High fat / high cholesterol diet does not provoke atherosclerosis in the ω3-and ω6-polyunsaturated fatty acid synthesis–inactivated Δ6-fatty acid desaturase–deficient mouse

Objective An increased ω6/ω3-polyunsaturated fatty acid ratio in the current Western diet is regarded as a critical epigenetic nutritional factor in the pathogenesis of several human lifestyle diseases, metabolic syndrome, cardiovascular disease, the central nervous system and the female and male reproductive systems. The impact of nutrient ω3-and ω6-PUFAs in the pathogenesis of dyslipoproteinemia and atherosclerosis has been a topic of intense efforts for several decades. Cellular homeostasis of the ω3-and ω6- PUFA pool is maintained by the synthesis of ω3-and ω6-PUFAs from essential fatty acids (EFA) (linoleic and α-linolenic acid) and their dietary supply. In this study, we used the auxotrophic Δ6-fatty acid desaturase- (FADS2) deficient mouse (fads2−/−), an unbiased model congenial for stringent feeding experiments, to investigate the molecular basis of the proposed protective role of dietary ω3-and ω6-PUFAs (Western diet) in the pathogenesis of multifactorial dyslipoproteinemia and atherosclerosis. We focused on the metabolic axis—liver endoplasmic reticulum (ER), serum lipoprotein system (Lp) and aorta vessel wall. Furthermore, we addressed the impact of the inactivated fads2-locus with inactivated PUFA synthesis on the development and progression of extended atherosclerosis in two different mouse mutants with disrupted cholesterol homeostasis, using the apoe−/− and ldlr−/− mutants and the fads2−/− x apoe−/− and fads2−/− x ldlr−/− double mutants. Methods Cohorts of +/+ and fads2−/− mice underwent two long-term dietary regimens: a) a PUFA-free standard chow diet containing only EFAs, essential for viability, and b) a high fat/high cholesterol (HFHC) diet, a mimicry of the human atherogenic “Western” diet. c) To study the molecular impact of PUFA synthesis deficiency on the development and progression of atherosclerosis in the hypercholesterolemic apoe−/− and ldlr−/− mouse models fed PUFA-free regular and sustained HFHC diets, we generated the fads2−/− x apoe−/− and the fads2−/− x ldlr−/− double knockout mutants. We assessed essential molecular, biochemical and cell biological links between the diet-induced modified lipidomes of the membrane systems of the endoplasmic reticulum/Golgi complex, the site of lipid synthesis, the PL monolayer and neutral lipid core of LD and serum-Lp profiles and cellular reactions in the aortic wall. Results ω3-and ω6-PUFA synthesis deficiency in the fads2−/− mouse causes a) hypocholesterolemia and hypotriglyceridemia, b) dyslipoproteinemia with a shift of high-density lipoprotein (HDL) to very low-density lipoprotein (VLDL)-enriched Lp-pattern and c) altered liver lipid droplet structures. d) Long-term HFHC diet does not trigger atherosclerotic plaque formation in the aortic arc, the thoracic and abdominal aorta of PUFA-deficient fads2−/− mice. Inactivation of the fads2−/− locus, abolishing systemic PUFA synthesis in the fads2−/− x apoe−/− and fads2−/− x ldlr−/− double knockout mouse lines. Conclusions Deficiency of ω3-and ω6-PUFA in the fads2−/− mutant perturbs liver lipid metabolism, causes hypocholesterolemia and hypotriglyceridemia and renders the fads2−/− mutant resistant to sustained atherogenic HFHC diet. Neither PUFA-free regular nor long-term HFHC-diet impacts the apoe- and LDL-receptor deficiency–provoked hypercholesterolemia and atherosclerotic plaque formation, size and distribution in the aorta. Our study strongly suggests that the absence of PUFAs as highly vulnerable chemical targets of autoxidation attenuates inflammatory responses and the formation of atherosclerotic lesions. The cumulative data and insight into the molecular basis of the pleiotropic functions of PUFAs challenge a differentiated view of PUFAs as culprits or benefactors during a lifespan, pivotal for legitimate dietary recommendations.


INTRODUCTION
Essential fatty acids (EFA) linoleic and a-linolenic acid and derived long-chain (LC) and very long-chain (VLC) u3-and u6-polyunsaturated fatty acids (PUFAs) are pivotal structural elements in the architecture of all mammalian membrane systems and precursors of structurally and functionally diverse signaling molecules operative in pathways regulating lipid and energy metabolism [1]. Homeostasis of the cellular pool of u3-and u6-PUFAs is maintained by dietary supply and the transformation of EFAs to LC-and VLC-u3-and u6-PUFAs in an orchestrated sequence at the tetrameric chain-elongation and trimeric desaturase complexes in the endoplasmic reticulum, the dominant site of PUFA synthesis. A variety of genetically conditioned dyslipoproteinemias, notably hypercholesterolemia and hypertriglyceridemia, play a crucial role in the pathogenesis of fatal atherosclerosis in coronary heart disease (CHD). For understanding the mechanism underlying the molecular pathogenesis, two genetically defined murine models have been described: the homozygous apoeÀ/À and ldlrÀ/À mouse mutants. They share the same or similar characteristic pathogenetic features leading to atherosclerosis in humans, despite their different genetic background.
Increased consumption of u6-PUFAs in Western diets has led to a considerable imbalance of the u3/u6-PUFA ratio over the last several decades and is regarded as critical in the pathogenesis of several human genetic and lifestyle diseases: cardiovascular disease, dyslipoproteinemia, obesity, neurodegenerative diseases and infertility. The role of nutrient u3-and u6-PUFAs as epigenetic factors in the pathogenesis of atherosclerosis has been assessed in numerous nutritional trials [2,3,5,6]. These studies led to dietary recommendations from the American Heart Association and FAO/WHO expert consultation on fats and fatty acids in human nutrition rating the impact of u3-and u6-PUFAs on physiological and pathophysiological conditions at the two highest levels of strength of evidence, "Convincing" and "Probable," to "conclude that u6-and u3-PUFAs affect major health and disease outcomes." u3-PUFAs are "convincing" in lowering fatal CHD events and u6-PUFAs are "probable" for lowering risk of metabolic syndrome components and diabetes [2]. However, due to the complexity and critical parameters, the validity of the outcome of the numerous dietary studies in different species (including humans) has provided only limited insight into the molecular basis of the underlying pathophysiology. Two recent extensive, contradictory reviews of randomized interventional trials [3,4] witness the current need for strategies for the molecular analysis of well-designed nutritional experiments in unbiased preclinical animal models to assess the role of u3-PUFAs in primary and secondary prevention of cardiovascular disease. To study underlying genetic and epigenetic mechanisms of this polygenic disorder, genetically defined null allelic murine strains, including the LDLR-deficient mutant (ldlrÀ/À), lacking the LDL-receptor for LDL uptake, and the ApoE-(apoeÀ/À) mutant, devoid of the remnant and VLDL-ligand for cholesterol clearance by the LDL and LRP (LDL-receptor related), have been generated by transgenic techniques [5]. These two strains rapidly develop extensive atherosclerosis throughout the entire aorta and are congenial models for nutrition studies implementing atherogenic HFHC ("Western") diets [5e18]. In this study, we used the unbiased, genetically defined auxotrophic fads2À/À mouse mutant [6] in sustained dietary studies to bypass potential confounding factors and to elaborate molecular roles of u3and u6-PUFA deficiency in the multifactorial pathogenesis of atherosclerosis. The molecular bases of several phenotypic features of the seemingly unrelated multi-variant phenotype of the fads2À/À mouse have been uncovered so far.
Loss of the D6-desaturase in the fads2À/À mutant leads to a) shutdown of the biosynthesis of PUFAs from their precursor EFAs, with the complete loss of these essential structural elements in the architecture of membrane lipid bilayers; b) D6-desaturase (fads2) gene inactivation activating an aberrant pathway of linoleic acid (18:2) metabolism, resulting in the synthesis of the uncommon eicosa-5Z,11Z,14Z-trienoic acid (20:3 5,11,14 ) end product; 20:3 5,11,14 systemically substitutes all PUFAs as single surrogates in the fatty acid profiles of phospholipid and sphingolipid classes; c) the conditioned loss of precursor u6-AA in the eicosanoid synthesis prohibiting thromboxane A2 and prostacyclin synthesis, required for platelet aggregation and thrombus growth in primary hemostasis of the clotting cascade, and leading to prolonged bleeding time and loss of leukotriene LTB4 synthesis in stimulated macrophages [6]; d) the altered lipid bilayer architecture of Sertoli and germ cells of testes and granulosa cells of ovaries disrupting intercellular junction systems and causing male and female infertility [6,7].
We focused our study on the impact of suppressed biosynthesis of u3and u6-PUFAs on the development of atherosclerosis in the fads2À/À mouse. To obtain insight into the underlying molecular role in the prevention or development of atherosclerotic lesions, we compared cohorts of male Original Article and female C57BL/6N and fads2À/À mice by exposition to two feeding regimens, which were started after weaning and sustained during the lifetime: 1) a standard chow diet, which contained the essential fatty acids (EFA), the precursors in the PUFA synthesis control C57BL/6N mice, and 2) a PUFA-free but EFA-containing HFHC ("Western") diet. The results of these studies, which combine the unbiased genetic model and stringent feeding conditions, unveiled molecular links between the suppression of u3-and u6-PUFA synthesis and prolonged PUFA-free regular and high fat/high cholesterol diets, perturbed membrane lipidomes and lipid metabolism of liver, lipid droplets and lipoproteins. They led to the surprising observation that suppression of PUFA-synthesis in the fads2À/À mouse on neither a PUFA-free regular diet nor a prolonged PUFA-free HFHC ("Western") diet triggered atherosclerotic plaque formation in the aortic arc, the thoracic and abdominal aorta. The fads2À/À mutant proved resistant to the development of atherosclerotic lesions by sustained atherogenic diet. Beyond this study, we pursued potential epistatic functions of the inactivated fads2 locus and the PUFA deficiency in the development of atherosclerotic lesions in the apoeÀ/À and ldlrÀ/À mutants, two model systems frequently used in the study of the multifactorial molecular pathogenesis of atherosclerosis, which is characterized by dysregulated cholesterol and lipoprotein metabolisms and completely different from our system. We generated the fads2À/À x apoeÀ/À and fads2À/À x ldlrÀ/À double mutants for long-term feeding experiments with a regular chow and HFHC diet to straightforwardly demonstrate the impact of suppressed fads2À/À activity by comparing plaque formation to that in apoeÀ/À and ldlrÀ/À mutants. Suppression of PUFA synthesis in fads2À/À x apoeÀ/À and fads2À/ À x ldlrÀ/À double knockout mice on normal chow or "Paigen" diet showed that neither enhanced nor reduced apoE and LDL-receptor deficiency mediated atherosclerotic plaque formation. Immuno-histochemical observations strongly suggest the deleterious role of PUFAs as highly vulnerable chemical targets of autoxidation, releasing reactive aldehydes (e.g., HNE (4-hydroxynonenal)) and covalently modifying proteins to novel autoantigens, which trigger inflammatory responses in atherosclerotic lesions. Our study demonstrates the necessity of a scrutinized view of the pleiotropic functions of u3-and u6-PUFAs as molecular culprits or benefactors during a lifespan before they are included in legitimate dietary recommendations.

Mouse mutants, generation and genotyping
Generation and genotyping of fads2þ/À and fads2À/À mice have been described before [6]. Fads2þ/À mice were back-crossed 10 times into the C57BL/N6 genetic background. Fads2þ/þ siblings of the heterozygous crossings were used as (þ/þ). ApoeÀ/À and ldlrÀ/ À mice were crossed with fads2þ/À mice to breed homozygous fads2À/À x apoeÀ/À and fads2À/À x ldlrÀ/À double mutants. Genotypes were led by PCR analysis of ear-punch DNA. The following oligonucleotide primers were used for genotyping of the apoeÀ/À and apoeÀ/À x fads2À/À double knockout loci: apoeÀ/À: Animals were housed in the specific pathogen-free (SPF) barrier mouse facility of the Center of Molecular Medicine (CMMC) with a 12h light/dark cycle and free access to water and chow. Cohorts of genderand weight-matched and mutant mice were used in this study. The animal studies followed ARRIVE Guidelines [8]. Animal breeding and test protocols followed the principles and practices outlined in the Guide for the Care and Use of Laboratory Animals. The procedures were approved by the Institutional Animal Care and Use Committee of the University of Cologne and by the Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen.

Cell fractionation of liver, isolation of lipid droplets and serum lipoproteins
Microsomal fraction of the 0.25M sucrose homogenate of liver was isolated as the 100,000Âg sediment of the 12,000Âg supernatant. Liver lipid droplets (LDs) were isolated from liver by established density gradient centrifugation adapted to liver homogenates [10]. In brief: mouse liver was homogenized in 2 mL disruption buffer (25 mM Trise HCl, pH 7.4; 100 mM KCl, 5 mM EGTA, 1 mM EDTA) and diluted with 2 mL 1.098M sucrose in disruption buffer. The homogenate was spun for 10 min at 1500Âg at 4 C. The post-nuclear supernatant fraction was overlaid with 2.5 mL 0.27M and 0.13M sucrose and filled with top solution (25 mM TriseHCl, pH 7.4, 1 mM EGTA and 1 mM EDTA) in the 11 mL centrifuge tube and spun at 36,000 rpm in the Beckman SW41 Ti rotor at 4 C for 2 h. The top buoyant LD fraction was removed with a bent Pasteur pipette and suspended in lysis buffer, total volume 1 mL, and 10 more 1 mL fractions were isolated. Serum lipoproteins were separated by agarose electrophoresis and cholesterol concentrations, determined using the Sebia HYDRASYS agarose gel electrophoresis system (Sebia, Inc. Norcross, Georgia, 30093, USA).

Laboratory measurements
Concentrations of blood glucose, serum insulin, triglycerides and total cholesterol of mice fasted overnight were determined by standard colorimetric assays [11].

Lipidome analysis
Total lipids from liver, subcellular fractions, LDs and serum of cohorts (n ¼ 5) of adult (4-month-old) þ/þ and fads2À/À female and male mice were extracted by homogenization in an Ultraturrax in 10 volumes of chloroform/methanol (CHCl 3 /CH 3 OH) at 2:1 (v/v) and re-extracted with CHCl 3 /CH 3 OH at 1:1 (v/v) and CHCl 3 /CH 3  Triple TOF 5600 mass spectrometer, was used. The system was equipped with a Thermo Fisher Scientific (Dreieich, Germany) Accucore C8 (2.6 mm particle size, 50 Â 3 mm) analytical column, and mobile phases consisted of aqueous formic acid 0.2% (pH 2) (solvent A) and acetonitrile as organic modifier (solvent B) for optimal ESI conditions. Thereby, the gradient was held for 0.5 min and then decreased from 50% A to 0% A with 0.325 mL/min within 6 min. The column was regenerated by washing for 1 min and equilibrated with 100% B for 3 min. The mass spectrometer was calibrated frequently (after 10 injections) via the Duo Turbo-V-Ion source by a calibrant delivery system containing the manufacturer's probes for positive ionization. The nitrogen for the ion source as well as collision gas supply was delivered by the nitrogen generator (CMC, Eschborn, Germany). Product ion experiments were acquired by isolating the respective [MþH] þ precursor ions in the quadrupole (unit resolution) and performing collision-induced fragmentation in the collision cell.

GC/MS analysis of fatty acid profiles of PL and NL classes
Total lipid extracts and isolated phospho-and neutral lipids were transesterified with 5% HCl-methanol at 80 C for 1 h. One volume of water was added, and fatty acid methyl esters (FAMEs) were extracted with hexane and concentrated under nitrogen. When indicated, double bond positions were determined in DMOX-derivatized fatty acids as described before [11]. FAMEs were separated, identified and quantified by combined gaseliquid chromatography/electrospray ionization mass spectrometry (ESI-MS) on an Agilent (Waldbronn, Germany) 6890/5973N instrument equipped with an HP-5MS fused silica column (length 17 m, i.d. 0.25 mm, film thickness 0.25 mm) or on a Carlo Erba Instrument Model GC8000. Samples of FAMEs and 2,2dimethyloxazoline derivatives were injected in a 10:1 split mode into the mass spectrometer, which was operated on full scan mode over a mass range of 50e500 u, and electron ionization (EI) was utilized at 70 eV. The injector temperature was set to 300 C.

Protein analysis
Freshly dissected liver or subcellular fractions, aliquots of the LD fraction and serum of þ/þ and fads2À/À mice were mechanically dispersed in lysate buffer containing protease inhibitor cocktail (Complete; Roche). Protein concentrations were measured using a Pierce BCA protein assay kit (Thermo Scientific).
2.8. Gene expression analysis by qRT-PCR RNA was isolated from control and fads2À/À liver using Trizol, Invitrogen. 10 mg of total RNA was reverse transcribed using a Transcriptase kit (Life Technologies, Darmstadt, Germany). Primer pairs used in quantitative PCR-reactions are listed in Hgprt was used as internal standard. qRT-PCR reactions were performed with the ABI Prism 7900HT, employing a 96-well format and the Fast SYBR Green Master Mix, Applied Biosystems, following the manufacturer's protocol. Data analysis was performed using the 2-DDCt method.

Histology and immunohistochemistry
Mice were perfused from the left ventricle with cold PBS, then with PBS and PBS-buffered 4% paraformaldehyde. Tissues were fixed and processed for light-and immunofluorescence microscopy as described before [12]. For the assessment of atherosclerotic plaques, the entire freshly prepared aorta was isolated from the arch to the iliac bifurcation, adventitia and adipose tissue were thoroughly removed and the aorta was opened lengthwise, pinned flat on a mossy rubber pad and fixed in 10% neutral buffered formalin overnight. The entire aorta was stained for 10 min with oil red and en face preparations were digitally photographed and quantified using the Zeiss Imager M1 with the Axiovision software. Cryo-and paraffin-embedded sections (3 mm) of liver and rolled aorta were stained for lipid deposits with oil red. Aortic plaque components were immunohistochemically characterized using anti-Apo E from Acris Antibodies GmbH (Cat# BP2046,

RESULTS
We investigated through long-term We then elaborated the relevance of the inactivated fads2À/À locus in the genome of the fads2À/À x apoeÀ/À double mutant for the development and progression of atherosclerotic lesions in the aorta. Cohorts of age-matched male and female control C57BL/6n, fads2À/ À, apoeÀ/À and fads2À/À x apoeÀ/À mice were subjected to a) regular chow and b) the HFHC diet for a period of 4 months. c) To enhance the formation and development of aortic lesions in apoeÀ/À and ldlrÀ/À mutants, a HFHC diet supplemented with 0.5% sodium cholate and 1% cholesterol ("Paigen" diet) was administered for a period of eight months after weaning.
3.2. Impact of u3-and u6-PUFAefree regular and HFHC ("Western") diet on liver lipid metabolism Cohorts of þ/þ, fads2À/À, apoeÀ/À, fads2À/À x apoeÀ/À and fads2À/À x ldlrÀ/À were studied in u3-and u6-PUFAefree regular and HFHC ("Western") diets for a period of four months after weaning. All animals survived the sustained, long-term chow-fed and HFHC-fed dietary interventions. Body weight of male and female cohorts, independent of the genotype, differed by approximately 25%. At the end of the feeding period the body weight of þ/þ mice exceeded that of the mutant strains by 20%.
Carbohydrate metabolism of þ/þ and fads2À/À on sustained regular diet ( Figure SI 1E) and male and female þ/þ, fads2À/À, apoeÀ/À and fads2À/À x apoeÀ/À mice on the HFHC diet at p120 (Figure SI G) remained unaffected by the absence of PUFAs, as indicated by unaltered fasting blood glucose concentration, and serum insulin concentration of þ/þ and fads2À/À mice on regular chow ( Figure SI 1F). Analysis of total serum cholesterol and triglycerides of fads2À/À mice on nd (normal diet) revealed hypocholesterolemia and hypotriglyceridemia: C and TG concentration were reduced to half of those of þ/þ control mice (Figure 2A,C). The HFHC diet increased serum cholesterol to twice the concentration of þ/þ, but did not change the low C concentration in cohorts of male and female fads2À/À mice ( Figure 2B,D). Inactivation of the fads2 locus in male and female fads2À/À x apoeÀ/À mice elevated serum C and TG concentrations to two-and threefold those in apoeÀ/À mice, respectively.
Our previous study has demonstrated the inactivation of the clotting system by the deprivation of arachidonic acid (20:4), the precursor of antithrombotic thromboxane A2 synthesis, leading to prolonged bleeding time, another hallmark of the fads2À/À phenotype [6]. Serum concentration of TXB2, the stable metabolite of TXA2, was not measurable by the immune-assay in the fads2À/À serum. Average bleeding times of nd-fads2À/À was 6e8 min, compared to 2e 2.5 min in þ/þ mice ( Figure 2E). HFHC diet had no impact on the prolonged bleeding time of the fads2À/À and fads2À/À x apoeÀ/À mutants ( Figure 2F). Microscopic images of sections of liver of male and female þ/þ, fads2À/À, apoeÀ/À and fads2À/À x apoeÀ/À mice maintained on the regular control and the HFHC diet were stained with oil red for quantitation of lipid deposition in lipid droplets (LD) (Figure 3AeD). Livers of the four different genotypes revealed different degrees of hepato-steatosis. We improved this quantification by immunestaining lipid droplets in paraffin sections of liver using the anti-perilipin2 antibody that recognizes LD-specific perilipin2 integrated into the LD membrane (Figure 3EeI). The sharp demarcation provided a detailed view of size, number, intra-and extracellular localization and LD. Quantitation of LD number and size is summarized ( Figure 3J,K).
3.3. Absence of u3-and u6-PUFAs perturbs membrane phospholipid bilayer of ER and of LD phospholipid monolayer of fads2À/À mouse liver A molecular hallmark of the fads2À/À phenotype is the systemic substitution of u3and u6-LCePUFA by surrogate 20:3 5,11,14 (sciadonic acid), synthesized from linoleic acid by an aberrant pathway [11]. We first investigated the impact of a sustained long-term regular chow dietary regimen on liver lipid metabolism of þ/þ and fads2À/À cohorts. Comparative experiments focused on the molecular architecture of the membrane lipid bilayer of the ER (microsomes), the site of lipid synthesis, the assembly of serum lipoproteins, and the ER-derived lipid monolayer of lipid droplets, the intracellular energy storage particles of the liver. ER and LDs were isolated by established gradient ultra-centrifugation [10]. PL and NL classes of the ER membrane lipid bilayer and LD lipid monolayer were separated by HPTLC for MS/MS analysis of the profile and stoichiometry of the hydrophobic DAG core species of individual PL classes (Figure 4), and their fatty acyl substituents were further characterized by GC/MS. NL classes of ER were isolated by HPTLC for GC/MS analysis (SI Figure 3). The stoichiometry of PL classes was very similar in the lipidomes of the ER and LD of þ/þ and fads2À/À liver ( Figure 4A,B).  Unlike the lipidome of the ER membrane, the LD-phospholipidome of þ/þ and fads2À/À liver was devoid of SM ( Figure 4B). Absence of SM and accumulation of PS, PI and PE further indicate the outer cytoplasmic lipid leaflet of the asymmetric ER lipid bilayer as donor of the LDePL monolayer. Regular chow raised only the TG concentration in the NL profile of the globular hydrophobic core of LDs in fads2À/À liver. DG and CE were present in minor concentrations and C was not detectable. Stoichiometry of CE, TG, C and DG in the NL fractions of ER and LDs of þ/þ and fads2À/À liver on regular diet were closely related, as were the fatty acid patterns of CE and TG. LDs of þ/þ and fads2À/À liver of HFHC-fed mice contained only CE and TG, the CE concentration being increased fourfold and TG concentration reduced threefold in fads2À/ À LDs ( Figure 6J). Gene expression of key transcription factors and enzymes in liver of þ/ þ and fads2À/À mice on regular, PUFA-deficient chow ( Figure 6M) indicated elevated steady-state mRNA concentration of lipogenesis regulating srebp1c, fatty acid synthase fas and remarkable overexpression of stearoyl-CoA desaturase (scd1) and suppression of the P450 cytochrome hydroxylase (cyp4A) involved in the metabolism of PUFAs leading to physiologically important metabolites.

PUFA deficiency impairs lipid droplet size and fusion
We next expanded the lipidomic analyses by immuno-histochemical ( Figure 7) and Western blot analysis of key proteins intimately involved in LD formation, uptake of fatty acids, and genes of enzymes regulating TG synthesis and lipolysis ( Figure 8). Western blot analysis of lysates of liver and LDs included the dominant LD surface-bound perilipins, members of the PAT protein family, PLIN1 (Perilipin1), PLIN2 (Perilipin-2, ADRP, Adipophilin) and PLIN3 (TIP47, tail interacting protein), which, associated with ATGL (adipose triglyceride lipase) and DGAT-1 (diacylglycerol acyl transferase-1), regulates LD formation, growth and fusion LD formation ( Figure 8) [14,15]. DGAT1 catalyzes the final step in TG synthesis and storage in hepatocytes and stellate   Sections of liver of PUFA-deficient male and female fads2À/À mice on regular chow ( Figure 7A,B) were free of LDs and, in response to HFHC diet ( Figure 7C,D), stored TG in micro-vesicular LD, indicated by LDspecific PLIN2 expression. We included small Rab5 GTPase, known as a key player in the regulation of LD multi-protein of membrane assembly, fusion and cytoskeletal transport [16]. Small GTPase Rab5 docking sites, distributed across the entire plasma membrane of control liver stellate cells and responsible for the fusion of small LD, were aggregated and LD fusioneinhibited in fads2À/À liver ( Figure 7D). In addition, immunohistochemistry of liver sections of mice on the regular diet stained with anti-ATGL and anti-DGAT1 ( Figure 7G-L) clearly visualized the LD membrane harboring ATGL and DGAT1 and indicated low expression of ATGL and DGAT1 similar to PLIN2 and Rab5 in merged images of the livers of control and male and female fads2À/À mice on regular chow, stained with ATGL and DGAT1 antibodies ( Figure 7G,H). Male and some female fads2À/À mice on the HFHC diet had developed micro-lipid droplet steatosis: (I) þ/þ, (J) fads2À/À. Strong expression was observed in (K) apoeÀ/À mice and micro-vesicular LDs and attenuated expression of ATGL and DGAT1 in (L) fads2À/À x apoe/-mice. The downregulation of PLIN2 (ADRP) and upregulation of ATGL expression in ER and LDs of fads2À/À mice were reflected in the decrease of LD-TG. Perilipin 1, located in the lipid droplet monolayer coat and known to serve as a recruitment site for lipases, blocks ATGLcatalyzed TG hydrolysis but triggers lipolysis upon phosphorylation by protein kinase A [17]. Western blot analysis (WB) of PAGE-separated protein lysates of þ/þ and fads2À/À liver and LDs revealed upregulated PLIN1 expression in fads2À/À LDs. Suppression of PLIN3 (TIP47) in fads2À/À liver correlated with the intra-hepatocyte accumulation of LDs.
Suppression of u3-and u6-PUFA synthesis elevated VLDL/LDL in female fads2À/À mice, but did not reduce the concentration of VLDL/ LDL in male and female fads2À/À x apoeÀ/À mutants ( Figure 9D). Quantification of total serum cholesterol (TC) and of C in HDL, VLDL, LDL and chylomicrons of control and fads2À/À, apoeÀ/À and fads2À/À x apoeÀ/À mice on regular and HFHC diet is summarized in (Figure 9E,F). The steady-state profile of the PL classes of total serum lipids in regular dietefed þ/þ and fads2À/À mice contained PC as main and PI as minor PL class, which remained rather stable during the feeding period, unlike that of NL lipidome, in which concentrations of C, CE and TG in serum of fads2À/À mice were attenuated to a low concentration ( Figure 10C). HFHC diet decreased the PC/SM ratio and PI to the abundance of 5 mol% in the phospholipidome; the low C, CE and TG concentrations in the NL pattern remained stable.
Species analysis by MS/MS uncovered SM species in fads2À/À Lp highly substituted with C12, C14 and C16 in addition to the prevailing 16:0 and 24:1-substituted SM species ( Figure 10K). DAG species of the major PL classes, PC, PE and PI, are characterized by the exchange of 20:4 with 20:3 in PC and PI in fads2À/À serum Lp, highlighted by the encasements (Figure 10DeG). HFHC-diet left the profile of PL-classes of þ/þ and fads2À/À lipoproteins unimpaired ( Figure 10H). The DAG core structures of PL species, however, revealed profound structural changes. HFHC diet strongly suppressed 20:3 5,11,14 synthesis, which otherwise substituted all u3-and u6-PUFAs in lyso-PC, PC, PI and PE.
Inactivation of the fads2À/À locus in the HFHC-fed diet apoeÀ/À and fads2À/À x apoeÀ/À double mutant had no impact on the high total C, CE, and TG concentrations in the profile of NL classes of serum Lps ( Figure 11C). GC/MS analyses revealed similar fatty acid profiles of total lipids of serum Lps of control ( Figure 11H), fads2À/À ( Figure 11J) apoeÀ/À ( Figure 11I) and fads2À/À x apoeÀ/À mice ( Figure 11K) on regular and HFHC diet.
3.7. PUFA deficiency has no effect on macrophage activation and inflammation in aortic lesions LDLox uptake by the endothelial lining of the aorta triggers chemotactic engulfment of monocytes and transformation into macrophages (foam cells), visible as fatty streaks. IHC with lysosomal markers LAMP1 and SMPD1 antibodies visualized only low numbers of phagosomes in macrophages within aortic lesions of þ/þ and fads2À/À mice, but there was an accumulation of lesions of the apoeÀ/À and fads2À/À x apoeÀ/À aorta ( Figure 13E, F). We further assessed the contribution of macrophages to the inflammation process by IHC using anti-NFkB, anti-TNFa and anti-TNFR1 antibodies, which did not reveal expression of these markers in Original Article 14 apoeÀ/À mice and revealed less-severe expression in fads2À/À x apoeÀ/À double mutants in the aortae of either control or fads2À/À male and female mice after the prolonged HFHC feeding period ( Figure 13 G-I).

DISCUSSION
The imbalanced ratio of u3-and u6-PUFAs in the current Western diet is regarded as a critical pathogenetic and epigenetic factor of the increasing incidence of metabolic syndrome and associated risk of developing dyslipoproteinemia and atherosclerosis in cardiovascular disease [2e4]. These conditions have been investigated in numerous complex interventional nutritional trials in human and animal models, with contradictory outcome and limited insight into the molecular pathophysiology. There is therefore a current need for unbiased preclinical animal models [3,4]. The current study uses the auxotrophic fads2À/À mouse mutant, characterized by the inactivation of D6-desaturase, the key enzyme of  u3-and u6-PUFA synthesis from EFAs linoleic and a-linolenic acid.
Bypassing potential confounding factors, this mutant proves congenial for elaborating the structural and metabolic roles of u3-and u6-PUFAs and PUFA deficiency [6,21] in C57BL/6 and fads2À/À mice in controlled long-term feeding regimens of a) regular chow and b) high fat/high cholesterol (HFHC) PUFA-free diet ("Western diet"). We describe   Our study underscores the pivotal role of the severely modified hydrophobic core structures of the ER lipid bilayer scaffold and perturbation of lipid metabolism of the liver. We concluded that the disrupted divinyl-methane (homo-allylic) rhythm in the 20:3 5,11,14 acid, the sole sn2 substituent of all u3-and u6-PUFA across the entire phospholipidome, critically impinges on the spatial arrangement of the DAG core structure and molecular architecture of the lipid bilayer, leading to asymmetry of the liver's ER membranes in the fads2À/À mutant. This interpretation is substantiated by comparative monolayer studies of 20:3 5 These finding support earlier observations that mice poorly convert, but preferentially b-oxidize, a-linolenic acid [26,27].
The current concept of membrane lipid bilayer asymmetry views SM, C and PC as abundant PL classes of the luminal leaflet and PS, PI and PE of the cytoplasmic leaflet of the ER membrane [28,29]. Our study provides strong lipid-based structural arguments supporting the concept that the PL monolayer of the cytoplasmic leaflet of the ERbilayer buds outwardly to form the monomolecular layer wrapping the NL (TG, C and CE) core of lipid droplets. PC, PS, PI and PE represent essential elements of the cytoplasmic leaflet of the ER bilayer and the LD monolayer. Their DAG species are strikingly similar in the cytoplasmic leaflet of ER and the monolayer of LDs of control (þ/þ) liver and likewise of fads2À/À liver, but devoid of PUFAs and substituted by the 20:3 5, 11, 14 surrogate. We observed that HFHC diet a) suppresses 20:3 5,11,14 surrogate synthesis, which includes chain elongation of 18:2 at the tetrameric elongase and a D5-desaturase reaction at the trimeric desaturase complex in the ER compartment, and b) surprisingly, strongly inhibits the depletion of all PUFAs in the lipidome of fads2À/À liver, documented here in the lipidomes of ER, LD ( Figure 5G,H, I, J K,L) and Lp ( Figure 10L). The disclosure of the molecular basis of this important metabolic and nutritional interplay awaits future experiments.
Our experiments suggest the fundamental importance of u3-and u6-PUFAs as PL constituents of the lipid bilayer structure, the scaffold of enzymes and transcription factor of lipogenesis in liver. Inactivation of the fads2 locus in the fads2À/À mouse perturbs liver lipid metabolism, lipid storage (including processing of inactive precursor proteins of sterol regulatory element-binding protein (SREBP) transcription factors [30,31,34] and causes dyslipoproteinemia. Studies in primary rat hepatocytes or in liver have demonstrated that liver X receptor a (LXRa) is not regulated by PUFAs as a ligand, and the suppression of SREBP-1 and its targeted lipogenic genes by unsaturated fatty acids is independent of LXRa [32]. The lipid monolayer is the scaffold of LD-specific proteins, the most abundant of which are Perilipins Plin 1, 2 and 3, equipped with protein binding motifs for targeting, binding and budding of lipid droplets from the ER [33]. As major constituents of the globular surface of LD, they assist DGAT-catalyzed TG synthesis and storage in NDs. ATGL mobilizes lipases and size of droplets and liver steatosis and assembly of lipoproteins.
The perturbed PL monolayer of LD in the liver of fads2À/À mice has a profound impact on the accessibility of PAT protein dynamics, mobilization of TG, hydrolysis and synthesis and providing C, CE and TG for loading the Lp transport system. Steady-state expression PLIN2 in ER and LD of fads2À/À liver is strongly downregulated and ATGL upregulated ( Figure 8A,B). HFHC diet elevates steady-state concentrations of PLIN1, 2 and 3 in apoeÀ/À liver, which is reduced in fads2À/À liver and normalized in the fads2À/À x apoeÀ/À double mutant. PL classes were present in liver ER, LD and Lp-phospholipidomes of control þ/þ and fads2À/À mice on regular diet in closely similar concentration, but differed severely in their modified DAG structures with different spatial requirements, transmitting a disturbed topology of the polar head group required for the cooperation of Rab5 and phosphatidylinositol (PI) and PI3P and effectors regulating membrane tethering and fusion of LD. Inactivation of the fads2 gene locus renders the fads2À/À mouse resistant to long-term PUFA-free regular and atherogenic HFHC (Western) dieteinduced development of atherosclerotic lesions. Inactivation of the fads2-locus and PUFAs in the fads2À/À x apoeÀ/À and fads2À/À x ldlrÀ/À double mutants neither accelerated nor retarded the rapid and extensive development of atherosclerotic lesions in the apoeÀ/À and ldlrÀ/À mouse.
The cis-double bond systems of u3-and u6-EFAs and PUFAs are chemical targets, highly vulnerable to autoxidation, which releases reactive aldehydes for derivation of proteins, including toxic LDLox, ligand of the endothelial scavenger receptor. Our immunehistochemical results strongly suggest that the absence of PUFAs as highly vulnerable chemical targets of autoxidation prohibits inflammatory responses and formation of atherosclerotic lesions.

CONCLUSION AND SYNOPSIS
The results of this comprehensive dietary study and the previous broad phenotypic characterization of the u3-and u6-PUFA synthesise deficient, unbiased genetic mouse model have widened and focused our molecular insight into the pleiotropic roles of u3-and u6-PUFAs.
They uncovered the necessity of further scrutinizing molecular studies of the pleiotropic functions of u3-and u6-PUFAs as molecular culprits or benefactors during the lifespan in the phase of growth and development [11,12], homeostasis [6,25] and degeneration before being included in legitimate dietary recommendations. and Sto 32/50-2, Lipoprotein separation and quantification by agarose gel electrophoresis using the Sebia HYDRASYS agarose gel electrophoresis system were kindly carried out by PD. Dr.T.Streichert and Dr. A Klatt, Institut für Klinische Chemie, University Hospital Cologne.