n-3 Polyunsaturated fatty acids

Although beyond the scope of the present review, a number of investigators have recently addressed the role of n-3 PUFA in suppressing chronic inflammation and cancer⁽Reference Chapkin, McMurray and Lupton^4–Reference Serhan, Yacoubian and Yang⁶⁾. With respect to mechanisms which functionally link the pleiotropic effects of bioactive dietary long-chain n-3 PUFA, inflammation and cancer, examples include (i) metabolic interconversion into novel bioactive eicosanoids⁽Reference Ariel and Serhan^7, Reference Bazan⁸⁾, (ii) modulation of nuclear receptor activation, gene transcription and translation⁽Reference deUrquiza, Liu, Sjoberg, Zetterström, Griffiths, Sjövall and Perlmann^9–Reference Jump, Botolin, Wang, Xu, Christian and Demeure¹²⁾ (Fig. 1), (iii) alteration of membrane phospholipid composition and functionality of self-organising lipid domains⁽Reference Ma, Seo, Switzer, Fan, McMurray, Lupton and Chapkin¹³⁾ (Fig. 2), (iv) effects on protein trafficking, including cytosol-to-membrane translocation⁽Reference Chapkin, McMurray and Lupton^4, Reference Seo, Barhoumi, Johnson, Lupton and Chapkin¹⁴⁾ and (v) interaction with SCFA to trigger lipid oxidation and intracellular Ca²⁺ compartmentalisation⁽Reference Kolar, Barhoumi, Lupton and Chapkin^15, Reference Kolar, Barhoumi, Callaway, Fan, Wang, Lupton and Chapkin¹⁶⁾.

Fig. 1 Nuclear receptor activation by conjugated linoleic acid (CLA). FABP, fatty acid-binding proteins (molecular chaperone); ER, endoplasmic reticulum; RXR, retinoid X receptors. CLA transactivates PPAR nuclear receptors. n-3 PUFA suppress NF-κB activation. All membranes incorporate EPA, DHA and conjugated PUFA to different degrees.

Fig. 2 Putative membrane microdomain-altering properties of n-3 PUFA and conjugated linoleic acid (CLA). Dietary DHA and CLA are incorporated into both the bulk phase of the plasma membrane as well as discrete heterogeneous cholesterol/sphingolipid-rich raft domains. This can alter plasma membrane organisation of inner leaflets and the dynamic partitioning of transduction proteins, thereby modulating their function.

The health benefits of long-chain PUFA are diverse and nutritional studies continue to demonstrate important benefits from the consumption of n-3 PUFA⁽Reference Chapkin, McMurray and Lupton^4, Reference Ariel and Serhan^7, Reference Bazan^8, Reference Chapkin, Wang, Fan, Lupton and Prior^17–Reference Harris, Poston and Haddock¹⁹⁾. Since the use of a health claim on labels for foods containing n-3 PUFA has been approved, food companies are now mobilising to incorporate these fatty acids into a range of novel commercial foods in order to provide for the wider public consumption of these bioactive compounds. Hence, it is now important to precisely determine how specific long-chain PUFA modulate cell phenotype and reduce the risk of developing cancer and inflammatory disorders.

Conjugated linoleic acid

There is growing interest with regard to the use and commercial availability of conjugated positional and geometric isomers of PUFA, particularly conjugated dienoic isomers of linoleic acid (conjugated linoleic acid; CLA). For example, in certain model systems, CLA is a powerful anti-cancer agent, capable of promoting growth arrest and apoptosis in tumour cells⁽Reference Ou, Ip, Lisafeld and Ip^20–Reference Song, Sneddon, Heys and Wahle²²⁾. In addition, CLA triggers adipose delipidation in rodent species⁽Reference Chung, Brown, Provo, Hopkins and McIntosh^23–Reference Pérez-Matute, Marti, Martínez, Fernández-Otero, Stanhope, Havel and Moreno-Aliaga²⁵⁾, and although there has been very little published clinical research⁽Reference Navarro, Fernandez-Quintela, Churruca and Portillo^26, Reference Tricon and Yaqoob²⁷⁾, recent preliminary evidence suggests that mixed-isomer CLA supplementation can alter fat oxidation and energy expenditure in human subjects⁽Reference Close, Schoeller, Watras and Nora^28, Reference Whigham, Watras and Schoeller²⁹⁾. However, the precise mechanism of action remains elusive. Since n-3 PUFA and CLA exhibit overlapping phenotypic properties, we propose a unifying molecular mechanism which may in part explain their protective effects.

Effect of bioactive polyunsaturated fatty acids on cell membranes

It is generally believed that the plasma membrane consists of a mosaic of functional microdomains that facilitate interactions between resident proteins and lipids⁽Reference Laude and Prior^30, Reference Hancock³¹⁾. Visible examples of these include caveolae, flask-shaped invaginations containing the structural protein caveolin-1 and many signal transduction proteins⁽Reference Anderson³²⁾. In addition, morphologically heterogeneous featureless microdomains, consisting mostly of cholesterol and sphingolipids, unable to integrate well into the fluid phospholipid bilayers, exist as ‘lipid rafts’⁽Reference Simons and Ikonen³³⁾. Although the existence of lipid rafts is still debated, new sophisticated imaging approaches have started to define cell surface nanoscale organisation⁽Reference Hancock³¹⁾. Significantly, both cholesterol-dependent microdomains, analogous to lipid rafts, and non-raft signalling microdomains have been observed using electron microscopic imaging of two-dimensional plasma membrane sheets⁽Reference Prior, Muncke, Parton and Hancock³⁴⁾. These studies have provided a template for further investigation into the effects of dietary PUFA on cell surface organisation and cell cytokinetics, apoptosis and disease progression.

With respect to the biological effects of n-3 PUFA, increasing evidence suggests that DHA is a unique fatty acid because it significantly alters basic properties of cell membranes, including acyl (ester-linked fatty acid) chain order and fluidity, phase behaviour, elastic compressibility, ion permeability, fusion, rapid flip-flop and resident protein function⁽Reference Stillwell and Wassall³⁵⁾. In part, due to the number of cis double bonds, DHA is sterically incompatible with sphingolipid and cholesterol and, therefore, appears to alter lipid raft behaviour⁽Reference Stillwell and Wassall³⁵⁾. Interestingly, a number of studies have recently demonstrated that dietary n-3 PUFA are incorporated into diverse cell types and appear to uniquely modulate cell membrane microdomains⁽Reference Ma, Seo, Switzer, Fan, McMurray, Lupton and Chapkin^13, Reference Fan, Ly, Barhoumi, McMurray and Chapkin^36–Reference Geyeregger, Zeyda, Zlabinger, Waldhäusl and Stulnig³⁹⁾. Indeed, we recently demonstrated that n-3 PUFA feeding can markedly alter lipid/protein composition of mouse colonic caveolae microdomains, thereby selectively modulating the localisation and function of caveolar proteins⁽Reference Ma, Seo, Switzer, Fan, McMurray, Lupton and Chapkin^13, Reference Seo, Barhoumi, Johnson, Lupton and Chapkin^14, Reference Ma, Seo, Davidson, Callaway, Fan, Lupton and Chapkin³⁷⁾. In addition, we demonstrated that H-Ras and endothelial NO synthase are displaced from caveolae in n-3 PUFA-fed mice, which was associated with the suppression of Ras-dependent signalling. In contrast, localisation of non-caveolae-resident proteins, K-ras and clathrin, was not affected, indicating selective displacement of acylated signalling proteins from caveolae by n-3 PUFA. Our findings highlight a novel modality by which n-3 PUFA influence membrane micro-organisation, thereby modulating biological responses.

Using T-cell-culture models, Stulnig et al. were the first to document the ability of PUFA enrichment to selectively modify the cytoplasmic layer of lipid rafts⁽Reference Stulnig, Berger, Sigmund, Raederstorff, Stockinner and Waldhäusl^40, Reference Stulnig, Huber, Leitinger, Imre, Angelisová, Nowotny and Waldhäusl⁴¹⁾. In complementary experiments, we investigated the effect of dietary n-3 PUFA on cholesterol/sphingolipid-rich plasma membrane microdomains (i.e. rafts) in mouse splenic T-cells⁽Reference Fowler, Chapkin and McMurray^5, Reference Fowler, McMurray, Fan, Aukema and Chapkin^42, Reference Switzer, McMurray, Morris and Chapkin⁴³⁾. A very novel and unexpected outcome from this effort was the demonstration that dietary n-3 PUFA reduced (by about 45 %) lipid raft sphingolipid content and altered raft fatty acid composition⁽Reference Fan, Ly, Barhoumi, McMurray and Chapkin^36, Reference Fan, McMurray, Ly and Chapkin^44, Reference Switzer, Fan, Wang, McMurray and Chapkin⁴⁵⁾. Therefore, we hypothesised that PUFA classes (n-6 v. n-3) differentially modulate T-cell membrane microdomains, which is supported by recent studies indicating that stimulation-induced protein kinase C θ translocation into T-cell lipid rafts is suppressed by dietary n-3 PUFA⁽Reference Fan, Ly, Barhoumi, McMurray and Chapkin³⁶⁾. In addition, in an attempt to further probe the effects of DHA on protein kinase C θ effector pathway signalling, we have recently demonstrated that the diet modification of lipid rafts is associated with the suppression of transcription factor NF-κB, AP-1 transcription site activation, IL-2 secretion and lymphoproliferation⁽Reference Fan, Ly, Barhoumi, McMurray and Chapkin³⁶⁾. With respect to lymphocyte subsets, recent studies indicate that the macromolecular complex organisation in lipid rafts is distinct in non-polarised, T helper cell, Th1 and Th2 polarised subsets⁽Reference Leitenberg, Balamuth and Bottomly⁴⁶⁾. This would suggest that these subsets, i.e. regulators of cell-mediated immunity (Th1) and humoral immunity (Th2), would respond differently to dietary PUFA-induced perturbation. However, the ability of DHA to influence membrane raft-mediated signalling in polarised T-cells has not been determined to date.

There is cogent evidence indicating that lipid raft integrity is a prerequisite for optimised signalling between T-cells and antigen-presenting cells⁽Reference Gaus, Chklovskaia, Fazekas de St Groth, Jessup and Harder^47, Reference Harder, Rentero, Zech and Gaus⁴⁸⁾. In addition, recent studies suggest that long-chain PUFA can block antigen presentation by interfering with lipid raft-dependent formation of the immunological synapse⁽Reference Zeyda, Säemann, Stuhlmeier, Mascher, Nowotny, Zlabinger, Waldhäusl and Stulnig^38, Reference Geyeregger, Zeyda, Zlabinger, Waldhäusl and Stulnig^39, Reference Shaikh and Edidin⁴⁹⁾. Overall, these findings provide evidence indicating that dietary n-3 PUFA can profoundly alter the biochemical make up of cell membrane lipid rafts/caveolae microdomains, which may directly or indirectly influence membrane fusion and cell–cell signalling. Interestingly, only a single study to date has examined the effects of CLA with regard to lipid raft/caveolae composition. Huot & Ma⁽Reference Huot and Ma⁵⁰⁾ demonstrated that mixed CLA isomers are concentrated in caveolae phospholipids, resulting in the reduction of caveolae-resident proteins, caveolin-1 and Her-2/neu, in Michigan Cancer Foundation (MCF)-7 breast cancer cells. Unfortunately, few studies to date have assessed the physical properties of CLA isomers⁽Reference Yin, Kramer, Yurawecz, Eynard, Mossoba and Yu⁵¹⁾. Therefore, future studies using purified CLA isomers are needed in order to elucidate how conjugated fatty acid structure affects membrane structure and function.

Docosahexaenoic acid alters the size and distribution of lipid rafts

In a proof of principle study, we sought to determine the effect of DHA on the size and distribution of lipid rafts in vivo ⁽Reference Chapkin, Wang, Fan, Lupton and Prior¹⁷⁾. Using immunogold electron microscopy of plasma membrane sheets coupled with spatial point analysis, morphologically featureless microdomains were visualised in HeLa cells. Cells were transfected with green fluorescent protein truncated H-ras (GFP-tH), which is located exclusively to inner leaflet rafts, and subsequently incubated with DHA and control fatty acid, for example, oleic acid (18 : 1n-9) for 48 h. Univariate K-function analysis of GFP-tH (5 nm gold) revealed that the interparticle distance was significantly reduced by DHA treatment compared with control fatty acid, indicating that select PUFA can increase clustering of proteins in cholesterol-dependent microdomains (GFP-tH), whereas non-raft microdomains are insensitive to DHA modulation. These novel findings suggest that the plasma membrane organisation of inner leaflets is fundamentally altered by DHA enrichment (Fig. 2).

‘Cholesterol-centric’ view of membranes

SFA compared with PUFA have a preferential affinity for cholesterol. This relationship provides the basis for a lipid-driven mechanism for the lateral segregation of membrane elements into cholesterol-rich and -poor microdomains⁽Reference Stillwell and Wassall^35, Reference Huster, Arnold and Gawrisch^52–Reference Niu and Litman⁵⁵⁾. For example, unfavourable interaction between cholesterol and PUFA chains has been clearly demonstrated by the exclusion of cholesterol from dipolyunsaturated phosphatidylcholine membranes where it is forced to directly contact polyunsaturated chains. Studies using a variety of techniques including differential scanning calorimetry⁽Reference Zerouga, Jenski and Stillwell⁵⁶⁾, ¹H NMR and nuclear Overhauser enhancement spectroscopy with magic angle spinning⁽Reference Huster, Arnold and Gawrisch⁵²⁾, determination of partition coefficients⁽Reference Kariel, Davidson and Keough⁵⁷⁾, measurements of lateral compressibility⁽Reference Needham and Nunn⁵⁸⁾ and fluorescence anisotropy⁽Reference Mitchell and Litman^53, Reference Mitchell and Litman⁵⁹⁾ indicate that the poor affinity of DHA and perhaps CLA for cholesterol provides a lipid-driven mechanism for lateral phase separation of cholesterol-rich lipid microdomains from the surrounding bulk membrane. This could in principle alter the size, stability and distribution of cell surface lipid microdomains such as rafts. Indeed, growing evidence from model membrane studies suggest that the energetically less favourable interaction between cholesterol and PUFA, especially DHA, promotes lateral phase segregation into sterol-poor/PUFA-rich and sterol-rich/SFA-rich microdomains⁽Reference Huster, Arnold and Gawrisch^52, Reference Mitchell and Litman^53, Reference Kariel, Davidson and Keough^57, Reference Brzustowicz, Cherezov, Caffrey, Stillwell and Wassall^60–Reference Pasenkiewicz-Gierula, Subczynski and Kusumi⁶²⁾.

Huang & Feigenson proposed the umbrella model to describe the solubility and condensing effect of cholesterol within membranes⁽Reference Huang and Feigenson⁶³⁾. According to this model, phospholipid head groups act as ‘umbrellas’ to prevent the energetically unfavourable contact of the non-polar part of cholesterol with interfacial water. This shielding will be less effective for DHA-containing phospholipids with a large molecular cross-sectional area, facilitating cholesterol precipitation at a lower concentration. In addition, this model allows for speculation that phosphatidylethanolamine with a smaller head group may enhance the DHA-associated reduction in shielding effects relative to phosphatidylcholine. Consistent with this notion, unlike phosphatidylcholine bilayers where a marked reduction in cholesterol solubility requires polyunsaturation at both sn-1 and sn-2 positions, DHA at the sn-2 position with a saturated sn-1 chain is sufficient in phosphatidylethanolamine to trigger cholesterol precipitation⁽Reference Shaikh, Cherezov, Caffrey, Stillwell and Wassall⁶⁴⁾.

Conjugated n-3 polyunsaturated fatty acids

There is growing evidence that the combination of conjugated double bonds and n-3 PUFA may have enhanced chemoprotective properties. Recent studies using a number of model systems suggest that conjugated EPA and conjugated DHA suppress tumour growth⁽Reference Danbara, Yuri, Tsujita-Kyutoku, Sato, Senzaki, Takada, Hada, Miyazawa, Okazaki and Tsubura^65, Reference Kimura and Sumiyoshi⁶⁶⁾, suppress topoisomerases⁽Reference Yonezawa, Tsuzuki and Eitsuka⁶⁷⁾, induce apoptosis⁽Reference Tsuzuki, Tanaka, Kuwahara and Miyazawa^68, Reference Tsuzuki, Kambe, Shibata, Kawakami, Nakagawa and Miyazawa⁶⁹⁾, inhibit lipid accumulation⁽Reference Tsuzuki, Kawakami, Nakagawa and Miyazawa⁷⁰⁾ and have potential use as therapeutic dietary supplements for minimising tumour angiogenesis⁽Reference Tsuzuki, Tanaka, Kuwahara and Miyazawa^68, Reference Tsuzuki, Shibata, Kawakami, Nakagaya and Miyazawa⁷¹⁾. Typically, conjugated EPA and conjugated DHA are generated by alkaline isomerisation, producing a mixture of isomers with conjugated double bonds, although small amounts are found in marine algae and seal oil. Since the exact structure of these novel fatty acid species has not been fully characterised, future studies are required to verify their safety and efficacy in humans. In addition, it remains to be determined whether conjugated EPA and conjugated DHA alter the size and distribution of cell surface microdomains.