intercellular adhesion molecule
monocyte chemoattractant protein-1
platelet-derived growth factor
vascular cell adhesion molecule
The World's most important cardiac societies have issued guidelines that recommend the intake of the two marine n-3 fatty acids EPA and DHA at 1 g/d for CVD prevention, treatment after a myocardial infarction and prevention of sudden death and secondary disease (De Backer et al. Reference De Backer, Ambrosioni, Borch-Johnsen, Brotons, Cifkova and Dallongeville2003; Priori et al. Reference Priori, Aliot, Blomstrom-Lundqvist, Bossaert, Breithardt and Brugada2003; Van der Werf et al. Reference Van der Werf, Ardissino, Betriu, Cokkinos, Falk and Fox2003; Smith et al. Reference Smith, Allen, Blair, Bonow, Brass and Fonarow2006). National cardiac societies have followed suit (Wirth & Gohlke, Reference Wirth and Gohlke2005). These recommendations are based not only on intervention trials with these n-3 fatty acids (Burr et al. Reference Burr, Fehily, Gilbert, Rogers, Holliday, Sweetnam, Elwood and Deadman1989, Reference Burr, Ashfield-Watt, Dunstan, Fehily, Breay, Ashton, Zotos, Haboubi and Elwood2003; The GISSI Prevenzione Group, 1999; Marchioli et al. Reference Marchioli, Barzi, Bomba, Chieffo, Di Gregorio and Di Mascio2002; Yokoyama et al. Reference Yokoyama and Origasea2003), but also on a wealth of literature describing mechanisms of actions, animal models, studies with surrogate and intermediate factors and other aspects (von Schacky, Reference von Schacky1987, Reference von Schacky2003). Taken together, the scientific basis of the current guidelines is so strong that they have been established despite a null result in a recent Cochrane analysis (Hooper et al. Reference Hooper, Thompson, Harrison, Summerbell, Ness and Moore2006).
EPA and DHA have been demonstrated not only to have an antiarrhythmic effect (Leaf et al. Reference Leaf, Albert, Josephson, Steinhaus, Kluger, Kang, Cox, Zhang and Schoenfeld2005; Raitt et al. Reference Raitt, Connor, Morris, Kron, Halperin and Chugh2005), but also to mitigate the course of coronary atherosclerosis (von Schacky et al. Reference von Schacky, Angerer, Kothny, Theisen and Mudra1999) and to stabilize unstable plaque, e.g. in carotid arteries (Thies et al. Reference Thies, Garry, Yaqoob, Rerkasem, Williams, Shearman, Gallagher, Calder and Grimble2003). These two findings are manifestations of an anti-inflammatory effect of EPA and DHA, currently used in the treatment of inflammatory disorders such as rheumatoid arthritis (Kremer, Reference Kremer2000). For some time atherosclerosis has been considered to be a disease with an inflammatory component (for example, see Ross, Reference Ross1999). The present review considers investigations aimed at unravelling the mechanisms by which EPA and DHA exert their anti-inflammatory effects through alterations of cytokine metabolism. Since differences exist in cytokine metabolism in vitro v. in vivo and in experimental animals v. in human subjects the present review will largely focus on work done in human subjects.
Cytokines related to inflammation and atherosclerosis
IL-1β is an important cytokine that has a plethora of actions, including a pronounced pro-inflammatory effect and increasing the expression of adhesion molecules (Dinarello, Reference Dinarello2006). Levels of IL-1β decrease by 61–90% when healthy volunteers and patients with rheumatoid arthritis ingest doses of between 2·7 and 5·8 g EPA and DHA/d (for review, see James et al. Reference James, Gibson and Cleland2000). Furthermore, IL-1β levels associated with stimulation by strenuous exercise (a marathon run) are not influenced by previous ingestion of 3·6 g EPA and DHA/d (Toft et al. Reference Toft, Thorn, Ostrowski, Asp, Moller, Iversen, Hermann, Sondergaard and Pedersen2000), which suggests that IL-1β levels clearly respond to a strong stimulus. These findings might partially explain why infectious complications are not seen more frequently in large-scale intervention trials using EPA and DHA (von Schacky & Harris, Reference von Schacky and Harris2006).
IL-6 is a cytokine that together with a soluble IL-6 receptor plays an important role in perpetuating an inflammatory state (Scheller et al. Reference Scheller, Ohnesorge and Rose-John2006). IL-6 is generally reduced, as assessed ex vivo, after supplementation of the human diet with n-3 fatty acids (for review, see Calder, Reference Calder2005). In the supernatant fraction of unstimulated mononuclear cells the reduction is more pronounced after 0·3 g EPA and DHA/d than after 1·0 or 2·0 g EPA and DHA/d (Trebble et al. Reference Trebble, Arden, Stroud, Wootton, Burdge, Miles, Ballinger, Thompson and Calder2003). However, in the supernatant fraction of stimulated mononuclear cells the reduction is clearly dose-related (Trebble et al. Reference Trebble, Arden, Stroud, Wootton, Burdge, Miles, Ballinger, Thompson and Calder2003).
An epidemiological study (Ferrucci et al. Reference Ferrucci, Cherubini, Bandinelli, Bartali, Corsi, Lauretani, Ma, Andres-Lacueva, Senin and Gualnik2006) has shown that low levels of this anti-inflammatory compound are associated with low levels of DHA. After treatment for 1 year with 3·4 g EPA and DHA/d, the IL-10 levels of forty-five recipients of heart transplants were found to be decreased (Holm et al. Reference Holm, Berge, Anderassen, Ueland, Kjekshus, Simonsen, Froland, Gullestad and Aukrust2001). In healthy volunteers after supplementation with 7 g EPA and DHA/d for 4 weeks IL-10 mRNA steady-state levels were found to be unaltered in either unstimulated mononuclear cells or monocytes that had been adherence-activated ex vivo (Baumann et al. Reference Baumann, Hessel, Larass, Müller, Angerer, Kiefl and von Schacky1999). There have been no other reports of studies in human subjects. Thus, there is at present no clear picture of the effects of EPA and DHA on levels of IL-10.
TNFα is a pro-inflammatory cytokine that has a large number of effects, among them increased body temperature, reduced appetite and stimulation of other immunmodulatory cytokines (Grimble, Reference Grimble1996). TNF is thought to be a propagator of atherosclerosis (Ross, Reference Ross1999). The effects of fish oil on TNFα production by peripheral blood mononuclear cells have been investigated in eleven studies of healthy volunteers, of which six studies have demonstrated a suppressive effect (for review, see Grimble et al. Reference Grimble, Howell, O'Reilly, Turner, Markovic, Hirrell, East and Calder2002). These apparently discrepant findings can be explained by the effects of inherent TNFα production and by polymorphisms in the TNFα and lymphotoxin α genes (Grimble et al. Reference Grimble, Howell, O'Reilly, Turner, Markovic, Hirrell, East and Calder2002). These polymorphisms might also explain an unexpected increase in TNF in recipients of heart transplants after supplementation with 3·4 g EPA and DHA/d for 1 year (Holm et al. Reference Holm, Berge, Anderassen, Ueland, Kjekshus, Simonsen, Froland, Gullestad and Aukrust2001). In a dose–response study of healthy volunteers levels of TNFα in the supernatant fraction of unstimulated and stimulated mononuclear cells was shown to decrease (Trebble et al. Reference Trebble, Arden, Stroud, Wootton, Burdge, Miles, Ballinger, Thompson and Calder2003). Interestingly, the decrease was found to be less pronounced after consuming 2·0 g EPA and DHA/d than after consuming 1·0 g EPA and DHA/d.
DHA, but not EPA, reduces the expression of pro-inflammatory IL-1, IL-6 and TNFα in vitro (De Caterina et al. Reference De Caterina, Zampolli, Del Turco, Madonna and Massaro2006).
Platelet-derived growth factor
Platelet-derived growth factor (PDGF) stimulates smooth muscle cell proliferation and plays a role in the migration of these cells into neointima following injury and in atherosclerosis (Raines, Reference Raines2004). PDGF is thought to play a major role in the proliferation of atherosclerotic lesions (Ross, Reference Ross1999). In volunteers ingesting 7 g EPA and DHA/d levels of mRNA coding for PDGF-A and -B were found to be reduced by 58% after 1 week and by 70% after 6 weeks, with these levels remaining constant in controls on an unaltered Western diet (Kaminski et al. Reference Kaminski, Jendraschak, Kiefl and von Schacky1993). This finding was the first demonstration that a deliberate change in diet can alter gene expression quantitatively. In a subsequent study, with the same dose of EPA and DHA but comparing it with 7 g n-6 fatty acids/d and 7 g n-9 fatty acids/d (Baumann et al. Reference Baumann, Hessel, Larass, Müller, Angerer, Kiefl and von Schacky1999), the reduction in the levels of PDGF-A and -B mRNA was found to be quantitatively less pronounced (for PDGF-A −25 (sd 10) %, for PDGF-B −31 (sd 13) %) in non-stimulated mononuclear cells. However, after cell adherence for 4 h or 20 h the reduction was found to persist to a quantitatively similar extent (Baumann et al. Reference Baumann, Hessel, Larass, Müller, Angerer, Kiefl and von Schacky1999). Lower doses of EPA and DHA (0·3, 0·6 or 0·9 g n-3 fatty acids/d) have no effect on serum mitogenic activity or serum PDGF levels (Wallace et al. Reference Wallace, McCabe, Roche, Higgins, Robson, Gilmore, McGlynn and Strain2000).
Monocyte chemoattractant protein-1
Monocyte chemoattractant protein-1 (MCP-1), acting through its receptor chemokine (C–C motif) receptor 2, appears to play an early and important role in the recruitment of monocytes to atherosclerotic lesions and in the formation of intimal hyperplasia after intimal injury (Charo & Taubman, Reference Charo and Taubman2004). Supplementation with dietary n-3 fatty acids (EPA and DHA), but not n-6 or n-9 fatty acids, at 7 g /d for 4 weeks reduces MCP-1 mRNA levels in unstimulated mononuclear cells ex vivo by 40% (Baumann et al. Reference Baumann, Hessel, Larass, Müller, Angerer, Kiefl and von Schacky1999). After stimulation of the cells by adherence for 4 h and 20 h MCP-1 mRNA levels are reduced by 35 (sd 20) % and 30 (sd 8) % respectively (Baumann et al. Reference Baumann, Hessel, Larass, Müller, Angerer, Kiefl and von Schacky1999). It is quite likely that the reductions in gene expression translate into reduced levels of MCP-1 before and after stimulation in vivo, although the direct proof will be quite difficult to obtain. Evidence from in vitro studies (De Caterina & Massaro, Reference De Caterina and Massaro2005) indicates that both EPA and DHA decrease agonist-induced activation of NF-κB and increase PPAR. This effect might very well play a role in the reduced gene expression for PDGF-A and -B and for MCP-1 in vivo after EPA and DHA.
C-reactive protein is a marker of systemic inflammation that is currently considered to be a risk factor for CVD (Tsimikas et al. Reference Tsimikas, Willerson and Ridker2006). Given the anti-inflammatory and anti-atherosclerotic effects of EPA and DHA, reduced levels of C-reactive protein would be expected to occur after supplementation. While an epidemiological study (Pischon et al. Reference Pischon, Hankinson, Hotamisligil, Rifai, Willett and Rimm2003) has shown an inverse relationship between the intake of EPA and DHA and levels of C-reactive protein, human intervention studies (for review, see Mori & Beilin, Reference Mori and Beilin2004) have not found a reduction in C-reactive protein levels after ingestion of EPA and DHA. Thus at present no clear picture of the relationship between EPA and DHA and C-reactive protein has emerged.
Cytokines related to endothelial activation
Vascular cell adhesion molecule (VCAM)-1, intercellular adhesion molecule-1 (ICAM-1) and E-selectin have a role in the pathogenesis of atherosclerosis (Hope & Meredith, Reference Hope and Meredith2003). Plasma levels can be measured because soluble (s) VCAM and sICAM are shed from the cell surface. Plasma levels of sE-selectin can also be measured (Roldan et al. Reference Roldan, Marin, Lip and Blann2003). Levels of VCAM and ICAM correlate to some extent with the presence of clinical atherosclerosis (Hope & Meredith, Reference Hope and Meredith2003). However, VCAM-1 and ICAM-1 have a role in the early phase of the development of the atherosclerotic lesion, i.e. in monocyte recruitment by endothelial cells, before the appearance of macrophages (foam cells) in the intima (De Caterina & Massaro, Reference De Caterina and Massaro2005). E-selectin is expressed on the endothelial surface slightly earlier in the sequence of events (Roldan et al. Reference Roldan, Marin, Lip and Blann2003). sVCAM-1, sICAM-1 and sE-selectin are considered to be markers of endothelial activation (Hope & Meredith, Reference Hope and Meredith2003; Roldan et al. Reference Roldan, Marin, Lip and Blann2003). Whether they are risk factors for atherosclerosis or its clinical events is a matter of debate.
Stimulated endothelial cells express less sVCAM, sICAM or sE-selectin in the presence of n-3 fatty acids, with DHA being more potent than EPA (De Caterina et al. Reference De Caterina, Madonna and Massaro2004). DHA reduces mRNA levels by inhibition of the activation of the NF-κB system of transcription factors (De Caterina et al. Reference De Caterina, Madonna and Massaro2004). Corresponding findings from studies in experimental animals have been published (for review, see Calder, Reference Calder2004).
Studies in human subjects have yielded less clear results. In a randomized study of forty-one male smokers with hyperlipidaemia (Seljeflot et al. Reference Seljeflot, Arnesen, Brude, Nenseter, Drevon and Hjermann1998) 4·8 g EPA and DHA/d was shown to increase levels of sVCAM-1 and sE-selectin. In middle-aged healthy volunteers (Thies et al. Reference Thies, Miles, Nebe-von-Caron, Powell, Hurst, Newsholme and Calder2001) supplementation with 1 g EPA and DHA/d for 12 weeks was found to reduce levels of sVCAM-1 (–28%) and sE-selectin (−17%), while levels of sICAM-1 were unaltered. In another 12-week randomized double-blind study of healthy subjects by the same group (Miles et al. Reference Miles, Thies, Wallace, Powell, Hurst, Newsholme and Calder2001) levels of all three compounds were found to be unaltered after fish oil supplementation. In a 1-year double-blind study of 300 patients after a myocardial infarction (Grundt et al. Reference Grundt, Nilsen, Mansoor, Hetland and Nordoy2003) treatment with EPA and DHA was not found to reduce levels of sICAM or sE-selectin. In individuals with type 2 diabetes, however, levels of sE-selectin are reduced after treatment with EPA or DHA (Nomura et al. Reference Nomura, Kanazawa and Fukuhara2003; Woodman et al. Reference Woodman, Mori, Burke, Puddey, Barden, Watts and Beilin2003). Administration of 2·4 g EPA and DHA/d to 171 elderly men at risk for coronary artery disease has been shown to increase sVCAM (Berstad et al. Reference Berstad, Seljeflot, Veierod, Hjerkinn, Arnesen and Pedersen2003). A randomized study (Eschen et al. Reference Eschen, Christensen, De Caterina and Schmidt2004) comparing EPA and DHA at doses of 2·0 g/d and 6·6 g/d in healthy subjects has shown a decrease in sE-selectin only when fed at a dose of 6·6 g/d. In a large (563 elderly subjects) 3-year randomized study that compared (using a factorial design) dietary advice, 2·4 g EPA and DHA/d and no treatment (Hjerkinn et al. Reference Hjerkinn, Seljeflot, Ellingsen, Berstad, Berstad, Hjermann, Sandvik and Arnesen2005) EPA and DHA was found to reduce levels of sICAM.
Thus, while a clear picture has emerged in vitro and in animal studies, demonstrating reductions in plasma levels of sVCAM, sICAM and sE-selectin, investigations in human subjects have yielded mixed results. This disparity may be related to differences in populations studied and doses used.
The measurement of urinary F2 isoprostane excretion by GC–MS is currently thought to best reflect oxidative stress in vivo (Mori & Beilin, Reference Mori and Beilin2004). In contrast to theoretical concerns and earlier observations with less-refined methodology, oxidative stress has been consistently shown to be reduced after ingestion of EPA and DHA both in combination and individually (Mori & Beilin, Reference Mori and Beilin2004). Reduced oxidative stress is thought to contribute to the anti-atherosclerotic actions of n-3 fatty acids, possibly through immunomodulation and decreased leucocyte activation (Mori & Beilin, Reference Mori and Beilin2004).
Taken together EPA and DHA reduce levels of the pro-inflammatory cytokines IL-1β, IL-6 and TNFα in human subjects. Moreover, mRNA levels of pro-atherosclerotic growth factors, such as PDGF-A and -B, and MCP-1 are reduced in mononuclear cells after supplementing the human diet with EPA and DHA. The cytokines and growth factors mentioned play a role in the propagation of the atherosclerotic lesion. In vitro the levels of sICAM, sVCAM and sE-selectin are reduced by the presence of EPA or DHA. These cytokines reflect endothelial activation. Data from human studies are less clear cut, probably because endothelial activation is transient. It is currently thought that the effects of EPA and DHA on cytokines and growth factors are the mechanisms responsible for the anti-atherosclerotic action of these n-3 fatty acids.