Main

Challenge of host tissues by microorganisms, tissue injury or surgical trauma leads to the release of exogenous and endogenous chemical mediators that in turn give rise to the initial cardinal signs of inflammation, rubor (redness), calor (heat), tumor (swelling) and dolor (pain), described by Celsus in the first century1 (Fig. 1). The exogenous mediators include microbial peptides that act as chemoattractants to recruit neutrophils to the site of challenge, where they phagocytose invading microorganisms and cellular debris. Within the neutrophils, newly formed phagosomes mature to form phagolysosomes by fusing with lysosomal granules, which contain degradative enzymes and produce reactive oxygen species (ROS) to kill trapped microorganisms or degrade cellular debris. So, the initial inflammatory response functions to protect the host and, ideally, its timely resolution ensures that it is self-limiting.

Figure 1: Decision paths in acute inflammation: resolution or chronic inflammation and the roles of endogenous chemical mediators.
figure 1

a | Microbial invasion of the host, injury from outside or the loss of barrier function initiates the release of exogenous chemoattractants (such as microbial peptides and endotoxins) and endogenous chemical mediators that can activate leukocyte recruitment and antimicrobial activities. Microorganisms taken up by phagocytosis activate bacterial killing mechanisms in, for example, neutrophil phagolysosomal vacuoles. Sometimes, during incomplete phagocytosis, neutrophil phagolysosomal vacuole contents (such as reactive oxygen species (ROS) and hydrolytic enzymes) can inadvertently spill into the extracellular milieu, and this leads to local tissue damage and/or unwanted amplification of pro-inflammatory responses. Surgical trauma and ischaemia–reperfusion injury also activate the release of endogenous chemical mediators, such as leukotriene B4, prostaglandin E2 and C5a. These mediators and others are involved in the development of the cardinal signs of acute inflammation, which involve vasoconstriction, vasodilation, vascular permeability and chemotaxis. The outcome of acute inflammation — resolution, chronicity and fibrosis — may be influenced by many factors, such as the type or site of injury, and the host response. The reestablishment of normal homeostasis (resolution) is an actively regulated programme. Prostaglandin E2 and prostaglandin D2, for example, stimulate the switching of arachidonic-acid-derived lipids from leukotriene B4 production to lipoxin A4 production and then the switching of lipid mediator families to produce anti-inflammatory and pro-resolution lipid mediators, such as E-series and D-series resolvins and protectins. Alternatively, chronic inflammation can result from excessive and/or unresolved inflammatory responses and can lead to chronic disorders. Arachidonic-acid-derived lipid mediators such as pro-inflammatory prostaglandins and leukotrienes can amplify this process. Fibrosis can occur when inflammatory injury causes substantial tissue destruction, connective tissue replacement occurs and results in loss of tissue function. b | Both exogenous and endogenous chemoattractant gradients stimulate the recruitment of neutrophils via diapedesis from postcapillary venules, an event that is amplified by the production of local leukotriene B4. During the tissue progression of inflammatory events, intravascular platelet–leukocyte interactions induce the formation of lipoxin A4 and lipoxin B4, which stop further recruitment of neutrophils and stimulate non-phlogistic monocyte infiltration. c | The chemical structures of pro-resolution mediators lipoxin A4, resolvin E1, resolvin D1 and protectin D1 are shown.

Full size image

Sometimes, however, neutrophil granule contents can inadvertently spill into the extracellular milieu before complete engulfment of microorganisms or debris by the neutrophil2. This leads to local tissue damage and amplification of acute inflammatory signals within minutes to hours of the challenge (Fig. 1). This unintentional spilling of granule contents, in particular the release of hydrolytic enzymes from phagolysosomes, also occurs when phagocytes encounter foreign surfaces that they fail to ingest, such as crystals, bacterial biofilms or other slimy surfaces1.

The endogenous chemical mediators that are intentionally released by cells that infiltrate the site of challenge include the eicosanoids (such as prostaglandin E2 and leukotriene B4) and complement components (such as C5a), which are important for host defence, but which can also lead inadvertently to tissue damage. Endogenous chemical mediators are released during sterile injury, such as in ischaemia–reperfusion injury3. This route of chemical mediator release can be amplified by overt activation and excessive recruitment of neutrophils to the site of injury that sustain the inflammatory response by further spilling of noxious granule contents and by the generation of other chemical mediators (Fig. 1).

Acute inflammation has several programmed fates including progression to chronic tissue fibrosis and the ideal outcome of complete resolution1. Challenges are met by infiltrating phagocytic cells of the innate immune system, primarily neutrophils, which traffic by sensing gradients of chemoattractants from postcapillary venules to the tissue via diapedesis (Fig. 1). These chemoattractants consist of endogenous lipid mediators, such as leukotrienes, and protein mediators, including chemokines and cytokines4, as well as exogenous chemoattractants released, for example, by microorganisms. Time-dependent progression of cell infiltration is led by specific leukocyte subtypes, namely, professional phagocytes, with neutrophils being the first to enter, followed by monocytes5.

Once the initiating noxious materials are removed via phagocytosis, the inflammatory reaction must be resolved to prevent the inflammation from spreading, becoming chronic or causing disease. Resolution of inflammation, or its catabasis6, is the reduction or removal of leukocytes and debris from inflamed sites, enabling the return to homeostasis. The resolution of inflammatory leukocytic infiltrates was previously considered to be a passive process7. Local chemotactic stimuli and gradients, for example, were thought to dissipate or simply 'burn out' with time, enabling tissues to drain, repair and return to normal function1. As we review here, recent findings indicate that resolution is not merely a passive termination of inflammation, but rather an active biochemical and metabolic process6,7. The resolution process is rapidly initiated after acute challenges by cellular pathways that actively biosynthesize local, specialized, dual-acting anti-inflammatory and pro-resolution lipid mediators, such as the lipoxins, resolvins and protectins8,9 (Fig. 1).

In this Review, we provide an update and overview of newly identified lipid mediators that have pivotal roles in the resolution of inflammation and in disease models. Unexpectedly, two of these recently described lipid mediator families, the resolvins and protectins, are biosynthesized from precursor essential omega-3 polyunsaturated fatty acids (PUFAs), eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)9,10,11 (Box 1).

Resolution of inflammation: new concepts

The resolution of inflammation was defined in modern molecular terms relatively recently6,8,12,13. Importantly, it is now considered to be a distinct process from anti-inflammatory processes5. This is because, in addition to serving as agonists to stop and reduce neutrophil infiltration to inflamed tissues, pro-resolution molecules promote the uptake and clearance of apoptotic cells and microorganisms by macrophages in inflamed sites14,15 (Fig. 2), and stimulate the antimicrobial activities of mucosal epithelial cells16,17. The resolution of inflammation is accompanied by an active switch in the mediators that predominate in exudates. Initially, mediators, such as classic prostaglandins and leukotrienes, which activate and amplify the cardinal signs of inflammation, are generated. Next, prostaglandin E2 and prostaglandin D2, by inducing the production of key enzymes, gradually promote the synthesis of mediators that have both anti-inflammatory and pro-resolution activities, such as the lipoxins8,14, resolvins and protectins9,10,11 (Fig. 1). These families of endogenous pro-resolution molecules are not immunosuppressive, but instead function in the resolution of inflammation by activating specific mechanisms to promote homeostasis. In general, pro-resolution molecules stimulate and accelerate resolution via mechanisms at the tissue level that are multi-factorial (Fig. 2). Specific lipoxins and members of the resolvin and protectin families provide potent signals that selectively stop neutrophil and eosinophil infiltration; stimulate non-phlogistic recruitment of monocytes (that is, without elaborating pro-inflammatory mediators); activate macrophage phagocytosis of microorganisms and apoptotic cells; increase the exit of phagocytes from the inflamed site through the lymphatics; and stimulate the expression of molecules involved in antimicrobial defence5,10,16,17.

Figure 2: Dual anti-inflammatory and pro-resolution actions of specific lipoxins, resolvins and protectins.
figure 2

The key histological feature in the resolution of inflammation is the loss of neutrophils from the local inflamed sites. This is a programmed process that is actively regulated at multiple levels: by reducing neutrophil infiltration into the exudate, increasing monocyte recruitment to the exudate, stimulating macrophage uptake of apoptotic neutrophils, and promoting phagocyte exit from the exudate via the lymphatics.

Full size image

Lipid mediators, such as prostaglandins and leukotrienes, are widely appreciated for their pro-inflammatory activities18,19. Using either cyclooxygenases or lipoxygenases, leukocytes rapidly biosynthesize these lipid mediators from membrane-derived arachidonic acid within seconds to minutes of acute challenge. Recent results, however, indicate that as inflammation proceeds the neutrophils in confined exudates stop producing chemoattractants (such as leukotriene B4) and within hours begin to convert arachidonic acid into protective lipoxins, thereby serving as agonists to actively terminate inflammation and promote resolution7,8,9. In contained exudates in certain inflamed tissues and inflammatory diseases1,20,21, both prostaglandin E2 and prostaglandin D2 have pro-inflammatory activities18, but each can promote a switch in the expression of biosynthetic enzymes by infiltrating neutrophils that changes their phenotype to a pro-resolution phenotype — a process known as lipid-mediator class switching8. It is worth noting that, in certain settings in which prostaglandin E2 is delivered pharmacologically, it can have anti-inflammatory activities, through the stimulation of cyclic AMP22. Therefore, in these settings, prostaglandin E2 is not acting as a pro-resolution mediator by definition23 because it does not stimulate the non-phlogistic responses of macrophages, namely, the uptake and clearance of apoptotic cells. In addition, in the presence of leukotriene B4 or C5a, prostaglandin E2 can enhance inflammation24. Prostaglandin D2 can also increase intracellular cyclic AMP levels in certain cell types and have anti-inflammatory actions25, whereas its non-enzymatic degradation products (that is, 15-deoxy-delta-12,14-prostaglandin J2 and related cyclopentaenones) can enhance resolution by promoting leukocyte apoptosis and macrophage clearance from inflamed sites26, possibly by inhibiting nuclear factor-κB (NF-κB) activation27.

Lipoxins: anti-inflammatory and pro-resolving

Production of lipoxins. Lipoxins were the first mediators recognized to have dual anti-inflammatory and pro-resolution activities14,28. Lipoxins (such as lipoxin A4 and lipoxin B4) are unique structures derived from arachidonic acid that have potent actions in vivo and in vitro. (For in-depth reviews on the biosynthesis, total organic synthesis and actions of lipoxins, aspirin-triggered lipoxins and their stable analogues, see Ref. 14.) Lipoxins are biosynthesized by the sequential actions of lipoxygenase(s) and other enzymes to produce bioactive trihydroxytetraenes, structures that are found in all eicosanoids of this class. In humans, initial oxygenation of arachidonic acid via 15-lipoxygenase type I and then by 5-lipoxygenase is one route of lipoxin biosynthesis that has been observed in mucosal tissues, such as the respiratory tract, gastrointestinal tract and oral cavity, and that results from the interactions between epithelial cells and leukocytes14. This activity is enhanced during inflammation but is also likely to provide a mechanism for restoring homeostasis, because mucosal surfaces are continuously exposed to microorganisms in vivo29,30,31. In the mucosa, lipoxins are generated by neutrophils from the 15-hydroxyeicosatetraenoic acid (15-HETE) precursor, which is provided by mucosal epithelial cells29. The blood vessels represent another main site where lipoxin biosynthesis occurs in humans; biosynthesis involves the initiation of arachidonic-acid oxygenation by 5-lipoxygenase in leukocytes and the release of the intermediate leukotriene A4, which is converted by the lipoxin-synthase activity of 12-lipoxygenase in platelets. This biosynthesis pathway is exemplified by platelet–leukocyte interactions in blood vessels or possibly in exudates that form a nidus for transcellular biosynthesis.

In inflamed sites, neutrophils can interact with other cells in their immediate vicinity, such as with other leukocytes, platelets, endothelial cells, mucosal epithelial cells and fibroblasts, and acquire the ability to produce lipoxins. More than 50% of the leukocyte-derived epoxide intermediate leukotriene A4 is released by cells for processing by platelet 12-lipoxygenase or mucosal 15-lipoxygenase to produce lipoxins32,33,34. Lipoxins generated by cell–cell interactions and transcellular biosynthesis stop neutrophil diapedesis and recruitment into the tissues29,30,31 (Fig. 1).

It is now clear that neutrophils change their phenotype to produce different profiles of lipid mediators depending on the cells and substrates present in their local environment8,9. For example, neutrophils in resolving inflammatory exudates switch from the production of leukotrienes to that of lipoxins and resolvins, whereas neutrophils in the peripheral blood, on activation, generate and release leukotriene B4 as one of their main bioactive products8. In this context, local prostaglandin E2 and prostaglandin D2 stimulate the processing of 15-lipoxygenase mRNA in leukocytes to produce a functional enzyme for lipoxin production8. Other cell types can acquire the ability to generate lipoxins when exposed to specific cytokines or growth factors32, or in the case of macrophages, for example, when they engulf apoptotic leukocytes35. These findings are of relevance to pro-resolution mechanisms because lipoxin A4 generated by macrophages probably contributes to the stimulation of their phagocytic activity15,17 without elaborating pro-inflammatory mediators — namely, the non-phlogistic process.

Pathogens can also contribute to the provision of the necessary components for lipoxin biosynthesis. Pseudomonas aeruginosa encodes the first identified secretory lipoxygenase that converts host arachidonic acid to 15-HETE for local lipoxin production36. High levels of lipoxins, greater than those considered to be physiologically standard, are produced by host cells infected by Toxoplasma gondii, which encode their own 15-lipoxygenase37,38. Therefore, 15-lipoxygenase expressed by pathogens may interact with endogenous biosynthetic circuits of the host to generate local 'stop signals' at levels that can divert the host immune defence.

Pro-resolution actions of lipoxins. Lipoxin A4 and lipoxin B4 inhibit neutrophil entry into inflamed sites and counter-regulate the main aspects of inflammation. They act on many cell types including blood cells, neural cells and stromal cells39,40,41,42 (Table 1). Lipoxin A4 regulates leukocyte responses in vitro and trafficking in vivo by activating its specific receptor, lipoxin A4 receptor (ALX) (Fig. 3). ALX is a G-protein-coupled receptor (GPCR) that is expressed by leukocytes and has cell-type-specific signalling pathways40. For example, in neutrophils, lipoxin-A4–ALX interactions stop neutrophil migration, whereas in monocytes lipoxin-A4–ALX interactions stimulate monocyte chemotaxis and non-phlogistic responses41. Unlike classic GPCRs for chemoattractants that mobilize intracellular Ca2+ to evoke chemotaxis, lipoxins instead induce changes in the phosphorylation of proteins of the cytoskeleton, resulting in cell arrest42,43. In addition to these effects on the resolution of inflammation, lipoxin A4 reduces organ fibrosis, acts directly on both vascular and smooth muscle (for reviews, see Refs 14, 40 ) and has direct action in reducing pain44.

Table 1 Key cellular actions of lipoxins, resolvins and protectins in innate immunity
Full size table
Figure 3: Mechanisms of action of lipoxin A4 and resolvin E1: regulation at multiple levels via G-protein-coupled receptors (GPCRs).
figure 3

There are two main 'classic' lipoxygenase-mediated pathways of lipoxin generation that are used in human cells and tissues (the 5-lipoxygenase–12-lipoxygenase pathway in neutrophils and platelets in the vasculature, and the 15-lipoxygenase–5-lipoxygenase pathway in epithelial cells and neutrophils at mucosal surfaces). The overall action of lipoxin A4 in vivo is likely to be attributed to its interactions with GPCRs and growth-factor receptors (GFRs). Direct activation of the lipoxin A4 receptor (ALX) by lipoxin A4 results in cell-type-specific signalling events that stop neutrophil migration and stimulate monocyte and macrophage activation. Indirect inhibition, via receptor crosstalk, of other GPCRs (such as leukotriene B4 receptor (BLT1)) and growth-factor receptors (such as VEGFR (vascular endothelial growth-factor receptor) expressed by endothelial cells and PDGFR (platelet-derived growth-factor receptor) and CTGFR (connective tissue growth-factor receptor) expressed by mesangial cells) by lipoxin A4 reduces angiogenesis and mesangial-cell proliferation and fibrosis. For the generation of resolvin E1, aspirin acetylates cyclooxygenase-2 (COX2) in vascular endothelial cells and generates 18R-hydroperoxyeicosapentaenoic acid (18R-HPEPE), which is further converted via 5-lipoxygenase in leukocytes and additional enzymatic reactions to form resolvin E1. Microbial cytochrome P450 enzymes can also contribute to resolvin E1 biosynthesis by converting eicosapentaenoic acid (EPA) to 18-HEPE. Resolvin E1 directly interacts with at least two GPCRs in a cell-type-specific manner. Resolvin E1 directly activates ChemR23 expressed by monocytes and dendritic cells and directly inhibits BLT1 that is expressed by human neutrophils. HUVEC, human umbilical vein endothelial cell; IL-12, interleukin-12; MAPK, mitogen-activated protein kinase; NF-κB, nuclear factor-κB.

Full size image

Aspirin impinges the endogenous lipoxin-generating system during cell–cell interactions. Inhibition of prostaglandin biosynthesis by aspirin is a well-appreciated mechanism in its anti-thrombotic and anti-inflammatory effect45. Of interest, aspirin triggers the endogenous formation of carbon-15 epimeric lipoxins, namely aspirin-triggered lipoxins (ATLs)46. Cells that express cyclooxygenase-2 (COX2), including vascular endothelial cells, epithelial cells, macrophages and neutrophils, are involved in ATL production. Acetylation of COX2 by aspirin blocks its ability to biosynthesize prostaglandins but does not impair its ability to produce ATL precursors. Accordingly, the production of ATLs is increased in aspirin-treated animals compared with control animals, and importantly in healthy humans taking low doses of aspirin47.

ATLs induce the expression of haem oxygenase 1 (HO1) by endothelial cells48, a key system in endogenous anti-inflammation and tissue protection. Mice lacking 15-lipoxygenase type I have an impaired HO1 response. Topical lipoxin A4 in these mice restores HO1 expression and protects them from inflammatory challenge49. When triggered by aspirin, lipoxins and 15-epi-lipoxins stimulate nitric oxide production by endothelial cells50, which blocks leukocyte adhesion to vascular endothelial cells. These mechanisms for lipoxins and 15-epi-lipoxins may be particularly relevant in the local resolution of inflammation. Compelling findings in experimental animal systems that demonstrate both the anti-inflammatory and pro-resolution actions of lipoxins and ATLs in vivo (Table 2) have established that preventative treatment with a pro-resolution mediator reduces inflammation and disease status15,51. Therefore, endogenous ATLs (15-epi-lipoxins) offer new and potentially important mechanisms that underlie the clinical benefits of aspirin47,48,49,50. 15-epi-lipoxin production may also be an essential mechanism of other widely used drugs, such as statins, which stimulate the production of endogenous 15-epi-lipoxin A4 (Ref. 52).

Table 2 Lipoxins, resolvins and protectins in complex disease models
Full size table

Of interest, both lipoxins and ATLs directly act on human T cells53 to block secretion of tumour-necrosis factor (TNF). Human peripheral T cells express ALX, and activation of this receptor blocks CD3-specific-antibody-induced activation of extracellular signal-regulated kinase (ERK), which is required for TNF secretion. Lipoxin B4 blocks phosphorylation of ERK in T cells, as does one of its analogues, 5-(R/S)-methyl lipoxin B4, which also inhibits TNF secretion by activating lipoxin B4 receptors. Lipoxins therefore provide a link between innate cells involved in resolution of inflammation and cells of the adaptive immune system54.

Resolvins and protectins

Previous studies have shown that essential omega-3 PUFAs administered at daily high doses (milligrams to grams) have beneficial actions in many inflammatory diseases, cancer and human health in general20,55,56. However, the molecular basis of omega-3 fatty acid action was not established until recently. To identify potential mechanisms that are actively involved in the resolution of inflammation, we devised a new lipid-mediator approach based on lipidomics9,10 and informatics57, with liquid chromatography-ultraviolet-tandem mass-spectrometry-based analyses, to map and profile the appearance and/or loss of mediators in resolving inflammatory exudates. When novel bioactive compounds were encountered, their structures were elucidated and their bioactivity and role(s) confirmed in vivo9,10,11,51,58. These studies uncovered two new families of bioactive mediators — termed resolvins and protectins — that are biosynthesized from omega-3 essential PUFAs.

Resolvins. The first resolvin was identified in exudates9 collected from inflamed murine dorsal air pouches in the spontaneous resolution phase59, and was so-named because it proved to be a potent regulator of resolution. Resolvins are derived from EPA and DHA with two chemically unique structural forms, the E-series and D-series, respectively. E-series member resolvin E1 reduces inflammation in vivo and blocks human neutrophil transendothelial migration9. Resolvin E1 can be produced in vitro, recapitulating events in vivo, by treating human vascular endothelial cells in a hypoxic environment with aspirin. These cells convert EPA to 18R-hydroperoxyeicosapentaenoic acid (18R-HPEPE) and release 18R-hydroxyeicosapentaenoic acid (18R-HEPE), which is rapidly transformed by activated human neutrophil 5-lipoxygenase9,60. Resolvin E1 is produced in healthy individuals and is increased in the plasma of individuals taking aspirin and/or EPA58.

Resolvin E2 is the second E-series member that reduces zymosan-initiated neutrophil infiltration, thereby displaying potent anti-inflammatory actions60. These EPA-derived E-series resolvins might contribute to the beneficial actions that have been attributed to omega-3 PUFAs in human diseases, such as skin inflammation, peritonitis9, periodontal disease23,61 and colitis62 (Table 2). Indeed, the production of 5-lipoxygenase in human leukocytes is pivotal for these beneficial effects and is controlled in part by temporal and spatial events in vivo to signal the production of either leukotrienes or anti-inflammatory mediators, such as lipoxins and resolvins. Of interest, microbial and mammalian cytochrome P450 enzymes convert EPA into 18-HEPE9, which can be transformed by human neutrophils to resolvin E1 and resolvin E2. Hence, it is likely that microorganisms at sites of inflammation or in the gastrointestinal tract can contribute to the production of E-series resolvins in humans.

Two receptors are known to be involved in the actions of resolvin E1 (Fig. 3). The GPCR chemokine-like receptor 1 (CMKLR1; also known as ChemR23) attenuates TNF-stimulated NF-κB activation in response to resolvin E1 binding, indicating a counter-regulatory action of this ligand–receptor pair58. Counter-regulation of TNF signalling was used to identify this GPCR because TNF is a key mediator in the early steps of acute inflammation9,58. Recently, a second GPCR that interacts with resolvin E1 was identified, the leukotriene B4 receptor (BLT1)63. Resolvin E1 interacts in a stereospecific manner with BLT1 expressed by human neutrophils as a receptor antagonist9, attenuating leukotriene-B4-dependent pro-inflammatory signals via BLT1 (Ref. 63). Accordingly, the anti-inflammatory actions of resolvin E1 are markedly reduced in BLT1-deficient mice. So, to counter-regulate inflammation and promote resolution, resolvin E1 selectively interacts with at least two GPCRs present on different cell types, namely CMKLR1 on monocytes and dendritic cells, and BLT1 on neutrophils.

DHA is the substrate for two groups of resolvins produced by different biosynthetic routes, referred to as the 17S and 17R D-series resolvins, during the resolution of inflammatory exudates10,11. D-series resolvins display potent anti-inflammatory actions10 and are particularly interesting because the brain, synapses and retina are highly enriched in DHA64,65,66. Endogenous DHA is converted in vivo via lipoxygenase-initiated mechanisms to the 17S-hydroxy-containing series of four resolvins, known as resolvin D1–resolvin D4 (Refs 10, 11). Each of these potent bioactive resolvins was first isolated in exudates from mice that had been given aspirin and DHA; this led to the identification of several new 17R-hydroxy-containing products in exudates during the resolution phase10. Their ability to stop neutrophil infiltration was used to assess their biological function for structural elucidation studies10 and their biosynthesis was reconstructed to establish their potential origins. Results from these studies showed that human recombinant COX2 converts DHA into a 13-hydro(peroxy)-containing product. In the presence of aspirin, the oxygen at carbon position 13 switches to position 17 with an R configuration that is a precursor for the aspirin-triggered 17R D-series resolvins. These are produced in exudates and in the brain in response to aspirin treatment10,11. Resolvin D1 and aspirin-triggered resolvin D1 are both potent regulators of human and mouse neutrophils10,67 (Table 1). In microglial cells, both the 17S and 17R D-series resolvins block the production of TNF-induced transcripts that encode the pro-inflammatory cytokine interleukin-1β (IL-1β), which is expressed rapidly in response to neural injury11,68. Resolvins control inflammation at many levels, by reducing peritonitis and skin inflammation10,11, protecting organs from reperfusion injury and neovascularization (Table 2). Therefore, the D-series resolvins are of interest in the control of inflammation resolution in host defence and in neural tissues.

Protectins. DHA in resolving exudates is converted to another molecule belonging to a new family of mediators named protectins. Protectins are distinguished by the presence of a conjugated triene double bond and by their potent bioactivity10,11. They are biosynthesized via a lipoxygenase-mediated pathway that converts DHA to a 17S-hydroperoxide-containing intermediate11, which is rapidly converted by human leukocytes into a 16(17)-epoxide that is enzymatically opened in these cells into a 10,17-dihydroxy-containing anti-inflammatory molecule10,11. This bioactive compound, initially referred to as 10,17-diHDHA or 10,17S-docosatriene11, is now known as protectin D1, owing to its potent protective activity in inflammatory51 and neural systems that has been documented in studies by Bazan and colleagues66,69,70. When produced by neural tissues, it is termed neuroprotectin D1 (Ref. 70); the prefix neuro is added to depict its biosynthetic origin51.

Several 10,17-dihydroxy-containing products are produced in vivo via different biosynthetic routes, the most potent of these products being protectin D1. The other natural protectin D1 isomers have different double-bond configurations and are less potent in dampening neutrophil recruitment and inflammation51. Protectin D1 is stereoselective and is log-orders of magnitude more potent in vivo than its precursor DHA. Protectin D1 is also produced by human peripheral blood mononuclear cells in T-helper-2-type conditions in a lipoxygenase-dependent manner via a 16(17)-epoxide intermediate. Protectin D1 blocks T-cell migration in vivo, reduces TNF and interferon-γ secretion and promotes T-cell apoptosis71.

Agonists of inflammation resolution. Specific resolvins, protectins and lipoxins stereoselectively stimulate resolution and reduce the magnitude of the inflammatory response in vivo (Tables 1,2). The clearance of apoptotic neutrophils by professional phagocytes, such as macrophages, is a cellular hallmark of inflammatory tissue resolution72, and this can be used to quantify resolution using specific indices6 (Fig. 4). Resolvin E1 initiates resolution of inflammation and causes a decrease in the number of neutrophils in exudates at earlier times than does spontaneous resolution. Protectin D1 shifts the onset of resolution to an earlier time point and in addition shortens the time taken to reduce the number of maximum neutrophils by half as indicated by the calculation of the resolution interval (Fig. 4). The pro-resolution actions of both resolvin E1 and protectin D1 are achieved by reducing neutrophil influx and stimulating macrophage ingestion of apoptotic neutrophils, as well as by enhancing the number of phagocytes present in the lymph nodes and spleen15. Disruption of biosynthesis of these pro-resolution mediators by either COX2 or lipoxygenase inhibitors gives rise to a 'resolution deficit' phenotype, which is characterized by impaired phagocytic removal, delayed resolution and prolonged inflammation. These findings emphasize a pivotal homeostatic function for lipoxygenase(s) and COX2 pathways in the timely resolution of acute inflammation. More importantly, pro-resolution mediators at lower doses than inhibitors of COX2 and lipoxygenase can rescue the deficit in resolution caused by these interventions15.

Figure 4: Resolution indices pinpoint the mechanism of action of anti-inflammatory and pro-resolution lipid mediators in tissues.
figure 4

The main events in resolution of acute inflammation in vitro can be quantified by introducing resolution indices: the magnitude (Ψmax, Tmax), that is the time point (Tmax) when neutrophil numbers reach maximum (Ψmax); the duration (R50, T50), that is the time point (T50) when the neutrophil numbers reduce to 50% of Ψmax (R50); the resolution interval (Ri), that is the time interval from the maximum neutrophil point (Ψmax) to the 50% reduction point (R50). The presence of aspirin-triggered lipoxin A4 analogue (ATLa) lowers the maximal neutrophil numbers (↓Ψmax); the presence of resolvin E1 and protectin D1, also initiate the resolution at an earlier time point (↓Tmax and T50); protectin D1 further shortens the resolution interval (↓Ri).

Full size image

Resolvin E1 and protectin D1 upregulate the expression of CC-chemokine receptor 5 (CCR5) by dying neutrophils73. CCR5 expression on late apoptotic neutrophils acts as a 'terminator' of chemokine signalling, by clearing the pro-inflammatory ligands of CCR5 (CCL3 and CCL5) from the inflammatory site. Resolvin E1, in keeping with a role in stimulating the clearance of inflammatory cells and the return to tissue homeostasis, selectively induces the expression of the anti-adhesion molecule CD55 on the apical surface of mucosal epithelial cells, and thereby promotes CD55-dependent clearance of neutrophils across mucosal surfaces17. So, these EPA- and DHA-derived mediators are potent resolution agonists that activate cell-type-specific programmes in, for example, neutrophils, macrophages and epithelial cells (Table 1) at multiple levels to accelerate resolution.

Resolution of inflammation in disease models

Uncontrolled inflammation is now appreciated in the pathogenesis of many diseases that were not previously considered classic inflammatory diseases. These include atherosclerosis, cancer, asthma and several neurological disorders, such as Alzheimer's disease and Parkinson's disease. Natural pro-resolution mechanisms involving lipoxins, resolvins and protectins were tested for their ability to promote resolution and control inflammation (Table 2). It is now clear that endogenous anti-inflammation alone is not an identical mechanism of action compared with that of mediators that possess dual anti-inflammatory and pro-resolution actions5. In this regard, lipoxins, resolvins and protectins have potent multi-level mechanisms of action in disease models and promote resolution in animal models of oral, lung, ocular, kidney, skin and gastrointestinal inflammation, as well as in ischaemia–reperfusion injury and angiogenesis (Table 2).

Inflammatory diseases. Periodontal diseases, such as gingivitis and periodontitis, are leukocyte-driven inflammatory diseases characterized by soft-tissue and osteoclast-mediated bone loss21. As a model of inflammatory diseases, periodontitis has several advantages in that many, if not all, of the tissues involved in inflammatory processes of other organ systems are affected, including the epithelium, connective tissue and bone. Indeed, there are many noteworthy similarities in the pathogenesis of periodontitis and arthritis. Results from rabbit models of periodontitis demonstrate an important role for resolution of inflammation in disease prevention. Overexpression of 15-lipoxygenase type I in transgenic rabbits increases the levels of endogenous lipoxin A4, protects against periodontitis, and reduces atherosclerosis74,75. In prevention studies using this model, topical treatment with resolvin E1 prevents >95% of alveolar bone destruction. Histological analysis of rabbits treated with resolvin E1 revealed few, if any, neutrophils in the tissue and little tissue damage. In addition, osteoclasts, which are cells responsible for bone resorption, were reduced in these rabbits61. In established disease, resolvin E1 prevents periodontitis tissue destruction; both soft tissue and bone that were lost during disease were regenerated23.

In humans, the differential actions of resolvin E1 were studied using neutrophils from patients with localized aggressive periodontitis (LAP) and healthy individuals. Resolvin E1 reduces neutrophil superoxide generation in response to TNF or the bacterial surrogate peptide N-formyl-methionyl-leucyl-phenylalanine. Neutrophils from both healthy subjects and LAP patients produced ∼80% less superoxide when treated with resolvin E1. In comparison, neutrophils from LAP patients do not exhibit inhibition of superoxide production following treatment with lipoxin A4, suggesting a molecular basis for excessive inflammation in these patients.

In murine dorsal air pouches, nanogram amounts of resolvin E1 (∼100 nM) reduce leukocyte infiltration by 50–70%; these levels are comparable to those achieved by administering microgram amounts of dexamethasone (∼30 μM) or milligrams of aspirin (∼6 mM)9,58. Similarly, in spontaneously resolving peritonitis induced by the yeast cell-wall component zymosan, both resolvin E1 and protectin D1 activate and accelerate resolution15. Resolvins and protectins reduce neutrophil infiltration and increase non-phlogistic recruitment of monocytes. Notably, resolvin E1, protectin D1 and an ATL analogue each display different kinetics and molecular profiles of action (Tables 1,2).

Organ-specific diseases. Lipoxins are produced in the human gut mucosa76, where they limit persistent inflammation. As in the oral cavity, this is important given the continuous exposure of this organ to commensal bacteria. Patients with ulcerative colitis exhibit low to absent levels of lipoxin A4 and have lower levels of mucosal 15-lipoxygenase76. Interestingly, in the trinitrobenzene sulphonic acid (TNBS)-induced mouse model of Crohn's disease, oral administration of a lipoxin A4 analogue has potent efficacy in promoting the resolution of colitis77. Resolvin E1 reduces colitis-associated mortality and protects animals from weight loss and shortening of the colonic mucosa62. Histological analysis of treated animals revealed fewer colonic ulcerations and reduced transmural infiltration of neutrophils, monocytes and lymphocytes. Serum levels of TNBS-specific IgG were also decreased by this treatment, suggesting diminished antigen presentation and antibody production. Resolvin-E1-treated mice with colitis have reduced production of the pro-inflammatory cytokines TNF and IL-12p40 and increased levels of inducible nitric oxide synthase and COX2, whereas the levels of interferon-γ, IL-4 and IL-10 remain essentially unchanged62. So, agonists of resolution of inflammation prevent immune-mediated tissue damage and restore tissue homeostasis.

In Alzheimer's disease, soluble amyloid precursor protein-α, which stimulates in vitro proliferation of neural embryonic stem cells, activates neuroprotectin D1 biosynthesis. The hippocampal cornu ammonis region 1, but not the thalamus or occipital lobes, has decreased levels of DHA and neuroprotectin D1 (Ref. 69). The hippocampus of individuals with Alzheimer's disease shows decreased expression of phospholipase A2 and 15-lipoxygenase — key enzymes in neuroprotectin D1 biosynthesis in this tissue — hence it produces less of this endogenous protective mediator than healthy tissues66,69. Neuroprotectin D1 reduces the expression of pro-inflammatory genes and upregulates the expression of anti-apoptotic genes, thereby favouring the survival of brain cells69,70.

Retinal pigment epithelium (RPE) of the eyes generates neuroprotectin D1 from endogenous DHA that protects them from oxidative-stress-induced apoptosis by limiting pro-inflammatory gene expression70. Photoreceptor-cell integrity depends on the RPE, and loss of integrity is characteristic of retinitis pigmentosa and age-related macular degeneration. RPE cells undergoing oxidative stress generate neuroprotectin D1, upregulate the expression of the anti-apoptotic proteins BCL-2 (B-cell lymphoma 2) and BCL-XL, and decrease the levels of pro-apoptotic proteins BAX (BCL-2-associated X protein) and BAD (BCL-2-antagonist of cell death). In addition, neuroprotectin D1 reduces leukocyte infiltration and pro-inflammatory gene expression in brain ischaemia–reperfusion injury66,69,70.

Independent of their neutrophil-directed actions, both lipoxin A4 and protectin D1 also influence wound healing. The mouse cornea generates lipoxin A4 and protectin D1 (Ref. 78). In corneal thermal injury, topical application of either lipoxin A4 or protectin D1 increases the rate of re-epithelialization by ∼75%. The removal of corneal epithelial cells initiates neutrophil infiltration and increases the levels of the pro-inflammatory chemokine CXCL1 (murine equivalent of human IL-8) produced by the corneal stroma. Local treatment of lipoxin A4 or protectin D1 decreased CXCL1 levels by ∼60% (Ref. 78).

Acute kidney injury is an inflammatory process in which repair and regeneration following acute inflammatory events can lead to interstitial fibrosis, scarring and chronic kidney failure with persistent leukocyte infiltration79. Ischaemia followed by reperfusion leads to endogenous mobilization and increased levels of DHA in the blood and the production of D-series resolvins and protectins80. When treated with resolvins before bilateral renal ischaemia, mouse kidneys were protected from injury; the serum levels of creatinine were lower in treated mice compared with control mice. Protectin D1 was also protective, and both protectin D1 and resolvins reduced the number of tissue neutrophils and limited the deposition of interstitial collagen, thereby protecting the mice against tissue fibrosis. Resolvin D1 when given after ischaemic kidney injury protected mice from acute renal failure; however, this was not the case with protectin D1. These results suggest that D-series resolvins and protectin D1 activate resolution circuits in acute kidney injury80.

Dietary incorporation of DHA is protective against mouse liver necro-inflammatory injury through increased local production of DHA-derived mediators81. Hepatocytes incubated with DHA or 17-HDHA show less hydrogen-peroxide-induced DNA damage and less cellular oxidative stress. Mice fed a DHA-enriched diet are protected from carbon-tetrachloride-induced necro-inflammatory hepatic damage81.

In the lungs, lipoxins generated in mouse models of asthma are potent regulators of airway inflammation and hyper-responsiveness82. Lipoxins and their stable analogues reduce pulmonary inflammation by decreasing neutrophil, eosinophil and lymphocyte recruitment and activation. They also block oedema formation and reduce the levels of pro-inflammatory mediators IL-5, IL-13, CCL11, prostanoids and cysteinyl leukotriene82. Consistent with this key counter-regulatory role, individuals with severe asthma have apparent defects in lipoxin biosynthesis83. Protectin D1 is present, albeit at reduced levels in asthmatic subjects and potentially reduces both airway inflammation and hyper-reactivity84. Therefore, it appears that pro-resolution lipid mediators have roles in both physiological and pathophysiological processes in specific tissues.

Concluding remarks and future directions

The recently identified resolvin and protectin families and the lipoxin class of eicosanoids constitute a new genus of pro-resolution mediators with dual actions. Their discovery opens new avenues for the development of treatment strategies for inflammatory diseases and the development of resolution-based pharmacology and lipidomics-based therapeutics.

As many current and widely used drugs were developed without knowledge of their impact in resolution processes, some agents, such as selective COX2 inhibitors and certain lipoxygenase inhibitors, have now been shown to impair the tissue programmes of resolution, thereby delaying the return to homeostasis5,15,85, whereas others, such as glucocorticoids86, aspirin87, cyclin-dependent kinase inhibitors88 and statins52, seem to work in concert with endogenous pro-resolution processes. COX inhibitors reduce the amplitude of and the cardinal signs of inflammation by inhibiting prostanoid biosynthesis. Therefore, combining pro-resolution molecules with lipoxygenase or COX pathway antagonists may be a useful strategy to restore resolution and control excessive inflammation. Pro-resolution mediators may also have therapeutic potential in settings in which sustained inflammation and impaired resolution are components of disease pathophysiology (Table 2). Indeed, susceptibility to chronic inflammatory disorders may arise from defects in the lipid-mediator receptors and/or their signalling pathways, or from defects in the de novo biosynthesis of pro-resolution molecules84,89.

Nevertheless, before resolution pathways can be harnessed for the treatment of human diseases, several important questions need to be addressed: will it be possible to treat organ-specific inflammatory disorders with different pro-resolution mediators and/or their analogue mimetics in a tissue-specific manner? Will dietary supplementation with intermediates and/or precursors be beneficial for the local production of pro-resolution mediators? Will knowledge of these new pathways lead to an increased appreciation of potential deleterious effects of excessive DHA and EPA intake that could lead to increases in auto-oxidation of these fatty acids that can damage cells and tissues? Finally, given that lipoxins, resolvin E1 and protectins act on T cells, dendritic cells and phagocytic cells, might nutrition represent a molecular link between the innate and adaptive immune systems?

The relationship between essential fatty acids in nutrition, dietary supplementation and the biosynthesis of resolvins and protectins is an area of active research. In this regard, our studies with Kang and colleagues of mice that overexpress fatty acid desaturase from Caenorhabditis elegans (fat-1-transgenic mice) help to address discrepancies in dietary studies that can arise from individual genetic and feeding variations. The fat-1-transgenic mice produce and store higher levels of EPA and DHA in their tissues than wild-type mice and as a result generate increased levels of resolvins and protectins. After challenge, fat-1-transgenic mice show reduced gastrointestinal inflammation90 and less tumour metastasis91. Whether these exciting findings extend to humans will be of interest in further studies.