Abstract
Rheumatoid arthritis (RA) and osteoarthritis (OA) are inflammatory joint diseases, characterized by pain and structural damage. Besides prostaglandins, usually targeted by non-steroidal anti-inflammatory drugs, other lipids, including fatty acids, phospholipids and other bioactive lipid mediators derived from fatty acids could also contribute to RA and OA.
In this review, we present evidence for the role of fatty acids and derivatives in RA and OA by summarizing findings related to their presence in serum and synovial fluid, as well as their association with clinical characteristics and effects on RA and OA tissues in vitro. Finally, a more direct evidence for their role in RA and OA derived from intervention studies in humans or mouse models of disease is summarized. Based on the presented data, we present a research agenda, in which some key unresolved questions regarding the role of lipids in RA and OA are formulated.
Introduction
Rheumatoid arthritis (RA) and osteoarthritis (OA) are joint diseases characterized by different pathophysiological mechanisms, but displaying common clinical characteristics, such as joint pain, functional impairment and structural damage which is hallmarked by bone erosions in RA and osteophytes in OA. Moreover, both diseases display joint space narrowing, reflecting cartilage loss. Another common feature of these diseases is the presence of inflammation in the majority of the patients. While the role of inflammation in the pathogenesis of RA has been established previously, its presence and possible role in OA has been only recently revealed. Several studies during the past 10 years have shown an association between synovial inflammation and pain on one hand and radiographic progression on the other hand, thus establishing the pivotal role of inflammation in OA (reviewed in Refs. ).
Fatty acids acquired through diet are usually transported through the body in triglycerides or phospholipids incorporated in lipoproteins, but can also be found in free form in blood. Moreover, they are present both in bound and free form in cells, where they perform various functions as energy source, membrane constituents or signalling molecules. They are the essential building blocks for higher-order lipids such as phospholipids, sphingolipids, glycerolipids and glycolipids. Moreover, they could be metabolized into bioactive lipid mediators such as oxylipins, including eicosanoids (prostaglandins, thromboxanes and leukotrienes) and other lipids with more anti-inflammatory and pro-resolving activity such as lipoxins, resolvins, maresins and protectins. Enzymes such as phospholipases (PLAs) which release fatty acids from phospholipids, cyclooxygenases (COXs) and lipoxygenases (LOXs) that oxidize fatty acids are involved in the generation of oxylipins. Fatty acids, higher-order lipids and oxylipins can interact with inflammatory as well as tissue-resident cells, thereby contributing to various processes in the body, including inflammation, wound healing, pain, etc, and potentially playing a role in RA and OA. In general, it is believed that saturated fatty acids, n-6 polyunsaturated fatty acids (PUFAs), such as arachidonic acid (AA), and AA derivatives (prostaglandins and leukotrienes) have a pro-inflammatory effect. By contrast, unsaturated fatty acids, n-3 PUFA, such as docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), and the oxylipins derived from them (resolvins, maresins and protectins) are believed to have an anti-inflammatory function. The latter are also believed to be pro-resolving, thereby actively helping wound healing and return to tissue homeostasis after an inflammatory response.
In this review, we will present data supporting the role of lipids in RA and OA. To this end, we will summarize findings from three lines of evidence. First, data related to the presence of various fatty acids and their derivatives (lipid mediators) in blood or synovial fluid (SF) of RA and OA patients will be summarized. Next, the association of these lipids with clinical disease parameters will be presented, as well as in vivo intervention studies both in humans and mouse models of disease. Finally, in vitro studies indicating the effects of these lipids on human joint tissues will be presented (summarized in Fig. 1 ). The role of cholesterol and lipoprotein metabolism in RA and OA has been reviewed elsewhere and will not be discussed in this review.
Lipids and lipid classes described in serum of healthy individuals
In an extensive study in which the major six lipid categories as defined by the Lipid Maps were measured in plasma of healthy individuals, more than 500 lipid species were identified . The measured sample was obtained from the National Institute of Standards and Technology (USA); the pooled plasma sample was obtained and stored after overnight fasting in a standardized fashion, from 100 healthy individuals (age: between 40 and 50 years) including an equal number of men and women whose ethnicity was representative of the US population. The most abundant (on a molar basis) was sterols (including cholesterol), followed by triglycerides (part of lipoproteins), glycerophospholipids, free fatty acyls and sphingolipids, and the least abundant was diacylglycerols and prenols. In terms of free fatty acids, oleic acid, followed by palmitic acid and stearic acid were the most abundant and comprised approximately 78% of all free fatty acids after overnight fasting. The most abundant PUFA were linoleic acid (LA) and AA, but EPA and DHA, which are derived from fish oil and are known for their anti-inflammatory effects, were also detectable. Lipid mediators such as oxylipins were also detected in plasma, with 15-deoxy-PGD 2 (prostaglandin D2) being the major metabolite generated by COX, while 5-HETE was the most prominent eicosanoid of the LOX pathway found in plasma .
This review focuses on selected lipids in the six lipid classes: fatty acids either in free form or incorporated in higher-order lipids (especially phospholipids), as well as their bioactive lipid mediators in RA and OA ( Table 1 ).
| Lipid class (according to Lipid Maps) | Lipids included in this review |
|---|---|
| Fatty acyls | Fatty acids, oxylipins: eicosanoids (prostaglandins, leukotrienes, lipoxins), derivatives of DHA, EPA (resolvins, maresins, protectins) |
| Glycerolipids | – |
| Glycerophospholipids | Phospholipids (e.g. PC) and lysophospholipids (e.g. LPC, LPA) |
| Sphingolipids | sphingomyelin |
| Sterol lipids | – |
| Prenol lipids | – |
Presence of fatty acids and derivatives in RA and OA patients
Lipids in serum of RA patients
Direct comparisons between different lipids in RA serum/plasma and healthy controls were only made in few studies and generally included relatively low numbers of patients (between 10 and 16). In a metabolomics study, it was found that the levels of lipids in general were lower in serum of newly presenting RA patients compared to healthy controls although a systematic investigation of different lipid classes was not performed. To exclude the effect of disease-modifying anti-rheumatic drugs (DMARDs) on metabolic status, patients who were administered DMARDs were excluded from this analysis . In another study in 166 RA patients, the levels of free fatty acids were found to be similar in RA patients and healthy controls , indicating that free fatty acids are not quantitatively different in the diseased state. Fatty acids in phospholipids and sphingomyelins were also described in serum of healthy individuals and RA patients . Of the different phospholipids studied, only the ratio phosphatidylcholine (PC)/lysophosphatidylcholine (LPC, generated from PC by the PLA enzyme) was found to be lower in serum of RA patients compared to healthy individuals , indicating a higher activation of PLA in RA patients. This could result in higher levels of free fatty acids that can further be metabolized into bioactive lipids. Indeed, lipid mediators such as prostaglandins, PGD 2 and PGE 2 , generated by COX enzymes, were shown to be present in serum of RA patients, indicating that not only precursor fatty acids might be higher in RA compared to healthy individuals, but also certain enzymes involved in generation of bioactive lipids. However, bioactive lipids were mainly studied in SF.
Lipids in SF of RA patients
Broad lipidomic studies in SF of RA patients were performed in which most of the lipid species described in serum/plasma could be detected ; several phospholipid and fatty acid species were found. Similarly to plasma, PC were the most abundant phospholipids in SF, followed by LPC and sphingomyelins and their levels were higher in RA SF than controls . The ratio of PC/LPC was higher in RA SF than in controls, in contrast to what was found in serum . A detailed analyses of the species of lipids revealed that RA SF was relatively enriched in LPC-containing saturated fatty acids, while saturated PC was lower than in controls, indicating that saturated fatty acids were possibly more efficiently released from PC in RA. The length of fatty acids contained in PC species was not different between RA SF and controls, but RA SF contained relatively more short-chain fatty acids in LPC compared to controls . As a longer chain length and higher saturation grade are believed to be beneficial for the lubricating properties of phospholipids, these data could indicate that the phospholipids present in RA SF are less potent lubricators and provide less protection against mechanical damage than in healthy controls. All studies were small and included approximately 20 patients.
Eicosanoids such as prostaglandins and leukotrienes were intensively investigated in SF of RA patients and were recently reviewed . Prostaglandin E2 (PGE 2 ) as well as PLA 2 and COX, enzymes involved in its generation, were found to be higher in SF or were higher expressed in synovium of RA patients (mPGES) . Likewise, the enzyme converting PGE 2 into the inactive 15-PGDH was higher in RA synovium, indicating that regulatory mechanisms targeting PGE 2 are also activated in this disease. Interestingly and in line with this observation, the more anti-inflammatory prostaglandins, PGD 2 and its metabolite 15-deoxy-PGJ 2 were also found in the RA joint. Expression of PGD 2 was detected in cells in RA joints; however, no evidence was found for enhanced biosynthesis of 15-deoxy-PGJ 2 . Next to prostaglandins, leukotrienes such as LTB 4 , LTD 4 and LTE 4 were found in SF of RA patients .
Interestingly, anti-inflammatory products of LOXs were also described in SF of RA patients. These included the anti-inflammatory and pro-resolving mediators lipoxin A4 (LXA 4 ) and the mediators derived from the n-3 PUFA DHA: maresin 1 (Mar1) and resolving D5 (RVD5) . Although this study was performed in only five patients and the effects of pro-resolving lipids on RA development or progression were not yet investigated, their potent immune modulatory functions identify them as promising therapeutic agents for chronic inflammatory diseases.
Besides the PLA, COX and LOX enzymes, autotaxin makes an important contribution to bioactive lipids in RA. It can convert membrane phospholipids such as LPC to lysophospatidic acid (LPA), as well as sphingosylphosphorylcholine to yield sphingosine-1-phosphate (S1P). S1P can act as an intracellular second messenger or extracellular lipid mediator via G-coupled receptors and is involved in regulation of pro-inflammatory pathways and cell migration. Its possible role in RA is discussed later. Both LPA and S1P, as well as autotaxin were shown to be present in SF of RA patients , while higher autotaxin levels were shown in RA SF ( n = 16) and serum ( n = 26) compared to controls .
Lipids in serum/SF of OA
Several lipidomic studies in OA patients have been performed over the recent years. Although most of them study SF, one study investigated the phospholipid profile of hip and knee OA patients ( n = 59 women) in plasma . Similarly to RA, the LPC/PC ratio was higher in both mild and moderate OA (based on radiographic damage) compared to controls, contradictory to that found in SF. In two studies with 30 and 48 OA patients, the levels of phospholipids were found to be elevated in SF of OA patients compared to post-mortem controls , similar to what was found for RA.
Few studies have investigated oxylipin levels in OA. In one study with 10 OA patients, 8-iso-PGF 2α and 15-keto-dihydroPGF 2α were detected in both serum and SF, but the levels did not correlate in these two fluids . This indicates that there might be local production of these lipids, rather than systemic diffusion. Noteworthy is that levels of both metabolites were higher in serum of OA patients compared to controls. Higher levels of these metabolites in serum of OA patients could indicate higher systemic oxidative damage and inflammatory responses. For some patients, the levels of both compounds, but especially 15-keto-dihydroPGF 2α tended to be somewhat higher in SF compared to serum. It is difficult to interpret these data due to the lack of healthy SF and low numbers of patients. More recently, a study investigated the 15-LOX product 15-HETE and the COX product PGE 2 in plasma of symptomatic OA patients and controls . Both metabolites were elevated in patients (three different cohorts with 291 patients in total) compared to controls, which could be a systemic reflection of local inflammation in the knee. Similarly, leukotriene B 4 (LTB 4 ) has also been described in SF of some OA patients, although no comparison with healthy controls was performed . Together, these data indicate that pro-inflammatory lipids are present in OA patients, although a more extensive investigation of the presence of other lipid classes, such as prostaglandins and anti-inflammatory oxylipins, in OA patients is warranted.
Presence of fatty acids and derivatives in RA and OA patients
Lipids in serum of RA patients
Direct comparisons between different lipids in RA serum/plasma and healthy controls were only made in few studies and generally included relatively low numbers of patients (between 10 and 16). In a metabolomics study, it was found that the levels of lipids in general were lower in serum of newly presenting RA patients compared to healthy controls although a systematic investigation of different lipid classes was not performed. To exclude the effect of disease-modifying anti-rheumatic drugs (DMARDs) on metabolic status, patients who were administered DMARDs were excluded from this analysis . In another study in 166 RA patients, the levels of free fatty acids were found to be similar in RA patients and healthy controls , indicating that free fatty acids are not quantitatively different in the diseased state. Fatty acids in phospholipids and sphingomyelins were also described in serum of healthy individuals and RA patients . Of the different phospholipids studied, only the ratio phosphatidylcholine (PC)/lysophosphatidylcholine (LPC, generated from PC by the PLA enzyme) was found to be lower in serum of RA patients compared to healthy individuals , indicating a higher activation of PLA in RA patients. This could result in higher levels of free fatty acids that can further be metabolized into bioactive lipids. Indeed, lipid mediators such as prostaglandins, PGD 2 and PGE 2 , generated by COX enzymes, were shown to be present in serum of RA patients, indicating that not only precursor fatty acids might be higher in RA compared to healthy individuals, but also certain enzymes involved in generation of bioactive lipids. However, bioactive lipids were mainly studied in SF.
Lipids in SF of RA patients
Broad lipidomic studies in SF of RA patients were performed in which most of the lipid species described in serum/plasma could be detected ; several phospholipid and fatty acid species were found. Similarly to plasma, PC were the most abundant phospholipids in SF, followed by LPC and sphingomyelins and their levels were higher in RA SF than controls . The ratio of PC/LPC was higher in RA SF than in controls, in contrast to what was found in serum . A detailed analyses of the species of lipids revealed that RA SF was relatively enriched in LPC-containing saturated fatty acids, while saturated PC was lower than in controls, indicating that saturated fatty acids were possibly more efficiently released from PC in RA. The length of fatty acids contained in PC species was not different between RA SF and controls, but RA SF contained relatively more short-chain fatty acids in LPC compared to controls . As a longer chain length and higher saturation grade are believed to be beneficial for the lubricating properties of phospholipids, these data could indicate that the phospholipids present in RA SF are less potent lubricators and provide less protection against mechanical damage than in healthy controls. All studies were small and included approximately 20 patients.
Eicosanoids such as prostaglandins and leukotrienes were intensively investigated in SF of RA patients and were recently reviewed . Prostaglandin E2 (PGE 2 ) as well as PLA 2 and COX, enzymes involved in its generation, were found to be higher in SF or were higher expressed in synovium of RA patients (mPGES) . Likewise, the enzyme converting PGE 2 into the inactive 15-PGDH was higher in RA synovium, indicating that regulatory mechanisms targeting PGE 2 are also activated in this disease. Interestingly and in line with this observation, the more anti-inflammatory prostaglandins, PGD 2 and its metabolite 15-deoxy-PGJ 2 were also found in the RA joint. Expression of PGD 2 was detected in cells in RA joints; however, no evidence was found for enhanced biosynthesis of 15-deoxy-PGJ 2 . Next to prostaglandins, leukotrienes such as LTB 4 , LTD 4 and LTE 4 were found in SF of RA patients .
Interestingly, anti-inflammatory products of LOXs were also described in SF of RA patients. These included the anti-inflammatory and pro-resolving mediators lipoxin A4 (LXA 4 ) and the mediators derived from the n-3 PUFA DHA: maresin 1 (Mar1) and resolving D5 (RVD5) . Although this study was performed in only five patients and the effects of pro-resolving lipids on RA development or progression were not yet investigated, their potent immune modulatory functions identify them as promising therapeutic agents for chronic inflammatory diseases.
Besides the PLA, COX and LOX enzymes, autotaxin makes an important contribution to bioactive lipids in RA. It can convert membrane phospholipids such as LPC to lysophospatidic acid (LPA), as well as sphingosylphosphorylcholine to yield sphingosine-1-phosphate (S1P). S1P can act as an intracellular second messenger or extracellular lipid mediator via G-coupled receptors and is involved in regulation of pro-inflammatory pathways and cell migration. Its possible role in RA is discussed later. Both LPA and S1P, as well as autotaxin were shown to be present in SF of RA patients , while higher autotaxin levels were shown in RA SF ( n = 16) and serum ( n = 26) compared to controls .
Lipids in serum/SF of OA
Several lipidomic studies in OA patients have been performed over the recent years. Although most of them study SF, one study investigated the phospholipid profile of hip and knee OA patients ( n = 59 women) in plasma . Similarly to RA, the LPC/PC ratio was higher in both mild and moderate OA (based on radiographic damage) compared to controls, contradictory to that found in SF. In two studies with 30 and 48 OA patients, the levels of phospholipids were found to be elevated in SF of OA patients compared to post-mortem controls , similar to what was found for RA.
Few studies have investigated oxylipin levels in OA. In one study with 10 OA patients, 8-iso-PGF 2α and 15-keto-dihydroPGF 2α were detected in both serum and SF, but the levels did not correlate in these two fluids . This indicates that there might be local production of these lipids, rather than systemic diffusion. Noteworthy is that levels of both metabolites were higher in serum of OA patients compared to controls. Higher levels of these metabolites in serum of OA patients could indicate higher systemic oxidative damage and inflammatory responses. For some patients, the levels of both compounds, but especially 15-keto-dihydroPGF 2α tended to be somewhat higher in SF compared to serum. It is difficult to interpret these data due to the lack of healthy SF and low numbers of patients. More recently, a study investigated the 15-LOX product 15-HETE and the COX product PGE 2 in plasma of symptomatic OA patients and controls . Both metabolites were elevated in patients (three different cohorts with 291 patients in total) compared to controls, which could be a systemic reflection of local inflammation in the knee. Similarly, leukotriene B 4 (LTB 4 ) has also been described in SF of some OA patients, although no comparison with healthy controls was performed . Together, these data indicate that pro-inflammatory lipids are present in OA patients, although a more extensive investigation of the presence of other lipid classes, such as prostaglandins and anti-inflammatory oxylipins, in OA patients is warranted.
Fatty acids and phospholipids
Dietary fatty acids, such as the n-3 PUFA EPA and DHA derived from fish oil. These n-3 PUFAs are believed to have anti-inflammatory properties, while the n-6 PUFA AA is believed to be more pro-inflammatory. Long-chain fatty acids such as AA, EPA and DHA are incorporated in phospholipids and these are essential constituents of membranes. In healthy individuals on a typical Western diet, about 10–20% of phospholipids in membranes of leukocytes are composed of AA, while 0.5–1% is EPA and 1.5–3% is DHA . Dietary intake of EPA and DHA leads to an increase of these fatty acids in cellular membranes and this occurs usually at the expense of AA. This incorporation begins within days and is dose dependent . Because erythrocytes have a lifespan of 100–120 days, their membrane phospholipid composition is generally used to monitor the dietary intake of lipids .
Association with clinical characteristics and intervention studies in RA
Most studies investigating the types of fatty acids present in RA patients and association with clinical parameters focussed on dietary fatty acids.
Several clinical trials have been performed with either fish oil or n-3 supplementation in RA patients. These randomized control trials have been recently summarized in a systematic review by Miles and Calder . The authors analysed 23 studies, in which EPA and DHA doses varied largely, between <1 and >7 g/day and were administered mostly orally as fish oil supplements. The duration of the studies varied between 4 and 52 weeks and the placebo controls were usually other types of oils, such as corn, olive oil and paraffin oil. The sample size of these studies was typically around 20–30 patients/group, excepting few studies with participation of less or more patients. Many studies had methodological shortcomings and no meta-analysis was performed. In general, the observed beneficial effects of n-3 PUFA were related to morning stiffness, the number of tender/swollen joints, grip strength, pain or disease activity. The effects were overall modest.
Since this systematic review, another study investigated the effect of intake of moderate amounts of n-3 PUFA (2.090 g EPA and 1.165 g DHA) in combination with regular anti-inflammatory therapy in 109 RA patients. High-oleic acid sunflower oil was used as control. Although there was an increase in n-3 PUFA and a relative decrease in n-6 PUFA AA in erythrocyte membranes in the treatment group, there was no significant effect on clinical symptoms, non-steroidal anti-inflammatory drug (NSAID) usage, cytokines, eicosanoids and bone turnover markers in this group , which might be attributable to the low dose of n-3 PUFA administered. Similar results regarding the incorporation of n-3 PUFA in erythrocyte membranes at the expense of AA were found in another study with a relative low number of patients (54 RA and six psoriatic arthritis patients) in which patients were randomized into three groups: one treated with 3 g n-3 PUFA/day, one with 3.2 g γ-linolenic acid (n-6 PUFA)/day and one with a combination of a low dose n-3 PUFA (1.6 g) and γ-linolenic acid (1.8 g)/day. The treatment duration was 12 weeks. A relative decrease of AA/EPA ratio was observed in the n-3 PUFA-treated group, similar to earlier studies. In addition, an enrichment in γ-linolenic acid in plasma lipids, cholesterol membranes and erythrocyte membranes was observed, and this enrichment was dependent on the dose indicating that this n-6 PUFA is also dose-dependently incorporated in cellular membranes upon intake .
Interestingly, several recent studies have investigated the effects of n-3 PUFA on early RA patients or on development of RA and related features in pre-RA individuals. Early RA patients ( n = 140) treated with DMARD therapy were additionally treated with 5.5 g/day EPA + DHA or with 0.4 g/day EPA and DHA (control group). The primary outcome was failure on triple DMARD after 3 months. Failure was lower and the rate of remission was higher in the n-3 PUFA-treated group compared to the control group after 3 months of treatment . Remarkably, EPA in plasma phospholipids was inversely associated with time to remission and to DMARD failure, while similar results, albeit not significant, were also obtained for DHA, indicating that EPA might be more effective than DHA as disease modulator.
Several studies investigated the effect of intake of fish containing n-3 PUFA on the risk of development of RA in healthy individuals. In a meta-analysis involving both prospective and case–control studies, the authors found that for each serving of fish per week, the relative risk of RA became 0.96 (95% confidence interval (CI) 0.91; 1.01), while there was a 20–24% lower risk (0.76 (95% CI 0.57; 1.02)) of RA for 1–3 servings of fish/week compared to no servings . In an observational population-based cohort study (32,232 women aged 54–89 years), self-reported intake of >0.21 g/day of n-3 PUFA led to a 35% (95% CI 10–52%) lower risk of developing RA than lower intake. Moreover, long-term consistent intake of >0.21 g/day of n-3 PUFA led to a 52% (95% CI 29–67%) lower risk of RA than lower intake, while long-term consistent intake of >1 serving of fish/week led to 29% lower risk (risk ratio (RR) 0.71; 95% CI 0.48–1.04) compared to <1 serving/week . Although the intake was based on self-reported data and no measurements of n-3 PUFA in phospholipids were presented, these studies suggest that n-3 PUFA could be beneficial for lowering the risk of RA development in pre-disease individuals and this effect is probably dose dependent. This would also be in line with a more recent study in which individuals at risk for developing RA (non-diseased first-degree relatives of RA patients and HLA-DR4+ individuals) and positive for anti-CCP antibodies ( n = 30) were compared to seronegative individuals ( n = 47) for the amount of n-3 PUFA intake (self-reported) and the percentage of n-3 PUFA in erythrocyte membranes. The findings indicate that anti-CCP + individuals were less likely to report n-3 PUFA intake, and the anti-CCP positivity was inversely correlated to the percentage of n-3 PUFA in erythrocyte membranes . Altogether, these findings suggest that n-3 PUFA might affect the risk for development of RA, as well as the clinical parameters of the disease in RA patients.
Supporting these data, interventions with DHA and EPA have been also shown to reduce the onset, incidence and severity of collagen-induced arthritis (CIA) in mice .
Other fatty acids have only scarcely been studied. Although not yet investigated in relation to human RA, α-lipoic acid (ALA) showed promising results in mice, as it inhibited joint inflammation and bone destruction in the CIA model both when administered intra-peritoneally and through diet . Although the mechanisms are unclear, ALA inhibited synovial inflammation in both studies, as well as human osteoclast differentiation in vitro and in mice in vivo .
Supporting a possible role for LPA/autotaxin axis in arthritis, inhibition of the LPA receptor inhibits development of disease in the K/BxN serum transfer model through effects on cartilage destruction and bone erosions, possibly through inhibition of osteoclast differentiation and activity and promotion of osteoblast differentiation . Similarly, mice lacking autotaxin in the mesenchymal cell compartment had less arthritis on a hTNF-α (tumour necrosis factor-α) transgenic background, as well as in the CIA model of arthritis and this effect was likely mediated by LPA effects on synovial fibroblasts . Interestingly, preventive oral administration of LPC precursor PC to CIA mice also inhibited severity of CIA, possibly through inhibition of leukocyte–endothelium interactions and nitric oxide (NO) production . These data indicate that the PC/LPC/LPA metabolic pathway might represent a promising target for intervention in RA patients.
Association with clinical characteristics and intervention studies in OA
Two studies have measured phospholipids in both early OA and late OA patients, with early and late being defined by the Outerbridge classification scale. One of these studies assessed the difference in lipids between early ( n = 17) and late OA ( n = 13) and showed that concentrations of 66 phospholipid species were different between these stages and that the total lipid content was higher in late OA compared to early OA . In addition, the PC/LPC ratio was higher in late OA then early OA, possibly indicating a higher activation of PLA 2 in early compared to late disease. A metabolomic approach can also be used to classify OA patients based on their metabolite profile. Using this approach, it was found that a group of 80 OA patients was divided into two distinct groups particularly based on the levels of carnitines and its acyl esters acylcarnitines (involved in fatty acid metabolism) in knee SF . The group with lower acylcarnitine levels could be further divided into two subgroups based on their glycophospholipid and sphingomyelin (SM) levels. Although the groups were not correlated to any clinical OA characteristics, these data indicate that OA patients can be divided in metabolically distinct groups. An earlier study by Kosinska et al. did correlate the levels of SM with relevant OA features . All SM species measured in SF were 2.4-fold higher in early OA patients ( n = 17) compared to controls and twofold higher in late ( n = 13) versus early OA patients. Early and late OA patients were classified with the Outerbridge score.
Most reports studying lipids and their association with clinical parameters have studied fatty acids, rather than phospholipids or other higher-order lipids. To the best of our knowledge, only one study investigated the association of plasma fatty acid levels with clinical features in OA patients. This study measured fasting plasma levels of fatty acids in 472 individuals, including OA patients, as well as individuals at risk of developing knee OA . N-3 fatty acids, in particular DHA, were inversely correlated with patellofemoral cartilage loss, but not tibiofemoral cartilage loss or synovitis at 30-month follow-up, suggesting a protective effect of this n-3 fatty acid on selected structural findings. The levels of the n-6 fatty acid AA, which is believed to be more pro-inflammatory, were positively correlated with synovitis. In line with this result, in another study on knee OA patients, the AA levels were found to correlate with histologic disease severity .
While fish oil or n-3 PUFA supplementation has been intensively studied in RA patients, only a few intervention studies have been performed in OA patient groups. Most of these studies aimed at reducing pain. Forty-seven patients with knee or hip OA, consuming 1200 mg fish oil (total n-3 18% EPA and 12% DHA) a day for 12 weeks in a randomized trial, showed no improvement on either visual analogue scale (VAS) for patient self-assessment of pain or the health assessment questionnaire (HAQ) for patient self-assessment of activity . In addition, a double-blind placebo-controlled trial in 86 OA patients failed to detect any benefit from taking either 10 ml cod liver oil (786 mg EPA) or olive oil per day for 24 weeks, next to the regular intake of NSAIDs . By contrast, one study reported the beneficial effects of fish oil in OA patients. A randomized, double-blind multicentre trial investigated the effect of a low dose of 0.45 g of fish oil (18% EPA and 12% DHA) per day versus a high dose of 4.5 g on pain and function scores in 202 knee OA patients after 2 years . Both groups benefitted from the treatment, although unexpectedly, the group receiving low-dose fish oil benefitted the most, with significantly lower pain scores (WOMAC index) at 18 and 24 months and better function limitation scores after 24 months. No beneficial effects were observed on cartilage volume or bone marrow lesions. It should be mentioned, however, that the low-dose fish oil preparation also contained Sunola oil which includes n-9 monounsaturated oleic acid. Therefore, it is difficult to conclude whether the observed effects were due to the n-3 fatty acids alone .
Despite the unclear effects of n-3 PUFA on pain in humans, they were shown to reduce radiographic damage in mouse models of OA, both in spontaneous and surgically induced models . Interestingly, their beneficial effect was correlated to a better wound healing, while saturated fatty acids and n-6 fatty acids had opposite effects on OA severity and wound healing , suggesting the existence of common mechanisms involved in both processes. Moreover, it suggests a possible benefit of n-3 PUFA on structural damage, which still needs to be investigated in humans.
Effects on RA human tissues in vitro
Free fatty acids have been described in serum and SF of RA patients and they could affect joint tissues and immune cells involved in disease pathogenesis. Indeed, in vitro treatment of RA synovial fibroblast with free fatty acids induced pro-inflammatory cytokines interleukin (IL)-6, IL-8, monocyte chemoattractant protein (MCP)-1 and matrix metalloproteinases (MMPs) and this effect was independent of the length or saturation degree of the fatty acid, but variable between donors . For chondrocytes, lipids isolated from SF could inhibit chondrocyte proliferation , while saturated fatty acids induced IL-6 . Endothelial cells only responded to high concentrations of fatty acids such as palmitic acid and linoleic acid, but not oleic acid . Taken together, these data indicate that different cell types respond differently to individual fatty acids and that combinations of fatty acids could have a different effect than the individual components. Moreover, free fatty acids and especially dietary fatty acids can potently influence bone metabolism through regulation of PGE 2 and leukotrienes, stimulating bone resorption or prostaglandin-mediated regulation of insulin-like growth factor 1 (IGF-1), a growth factor that stimulated bone formation . The role of these processes in bone metabolism in RA remain to be investigated.
Besides their effect on joint tissues, fatty acids also have immune modulatory effects. Effects of n-3 PUFA on cells derived from RA patients have been investigated, while much less attention has been given to other fatty acids. N-3 PUFA had an anti-inflammatory effect on cytokine secretion by monocytes derived from RA patients , while inhibiting formation of reactive oxygen species and the AA-mediated induction of TNFRI and –II on RA neutrophils . Fatty acids can also affect T cell function and B cell function , although the effects of fatty acid type, saturation and length, as well as the specific effects on T and B cells from RA patients remain to be addressed.
Some of the phospholipids found in serum or SF have been shown to affect joint tissues and cells thereby potentially contributing to RA. Among these, LPC and LPA have received special attention, as they were described to have potent immune modulatory effects (reviewed in Ref. ). For example, LPA has been shown to induce COX2 expression in RA synovial fibroblasts, either alone or in combination with pro-inflammatory cytokines such as IL-1α or β , to induce their proliferation and enhance production of IL-6, VEGF, CCL 2 and MMP3, as well as expression of VCAM and migration . Although the effects of LPC on human RA tissues were not yet investigated, LPC was shown to induce COX2 expression in vascular endothelial cells and macrophages , which would suggest a possible pro-inflammatory role of this lipid in RA.
Another product of autotaxin found in RA SF, S1P, was also shown to have effects on RA-derived tissues, including stimulation of proliferation of RA synovial fibroblasts and enhancing COX2 and PGE 2 production in synovial fibroblasts and chondrocytes , or by decreasing aggrecan production by chondrocytes and stimulating osteoblast differentiation .
Together, these data indicate a pro-inflammatory effect of the PC/LPC/LPA axis in RA.
Effects on OA human tissues in vitro
While intervention studies with fatty acid supplementation mainly focussed on reducing pain in patients, the in vitro studies performed focussed predominantly on the effects of fatty acids on OA tissues and the mechanisms underlying these effects. Chondrocytes were most studied in this respect and they were usually stimulated with cytokines believed to have a prominent role in OA. Chondrocytes from knee cartilage explants of OA patients were shown to take up linoleic, oleic and palmitic acid, regardless of whether they were stimulated with TNF-α . Upon oleic acid exposure, TNF-α-stimulated chondrocytes secrete less GAG and downregulate the expression of MMP-1 and COX-2, indicating an anti-inflammatory effect. Palmitic acid had similar effects on GAG release and MMP-1 expression but did not affect PGE 2 release or COX-2 expression. By contrast, AA precursor linoleic acid increased the release of PGE 2 by TNF-α-stimulated chondrocytes. These results suggest that local fatty acid concentrations can contribute to cartilage damage. In another study, the effect of palmitate and oleic acid on chondrocytes of OA patients was investigated in comparison to post-mortem control chondrocytes . The data indicated that IL-1β together with palmitate synergistically increased IL-6 and COX-2 expression, whereas oleic acid did not in both donor types, indicating a rather pro-inflammatory effect of palmitic acid. In addition, increased apoptotic cell death was observed in chondrocytes stimulated with both IL-1β and palmitate in the post-mortem cartilage. The contrasting effects of palmitic acid on chondrocytes in these two studies could be due to differences in stimulus used to mimic OA-related inflammation in chondrocytes.
Related posts:
HLA and anti-citrullinated protein antibodies: Building blocks in RA
Translating genomics into mechanisms of disease: Osteoarthritis
From mechanisms of action to therapeutic application: A review on current therapeutic approaches and future directions in systemic sclerosis
New biological therapies in Sjögren’s syndrome
Metabolomics to identify biomarkers and as a predictive tool in inflammatory diseases
Corrigendum to “Risk of infection with biologic antirheumatic therapies in patients with rheumatoid arthritis” [Best Pract Res Clin Rheumatol 29 (2) (2015) 290–305]
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
