Interplay of Large Neutral Amino Acids, Metabolic Syndrome, and Apolipoprotein E ε4 on Brain Integrity at Midlife Open Access

The pathogenesis of dementia is characterized by its heterogeneous and multifactorial nature contributed by both genetic predispositions and modifiable risk factors [1‒4]. The apolipoprotein E gene (ApoE ε4) allele emerges as the most prevalent genetic risk factor for late-onset Alzheimer’s disease. However, up to 40% of dementia cases could be attributed to modifiable risk factors such as hypertension, obesity, and dietary patterns at midlife [5]. These factors also play roles in vascular cognitive impairment development [6, 7]. Moreover, the interactions and cumulative effects of more than one preclinical elevation of cardiovascular risk factor may have a higher predictive value for cognitive impairment above and beyond the sum of individual conditions [8, 9]. Such a cluster of risk factors is known as metabolic syndrome (MetS) [10]. MetS is also recognized for its association with cognitive decline and brain abnormalities [11]. Such associations between MetS and cognitive dysfunction are particularly strong among ApoE ε4 carriers [12], suggesting a potential synergistic effect between MetS and ApoE ε4 in predicting cognitive decline. This underscores the critical need for early detection and management of MetS to mitigate its impact on brain health and cognitive function.

Given that MetS and neurodegenerative diseases linked to cognitive decline share many common cardiometabolic pathways, improving nutrition could potentially lower the risk for both MetS and MetS-related cognitive impairment [13]. In particular, proteins and their constituent amino acids are crucial for maintaining cellular function, especially in brain cells. Large neutral amino acids (LNAAs), such as tryptophan and phenylalanine, are essential amino acids that must be obtained through diet. The changes in their amount play a role in cognitive decline [14], as variations in brain tryptophan and phenylalanine concentrations modify the synthesis and release of serotonin and the catecholamine neurotransmitters dopamine, norepinephrine, and epinephrine, respectively [15]. Yet exploration of the interactions between LNAA, MetS, and ApoE ε4 status in influencing brain structure has remained largely unexplored.

The potential for cognitive impairment in individuals bearing metabolic risk factors extends beyond the direct impact of MetS components on cognition to encompass the cerebrovascular alterations that may compromise brain structure [8, 16]. White matter hyperintensities (WMHs), which are macrostructural lesions predominantly of vascular origin, serve as clinical predictors of cognitive decline and have demonstrated a positive association with MetS components [17‒19]. The presence of WMH underscores the crucial role of cerebrovascular health in cognitive integrity and highlights the necessity of investigating how MetS and its individual components may exacerbate white matter alterations. This underscores the imperative to examine not only the individual impact of MetS components, but also their collective influence on white matter integrity.

Building on this existing body of knowledge, this study aims to fill critical gaps by delving into the complex interrelations among dietary LNAA, MetS components, and genetic predispositions in the context of brain health. By leveraging advanced neuroimaging techniques, we aim to better understand the underlying physiological processes that influence white matter brain structure in cognitively unimpaired individuals at midlife with varied metabolic and genetic risks.

This study included 68 middle-aged adults aged 40–61, from the Neural Consequences of Metabolic Syndrome study, a project that assessed the impact of metabolic and cardiovascular health on brain health in adults in midlife. Participants with a history of neurological disease, major psychiatric illness, previous hospitalization of substance use, or contraindications of MRI were excluded. The study protocol consisted of two laboratory visits: a health assessment visit and a neuroimaging visit completed within 1 month of each other. Global cognitive functioning was assessed by Mini-Mental State Examination (MMSE) [20] to exclude participants with scores below 27 based on the scoring range to increase confidence in our sample includes mostly cognitively unimpaired midlife adults [21, 22]. All participants provided written informed consent for all study procedures, which were approved by the University of Texas at Austin Institutional Review Board (Approval No. 2011070025).

Participants underwent a general health assessment that included blood pressure monitoring and fasted blood draw to obtain concentrations of glucose, triglycerides, total cholesterol, and high-density lipoprotein cholesterol using a standard enzymatic technique during their first visit. Based on the examination completed during this visit, the number of MetS components (0–5) was calculated according to the unified criteria [23]. The categorical cut points for each component include the following: (a) waist circumference >102 cm in males and >88 cm in females; (b) triglyceride levels ≥150 mg/dL (1.7 mmol/L); (c) high-density lipoprotein-cholesterol levels <40 mg/dL (1.0 mmol/L) in males and <50 mg/dL (1.3 mmol/L) in females or drug treatment for dyslipidemia; (d) systolic blood pressure ≥130 mm Hg or diastolic blood pressure ≥85 mm Hg or antihypertensive medication; (e) fasting blood glucose levels ≥100 mg/dL or drug treatment for hyperglycemia. Serum concentrations of tryptophan, kynurenine, phenylalanine, and tyrosine were measured using liquid chromatography [24]. The ratios of (a) kynurenine and tryptophan, and (b) phenylalanine and tyrosine were calculated as indexes of tryptophan degradation and phenylalanine 4-hydroxylase (PAH) activity, respectively.

Structural MRI images were collected during the second visit. MRI acquisitions were performed on a 3-T Siemens Skyra scanner (Siemens Medical Solutions, Malvern, PA, USA) equipped with a 32-channel head coil. Anatomical scans (T₁ weighted images) of the entire brain were collected using high-resolution Magnetization-Prepared Rapid Acquisition Gradient-Echo (MPRAGE) sequence. T₂ images were acquired using a fluid attenuated inversion recovery (FLAIR) sequence.

The FLAIR images were processed using the Lesion Segmentation Tool version 1.2.3, an automated algorithm in Statistical Parametric Mapping 8. Based on spatial and intensity probabilities from T₁ images and hyperintensity outliers on T₂ FLAIR images, we assigned voxels to tissue probability maps and gave them a probability of being a white matter lesion. We applied an initial threshold of 0.30 to create lesion seeds, and it was used to generate the conservative lesion belief map from the gray and white matter voxels. Next, a growth algorithm grew these seeds toward a liberal lesion belief map that contained gray matter, white matter, and cerebrospinal fluid lesion belief maps. Finally, we used a threshold of 0.99 on the resulting lesion belief map to remove any voxels with a lower probability of being a lesion. The resulting total WMH volume was divided by total intracranial volume, obtained through the FreeSurfer Imaging Analysis Suite (http://surfer.nmr.mgh.harvard.edu), and multiplied by 100 to give a WMH ratio in units of percentage of intracranial volume.

Saliva samples were collected using the Oragene Discover (OGR-500) kit, and DNA was extracted using 500 μL of saliva with the prepIT·L2P kit from DNAgenotek. Genomic DNA was extracted and purified according to the manufacturer’s instructions. All purified samples were stored at −40°C before the genotyping process. Polymerase chain reaction (PCR) was conducted using APOE-Fwd4 (GCT GAT GGA CGA GAC CAT GAA GGA GTT) and APOE-snapR (GCC CCG GCC TGG TAG ACT GCC A) primers, and performed with 10 ng of DNA and 10 pmol of primer prior to amplification. The PCR amplification process consisted of initial denaturation at 95°C for 15 min, followed by 35 cycles of 95°C for 30 s, 65°C for 30 s, and 72°C for 30 s, with a concluding hold at 4°C.

Using PCR amplification and Sanger sequencing [25], ApoE genotyping was conducted with Variant Reporter Software from Life Technologies (Thermo Fisher Scientific) at the DNA Sequencing Facility at the University of Texas at Austin. In total, the ApoE genotypes of 68 human genomic DNA samples were determined. The study participants were divided into two categories: individuals carrying one or two copies of the ApoE ε4 allele (ApoE ε4 carrier) or those not carrying any copy of the ApoE ε4 allele (ApoE ε4 non-carrier). Due to the small sample size, this study combined ApoE4 heterozygous and homozygous individuals (n = 16) and compared the ApoE4 carrier group to the ApoE4 non-carrier group (n = 52).

To test the interaction effect of LNAA, MetS, and ApoE4 in the statistical prediction of WMH, this study conducted general linear models. As the statistical significance of the interaction effects may be undermined by the high correlations among interaction terms, we ran preliminary correlation analyses to examine multicollinearity between LNAA and ApoE4 and found that there was no significant relationship between ApoE4 and any of the LNAA. Hence, we are reasonably assured that the results of the interaction effects will not be vulnerable to multicollinearity. A natural log transformation was utilized on WMH resulting in the normality of the variable. Sex, age, and years of education were included as covariates. All results were analyzed using the R statistical software package, version 4.1.0 [26].

Selected subject characteristics are shown in Table 1 describe a cognitively unimpaired, ethnically diverse, middle-aged sample, representative of the population of the state of Texas. The serum concentrations of kynurenine, tryptophan, phenylalanine, and tyrosine for the sample are reported in Table 1. While the sample maintained generally healthy levels of LNAA, it is of note that the mean tyrosine level for our sample was slightly higher (>75th percentile), and the mean tryptophan level was below-average (∼25th percentile overall) relative to a cohort of healthy blood donors [27]. Table 2 summarizes intercorrelations among all variables of interest. The number of MetS components was significantly correlated with all LNAA except for tryptophan. The WMH volume was not associated with any LNAA or the number of MetS components. Aside from the expected correlations between kynurenine and tryptophan (r = 0.34, p < 0.005) and phenylalanine and tyrosine (r = 0.43, p < 0.001), tryptophan was positively and significantly associated with tyrosine (r = 0.42, p < 0.001).

Next, we explored the complex interplay between LNAA, the number of MetS components, the presence of the ApoE ε4 allele, and their combined effects on WMH in midlife adults, while adjusting for age, sex, and education. The analysis revealed a significant 3-way interaction effect between tryptophan levels, the number of MetS components, and ApoE ε4 on WMH (β = 0.042, SE = 0.018, p < 0.05; Fig. 1). Additionally, significant main effects were observed for the presence of ApoE ε4 allele (β = 6.203, SE = 2.816, p < 0.05) and age (β = 0.0416, SE = 0.012, p < 0.05) on WMH. Furthermore, significant interactions were found between tryptophan levels and ApoE ε4 (β = −0.107, SE = 0.051, p < 0.05), as well as between the number of MetS components and ApoE ε4 (β = −2.404, SE = 1.014, p < 0.05). The directionality of these effects suggests that worse metabolic health (indicated by a higher number of MetS components) and the presence of the ApoE ε4 allele are disadvantageous, resulting in higher WMH volumes. These findings imply that individuals with poorer metabolic health and a genetic predisposition (ApoE ε4 carriers) are more susceptible to changes in WMH volume influenced by tryptophan levels. Furthermore, these relationships are complex and are significantly influenced by the combined presence of metabolic and genetic risk factors. There were no direct effects of tryptophan levels, the number of MetS components, sex, and years of education on WMH.

Similarly, phenylalanine also showed a significant three-way interaction between phenylalanine levels, the number of MetS components, and ApoE ε4 on WMH (β = 0.044, SE = 0.013, p < 0.01; Fig. 2). Significant two-way interactions were also observed. Specifically, the interaction between MetS and the ApoE ε4 allele was found to significantly predict WMH (β = −3.244, SE = 0.971, p < 0.01). Additionally, individual effects of age (β = 0.043, SE = 0.017, p < 0.05) and sex (β = −0.642, SE = 0.274, p < 0.05) were significant. The directionality of the findings suggests that higher phenylalanine levels are associated with greater WMH volumes, indicating a potential adverse effect on white matter lesions particularly for individuals with higher number of MetS components and ApoE ε4 allele. The presence of the ApoE ε4 allele generally results in higher WMH volumes, indicating greater susceptibility to white matter lesions in these individuals. The interaction effects imply that individuals with poorer metabolic health and genetic predisposition to AD (ApoE ε4 carriers) show varying susceptibilities to changes in WMH volume influenced by phenylalanine levels. Specifically, the protective effect of phenylalanine on reducing WMH volume may be most evident in those with fewer MetS components and without the ApoE ε4 allele. The tyrosine analysis only showed marginally significant three-way interactions (β = 0.010, SE = 0.005, p = 0.059). However, the interaction between MetS and ApoE ε4 (β = −1.603, SE = 0.737, p < 0.05), as well as age (β = 0.041, SE = 0.019, p < 0.05), was significant predictors. The kynurenine analysis did not demonstrate significant three-way interactions, with only age showing a significant association with WMH (β = 0.048, SE = 0.020, p < 0.05).

The current study sought to understand the interactions between large natural amino acids, MetS, and the ApoE ε4 allele in relation to WMH volume in a midlife adult population. Our findings underscore the nuanced interplay between these factors, revealing a significant three-way interaction effect on WMH volume that could not be attributed to any single factor alone. These results align with existing literature that identifies the ApoE ε4 allele as a significant risk factor for brain pathology and vulnerability to WMH [11, 12].

Notably, our findings demonstrated that the interaction between serum tryptophan levels and the number of MetS components was significantly moderated by the ApoE ε4 allele, underscoring the potential for genetic factors to influence the risk profile of metabolic abnormalities on brain health. This observation aligns with prior studies indicating that ApoE ε4 carriers exhibit a heightened sensitivity to other risk factors for neurodegeneration [11]. Similarly, phenylalanine levels also presented a significant three-way interaction with MetS components and the ApoE ε4 allele, further highlighting the complexity of metabolic influences on brain integrity.

Interestingly, the absence of a direct relationship between WMH volume and any single LNAA or MetS component alone suggests that it is the synergistic effect of these variables, rather than their individual effects, that plays a crucial role in the development of midlife cerebral white matter changes. This aligns with the idea that the biological pathways leading to cognitive impairment and dementia are multifactorial and that a cumulative risk factor approach may be necessary for early detection and intervention [8, 9, 28‒30].

Given the observed interactions between LNAA levels, MetS, and ApoE ε4 allele status on WMH volume, there is a compelling case for investigating dietary interventions targeting LNAA modifications especially among individuals with genetic predisposition to dementia. Our findings underscore the importance of a multifactorial approach in both the research and management of neurodegenerative diseases, focusing on individualized risk profiles and targeted interventions [11, 12, 30, 31]. This means that a single, universal diet is unlikely to be effective for everyone. Rather, personalized nutrition strategies should be developed, taking into account individual variations in genetic makeup and metabolic health. Dietary interventions could specifically aim to optimize levels of tryptophan and phenylalanine, which have shown significant associations with brain structure alterations in our study. Specifically, individuals with the ApoE ε4 allele and more MetS components may benefit from diets that modulate phenylalanine and tryptophan levels, potentially reducing the risk of white matter lesions. On the other hand, enhancing the dietary intake of foods rich in these amino acids, such as chicken, turkey, dairy products, nuts, and seeds, might benefit those without MetS and without the ApoE ε4 allele by influencing their serum levels and potentially improving brain health [29].

Ultimately, our findings suggest a need for precision nutrition approaches that consider individual variations in genetic and metabolic conditions when devising dietary strategies aimed at preserving brain health [32, 33]. Such studies would not only validate the role of LNAA in neuroprotection but could also lead to more targeted dietary guidelines that effectively address the complex interplay of nutrition, metabolism, and genetics in aging populations.

Despite the advances made by this study, some limitations warrant consideration. First, the cross-sectional nature of the study precludes the establishment of causality. Longitudinal studies are needed to determine the temporal relationship between LNAA levels, MetS components, and the ApoE ε4 allele in the progression of WMH and cognitive decline. Second, while the sample size may be adequate for the exploratory nature of our pilot study, it is understood that such a small cohort may constrain the generalizability of our results. Additionally, we did not assess certain dietary variables such as macro-nutrient intake levels, which means that we cannot determine whether the effects of LNAA are specific to their biochemical properties or if they are indicative of a generally healthier diet. This limitation affects our ability to generalize the findings to broader dietary contexts and understand the precise role of dietary LNAA within participants’ overall food intake patterns. Lastly, statistical power is a recognized constraint in detecting the subtle yet potentially significant interaction effects in complex biological relationships [34], a limitation that is accentuated in studies with smaller sample sizes like ours. Future research with larger, more diverse populations is necessary to validate and extend these findings.

In conclusion, our study adds to the growing body of evidence that the relationship between diet, metabolic health, and genetics is pivotal in influencing brain health. The significant interactions identified between LNAA, the number of MetS components, and ApoE ε4 on WMH volume highlight the importance of considering a multifaceted approach between nutritional, metabolic, and genetic factors to understanding the pathophysiology of WMH and potentially mitigating the risk of brain atrophy. These insights support future investigations toward personalized prevention and management interventions for those with heightened risk for neurodegenerative conditions, particularly focusing on individuals with high MetS components and the ApoE ε4 allele. Our findings suggest that these groups might benefit from targeted dietary recommendations to optimize LNAA levels and potentially mitigate their increased risk of white matter lesions.

The authors would like to thank the MetS study participants and team for their work in collecting, processing, and disseminating these data for analysis.

This study protocol was reviewed and approved by the University of Texas at Austin Institutional Review Board (Approval No. 2011070025). Written informed consent for all study procedures was obtained from all participants to participate in the study. Consent for the publication of anonymized data and relevant information was obtained from all participants prior to their inclusion in this study.

The authors have no conflicts of interest to declare.

This work was supported by the National Institutes of Health grant R01NS075565.

Study conception and design: B.S., H.T., and A.P.H.; analysis and interpretation of data: C.Y. and A.P.H.; drafting of manuscript: C.Y. and A.P.H.; critical revision: C.Y., M.L.C., Y.L., I.A.G., B.S., H.T., and A.P.H.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from A.P.H. upon reasonable request.