n-3 Polyunsaturated Fatty Acids and Their Derivates Reduce Neuroinflammation during Aging

Aging is associated to cognitive decline, which can lead to loss of life quality, personal suffering, and ultimately neurodegenerative diseases. Neuroinflammation is one of the mechanisms explaining the loss of cognitive functions. Indeed, aging is associated to the activation of inflammatory signaling pathways, which can be targeted by specific nutrients with anti-inflammatory effects. Dietary n-3 polyunsaturated fatty acids (PUFAs) are particularly attractive as they are present in the brain, possess immunomodulatory properties, and are precursors of lipid derivates named specialized pro-resolving mediators (SPM). SPMs are crucially involved in the resolution of inflammation that is modified during aging, resulting in chronic inflammation. In this review, we first examine the effect of aging on neuroinflammation and then evaluate the potential beneficial effect of n-3 PUFA as precursors of bioactive derivates, particularly during aging, on the resolution of inflammation. Lastly, we highlight evidence supporting a role of n-3 PUFA during aging.


Introduction
Aging is a world concern as the elderly population tripled from 4% to 13% in the last century and is expected to grow sharply to reach 20% of the population in 2025 and 33% in 2050 [1]. Aging is associated to cognitive decline for 15-20% of the elderly >65 [2][3][4]. These cognitive alterations can lead to age-related disease such as neurodegenerative diseases. Alzheimer's disease is the predominant one, affecting 24 million people in the world [5]. Thus, healthy aging constitutes a real economic challenge of the 21st century for the nations. The mechanisms explaining this process are still not fully elucidated, but neuroinflammation seems largely involved. Then, strategies to reduce and resolve neuroinflammation in a time-limited manner are encouraged. Recent studies suggest that nutrition, particularly fish oil, has promising anti-inflammatory effects. Fish oil contains n-3 long chain polyunsaturated fatty acids (LC-PUFAs), which are precursors of bioactive lipids called specialized pro-resolving mediators (SPMs) that largely contribute to this beneficial effect. Here, we review the effect of aging on neuroinflammation, in particular microglia activity and cognitive decline, and how n-3 LC-PUFAs and their derivates impact neuroinflammation, especially during aging. We discuss that nutrition, an environmental factor to which individuals are exposed throughout life, plays a key role to prevent or delay neuroinflammation during aging.
but have also been detected in the brain. RvD1 was measured in mouse brain following cerebral ischemia [118]. Its level is modulated by a DHA intravenous injection [119] and during inflammation; it decreases at the beginning and then increases during the resolution phase [120]. RvD1 acts at picomolar range, but exerts its biological effects at nanomolar range [117,121]. The receptor of RvD1 is lipoxin A4 receptor/formyl peptide receptor 2 (ALX/FPR2) in rodents and G protein coupling receptor 32 (GPR32) in humans [116]. Several brain structures express ALX/FPR2: brainstem, spinal cord, hypothalamus, cortex, hippocampus, cerebellum, and striatum [122]. At the cellular level, these receptors are expressed in microglial cells [123], neurons [122,124], and astrocytes [125,126]. Via these receptors, RvD1 regulates micro-RNAs (miRNAs), which play a key role in modulating the expression of target genes such as inflammatory genes [123,125,[127][128][129].

Role of Lipid Mediators in the Resolution of Inflammation
A large number of studies support the beneficial role of n-3 LC PUFAs in inflammation in human and animal models of acute and chronic inflammation, including in the brain (for recent reviews, see [82,83]). Here, we will review the biological roles at the brain level of RvD1 and RvE1, two distinct lipid mediators generated from the n-3 LC-PUFAs DHA and EPA, known for their powerful anti-inflammatory and pro-resolutive properties.

In Humans
The effect of RvD1 was mainly studied in Alzheimer's and Parkinson's patients (Table 1). In patients with dementia, the levels of RvD1 in cerebrospinal fluid are positively associated with the improvement of cognitive functions [126]. RvD1 promotes Aβ phagocytosis by macrophages isolated from Alzheimer's patients, reducing the amyloid load [151,152]. Moreover, as cited in Krashia et al., endogenous RvD1 is decreased in patients diagnosed with early-Parkinson's disease [153]. As a result, the decrease of RvD1 levels in Alzheimer's and Parkinson's disease patient's brain could contribute to the disease development and progression. Conversely, an increased anti-inflammatory RvD1 activity has been reported in maniac and depressive patients, suggesting that RvD1 may serve to improve inflammatory imbalance [154].
The effect of RvE1 was reported in patients at the periphery (Table 1) [155][156][157], but not at the brain level. Hence, more studies are needed to develop this area.

In Animals
Several studies report that, in rodent models of inflammation, RvD1 and RvE1 display anti-inflammatory activities in the CNS (Table 2). Indeed, RvD1 reduces the activation of NFκB and the expression of pro-inflammatory factors such as IL-1β, IL-6, TNF-α, and iNOS in rats with hemorrhagic shock or in streptozotocin (STZ)-induced diabetic retinopathy [158,159]. RvD1 attenuates neuroinflammation through ALX-FPR2 receptor via miRNA in a neonatal hypoxia-ischemia rat pup model or in a remote damage model [125,160]. Moreover, RvD1 induces the polarization of macrophages and microglia toward an M2 phagocytic phenotype [161][162][163]. Chronic and early RvD1 administration in a rat model of Parkinson's disease prevents central and peripheral inflammation, as well as neuronal dysfunction and motor deficits [153]. In addition, the precursors of RvD1, 17R-HDHA and 17S-HDHA, reduce the production of pro-inflammatory cytokines in the spinal cord and in the hippocampus [135,164].
RvE1 reduces the expression of pro-inflammatory cytokines IL-1β and IL-6 in the prefrontal cortex and decreases the neuropathological features of Aβ pathology in a murine model of Alzheimer's disease [165]. Furthermore, repeated RvE1 administration moderates the activation of microglia by promoting ramified microglia following traumatic brain injury or peripheral brain injury [166].
The effect of RvD1 on neuroinflammation is associated to effects on cognition. Indeed, RvD1 prevents cognitive deficits. In a rodent model of systemic inflammation or traumatic brain injury, an intraperitoneal administration of 17R-RvD1 prevents cognitive decline [166,167]. Of note, higher levels of brain RvD1 in Fat-1 mice, owing to higher brain n-3 LC-PUFAs induced by genetic means, are linked to less cognitive deficits, a reduction in microglial activation, and in pro-inflammatory status following brain ischemia [168,169]. Conversely, lower levels of brain RvD1, owing to 15-LOX inhibition, are related to alterations in working memory and synaptic plasticity in rats [97].
RvD and RvE have been reported to prevent emotional behavior alterations in rodent models of mood disorders in the review of Furuyashiki et al. [170]. These SPMs have positive effects in LPS-induced or chronic stress-induced or post-myocardial infarct depression [164,[171][172][173][174][175][176].  The increase in brain n-3 PUFA reduces LPS-induced pro-inflammatory cytokine production and subsequent spatial memory alteration [169] Luo et al.

In Vitro
The effects of RvD1 and RvE1 were tested on different brain cells, highlighting their pro-resolutive properties (Table 3). In microglial cells, RvD1 enhances the effect of the anti-inflammatory cytokines IL-4, Arg1, and Ym1 and reduces the activation of microglia by decreasing CD11b expression, leading to a more anti-inflammatory phenotype of microglia [163,177,178]. Moreover, RvD1 reduces LPS-induced pro-inflammatory cytokine (TNF-α, IL-6, and IL-1β) gene expression in microglial BV2 cells by regulating miRNA expression [123]. It was also reported that RvD1 down-regulates Aβ-induced inflammation in human microglia [136]. RvD2 decreases the expression of toll like receptor 4 (TLR4, the receptor of LPS) following LPS treatment, and consequently its downstream signaling pathway NFκB [179]. RvE1 also reduces microglial activation and pro-inflammatory cytokine release in microglial cells [123,177]. In astrocytes, RvD1 decreases LPS-induced TNF-α production [164]. In neurons from spinal nods, RvD1 increases neurite outgrowth [180]. In PC12 neural cells, used as an in vitro model of Parkinson's disease, RvD1 reduces TNF-α and IL-6 mRNA expression [181]. The anti-inflammatory properties of RvD1 were also tested in macrophage cells. RvD1 reduces the expression of pro-inflammatory markers (cytokines, PGE2) and increases anti-inflammatory cytokine IL-10 in murine macrophages stimulated by LPS [182]. RvD1 polarizes primary human macrophages toward a pro-resolutive phenotype through GPR32 receptor [183].

Defects in Lipid Metabolism and Lipid Mediator Production during Aging
During aging, brain levels of n-3 LC-PUFAs decrease, although all brain structures are not affected in the same way [30,32,70,184]. This reduction was observed in human [185,186], especially in the cortex, the hippocampus, and the cerebellum [73,[187][188][189], and in rodents [30,32,190,191], in particular in the hippocampus [191] and the cortex [73], which are key structures in mnesic processes. This decrease is mainly because of changes in lipid metabolism: alteration in the intestinal absorption of essential fatty acids [192][193][194]; impairment in the enzymes of phospholipid synthesis [195]; reduced conversion rates of the precursors into LC-PUFAs owing to reduced activity of the enzymes involved in their synthesis, in particular of ∆6 desaturase [186,196,197]; and modifications in the expression of the genes implicated in the metabolism of PUFAs. Indeed, single nucleotide polymorphisms (SNPs) in desaturase genes FADS1 (∆5 desaturase), FADS2 (∆6 desaturase), as well as ELOVL2 (elongase 2) are related to higher ALA and lower EPA plasma phospholipid levels with age, suggesting different rates of conversion [198]. Moreover, another possible reason of the decrease of n-3 LC-PUFAs in the membranes is their high propensity to oxidation to generate peroxidation products such as malonaldehyde (MDA), 4-hydroxy-2-nonenal (4-HNE), or 4-hydroxy-2-hexenal (4-HHE). Indeed, levels of MDA and 4-HNE are increased in aged brain of humans and rodents [199,200].
Aging-associated DHA metabolism disturbance could participate in cognitive decline ( Figure 2). This has been demonstrated both in humans and animals. In elderly, decreased n-3 PUFA consumption associated to reduced erythrocyte DHA levels are inversely correlated with age-related cognitive decline [201][202][203]. In rats, a low-DHA dietary supply for one or more generations is related to alterations in cognitive function [204][205][206]. In addition, we showed in aged mice that an n-3 PUFA deficient diet impairs memory as well as neuroinflammation and synaptic plasticity [32,[207][208][209][210]. Furthermore, the decrease in brain DHA content induced by a n-3 PUFA deficient diet increases vulnerability to inflammation, which trigger both synaptic and memory alteration [211,212]. On the contrary, a two-month n-3 LC-PUFA supplementation in aged mice (between 20 and 22 months old) reverses age-induced spatial memory deficits [30].
Age-related alteration of n-3 PUFA metabolism contributes to reducing the n-3 LC-PUFA content in brain phospholipids. As n-3 LC-PUFAs are precursors of bioactive mediators involved in the resolution of inflammation, it may have consequences on SPM profile and production. Indeed, it was recently shown that blood oxylipin profile is altered in 45-64-year-old healthy men and women versus 19-28-year-old young people [213,214]. Moreover, Gangemi et al. (2005) demonstrated that aging is associated to a decrease in urinary LxA4/leukotriene, a ratio of anti-inflammatory/pro-inflammatory mediators synthesized from arachidonc acid and considered as an index of the endogenous anti-inflammatory potential [215]. Moreover, LxA4 is significantly lower in cerebrospinal fluid (CSF) of humans with Alzheimer's disease as compared with humans with mild cognitive impairment or subjective cognitive impairment, with a positive correlation between CSF LxA4 and cognitive performance [126].
The modifications of oxylipin profile are linked to changes in the expression of the enzymes involved in oxylipin formation. In humans, the expression of PLA2 and LOX increases with aging in post-mortem brain [214]. Similar results were obtained in 70-year-old versus 41-year-old patients concerning PLA2 and CYP [214]. In Alzheimer's disease patients, 15-LOX level is also increased in the hippocampus [126]. Aging-associated DHA metabolism disturbance could participate in cognitive decline ( Figure 2). This has been demonstrated both in humans and animals. In elderly, decreased n-3 PUFA consumption associated to reduced erythrocyte DHA levels are inversely correlated with age-related cognitive decline [201,202,203]. In rats, a low-DHA dietary supply for one or more generations is related to alterations in cognitive function [204,205,206]. In addition, we showed in aged mice that an n-3 PUFA deficient diet impairs memory as well as neuroinflammation and synaptic plasticity [32,207,208,209,210]. Furthermore, the decrease in brain DHA content induced by a n-3 PUFA deficient diet increases vulnerability to inflammation, which trigger both synaptic and memory alteration [211,212]. On the contrary, a two-month n-3 LC-PUFA supplementation in aged mice (between 20 and 22 months old) reverses age-induced spatial memory deficits [30].
Age-related alteration of n-3 PUFA metabolism contributes to reducing the n-3 LC-PUFA content in brain phospholipids. As n-3 LC-PUFAs are precursors of bioactive mediators involved in the resolution of inflammation, it may have consequences on SPM profile and production. Indeed, it was recently shown that blood oxylipin profile is altered in 45-64-year-old healthy men and women versus 19-28-year-old young people [213,214]. Moreover, Gangemi et al (2005) demonstrated that aging is associated to a decrease in urinary LxA4/leukotriene, a ratio of anti-inflammatory/proinflammatory mediators synthesized from arachidonc acid and considered as an index of the endogenous anti-inflammatory potential [215]. Moreover, LxA4 is significantly lower in cerebrospinal fluid (CSF) of humans with Alzheimer's disease as compared with humans with mild cognitive impairment or subjective cognitive impairment, with a positive correlation between CSF LxA4 and cognitive performance [126].
In animals, oxylipin profile modification was also reported with aging. Aged rodent brains display higher levels of TxB2, 6-keto-PGF1, and PD1-like metabolites [214]. In a model of senescence-accelerated prone mice (SAMP8), the cortex contains higher levels of PGE2, TxB2, and In animals, the expression of 5-LOX is increased with aging [214] whereas the expression of 12-LOX is decreased in nine-month-old SAMP8 mice [216].
The changes in oxylipin profile may have compensatory consequences on their receptors. Indeed, in humans, ALX/FPR2 and ChemR23 levels are higher in the hippocampus of Alzheimer's disease patients as compared with controls [126]. A similar result was obtained for ALX/FPR2 in SAMP8, despite that its level is similar to the SAMR1 controls [216].
All these results suggest an altered resolution of inflammation during aging that may contribute to the age-related cognitive decline, as high inflammation is associated to altered cognition.

Evidence Supporting a Role of Dietary n-3 PUFAs during Aging
Bioactive nutrients such as n-3 PUFAs constitute an interesting potential way to prevent or delay neuroinflammation that occurs during aging. Here, we will focus on dietary n-3 PUFAs because they modify the levels of brain n-3 LC-PUFAs [83,84,218] that are both anti-inflammatory and pro-resolutive and prevent cognitive decline associated to aging.
Evidence in humans (Table 4) and animals (Table 5) supports a powerful role of n-3 LC-PUFAs in the regulation of both inflammatory pathways, and in fine, in the resolution of inflammation, including in the brain (recently reviewed in [83]). Here, we will focus on dietary supplementation using n-3 LC PUFAs during aging. Barberger-Gateau highlighted in elderly that the more they eat n-3 PUFAs, the less they are subjected to cognitive decline [219]. Similarly, Tan et al. showed in the Framingham Study participants that lower erythrocyte DHA levels are related to cognitive impairment [220]. Moreover, in a prospective observational study, baseline dietary DHA intake levels at 70 years old are positively correlated with a better declarative memory test performance at the age of 75 in a healthy population [221]. Dietary supplementation with n-3 PUFAs conducted in humans has been motivated by observational studies showing the link between dietary consumption of DHA and improved cognitive function and/or reduced cognitive decline in the elderly. Indeed, fish oil consumption, leading to increased levels of DHA in erythrocytes, has been associated with better cognitive performance in elderly [222] and with a lower risk of developing neurological disorders [223][224][225]. DHA dietary supply is associated to better performance and speed in a verbal learning test in a cohort of 45-70-year-old healthy individuals [226] and to improved mini mental state examination (MMSE) scores, used to evaluate cognitive functions and memory abilities, in a cohort of elderly of 75-year-olds [227]. Yurko-Mauro et al. have shown in a systematic meta-analysis that DHA intake improves episodic, working and semantic memories [228]. More recently, McNamara et al. have revealed that fish oil consumption decreases self-reported inefficiencies in everyday functioning as well as improves cognition in elderly with cognitive complaints [229]. Moreover, circulating n-3 PUFAs (including DHA) have been negatively associated to the level of cytokines [230][231][232].  Beneficial effects of n-3 LC-PUFAs have also been found in animals. Administration of a DHA/EPA diet to aged mice protects against neuroinflammation and cognitive impairment [30] and improves spatial cognition and learning ability and memory [233,234]. Interventional studies in aged rodents have demonstrated that the ingestion of a fish oil-enriched diet decreases the ex vivo production of IL-1β, TNF-α, and IL-6 by monocytes and macrophages [235][236][237]. Moreover, circulating concentrations of IL-1β, TNF-α, and IL-6 following LPS injections are lower in rats and mice fed a fish oil-enriched diet [238][239][240]. Furthermore, age-related brain expression of pro-inflammatory cytokines in rodents is reduced with high levels of DHA [241].
In addition, it is possible to modulate oxylipin profile via dietary interventions. Indeed, as reviewed by Caligiuri et al. in human blood, the oxylipin profile is changed towards a less inflammatory profile after n-3 LC-PUFA consumption [214]. We found that in mice treated with LPS, a brain n-3 LC-PUFA increase by dietary supplementation promotes the synthesis of n-3 PUFA derived SPMs and decreases n-6 PUFA-derived SPMs displaying an anti-inflammatory profile [100]. Moreover, increased plasmatic pro-inflammatory oxylipins in elderly is reversed by dietary n-3 PUFA (alpha-linolenic acid, the precursor of n-3 LC-PUFAs) [213]. The OmegAD study revealed that Alzheimer's disease patients treated with n-3 PUFAs preserve their RvD1 levels as compared with placebo-treated patients [242].
In aged rats, n-3 LC-PUFA supplementation increases DHA-derived oxylipins in the cortex and improves the reference memory-related ability learning [243].
The modification of SPM levels in blood and brain cells of aged human and rodents is accompanied by some modification of the expression of their enzymes involved in their synthesis. 15-LOX mRNA expression increases in n-3 LC-PUFA supplemented group and decreases in n-3 LC-PUFA deficient diet [100,244,245]. 15-LOX generates both 15-HETEs that inhibit NFκB [103] as well as RvD1 that contributes to the preservation of cognitive performance [97].
These results suggest that dietary habits may be essential regulators of oxylipin profile and reinforce the importance of the recommendation of n-3 PUFA rich diet.

Conclusions
In conclusion, aging is characterized by low-grade neuroinflammation, in particular, activation of microglial cells and increase in the production of brain pro-inflammatory factors, such as cytokines. This neuroinflammation is associated with cognitive decline (15-20% of the >65-year-old elderly), which affects life quality and has a major economic and social impact. In this context, it is a priority to find strategies to delay the evolution towards neurodegenerative diseases. n-3 LC-PUFAs and their bioactive lipid derivates (SPMs) are promising as they reduce and resolve inflammation. SPMs are modulated by aging and dietary means reinforcing the importance of nutrition in the regulation of inflammation. Changes in dietary n-3 PUFA balance should have dramatic consequences in brain PUFA metabolism, and finally in the response to neuroinflammation particularly during aging. More studies are needed to confirm the role of SPMs in age-related changes, with this field being yet in emergence, and to investigate the interest to combine different oxylipins to potentiate their beneficial effects during aging. The clinical form (encapsulated SPMs or more stable-SPM analogues), the doses, and the way of administration should also be defined.