Effects of various dietary supplements on inflammatory processes in primary canine chondrocytes as a model of osteoarthritis

Abstract

The use of dietary supplements as an alternative treatment for joint-related pathologies such as osteoarthritis (OA) is increasing. However, there is little scientific evidence to support the intended use. The aim of this study was to evaluate the anti-inflammatory effects of creatine- and amino acid-based supplements in primary cultured canine chondrocytes (CnCs) as an in-vitro model of OA and compare the effects to more commonly used agents, such as the non-steroidal anti-inflammatory drug (NSAID), carprofen, and the joint supplement, glucosamine (GS). CnCs were stimulated with interleukin-1β (IL-1β) and the subsequent release of prostaglandin E2 (PGE2) and tumor necrosis factor alpha (TNFα) was measured using an enzyme-linked immunosorbent assay (ELISA). Changes in oxylipins were also assessed using high-performance liquid chromatography/tandem mass spectrometry (HPLC/MS/MS). All compounds examined were able to significantly reduce the release of PGE2 and TNFα and were associated with reductions in cyclooxygenase-2 (COX-2) expression and nuclear factor-kappaB (NF-κB) phosphorylation. The creatine- and amino acids-based supplements also altered the profile of oxylipins produced. All compounds examined were less effective at reducing the release of PGE2 than carprofen. Carprofen significantly increased release of TNFα from CnCs, however, while the other agents reduced TNFα release. This study suggests that creatine- and amino acid-based supplements may have a beneficial role in preventing inflammation within the joint and that further studies are warranted.

Introduction

Osteoarthritis (OA), also known as degenerative joint disease, is a common clinical disease in humans, dogs, cats, and other companion animals and is one of the leading causes of disability and morbidity around the world (1,2). OA is characterized clinically by joint pain, stiffness, and functional disability and radiographically by narrowing of joint spaces and formation of osteophytes (bone spurs) (3). While OA can affect any joint in the body, it most commonly affects joints in knees, feet, hips, and spine (4).

OA results in the loss of articular cartilage within the joint of a bone. The articular cartilage is made up of a sparse distribution of chondrocytes within a dense extracellular matrix (ECM) (5). Under normal conditions, chondrocytes maintain a balance between the synthesis and the degradation of the ECM (6). In OA, however, altered chondrocyte activity leads to a gradual loss of the articular cartilage (6). Inflammatory mediators, such as interleukin-1β (IL-1β), tumor necrosis factor alpha (TNFα), IL-6, IL-8, and prostaglandins, which are also produced and released by the chondrocytes, activate the degradation of the joint, leading to collagen and proteoglycan breakdown (7).

There is no known cure for OA and medical therapy has focused on providing symptomatic relief and maintaining joint function. For symptomatic relief, non-steroidal anti-inflammatory drugs (NSAIDs) are a mainstay of OA therapy. Due to the adverse effects of long-term NSAIDs therapy, including gastrointestinal irritation and bleeding (3,8), alternative approaches to relieving joint pain have been explored. Nutraceuticals, such as glucosamine (GS) and chondroitin, represent the largest category of natural supplements for veterinary medicine (9,10). In humans, these nutraceuticals have been reported to reduce pain and stiffness of the joints in patients suffering from OA (3). Interest in creatine and amino acid supplements as potential adjuvants in the treatment of OA is also growing (11,12). Clinical trials examining the use of GS and chondroitin in OA have conflicting results, however, which highlights the need for more basic fundamental studies on the effects of nutraceuticals in OA (3,13).

The present study examined the anti-inflammatory effects of several different dietary supplements in cultured canine chondrocytes (CnCs) as an in-vitro model of OA. The dietary supplements examined included creatine monohydrate (CM), creatine hydrochloride (CHCl) (14,15), GS, and the conjugated amino acid supplement, Ethyl-α(guanido-methyl)ethanoate (AlphaGEE). Creatinine (CRN), a stable metabolite of creatine, was also examined. Effects on traditional inflammatory mediators, such as prostaglandin E2 (PGE2) and TNFα, were evaluated along with a more extended oxylipin analysis. The oxylipins are produced by oxygenation of polyunsaturated fatty acids and have both inflammatory and anti-inflammatory activity (16).

While all the dietary supplements examined were able to inhibit the release of PGE2, AlphaGEE was the most potent, with significant effects on PGE2 and TNFα release observed at 10 μM. The dietary supplements examined also altered the oxylipin profile of chondrocytes after inflammatory activation. Interestingly, CRN, which is generally considered to be a biologically inactive metabolite of creatine, also displayed anti-inflammatory activity in multiple assays. These results demonstrate that inflammatory processes within chondrocytes can be modified by treatment with selected amino acids and GS and suggest that further studies with these nutraceuticals as potential alternative or complementary treatments for OA are warranted.

Materials and methods

Primary CnCs derived from normal canine articular cartilage were purchased from Cell Applications (San Diego, California, USA) along with complete culture media and culturing reagents. Cell culture flasks, plates, and dishes were obtained from either Corning Incorporated (Corning, New York, USA) or Thermo Fisher Scientific (Waltham, Massachusetts, USA). Recombinant canine IL-1β was purchased from R&D Systems (Minneapolis, Minnesota, USA). CHCl and AlphaGEE were provided by Vireo Systems (Madison, Tennessee, USA). CM, CRN, and carprofen (Rimadyl) were obtained from Acros Organics (Geel, Belgium) and Fluka Analytical (St. Gallen, Switzerland), respectively. Canine TNFα ELISA kit and Prostaglandin E2 Parameter Assay kit were purchased from Raybiotech (Norcross, Georgia, USA) and R&D Systems, respectively. Pierce protease inhibitor mini-tablets and the Pierce bicinchoninic acid (BCA) protein assay kit were obtained from Pierce Biotechnology (Rockford, Illinois, USA). PageRuler Plus Prestained protein ladder, 10 to 250 kDa, was obtained from Thermo Scientific. Immuno-blot polyvinyldifluoride (PVDF) membrane was purchased from Bio-Rad Laboratories (Mississauga, Ontario). Primary rabbit polyclonal antibody for cyclooxygenase-2 (COX-2) was obtained from Abcam (Cambridge, UK). Mouse monoclonal antibody for B-actin was purchased from Sigma Aldrich (St. Louis, Missouri, USA). The secondary horseradish peroxidase (HRP)-linked antibodies were purchased from Novus Biologicals (Littleton, Colorado, USA). All other reagents and chemicals were purchased from Sigma Aldrich.

Canine chondrocyte culture

Primary CnCs were received as passage-1 cells and were plated in a T-75 flask and grown in CnC complete medium. When reaching approximately 80% confluency, CnCs were passaged at a seeding density of 8000 to 15 000 cells/cm². Cells were maintained in a humidified environment at 37°C with 5% carbon dioxide (CO₂), with media replacement every other day until the cells reached 60% to 70% confluence and every day upon reaching confluency. All experiments were conducted on confluent monolayers of CnC (passage number 4).

Cell viability studies

CnCs were seeded in 96-well plates with 5000 cells/well and incubated for 24 h in culture media. Cells were treated with culture media containing IL-1β (10 ng/mL) in the presence or absence of CM, AlphaGEE, CRN, CHCL, or GS (10 to 1000 μM), or with carprofen (1 to 100 μM). Cells were exposed to these treatments for 72 h. Additional cells treated with culture media alone were used as control. For every control and experimental treatment, 4 wells were used. Cell viability was determined using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. After 72 h, 25 μL of the MTT substrate [5 mg/mL in phosphate-buffered saline (PBS)] were added to each well and incubated for 3 h at 37°C. Viable cells with active metabolism convert MTT into a purple-colored formazan product with an absorbance maximum of approximately 570 nm. After the MTT substrate was added, dimethylsulfoxide (DMSO) was added to solubilize the cells and the resulting absorbance of the formazan product in the cell lysates was measured at a wavelength of 570 nm using a Synergy HT plate reader (BioTek, Winooski, Vermont, USA). Cell viability was expressed as a percentage of the absorbance levels observed in the control cells (media alone).

ELISA for PGE2 and TNFα analysis

CnCs were seeded in 96-well plates with 5000 cells/well and incubated for 24 h in culture media. Cells were treated with culture media containing IL-1β (10 ng/mL) alone or in combination with CM, CRN, CHCL, AlphaGEE, GS (10 to 1000 μM), or carprofen (1 to 100 μM). Additional cells were treated with culture media alone. For every control and experimental treatment, 4 wells were used. Culture media was collected at various time points (4 to 72 h) and stored at −80°C until further analysis. Quantitative analysis of PGE2 and TNFα in culture media samples was determined using enzyme-linked immunosorbent assay (ELISA). The ELISA assays were carried out as detailed in the manufacturer’s protocol and the optical density was determined using a Synergy HT plate reader (BioTek) set at 450 nm.

Western blot assays for COX-2 and NF-κB

Changes in COX-2 and phosphorylated nuclear factor-kappaB (NF-κB) protein expression in response to inflammatory stimuli were examined using Western blot analysis. Confluent monolayers of primary CnCs grown on 100-mm petri dishes were treated with culture media containing IL-1β (10 ng/mL), alone or with 1000 μM CM, CHCl, CRN, AlphaGEE, or GS, or with 100 μM carprofen. Additional cells were treated with culture media alone. For every control and experimental treatment, 3 petri dishes were used. At various treatment times, cell media was aspirated and the cells were washed twice with ice-cold PBS. Cells were lysed with 100 μL radioimmunoprecipitation assay (RIPA) buffer [150 mM sodium chloride (NaCl), 1.0% IGEPAL CA-630, 0.5% sodium deoxycholate, 0.1% sodium docecyl sulfate (SDS), and 50 mM Tris, pH 8.0], supplemented with a protease inhibitor cocktail. The cell lysate was sonicated and centrifuged at 15 000 rpm for 10 min at 4°C. Total cell protein supernatant was collected and prepared for electrophoresis by adding to 20 μg of total protein in a 1:1 ratio. Samples were then denatured by heating at 95°C for 5 min and loaded onto a 10% polyacrylamide gel for electrophoresis at 120 V at room temperature. The protein bands were transferred to an immunoblot PVDF membrane by applying a constant current of 200 mA for 2 h at 4°C.

Subsequently, membranes were washed with 5% (w/v) non-fat skim milk in tris-buffered saline with 0.1% tween-20 (TBS-T) to block non-specific binding sites under constant agitation for 1 h at room temperature. Membranes were then incubated with rabbit polyclonal primary antibody for COX-2 (1 mg/mL) or phospho-NF-kappaB p65 pSer536 antibody (1:1000 dilution) (Thermo Fisher Scientific) overnight at 4°C. Membranes were washed 3 times and incubated with mouse anti-rabbit immunoglobulin G-horseradish peroxidase (IgG-HRP) secondary antibody, diluted to 1:10 000 in blocking buffer for 1 h at room temperature. The specific bands were visualized using a chemiluminescence kit (Bio-Rad) and densitometry analysis was carried out using the Bio-Rad ChemiDoc MP imaging system and ImageLab 5.2.1 build 11 software. Protein expression was normalized to β-actin by stripping the membranes using a re-blot solution (Chemicon International, Billerica, Massachusetts, USA). The stripped membranes were then re-blotted with mouse monoclonal primary antibody for β-actin (0.5 mg/mL) and rabbit anti-mouse IgG-HRP.

Oxylipin analysis

The production of oxylipins in response to inflammatory mediators and the various treatments was examined using high-performance liquid chromatography/tandem mass spectrometry (HPLC/MS/MS). Confluent monolayers of primary CnCs were grown on 6-well plates and treated with culture media containing IL-1β (10 ng/mL) in the presence and absence of CM, CHCl, CRN, or AlphaGEE or GS (1000 μM), or with 100 μM of carprofen. Additional cells were treated with culture media alone. For every control and experimental treatment, 4 wells were used. After treatment, cells were placed in a humidified CO₂ incubator maintained at 37°C. After 24 h, 3-mL samples were collected from culture media and antioxidant solution [0.2 mg/mL butylated hydroxytoluene (BHT), 0.2 mg/mL ethylenediamine tetra-acetic acid (EDTA), 2 mg/mL triphenylphosphine, and 2 mg/mL indomethacin in methanol/ethanol/water (2:1:1, v/v/v)] was added to each sample to protect it from degradation and the samples were stored at −80°C until further analysis.

Oxylipin concentrations in media samples were determined as described by Aukema et al (17). Briefly, after adding 10 ng of deuterated internal standards (Cayman Chemical, Michigan, USA), samples were adjusted to pH < 3. Solid-phase extraction (SPE) was carried out with Strata-X SPE columns (Phenomenex, California, USA) that were preconditioned with methanol and pH 3 water, loaded with sample, rinsed with 10% methanol, and eluted with methanol. Samples were dried and resuspended in solvent for analysis by HPLC/MS/MS (API 6500; Sciex, Concord, Ontario). Quantification of oxylipins was determined using the stable isotope dilution method (17).

Statistical analysis

All data were expressed as mean ± standard error of the mean (SEM). Comparison among the groups of data was evaluated by either 1-way analysis of variance (ANOVA), if only 1 variable was present, or 2-way ANOVA, if more than 1 variable was present, using Fisher’s least significant difference test for multiple comparisons of the means. The tests were done using GraphPad Prism, Version 6 software (San Diego, California, USA), with P-values < 0.05 considered significant.

Results

Cell viability of primary cultured CnCs was evaluated after exposure to both the inflammatory stimulant (IL-1β) and the various treatment agents (Figure 1). No significant effects on cell viability were observed after CNCs were exposed to IL-1β for 72 h. Furthermore, none of the treatments examined changed cell viability within a 72-h exposure period.

Figure 1.

Effects of various treatments on cell viability. MTT assay to assess cell viability of primary cultured canine chondrocytes (CnCs) after 72-hour exposure to the inflammatory stimulant (IL-1β) and carprofen (A), CM (B), AlphaGEE (C), CRN (D), CHCL (E), or glucosamine (GS) (F). Values represent the mean ± SEM of 8 CnC monolayers.

Release of PGE2 from primary CnC monolayers under control and stimulated conditions

Treatment of cultured CnCs with IL-1β produced a time-dependent increase in PGE2 release, with significant increases observed from 8 to 48 h compared to unstimulated cells (Figures 2 and 3). Treatment of CnCs with the COX inhibitor, carprofen (1 to 100 μM), substantially attenuated the release of PGE2 induced by IL-1β. The reductions in PGE2 release were observed as early as 8 h after carprofen and at all concentrations examined (Figure 2A). Exposure of cells to GS produced significant decreases in PGE2 release from CnC monolayers. The effects were both time- and concentration-dependent, with 58% reduction at 8 h after exposure at the highest concentration examined and more modest reductions at 48 h (Figure 2B). Treatment of CnCs with CM resulted in a significant decrease in PGE2 concentrations compared to IL-1β alone at both 8 and 48 h, up to 55% and 38%, respectively (Figure 3A). Similar reductions in PGE2 release were obtained with CHCl (Figure 3B). The creatine metabolite, CRN, was also able to reduce PGE2 release, with greatest effects observed at 8 h after exposure (Figure 3C). Interestingly, of the dietary supplements examined, AlphaGEE appeared to be the most efficacious and potent at reducing PGE2 release from CnCs, with responses observed at concentrations as low as 10 μM and maximal inhibition of 76% at the 8-hour time point (Figure 3D). It should be noted that, while all the dietary supplements examined were able to reduce PGE2 release in response to the inflammatory stimulus, the level of reduction was not as great as with carprofen.

Figure 2.

Effects of carprofen and glucosamine (GS) on the release of inflammatory mediators from primary cultured canine chondrocytes (CnCs). Release of PGE2 (A and B) or TNFα (C and D) under control conditions and after stimulation with IL-1β with and without carprofen (A and C) or GS (B and D). Values represent the mean ± SEM of 4 monolayers per treatment group.

*P < 0.05, **P < 0.01, ***P < 0.001, compared to IL-1β treatment group (2-way ANOVA), ****P < 0.0001.

Figure 3.

Release of PGE2 from primary cultured canine chondrocytes (CnCs) under control conditions and after stimulation with IL-1β with and without CM (A), CHCL (B), CRN (C), or AlphaGEE (D). Values represent the mean ± SEM of 4 monolayers per treatment group.

*P < 0.05, **P < 0.01, ***P < 0.001, compared to IL-1β treatment group (2-way ANOVA), ****P < 0.0001.

COX-2 expression in primary CnC monolayers

The expression of COX-2 was also examined in primary CnC monolayers under control conditions and after stimulation with IL-1β alone or in combination with various potential anti-inflammatory agents (Figure 4A). There was no detectable COX-2 in CnC lysates under control conditions. In contrast, COX-2 expression increased significantly when the monolayers were exposed to the pro-inflammatory cytokine, IL-1β (Figure 4A). All the dietary supplements examined were able to significantly attenuate COX-2 expression in response to IL-1β (Figure 4A). This was in contrast to the effects of carprofen exposure in CnCs, which significantly enhanced COX-2 expression in response to IL-1β stimulation (Figure 4A).

Figure 4.

Western blot analysis of COX-2 (A) and NF-κB (B) in stimulated cells after 48 h for COX-2 and 1 h for NF-κB under the various treatment conditions. Values were expressed as mean ± SEM of 3 samples per treatment group.

*P < 0.05, **P < 0.01, ***P < 0.001, compared to IL-1β treatment group (1-way ANOVA).

Release of TNFα and activation of NF-κB in primary CnC monolayers under control and stimulated conditions

Under control conditions, little TNFα was detected in the culture media (Figure 2C and 2D). After stimulation with IL-1β, however, significant increases in TNFα release (2-fold) were observed at the various time points examined (Figure 2C and 2D). Those CnC monolayers stimulated with IL-1β and receiving carprofen had even greater TNFα release than IL-1β alone (Figure 2C). In contrast, treatment of primary cultured CnCs with GS produced slight reductions in TNFα release in response to IL-1β stimulation, which was most apparent after 48 h (Figure 2D). For the creatine compounds and AlphaGEE, there was a reduction in TNFα release at 24 h at the highest concentration of CM examined and an even more pronounced reduction at all concentrations at 48 h (Figure 5A). Similar effects were observed with CHCL, although maximal effects occurred at 24 h (Figure 5B). Significant reductions in TNFα release after IL-1β stimulation were also observed at 24 and 48 h in the CRN treatment group, but only at the highest concentration of CRN examined (Figure 5C). Treatment of cells with AlphaGEE led to significant reductions in TNFα release in response to IL-1β stimulation, with maximal responses observed at 24 h (Figure 5D).

Figure 5.

Release of TNFα from primary cultured canine chondrocytes (CnCs) under control conditions and after stimulation with IL-1β with and without CM (A), CHCL (B), CRN (C), or AlphaGEE (D). Values represent the mean ± SEM of 4 monolayers per treatment group.

*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, compared to IL-1β treatment group (2-way ANOVA).

As TNFα produces its effects by activating NF-κB transcription factor, the expression of the phosphorylated form of NF-κB was examined under normal conditions and after IL-1β stimulation. As shown in Figure 4B, treating the cells with IL-1β alone (10 ng/mL) for 1 h induced significant expression of phosphorylated-NF-κB compared to untreated control cells. All the creatine compounds, as well as AlphaGEE, GS, and carprofen, significantly reduced the expression of phosphorylated-NF-κB compared to the positive control, with CRN causing the most reduction (85%) (Figure 4B).

Oxylipin production in CnCs under control and stimulated conditions

Of the 135 oxylipins analyzed, 90 were present at or above the quantitation level of the assay and over 50 were present at a level of ≥ 0.1 ng/mL in media taken from control cells and those stimulated with IL-1β. Of those oxylipins that were present in measurable amounts in the culture media, 14 showed at least a 2-fold change after a 24-h IL-1β challenge, including PGE1, d17-6k PGF1α, and dihomo PGF2α derived from dihomo gamma linolenic acid, eicosapentaenoic acid, and adrenic acid precursors, respectively, in addition to PGE2, 11 PGE2, and 6-keto Prostaglandin F1α, 6,15-diketo-,dihydro-PGF1α (6,15-dk-dh-PGF1α) originating from arachidonic acid metabolism.

The remaining oxylipins with at least a 2-fold elevation after IL-1β exposure were from the lipooxygenase (LOX) metabolic pathways and included 11-hydroxy-eicosatetraenoic acid (11-HETE), 12-HETE, and 13-hydroxy-docosahexaenoic acid (13-HDoHE). In addition, 4 oxylipins showed at least a 2-fold decrease in release after IL-1β stimulation, including 14,15 epoxy-eicosatrienoic acid (14,15 EpETrE), 20-hydroxy-docosahexaenoic acid (20-HDoHE), 16,17 epoxy-docosapentaenoicacid (16,17 EpDPE), and 18-hydroxy-eicosapentaenoic acid (18-HEPE). All the oxylipins showing a reduction after IL-1β were derived from the cytochrome P450 (CYP) metabolic pathway. Those oxylipins that displayed significant changes in response to the various treatment groups compared to IL-1β stimulation alone were 12-HETE, 18-HEPE, 20-HDoHE, and dihomo PGF2α (Figure 6). In the case of 12-HETE and dihomo PGF2α, the increases observed in response to IL-1β stimulation were significantly reduced with all treatments examined (Figure 6). The various treatments were able to restore 18-HEPE and 20-HDoHE, which were significantly reduced after IL-1β stimulation alone, to levels similar to unstimulated control groups (Figure 6).

Figure 6.

Levels of 12-HETE (A), 18-HEPE (B), dihomo PGF2α (C), and 20-HDoHE (D) in stimulated cells after 24 h under the various treatment conditions. Values were expressed as mean ± SEM of 4 samples per treatment group. Asterisks above the bars indicate significance compared to IL-1β treatment group; square brackets indicate significance compared to control (untreated group).

*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

ns — nonsignificant (1-way ANOVA).

Discussion

While OA has long been considered a non-inflammatory joint disease, increasing evidence supports the role of inflammatory mediators produced by the articular tissue in the early initiation and progression of the disease (7,18,19). OA is therefore best characterized as a condition involving chronic, low-grade inflammation within the afflicted joint. The inflammatory mediators associated with both pain and cartilage degeneration in OA include the inflammatory cytokines, IL-1β, and TNFα, as well as prostaglandins and other oxylipins produced from the metabolism of arachidonic acid (20–22). Agents capable of modulating the multiple cytokine and oxylipin pathways activated within the cells of the affected joint are of interest and may potentially influence the progression of OA.

In the current study, primary CnCs derived from normal canine articular cartilage were used as an in-vitro model of OA for examining the anti-inflammatory effects of various nutraceutical products. Primary CnCs are considered a well established in-vitro model for studying the inflammatory mechanisms of OA and have been used to identify the key inflammatory targets and to evaluate the potential effectiveness of therapeutic agents (23–25). Chondrocytes from patients with OA show elevated levels of IL-1β and TNFα release compared to chondrocytes isolated from non-OA patients, which helps to establish the role of these cytokines in disease (26). Primary chondrocyte cultures have also been used to screen for natural products that could be important in reducing inflammation in OA (24,25). The ability of cultured chondrocytes to activate inflammatory pathways after exposure to IL-1β, which is a key inflammatory mediator in the pathogenesis of OA, is important for the present study (24,25).

Complementary approaches for treating OA are of interest due in part to the absence of disease-modifying therapies for the disease. In the present study, we examined several dietary supplements for their potential anti-inflammatory and disease-modifying activity. The supplements selected were based on both manufacturer’s claims and published studies supporting their potential anti-inflammatory effects. In the case of creatine compounds, previous studies have reported an anti-inflammatory effect in rodent models of inflammation (27). More recent studies have shown that high-dose CM supplementation reduced both plasma TNFα and PGE2 in athletes (28,29). As both the rodent and human studies involved a TNFα component, we hypothesized that creatine compounds would also be effective anti-inflammatory agents in our in-vitro OA model. Previous studies have also reported anti-inflammatory effects of several amino acid and modified, i.e., branched chain, amino acid compositions (30–33). While the mechanisms proposed for the anti-inflammatory effects of various amino acids differ based on the inflammatory model examined, there is evidence to suggest that the cellular pathways involved affect those examined in the present study, i.e., oxylipins and TNFα (30–32).

All the nutritional supplements examined in this study displayed some ability to reduce the release of PGE2 and TNFα into the culture media after IL-1β stimulation. The effects were both time- and concentration-dependent, with significant decreases in the concentrations of PGE2 observed at 8 and 48 h after stimulation and significant reductions in TNFα release occurring at 24 and 48 h. The reason(s) for the biphasic effects observed with the creatine compounds is(are) not known. These compounds could affect PGE production through multiple pathways and thus different exposure times could result in different effects on PGE release. The studies examining the pharmacological effects of creatine in a rodent carrageenan inflammation model suggested a lysosomal membrane stabilization effect and potential downstream effects on formation of prostaglandin (27). Such effects may be responsible for the early reductions in PGE2 release, i.e., at 8 h, observed in the present study, while the later effects may be driven by decreases in TNFα release.

The increased PGE2 found in chondrocytes from cartilage of human patients suffering from OA appears to be associated with increased expression of COX-2 (26). In addition, COX-2 expression is influenced by transcription factors activated through IL-1β and TNFα (25,27). In the present study, the expression of COX-2 was significantly increased after the cells were stimulated with IL-1β compared to control cells. Furthermore, with the exception of carprofen, all the tested compounds were able to decrease COX-2 expression at 48 h. A similar effect was observed when examining TNFα release in the media, with all the tested compounds, except carprofen, significantly reducing the release of TNFα.

The fact that carprofen increased the expression of COX-2 and the release of TNFα is consistent with the findings of other studies on the effects of COX-2 inhibitors. Page et al (34) showed that NSAIDs, specifically celecoxib and rofecoxib, which are COX-2-specific NSAIDs, and diclofenac, which is a COX-1-and COX-2-bispecific NSAID, significantly increased the spontaneous release of TNFα in arthritic human synovial membrane cultures. The same increase of TNFα with NSAIDs was observed in primary human monocytes after stimulating the cells with lipopolysaccharide (LPS) (34). These findings might explain the ability of NSAIDs to relieve joint pain and tenderness and their lack of disease-modifying anti-rheumatic activity, as increased TNFα release would be expected to exacerbate the inflammation and the pathologic processes in arthritic patients.

As the inflammatory effects of TNFα are associated with activation of NF-κB, the effects of the compounds on NF-κB activation were also examined. As expected, given the effects of the various nutraceutical compounds on TNFα release, there was also a reduction in NF-κB activation observed compared to IL-1β stimulation alone. The ability of the nutraceutical compounds to influence multiple pathways in the inflammation process suggests that the anti-inflammatory properties may be a result of the beneficial effects of creatine on the mitochondria and its antioxidant properties. A growing body of evidence suggests an interplay between inflammation and mitochondrial dysfunction in OA. For example, Kim et al (35) have shown that pro-inflammatory cytokines such as IL-1β and TNFα disturb normal mitochondrial function in primary human osteoarthritic chondrocyte cultures by damaging mitochondrial DNA (mtDNA), decreasing energy production, and increasing production of reactive oxygen species (ROS). It has also been shown that mitochondrial dysfunction induced by commonly used inhibitors, such as oligomycin and antimycin A, which inhibit mitochondrial respiratory chain complexes V and III, respectively, increases inflammatory responsiveness to IL-1β and TNFα, i.e., increases IL-8, COX-2, and PGE2 production, in normal human chondrocytes (36). This suggests that any therapeutic intervention that has the ability to stabilize and improve mitochondrial function could also decrease the inflammation observed in OA.

Although the role of prostaglandins and specifically PGE2 in the pathophysiology of OA has been the subject of extensive research, the role of other oxylipins is not well-understood. Since inflammation is an important component in the pathophysiology of OA, changes in oxylipin release in response to inflammation in the CnCs might provide important insight into the development of OA, as well as provide potential biomarkers for treatment response. In the present study, the production of oxylipins in CnCs was examined in response to exposure to IL-1β. It was demonstrated that exposure of the cells to IL-1β, which is an early endogenous pro-inflammatory mediator, increased the levels of 10 oxylipins formed via COX and LOX pathways and decreased the levels of 4 oxylipins derived from the CYP pathway. The pro-inflammatory PGE2 was the COX-derived oxylipin present at the highest level upon incubation with IL-1β. Several other COX-derived oxylipins showed substantial elevations (up to 32-fold) after IL-1β exposure, although in smaller concentrations. While the role of these oxylipins in OA and inflammation remains to be determined, recent studies point to their potential value as biomarkers for symptomatic OA (37).

Two oxylipins that were increased after IL-1β stimulation were 11-HETE and 12-HETE. Both of these oxylipins are pro-inflammatory, derived via the LOX pathway from AA (16), and are strongly associated with inflammation (16). A role for these oxylipins in OA is suggested by Wong et al (37) who reported increased levels of 12-HETE in knee joints of a rat model of OA compared to healthy controls. It should be noted that all the creatine- and amino-acid based supplements examined in this study were able to reduce the release of 12-HETE in CnCs stimulated with IL-1β.

In addition to increasing the release of pro-inflammatory oxylipin, stimulating the cells with IL-1β decreased the levels of some anti-inflammatory oxylipins derived via the CYP pathway, including 18-HEPE (16). Although the importance of this particular oxylipin in OA is unknown, 18-HEPE has the ability to decrease LPS-induced TNFα secretion in the murine macrophage cell line (38). The ability of CM, CHCl, CRN, and AlphaGEE to restore 18-HEPE after IL-1β stimulation in order to control levels may therefore be involved in the effects of these agents on TNFα release reported here.

The finding that CRN also has anti-inflammatory properties was unexpected since this compound has been considered to be the inactive metabolite of creatine, which is excreted in the urine (39). Previous studies using the rodent carrageen inflammation model have reported anti-inflammatory actions for both creatine and CRN (27,30). The finding of anti-inflammatory effects of CRN in the CnC model in the present study is also consistent with the more recent findings of Leland et al (40). This study reported that short-term exposure (10 to 60 min) of a mouse macrophage cell line (RAW 264.7) to 100 μM of CRN significantly down-regulated levels of toll-like receptor 2 (TLR2), TLR3, TLR4, and TLR7 messenger ribonucleic acid (mRNA) in the cells, which suggests that CRN may have immunosuppressive properties (40).

In summary, the present study demonstrates that various creatineand amino acid-based compounds can produce a significant anti-inflammatory effect. The exact mechanism(s) by which the creatine and amino acid compounds produce their anti-inflammatory effects are not known. Unlike the COX-2 inhibitor, the dietary supplements examined here appeared to affect multiple pathways in the process of inflammation in the CnC model. The creatine- and amino acid-based compounds examined interfered with NF-κB and TNFα release, leading to reductions in COX-2 expression and the release of PGE2, as well as alterations in both pro-inflammatory and anti-inflammatory oxylipins in response to IL-1β activation.

An important consideration for translating these initial findings in a cell culture model of OA to in-vivo animal models and human trials is the concentrations required to obtain the effects observed. While the concentrations required for anti-inflammatory response to GS in the present study were much higher than the peak plasma levels reported with standard dietary supplementation with GS (41), the concentrations required for creatine- and amino acid-based compounds were certainly within the range of blood levels observed after dietary supplementation (42). This study suggests that supplementation with creatine- and amino acid-based supplements may provide beneficial effects in OA and warrants further research into their clinical effectiveness and mechanism of response.