Undetected Association Between Fatty Acids and Dementia with Lewy Bodies: A Bidirectional Two-Sample Mendelian Randomization Study

Abstract

Background:

Although observational studies indicated connections between fatty acids (FAs) and Alzheimer’s disease and dementia, uncertainty persists regarding how these relationships extend to dementia with Lewy bodies (DLB).

Objective:

To explore the potential causal relationships between FAs and the development of DLB, thus clarifying these associations using genetic instruments to infer causality.

Methods:

We applied a two-sample Mendelian randomization (MR) and multivariable Mendelian randomization (MVMR) approach. Genetic data were obtained from a DLB cohort, comprising 2,591 cases and 4,027 controls of European descent. Eight FAs, including linoleic acid, docosahexaenoic acid, monounsaturated fatty acid, omega-3 fatty acid, omega-6 fatty acid, polyunsaturated fatty acid, saturated fatty acid, and total fatty acid, were procured from a comprehensive GWAS of metabolic biomarkers of UK Biobank, conducted by Nightingale Health in 2020 (met-d), involving 114,999 individuals. Our analysis included inverse-variance weighted, MR-Egger, weighted-median, simple mode, and weighted-mode MR estimates. Cochran’s Q-statistics, MR-PRESSO, and MR-Egger intercept test were used to quantify the heterogeneity and horizontal pleiotropy of instrumental variables.

Results:

Only linoleic acid showed a significant genetic association with the risk of developing DLB in the univariate MR. The odds ratio for linoleic acid was 1.337 with a 95% confidence interval of 1.019–1.756 (p_IVW = 0.036). Results from the MVMR showed that no FAs were associated with the incidence of DLB.

Conclusions:

The results did not support the hypothesis that FAs could reduce the risk of developing DLB. However, elucidating the relationship between FAs and DLB risk holds potential implications for informing dietary recommendations and therapeutic approaches in DLB.

Keywords

Alzheimer’s disease dementia dementia with Lewy bodies fatty acids linoleic acid Mendelian randomization

INTRODUCTION

Dementia with Lewy bodies (DLB) is the second most common neurodegenerative disease [1, 2], following Alzheimer’s disease (AD) among aging individuals. Its diagnosis relies on a combination of symptoms, including visual hallucinations, fluctuating cognitive impairments, rapid eye movement, sleep behavioral disorder, and parkinsonism [3]. The classical hallmark of DLB is Lewy pathology, which includes Lewy bodies and Lewy neurites, primarily composed of misfold α-synuclein (αS) [4]. Moreover, over half of DLB pathological changes exhibit co-pathological features similar to AD, characterized by amyloid-β (Aβ) plaques and tau tangles [5]. With improving living conditions and increasing awareness about healthy life expectancy, DLB incidence increases with age, accompanying multiple age-related neurodegenerative and cerebrovascular pathologies [6], positioning DLB as a growing global health challenge.

Fatty acids (FAs) are major components of lipid species, playing a pivotal role in the structure and function of almost all cells [7]. They undergo diverse chemical modifications, which significantly expand their functional spectrum and biological activities [8], including involvement in various metabolic pathways, energy expenditure processes [9], and antibacterial functions [10]. Prior studies have established that a high intake of FAs can mitigate the risk of cardiovascular diseases and influence body composition [11] through their lipid-lowering, anti-hypertensive, and anti-atherosclerotic effects [12]. In the brain, which is rich in lipids, FAs are crucial for neuronal cell growth and development. Specifically, in the central nervous system, omega-3 fatty acids, notably docosahexaenoic acid (DHA), are known to induce an anti-inflammatory phenotype in microglia and enhance the expression of phagocytic receptors [13]. Linoleic acid (LA), a precursor of DHA and a major brain component, emerges as a potential protector of brain health by improving the functionality of the blood-brain barrier [14], enhancing the migratory ability of microglia, thereby promoting phagocytosis and degradation of internalized tau, leading to a reduction of extracellular tau seeds [13]. LA also influences tau protein by inducing its spontaneous assembly and inhibiting its aggregation [15]. αS is a presynaptic protein involved in neurotransmission and neural signaling [16]. In the context of DLB pathology, αS accumulates as brain inclusions. Polyunsaturated fatty acid (PUFA) has been shown to promote the oligomerization of αS, increasing its levels in mesencephalic neuronal cells, while saturated fatty acids (SFA) appear to reduce the level of oligomerized αS [17]. In transgenic DLB mice, soluble αS oligomers, encouraged by PUFA, accumulate with age and precede the formation of neurodegenerative insoluble aggregates [17]. α-Synucleinopathy is a term used to describe a group of neurodegenerative disorders characterized by abnormal deposits of a protein named αS in the brain [18]. Type-3 fatty acid-binding protein (FABP3) is a critical factor for facilitating the uptake, propagation, and neurotoxicity of αS in neuronal cells and contributing to mitochondrial synthesis [19, 20]. Additionally, many other FABPs are involved in α-synucleinopathies via the production of reactive oxygen species or modulation of the inflammatory response in glial cells [18]. These findings contrast with those of earlier studies, indicating a complex relationship between FAs and DLB pathology.

Observational studies have indicated that diets rich in FAs such as LA, omega-3 fatty acid, and PUFA might protect against cognitive impairment [21]. However, specific studies involving 3,221 and 1,043 individuals found no significant differences in cognitive function between the experimental and control groups when assessed using tests like the Mini-Mental State Examination, word learning, digit span, and verbal fluency [22]. Furthermore, another observational study on AD and FAs suggested that supplementation with FAs like DHA, LA, or eicosapentaenoic acid (EPA) could reduce the risk of AD [23]. Yet, Mendelian randomization (MR) analysis of AD and FAs revealed that none of the PUFA showed a statistically significant association with the risk of AD [24]. Through liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis from the parietal cortex of 25 DLB patients and 17 age-matched controls, results related that altered immunoreactivities of brain lysolipid DHA transporter, MFSD2A, and the FABP5 in DLB parietal cortex [25]. Recent research has suggested a possible connection between dietary factors and the risk of developing DLB [26]. Omega-3 fatty acids from fish and plant sources are considered to be promising non-medical alternatives to improve brain functions for their protective antioxidative effect on cell membranes and potential neurochemical mechanisms related to amyloid pathology [26]. However, no convincing evidence was found in comparable randomized, placebo-controlled investigations of omega-3 PUFA in patients with AD from mild to moderate symptoms [26]. Additionally, the effects of FAs on DLB remain unclear and require further investigation. Given DLB’s status as the second most common type of cognitive impairment, further understanding of the relationship between DLB and FAs holds significant implications for the clinical diagnosis, treatment, healthy guidance, disease prevention, and treatment target discovery in patients with DLB. To address this gap, we employed univariate and multivariate MR analysis to verify the relationship between FAs and DLB, providing a theoretical basis for subsequent relevant studies.

MR analysis employs genetic polymorphisms and is naturally randomized to minimize confounding factors. This method is advantageous because the random allocation of alleles at conception inherently reduces the influence of confounding variables, and their presence from birth eliminates the possibility of reverse causation. Our research aimed to explore the potential associations between FAs and the risk of DLB using MR analysis and to overcome the limitations inherent in traditional epidemiological approaches. Furthermore, we harnessed recent genetic data from a whole-genome sequencing in a cohort comprising 2,591 patients diagnosed with DLB and 4027 neurologically healthy individuals [27], along with another European cohort of metabolic biomarkers from the UK Biobank. This comprehensive approach allowed for a more robust and nuanced investigation into the possible links between FAs and DLB risk.

MATERIALS AND METHODS

We conducted a bidirectional two-sample MR analysis to explore the causal relationships between each FAs and DLB. The MR analysis relied on three assumptions, and genetic variants used as instruments had to satisfy these critical assumptions as well (Fig. 1). First, the genetic variants used as instruments had to be associated with the relevant risk factor. Second, there should be no hidden variables capable of confounding the relationship between these genetic variants and the outcome. Finally, the genetic variants’ impact on the outcomes should be exclusively due to their influence on the risk factor [28]. Additionally, to avoid the influence of confounding factors between collinearity and FAs, the multivariable Mendelian randomization (MVMR) method was utilized to eliminate the bias [29].

Fig. 1

Major assumptions and procedures for the MR Analysis. Including three assumptions for the two-sample MR study and the procedures for the MR Analysis. LD, linkage disequilibrium; F-statistics, “strength” of the instrumental variable; MR, Mendelian randomization; r², proportion of variance in exposure variable explained by SNPs; LA, linoleic acid; DHA, docosahexaenoic acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid; TFA, total fatty acid.

This study adhered to the recommendations of the Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization (STROBE-MR) reporting guideline [30].

Data sources

Genome-wide association study (GWAS) summary statistics for DLB were sourced from a comprehensive whole-genome sequencing project. A total of 5,154 participants of European ancestry (2,981 DLB and 2,173 healthy controls) were recruited across 17 European and 27 North American sites/consortia to create a genomic resource for DLB research. Additionally, they obtained healthy controls from the Wellderly cohort (n = 1,202) and European-ancestry control genomes (n = 1,016). This brought the cohort comprising 2,981 individuals diagnosed with DLB and 4,391 neurologically healthy European individuals [27]. After quality control, 2,591 DLB patients and 4,027 healthy controls were available for study. The appropriate institutional review boards of participating institutions approved the study (03-AG-N329, NCT02014246), and informed consent was obtained from all subjects or their surrogate decision-makers, according to the Declaration of Helsinki [27]. Genome sequence data were processed using the pipeline standard developed by the Centers for Common Disease Genomics (CCDG https://www.genome.gov/27563570/). For sample-level and variant-level quality control checks, genomes were excluded from analysis for multiple reasons [27].

Instruments for FAs were selected from another large-scale IEU GWAS of Metabolic Biomarkers in the UK Biobank measured by Nightingale Health 2020 (met-d), which contained 114,999 individuals and covered eight FA indices: LA, DHA, omega-3 fatty acid, omega-6 fatty acid, PUFA, monounsaturated fatty acid (MUFA), total fatty acid (TFA), and SFA. The UKB has research tissue bank approval from the North West Multi-center Research Ethics Committee (https://www.ukbiobank.ac.uk/learn-more-about-uk-biobank/about-us/ethics) and provided oversight for this study. Confounders such as age, sex, race, education, etc., were adjusted. Notably, there was no overlap of participants between the exposure and outcome databases. Table 1 provides the basic information of the GWAS for both DLB and FAs. All data are accessible at https://gwas.mrcieu.ac.uk/.

Table 1

GWAS dataset used in the Mendelian randomization

	Sample size
Phenotype	Cases, n	Controls, n	Population	PMID
(% European)
LA	114,999	NA	100	NA
DHA	114,999	NA	100	NA
Omega-3 fatty acid	114,999	NA	100	NA
Omega-6 fatty acid	114,999	NA	100	NA
PUFA	114,999	NA	100	NA
MUFA	114,999	NA	100	NA
TFA	114,999	NA	100	NA
SFA	114,999	NA	100	NA
DLB	2,591	4,027	100	33589841

LA, linoleic acid; DHA, docosahexaenoic acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid; TFA, total fatty acid.

Selection of instrumental variable

We utilized single-nucleotide polymorphisms (SNPs) significantly associated with each trait, specifically those with a genome-wide significance level (p < 5.0×10^–8). To ensure the independence of these loci, SNPs were clumped using a linkage disequilibrium (LD) threshold of r² < 0.001 and a maximum distance of 1000 kb in PLINK [31]. We excluded SNPs without proxies and also removed palindromic SNPs due to the unavailability of allele frequency information in the FAs GWAS, preventing certainty about the alignment of these SNPs in the same direction for both exposure and outcome. The “TwoSampleMR” package was utilized to harmonize the exposure-outcome datasets. This process involved inferring forward strand ambiguous SNPs using allele frequency information. However, strand ambiguous SNPs with intermediate allele frequencies (AF > 0.42-the default value in TwoSampleMR) were excluded from the analysis [32]. An F-statistic >10 indicated a low risk of weak instrument bias using the formula [33, 34]. Additionally, SNPs in outcomes with p-values <10^–8 were excluded.

Statistical analysis

We conducted a comprehensive bidirectional two-sample MR and MVMR analysis with multiple methods to explore the causal relationship between FAs and DLB, such as the inverse variance weighted (IVW) method [35], MR-Egger regression [36], weighted-median estimator [37], and both simple mode and weighted-mode analyses. Each method offered unique insights and controls for different biases, thereby providing a robust and comprehensive understanding of the potential causal links between FAs and DLB. The IVW method is a statistical approach predominately utilized for combining results from various studies or data sources, especially in meta-analyses. The fundamental concept of IVW lies in allocating weights to individual studies in proportion to their respective variances or the degree of uncertainty they embody [35]. The IVW method will return an unbiased estimate in the absence of horizontal pleiotropy or when horizontal pleiotropy is balanced [38]. All causal estimates were converted to odd ratios (OR), as the outcome was dichotomous. This process allows for the synthesis of a consolidated effect estimate, which reflects a balanced consideration of all included studies, prioritizing those with greater precision and stability in their results. In situations where only a single IV was available, we employed the Wald ratio for our analysis. This method is particularly suited to instances with limited instrumental variables. Our primary analytical strategy focused on determining the correlation between genetic predisposition to a specific exposure and its associated outcomes. To accomplish this, we utilized the IVW method, which was integrated within a multiplicative random-effects model framework. The MR-Egger method, which is comparable to Egger’s regression used in conventional meta-analysis, offers a robust mechanism for detecting and correcting pleiotropic effects [39]. Horizontal pleiotropy occurs when genetic variants influence multiple traits, which can potentially confound analytical results. By adjusting for them, MR-Egger ensures a more accurate and reliable interpretation of the genetic associations and their potential causal relationships [36]. When the Egger intercept significantly deviated from zero, there was evidence of horizontal pleiotropy. In MR studies, this method is particularly valuable as it helps to address concerns about the causal inference being biased due to these pleiotropic effects. The weighted-median approach offers a robust estimate that is typically less susceptible to certain types of biases in comparison to methods like simple averages or weighted means. In MR studies, the weighted-median estimator can be used to provide a causal estimate when multiple genetic variants are employed as instrumental variables. Each variant’s effect estimate is weighted by its precision (inversely proportional to the variance) when calculating the median [37]; it helps in providing a more reliable estimate when some genetic variants are invalid instruments (e.g., due to pleiotropy). Additionally, to further enhance the robustness and reliability of our causal relationship estimations, we incorporated both the simple and weighted mode methods. For MVMR, SNPs in linkage disequilibrium (r²≥ 0.001) were excluded. Conditional F-statistics were used to assess weak instrument bias, assuming no covariance between the exposures, which likely gives a lower bound [40]. Multivariable Q-statistics were employed to evaluate instrument pleiotropy under the assumption of no covariance between the exposures, thereby likely establishing an upper boundary [40, 41]. Additionally, the Benjamini-Hochberg procedure (false discovery rate [FDR] correction) was utilized to justify these assessments [41]. Results with p-values smaller than 0.05 that did not withstand the FDR correction were reported as nominal associations [42].

IVW, MR-Egger, and MVMR methods all demand NO Measurement Error (NOME) assumption, where SNP-exposure associations are already known rather than estimated. F-statistic was calculated using the following formula [43]: $F = \frac{R^{2} (N - 2)}{1 - R^{2}}$ . Here, R² is the variance of the risk of Parkinson’s disease (PD) explained by each instrument and N is the sample size of FAs GWAS database. R² was calculated as follows [16]: $R^{2} = \frac{2 \times EAF \times (1 - EAF) \times {beta}^{2}}{2 \times EAF \times (1 - EAF) \times {beta}^{2} + 2 \times EAF \times (1 - EAF) \times N \times SE ({beta}^{2})}$ . Here, EAF is the effect allele frequency, beta is the estimated genetic effect on FAs, and SE (beta) is the standard error of the genetic effect. The mean F-statistic is widely used to measure the weak instrument bias whenever the genetic variants violate the NOME assumption. Generally, the mean F-statistic is required to be more than 10 to avoid weak instrument bias [16]. We conducted heterogeneity analysis using Cochran’s Q statistic in IVW and MR-egger (Q-p<0.05 indicating significant heterogeneity) [44]. Upon detecting substantial global heterogeneity, we proceeded with a local outlier test to detect outlying SNPs. Subsequently, the Mendelian Randomization Pleiotropy RESidual Sum and Outlier (MR-PRESSO) method [45] was utilized to detect and manage pleiotropic outlier SNPs [45]. This method involves three parts: 1) a “global test” assessing the extent of horizontal pleiotropy (with significant pleiotropy indicated by p < 0.05), 2) an “outlier-corrected causal estimate” providing adjustments for any significant pleiotropy identified, and 3) a “distortion test” evaluating how much the original and corrected estimates differ (p < 0.05 here suggests a significant alteration following corrections for pleiotropy) [45]. The latter two components of the test are implemented only if the global test yields p < 0.05 [45].

Leave-one-out analyses were performed to evaluate whether the MR results were overly dependent on any specific SNP. To mitigate issues of pleiotropy and avoid collinearity, we also performed MVMR to ensure the robustness of our findings. Visual inspections of heterogeneity and horizontal pleiotropy were conducted using forest plots and funnel plots. R software (Version 4.3.2) was used for all statistical analyses, with the MR Analysis package comprising “TwoSampleMR (version 0.5.6)” [46] and “MR-PRESSO (version 1.0)” [45].

Ethical considerations

This study employed study-level summary data, and ethical approval and participant informed consent were obtained.

RESULTS

In our study, we meticulously explore the bidirectional causal relationships between each identified trait. This approach allows us to understand not only how one trait may influence the other but also the possibility of reverse causation. By examining these relationships from both directions, we gain a more comprehensive and nuanced understanding of the interplay between these traits.

Effects of FAs on DLB

A total of 382 SNPs were selected as instrumental variables for various FAs, including LA, DHA, omega-3 fatty acid, omega-6 fatty acid, PUFA, MUFA, TFA, and SFA. The effects of these FAs on DLB are detailed in Table 2. Through eight common MR analyses, we observed that the genetic association of LA might be inversely related to DLB, as indicated by the IVW method ([odds ratio (OR) 1.337], [95% confidential interval (CI) 1.019–1.756], p_IVW = 0.036), suggesting a higher risk of DLB with increased levels of LA (Fig. 2A). No evidence of heterogeneity (p_{_IVWQ} = 0.075, p_{Egger _ Q} = 0.071) or pleiotropy (p_{MR - PRESSO} _globaltest = 0.056) exists (Supplementary Table 1: FAs-DLB). Additionally, a leave-one-out analysis confirmed the absence of outliers, indicating that the causal effect was not driven by any particular instrumental SNP (Fig. 2B). However, no significant causal relationships between DLB risk and other genetically predicted plasma FAs were found (Supplementary Table 1: FAs-DLB). The results for DHA, omega-3, omega-6, PUFA, MUFA, TFA, and SFA from MVMR corroborated those from the initial bidirectional two-sample MR. The MVMR results for LA showed no significant statistical association with DLB (p_IVW = 0.720, se = 2.39) (Supplementary Table 2: FAs-DLB MVMR), indicating a complex genetic association of LA with DLB potentially influenced by other factors.

Fig. 2

Two-sample MR analyses of LA on DLB. A) Scatter plots for the causal association between LA and DLB using IVW, MR egger, weighted median, and weighted mode. B) Leave-one-out plots for the causal association between LA and DLB.

Table 2

Full result of MR estimates for the association between LA and DLB

Exposure	Outcome	No. of SNPs	Method	β	OR	95% CI	SE	P	Heterogeneity p	Pleiotropy p
LA	DLB	42	MR-Egger	0.494	1.639	0.941 to 2.856	0.283	0.089	0.071	0.414
			WM	0.258	1.295	0.879 to 1.906	0.197	0.190	NA	NA
			IVW	0.291	1.337	1.019 to 1.756	0.139	0.036	0.075	NA
DHA	DLB	40	MR-Egger	–0.405	0.447	0.285 to 1.563	0.434	0.357	0.004	0.200
			WM	–0.062	0.435	0.572 to 1.544	0.253	0.807	NA	NA
			IVW	0.095	0.499	0.734 to 1.648	0.206	0.644	0.002	NA
Omega-3	DLB	42	MR-Egger	–0.245	0.782	0.599 to 1.022	0.137	0.080	0.109	0.029
			WM	–0.169	0.845	0.678 to 1.052	0.112	0.132	NA	NA
			IVW	–0.027	0.973	0.797 to 1.187	0.102	0.787	0.042	NA
Omega-6	DLB	51	MR-Egger	0.104	1.110	0.666 to 1.848	0.260	0.691	0.152	0.968
			WM	0.038	1.039	0.729 to 1.482	0.180	0.832	NA	NA
			IVW	0..095	1.100	0.853 to 1.418	0.130	0.465	0.176	NA
PUFA	DLB	54	MR-Egger	0.198	1.219	0.763 to 1.946	0.239	0.411	0.053	0.269
			WM	–0.004	0.996	0.717 to 1.384	0.168	0.982	NA	NA
			IVW	–0.030	0.970	0.760 to 1.240	0.125	0.810	0.048	NA
MUFA	DLB	51	MR-Egger	0.068	1.070	0.751 to 1.526	0.181	0.709	0.901	0.513
			WM	–0.005	0.995	0.723 to 1.369	0.163	0.973	NA	NA
			IVW	–0.026	0.974	0.783 to 1.213	0.112	0.817	0.910	NA
TFA	DLB	54	MR-Egger	0.035	1.036	0.706 to 1.520	0.195	0.857	0.739	0.900
			WM	–0.005	0.995	0.714 to 1.387	0.169	0.977	NA	NA
			IVW	0.015	1.015	0.814 to 1.266	0.113	0.893	0.770	NA
SFA	DLB	48	MR-Egger	–0.097	0.908	0.584 to 1.409	0.225	0.668	0.683	0.476
			WM	–0.010	0.990	0.682 to 1.436	0.190	0.957	NA	NA
			IVW	0.037	1.038	0.811 to 1.327	0.125	0.769	0.700	NA

LA, linoleic acid; DHA, docosahexaenoic acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid; TFA, total fatty acid; OR, odds ratio; CI, confidential interval; IVW, inverse variance weighted; WM, weighted median.

Effects of DLB on FAs

In the reverse analysis, the genetic predisposition to DLB did not significantly affect plasma FA levels (OR 0.996, 95% CI 0.978–1.015 p_LA - IVW = 0.691) (Supplementary Table 3: DLB-FAs).

DISCUSSION

The rise in the global aging population contributes to an upsurge in age-related diseases such as AD, PD, and DLB. DLB is a complex neurodegenerative disease with pathological features of both AD and PD. Previous preclinical and epidemiological studies have suggested FA supplements, derived from plants or fish, could serve as a non-pharmacological approach to maintaining neuronal integrity and slowing dementia progression [47]. However, MR analyses investigating the relationship between FAs and AD rejected the hypothesis that PUFA can reduce AD risk [24]. Similarly, a two-sample MR study exploring FAs and PD indicated that higher levels of arachidonic acid and possibly EPA are associated with an elevated risk of PD, while no significant association was found between PD risk and LA, DHA, docosapentaenoic acid, or α-linolenic acid (ALA, n3-PUFA) [29]. These MR findings all contrast with results from randomized controlled studies. To further investigate the correlation between FAs and DLB, we conducted this MR analysis, revealing a potential association between increased LA levels and a higher risk of developing DLB. Our MR analyses found that only LA showed a significant genetic association with the risk of DLB. The odds ratio (OR) for LA was 1.337 with a 95% confidence interval (CI) of 1.019–1.756 (p_IVW = 0.036). The absence of heterogeneity or pleiotropy indicates the absence of significant systemic variation or batch effects among the samples [39, 44]. This implies that experimental outcomes are reliable and stable, providing a solid foundation for further biological interpretation and clinical application.

FAs serve as biomarkers for dietary fat intake and metabolism and are implicated in various chronic diseases. PUFA addresses macrophage deficiencies in subjective cognitive impairment, mild cognitive impairment, and patients with AD by influencing transcriptome, enriching glycome, enhancing Aβ clearance, and finally benefiting the cognition of patients with early-stage AD [48, 49]. Prospective cohort studies investigating the link between FAs and cognitive function provide evidence that replacing SFA with LA is associated with an increased risk of dementia and cognitive decline while substituting marine omega-3 fatty acid for LA appears to decrease dementia risk [50]. Additionally, a different combination of FABP3 with other biomarkers like αS, t-tau, and p-tau in cerebral spinal fluid yielded a higher AUC in a differential diagnosis between AD and DLB [51]. Our study is the first MR analysis to investigate the causal relationship between FAs and DLB. We found that none of the FAs, except for LA, were associated with DLB risk. The results from our MR analysis, which are generally less susceptible to confounding factors than observational studies, do not support the previously suggested notion that FAs are protective against DLB.

αS is a prevalent neuronal protein that accumulates in insoluble inclusions in PD and DLB, contributing to neuronal dysfunction. αS is disordered in solution but can adopt an α-helical conformation upon binding to lipid membranes. This lipid-protein interaction plays a vital role in its biological function, such as synaptic plasticity [52]. Research indicates that PUFA can enhance the oligomerization of αS in mesencephalic neuronal cells. Furthermore, elevated levels of PUFA have been detected in the brain’s soluble fractions in both PD and DLB, with this increase associated with heightened membrane fluidity in neurons overexpressing αS [53]. In contrast, membrane fluidity and specific PUFA levels were found to decrease in mice models lacking αS [53]. These findings suggest that interactions between αS and PUFA play a critical role in regulating neuronal PUFA levels and the oligomerization state of αS, both under normal conditions and in synucleinopathies like DLB. On the one hand, alterations in PUFA levels can lead to aggregation of phosphorylated-αS and subsequent deposition into potentially cytotoxic oligomers that precede inclusions in dopaminergic cells [54]. The interaction between PUFA and αS facilitates the formation of soluble αS oligomers, which are linked to cytotoxicity, implying a detrimental role of PUFA in disease progression [54]. Intriguingly, these soluble αS oligomers precede the development of Lewy-like inclusions, which seem to possess protective characteristics [54]. This suggests that PUFA might also contribute to protective mechanisms, possibly by facilitating the transformation into less harmful inclusion forms. Besides the bidirectional of PUFA and the state of αS, PUFA also contributes to inflammation-related psychiatric and neurological disorders via the soluble epoxide hydrolases, enzymes present in all living organisms [55]. Thus, the role of PUFA in DLB is complex, potentially exerting both harmful and beneficial effects, depending on how they interact with αS and other cellular components. In a phospholipidomics analysis of postmortem parietal cortex samples from 25 DLB patients and 17 controls, a notable decrease was observed in 6 out of 46 DHA-containing phospholipid species in DLB [21]. Five of these species showed a negative correlation with soluble Aβ₄₂ levels and three with phosphorylated αS. Furthermore, these phospholipids were associated with presynaptic Rab3A and postsynaptic neurogranin deficits [21]. These findings indicate specific alterations in the DHA-containing phospholipid profile in DLB, which are linked to neuropathological burden and synaptopathy. α-Syn oligomers modified by protein with dopamine and DHA showed differentiation in aggregate size, conformation, tryptic digestion, and toxicity. They also exhibited different interactions with tau protein: the tau aggregate cross-seeded with the dopamine-modified αS oligomeric strain possessed a potent intracellular tau seeding propensity [56]. The DLB-causing protein αS interacts with lipids and FAs physiologically and pathologically, suggesting that targeting FA homeostasis for therapeutics is in its infancy [57].

A gas chromatography-mass spectrometry study analyzed FA profiles in the human parietal cortex of 27 cognitively normal age-matched controls, 16 with moderate AD, 30 with severe AD and DLB, a total of 24 FAs were identified in the brains of those with moderate AD and DLB [58]. Subjects with moderate AD and DLB showed significantly elevated levels of various FAs, including nervonic, lignoceric, cis-13,16-docosadienoic, arachidonic, cis-11,14,17-eicosatrienoic, erucic, behenic, α-linolenic, stearic, oleic, cis-10-heptanoic, and palmitic acids, indicating an association between AD or DLB pathology development and alterations in brain FA composition [58]. LA, an 18-carbon n-6 PUFA with a variety of effects on human physiology [59], is a bioactive FA with diverse effects on human physiology and pathophysiology [59], including skin barrier [60], immune system [61], cardiovascular [62, 63], and neurobiological functions [64]. As a natural ligand for peroxisome proliferator-activated receptors (PPARs), LA has the potential to moderate the pro-inflammatory state of microglia and rejuvenate their pro-regenerative properties, leading to improved recovery from demyelination injuries and enhanced spatial learning functions through the activation of PPAR-γ signaling pathways [65]. In recent decades, with improvements in living conditions and growing awareness about healthy life expectancy, there has been a significant increase in dietary intake of LA due to a shift towards polyunsaturated seed oils in our diets, coinciding with the increase in diet-induced obesity. Obesity, across all ages, has emerged as a significant health and economic challenge. Individuals with obesity face a heightened risk of developing chronic diseases, particularly metabolic disorders and neurodegenerative diseases like DLB [66]. Studies have shown that in certain brain regions, obese animals exhibit neurons with reduced spine density [67], disrupted myelin, and activated microglia [68]. These cellular changes could be attributed to the release of various pro-inflammatory molecules by expanding adipose tissue, including saturated and monounsaturated FAs, as well as cytokines like IL-6 and TNFα, which enter the bloodstream [69]. Consequently, this disrupts the functional and structural integrity of the blood-brain barrier and results in an increasing influx of lipids, pro-inflammatory molecules, and immune cells into the brain parenchyma, thereby fostering the onset of neuroinflammation [70].

Our bidirectional two-sample MR analysis indicated that a high level of LA may be associated with an increased risk of developing DLB. However, this result is inconsistent with the MVMR, highlighting the complexity of the underlying genetic architecture. The discrepancy between the two-sample MR and MVMR results can be interpreted in several ways. MVMR allows for the adjustment of multiple genetic variants simultaneously, potentially uncovering the influence of confounding factors not apparent in the two-sample MR analysis [40]. The incorporation of additional genetic instruments in MVMR might dilute the individual contribution of each variant, leading to a reduction in statistical power [71]. Furthermore, the possibility of collinearity between genetic variables in MVMR could affect the stability and interpretability of the model [72]. It also suggests that the initial associations observed might be driven by other genetic or environmental factors not accounted for in the two-sample MR analysis. The absence of a significant association in the MVMR analysis suggests that the initial relationship observed between LA and DLB risk may be more complex and influenced by a network of interrelated genetic and environmental factors rather than a direct causal pathway. Among all FAs, LA and ALA are two essential FAs that should be provided by food sources (e.g., vegetable oils and seeds) [73], hinting that the circulating level of FAs may influenced by the economic level and quality of life of participants. Additionally, although our study shows that there is no causal relationship between other FAs and DLB, different levels of FAs may influence the progression of DLB [74], which was not confirmed in the current studies. For example, conjugated linoleic acid, a special isomeric structure of LA, can increase the level of glucose transporters, along with its antioxidant and anti-inflammatory effects [75], which expands the therapeutic targets of these molecules and comes out as a suitable candidate for the treatment of multifactorial disease. Besides, circulating factors, including LA and glycocholic acid, have a direct effect on mitochondrial bioenergetics, and individual circulating factors identified to be associated with mitochondrial function are differently expressed in patients with DLB [76]. Metabolic syndrome, which is characterized by insulin resistance, high blood glucose, obesity, and dyslipidemia, is known to increase the risk of dementia accompanied by memory loss and depression [67]. The inadequate intake of LA and ALA is reported to be involved in neuropathology and neuropsychiatric diseases as well as imbalanced metabolic conditions. LA and ALA on metabolic-related dementia focus on insulin resistance, dyslipidemia, synaptic plasticity, cognitive function, and neuropsychiatric issues [77], suggesting that LA is an important FA for concurrent treatment of both metabolic syndrome and neurological problems.

Although MR provides less confounded estimates than observational studies, the results from this study should be interpreted in conjunction with some limitations. Firstly, we cannot be certain the selected SNPs do not violate the exclusion-restriction assumption. However, we used MR-PRESSO and MR-Egger regression to estimate the extent to which heterogeneity and pleiotropy may bias the reported results, and we have reported those results that are robust to the violations of MRs assumptions [55]. Secondly, in bidirectional two-sample MR, it is assumed that both samples come from comparable populations. Although the cohorts in our study are all from European ancestry, confounding due to population stratification cannot be ruled out [71]. Consequently, this situation limits the interpretation of these results to other ethnic groups. Furthermore, genetic instruments (SNPs) need to predict exposure strongly. We ensured this by strict criteria for choosing genetic variants as instrumental variables. Moreover, we considered the presence of selection bias due to selective survival on genetic makeup and competing risk of DLB, which may also violate the exclusion-restriction assumption. Finally, although bidirectional MR analysis proved the relationship between LA and DLB, the results from MVMR does not support this conclusion, thus the evidence for the positive results may be merely suggestive, and it should be approached with caution. It would be beneficial to conduct similar studies with diverse populations in order to confirm or refute the findings of this study. Other FAs, except these eight FAs in our research, like derenic acid and dihomo-gama-linolenic acid, may need further investigation.

In conclusion, this study underscores the importance of using comprehensive approaches like MR and MVMR to unravel the intricate web of genetic associations. It also highlights the need for cautious interpretation of MR findings, especially when examining complex diseases like DLB, where multiple genetic and non-genetic factors are likely at play. Our results warrant further investigation into the role of FAs and their genetic determinants in DLB, considering broader biological pathways and interactions.

AUTHOR CONTRIBUTIONS

Weijie Zhai (Data curation; Formal analysis; Methodology; Writing – original draft); Anguo Zhao (Data curation; Methodology; Writing – review & editing); Chunxiao Wei (Investigation; Writing – review & editing); Yanjiao Xu (Investigation; Writing – review & editing); Xinran Cui (Writing – review & editing); Yan Zhang (Writing – review & editing); Lingjie Meng (Writing – review & editing); Li Sun (Funding acquisition; Supervision; Visualization).

Footnotes

ACKNOWLEDGMENTS

We would like to thank the relevant consortia for making their data available. We would like to thank Editage (www.editage.cn) for English language editing.

FUNDING

The funding was provided by the General Program of the National Natural Science Foundation of China (Grant No. 82071442); the Jilin Provincial Department of Finance (Grant No. JLSWSRCZX2021-004); the Major Chronic Disease Program of the Ministry of Science and Technology of China (Grant No. 2018YFC1312301); STI2030-Major Projects (No. 2021ZD0201802); and Cohort study and mechanism exploration on the relationship between visceral fat and cognitive disorders (NO.2023FP0202706).

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

Every component of data used in this Mendelian randomization-based genome-wide genetic association study is identified as publicly available. For further inquiries, please contact the corresponding author.

The supplementary material is available in the electronic version of this article: .

References

Attems

, Toledo

, Walker

, Gelpi

, Gentleman

, Halliday

, Hortobagyi

, Jellinger

, Kovacs

, Lee

, Love

, McAleese

, Nelson

, Neumann

, Parkkinen

, Polvikoski

, Sikorska

, Smith

, Grinberg

, Thal

, Trojanowski

, McKeith

(2021) Neuropathological consensus criteria for the evaluation of Lewy pathology in post-mortem brains: A multi-centre study. Acta Neuropathol 141, 159–172.

Vann Jones

, O’Brien

(2014) The prevalence and incidence of dementia with Lewy bodies: A systematic review of population and clinical studies. Psychol Med 44, 673–683.

Toledo

, Abdelnour

, Weil

, Ferreira

, Rodriguez-Porcel

, Pilotto

, Wyman-Chick

, Grothe

, Kane

JPM

, Taylor

, Rongve

, Scholz

, Leverenz

, Boeve

, Aarsland

, McKeith

, Lewis

, Leroi

, Taylor

, Group

ILbdTMW

(2023) Dementia with Lewy bodies: Impact of co-pathologies and implications for clinical trial design. Alzheimers Dement 19, 318–332.

McKeith

, Boeve

, Dickson

, Halliday

, Taylor

, Weintraub

, Aarsland

, Galvin

, Attems

, Ballard

, Bayston

, Beach

, Blanc

, Bohnen

, Bonanni

, Bras

, Brundin

, Burn

, Chen-Plotkin

, Duda

, El-Agnaf

, Feldman

, Ferman

, Ffytche

, Fujishiro

, Galasko

, Goldman

, Gomperts

, Graff-Radford

, Honig

, Iranzo

, Kantarci

, Kaufer

, Kukull

, Lee

VMY

, Leverenz

, Lewis

, Lippa

, Lunde

, Masellis

, Masliah

, McLean

, Mollenhauer

, Montine

, Moreno

, Mori

, Murray

, O’Brien

, Orimo

, Postuma

, Ramaswamy

, Ross

, Salmon

, Singleton

, Taylor

, Thomas

, Tiraboschi

, Toledo

, Trojanowski

, Tsuang

, Walker

, Yamada

, Kosaka

(2017) Diagnosis and management of dementia with Lewy bodies: Fourth consensus report of the DLB Consortium. Neurology 89, 88–100.

Ruffmann

, Calboli

, Bravi

, Gveric

, Curry

, de Smith

, Pavlou

, Buxton

, Blakemore

, Takousis

, Molloy

, Piccini

, Dexter

, Roncaroli

, Gentleman

, Middleton

(2016) Cortical Lewy bodies and Abeta burden are associated with prevalence and timing of dementia in Lewy body diseases. Neuropathol Appl Neurobiol 42, 436–450.

Vogel

, Young

, Oxtoby

, Smith

, Ossenkoppele

, Strandberg

, La Joie

, Aksman

, Grothe

, Iturria-Medina

, Alzheimer’s Disease Neuroimaging I Pontecorvo

, Devous

, Rabinovici

, Alexander

, Lyoo

, Evans

, Hansson

(2021) Four distinct trajectories of tau deposition identified in Alzheimer’s disease. Nat Med 27, 871–881.

Kyselova

, Vitova

, Rezanka

(2022) Very long chain fatty acids. Prog Lipid Res 87, 101180.

Bhatt

, Budoff

, Mason

(2020) A Revolution in Omega-3 Fatty Acid Research. J Am Coll Cardiol 76, 2098–2101.

Behrouz

, Yari

(2022) A review on differential effects of dietary fatty acids on weight, appetite and energy expenditure. Crit Rev Food Sci Nutr 62, 2235–2249.

10.

Arellano

, Nardello-Rataj

, Szunerits

, Boukherroub

, Fameau

(2023) Saturated long chain fatty acids as possible natural alternative antibacterial agents: Opportunities and challenges. Adv Colloid Interface Sci 318, 102952.

11.

Monnard

, Dulloo

(2021) Polyunsaturated fatty acids as modulators of fat mass and lean mass in human body composition regulation and cardiometabolic health. Obes Rev 22 Suppl 2, e13197.

12.

Bertoni

, Abodi

, D’Oria

, Milani

, Agostoni

, Mazzocchi

(2023) Alpha-linolenic acid and cardiovascular events: A narrative review. Int J Mol Sci 24, 14319.

13.

Desale

, Chinnathambi

(2021) alpha- Linolenic acid modulates phagocytosis and endosomal pathways of extracellular Tau in microglia. Cell Adh Migr 15, 84–100.

14.

Leikin-Frenkel

, Schnaider Beeri

, Cooper

(2022) How alpha linolenic acid may sustain blood-brain barrier integrity and boost brain resilience against Alzheimer’s disease. Nutrients 14, 5091.

15.

Desale

, Dubey

, Chinnathambi

(2021) alpha-Linolenic acid inhibits Tau aggregation and modulates Tau conformation. Int J Biol Macromol 166, 687–693.

16.

Burgess

, Thompson

, Collaboration

CCG

(2011) Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol 40, 755–764.

17.

Sharon

, Bar-Joseph

, Frosch

, Walsh

, Hamilton

, Selkoe

(2003) The formation of highly soluble oligomers of alpha-synuclein is regulated by fatty acids and enhanced in Parkinson’s disease. Neuron 37, 583–595.

18.

Kawahata

, Sekimori

, Oizumi

, Takeda

, Fukunaga

(2023) Using fatty acid-binding proteins as potential biomarkers to discriminate between Parkinson’s disease and dementia with Lewy bodies: Exploration of a novel technique. Int J Mol Sci 24, 13267.

19.

Bereczki

, Francis

, Howlett

, Pereira

, Hoglund

, Bogstedt

, Cedazo-Minguez

, Baek

, Hortobagyi

, Attems

, Ballard

, Aarsland

(2016) Synaptic proteins predict cognitive decline in Alzheimer’s disease and Lewy body dementia. Alzheimers Dement 12, 1149–1158.

20.

Kawahata

, Fukunaga

(2020) Degradation of tyrosine hydroxylase by the ubiquitin-proteasome system in the pathogenesis of Parkinson’s disease and dopa-responsive dystonia. Int J Mol Sci 21, 3779.

21.

Chong

, Chai

, Xing

, Herr

, Wenk

, Francis

, Ballard

, Aarsland

, Silver

, Chen

, Cazenave-Gassiot

, Lai

MKP

(2023) Decreased DHA-containing phospholipids in the neocortex of dementia with Lewy bodies are associated with soluble Abeta(42), phosphorylated alpha-synuclein, and synaptopathology. Brain Pathol 33, e13190.

22.

Fratesi

, Hogg

, Young-Newton

, Patterson

, Charkhzarin

, Block Thomas

, Sharratt

, Stark

(2009) Direct quantitation of omega-3 fatty acid intake of Canadian residents of a long-term care facility. Appl Physiol Nutr Metab 34, 1–9.

23.

Wei

, Li

, Dong

, Tan

, Alzheimer’s Disease Neuroimaging Initiative Xu

(2023) The relationship of omega-3 fatty acids with dementia and cognitive decline: Evidence from prospective cohort studies of supplementation, dietary intake, and blood markers. Am J Clin Nutr 117, 1096–1109.

24.

Tomata

, Larsson

, Hagg

(2020) Polyunsaturated fatty acids and risk of Alzheimer’s disease: A Mendelian randomization study. Eur J Nutr 59, 1763–1766.

25.

Cloos

, Schissel

, Batra

, Donahue

, Wenos

, Kumar

, 2nd, Leinicke

, Thompson

, Langenfeld

(2023) Characteristics of pathologic complete response for locally advanced rectal cancer. Am J Surg 226, 873–877.

26.

Phillips

(2017) Omega 3 fatty acids for the treatment of dementia: A Cochrane review summary. Int J Nurs Stud 67, 83–84.

27.

Chia

, Sabir

, Bandres-Ciga

, Saez-Atienzar

, Reynolds

, Gustavsson

, Walton

, Ahmed

, Viollet

, Ding

, Makarious

, Diez-Fairen

, Portley

, Shah

, Abramzon

, Hernandez

, Blauwendraat

, Stone

, Eicher

, Parkkinen

, Ansorge

, Clark

, Honig

, Marder

, Lemstra

, St George-Hyslop

, Londos

, Morgan

, Lashley

, Warner

, Jaunmuktane

, Galasko

, Santana

, Tienari

, Myllykangas

, Oinas

, Cairns

, Morris

, Halliday

, Van Deerlin

, Trojanowski

, Grassano

, Calvo

, Mora

, Canosa

, Floris

, Bohannan

, Brett

, Gan-Or

, Geiger

, Moore

, May

, Kruger

, Goldstein

, Lopez

, Tayebi

, Sidransky

, American Genome

, Norcliffe-Kaufmann

, Palma

, Kaufmann

, Shakkottai

, Perkins

, Newell

, Gasser

, Schulte

, Landi

, Salvi

, Cusi

, Masliah

, Kim

, Caraway

, Monuki

, Brunetti

, Dawson

, Rosenthal

, Albert

, Pletnikova

, Troncoso

, Flanagan

, Mao

, Bigio

, Rodriguez-Rodriguez

, Infante

, Lage

, Gonzalez-Aramburu

, Sanchez-Juan

, Ghetti

, Keith

, Black

, Masellis

, Rogaeva

, Duyckaerts

, Brice

, Lesage

, Xiromerisiou

, Barrett

, Tilley

, Gentleman

, Logroscino

, Serrano

, Beach

, McKeith

, Thomas

, Attems

, Morris

, Palmer

, Love

, Troakes

, Al-Sarraj

, Hodges

, Aarsland

, Klein

, Kaiser

, Woltjer

, Pastor

, Bekris

, Leverenz

, Besser

, Kuzma

, Renton

, Goate

, Bennett

, Scherzer

, Morris

, Ferrari

, Albani

, Pickering-Brown

, Faber

, Kukull

, Morenas-Rodriguez

, Lleo

, Fortea

, Alcolea

, Clarimon

, Nalls

, Ferrucci

, Resnick

, Tanaka

, Foroud

, Graff-Radford

, Wszolek

, Ferman

, Boeve

, Hardy

, Topol

, Torkamani

, Singleton

, Ryten

, Dickson

, Chio

, Ross

, Gibbs

, Dalgard

, Traynor

, Scholz

(2021) Genome sequencing analysis identifies new loci associated with Lewy body dementia and provides insights into its genetic architecture. Nat Genet 53, 294–303.

28.

Davies

, Holmes

, Davey Smith

(2018) Reading Mendelian randomisation studies: A guide, glossary, and checklist for clinicians. BMJ 362, k601.

29.

Zhu

, Huang

, Kang

, Chen

, Liu

(2023) Associations between polyunsaturated fatty acid concentrations and Parkinson’s disease: A two-sample Mendelian randomization study. Front Aging Neurosci 15, 1123239.

30.

Skrivankova

, Richmond

, Woolf

BAR

, Davies

, Swanson

, VanderWeele

, Timpson

, Higgins

JPT

, Dimou

, Langenberg

, Loder

, Golub

, Egger

, Davey Smith

, Richards

(2021) Strengthening the reporting of observational studies in epidemiology using mendelian randomisation (STROBE-MR): Explanation and elaboration. BMJ 375, n2233.

31.

Purcell

, Neale

, Todd-Brown

, Thomas

, Ferreira

, Bender

, Maller

, Sklar

, de Bakker

, Daly

, Sham

(2007) PLINK: A tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 81, 559–575.

32.

Hemani

, Zheng

, Elsworth

, Wade

, Haberland

, Baird

, Laurin

, Burgess

, Bowden

, Langdon

, Tan

, Yarmolinsky

, Shihab

, Timpson

, Evans

, Relton

, Martin

, Davey Smith

, Gaunt

, Haycock

(2018) The MR-Base platform supports systematic causal inference across the human phenome. Elife 7, e34408.

33.

Pierce

, Ahsan

, Vanderweele

(2011) Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants. Int J Epidemiol 40, 740–752.

34.

McCarthy

, Das

, Kretzschmar

, Delaneau

, Wood

, Teumer

, Kang

, Fuchsberger

, Danecek

, Sharp

, Luo

, Sidore

, Kwong

, Timpson

, Koskinen

, Vrieze

, Scott

, Zhang

, Mahajan

, Veldink

, Peters

, Pato

, van Duijn

, Gillies

, Gandin

, Mezzavilla

, Gilly

, Cocca

, Traglia

, Angius

, Barrett

, Boomsma

, Branham

, Breen

, Brummett

, Busonero

, Campbell

, Chan

, Chen

, Chew

, Collins

, Corbin

, Smith

, Dedoussis

, Dorr

, Farmaki

, Ferrucci

, Forer

, Fraser

, Gabriel

, Levy

, Groop

, Harrison

, Hattersley

, Holmen

, Hveem

, Kretzler

, Lee

, McGue

, Meitinger

, Melzer

, Min

, Mohlke

, Vincent

, Nauck

, Nickerson

, Palotie

, Pato

, Pirastu

, McInnis

, Richards

, Sala

, Salomaa

, Schlessinger

, Schoenherr

, Slagboom

, Small

, Spector

, Stambolian

, Tuke

, Tuomilehto

, Van den Berg

, Van Rheenen

, Volker

, Wijmenga

, Toniolo

, Zeggini

, Gasparini

, Sampson

, Wilson

, Frayling

, de Bakker

, Swertz

, McCarroll

, Kooperberg

, Dekker

, Altshuler

, Willer

, Iacono

, Ripatti

, Soranzo

, Walter

, Swaroop

, Cucca

, Anderson

, Myers

, Boehnke

, McCarthy

, Durbin

, Haplotype Reference C (2016) A reference panel of 64,976 haplotypes for genotype imputation. Nat Genet 48, 1279–1283.

35.

Burgess

, Butterworth

, Thompson

(2013) Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol 37, 658–665.

36.

Burgess

, Thompson

(2017) Interpreting findings from Mendelian randomization using the MR-Egger method. Eur J Epidemiol 32, 377–389.

37.

Bowden

, Davey Smith

, Haycock

, Burgess

(2016) Consistent estimation in Mendelian randomization with some invalid instruments Using a weighted median estimator. Genet Epidemiol 40, 304–314.

38.

Bowden

, Del Greco

, Minelli

, Davey Smith

, Sheehan

, Thompson

(2017) A framework for the investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat Med 36, 1783–1802.

39.

Bowden

, Davey Smith

, Burgess

(2015) Mendelian randomization with invalid instruments: Effect estimation and bias detection through Egger regression. Int J Epidemiol 44, 512–525.

40.

Sanderson

, Davey Smith

, Windmeijer

, Bowden

(2019) An examination of multivariable Mendelian randomization in the single-sample and two-sample summary data settings. Int J Epidemiol 48, 713–727.

41.

Sesia

, Bates

, Candes

, Marchini

, Sabatti

(2021) False discovery rate control in genome-wide association studies with population structure. Proc Natl Acad Sci U S A 118, e2105841118.

42.

Yeung

CHC

, Au Yeung

, Schooling

(2022) Association of autoimmune diseases with Alzheimer’s disease: A mendelian randomization study. J Psychiatr Res 155, 550–558.

43.

Papadimitriou

, Dimou

, Tsilidis

, Banbury

, Martin

, Lewis

, Kazmi

, Robinson

, Albanes

, Aleksandrova

, Berndt

, Timothy Bishop

, Brenner

, Buchanan

, Bueno-de-Mesquita

, Campbell

, Castellvi-Bel

, Chan

, Chang-Claude

, Ellingjord-Dale

, Figueiredo

, Gallinger

, Giles

, Giovannucci

, Gruber

, Gsur

, Hampe

, Hampel

, Harlid

, Harrison

, Hoffmeister

, Hopper

, Hsu

, Maria Huerta

, Huyghe

, Jenkins

, Keku

, Kuhn

, La Vecchia

, Le Marchand

, Li

, Lindblom

, Lindor

, Lynch

, Markowitz

, Masala

, May

, Milne

, Monninkhof

, Moreno

, Newcomb

, Offit

, Perduca

, Pharoah

PDP

, Platz

, Potter

, Rennert

, Riboli

, Sanchez

, Schmit

, Schoen

, Severi

, Sieri

, Slattery

, Song

, Tangen

, Thibodeau

, Travis

, Trichopoulou

, Ulrich

, van Duijnhoven

FJB

, Van Guelpen

, Vodicka

, White

, Wolk

, Woods

, Wu

, Peters

, Gunter

, Murphy

(2020) Physical activity and risks of breast and colorectal cancer: A Mendelian randomisation analysis. Nat Commun 11, 597.

44.

Greco

, Minelli

, Sheehan

, Thompson

(2015) Detecting pleiotropy in Mendelian randomisation studies with summary data and a continuous outcome. Stat Med 34, 2926–2940.

45.

Verbanck

, Chen

, Neale

, Do

(2018) Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet 50, 693–698.

46.

Hemani

, Tilling

, Davey Smith

(2017) Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet 13, e1007081.

47.

Lin

, O’Connor

, Rossom

, Perdue

, Burda

, Thompson

, Eckstrom

(2013) In Screening for Cognitive Impairment in Older Adults: An Evidence Update for the U.S. Preventive Services Task Force, Rockville (MD).

48.

Dover

, Moseley

, Biskaduros

, Paulchakrabarti

, Hwang

, Hammock

, Choudhury

, Kaczor-Urbanowicz

, Urbanowicz

, Morselli

, Dang

, Pellegrini

, Paul

, Bentolila

, Fiala

(2023) Polyunsaturated fatty acids mend macrophage transcriptome, glycome, and phenotype in the patients with neurodegenerative diseases, including Alzheimer’s disease. J Alzheimers Dis 91, 245–261.

49.

Marin

, Fabelo

, Martin

, Garcia-Esparcia

, Ferrer

, Quinto-Alemany

, Diaz

(2017) Anomalies occurring in lipid profiles and protein distribution in frontal cortex lipid rafts in dementia with Lewy bodies disclose neumical traits partially shared by Alzheimer’s and Parkinson’s diseases. Neurobiol Aging 49, 52–59 roche.

50.

Koch

, Furtado

, DeKosky

, Fitzpatrick

, Lopez

, Kuller

, Mukamal

, Jensen

(2021) Case-cohort study of plasma phospholipid fatty acid profiles, cognitive function, and risk of dementia: A secondary analysis in the Ginkgo Evaluation of Memory Study. Am J Clin Nutr 114, 154–162.

51.

Chiasserini

, Biscetti

, Eusebi

, Salvadori

, Frattini

, Simoni

, De Roeck

, Tambasco

, Stoops

, Vanderstichele

, Engelborghs

, Mollenhauer

, Calabresi

, Parnetti

(2017) Differential role of CSF fatty acid binding protein 3, alpha-synuclein, and Alzheimer’s disease core biomarkers in Lewy body disorders and Alzheimer’s dementia. Alzheimers Res Ther 9, 52.

52.

Galvagnion

(2017) The role of lipids interacting with alpha-synuclein in the pathogenesis of Parkinson’s disease. J Parkinsons Dis 7, 433–450.

53.

Sharon

, Bar-Joseph

, Mirick

, Serhan

, Selkoe

(2003) Altered fatty acid composition of dopaminergic neurons expressing alpha-synuclein and human brains with alpha-synucleinopathies. J Biol Chem 278, 49874–49881.

54.

Assayag

, Yakunin

, Loeb

, Selkoe

, Sharon

(2007) Polyunsaturated fatty acids induce alpha-synuclein-related pathogenic changes in neuronal cells. Am J Pathol 171, 2000–2011.

55.

Hashimoto

(2019) Role of Soluble Epoxide Hydrolase in Metabolism of PUFAs in Psychiatric and Neurological Disorders. Front Pharmacol 10, 36.

56.

Sengupta

, Puangmalai

, Bhatt

, Garcia

, Zhao

, Kayed

(2020) Polymorphic alpha-synuclein strains modified by dopamine and docosahexaenoic acid interact differentially with tau protein. Mol Neurobiol 57, 2741–2765.

57.

Fanning

, Cirka

, Thies

, Jeong

, Niemi

, Yoon

, Ho

GPH

, Pacheco

, Dettmer

, Liu

, Clish

, Hodgetts

, Hutchinson

, Muratore

, Caldwell

, Selkoe

(2022) Lipase regulation of cellular fatty acid homeostasis as a Parkinson’s disease therapeutic strategy. NPJ Parkinsons Dis 8, 74.

58.

Nasaruddin

, Pan

, McGuinness

, Passmore

, Kehoe

, Holscher

, Graham

, Green

(2018) Evidence that parietal lobe fatty acids may be more profoundly affected in moderate Alzheimer’s disease (AD) pathology than in severe AD pathology. Metabolites 8, 69.

59.

Guyenet

, Carlson

(2015) Increase in adipose tissue linoleic acid of US adults in the last half century. Adv Nutr 6, 660–664.

60.

Ziboh

(1992) Prostaglandins, leukotrienes, and hydroxy fatty acids in epidermis. Semin Dermatol 11, 114–120.

61.

Harbige

(1998) Dietary n-6 and n-3 fatty acids in immunity and autoimmune disease. Proc Nutr Soc 57, 555–562.

62.

Farvid

, Ding

, Pan

, Sun

, Chiuve

, Steffen

, Willett

, Hu

(2014) Dietary linoleic acid and risk of coronary heart disease: A systematic review and meta-analysis of prospective cohort studies. Circulation 130, 1568–1578.

63.

Ramsden

, Zamora

, Leelarthaepin

, Majchrzak-Hong

, Faurot

, Suchindran

, Ringel

, Davis

, Hibbeln

(2013) Use of dietary linoleic acid for secondary prevention of coronary heart disease and death: Evaluation of recovered data from the Sydney Diet Heart Study and updated meta-analysis. BMJ 346, e8707.

64.

Gibson

, Muhlhausler

, Makrides

(2011) Conversion of linoleic acid and alpha-linolenic acid to long-chain polyunsaturated fatty acids (LCPUFAs), with a focus on pregnancy, lactation and the first 2 years of life. Matern Child Nutr 7 Suppl 2, 17–26.

65.

Zhou

, Dong

, Hu

, Tang

, Chu

, Chen

, Yang

, Chen

, Wu

, Wang

, Qin

, Tian

(2023) Staged suppression of microglial autophagy facilitates regeneration in CNS demyelination by enhancing the production of linoleic acid. Proc Natl Acad Sci U S A 120, e2209990120.

66.

Misiak

, Leszek

, Kiejna

(2012) Metabolic syndrome, mild cognitive impairment and Alzheimer’s disease–the emerging role of systemic low-grade inflammation and adiposity. Brain Res Bull 89, 144–149.

67.

Huang

, Wang

, Xiang

(2022) Bidirectional communication between brain and visceral white adipose tissue: Its potential impact on Alzheimer’s disease. EBioMedicine 84, 104263.

68.

Alkan

, Altunkaynak

, Gultekin

, Baycu

(2021) Hippocampal neural cell loss in high-fat diet-induced obese rats-exploring the protein networks, ultrastructure, biochemical and bioinformatical markers. J Chem Neuroanat 114, 101947.

69.

Alexaki

(2021) The impact of obesity on microglial function: Immune, metabolic and endocrine perspectives. Cells 10, 1584.

70.

Rudge

(2022) A new hypothesis for Alzheimer’s disease: The lipid invasion model. J Alzheimers Dis Rep 6, 129–161.

71.

Sanderson

, Spiller

, Bowden

(2021) Testing and correcting for weak and pleiotropic instruments in two-sample multivariable Mendelian randomization. Stat Med 40, 5434–5452.

72.

Lee

, Lee

, Paik

(2008) Polyunsaturated fatty acid levels in the cerebrospinal fluid of patients with Parkinson’s disease and multiple system atrophy. Mov Disord 23, 309–310.

73.

Simopoulos

(2008) The importance of the omega-6/omega-3 fatty acid ratio in cardiovascular disease and other chronic diseases. Exp Biol Med (Maywood) 233, 674–688.

74.

Wieckowska-Gacek

, Mietelska-Porowska

, Wydrych

, Wojda

(2021) Western diet as a trigger of Alzheimer’s disease: From metabolic syndrome and systemic inflammation to neuroinflammation and neurodegeneration. Ageing Res Rev 70, 101397.

75.

Cuciniello

, Luongo

, Ferramosca

, Lunetti

, Rotondi-Aufiero

, Crispi

, Zara

, Maurano

, Filosa

, Bergamo

(2022) Conjugated linoleic acid downregulates Alzheimer’s hallmarks in aluminum mouse model through an Nrf2-mediated adaptive response and increases brain glucose transporter levels. Free Radic Biol Med 191, 48–58.

76.

Amick

, Mahapatra

, Gao

, Dewitt

, Craft

, Jain

, Molina

AJA

(2022) Plasma glycocholic acid and linoleic acid identified as potential mediators of mitochondrial bioenergetics in Alzheimer’s dementia. Front Aging Neurosci 14, 954090.

77.

Kim

, Song

(2024) Important roles of linoleic acid and alpha-linolenic acid in regulating cognitive impairment and neuropsychiatric issues in metabolic-related dementia. Life Sci 337, 122356.