The classical amyloid cascade hypothesis has played a leading role in AD research but has not been able to account for the heterogeneity of AD in sporadic and late-onset cases or the discordance between amyloid load and cognitive impairment in a minority of individuals [7, 23]. It is emerging that the metabolic and lipid abnormalities play an upstream role in A$\beta$ and tau pathology and contribute to the genetic, inflammatory, and energetic failures to a singular disease model [10, 15, 24].

The metabolism of lipids intersects with the AD biology on several levels. The localization and activity of $\beta$ and $\gamma$-secretases are regulated in membrane microdomains (lipid rafts), which are cholesterol and sphingolipid rich membrane microdomains and enhanced amyloidogenic processing of amyloid precursor protein (APP) occurs when cholesterol levels are elevated [25, 26, 27]. In the meantime, oxidative stress and apoptotic signals are enhanced by ceramides and oxidized phospholipids and oxidative damages membrane integrity and antioxidant protection by disrupting plasmalogen depletion [22, 28, 29].

Lipid biology is associated with disease susceptibility, which is connected to genetic factors, especially the APOE $\epsilon$4 allele, that modulate lipid transport, A$\beta$ clearance, and neuroinflammatory tone [14, 15, 18]. Moreover, the multi-omics analysis unveils the APOE $\epsilon$4-mediated lipidomic signatures to be associated with the impaired mitochondrial dysfunction, deteriorated lipid droplet turnover, and energy deficiency in neurons [10, 30]. Taken together, these data put lipid metabolism in the realm of an integrative upstream mechanism that affects various dimensions of AD pathology, fatigued in amyloid and tau aggregation as well as glial activation and oxidative damage. In the therapeutic approach, lipid pathway targeting has already been showing promising initial outcomes. Lipid homeostasis and A$\beta$ clearance are restored with the help of APOE lipidation enhancers, liver X receptor (LXR)/retinoid X receptor (RXR) agonists, and ACAT (Acyl-CoA: cholesterol acyltransferase) inhibitors in the preclinical models [31, 32, 33]. Equally, omega-3 supplemented diets and ketogenic diets improve the composition of the membrane and energy metabolism and slow the cognitive impairment in mild cognitive impairment (MCI) and in the early AD [33, 34, 35]. Collectively, these developments underscore lipid metabolism as a promising avenue of therapeutics that serves as an interface between the molecular pathways and clinical therapy.

Lipid metabolism has a fundamental effect on the metabolic formation, maturation and maintenance of the CNS. In addition to playing structural roles in cellular membranes, lipids control the membrane fluidity, vesicles in the synaptic cleft, intracellular movement, and intracellular signal transmission. The second-richest organ of the body in terms of lipid content is the brain, in which lipids constitute about 50–60% of the dry mass [8, 11, 12]. Phospholipids, sphingolipids, cholesterol and fatty acids are major lipid classes in the CNS, and each have their specific physiological roles to play. The structural framework of neuronal and glial membranes is made up of phospholipids, including phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, and phosphatidylinositol that also act as the precursors of intracellular signaling molecules [36]. The specific subclass of phospholipids, plasmalogens, is an enriched form of phospholipids, which is found in large amounts in neural tissue and plays a role in membrane stability, vesicular fusion, and antioxidant defense [37]. Sphingolipids such as sphingomyelin and gangliosides are inbuilt parts of membrane microdomains, which structure receptor localization and signal transmission [38]. Cholesterol synthesized de novo in the CNS plays a role in membrane organization, myelination and functioning of neurotransmitter receptors [39], and polyunsaturated fatty acids, including docosahexaenoic acid (DHA) in the maintenance of the structure of synaptic membranes and membrane-associated processes [40]. Table 1 (Ref. [8, 10, 13, 15, 18, 22, 28, 32, 33, 34, 38, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52]) summarizes the major lipid classes, their mechanistic roles, and key study outcomes in Alzheimer’s disease pathophysiology.

Since most circulating lipids do not enter the CNS due to the blood-brain barrier, autonomous lipid production, transport, and recycling processes assist the CNS to stay at homeostasis [39]. The production of cholesterol and phospholipids and their transport into neurons occurs through lipoprotein particle-containing apolipoproteins produced by astrocytes and transported to the neurons by receptor-mediated endocytosis [53]. ATP (adenosine triphosphate) binding cassette transporters such as ABCA1 (ATP-binding cassette subfamily A member 1) and ABCG1 (ATP-binding cassette subfamily G member 1) mediate cholesterol efflux and loading lipids onto apolipoproteins and have central roles in the maintenance of lipid balance across neural cell types [54]. Other than the transport of lipids, lipid storage is also another facet of CNS lipid homeostasis. Lipid droplets are cytoplasmic organelles that store neutral lipids and serve as dynamic stores that compensate for changes in lipid supply and safeguard cells against lipotoxic stress [55].

The lipid brain metabolism involves the well-coordinated mechanisms of synthesis, turnover, and consumption. The cholesterol biosynthesis is regulated by rate limiting enzymes and transcriptional regulators that are responsive to cellular lipid requirements through the mevalonate pathway [56, 57]. Cholesterol 24-hydroxylase (CYP46A1) converts excess cholesterol to 24S-hydroxycholesterol, and this allows the body to regulate the amount of cholesterol across the blood-brain barrier [56]. Fatty acid synthesis is done by fatty acid synthase and elongation-desaturation systems, and the breakdown of fatty acids is done by $\beta$-oxidation in the mitochondria and peroxisomes [58]. A complex of transcriptional regulators directs lipid homeostasis in the CNS, regulating lipid synthesis, oxidation and efflux. SREBPs (sterol regulatory element-binding proteins) and the LXRs and RXRs mediate cholesterol biosynthesis, apolipoprotein lipidation and cholesterol efflux, and the PPARs (peroxisome proliferator-activated receptors) mediate fatty acid oxidation and lipid-responsive gene expression [41]. These regulatory systems combine to ensure cellular lipid balance in physiological conditions.

There are several different types of neural cells with lipid metabolic programmes that are dependent on the CNS functions. Astrocytes are the main place of cholesterol production and lipid transport that provides neurons with the necessary lipids to turnover and maintain their membrane and synapses [59]. The neurons are extremely dependent on lipids delivered to support the rapid membrane restructuring and synaptic vesicle cycle [60]. It is observed that microglia change the dynamic organization of lipid metabolism depending on the environmental cues, which are indicators of the change in the cellular energetic and functional state [61]. As oligodendrocytes, which are highly cholesterol and sphingolipid-enriched, rely on a strong lipid biosynthesis to sustain and/or form myelin, their presence in the brain influences myelin regeneration [62]. Interconnectedness of these cell-type-specific lipid metabolic pathways is the basis of CNS structural integrity and functional resilience. Fig. 1 summarizes the lipid metabolic programs of neurons, astrocytes, microglia, and oligodendrocytes, and their roles in maintenance of synapses, neuroinflammation, and maintenance of myelin.

Fig. 1.

Cellular crosstalk of lipid metabolism in the Alzheimer’s brain. Astrocytes, neurons, microglia, oligodendrocytes, and endothelial cells coordinate lipid trafficking essential for brain homeostasis. Astrocytes supply cholesterol and fatty acids via APOE lipoproteins, while neurons regulate lipid oxidation and APP processing. Microglia sense extracellular lipids and amyloid through TREM2, forming lipid droplets during activation. Oligodendrocytes maintain myelin lipid turnover, and endothelial cells mediate lipid transport across the blood–brain barrier. Disruption of this lipid network contributes to neurodegeneration and inflammation in Alzheimer’s disease. TREM2, Triggering receptor expressed on myeloid cells 2; BBB, blood–brain barrier.

The amyloid cascade hypothesis postulates that aberrant cleavage of APP by $\beta$-secretase and $\gamma$-secretase leads to the formation of A$\beta$ peptides, which subsequently aggregate into soluble oligomers and insoluble amyloid plaques [2]. Nevertheless, in some rare cases of AD, upstream metabolic and inflammatory conditions play a critical role in A$\beta$ production and clearance. The primary determinant of APP processing is membrane lipid composition. Lipid rafts containing cholesterol concentrate $\beta$-site amyloid precursor protein-cleaving enzyme 1 (BACE1) and $\gamma$-secretase complexes, which prefer amyloidogenic cleavage. High cholesterol in the membrane stabilizes such microdomains and increases A$\beta$ production, and redistribution of cholesterol to non-raft areas, transcription of APP processing to the $\alpha$-secretase [8, 26]. In this way, cholesterol homeostasis has a direct impact on the ratio between neurotoxic and non-amyloidogenic processing. Gangliosides, notably GM1 (Monosialotetrahexosylganglioside), mediate the early conformational changes of A$\beta$ by GM1-A$\beta$ complex formation that promotes oligomer formation [27]. These connections are made at the surface of the membrane, which emphasizes the importance of lipid microdomains at the beginning of plaque.

Sphingolipid imbalance also regulates amyloidogenic mechanisms. Ceramide stabilization makes BACE1 steady and increases its enzymatic performance. High levels of ceramide species are observed in the preclinical AD and associated with amyloid load and cognitive impairment [42, 43]. Ceramides, mechanistically, make membranes stiffer and change endosomal trafficking, which promotes amyloidogenic APP pathways. Notably, lipid changes can lead to dysfunction of A$\beta$ clearance. Maladaptive lipidation of APOE, especially in APOE $\epsilon$4 carriers, inhibits receptor-mediated A$\beta$ uptake and extracellular A$\beta$ elimination. Consequently, lipid dysregulation plays a role in enhancing both the production and deposition of A$\beta$.

Tau is a microtubule related protein that stabilizes the architecture of the axonal cytoskeleton. During AD, tau hyperphosphorylates, dissociates off microtubules and forms NFT [63]. Tau pathology can be modified by lipid perturbations via kinase activation, oxidative stress, and membrane signaling changes. The activation of the stress-responsive kinases such as glycogen synthase kinase 3 beta (GSK3$\beta$) and cyclin-dependent kinase 5 (CDK5) by ceramides and oxidized lipids leads to tau phosphorylation of pathological epitopes [42]. High ceramide concentrations can thus be linked to membrane stress with cytoskeletal destabilization. Oxysterols 24S-hydroxycholesterol and 27-hydroxycholesterol are accumulated in the parts of the AD brain and enhance the activation of kinases through oxidative stress [32]. The lipid peroxidation products, such as 4-hydroxynonenal (4-HNE), may covalently alter tau and increase aggregation potential and worsen microtubule binding. On the other hand, the lipid derivative of DHA like neuroprotectin D1 inhibits the activation of GSK3$\beta$ and decreases tau phosphorylation [64]. These results indicate that the membrane lipid composition influences tau stability based on both degenerative and protective signaling pathways. In this way, lipid imbalance can determine tau pathology by altering the kinase signaling, oxidative modification, and membrane-cytoskeleton interactions.

AD is characterized by chronic neuroinflammation, which is mainly caused by the activation of the microglial and astrocytic functions [65]. Metabolism of lipids is critical in the establishment of the tone of inflammatory processes. Nuclear factor kappa B (NF-$\kappa$B) and cytokine releases are mediated by arachidonic acid-derived eicosanoids, which are inflammatory cascades, including prostaglandins and leukotrienes. Enhanced phospholipase A2 in AD releases arachidonic acid of membrane phospholipids, maintaining pro-inflammatory signaling. Conversely, the specific pro-resolving mediators (SPMs) derived out of DHA and eicosapentaenoic acid (EPA), such as resolvins and protectins, stimulate the resolution of inflammation by inhibiting the activation of inflammasomes and stimulating the repair of tissues [66]. The ratio of omega-6-produced pro-inflammatory lipids and omega-3 produced pro-resolving mediators determines microglial phenotype and persistence of inflammation. Lipid droplet-accumulating microglia (LDAM) is a pathologic form that is marked by the inability to phagocytose, a large number of ceramide and cholesterol deposition and elevated production of cytokines [44]. This lipid overload condition decreases the ability to clear the A$\beta$ efficiently and facilitates complement-mediated synaptic pruning. In such a way, lipid composition defines the microglia taking on protective versus destructive identities, making lipid metabolism a determinant of neuroimmune homeostasis.

Excessive production of reactive oxygen species (ROS) and dysfunctional antioxidant defenses are the causes of oxidative stress. The polyunsaturated fatty acids in the neuronal membranes are also highly susceptible to peroxidation, which produces the reactive aldehydes like 4-HNE and malondialdehyde [67]. These products of lipid peroxidation alter the proteins of synaptic transmission and mitochondrial respiration. The lipid composition of the mitochondria plays a vital role in the bioenergetic stability. Anchor Electron transport chain complexes are anchored by cardiolipin, a mitochondria-specific phospholipid. The oxidative damage of cardiolipin interferes with the respiratory chain assembly, decreases the production of ATP and facilitates the release of cytochrome c [68]. Synaptic terminal energy breakdown is a precursor to neuronal loss and leads to cognitive impairment. The further effect of the ceramide accumulation is the impairment of the mitochondrial membrane integrity and the amplification of the apoptotic signaling pathways. A change in the sterol structure affects the fluidity of the mitochondrial membrane and respiratory efficiency, too [10]. Accordingly, lipid dysregulation is a feed-forward mechanism: oxidative damage to lipids causes lipid damage, lipid damage causes mitochondrial impairment, and mitochondrial impairment produces further ROS.

Loss of synapses is more strongly related to cognitive impairment than plaque or tangle load [69]. The phospholipids and cholesterol in synaptic membranes are very abundant in DHA, which control vesicle fusion, receptor movement and dendritic spine structural formation [12, 40]. Membrane fluidity is reduced by lowering DHA and plasmalogen loss, which inhibits the assembly of the soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) complex, which inhibits neurotransmitter release, and reduces the plasticity of the synapse. The N-methyl-D-aspartate (NMDA) and $\alpha$-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors are clustered by the cholesterol imbalance and disrupt the glutamatergic signaling [70]. Isoform dependent lipid transport also varies, aiding in regulating synaptic resilience by way of APOE. APOE $\epsilon$4 has less lipidation efficiency, which affects the redistribution of cholesterol to neurons and the process of synaptic repair [15]. Also, ceramide build-up enhances rigidity in the membrane and disturbs the provision of ATP in the mitochondrion in the presynaptic terminals, worsening synaptic failure. The GM1-enriched synaptic membranes are preferentially bound by A$\beta$ oligomers, enhancing local toxicity. At the same time, complement-mediated synaptic pruning is increased by lipid-mediated microglial dysfunction. Thus, the function of the synaptic point is a meeting point of lipid imbalance with amyloid toxicity, tau pathology, inflammatory signalling, and mitochondrial failure. Membrane destabilization earlier in life can dictate the pathway between triggers of functional impairment and neurodegeneration that cannot be reversed.

The distribution of cholesterol in the membranes of neurons is a vital regulator of APP processing. More cholesterol in the membrane prefers $\beta$ and $\gamma$-secretases localization to cholesterol-rich microdomains, which also increases amyloidogenic cleavage. On the other hand, a-secretase-mediated non-amyloidogenic processing is favored by increased cholesterol efflux or redistribution, decreasing the production of A$\beta$ [10, 13, 32]. The brain has a large amount of cholesterol turnover that is regulated primarily by CYP46A1, which is the only way to convert cholesterol to 24S-hydroxycholesterol and then allow the substance to be taken across the blood-brain barrier. Lower cholesterol turnover enhances membrane rigidity and secretase clustering, and CYP46A1 activation enhances synaptic plasticity and decreases the amyloid burden in models. Likewise, defects in cholesterol efflux transporter including ABCA1 and ABCG1, reduce the effectiveness of APOE lipidation, which disrupts extracellular lipid transport and encourages the aggregation of A$\beta$ [15, 71]. Therefore, an unregulated cholesterol secretion and not the total levels of cholesterol, is what determines the balance between protective and pathogenic APP processing.

The metabolism of sphingolipids has a strong impact on the survival of neurons and inflammatory signaling. The sphingomyelin hydrolysis or de novo synthesis produces ceramides that aid amyloidogenic processing and at the same time stimulate the mitochondrial permeability, oxidative stress, and apoptotic signalling [42, 43]. Early lipidomic profiles in preclinical AD show that the species of ceramides such as C16:0 and C18:0 accumulate early [22]. These species disrupt membranes of mitochondria, elevate the production of reactive oxygen species and enhance inflammatory cascades. Conversely, the pro-survival sphingolipid metabolite, sphingosine-1-phosphate (S1P), is lower in AD, the loss of which hinders neurogenesis and synaptic plasticity [72]. Sphingomyelin to bioactive ceramide species transition is thus a directional metabolic imbalance, which strengthens amyloid toxicity and inflammatory susceptibility.

Membrane curvature, vesicle fusion, and receptor trafficking is supported by phospholipids which are phosphatidylcholine (PC), phosphatidylethanolamine (PE), and plasmalogens. Polyunsaturated PC and PE species depletion impairs the fluidity of the membrane and decreases the formation of synaptic vesicles in AD [22]. Reactive aldehydes produced through oxidative modification of phospholipids (4-hydroxynonenal) covalently modify APP and tau and promote aggregation propensity [10, 73]. The depletion of plasmalogens also impairs antioxidant buffering capacity, making one more vulnerable to lipid peroxidation. These structural changes of phospholipids destabilize synaptic membranes and provide a biochemical environment that promotes protein misfolding and neuronal stress.

Lipid rafts are used in receptor clustering and intra-cellular signaling. Cholesterol and sphingolipid changes in composition impair raft integrity, which influences the localization of NMDA receptors, Tropomyosin receptor kinase B (Trk$\beta$), insulin receptors, and other signaling proteins [74, 75]. Different isoforms of APOE have varying effects on raft organization, with APOE $\epsilon$4 disrupting membrane microdomains and disrupting receptor clustering, rendering cells excitotoxically vulnerable [76]. The oxidation of sterols accumulating in rafts, including 7-ketocholesterol, also worsens receptor signaling and neuronal stress [32]. Raft remodeling is thus the connection between membrane lipid composition and defective synaptic signaling and metabolic dysfunction.

Neuroinflammation and lipid metabolism are closely coupled with lipid-sensing receptors and bioactive lipid mediators. Toll-like receptor 2 (TLR2), TLR4, and the lipid-sensing receptor Triggering receptor expressed on myeloid cells 2 (TREM2) are activated by oxidized phospholipids and ceramides to induce the activation of microglia and the assembly of inflammasomes [65, 77]. TREM2 mutations affect lipid uptake and phagocytic capacity and cause the build-up of extracellular debris and increased inflammatory signaling. At the same time, the elimination of specialized pro-resolving lipid mediators based on omega-3 fatty acids impairs the resolution stage of inflammation, which prolongs chronic microglial activation [78]. Therefore, lipid imbalance not only initiates inflammatory activation but also inhibits its measurement, establishing a self-sustaining neuroimmune response cycle.

Lipid composition is very sensitive in mitochondria. Mitochondria-specific phospholipid (cardiolipin) stabilizes respiratory chain complexes. Its oxidation interferes with the efficiency of electron transport and favors the release of cytochrome c [68]. An increase in the levels of ceramides in the mitochondrial membranes increases permeability transition and apoptotic signaling, whereas a breakdown in $\beta$-oxidation contributes to the accumulated toxic lipid intermediates [10]. Additional changes in lipid trafficking that occur in association with APOE $\epsilon$4 worsen the effects of fragmentation and bioenergetic decline in mitochondria [45]. This reciprocal interaction in which lipid imbalance leads to decreased mitochondrial function, and mitochondrial dysfunction increases lipid peroxidation, leads to a vicious circle where energetic failure progressively increases synaptic degeneration.

The lipid dysregulation offers a mechanistic intermediation of the amyloid and tau pathology. Ceramides and oxidized lipids stimulate oxidation-stimulated lipids GSK3$\beta$ and CDK5, facilitating tau phosphorylation and increasing amyloidogenic processing at the same time [10]. Lipid aldehydes reacting with APP and tau induce aggregation propensity and seeding potential [46]. The clustering of membrane microdomains also promotes pathological interactions of A$\beta$ oligomers and tau species. Early metabolic imbalance is converted to convergent proteinopathy by common lipid-sensitive signaling pathways.

The peripheral lipid metabolism and gut-derived metabolites have an impact on central lipid homeostasis. Gut microbiota control the bile acid composition and production of short-chain fatty acids that adjust the lipid metabolism in the system and neuroinflammatory tone [47]. Bile acids and trimethylamine-N-oxide (TMAO) cross the blood-brain barrier and encourage oxidative stress and microglial activation [48]. Lipid changes related to dysbiosis thus play roles in systemic-central metabolic interrelation with dietary habits and peripheral lipid dysbalance, being related to neurodegenerative susceptibility.

In addition to the individual lipid species, coordinated changes in the lipid pathways give information on the AD progression. Directional metabolic flux ratios, e.g., sphingomyelin-to-ceramide, are associated with cognitive deterioration [79]. Lipidomic analyses combined with other techniques have shown that cholesterol turnover, sphingolipid remodeling, and phospholipid oxidation are not cascades but networks of interrelated processes [10]. These perturbations at the pathway level affect the amyloid processing, phosphorylation of tau, and activation of inflammation. The mechanistic basis of metabolic intervention, along with stratification of precision, of lipid metabolism can be achieved by seeing lipid metabolism as a dynamic systems network, as opposed to a set of discrete abnormalities.

APOE $\epsilon$4 allele is the most potent genetic risk factor of sporadic AD that presents dose-related contributions to the risk and earlier onset [14, 15, 80]. Contrary to the mutations that directly augment the production of A$\beta$, the APOE affects the disease by the dysfunction of lipid transport. The lipid carrier in the brain is APOE, which facilitates the redistribution of cholesterol and phospholipids between astrocytes and neurons by a process of lipidation via ABCA1. APOE $\epsilon$4 isoform has weakened lipid-binding capacity, a change in structural stability, and proteolytic vulnerability [16, 18]. Unstable lipoprotein particles are the result of poor lipidation, and they disrupt the maintenance of the membrane and the synaptic integrity. In APOE $\epsilon$4 carriers, the defective lipid conveyance impacts the receptor-mediated A$\beta$ clearance and the organization of membrane microdomains, indirectly preferentially supporting amyloidogenic processing. Markedly, this is a sign of defective lipid-dependent trafficking but not mere overproduction of A$\beta$. In addition to the neuronal action, APOE interferes with the lipid metabolism of glial populations. Disturbed cholesterol movement in oligodendrocytes disrupts lipid synthesis of the myelin [81]. APOE $\epsilon$4 enhances the lipid droplet accumulation, decrease $\beta$-oxidation, and increases inflammatory activation in microglia [44, 82, 83]. These cell-type selective effects define APOE $\epsilon$4 as a cause of extensive metabolic susceptibility and not a specific amyloid modulator.

Other susceptibility loci support the lipid regulation vicinity in AD. ABCA1 and ABCG1 regulate cholesterol and phospholipid efflux that is essential in the normal APOE lipidation. The decreased ABCA1 activity leads to the production of lipidated APOE particles of poor quality and to disruption of cholesterol distribution in the membrane [84, 85]. ABCA7 combines lipid translocation and microglial phagocytosis; the loss-of-function variants disrupt the uptake of A$\beta$ and modify the phospholipid composition [86]. CLU is a lipid chaperone that maintains extracellular lipid-protein complexes and complement activation, which connects lipid stress with amplification of inflammation [87]. SORL1 regulates the APP intracellular trafficking between the trans-Golgi and endosomes; impaired function increases the amyloidogenic sorting in the cholesterol-rich compartments [88]. TREM2 plays a vital role in microglial lipid sensing as it binds phospholipids and APOE-containing particles. Lipid uptake and metabolic adaptation are impaired by risk variants, which decreases the phagocytic efficiency and increases the activation of inflammation [77]. APOE-TREM2 functionality creates a lipid-sensing circuit that identifies microglial state changes with deposition of amyloid. Phospholipase C gamma 2 (PLCG2), bridging integrator 1 (BIN1), and phosphatidylinositol-binding clathrin assembly protein (PICALM) variations also modify phosphoinositide breakdown and clathrin-mediated endocytosis, which changes vesicular transport and receptor intake [89]. The combination of these loci creates a network of interconnected lipid trafficking and membrane remodeling that regulates neuronal and microglial activity in response to metabolic stress.

Metabolic conditions and exposure to the environment make significant adjustments to genetic susceptibility. Lipid metabolism is a primary nexus by which diet, insulin sensitivity, and endocrine contributions to the disease path are mediated and especially in APOE $\epsilon$4 carriers. APOE $\epsilon$4 carriers regulate systemic lipid processing and neuronal membrane lipid incorporation. APOE $\epsilon$4 carriers have less incorporation of the DHA into neuronal membranes when supplemented [34, 49]. Since DHA promotes membrane fluidity and anti-inflammatory communication, reduced delivery could be a limiting factor to synaptic resilience and therapeutic responsiveness. This is increased by metabolic disorders. The resistance to insulin increases the synthesis of ceramide and interferes with Akt signaling, which favors the accumulation of mitochondrial stress [90]. Such metabolic stress increases the rate of lipid droplet formation in glial cells and maintains cellular inflammatory activity in genetically susceptible subjects. Gene expression is also controlled by lipid metabolites. The availability of acetyl-CoA regulates histone acetylation, which affects the transcription of lipid synthesis and mitochondrial functioning genes [76, 91]. Ketone bodies that increase the resistance to the resistance pathways of oxidative stress are b-hydroxybutyrate as signaling metabolites [92]. Nuclear receptors such as PPARs and LXRs are activated using fatty acids and coordinate transcriptional programs that control lipid oxidation and immune tone. Therefore, the AD risk is a combination of variants inherited in lipid regulation and metabolism.

Multi-omics studies indicate that AD is caused by the concerted breakdown of lipid-immune-metabolic modules, but not individual gene mutations. The co-expression network of APOE, ABCA1, SORL1, and CLU is associated with inflammatory and mitochondrial genes in large data sets, such as Alzheimer’s Disease Neuroimaging Initiative (ADNI) [10, 22]. The modules linked to ceramide are also related to complement activation and synaptic vulnerability, whereas the decrease in phosphatidylcholine and plasmalogen species correlates with the inhibition of oxidative phosphorylation genes. These synchronized changes can be identified before more complex neurodegeneration, and this shows that lipid network destabilization is a systems-level process. Lipid metabolism is also further tied to the process of gene regulation by epigenetic means. The histone acetylation under the influence of acetyl-CoA affects the transcription of lipid and mitochondrial genes [93]. The products of lipid peroxidation can alter transcriptional regulators, which strengthens expression designations by oxidative stress [94]. Lipid-sensitive nuclear receptors are activated to remodel chromatin accessibility and orchestrate the cholesterol efflux and immune regulation. There is another modulatory layer of sex hormones. The estrogen controls the expression of APOE and ABCA1 and modulates the activity of the mitochondrial function [95]. Reduced estrogen levels can thus increase lipid regulation in those genetically predisposed, which explains sex-specific differences in AD risk. Taken together, these results place AD as a disease of genetically regulated lipid network instability, which is influenced by metabolic and epigenetic contexts.

Consistent with a large-scale decrease in PC and PE species enriched with DHA, such as PC 38:6 and PE 40:6, postmortem lipidomic analyses of AD cortex and hippocampus show such decreases [20, 28]. Plasmalogens, especially plasminyl-PE species, are significantly decreased, that weaken the stability of the membrane curvature and antioxidant buffering activity. Ceramide species, including Cer d18:1/16:0 and Cer d18:1/18:0, on the contrary, are increased [22, 42, 50], indicating increased sphingomyelin hydrolysis and transition to pro-apoptotic lipid signaling. The sphingomyelin-to-ceramide ratio is consequently shifted in favor of pathway activation that is congruent with inflammatory and mitochondrial stress. The total cholesterol can be quite stable, but the decreased level of 24S-hydroxycholesterol and the enhanced level of oxidized sterols such as 7-ketocholesterol show evidence of the disrupted cholesterol turnover and oxidative modification [96]. Spatial lipidomics proves the idea of enrichment of ganglioside GM1 in amyloid plaques and peri-plaque areas, accumulation of ceramide to show the anatomically localized lipid remodeling related to the pathological lesions [97]. Together, brain tissue lipidomics can show a predictable course of lipid degradation, plasmalogen disappearance, ceramide increase, and sterol oxidation that may be observed before severe neuronal damage.

Peripheral lipid profiling detects consistent patterns that are related to cognitive decline and phenoconversion. Lipidomic profiles reported by Huynh K et al. [98] (2020), encompassing multiple lipid classes such as phospholipids and sphingolipids, are associated with both prevalent Alzheimer’s disease and future disease risk. Ceramide species are one of the strongest predictors of the plasma. The high Cer d18:1/16:0 and Cer d18:1/18:0 levels are associated with hippocampal atrophy and progressive acceleration [99]. Notably, the pathway-level measures are better than those based on individual species. A decrease in the ratios of sphingomyelin-to-ceramide, especially SM d18:1/24:0, reflects a higher activity of the sphingomyelinase and metabolic directionality [51]. AD plasma raises the levels of oxidized phospholipids including POVPC (1-palmitoyl-2-(5-oxovaleroyl)-sn-glycero-3-phosphocholine) and PGPC (1-palmitoyl-2-glutaroyl-sn-glycero-3-phosphocholine) [100], which is associated with systemic oxidative stress and the activation of inflammation. Reduced levels of 24S-hydroxycholesterol and disturbed levels of desmosterol in CSF indicate impaired neuronal cholesterol metabolism [101]. Even though the effect of peripheral lipids on predictive performance can be enhanced by systemic metabolism, their combination with genotype and longitudinal imaging data is more effective.

The genotype of APOE plays a significant role in altering the systemic and central lipidomic architecture that can modify the baseline lipid transport efficiency and the susceptibility to metabolic stress. $\epsilon$4 carriers have an increased level of circulating ceramide species, especially Cer d18:1/16:0 and Cer d18:1/18:0, and more oxidized phospholipids and lower levels of plasmalogen [10, 45]. These changes are in line with the dysfunction of the APOE lipidation, the decreased cholesterol efflux via ABCA1, and the modified redistribution of cholesterol and phospholipids among astrocytes and neurons. Ineffective lipidation of APOE $\epsilon$4-carrying lipoproteins results in reduced capacity of aiding membrane repair and synaptic remodeling with the lipid needed, making them more vulnerable to lipid peroxidation and to inflammatory signalling. Notably, genotype not only alters lipid abundance but also pathway directionality. $\epsilon$4 carriers simply show reduced sphingomyelin-to-ceramide ratios, which indicates an increase in sphingolipid metabolism, sphingomyelinase activity and a shift to a more pro-apoptotic sphingolipid metabolism. Combining plasma ceramide levels with APOE genotype allows predicting atrophy of the hippocampal and cognitive decline better than either of these factors alone [102]. It is an interaction that implies that genotype sets a fundamental lipid transport structure, and that lipidomic profiling records dynamic metabolism perturbation imposed on that structure.

Lipid species do not tend to work individually. Systems-level analyses of ADNI and associated multi-cohort data reveal that lipid changes come together in coordinated metabolic modules, which combine inflammatory, mitochondrial, and complement pathways [30, 103]. Such modules are more biologically resolved than individual biomarkers and represent pathway-wide dysregulation. A reproducible module is based on ceramide enrichment and complement cascade proteins (C1q, C3), and inflammatory mediators. Higher levels of ceramide species are associated with higher complement activation levels and microglial inflammatory signals, which are predictors of faster cognitive impairment and increased cortical atrophy. Mechanically, the presence of ceramide facilitates rigidification of membranes, the activation of the inflammasome, and mitochondrial strain, which strengthens complement-mediated synaptic pruning. A second modular association is between loss of polyunsaturated phosphatidylcholine and plasmalogen species in connection with decreased oxidative phosphorylation proteins and abnormal acylcarnitine. This trend indicates that there is impaired membrane remodeling and reduced mitochondrial $\beta$-oxidation capacity. A decreased supply of phospholipids interferes with the process of synaptic vesicle dynamics, and the changes in acylcarnitines show that fatty acid oxidation is not fully performed, and the process is not efficient. Combined with genomic data, this again indicates that these lipid-protein modules agree with risk loci such as APOE, ABCA1, and TREM2 that genetic vulnerability is manifested via a network of coordinated metabolism. Compared to single-analytes techniques, module-based modelling enhances the ability to measure prognostic and increases the ability to interpret biologically.

Regardless of technological progress, the clinical application of lipidomics is limited by the lack of methodological consistency. The pre-analytical variables, such as diet, fasting condition, and sample manipulation, cause a lot of variability [104]. Differences between analytical procedures, ionization platforms, and internal standardization cause problems in cross-study comparability. Biological heterogeneity also makes interpretation difficult. Lipid baseline is dependent on age, sex, metabolic comorbidities, and APOE genotype. Clinical implementation is restrained by the lack of standardized reference ranges. Efforts are being undertaken to standardize nomenclature, reporting, and data processing systems like the Lipidomics Standards Initiative [104]. A reproducible panel of pathways at the panel level, and established cutoffs will be required to move lipidomics past exploratory profiling and establish lipidomics as the precision diagnostic and therapeutic monitoring instrument in AD.

Statins, 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors, have also been suggested as possible neuroprotective agents due to their effects on the lowering of peripheral cholesterol and their possible impact on cerebral cholesterol turnover. Epidemiological research indicates that statins can prevent the onset of dementia over the long term, and greater evidence indicates this effect is stronger in APOE $\epsilon$4 non-carriers [105, 106]. Statins have also been reported to have a mechanistic effect in the modulation of secretase activity, preferential processing of non-amyloidogenic APP, and enhancement of endothelial functioning and cerebral perfusion [107]. Nevertheless, the results of randomized controlled trials have been inconclusive in terms of cognitive effects, probably because of a heterogeneity of blood-brain barrier permeability to statins, intervention timing, underlying vascular risk, and disease severity [108]. There are some results in emerging lipidomic and imaging, which suggest that lipophilic statins (e.g., simvastatin, atorvastatin) have the potential to normalize features of cholesterol turnover and decrease A$\beta$ -related burden when they are administered at early or preclinical stages [10, 32], but these claims cannot be conclusively proven by clinical methods. In general, the data available indicate stage and subgroup effect and not homogeneous benefit in established AD.

Cholesterol efflux, APOE lipidation and inflammatory tone are controlled by LXRs and RXRs. Expressions of ABCA1 and ABCG1 are augmented by pharmacologic activation, which boosts the export of cholesterol and possibly facilitates A$\beta$ clearance. Proper agonist bexarotene also decreases the load of the plaque and enhances cognition in transgenic mouse brains [109], although further clinical trials in humans showed inconsistent results, which could be due to the pharmacokinetic properties and dose-related toxicity. Emerging data in preclinical research indicate that integrating the actions of LXR and Peroxisome Proliferator-Activated Receptors (PPARs) can have more extensive effects on lipid homeostasis, mitochondrial dysfunction and neuroinflammation [110, 111]. Since hepatic steatosis can be caused by pan-LXR activation, novel LXR$\beta$ (Liver X receptor beta) modulators are underway to enhance safety in the long-term whilst retaining central lipid regulation functions.

The goal of ketogenic dietary interventions is to reverse the cerebral hypometabolism in AD by enhancing the levels of circulating $\beta$-hydroxybutyrate and acetoacetate, which are relatively accessible as substrates to brain oxidation in case glucose metabolism is compromised [33, 52]. The ketone bodies join the tricarboxylic acid cycle as acetyl-CoA to aid with ATP-generation and $\beta$-hydroxybutyrate also acts as a signaling metabolite to inhibit class I histone deacetylases and activate transcriptional programs associated with antioxidant defense and mitochondrial biogenesis [112]. Preclinical trials indicate a reduction in amyloid load and inflammatory biomarkers, whereas small clinical trials in mild cognitive impairment indicate slight positive cognitive effects [33]. Disadvantages are the problem of compliance, fluctuating cardiovascular outcomes based on the fat composition, and a lack of long-term disease-modifying data. Its therapeutic utility is best likely to be used in the early or prodromal phase and preferably directed by biomarkers of hypometabolism and stratified by genotype and metabolic situations.

The Mediterranean and MIND diets, which are based on monounsaturated fats, omega-3 polyunsaturated fatty acids, antioxidants, and polyphenol-rich diets, are linked with decreased AD risk and slower cognitive deterioration [48]. Mechanistically, oleic acid of olive oil and omega-3 fatty acids are membrane-stabilising in nature and can re-establish lipid mediator balance towards anti-inflammatory eicosanoids. The components of Polyphenols and antioxidants minimize lipid peroxidation and oxidative alteration of membrane phospholipids. The MIND diet also focuses on factors that are associated with vascular and metabolic well-being (leafy greens, berries, whole grains) and minimizes the intake of saturated fats, which can indirectly lower the levels of insulin resistance and lipid-induced inflammatory priming. Nevertheless, causality is challenging to determine because of confounding lifestyle factors, whereas evidence in the form of randomized trials is still scarce. Individual effects of genotypes can be different; APOE $\epsilon$4 carriers tend to exhibit less incorporation of DHA into brain phospholipids, which can counteract the action of omega-3-dependent pathways [34]. Overall, these diets can be classified as the low-risk preventive or modulatory measures with the possibility of increased accuracy when used with lipidomic and genetic stratification.

Neuroprotectin D1 (NPD1) is a DHA-derived specialized pro-resolving mediator that is produced through the action of 15-lipoxygenase and serves as a bioactive signal to counteract the action of A$\beta$ -induced stress response. NPD1 enhances the anti-apoptotic response via boosting the Bcl-2 family level of protective proteins (Bcl-2, Bcl-xL) and inhibiting the pro-apoptotic mediators (Bax, Bad), which stabilize mitochondrial integrity and inhibit caspase-3 activation [113]. NF-$\kappa$B activation is also reduced by NPD1, as well as the expression of pro-inflammatory cytokines such as tumor necrosis factor alpha (TNF$\alpha$) and interleukin-1 beta (IL-1$\beta$) [12, 114]. Simultaneously, NPD1 is also reported to suppress BACE1 and facilitate non-amyloidogenic APP processing [115], which is a viable way of lowering amyloid synthesis and cellular susceptibility. It is interesting to note that the lowering of NPD1 in the AD hippocampus with the presence of DHA indicates that pro-resolving signaling changes are impaired by DHA [12]. Lipid mediator stability and delivery issues limit translational development and genotype-specific variations in DHA delivery (especially through APOE $\epsilon$4 carriers) can affect NPD1 bioavailability and response.

Membrane phospholipid-released arachidonic acid (AA) is a key precursor of bioactive eicosanoids that regulate inflammatory and excitotoxic signalling. Higher activity of PLA2 (Phospholipase A2) and free AA was also reported in AD-affected areas [116]. AA oxidation through cyclooxygenase (COX) and lipoxygenase (LOX) produces prostaglandins and leukotrienes that encourage microglial action, cytokine production, and oxidative stress, which amplifies inflammatory responses that depend on NF-$\kappa$B. The amyloidogenic processing is also affected by AA signaling, changing membrane properties and microdomain organization, and evidence indicates that AA metabolites can also elevate BACE1 expression and stabilize its localization to lipid rafts, promoting APP cleavage to produce A$\beta$ [21]. Oxidative stress produced by AA can further enhance the formation of reactive aldehydes, supporting the oligomerization of A$\beta$ and leukotriene signaling is elevation of calcium influx and excitotoxic susceptibility, which contributes to the downstream activation of tau kinases. Since AA and DHA are in competition on incorporation in the membrane, a high omega-6/omega-3 ratio encourages the production of pro-inflammatory mediators, and DHA enrichment encourages pro-resolving pathways. Selective COX inhibition has not been very effective in developed AD, as predicted by the timing of AA-mediated amplification of this inhibitor and the physiological functions of AA signaling in synaptic activity.

Lipid responder phenotypes, where the exceptional pathway-wide lipid dysfunction can be used to predict treatment responses, are supported by integrating the lipidomics of plasma with genetic stratification [10, 22]. High levels of ceramides and low levels of sphingomyelin-ceramide ratios accompanied by complement activation and rapid progression are known to be associated with a ceramide-dominant inflammatory phenotype, often based on APOE $\epsilon$4 carriers. Due to the presence of the signs of disrupted oxidative phosphorylation, a phospholipid-deficient bioenergetic phenotype is characterized by depletion of plasmalogens and polyunsaturated phosphatidylcholine species [30]. Altered oxysterols and lower signs of efficiency of lipidation of APOE are seen in a cholesterol-carrying-deficient phenotype. These patterns do not exclude each other but are rather dominant network imbalances that may be determined using systems modeling. The machine learning methods that combine lipid species, genotype and proteomic markers enhance the prediction of the progression and it has the potential to decrease heterogeneity in intervention studies [117]. Potential validation and standardized phenotype definitions are important, followed by an ordinary clinical translation.

Peroxisomes control very-long-chain fatty acid $\beta$-oxidation, plasmalogen biosynthesis, and ROS detoxification, which makes them significant factors in lipid homeostasis and redox regulation in AD. Damaged peroxisomal oxidation may cause the build-up of the very-long-chain fatty acids that destabilize the membrane properties and can indirectly influence the microdomain organization [118]. Peroxisomal dysfunction is also damaging to plasmalogen synthesis, in line with the plasmalogen depletion found in AD cortex, impairing antioxidant buffering, and making them susceptible to lipid peroxidation. The lowered detoxification rate accompanied by catalase-dependence can further allow the build-up of hydrogen peroxide, which escalates oxidative injuries. There is an essential crosstalk with mitochondria; the inability to perform peroxisomal oxidation may augment lipid load to mitochondria to strengthen oxidative stress [119]. Transcriptomic research findings of alteration of peroxisome biogenesis factors and ether lipid enzymes in AD brain tissue indicate peroxisome implication other than a secondary effect [120]. Plasmalogen supplementation and regulation of PPAR-controlled lipid oxidation signalling have been found to be therapeutically promising in preclinical trials, with larger clinical trials required.

The glymphatic system helps in the exchange of CSF-interstitial fluid and the clearance of soluble metabolites. Although it is best examined in the context of A$\beta$ and tau removal, dysfunctional glymphatic activity may also be a contributing factor to the build-up of oxidized lipids, inflammatory lipid mediators, and cholesterol metabolites [121]. Polarized expression of the aquaporin-4 (AQP4) on astrocytic end feet depends on glymphatic flow. The mislocalization of AQP4 and decreased effectiveness in clearance are linked to aging and AD and could contribute to the heightened extracellular retention of lipid peroxidation end products and lipid pro-inflammatory mediators [122]. Since the extracellular membrane lipid transport is partially achieved through APOE-carrying lipoproteins, defective APOE lipidation in APOE $\epsilon$4 carriers may augment defective clearance capability. Loss of sleep also impairs glymphatic functioning and could increase the rate of lipid-derived oxidative species, which connects sleep biology with lipid homeostasis [123]. Stiffness of the vasculature and decreased plasticity also impair the perivascular fluid dynamics, creating another point of convergence between vascular risk and metabolic condition and inefficient clearance. Even though human biomarkers of glymphatic functionality are still in development, a promising avenue of approach to indirectly change lipid clearance dynamics in AD is through targeting sleep and vascular health.