Fairweather Friends: Evidence of Lipoxin Dysregulation in Neurodegeneration

Abstract

Lipoxins (LXs) are autacoids; specialized proresolving lipid mediators (SPMs) acting locally in a paracrine or autocrine fashion. They belong to a complex superfamily of dietary small polyunsaturated fatty acid (PUFA)- metabolites, which direct potent cellular responses to resolve inflammation and restore tissue homeostasis. Together, these SPM activities have been intensely studied in systemic inflammation and acute injury or infection, but less is known about LX signaling and activities in the central nervous system. LXs are derived from arachidonic acid (AA), an omega-6 PUFA. In addition to well established roles in systemic inflammation resolution, they have increasingly become implicated in regulating neuroinflammatory and neurodegenerative processes. In particular, chronic inflammation plays a central role in Alzheimer’s disease (AD) etiology, and dysregulated LX production and activities have been reported in a variety of AD rodent models and clinical tissue samples, yet with complex and sometimes conflicting results. In addition, we recently reported reduced LX production following retinal injury, and demonstrated an intriguing direct neuronal activity promoting survival and homeostasis in retinal and cortical neurons. Here we review and clarify this growing literature and suggest new research directions to further elaborate the role of lipoxins in neurodegeneration.

Graphical Abstract

Lipoxins (LXs) are lipid mediators derived from arachidonic acid (AA), an omega-6 PUFA, and are potent signals to resolve inflammation and restore tissue homeostasis. LXs are increasingly implicated in neuroinflammation and neurodegeneration through dysregulated lipoxygenase (LOX) activity and receptor signaling, as well as directly affecting neuronal survival. We synthesize and clarify this growing literature.

INTRODUCTION

Lipoxins (LXs) are autacoids; specialized proresolving lipid mediators (SPMs) acting locally in a paracrine or autocrine fashion and metabolized rapidly via dehydrogenization [1–3]. They belong to a complex superfamily of omega-6 and omega-3 polyunsaturated fatty acid (PUFA)-derived metabolites, which direct potent cellular responses, particularly to resolve inflammation and restore homeostasis[1, 4, 5]. Most studies on SPM functions have therefore focused on immune and endothelial cells. However, SPM activities have been increasingly implicated in the central nervous system, particularly with regards to chronic injury and neurodegenerative diseases. For example, SPMs derived from the omega-3 PUFA docosahexaenoic acid (DHA) have been a focus of extensive investigations for their roles in neuroinflammation [6–8]. Recent studies have also implicated the structurally distinct LXs, derived from arachidonic acid (AA), in regulating neuroinflammation and neurodegeneration. These studies suggest that LX production and function becomes dysregulated in chronic neurodegenerative diseases, yet with complex and sometimes conflicting results. Additionally, we have described an intriguing new context in which lipoxins can act directly on central nervous system (CNS) neurons in a protective and homeostatic capacity. Here we review and clarify this growing literature and suggest new directions to further elaborate the role of lipoxins in CNS injury and repair.

The LX precursor AA, is itself a dietary omega 6 conditionally-essential PUFA that can also be derived from abundant sources of linoleic acid [9]. Both AA and linoleic acid are commercially available through supplements, or are rich in foods such as vegetable oils, nuts, seeds and meats. Literature regarding nutritional influence specifically on production of LX remains scarce. However, changes in AA or linoleic acid dietary intake can alter formation of other proinflammatory metabolites, such as leukotrienes and prostaglandins [10–13]. We have also recently shown that dietary changes in omega 6 and omega 3 PUFA intake can alter enzymatic LX production in the contexts of parenteral nutrition[14] and ocular inflammation[15].

Two naturally occurring LX family members are enzymatically generated from AA; lipoxin A₄ (LXA₄) and lipoxin B₄ (LXB₄), which are produced in multiple tissues including the CNS [16, 17]. Their production involves release of AA from membrane phospholipids via cytosolic and secretory phospholipases A2. Free AA is oxidized by the coordinated and sequential interactions of 5- and 12/15-lipoxygenases (LOXs) to generate hydroperoxy-eicosatetraenoic acid intermediates (H(p)ETEs), that are further catalyzed to generate each of the LXs; LXA₄ (5S,6R,15S-trihydroxy-7,9,13-trans-11-cis-eicosatetraenoic acid), or LXB₄ (5S,14R,15S-trihydroxy-6,10,12-trans-8-cis-eicosatetraenoic acid) [18]. Currently, there are three canonical mechanisms of endogenous LX synthesis: 1) In a two-step reaction, 15-LOX generates the intermediate 15-H(p)ETE, which is a substrate for 5-LOX [9] to generate LX. 2) A similar two-step reaction starts with the oxygenation of AA via 5-LOX to generate the epoxide intermediate leukotriene A, which is converted by either 12-LOX or 15-LOX to LX [19]. 3) Interestingly, an aspirin-mediated mechanism was identified as a third route of LX formation. Aspirin acetylates cyclooxygenase-2 (COX-2) to inhibit formation of the prostaglandins from AA [20]. Acetylation only partially inhibits COX-2 activity and leads the generation of 15R-HETE, a positional isomer of the LOX generated 15S-HETE. 5-LOX converts the aspirin-triggered 15R-HETE to LX isomers, 15-epi-LXA₄ (ATLXA₄) and 15-epi-LXB₄ (ATLXB₄)[20] (Figure 1). These aspirin-triggered LX isomers are more resistant to enzymatic inactivation and have longer half-lives relative to LXs [21].

Fig 1. Overview of LX biosynthesis.

LXA₄ and LXB₄ are synthesized from arachidonic acid (AA) through sequential oxygenation by actions of 5-LOX and 12/15-LOX enzymes, or alternatively via COX-2 to form ATLs.

LXs were the first identified SPMs and drive the resolution of acute inflammation by inhibiting neutrophil infiltration and activation, recruiting nonphlogistic macrophages to drive efferocytosis, inhibiting and counter-regulating pro-inflammatory cytokines/chemokines and decreasing vascular permeability, all of which restore tissue homeostasis [5, 22–27]. LXA₄ primarily exerts its activities through binding to the G-protein coupled formyl peptide receptor 2 (ALX/FPR2) [28] and GPR32 in humans, and ALX/FPR2 in rodents. However, LXA₄ or ATLXA₄ have also been reported to bind to the cysteinyl leukotriene receptor 1 (CysLT1) as an antagonist or inverse agonist [23, 29], act as allosteric modulators of the cannabinoid receptor CB1 [30], and a partial agonist for the PPARγ nuclear receptor [31], which elicit pro-resolution or anti-inflammatory responses. Thus, LXA₄ mediates its actions via complex and cell-specific signal pathways that remain to be fully defined. For example, LXA₄ via ALX/FPR2 activates a downstream cytokine signaling suppressor (SOCS-2), which is required to drive migration and cytokine production of dendritic cells in response to infection [32]. In comparison, although LXB₄ shows potent bioactivities, it currently has no known receptor, making functional investigations challenging. Consequently, the vast majority of LX studies focus on LXA₄ and ALX/FPR2 signaling. Yet, the activities of LXB₄ are distinct from LXA₄, and include potent effects on macrophages and nonphlogistic monocyte activation [26, 27, 33]. Together, these LX effects have been intensely studied in systemic inflammation and acute injury or infection models, but less is known about LX signaling and activities in the central nervous system and their roles in neurodegeneration.

LIPOXIN DYSREGULATION IN NEURODEGENERATION AND NEUROINFLAMMATION

Neuroinflammation can be initiated by a variety of stressors, such as infection, toxins, metabolic stress, invasive injury and autoimmunity. In the CNS, glial cells, such as astrocytes, oligodendrocytes, Müller cells, and microglia, play critical homeostatic roles by maintaining extracellular ion and neurotransmitter balance, phagocytosing cell debris, and providing nutrients [34–38]. However, dysfunction or chronic activation of these cells can also trigger detrimental effects, resulting in neuroinflammatory cascades that induce or exacerbate pathological mechanisms leading to neurodegeneration [39–44]. In the chronically stressed CNS, glial dysregulation of the LX synthetic circuit has been increasingly implicated in promoting this neuroinflammatory cycle and, subsequently contributing to neurodegenerative mechanisms.

A majority of studies regarding lipid mediators in neurodegeneration focus on the context of Alzheimer’s disease (AD). A contemporary view of AD pathogenesis integrates chronic neuroinflammatory events with accumulations of extracellular amyloid β (Aβ) aggregates and neurofibrillary tangles (NFT) of hyperphosphorylated tau protein [43, 45–47]. The limited clinical success, to date, of direct plaque targeting strategies has led to increased interest in targeting the neuroinflammatory components of the disease and reduce chronic activation of microglia and astrocytes [43, 48]. Strategies to modulate these processes have suggested therapeutic benefits to this approach [49–52]. In this regard proresolution strategies may compliment other anti-inflammatory approaches.

Levels of several SPM are reduced in AD cortex and have been demonstrated to reduce parainflammatory parameters when restored in animal models [53]. In cerebrospinal fluid and post-mortem hippocampus samples from AD patients levels of LXA₄ were significantly lower compared to healthy control samples [45], and these levels correlated with the increased accumulation of tau in same brain tissues. Interestingly, these reduced levels also correlated with mini-mental state examination (MMSE) scores [45]. However, a follow up study showed no significant differences in LXA₄ or LXB₄ in post-mortem entorhinal cortex [51]. Reduced LXA₄ levels were also reported in an AD mouse model harboring a pathogenic double mutation in the Aβ precursor protein (APP) gene [54, 55]. Corresponding rescue with ATLXA₄ treatment reduced both Aβ and phosphorylated tau levels [54, 56]. In one study twice daily ATLXA₄ administration to AD mice also demonstrated reduced levels of TNF-α, interleukin 1B (IL-1B), interferon-γ and IL-6, which correlate with AD progression [56]. The study also showed that ATLXA₄ increased anti-inflammatory interleukin 10 (IL-10) and transforming growth factor-β (TGF- β), resulting in the recruitment of a non-classical, nonphlogistic microglia subtype, which reduced Aβ levels. Additionally, ATLXA₄ is reported to attenuate proinflammatory cytokine production from microglia through NADPH oxidase inhibition [57], and supplemented LXA₄ levels promote antioxidant thioredoxin production [58, 59]. Downstream of LXA₄ signaling, expression of the cytokine signaling inhibitor (SOCS-2) is increased [32], which is also observed in AD patients [60]. Mice deficient in SOCS-2 have elevated production of pro-inflammatory cytokines and increased mortality [32]. Therefore, LXA₄ treatment drives proresolving effects in a variety of AD-relevant contexts. (These findings and additional studies relevant to LX activity and production in AD are summarized in Table 1).

Table 1.

LX and LOX studies in AD degeneration

Study Design	Tissue/Study Type	Findings	Citations
Human Pathology
AD brain, CSF	Post-mortem, biopsy	LXA4 ↓ in hippocampus & CSF, 15-LOX ↑ Reduced levels correlate with cognition	Wang et al, 2015
AD brain	Post-mortem	No significant change in LXA4, LXB4 in ENT	Zhu et al, 2016
AD brain	Post-mortem	5LOX protein ↑	Firuzi et al, 2003
AD brain	Post-mortem	5LOX protein ↑	Ikonomovic et al, 2008
AD brain	Post-mortem	12HETE & 15HETE ↑	Pratico et al, 2004
AD brain	Post-mortem	15-LOX ↓	Lukiw et al, 2005
Animal Models
Tg2576 mice	In vivo	5-LOX mRNA, protein ↑	Firuzi et al, 2003
Tg2576 mice overexpressing 12/15LO	In vivo	12/15-LOX protein ↑ BACE1 mRNA & protein ↑ Memory retention ↓ Aβ40 & Aβ42 ↑ Aβ plaques ↑	Chu et al, 2012
Tg2576 mice overexpressing 12/15LO	In vivo	p-Tau cdk5, p35, & p25 ↑ PSD-95 & synaptophysin ↓	Giannopoulos et al, 2013
Tg2576 mice overexpressing 5LO	In vivo	P-tau ↑ PSD-95 & synaptophysin & MAP2 ↓ P35 & p25 ↑	Chu et al, 2013
Tg2576 × 12/15LO−/− mice	In vivo	Aβ40, Aβ42, Aβ plaques ↓ CTF-β, & BACE-1 ↓ Memory retention ↑	Yang et al, 2010
Tg2576 × 5LO−/− mice	In vivo	Aβ42 ↓, γ-secretase ↓	Firuzi et al, 2003
Tg2576 mice treated with ATLA4	In vivo	Cognitive impairment ↓ Aβ40 & Aβ42 ↓, Aβ42 plaques ↓	Medeiros et al, 2013
3xTg mice overexpressing 5-LO	In vivo	Aβ40 & Aβ42 ↑, p25 ↑ Fear response ↓	Chu et al, 2012
3xTg mice treated with ATLA4	In vivo	Aβ40 & Aβ42 ↓, Aβ size & plaques ↓ Cognition ↑	Dunn et al, 2015
3xTg x 5LO−/− mice	In vivo	Aβ40 & Aβ42 ↓, Aβ plaques ↓ cdk5, p35, p25 ↓ PSD-95 & synaptophysin ↑ Long-term potentiation ↑ Working memory & retention ↑	Giannopoulos et al. 2014
12/15LO−/− 3-NP stressed mice	In vivo	Striatal lesion number ↑ Lesion incidence rate ↑	He et al, 2017
5xFAD mice	In vivo	LXA4 levels ↓, Aβ40 plaques ↓ IL-1b & IL-10 ↓	Kantarci et al, 2017
Cellular Models
HEK293-APPC99 cells incubated with 5-HPETE	In vitro	γ-secretase ↑	Firuzi et al, 2003
APP x 5LO MEF cells treated with A23187	In vitro	Aβ42 & γ-secretase ↑	Firuzi et al, 2003
N2A-APPswe x 12/15LO cells	In vitro	p-Tau↑, cdk5, p35 & p25 ↑	Giannopoulos et al, 2013
N2A-APPswe cells incubated with 12/15-HETE	In vitro	CTF-β, BACE-1 & sAPPβ ↓ BACE-1 mRNA ↓	Yang et al, 2010
N2A-APPswe x 12/15LO overexpression	In vitro	BACE1 mRNA & protein ↑	Chu et al, 2012
N2A-APPswe x 5LO overexpression	In vitro	Tau & p-tau ↑ PSD-95, & synaptophysin ↓	Chu et al, 2012
SHSY5Y cells treated with LXA4 & staurosporine	In vitro	Cell viability ↑	Zhu et al, 2016
Cortical neurons & HT22 cells treated with LXA4 and LXB4	In vitro	Cell viability, neurites, mitochondria ↑	Livne-bar et al, 2017

While LXA₄ demonstrates a proresolving activity, its primary target receptor ALX/FPR2 binds also binds other ligands, such as N-Formylmethionine peptides and annexin 1, that can induce specific and sometimes opposing bioactions [61]. Interestingly, Aβ can also bind to ALX/FPR2 with antagonistic effects to LXA₄. Le and colleagues reported Aβ_1–42 inducing chemotactic activity in human leucocytes through ALX/FRP2 activation [62]. Other studies demonstrated similar chemotactic activity, as well as superoxide production in mouse neutrophils and stimulation of murine microglial cells in vitro [63, 64]. Furthermore, expression of the ALX/FPR2 receptor is elevated in microglia and astrocytes of AD hippocampal samples [45]. This counterintuitive result was attributed to increased receptor availability to transduce resolution signals. However, the underlying mechanism remains unclear. The increased ALX/FPR2 receptor relative to reduced LXA₄ in the AD brain could indicate a compensatory mechanism to upregulate homeostatic and protective LX signaling.

LOX ENZYME DYSREGULATION IN AD

AD is also associated with dysregulated levels of the two key LOX enzymes and intermediates involved in LX production. Post-mortem analysis of AD patients revealed upregulation of 12/15 LOX and increased levels of the LX precursors 12-HETE and 15-HETE in the frontal and temporal lobes [65]. These findings are corroborated by another study which showed increase of the same precursors in cerebrospinal fluid of AD patients, as well as at-risk patients with mild cognitive impairment [66], suggesting 12/15 LOX may be involved in the initiation of AD [67]. However, another study has also demonstrated strongly reduced expression and activity of 15-LOX in the hippocampus of AD patients compared to the hippocampus of matched postmortem controls[68].

In genetic models, 12/15-LOX deficient mice show reduced production and deposition of Aβ [50], while overexpression of the 12/15-LOX gene (12/15-LO) in Tg2576 APP mutant mice activated astrocytes and microglial cells [69], and resulted in increased tau phosphorylation[70]. It is important to note that the mouse 12/15-LOX is the closest homolog to human 15-LOX, however, unlike the human enzyme the mouse 12/15-LOX primarily functions as a 12-LOX enzyme generating 12-HETE as its primary metabolite. Hence, data from mice need to be interpreted with caution. In addition, a body of work has demonstrated that 15-LOX metabolites in mouse, rat and rabbit models of neuro-inflammation and degeneration are protective in the retina and brain[71]. AD is similarly associated with increases in 5-LOX. For example, 5-LOX protein [72, 73] and mRNA levels [73] were increased in the hippocampus and frontal cortex of human AD patients and in pathogenic APP transgenic mice. Injections of inflammatory lipopolysaccharide (LPS), which have been proposed to play a role in AD pathogenesis [74], also induces expression of the 5-LOX gene (5-LO) in mouse hippocampus [75]. Increased 5-LOX has also been associated with activation of astrocytes and microglia [76], as well as increased Aβ levels through a γ-secretase dependent mechanism [72, 77], whereas absence of 5-LOX resulted in reduced NFTs and Aβ through γ-secretase inhibition [72, 76], and improved memory and synaptic integrity [78]. However, the role of 5-LOX is not clear as the enzyme can generate four leukotrienes that amplify or initiate inflammation, but is also required for the formation of a large number of protective and anti-inflammatory SPM. In a mouse model of allergic encephalomyelitis both 5-LOX and 12/15-LOX KO mice demonstrated significantly worse clinical scores compared to wild type controls [79]. In addition, relevant leukotriene formation in the brain or retina has not been clearly established and clinically the 5-LOX inhibitor zileuton has shown limited or no efficacy in human inflammatory diseases. Hence, further research is required to elucidate the role and function of 5-LOX in the brain.

Given the positive effects of LXA₄ and ATLXA4 on neuroinflammation, it may appear paradoxical that reduction in LOXs, which catalyze LX production, would also reduce Aβ, phosphorylated tau, and reactive glia. However, LOX enzymes can use both omega-3 and omega-6 PUFA as substrates, including AA, DHA and eicosapentaenoic acid (EPA), and LOX intermediates. Therefore, LXs represent only one of several distinct classes of LOX-derived lipid mediators. For example, proinflammatory leukotrienes and HETES are eicosanoids that can be generated by leukocytes. Ultimately, LOX enzymes have cell and tissue specific expression and generate a large number of structurally distinct lipid mediators with diverse bioactions. Therefore, additional research is required to better understand the contributions of LOX products to this neuroinflammatory milieu.

THE LX CIRCUIT IN OTHER NEURODEGENERATIVE CONTEXTS

In addition to AD, levels of LX and other PUFA derived lipid mediators have recently been reported in a growing literature of other neurodegenerative contexts. In a study of Multiple Sclerosis patients lipid mediator levels were compared in serum and cerebrospinal fluid between highly active and less active MS patients [80]. While levels of the LXA₄ precursor, 15-HETE, was increased in the active MS group, LXA₄ itself was not significantly elevated in patients with active disease compared to inactive disease. In a rodent stroke model, studies have shown ALX/FPR2 mediated LXA₄ effects on infarct size and cognition, suggesting a potential dysregulation of the LX pathway [81]. In clinical results the 12/15LOX enzyme is increased in both ischemic stroke patients and in animal models [82]. LX dysregulation may also affect neuronal health as a consequence of systemic inflammatory conditions that may impact neural tissues, such as diabetes and metabolic syndrome [83–85]. However, it remains unclear how LX actions may impact damage to these attendant neural tissues.

LIPOXINS CAN MEDIATE DIRECT NEURONAL ACTIONS

Studies assessing LXs in the CNS have generally focused on their proresolution roles. However, recent findings from our laboratory and others have also intriguingly uncovered a direct neuronal activity [51, 86, 87]. Previously, neuroprotective effects have been described for the structurally distinct DHA product NPD1 [88–90]. But, such activity had not been identified for LXs. Zhu and colleagues observed LXA₄ protective effects on immortalized SH-SY5Y neuroblastoma cells [51]. Our studies focused on astrocyte-neuron interactions in retinal injury models, relevant to the neurodegeneration of retinal ganglion cells (RGCs) in the chronic blinding disease glaucoma.

As for the brain, LX and LOX activities in the eye have been primarily studied in the context of ocular inflammation, wound repair and angiogenesis [91–93]. However, we reported that LXA₄ and LXB₄ are produced by retinal astrocytes and are regulated under stress. Both LXs are detected in mouse retina and optic nerve, but their levels and 5-LOX expression decrease following injury. Intriguingly, either LXA₄ or LXB₄ were sufficient to protect primary RGCs and cortical neurons in a dose dependent manner, but related SPMs had no activity. In contrast, inhibition of LX synthetic and signaling pathway components exacerbated RGC loss following metabolic injury. Finally, therapeutic LXB₄ treatment starting after eight weeks of increased intraocular pressure preserved RGC function and survival in a chronic rodent glaucoma model.

Surprisingly, although LXA₄ signaling has been much more extensively investigated, LXB₄ was substantially and consistently more potent in protecting RGC survival in vitro and in vivo. Additionally, only LXB₄ treatment stabilized neuronal mitochondrial membrane potential. As described in the introduction, relatively little is known about LXB₄ signaling, and its receptor has yet to be identified. However, the evidence that LXs operate through divergent pathways is suggested by reports indicating they also induce distinct proresolving pathways. For example, monocyte interaction with LXA₄, but not LXB₄, increases intracellular calcium in leukocytes [33] Also, treatment with LXB₄, but not LXA₄ conferred radioprotection of hematopoetic stem cells in mice [94]. These findings, combined with our own in neurons, suggest that the LXB₄ receptor and signaling cascade is distinct from LXA₄ (Figure 2). More research is needed to further describe the LXB₄ signaling pathway in neuronal and non-neuronal contexts.

Fig 2. LX’s promote neuronal homeostasis.

Proposed LX signaling to promote neuroprotective outcomes. LXA₄ signaling through ALX/FPR2 is defined, but the more potent LXB₄ signaling is unknown.

CONCLUSIONS

LXs have well established proresolution mechanisms in several types of activated immune cells, and have been increasingly implicated in neurodegenerative contexts. Accumulating evidence describes consistent dysregulation of LX synthesis and LOX function, as well as ALX/FPR2 signaling in the chronically injured CNS. Consequently, restoration or supplementation of LX signaling may also present promising therapeutic potential, as they induce a proresolving effect in models of neuroinflammation, while also exerting direct neuroprotective activity. Recent interesting evidence suggests that dietary supplementation of omega-3 PUFA can sustain the levels of SPMs in AD patients [95] and amplifies LXA₄ levels in several tissues in mice [15].

However, more research is needed to understand the signaling induced by each LX in specific diseases. There is particular relevance for these effects in AD pathology, as LXA₄ and ATLXA4 have been shown to reduce Aβ aggregates and neurodegeneration in transgenic mice expressing pathogenic APP. These beneficial effects have been attributed to their inflammation-resolution activity but a direct neuroprotective effect has not been ruled out. Our studies have also demonstrated that LXs exert direct neuroprotective effect on retinal and cortical neurons. Moreover, in vivo protection of retinal neurons was observed following acute injury which is independent of a classical inflammatory response. In this neuroprotective role, both in vivo and in vitro data surprisingly demonstrate that LXB₄ is more potent than LXA₄. These observations indicate that further investigations into LXB₄ signaling will provide insights into novel neuroprotective strategies.

Common Abbreviations:

15-LOX

15 lipoxygenase

5-LOX

5 lipoxygenase

Aβ

amyloid β

AA

arachidonic acid

AD

Alzheimer’s disease

ALX/FPR2

LXA₄ receptor/formyl peptide receptor 2

APP

amyloid β precursor protein

ATLXA₄

aspirin-triggered LXA₄

ATLXB₄

aspirin-triggered LXB₄

CNS

central nervous system

DHA

docosahexaenoic acid

H(p)ETE

hydroperoxy-eicosatetraenoic acid

LX

lipoxin

LXA₄

lipoxin A4

LXB₄

lipoxin B4

MMSE

mini-mental state examination

NFT

neurofibrillary tangles

PUFA

polyunsaturated fatty acid

SOCS-2

suppressor of cytokine signaling

SPM

specialized proresolving lipid mediator