Lipidomics of Bioactive Lipids in Alzheimer’s and Parkinson’s Diseases: Where Are We?

2.1. Glycerophospholipids and Sphingolipids

Abnormal phospholipid composition characterizes the brain of AD patients and mainly affects the cortex area. In the brain of AD patients, phosphatidylcholine (PC) and phosphatidylethanolamine (PE), the two most predominant phospholipids, were significantly decreased, and phospholipid diacylation products glycerophosphocholine were increased in the frontal, primary auditory, and parietal cortices [13]. Three PCs were found significantly diminished in AD: PC (16:0/20:5), PC (16:0/22:6), and PC (18:0/22:6) [14]. Mapstone et al. conducted a 5-year, observational study in healthy, elderly patients and identified ten metabolites, comprising seven PCs, one lysophosphatidylcholine, and two acylcarnitines, that were depleted in the plasma of patients with mild cognitive impairment (MCI) or AD, and the depletion could identify (with accuracy above 90%) cognitively normal individuals who, on average, will convert to MCI or AD within 2–3 years [15]. Although most studies reported a reduction of PC levels in AD, contradictory findings have also been reported. Proitsi et al. found that the lipids most strongly associated with AD were PC 40:4 and PC 36:3, both of which were increased in AD [16]. An increase in CSF PC was observed in AD brains, as compared to control brains. During normal aging, the plasma levels of lysophosphatidylcholine, choline plasmalogen, and lyso-PAF increase significantly; similar but more pronounced changes in these choline-containing phospholipids were observed in AD patients [58].

Cardiolipin is another glycerophospholipid, mainly found in the inner mitochondrial membrane, and responsible for the maintenance of the fluidity and activity of mitochondrial electron transport chain enzymes. A reduction of cardiolipin in the synaptic mitochondrial membranes was reported in the brains of AD patients [59].

Concerning sphingolipids, several studies report variations of these subtype of lipids although there is some inconsistency among them. Levels of sphingomyelins (SMs) and ceramides were studied in two different post-mortem brain regions of patients with AD. The region with extensive Aβ has increased ceramide and decreased sphingomyelin, while in the region with only diffuse Aβ deposits, the ceramide/sphingomyelin ratio was reversed [17]. In the white matter, no differences in sphingomyelin between AD and healthy subjects were found, but significant decreases in ceramide C16:0, C22:0, and C24:1 were observed [18]. Additionally, metabolomic assays reported increased SM levels in the brain tissue of AD subjects, and these were associated with the severity of AD pathology and an increased risk of abnormal cognition, including three SMs with acyl residue (SM C16:0, SM C16:1, and SM C18:1) and one hydroxysphingomyelin with acyl residue (SM (OH) C14:1) [19]. These findings came as support of former evidence in a mouse model of AD, where the intracerebral injection of Aβ promoted the catalysis of sphingomyelin to ceramide by hydrolysis and thus an increase in ceramide levels [60].

Lipidomic studies found increased ceramide levels in AD brains, in particular, ceramides Cer16, Cer18, Cer20, and Cer24 [20]. Senile plaques were abundant in saturated ceramides Cer(d18:1/18:0) and Cer(d18:1/20:0) [61]. Increased levels of ceramides were also found in the CSF [21]. Han et al. observed an enhancement of ceramide in the early AD stages in brain tissue lipid extracts by electrospray ionization mass spectrometry, while its concentration reduced with disease severity [22]. So far, few studies have measured glycerophospholipids and sphingolipids in the serum of AD patients. In particular, only three specific ceramides (containing C16:0, C20:0, and stearoyl as fatty acids) were increased, indicating that they are important predictors of cell damage and memory impairment [23]. The major metabolic product derived from ceramide is Sphingosine 1-Phosphate, and its level was decreased in the AD brain [24,62].

Over the past 15 years, it has been useful to analyze the concentration of lipids in CSF to better understand levels of brain impairment. A research team in Pasadena, California has described the presence of lipid-rich nanoparticles (NPs) in the cerebrospinal fluid of older adults [25]. During LC/MS analysis of the CSF fraction, they revealed that the NPs had a higher glycerophospholipid (PC, PE, PS, and LPS) levels than the supernatant fluid in AD patients [25]. Specifically, a subsequent study revealed that the NPs fraction composition showed increased levels of two odd-numbered MUFAs (C15:1 and C19:1) [63]. The CSF levels of sphingomyelins and phosphatidylcholine-containing MUFAs significantly correlated with p-tau levels, suggesting that this family of bioactive lipids might be associated with disease severity [64].

In search of biomarkers for the non-clinical recognition of AD stages, a method was suggested to distinguish cognitively healthy individuals with normal (CH-NAT) from pathological Aβ42/tau (CH-PAT) and AD. This study revealed that phosphatidylcholine molecular species from the supernatant fraction of CH-PAT were higher than in the CH-NAT AD participants. Furthermore, sphingomyelin levels in the supernatant fraction were lower in the CH-PAT and AD than in the CH-NAT group. The decrease in sphingomyelin corresponded with an increase in ceramide and dihydroceramide and an increase in the ceramide to sphingomyelin ratio in AD. In contrast with the supernatant fraction, sphingomyelin was higher in the nanoparticle fraction from the CH-PAT group, accompanied by lower ceramide and dihydroceramide and a decrease in the ratio of ceramide to sphingomyelin in CH-PAT, as compared with CH-NAT [26]. Other types of sphingolipids, such as sulfatides and gangliosides, although being reported to significantly vary in AD, were not the focus of the present review since they are more constituent of myelin or lipid rafts than actual bioactive lipids.

Since both glycerophospholipids and sphingolipids respectively have two or one fatty acid chains attached to the molecule of glycerol or sphingosine, a great part of literature has focused on measuring levels of FAs. The brain is highly enriched in the PUFAs DHA (22:6n-3) and AA (20:4n-6), and linoleic acid, EPA and DHA account for ~10% of brain lipids.

Patients with MCI and AD had elevated levels of AA, but they had reduced levels of its precursor, linoleic acid (LA), as compared with healthy controls, with the latter progressively decreasing in cognitive performance [27]. Oleic acid (OA), an omega-9 FA and the most abundant dietary FA, was decreased in the frontal cortex and hippocampus of AD brains [28]. Furthermore, a case-control study performed on 148 AD patients reported significantly lower serum levels of DHA, and this reduction was associated with the severity of clinical dementia [65]. A more complete study analyzed the level of FFAs in the serum of AD patients and found that several of them significantly decreased when compared to the control; the study included 3 saturated fatty acids (C14:0, C16:0, and C18:0) and 6 unsaturated fatty acids (C16:1, C18:1, C18:2, γ-C18:3, C20:2, and C22:6). The serum level of C18:3 was significantly higher in AD patients [30]. In an elegant study by Martin et al., lipid composition was analyzed in lipid rafts of human frontal brain cortex obtained between 3- and 18-h post-mortem, revealing that lipid rafts from AD brains exhibit aberrant lipid profiles compared to healthy brains. In particular, lower levels of omega-3 long-chain PUFAs (mainly DHA) and oleic acid were observed [31].

Studies testing six unsaturated FAs, including linoleic acid (LA), AA, α-linolenic acid (ALA), DHA, EPA, and OA, showed that all these unsaturated FAs were positively associated with neuritic plaques and neurofibrillary tangle burdens, and they were negatively correlated with cognitive performance. In brain regions vulnerable to AD pathology—the middle frontal and inferior temporal gyri, there were decrements in LA, ALA, and AA, and increases in DHA [32]. All these unsaturated FAs can directly interact with Aβ40 and Aβ42 peptides, and they display excellent anti-aggregation properties by preventing amyloid fibril formation, especially OA and DHA [66]. However, when investigating the role of different unsaturated FAs in modulation of neuroprotective α-secretase-cleaved soluble APP (sAPPα) secretion and cell membrane fluidity, only AA, EPA, and DHA with four or more double bonds are capable of increasing membranous fluidity and sAPPα secretion, whereas stearic acid (SA, 18:0), LA, ALA, and OA cannot [67].

Among all PUFAs, DHA is assuredly the most investigated in AD, and since the end of the 1990s, many studies have concurred that a reduction of this lipid in several tissues [10,28] suggests a key role in AD pathology. AD patients have decreased DHA levels throughout their brains, including the disease-resistant regions, but the most prominent reduction occurs in the hippocampus, and DHA content has been found to show a positive correlation with dementia [33]. Livers from AD patients also contain lower levels of DHA but higher levels of short-chain n-3 precursors, including tetracosahexaenoic acid, suggesting a defect in its bioconversion into DHA [34]. However, some studies reported no significant changes in DHA levels between erythrocytes or the brain tissue of AD and control subjects [68,69].

In contrast, a very recent study with 2 years of follow-up reported that DHA is a strong protective factor for cognitive decline, inasmuch as AD patients who had stable cognitive impairment for 2 years resulted in higher baseline serum DHA levels than patients with declining AD [70].

2.2. Classical Eicosanoids

Most of the eicosanoids derived from arachidonic acid, i.e., PGs, LTs, HETEs and EETs, have been detected and associated with AD, as elegantly reviewed by Biringer [40] and reported by other studies that analyzed their oxidation products [71] or revealed specific increases in 12-HETE in plasma and brain and decreases in 15-HETE and PGD₂ during the early phase [35]. An eicosanoid known to be a specific marker of in vivo lipid peroxidation is isoprostane-F2a, which is present in elevated concentrations in the CSF of AD [36] and in plasma, too [37,72]. Although plasma levels of F2a-isoprostanes do not qualify as robust biomarkers for AD diagnosis as in the CSF, even if plasma measurements may still have value in clinical trials. A hypothetical classification system for AD diagnosis is based on CSF, Aβ42, and tau levels, which allowed the improvement of the diagnostic accuracy for AD. For discriminating AD from non-AD, the level of CSF F2-IsoP needs to be greater than 25 pg/mL [38]. However, in another study, it was found that peripheral F2A and F4-Neuroprostatene levels in urine and plasma are not increased in AD [73,74].

The alteration of neurotransmitters, lipids, and sterol metabolism are the pathways impaired in the CSF of AD patients. The major metabolites altered are prostaglandins (PGG₂ and PGJ₂), hydrocortisone, and tetrahydrocortisone [39]. In a recent work on the CSF of AD patients, the levels of both pro-inflammatory and pro-resolving LMs seem to be associated with cognitive dysfunction [50].

Although analyses of leukotrienes in CSF have been performed since the 1980s, no significant alteration in the context of AD is known. The levels of PGs in post-mortem brains revealed regional differences in the patterns of PG profiles between the parieto-occipital cortex and the other cortical areas; only TBX₂ and PGD₂ were greater than in controls [41]; specifically, PGD₂ was significantly increased in the frontal cortex [40], and PGE₂ and PGF₂ were reduced [29].

Regarding those fluid samples more easily accessible from the patient, such as CSF, plasma, and urine, the levels of eicosanoids were significantly different from those of control subjects. Many studies observed alterations in the levels of diacylglycerols, prostaglandins, and phospholipids, which can be related to oxidative stress and membrane breakdown.

Elevations in serum prostaglandin levels are important markers of oxidative stress [75]. PGE₂ concentrations in CSF were approximately 5 times higher, and 6-keto-PGF1a levels were 4-fold less than those of control subjects [42]. In contrast, the detection of PGF_2a, PGD₂, and TXB₂ did not reveal differences between AD and the control. Subsequently, a study investigated the correlation between PGE2 and cognitive scores, and they found that PGE2 levels were higher when learning scores were just below the normal range (early Alzheimer’s disease), but declined with progressive learning impairment [39,43,76]. These findings support the hypothesis that inflammatory processes predominate in early AD.

In urinary samples, the concentrations of F2-isoprostanes, PGF₂ and 8-isoPGF₂, significantly increased in AD patients when compared to those of the healthy subjects [44]. Thromboxane (TX), the last class of eicosanoids, did not seem to change in plasma or CS; instead, urinary samples with higher concentrations of 11-dehydro-TXB₂ were reported in AD patients, as compared to healthy controls [77]. Concerning this lipid class, an Alzheimer’s Disease Anti-inflammatory Prevention Trial (ADAPT) analyzed the urine and plasma of participants to determine whether treatments for disease produced serious, adverse cardiovascular (CV) events, and an association between high urine Tx-M/PGI-M ratios and CV events was observed [78].

2.3. Specialized Pro-Resolving Mediators

Despite the relatively recent discovery of this superfamily of bioactive lipids, which is constantly flourishing and up-to-date counts include more than 30 members, measurement and quantification of SPMs in AD started as early as 2005 with the work by the group of Prof. Nicolas Bazan, when the levels in post-mortem AD brain tissues of the DHA-derived neuroprotectin D1 (NPD1) were found to be reduced in the post-mortem hippocampal CA1 region, but not in the thalamus or occipital lobes of AD patients, as compared to age-matched controls [45]. After this study, the reduction of NPD1 levels during the AD course have been corroborated in the hippocampus of the 3xTg mouse model of AD [46]. Although only two studies used the less-standardized enzyme immune assay (EIA) developed for a few SPMs by Cayman, whereby LXA₄, RvD2, and RvE1 were reduced in the brain or hippocampus of 3xTg mice and 5xFAD mice, respectively [79,80], the first complete lipidomic profiling was undertaken in 2015 by the group of Prof. Serhan. This study found a reduction of RvD1 in CSF but not in the hippocampus, and it found decreased levels of MaR1 in the hippocampus in post-mortem human AD brains [47]. The same group conducted randomized, double-blind, and placebo-controlled clinical trial on AD patients—the OmegAD study—in which a placebo or a supplement of 1.7 g DHA and 0.6 g EPA was taken daily for 6 months. After the treatment, RvD1 and LXA₄ levels were measured by EIA on the peripheral blood mononuclear cells (PBMCs), and their levels remained unchanged, suggesting that omega-3 supplementation prevented a reduction in SPMs [81]. A similar result was obtained in another clinical trial in human patients affected by mild cognitive impairment (MCI) and pre-MCI conditions, whereby the supplementation with omega-3 fatty acids and antioxidants led to increased levels of RvD1 in macrophage cultures isolated from PBMCs, as quantified by EIA [82]. The search for SPM measurements as potential biomarkers of AD pathology was then performed on a brain area different from the previously analyzed hippocampus, namely the entorhinal cortex, an anatomical region of importance for memory which is affected early in AD pathogenesis. In this study, several SPMs were significantly reduced as compared to healthy controls, such as PD1, MaR1, and RvD5, while LXA₄, LXB₄, RvD1, RvD2, RvE1, and RvE2 did not show relevant changes [48]. Subsequently, SPMs were analyzed in 3 different mouse models of AD: Fat-1 mice, Tg2576 mice, and APP/PS1/SphK1 mice. As expected, these studies led to contrasting results, and several SPMs were undetectable in the brain [83], increased in the plasma [84]—in both cases following a DHA-enriched diet, or reduced in neuronal cells [49].

Recently, the first lipidomic profiling of cerebrospinal fluid in AD was undertaken to simultaneously analyze both SPMs and classical eicosanoids in patients with cognitive impairment, ranging from subjective impairment to a diagnosis of AD, and correlated to cognition, CSF tau, and β-amyloid. In this study, RvD4, RvD1, PD1, MaR1, and RvE4 were lower in AD and/or MCI, as compared to patients with spinal cord injuries. Furthermore, levels of RvD1 showed a negative correlation with p-tau levels, while RvD4 negatively correlated with AD tangle biomarkers, and positive correlations with cognitive test scores were observed for both SPMs and their precursor fatty acids [50].

3.1. Glycerophospholipids and Sphingolipids

Similar to AD, PD changes in the lipid membrane composition are strictly associated with its pathology, and superfamilies of several bioactive lipids seem to be involved in the pathogenesis of PD [133,134].

One prominent gene involved in the pathogenesis of PD is the GBA gene, which encodes for the β-glucocerebrosidase enzyme, which hydrolyzes glucosylceramide. Mutations in this gene are found in up to 10% of patients with PD, thus implicating aberrant sphingolipid metabolism [135]. Indeed, the CSF lipidome signatures of Parkinson’s patients at early stages, de novo PD patients with abnormal dopamine transporters, and healthy controls show distinct lipidome changes, with a significant increase in glucosylceramide (GlcCer), while sphingomyelin (SM) was significantly reduced in GBA-PD patients [93]. Furthermore, applying the ratio of GlcCer to SM for stratifying cases of idiopathic PD revealed that patients with a high GlcCer/SM ratio present increased cognitive deterioration compared to those in the lowest quartiles. However, other studies did not find any significant accumulation of GluCer in the CSF of PD subjects [136]. Of note, PD patients with neuropathic and dyskinetic pain had higher plasma levels of GlcCer compared to PD patients with no sensory pain [116].

Interestingly, plasma ceramide lipidomics linked cognitive impairments among PD patients with higher levels of the plasma ceramides C16:0, C18:0, C20:0, C22:0, and C24:1, and the monohexosylceramide species C16:0, C20:0, and C24:0 [94]. Thus, plasma ceramide and monohexosylceramide metabolism seem to be altered in PD and also in non-GBA mutation carriers, with higher levels being accompanied by worse cognition. These findings were confirmed by shotgun lipidomics of L444PGBA mutation fibroblasts, an ex vivo PD system that presents a mutation of the GBA gene, showing distinct, increased proportions of ceramides and hexosylceramide, while total phospholipids were significantly decreased [95]. Intriguingly, isolated membrane lipids from PD-GBA cultures proved to be more potent triggers of recombinant human α-syn in vitro fibrillation compared to lipids from healthy controls [95]. LRRK2 is another important gene for the development of PD, and it has been implicated in a variety of pathways, including mitochondrial dynamics. LRRK2 is mutated in families with autosomal-dominantly inherited PD, and it has been acknowledged as a susceptibility factor for PD [137]. Although in vivo models of LRRK^−/− show an intact dopaminergic system [138], targeted lipidomics approaches identified altered sphingolipid composition in LRRK2^−/− mouse brains, with significantly increased levels of Cer in LRRK2^−/− mice and direct alterations in GBA1 enzymatic activity [96].

Furthermore, new machine learning analysis has been introduced to link lipidome data of PD patients to the course of disease severity by using whole blood lipidomics and an established lipid panel containing dihydrosphingomyelin (dhSM), GlcCer, dihydro globotriaosylceramide (dhGB3), and dihydro GM3 ganglioside (dhGM3). [139]. In addition, untargeted lipidome assessments were utilized to detect incipient dementia in PD. Lipidome analysis by isotope-standard-assisted liquid chromatography of PD patients’ serum revealed that an association with patients transitioning to the development of dementia could be characterized by a 5-lipid biomarker panel [97].

A multi-omic integration analysis on 30 drug-naive, de novo PD patients and 30 matched controls revealed that a longitudinal trajectory with 3 long-chain fatty acids (FA) (FA C14:0, FA C17:1, and FA C20:1) and dopaminergic medication exerted a strong change in lipidomic signature, with a prevalence of phospholipid species and breakdown products of phospholipids [98].

The substantia nigra forms the major spot of interest in PD research, as this brain area is characterized by the loss of dopaminergic neurons, resulting in the clinical manifestation of PD. However, information regarding the lipid alterations in established PD animal models is still scarce. HPLC-ESI-MS/MS measurements of the substantia nigra from rats 21 days after an infusion of 20 µg of 6-OHDA, or a saline vehicle into the anterior dorsal striatum, displayed a relative down-regulated abundance of several PC species, while known neuroinflammatory signaling lipids, namely lysophosphotidylcholine (16:0 and 18:1), were up-regulated [99].

In post-mortem human substantia nigra samples, 5 lipid species have shown different abundances between PD patients and controls, namely Bis (Monoacylglycerol) Phosphate (BMP) 42:8, PC 36:3, PE A36:2, PI 42:10, and PS 36:3, with BMP and PI showing a significant upregulation in PD samples [100]. The same study also showed a saturated sphingomyelin species depletion in the putamen of PD patients. Post-mortem-acquired sphingolipidome data of the caudate, putamen, and globus pallidus revealed altered levels of sphingolipid species, including ceramides (Cer), dihydroceramides (DHC), hydoxyceramides (OH-Cer), phytoceramides (Phyto-Cer), and phosphoethanolamine ceramides (PE-Cer) in the PD subjects compared to healthy controls. The putamen showed the strongest effect of depletion of Cer and Cer-OH, while sphingomyelin levels remained largely unchanged across the basal ganglia of both PD and control samples [101]. These results are in line with prior studies reporting unchanged total glucosylceramide levels in the putamen and cerebellum in both GBA-PD and sporadic PD brain samples [140]. However, other studies have reported decreased levels of SM in the anterior cingulate cortex [102] and increased levels in the substantia nigra [103], while lipidome analysis of the visual cortex, a brain area that does not show any formation of Lewy bodies in PD, showed increased overall sphingolipid levels [104]. Applied lipidomics on the Lewy bodies in the post-mortem mesencephalon and corpus callosum of PD patients with dementia confirmed a high cell-membrane-related lipid content within the Lewy bodies, with the mass spectra showing strong peaks corresponding to SM and PC for positive immunostained α-syn inclusions isolated from both the SN and hippocampal CA2 sector [141]. However, the profile of SM/PC does not seem to be specific for LB.

It is important to note that, while post-mortem data acquired from PD human brain samples show distinct changes in lipid profiles, post-mortem-acquired parkinsonian CSF seems to present increased lipid levels of almost the complete profile of PC and the majority of CM and SM [105]. This effect might result from the breakdown of the blood–brain–barrier and the subsequent influx of non-CSF lipids into the CSF, thus potentially leading to an impaired lipid homeostasis.

Although most lipidomics approaches in PD are focused on the CSF and blood, lipidomics have also been introduced for the examination of other biological substances and tissues, such as the sebum or skin fibroblasts. Sebum is an oil-like substance produced by the sebaceous glands and is predominantly composed of triglycerides, fatty acids, wax esters, squalene, and cholesterol. LC-MS analyzation of sebum samples, including drug-naïve PD, medicated PD, and healthy controls, revealed an enrichment of the sphingolipid metabolism in both drug-naïve and medicated PD [106]. Lipid dysregulation in the sphingolipid metabolism in the sebum of PD patients reflected an overall lipid dysregulation. Additionally, lipid profiling of parkin-mutant, human skin fibroblasts also found these lysophophatidylcholine species to be upregulated [107]. Also, while a majority of the presented lipidomic data were acquired by mass spectrometry, different lipidomic approaches involving lipoprotein profiling by the implementation of NMR spectroscopy, bundled with multivariate data pre-processing, show promising results in discriminating between early-stage PD and PD-related dementia [142].

3.2. Classical Eicosanoids

Higher PUFA levels were detected in the supernatants and high-speed membrane fractions of neuronal cells over-expressing wild-type or PD-causing mutant α-syn. This increased PUFA content in the membrane fraction was accompanied by increased membrane fluidity in the α-syn overexpressing neurons. Membrane fluidity and the levels of certain PUFAs were lower in the brains of mice that have had α-syn genetically deleted [108]. Furthermore, lipid rafts were purified from human frontal cortex from the normal, early motor stages of PD and incidental PD subjects, and the lipid composition was analyzed. Lipid classes were separated from a fraction of total lipid by one-dimensional, double-development, high-performance, thin-layer chromatography. The results show that lipid rafts from both cohorts of patients exhibited dramatic reductions in their contents of omega-3 and omega-6 long-chain PUFAs, especially DHA and AA [109].

Metabolomic approaches using UPLS-MS measured the plasma levels of 158 fatty-acid metabolites in a cohort including 42 PD patients and 54 healthy volunteers. Results showed an increase in 2 eicosanoids (arachidonic acid and 13-hydroxy-octadecatrienoic acid) and a decrease in 11 eicosanoids (DHA, lyso-platelet-activating factor, 12-hydroxy-eicosatetraenoic acid, dihydroxy-eicosatrienoic acids, dihidroxy-octadecenoic acids, 17,18-dihydroxy-eicosatetraenoic acid, and hydroperoxy-octadecadienoic acids) in PD patients compared to healthy controls [110]. A similar study identified LTB₃ as a potential biomarker for PD [111].

Among PGs, PGE₂ is the most associated with PD, and several lipidomic studies identified this eicosanoid in PD patients. Indeed, PGE₂ levels were increased in the substantia nigra [143] and the CSF of PD patients with mild cognitive impairment and dementia although, in the first study, it was measured by enzyme immunoassay, whereas in the second study, its levels were not associated with elevated total-tau, phosphorylated-tau, or other inflammatory mediators [144]. In an α-syn model of PD, among all prostaglandins assayed in the brain (PGE₂, PGD₂, and PGF_2α), PGD₂ showed the greatest increase (two-fold) compared to wild-type animals [112]. Furthermore, analysis of brain extracts by electrospray ionization–Fourier transform resonance mass spectrometry (ESI–FT-ICR-MS) revealed an increase in two different prostaglandins (PGB₁ and PGH₂) and in 15(S)-HETE in brain extracts of a Parkinson-like, rat model chronically exposed to manganese [113].

Other eicosanoid members, such as isoprostanes and HETEs, were also identified in PD. Indeed, the levels of plasma F₂-isoprostanes (F₂-IsoPs), neuroprostanes (F₄-NPs), and HETEs were found to be higher in PD patients compared to controls, especially in the early stages of PD. In the same study, a significant negative correlation between the cumulative intake of L-DOPA and plasma total HETEs was also observed [114]. Changes in PGE₂ and PGD₂ levels were corroborated in other models of PD. Indeed, while mice carrying the mutated disease gene DJ-1 showed reduced PGE₂ and PGD₂ levels [145,146], LPS-injected mice displayed increased PGE₂ levels [147,148], and MPTP-injected mice presented enhanced striatal and nigral levels of LTB₄ [149]. However, 6-OHDA-treated rats showed lower PGE₂ levels 4 weeks after treatment compared to controls [150]. Of note, in these studies, eicosanoids were always measured with the commercially available enzyme immunoassay kits from Cayman, and not by lipidomics.

Additionally, LC/MS analysis in rotenone-treated rats showed a significant reduction of oxidizable PUFA-containing phospholipid cardiolipin species and increased contents of mono-oxygenated CL species in later stages of exposure in the substantia nigra. Notably, linoleic acid in the sn-1 position was the major oxidation substrate yielding its mono-hydroxy- and epoxy-derivatives, whereas more readily “oxidizable” fatty-acid residues, AA and DHA, remained non-oxidized in the substantia nigra. On the other hand, elevated levels of PUFA CLs were detected in plasma [115].

3.3. Specialized Pro-Resolving Mediators

As for this superfamily of bioactive lipids, so far, only 1 study from our group measured 2 specific SPMs, i.e., RvD1 and RvD2, in both the plasma and CSF of the transgenic and over-expressing α-syn (Syn) rat model of PD, as well as in early-PD patients. In particular, although no changes were observed in RvD2 levels, the content of RvD1 was higher in 2-month-old Syn rats and progressively decreased in 4-month-old and 18-month-old rats compared to their wild-type littermates, suggesting an impairment in the production of this SPM along disease progression, starting at an early phase when neuroinflammatory, synaptic, and motor-behavioral dysfunctions started to appear. Early-onset PD patients also showed strong reductions in RvD1 levels in both plasma and CSF [128]. However, the measurement of RvD1 and RvD2 were performed by commercially available EIA kits from Cayman, and so far, no study has ever attempted to measure all of the other SPMs through targeted-lipidomics analysis.