Dietary Inulin Supplementation Affects Specific Plasmalogen Species in the Brain

Plasmalogens (Pls) are glycerophospholipids that play critical roles in the brain. Evidence supports the role of diet and that of the gut microbiota in regulating brain lipids. We investigated the impact of dietary intake of inulin—a soluble fiber used as prebiotic—on the Pl content of the cortex in mice. No global modification in the Pl amounts was observed when evaluated by gas chromatographic analysis of dimethyl acetals (DMAs). However, the analysis of individual molecular species of Pls by liquid chromatography revealed a reduced abundance of major species of ethanolamine Pls (PlsEtn)―PE(P-18:0/22:6) and PE(P-34:1)―in the cortex of mice fed a diet supplemented with inulin. DMA and expression levels of genes (Far-1, Gnpat, Agps, Pla2g6 and Tmem86b) encoding key enzymes of Pl biosynthesis or degradation were not altered in the liver and in the cortex of mice exposed to inulin. In addition, the fatty acid profile and the amount of lyso forms derived from PlsEtn were not modified in the cortex by inulin consumption. To conclude, inulin affects the brain levels of major PlsEtn and further investigation is needed to determine the exact molecular mechanisms involved.


Introduction
The brain is the second-richest organ in terms of lipid content after adipose tissue. Lipids account for about half of the dry weight of the brain and are essential components in the structure and function of this organ [1][2][3]. The crucial role of lipids in maintaining the health status of the brain is well illustrated by the existence of neurological disorders (e.g., mood disorder, bipolar disorders and schizophrenia) and neurodegenerative diseases (e.g., Alzheimer's disease (AD) and Parkinson's disease (PD)) that are associated with alterations in lipid homeostasis in the brain [1,2]. In addition to displaying a high lipid content, the brain is also characterized by a high lipid diversity, which relies mainly on fatty acids [4,5]. In the brain, phospholipids are the reservoirs of fatty acids, and particularly of arachidonic acid (ARA, C20:4n-6) and docosahexaenoic acid (DHA, C22:6n-3), which are polyunsaturated fatty acids (PUFAs) involved in the regulation of the structure and functions of brain cells [6,7]. Indeed, in addition to serving as an energy source, fatty acids  It is well documented that the gut microbiota influences the physiology of organs at distance from the gut mucosa, including the nervous tissues [24,25]. In particular, there is evidence that the gut microbiota modulates the lipid composition of both the brain and the retina-the neurosensorial tissue that lines the back of the eye and that is known to be an extension of the central nervous system. Indeed, analysis of the retinal lipidome of germ-free mice and conventionally raised mice showed that the gut microbiota influences the PlsEtn content of the retina [26]. In addition, comparison of the lipid profile of germfree mice colonized with the gut microbiota of young or old donor mice revealed that the composition of the gut microbiota affects the cholesterol and phospholipid content of the cortex, including phosphatidylcholine (PChol), phosphatidylethanolamine (PEtn) and PlsEtn species [27].
Diet and the gut microbiota are intrinsically linked [28]. Among dietary factors shaping the gut microbiota and influencing its functions is the consumption of dietary fibers [29]. Dietary fibers can be categorized according to their water solubility. Whereas insoluble fibers (e.g., cellulose or hemicellulose) are poorly digested in the colon by the gut microbiota, soluble fibers (e.g., inulin-type fructans) can be fermented by gut bacteria. The fermentation of soluble fibers by bacteria generates metabolites (e.g., short-chain fatty acids (SCFAs)) that can have biological effects on the host, including effects on lipid metabolism [30]. A lack of fibers has been shown to alter the composition, diversity and richness of the gut microbiota [31][32][33]. Soluble dietary fibers may influence the gut microbial ecosystem in several ways. The consumption of soluble dietary fibers favors not only the expansion of gut bacteria that are enzymatically equipped to degrade these substrates, but also that of gut bacteria that will take advantage of the physicochemical changes associated with the presence of fibers (e.g., acid environment) and/or benefit from the intermediate products or metabolites arising from the fiber degradation. The influence of inulin on the gut microbiota is particularly well documented. Data obtained from mouse models as well as from studies of humans showed that inulin consumption is associated with the expansion of bacteria that are described as conferring health benefits and with a reduction in pathobionts [34][35][36][37]. Modulation of the host lipid metabolism is also associated with inulin consumption. Indeed, effects of inulin on triglyceride and cholesterol blood levels have been reported, but these findings are still controversial [38][39][40]. In addition, we recently showed that supplementation of a low-or high-fat diet with inulin affects the fatty acid content of mouse liver [34]. Although no direct causal relationship has been established, some inulin-induced changes in the gut microbiota were correlated with modification of the expression of genes encoding enzymes involved in fatty acid biosynthesis [34]. The aim of this study was to investigate whether dietary intake of inulin affects the Pl content of the brain. To this end, mice were exposed to a diet supplemented with either cellulose or inulin. The abundance and the diversity of Pls were explored in the liver and the cortex of mice through gas and liquid chromatographic techniques. The expression levels of the key enzymes involved in Pl biosynthesis and cleavage/degradation were also determined.

Mice and Diets
For this study, 5-week-old male C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Mice were housed at Georgia State University, Atlanta, GA, USA until euthanasia under institutionally approved protocols (Institutional Animal Care and Use Committee IACUC #A18006). Mice were maintained on 12 h light:dark cycles with ad libitum access to food and water. After 1 week of acclimation, mice were randomly divided into two groups: a control group (CTRL; n = 12) received a purified diet supplemented with 50 g cellulose/kg (Research Diet; #D12450J) and an inulin group (INU; n = 11) received a purified diet supplemented with 200 g inulin/kg (Research Diet; #D13081108) [34]. Cellulose as a source of fiber is generally poorly fermented by the gut. The diet containing cellulose served as a control. The source of inulin was chicory (average degree of polymerization ≥ 23; Orafti ® HP; BENEO-Orafti, Tienen, Belgium). Mice were maintained on these respective diets for 11 weeks. Blood was collected by retrobulbar venous plexus puncture in heparinized tubes and plasma was isolated after centrifugation (1800× g, 10 min, 4 • C). They were then euthanized by cervical dislocation and the cortex and liver were collected.

Lipid Extraction and Determination of Fatty Methyl Ester and Dimetyl Acetal Profiles
Total lipids from cortex, plasma and livers were extracted using Folch's procedure [41]. Boron trifluoride in methanol was used for transmethylation [42]. Hexane was used to extract fatty acid methyl esters (FAMEs) and dimethyl acetals (DMAs). Analyses were performed on a GC Trace 1310 (Thermo Scientific, Illkirch, France) gas chromatograph (GC) using a CPSIL-88 column (100 m × 0.25 mm inside diameter, film thickness 0.20 µm; Agilent, CA, USA). This device was coupled to a flame ionization detector (FID). The configuration was: inlet pressure of hydrogen 210 kPa, oven temperature 60 • C for 5 min + 165 • C at 15 • C per min and upholding for 1 min, +225 • C at 2 • C per min and upholding at 225 • C for 17 min. The injector and the detector were maintained at 250 • C. Comparisons with commercial and synthetic standards enabled the identification of FAMEs and DMAs. The ChromQuest 5.0 version 3.2.1 software (Thermo Scientific, Illkirch, France) was used to process the data.
The process of identification and quantification of phospholipid species was performed on a Thermo UltiMate™ 3000 coupled to an Orbitrap Fusion TM Tribrid Mass Spectrometer equipped with an EASY-MAX NGTM Ion Source (H-ESI) (Thermo Scientific, Waltham, MA, USA).
Separation of phospholipid classes was achieved under hydrophilic interaction liquid chromatography (HILIC) conditions using a Kinetex HILIC 100 m × 2.1 mm, 1.7 µm column (Phenomenex, Sydney, Australia), with a flow of 0.5 mL/min. The mobile phase consisted of (A) acetonitrile/water (CH 3  Phospholipid species were detected by high-resolution mass spectrometry (HRMS) analysis. H-ESI source parameters were optimized and set as follows: ion transfer tube temperature of 285 • C, vaporizer temperature of 370 • C, sheath gas flow rate of 35 au, sweep gas of 1 au, auxiliary gas flow rate of 25 au. Positive and negative ions were monitored alternatively by switching the polarity approach with a static spray voltage at 3500 V and 2800 V in positive and negative mode, respectively. Mass spectra in full scan mode were obtained using the Orbitrap mass analyzer with the normal mass range and a target resolution of 240,000 (full width at half maximum (FWHM) at m/z 200), in a mass-to-charge ratio m/z ranging from 200 to 1600 using a Quadrupole isolation in a normal mass range. All mass spectrometry (MS) data were recorded using a maximum injection time of 100 ms, automatic gain control (AGC) target (%) at 112.5, radio frequency lens (%) at 50 and one microscan. An intensity threshold filter of 1.103 counts was applied.
For tandem mass spectrometry (MS/MS) analyses, the data-dependent mode was used for the characterization of phospholipid species. Precursor isolation was performed in the Quadrupole analyzer with an isolation width of m/z 1.6. Higher-energy collisional dissociation was employed for the fragmentation of phospholipid species with an optimized stepped collision energy of 27%. The linear ion trap was used to acquire spectra for fragment ions in data-dependent mode. The AGC target was set to 2.104 with a maximum injection time of 50 ms. All MS and MS/MS data were acquired in the profile mode.
The Orbitrap Fusion was controlled by Xcalibur TM 4.1 software (Thermo Scientific, Waltham, MA, USA). Data of high accuracy and the information collected from fragmentation spectra, with the help of the LipidSearch TM 2.0 software (Thermo Scientific, Waltham, MA, USA) and the LIPID MAPS ® database [44], were used for phospholipid species identification.

Gene Expression
Total RNA was extracted using TRIzol reagent (Fisher Scientific, Illkirch, France). Reverse transcription was performed with the PrimeScript RT reagent kit containing gDNA Eraser (Takara Bio Europe, Saint Germain-En-Laye, France) and using 500 ng of total RNA. Gene expression was determined by real-time polymerase chain reaction (PCR) using SYBR Green (Bio-Rad, Marnes-La-Coquette, France) and a CFX96 Real-Time PCR system (Bio-Rad, Marnes-La-Coquette, France). Hprt was used as the internal control for normalization. Fold induction was calculated with the delta-delta Ct (ddCt) method. Primer sequences are given in Table 1.

Statistical Analysis
Statistical analyses were performed using Prism 6 software (GraphPad Software Inc., San Diego, CA, USA). The non-parametric Mann-Whitney test was used to compare data from the two groups. All p values of less than 0.05 were considered statistically significant (* p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001).

Effect of Inulin on the Level of Total Pls in the Liver, in the Plasma and in the Cortex
The liver has been proposed as the primary organ of Pl biosynthesis. However, in contrast to the brain whose Pl content is very high, the hepatic level of Pls is very low due to a low storage rate and a high rate of export to other organs [8,12]. The amount of Pls in the liver and in the cortex was measured by GC-FID. Acid-catalyzed transmethylation of the aldehyde aliphatic groups from the sn-1 position of Pls resulted in the production of DMAs (DMA 16:0, DMA 18:0, DMA 18:1n-7 and DMA 18:1n-9) whose amounts could be determined concomitantly with FAMEs by GC-FID. As expected, we observed that the amount of DMAs in the liver of control mice represented only 0.06% ± 0.005% of the total FAMEs and DMAs ( The level of total Pls was also measured in the plasma. As for the liver, the mean level of DMAs in the plasma was low (0.80% ± 0.10% of total DMAs and FAMEs in CTRL mice; Figure 2b). Inulin did not modify the total amount of DMAs in this transport fluid ( Figure 2b). However, the analysis at the species level revealed that the relative abundance of the two DMA species detected in the plasma (DMA 16:0 and DMA 18:0) was modified by inulin consumption: the plasma level of DMA 16:0 was significantly decreased and that of DMA 18:0 significantly increased in the plasma of INU mice compared to CTRL mice (Figure 2c,d).
In the cortex, the amount of total DMAs represented 9.24% ± 0.11% of the total

Overview of Plasmalogen Species
In the cortex, PlsEtn are the most abundant Pls [8,45]. A total of 102 glycerophosp lipid species were identified in the cortex of control mice by liquid chromatographydem mass spectrometry method (HPLC-MS 2 ) analyses. Among them, five were al These data indicate that, despite its effects on Pl classes in the plasma, inulin had no impact on the total amount of Pls in the cortex. In the cortex, PlsEtn are the most abundant Pls [8,45]. A total of 102 glycerophospholipid species were identified in the cortex of control mice by liquid chromatographytandem mass spectrometry method (HPLC-MS 2 ) analyses. Among them, five were alkylglycerophospholipids (AKGs), which are intermediate molecules in the biosynthesis of Pls, and 16 were alkenyl-glycerophospholipids, namely, Pls ( Figure 1 and Table 2). As expected, the large majority (76.2%) of AKGs and Pls belonged to the ethanolamine subclass ( Table 2). PlsEtn represented 46.155 ± 1.303% of the overall ethanolamine glycerophospholipid species. The three most abundant PlsEtn were PE(P-18:0/22:6), PE(P-16:0/22:6), and PE(P-18:0/20:4), which represented 10.857 ± 0.530%, 5.189 ± 0.299%, and 5.001 ± 0.230% of total ethanolamine glycerophospholipids in CTRL mice, respectively ( Table 2). Table 2. Relative amounts of alkyl-glycerophospholipid and plasmalogen species in the different classes of glycerophospholipids measured in mouse cerebral cortex.

Glycerophospholipids
Relative Abundance (%) Among the other glycerophospholipids, we identified three AKG species in the choline (AKGChol) and one in the inositol subclasses (Table 2). Only one Pl species, PC(P-32:0), was detected in the class of choline glycephospholipids. This represented only 0.125 ± 0.010% of total choline glycerophospholipids ( Table 2). For each species and each glycerophospholipid class, we compared the abundance of the individual species of AKGs and Pls in the cortex of INU mice relative to that of CTRL mice (Figure 3).
Altogether, these results show that supplementation of the diet with inulin modifies the abundance of specific AKG and individual Pl species in the cortex.

Effect of Inulin on the Expression of Genes Encoding Enzymes Involved in Plasmalogen Biosynthesis
Fatty acyl-CoA reductase 1 (encoded by Far1), alkyl-DHAP synthase (encoded by Agps), and DHAP-AT/DAP-AT (encoded by Gnpat) are key enzymes involved in Pl biosynthesis ( Figure 1). As their level of expression could be a factor modulating the amount of Pl, we compared the mRNA levels encoding these enzymes in the liver (Figure 4a) and in the cortex (Figure 4b) of INU and CTRL mice. As estimated by the comparison of the DeltaCt (∆Ct), the expression levels of Far1, Agps and Gnpat in CTRL mice were significantly lower in the liver than in the cortex (Appendix A Figure A1). Diet supplementation with inulin did not modulate gene expression in either organ (Appendix A Figures 4 and A1).  Altogether, these results show that supplementation of the diet with inulin modif the abundance of specific AKG and individual Pl species in the cortex.

Effect of Inulin on the Expression of Genes Encoding Enzymes Involved in Plasmalogen Biosynthesis
Fatty acyl-CoA reductase 1 (encoded by Far1), alkyl-DHAP synthase (encoded Agps), and DHAP-AT/DAP-AT (encoded by Gnpat) are key enzymes involved in Pl b synthesis ( Figure 1). As their level of expression could be a factor modulating the amou of Pl, we compared the mRNA levels encoding these enzymes in the liver (Figure 4a) a in the cortex (Figure 4b) of INU and CTRL mice. As estimated by the comparison of DeltaCt (ΔCt), the expression levels of Far1, Agps and Gnpat in CTRL mice were sign cantly lower in the liver than in the cortex (Appendix A Figure A1). Diet supplementat with inulin did not modulate gene expression in either organ (Appendix A Figure A1 a Figure 4).

Modulation of the Fatty Acid Content of the Cortex by the Dietary Intake of Inulin
Another limiting factor that could have affected the amounts of PlsEtn PE 18:0/22:6) and PE(P-34:1) [PE(P-16:0/18:1); PE(P-18:1/16:0)] in the cortex is the bioava bility of fatty acids entering the biosynthesis of these lipid species. Therefore, we analyz the fatty acid composition of the cortex by GC-FID in INU mice compared with CT mice (Table 3). We observed that inulin supplementation has a weak effect on the sa rated fatty acid (SFA) content of the cortex, since only the abundance of two minor SF (C15:0 and C17:0) was significantly modified (Table 3). Among monounsaturated fa acids (MUFAs), a trend toward a decrease in the amount of total MUFAs of the n-7 ser (p = 0.0572) and a significant decrease in the abundance of C16:1n-7 were observed in cortex of mice exposed to inulin compared to those fed a control diet (Table 3). Howev the dietary intake of inulin modulated the abundance of several PUFAs in the cortex. T abundance of C22:5n-3, C18:2n-6 and C20:3n-6 was decreased whereas that of C22:5

Modulation of the Fatty Acid Content of the Cortex by the Dietary Intake of Inulin
Another limiting factor that could have affected the amounts of PlsEtn PE(P-18:0/22:6) and PE(P-34:1) [PE(P-16:0/18:1); PE(P-18:1/16:0)] in the cortex is the bioavailability of fatty acids entering the biosynthesis of these lipid species. Therefore, we analyzed the fatty acid composition of the cortex by GC-FID in INU mice compared with CTRL mice (Table 3). We observed that inulin supplementation has a weak effect on the saturated fatty acid (SFA) content of the cortex, since only the abundance of two minor SFAs (C15:0 and C17:0) was significantly modified (Table 3). Among monounsaturated fatty acids (MUFAs), a trend toward a decrease in the amount of total MUFAs of the n-7 series (p = 0.0572) and a significant decrease in the abundance of C16:1n-7 were observed in the cortex of mice exposed to inulin compared to those fed a control diet (Table 3). However, the dietary intake of inulin modulated the abundance of several PUFAs in the cortex. The abundance of C22:5n-3, C18:2n-6 and C20:3n-6 was decreased whereas that of C22:5n-6 and C20:3n-9 was increased in the cortex of INU mice compared to CTRL mice (Table 3). The percentage of each fatty acid methyl ester (FAME) relative to that of total FAMEs (100%) was determined. Data are expressed as mean ± SEM. Mann-Whitney test for comparison of the abundance of each fatty acid between control group (CTRL) and inulin group (INU) mice, * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. SEM, standard error of the mean.

Influence of Inulin on the Production of Lyso-Glycerophospholipids in the Cortex
A decrease in the amounts of glycerophospholipids can result from an enhanced production of metabolic intermediates termed "lyso-glycerophospholipids" that are generated by the release of the fatty acid esterified at the sn-2 position of the glycerol molecule following the action of the enzyme phospholipase A(2) encoded by the Pla2g6 gene [46].
The vinyl-ether bond of lysoplasmalogens can then be cleaved by the enzyme lysoplasmalogenase encoded by the Tmem86b gene. As estimated by the analysis of the ∆Ct levels, the expression levels of these genes was significantly higher in the liver than in the cortex (Appendix A Figure A1). We observed no modification of the expression levels of Pla2g6 and Tmem86b in liver and cortex of INU mice compared to CTRL mice ( Figure 5). generated by the release of the fatty acid esterified at the sn-2 position of the glyc ecule following the action of the enzyme phospholipase A(2) encoded by the Pl [46]. The vinyl-ether bond of lysoplasmalogens can then be cleaved by the enz plasmalogenase encoded by the Tmem86b gene. As estimated by the analysis o levels, the expression levels of these genes was significantly higher in the liver t cortex (Appendix A Figure A1). We observed no modification of the expression Pla2g6 and Tmem86b in liver and cortex of INU mice compared to CTRL mice (F Using HPLC-MS 2 , we analyzed and compared the amounts of lyso-etha glycerophospholipids in the cortex of mice fed control or inulin-supplemented d 4). In total, 21 species of lyso-phosphatidylethanolamine (LPEs) species were but no lyso form of PlsEtn was detected (Table 4). No significant modification o of total LPEs/total ethanolamine glycerophospholipids was observed in the CTRL mice compared to that of INU mice (Table 4). In addition, we observed a vidual species level that the INU diet affected the abundance of LPE 14:0 (Table Altogether, these results suggest that dietary intake of inulin is not associ an increase in LPEs in the cortex.  Using HPLC-MS 2 , we analyzed and compared the amounts of lyso-ethanolamine glycerophospholipids in the cortex of mice fed control or inulin-supplemented diet (Table 4). In total, 21 species of lyso-phosphatidylethanolamine (LPEs) species were identified but no lyso form of PlsEtn was detected (Table 4). No significant modification of the ratio of total LPEs/total ethanolamine glycerophospholipids was observed in the cortex of CTRL mice compared to that of INU mice (Table 4). In addition, we observed at the individual species level that the INU diet affected the abundance of LPE 14:0 (Table 4).
Altogether, these results suggest that dietary intake of inulin is not associated with an increase in LPEs in the cortex.

Influence of Inulin on Oxidative Stress-Related Mechanisms in the Cortex
As oxidative-stress-related molecules could cause Pl degradation by attacking their vinyl-ether bond [47,48], we compared the expression level of a set of genes involved in oxidative stress-related mechanisms: Cat encoding catalase, Gpx1 encoding for glutathione peroxidase 1, Nos2 encoding for inducible NO synthase, Sod1 encoding for superoxide dismutase (Cu-Zn), Cox-2 encoding for cyclooxygenase-2 and Sqstm1 encoding for sequestosome-1 (ubiquitin-binding protein p62). As presented in Figure 6, we did not observe any modification in the expression levels of these genes in the cortex of mice from the INU group compared to CTRL mice, suggesting that oxidative stress-related mechanisms were not modulated by inulin.

Influence of Inulin on Oxidative Stress-Related Mechanisms in the Cortex
As oxidative-stress-related molecules could cause Pl degradation vinyl-ether bond [47,48], we compared the expression level of a set of oxidative stress-related mechanisms: Cat encoding catalase, Gpx1 enco one peroxidase 1, Nos2 encoding for inducible NO synthase, Sod1 encod dismutase (Cu-Zn), Cox-2 encoding for cyclooxygenase-2 and Sqstm1 en tosome-1 (ubiquitin-binding protein p62). As presented in Figure 6, w any modification in the expression levels of these genes in the cortex of m group compared to CTRL mice, suggesting that oxidative stress-related not modulated by inulin. Figure 6. Effect of inulin on the cortex expression of gene-encoding proteins in stress-related mechanisms. Cat encodes catalase, Gpx1 encodes glutathione pe codes inducible nitric oxide (NO) synthase, Sod1 encodes superoxide dismutas codes cyclooxygenase-2, and Sqstm1 encodes sequestosome-1 (ubiquitin-bindin levels of mRNA were normalized to Hprt mRNA level for calculation of the re scripts. mRNA levels are illustrated as fold change. Data are presented in bo format (median; min. to max.). The Mann-Whitney test was used for compariso mRNA between CTRL and INU mice.

Discussion
The brain is highly enriched in Pls, where they are essential in mai

Discussion
The brain is highly enriched in Pls, where they are essential in maintaining structure (e.g., myelination) and homeostasis (e.g., anti-oxidative properties, regulation of inflammation) as well as the functioning of specific processes (e.g., neurotransmission) [16]. It is now well recognized that the lipid composition of the brain is modulated by the lipid composition of the diet [49][50][51][52][53]. Thanks to the efforts made over the past decades to understand the link between diet, gut microbiota, and host metabolism, it has become evident that diet modulates not only host lipids by bringing lipids and their precursors to the host but also by acting through the gut microbiota [54]. A body of evidence indicates that the gut microbiota influences different aspects of host lipid metabolism as well as the lipid composition of organs, including that of the brain [27,[55][56][57][58]. In this study, we investigated the effect of inulin, a soluble dietary fiber with prebiotic properties, on the content and composition of Pls in the brain.
Determination of the DMA profile in the cortex of mice fed a control diet revealed that Pls represent approximately 9.3% of total fatty acids in this brain structure and that they are distributed into four classes (DMA 16:0, DMA 18:0, DMA 18:1n-7 and DMA 18:1n-9), with DMA 18:0 and DMA 16:0 being the most abundant. These results are in agreement with previous studies [59]. We found that the dietary supplementation with inulin did not alter the DMA content or the distribution of DMAs into the different classes, suggesting that dietary intake of this prebiotic does not affect the whole Pl content of the cortex, or the distribution of Pls according to their sn-1 position.
In addition to the fatty alcohol moiety linked by a vinyl-ether bond at the sn-1 position (whose trans-methylation yields the DMA derivatives), the diversity of Pl species is also ensured by the fatty acid esterified at the sn-2 position as well as by the polar head group at the sn-3 position of glycerol. Analysis by liquid chromatography coupled to MS/MS of the diversity of glycerophospholipids in the cortex of mice enabled the identification of five AKG species that are intermediate metabolites of Pl synthesis, as well as 16 species of Pls. As reported in other studies, we observed that most of them (76.2%) were PlsEtn [8]. It has been shown in the context of PlsEtn deficiency that the level of PEtn is adjusted to keep the level of PlsEtn + PEtn constant [60]. In our study, no modification in the total amount of PlsEtn or PEtn was observed in the cortex of mice fed an inulin-supplemented diet compared to those fed a control diet. However, two Pl species were affected by supplementation of the diet with inulin, namely, PE(P-34:1) [PE(P-16:0/18:1); PE(P-18:1/16:0)] and PE(P-18:0/22:6), the latter being the most abundant Pl species of the cortex (10.9% of the PlsEtn). Their abundance was decreased in the cortex of mice fed a diet supplemented with inulin compared to those fed a control diet.
The inulin-dependent effect on PlsEtn could have deleterious effects on the brain tissue since PE(P-18:0/22:6) constitutes a major reservoir of C22:6n-3 (DHA). Indeed, DHA and its derivatives are essential for the development and maintenance of brain structure and function [61]. Epidemiological studies also support a link between dietary intake of DHA and the development of brain diseases and disorders such as AD [61,62]. In addition, decreased amounts of PE(P-18:0/22:6) and PE(P-16:0/18:1) have been reported in the cerebrum of patients with AD [63]. Potential harmful effects of inulin have already been described. Dietary intake of inulin has been shown to aggravate colitis, exacerbate atherosclerosis, enhance hepatic inflammation and fibrosis, disturb hepatic and bile acid metabolism, and cause hepatocellular carcinoma in specific genetic contexts associated with dysbiosis [39,[64][65][66][67]. In addition, we have recently shown that although inulin prevents some of the alterations in the hepatic fatty acid metabolism caused by chronic consumption of a high-fat diet (HFD), it also exacerbates others [34]. Indeed, inulin consumption prevented the HFD-induced increase in C16:1n-9 and C20:3n-6 as well as the HFD-induced modulation of expression and/or activity of enzymes involved in fatty acid biosynthesis (Elovl2, Elovl5 and FADS2) in mouse liver. However, this dietary fiber also exacerbated the HFD-induced increase in the hepatic amount of C17:0.
To expand our understanding of the mechanisms underlying the inulin-dependent decrease in some Pl species in the cortex, we explored several hypotheses. Since the liver has been proposed as the primary site of Pl biosynthesis, we investigated whether alterations of Pls in the cortex could have a hepatic origin. To this end, we evaluated the DMA content of the liver as well as the expression level of genes encoding key enzymes involved in the initial three steps of Pl biosynthesis (Far1, Gnpat and Agps) and compared them between mice fed a control diet and those fed an inulin-supplemented diet. Only DMA 16:0 was detected in the liver of mice at very low levels, which is consistent with previous studies [8]. No effect of inulin was observed, neither on the DMA content nor on the expression levels of Far1, Gnpat and Agps in the liver. This is in line with previous results showing no modification in the expression levels and activities of enzymes involved in the biosynthesis of fatty acids following inulin supplementation [34]. To go further in the exploration of a hepatic origin for the changes we observed in the Pl content of the cortex, we analyzed the Pls in the plasma. Indeed, we previously showed that the inulin supplementation, as we provided in the diet of this study, induced changes in the composition of the gut microbiota [34], and changes in the composition of the gut microbiota following inulin consumption have been associated with serum Pl levels [68]. No modification of the total DMA content was observed in the plasma of mice fed a diet supplemented with inulin. However, intra-class modifications were observed: the relative abundance of DMA 16:0 was decreased and counterbalanced by an increase in DMA 18:0. Whereas this result might account for the decrease in PE(P-34:1) [PE(P-16:0/18:1); PE(P-18:1/16:0] in the cortex, it does not explain that in PE(P-18:0/22:6). Altogether, these data suggest that it is unlikely that the Pl changes observed in the cortex have an extra-brain or hepatic origin.
The bioavailability of the fatty acids in the cortex required for their biosynthesis was also analyzed. We observed no modification in the abundance of C16:0, C18:0, C18:1n-7, C18:1n-9 or C22:6n-3 in the cortex of mice fed an inulin-supplemented diet. However, the level of docosapentaenoic acid (DPA) from the n-3 series (C22:5n-3), which is an intermediate between eicosapentaeinoic acid (EPA, C20:5n-3) and DHA (C22:6n-3), was decreased. Finally, as the decrease in the abundance of some PlsEtn could also be the consequence of their hydrolysis, the expression levels of enzymes involved in Pl cleavage/degradation and the level of lyso species were evaluated. No modification of the expression level of Pla2g6 and Tmem86b genes, encoding phospholipase A(2) and lysoplasmalogenase, respectively, was observed in mice fed an inulin-supplemented diet. Another cause of Pl degradation could be an attack on the vinyl-ether bond by oxidative stress-related molecules [47,48]. To test this hypothesis, the level of oxidative stress as well as the amount of oxidized derivatives of Pls should be evaluated. However, our results showed that dietary intake of inulin did not modify the expression level of a set of genes involved in oxidative stressrelated mechanisms (Cox-2, Cat, Gpx1, Sod1, Nos2 and Sqstm1) in the cortex. In addition, no lyso form of PlsEtn was detected in the cortex of mice fed a control diet or an inulinsupplemented diet and no modification of the ratio of LPEs/PEs was observed in the cortex of mice exposed to inulin. Taken together, these data suggest that the dietary intake of inulin does not enhance glycerophospholipid hydrolysis. However, as intermediate products arising from Pl degradation may only have a short-lived existence, further experiments such as assessment of phospholipase A(2) and lysoplasmalogenase activities are needed to rule out the existence of an impact of inulin consumption on PlsEtn degradation.
Finally, despite the use of compositionally controlled diets, we cannot exclude that the amount of fiber consumed by mice fed an inulin-supplemented diet was different to that of the control mice that received a cellulose-containing diet. Indeed, we have previously reported that inulin supplementation can slightly decrease food consumption, very likely linked to the energy provided by fermentable fiber compared to non-fermentable fiber [32]. More importantly, the dose of inulin used in the current study is relatively high and cannot be transposed to human nutrition. Hence, future studies appear warranted to investigate the effect of lower doses of inulin on cortex Pls.

Conclusions
In this study, we showed that dietary supplementation with inulin do not modify the global amount of Pls in the cortex of mice but affects its content at the species level. In particular, dietary intake of this prebiotic induces a decrease in the abundance of the most widely represented PlsEtn species, PE(P-18:0/22:6), which represents a major reservoir of DHA, a fatty acid essential for brain development and function. This study joins others that suggest inulin may have deleterious effects. The consequences of these alterations on the physiology and the functioning of the brain, as well as the molecular mechanisms that link inulin/gut microbiota and Pl levels in the brain, remain to be elucidated.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.