflammatory signaling cascades induced by cytokines in synoviocytes and chondrocytes. These cascades lead to the production of several mediators of cartilage damage in arthritic joints (27). An in vivo study showed that selective inhibitors of extracellular signal-regulated kinase (ERK) p44/p42, such as PD 198306, partially decrease the development of some structural changes in experimental osteoarthritis (OA). This effect was associated with reduced ERKp44/p42 (also known as ERK1/2) activity in OA chondrocytes, which probably explains the action of the drug (25). MAPK activity appears to be critical for regulating chondrocyte and synoviocyte apoptosis and matrix metalloproteinase (MMP) genes (22).

Recently, β-endorphin (β-END) has been detected not just in the central nervous system (CNS), where it has various functions (30), but also in the periphery regulating processes such as cell differentiation, cell communication, and modulation of inflammation (4). β-END binding sites have been detected in rat chondrocytes (6), and β-END has been found in articular chondrocytes, synovial membrane, and fibrous joint capsule of dogs (16), suggesting the possible existence of an opiate modulation of articular cartilage. β-END had an anti-inflammatory effect when injected into injured canine joint (23). Simultaneously, it has been recently demonstrated by our group that β-END increase the IL-1β protein levels in human articular chondrocytes (2). There is no evidence that human articular chondrocytes produce β-END (1), but they can be affected by β-END via a μ -opioid receptor (9).

The effect of β-END stimulation on cultured human articular chondrocytes expressing μ -opioid receptor has been studied by examining the phosphorylation of three major MAPKs: ERKp44/p42, p38, and JNK. Furthermore, we have observed the effect of specific MAPK inhibitors on IL-1β protein levels during β-END incubation.

MATERIALS AND METHODS

Chondrocytes/Cell Cultures

Human chondrocytes were obtained from patients undergoing total knee arthroplasty due to OA. The patients participated with informed consent and the study was approved by the Regional Ethical Committee. Cell cultures from three different patients were used. The experiments were performed separately for each patient. Biopsies were taken from the mainly weight-bearing areas, where the cartilage had a quality appropriate for harvesting (5, 11, 18).

Cartilage biopsies were kept in 0.9% NaCl for approximately 2 h, and then cut into 1–1.5 mm³ pieces. These were incubated for 18 h in 2–5 ml DMEM/HAM'S F-12 (Cat. No. T 481-50, BioChrom Labs, Terre Haute, IN) containing collagenase (Cat. No. C-9407, Sigma-Aldrich Norway AS, Oslo, Norway) at a final concentration of 0.8 mg/ml. The enzyme solution was removed by centrifugation at 200 g and by consecutive washing steps with DMEM/HAM'S F-12 (containing 10% serum during the first wash) and the pellet was resuspended in fresh growth medium (DMEM/HAM'S F-12 supplemented with 20% human serum obtained from healthy volunteers). Cultures were further expanded by trypsinization (Cat. No. T-3924, Sigma), and after repeated washing were resuspended in DMEM/HAM'S F-12 supplemented with 10% serum. Experiments were performed when an approximate number of 2–3 × 10⁶ cells per patient was achieved.

Detection of Activated MAPKs by Western Blot Analysis

Human chondrocytes were seeded in a 6-well plate, at approximately 0.5 × 10⁶ cells per well. They were incubated for at least 48 h in DMEM/HAM'S F-12 medium (BioChrom Labs) supplemented with gentamicin (Cat. No. G-1264, Sigma-Aldrich), amphotericin B (Cat. No. A-2942, Sigma-Aldrich) and 10% filter-sterilized human serum. The cells reached confluence before the experiments. In view of the cytotoxicity of amphotericin B, the manufacturer's recommendations were strictly followed (10 ml/L, concentration established as safe in cell culture applications). Thereafter, the cells were washed twice with PBS, and fresh serum-free medium containing human β-END (β-lipotropin 61-91, Cat. No. E 6261, Sigma-Aldrich) at 60 ng/ml, 600 ng/ml, and 6000 ng/ml was added to three separate wells. The fourth well contained 600 ng/ml β-END and 100 ng/ml naltrexone (competitive antagonist of β-END; Cat. No N-3136, Sigma-Aldrich). The fifth well was a control. After 15 min, 8 h, and 24 h incubation the cells were lysed with NuPAGE LDS sample buffer (Invitrogen): they were scraped from the wells, transferred to 1.5 ml Eppendorf tubes, kept on ice, and sonicated for 2 × 10 s. The lysates were frozen (−20°C) and SDS−PAGE was run using NuPAGE 10% Bis-Tris Gel (Invitrogen) and NuPAGE MOPS SDS Running Buffer (Cat. No. NP0001, Invitrogen). The manufacturer's instructions were followed for the use of each separate antibody mentioned later in the text. The proteins were electro-transferred on to microporous PVDF membranes, which were blocked for 1 h, washed 3 × 5 min in TBST (10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Tween-20), and incubated with primary antibodies (see below) overnight at 4°C. After washing, immuno-complexes were detected using the anti-rabbit Chemiluminiscent Phosphotope® -HRP Western Blot Detection Kit (Cat. No. 9190, Cell Signaling Technology, Beverly, MA). The following pairs of rabbit polyclonal antibodies were used: anti-ACTIVE® MAPK (also known as extracellular signal-regulated kinase p44/p42, i.e., ERKp44/p42) antibody (Cat. No. V8031, Promega), which detects the dually phosphorylated, active form of MAPK, and anti-ERK 1/2 antibody (Cat. No. V1141, Promega) that detects ERK1 and ERK2 in the nonphosphorylated, monophosphorylated and dually phosphorylated forms; the anti-ACTIVE® JNK antibody (Cat. No. V7931, Promega), which recognizes the dually phosphorylated, active form of JNK, and anti-JNK1 pan antibody (Cat. No. 44-690, BioSource International, Inc., Camarillo, CA) that determines total levels of JNK1 protein; finally the anti-ACTIVE® p38 antibody (Cat. No. V1211, Promega), which detects the active form of p38, and p38 MAPK [pTpY180/182] Western Blotting Kit (Cat. No. BHO0011, Biosource), containing both p38 MAPK [pTpY180/182] phosphorylation site specific antibody and p38 MAPK pan antibody. The membranes were stained with Ponceau S (Cat. No. P-3504, Sigma-Aldrich) prior to antibody incubation. Western blots were first probed for phosphorylated ERKp44/p42, p38 and JNK MAPKs, stripped using standard protocols and then reprobed for total ERKp44/p42, p38 and JNK MAPKs.

Immuno-reactivity was quantified with the LUMI-IMAGER F1TM system (Boehringer, Manheim, Germany) and analyzed by LumiAnalystTM (Version 3.0 for Windows) computer software.

Determination of IL-1β Protein Levels by ELISA

Human chondrocytes were seeded in a 6-well plate, at approximately 0.5 × 10⁶ cells per well. They were incubated for at least 48 h in DMEM/HAM'S F-12 medium (BioChrom Labs) supplemented with gentamicin (Cat. No. G-1264, Sigma-Aldrich), amphotericin B (Cat. No. A-2942, Sigma-Aldrich), and 10% filter-sterilized human serum. The cells reached confluence before the experiments. In view of the cytotoxicity of amphotericin B, the manufacturer's recommendations were strictly followed (10 ml/L, concentration established as safe in cell culture applications). Thereafter, the cells were washed twice with PBS, and fresh medium containing 0.5% human serum was added for another 24 h. On the day of the experiments, the cells were washed twice with PBS, and new serum-free medium containing 20 μ M PD98059 (specific ERK1/2 inhibitor, Cat. No. P1019, A.G. Scientific, Inc., San Diego, CA), 10 μ M SB203580 (p38 inhibitor, Cat. No. S1014, A.G. Scientific), and 100 μ M SP600125 (specific JNK inhibitor, Cat. No. S2022, A.G. Scientific) were added to three separate wells and incubated for 1 h. The manufacturer's recommendations concerning effective concentrations and incubation time were followed. Control group remained untreated. After 1 h of pre-incubation with specific MAPK inhibitors, the cells were washed twice with PBS, and new serum-free medium containing 6000 ng/ml human β-END was added both to the wells pre-incubated with specific MAPK inhibitors and to a separate well previously untreated with one of MAPK inhibitors. Control group remained again untreated. Following 1 h of incubation with β-END, the supernatants were aspirated and briefly centrifuged. Subsequently, the samples were frozen and stored at −20°C until further experiments.

Repeated thaw and freeze cycles were avoided. Human IL-1β ELISA Kit (Cat. No. 88-7010, eBioscience, San Diego, CA) was used for protein level determination by following the manufacturer's instructions. Absorbance readings were done by using a plate reader with a 450 nm filter.

STATISTICS

Two-way ANOVA was performed to analyze differences between samples. A p-value of < 0.05 was considered significant. Results are represented as mean ± standard error of the mean (SEM).

RESULTS

The results demonstrate that β-END, dependent on concentration and time of incubation significantly influences the activation of three MAPKs, p38, ERKp44/p42, and JNK in cultured human articular chondrocytes (Figure 1). ERKp44/p42 phosphorylation was significantly increased during 15 min incubation (Figure 2) at concentrations of 60 ng/ml (p = 0.009) and 600 ng/ml (p = 0.009), and also in the group treated with the competitive antagonist naltrexone in addition to 600 ng/ml β-END. Furthermore, an increase in ERKp44/p42 activation was noticed after 8 h incubation (Figure 2) with 60 ng/ml (p = 0.0005), as well as after 24 h incubation (Figure 2) with 6000 ng/ml (p = 0.0009) β-END. When naltrexone and β-END (600 ng/ml) were combined in the 24 h group, ERKp44/p42 phosphorylation increased (p = 0.03). The levels of activated p38 were significantly lower during 15 min incubation with a combination of naltrexone and β-END (Figure 3), (p = 0.03), and significantly higher than control (p = 0.001) when stimulated with 6000 ng/ml β-END for 15 min. The phosphorylation of p38 was significantly lower than control when stimulated with 60 ng/ml (p = 0.03) and 600 ng/ml (p = 0.02) β-END during 8 h incubation (Figure 3). Finally, phosphorylated JNK was significantly higher in the group stimulated with 6000 ng/ml (p = 0.03) β-END when incubated for 15 min (Figure 4), and significantly lower in the 60 ng/ml (p = 0.03) and 600 ng/ml (p = 0.03) groups during 24 h stimulation (Figure 4). In all β-END untreated groups, there were no statistical differences between the non-activated and MAPK activated groups (Figure 2–4, 0(−) and 0(+), respectively). JNK stimulation under 8 h and p38 during 24 h showed no statistical differences among the groups and these results are shown as well (Figure 3 and Figure 4).

Figure 1 Determination of ERKp44/p42, p38, and JNK phosphorylation by Western blots. Lane 1: biotinylated protein ladder. Lane 2: MAPK non-activated control group. Lane 3: MAPK activated control group. Lane 4: chondrocyte cultures stimulated with 600 ng/ml β-END and blocked with 100 ng/ml naltrexone. Lanes 5–7: cell lysates from cultured articular chondrocytes stimulated with β-END in the following order: (5) 6000 ng/ml, (6) 600 ng/ml, and (7) 60 ng/ml and stressed with 50 ng/ml NGF (ERKp44/p42), or 0.5M sorbitol (p38 and JNK) for MAPK activation.

Figure 2-4 The effect of 15 min, 8 h, and 24 h β-END stimulation on MAPK phosphorylation in cultured articular chondrocytes. The group marked as 0(−) represents control, MAPK non-activated cell culture; 0(+) identifies control, MAPK activated control group. NAL (+) stands for 600 ng/ml β-END blocked with 100 ng/ml naltrexone. Different concentrations of β-END used in three separate culture-plate wells were designated in this order: 6000 ng/ml, 600 ng/ml, and 60 ng/ml; all of the β-END stimulated groups were also treated with 50 ng/ml NGF for ERKp44/p42 activation, i.e., 0.5 M sorbitol for p38 and JNK activation for 5 minutes, and therefore marked as well with (+). The results were expressed as duplicates of three independent experiments (mean ± SEM), or chondrocytes derived from three different patients (n = 6).* p < 0.05.

MAPK inhibitors prevented the effect of β-END on IL-1β protein levels in human articular chondrocytes (Figure 5). The group incubated with β-END for 1 h showed significantly higher levels of IL-1β than the untreated group marked as CONTROL (p = 0.02). None of the β-END treated groups pre-incubated for 1 h with a specific MAPK inhibitor, i.e., ERK1/2, p38, and JNK, showed a statistically significant difference when compared to the untreated control group (Figure 5). The group pre-incubated with specific ERK1/2 inhibitor and then treated with β-END did not differ significantly either when compared to β-END or untreated control group. Both p38- and JNK-specific inhibitor pre-incubated groups that were subsequently treated with β-END demonstrated significantly lower levels of IL-1β when compared to β-END stimulated group (p = 0.02, respectively, p = 0.03).

Figure 5 The influence of specific MAPK inhibitors on IL-1β protein levels during β-END incubation in cultured articular chondrocytes. All incubations with β-END were performed for 1 h and at a concentration of 6000 ng/ml. The pre-incubations with specific MAPK inhibitors were done for 1 h using 20 μ M PD98059 (ERK1/2 inhibitor), 10 μ M SB203580 (p38 inhibitor), and 100 μ M SP600125 (JNK inhibitor). The group marked as END stands for β-END incubated group, while CONTROL represents untreated group. ERK1/2, p38, and JNK identify the groups that were pre-incubated with respective specific inhibitor for 1 h, and then stimulated with β-END for 1 h. The results were expressed as triplicates of three independent experiments (mean ± SEM), i.e., chondrocytes derived from three different patients (n = 9). *p < 0.05 vs. CONTROL, †p < 0.05 vs. END.

DISCUSSION

In this study on chondrocytes in culture β-END was found to affect significantly ERKp44/p42, p38, and JNK phosphorylation.

Only stimulatory effects were seen in this study from β-END on ERKp44/p42. Morphine, a μ -opioid receptor agonist, induces ERK-p44/p42 and modulates MAPK in a rat and mouse animal model after 3 and 30 min, respectively (3, 10), while it inhibits ERK-p44/p42 in T-cells (28). Another study (21) reported activation of ERK-p44/42 in rat neurons after chronic administration of morphine (10^{− 5} M). An interesting observation in our study is that when the highest β-END concentration, 6000 ng/ml, was used in the 15 min group, there was no activation of ERKp44/p42, which was found with the lower concentrations 600 and 60 ng/ml. In the 8 h group the lowest and in the 24 h group the highest β-END concentration increased ERKp44/p42 phosphorylation.

Naltrexone did not block the effect on ERKp44/p42 in the 15 min group where β-END in the concentration 600 ng/ml significantly activated enzyme phosphorylation. Naltrexone in combination with 600 ng/ml β-END increased ERKp44/p42 after 15 min and 24 h, even though β-END 600 ng/ml alone did not change ERKp44/p42 significantly in the 24 h group. Naltrexone may have opioid agonist effects in some cases (19). Naloxone, whose effect is similar to naltrexone, may activate ERK-p44/p42. This observation was explained by cell-specific properties of receptor activation or participation of other receptor systems (19). Both p38 and JNK were activated after 15 min stimulation with 6000 ng/ml β-END when compared to control (Figure 3 and 4), but p38 and JNK activation decreased after 8 h and 24 h when the two lower concentrations 60 and 600 ng/ml β-END were used (Figure 3 and Figure 4).

The effect on JNK following 15 min incubation seen in our study is consistent with the findings following acute morphine administration in T cells in humans (28). Since IL-1β-induced catabolic effects in OA chondrocytes depend on JNK activity (27), it appears that μ -opioid receptor stimulation may influence these effects. It has also been reported that the induction of MMP-13 gene expression, which has a prominent role in arthritic joint inflammation and resorption of cartilage, is mediated by, among others, ERK-p44/p42, p38, and JNK (20).

The effects of different β-END concentrations could be explained by activation of the μ -opioid receptor, which is known to be coupled to G-proteins of the G_i, G_o, and also G_s (13) subtypes, and can therefore lead to both increase and decrease of cellular activity.

It has been stressed that ERKp44/p42 selective inhibition partially decreases some structural changes in experimental OA (25). The stimulatory effect on ERKp44/p42 phosphorylation by β-END in certain concentrations seen in our study could therefore speak for a modulation of inflammatory action in human OA chondrocytes. Similarly, the ERKp44/p42 mediated chondrocyte proliferation (33) might be regulated by the concentrations tested. In the same study, it has been stated that p38 mediates the induction of chondrocyte differentiation by connective tissue growth factor/hypertrophic chondrocyte specific gene product 24 (CTGF/Hcs24) that promotes proliferation and differentiation of chondrocytes in culture (19). Finally, the apoptotic signaling and differentiation in articular chondrocytes are mediated via ERKp44/p42 and p38 (17, 24). Another study has reported morphine-induced murine macrophage apoptosis through opiate receptors via p38 phosphorylation (29). The concentrations used were higher than those we worked with (10^{− 11} M–10^{− 9} M), i.e., 10^{− 8} M–10^{− 6} M morphine during 60 min.

A dose-dependent inhibition has been observed for p38 and JNK after incubation with β-END for a longer period of time (p38 for 8 h, JNK for 24 h). A similar inhibitory effect from an opiate has been seen in T cells stimulated with morphine (31). The amounts of β-END used in one of the groups (60 ng/ml ≈ 10^{− 11} M) are similar to those that could be found in normal physiological conditions. No more than 10^{− 11}M circulating β-END has been detected in patients with a radius fracture (15). In our opinion, the physiological concentration used in the study strengthens the relevance of the data obtained. When interpreting the results, it should be noted that different MAPKs have different activation timing (14). This is apparently cell- and species-specific (14, 24), and these data were missing for our specific cell of interest at the time the authors conducted the study.

It could be noticed that MAPKs were not significantly increased in any of the MAPK activated groups (Figure 2, Figure 3, Figure 4, 0(−) vs. 0(+)) when compared to the control. We have used the standard procedure for MAPK activation described in the Anti-ACTIVE® MAPK, p38, and JNK Polyclonal Antibodies and Anti-ACTIVE® Qualified Secondary Antibody Conjugates (Technical Bulletin No. 262, Promega) that means 5 min stimulation with 50 ng/ml NGF for ERKp44/p42 and 0.5 M sorbitol for p38 and JNK activation. Here we have to stress the fact that the source tissue (though structurally normal) was from osteoarthritic joints where MAPK are to be activated (32). MAPKs are also known to be activated by a temperature change (12). It has also been reported that MAPKs could be constitutively expressed in chondrocytes (7).

Cytokines, among others IL-1β, are known to be regulated via MAPKs in synoviocytes and chondrocytes (27). Once we had demonstrated the modulation of MAPKs via β-END, we wanted to examine the possibility of MAPKs being responsible for β-END regulation of IL-1β (2). The rationale for choosing 1 h of β-END incubation at 6000 ng/ml has been, in general, most significant effect at that concentration and 15 min incubation on MAPK phosphorylation (Figure 2, Figure 3, Figure 4). Since we have demonstrated that both ERK1/2, p38 and JNK specific inhibitors hindered β-END to increase the IL-1β protein levels significantly (Figure 5), it appears that this action is mediated via MAPK activation.

Overall, the results indicate that β-END significantly regulates MAPKs in cultured human articular chondrocytes, dependent on concentration and duration of incubation. Consequently, imbalance or destruction of homeostasis regulating MAP kinase activity is related to the pathogenesis of cartilage diseases such as OA (7). This observation could contribute to a better understanding of the β-END's possible role in cartilage damage in inflammatory diseases (8), and chondrocyte function modulated via MAPKs (7). In addition, β-END mediated increase of IL-1β seems to be dependent on MAPK activation, which could be of potential therapeutic interest if opiates were to play some role in modulation of articular cartilage

ACKNOWLEDGMENT

The authors thank Geir Tore Abrahamsen, MD (Dept. of Orthopaedics, UNN, Tromsoe, Norway), Kirsti Rønne, and Inigo Martinez, BSc, PhD (Dept. of Orthopaedics, University of Tromsoe, Norway) for technical assistance and useful comments.