Serum and CSF metabolomics analysis shows Mediterranean Ketogenic Diet mitigates risk factors of Alzheimer's disease

Abstract

Alzheimer's disease (AD) is influenced by a variety of modifiable risk factors, including a person's dietary habits. While the ketogenic diet (KD) holds promise in reducing metabolic risks and potentially affecting AD progression, only a few studies have explored KD's metabolic impact, especially on blood and cerebrospinal fluid (CSF). Our study involved participants at risk for AD, either cognitively normal or with mild cognitive impairment. The participants consumed both a modified Mediterranean Ketogenic Diet (MMKD) and the American Heart Association diet (AHAD) for 6 weeks each, separated by a 6-week washout period. We employed nuclear magnetic resonance (NMR)-based metabolomics to profile serum and CSF and metagenomics profiling on fecal samples. While the AHAD induced no notable metabolic changes, MMKD led to significant alterations in both serum and CSF. These changes included improved modifiable risk factors, like increased HDL-C and reduced BMI, reversed serum metabolic disturbances linked to AD such as a microbiome-mediated increase in valine levels, and a reduction in systemic inflammation. Additionally, the MMKD was linked to increased amino acid levels in the CSF, a breakdown of branched-chain amino acids (BCAAs), and decreased valine levels. Importantly, we observed a strong correlation between metabolic changes in the CSF and serum, suggesting a systemic regulation of metabolism. Our findings highlight that MMKD can improve AD-related risk factors, reverse some metabolic disturbances associated with AD, and align metabolic changes across the blood-CSF barrier.

Keywords: Diseases; Metabolomics.

Fig. 1. Crossover design of the diet study.

In the study, 19 participants were enrolled. They were randomized into two groups: one followed the modified Mediterranean ketogenic diet (MMKD) and the other followed the American Heart Association diet (AHAD). Each group adhered to their respective diet for 6 weeks, followed by a 6-week washout period, after which they switched to the other diet. Throughout the study, samples from cerebrospinal fluid (CSF), serum, and microbiome were collected from each participant at various timepoints.

Fig. 2. Serum metabolite associated with the modified Mediterranean ketogenic diet (MMKD).

a Fold change of significantly altered metabolites (FDR adjusted P value < 0.2) after 6 weeks on MMKD (n = 19). The arrows next to the metabolite names indicate the effect direction of these metabolites in dementia or AD based on the UKBB cohort. b Changes in the levels of the three branched-chain amino acids (BCAAs) pre- and post-diet. Each line represents the trend in a single individual. c (i) Microbiome species that show differential abundance by diet and are significantly correlated with serum BCAA (P = 3.9e-13). (ii) Species whose genome encodes BCAA biosynthesis have a significantly higher association to serum BCAA and are significantly more enriched post-MMKD as compared to AHAD. d Pre- and post-diet changes in total HDL and size-specific HDL particle concentrations, cholesterol ester, and free cholesterol content (mmol/l). HDL high-density lipoprotein, PLs phospholipids, CEs cholesteryl esters, VLDL very low-density lipoprotein, FC free cholesterol, C cholesterol, TGs triglycerides, PGs phosphoglycerides, SFAs saturated fatty acids, FA fatty acids, BCAAs branched-chain amino acids.

Fig. 3. CSF metabolites associated with the modified Mediterranean Ketogenic Diet (MMKD).

a Fold change of significantly altered metabolites after 6 weeks on MMKD (n = 19) with 95% confidence interval. Light blue circles represent the metabolites altered in CSF. The corresponding change of each metabolite measured in serum is shown in red. b BCAA metabolism network with MMKD-altered metabolites colored to reflect effect direction. Blue represents a decrease post-diet, red represents an increase post-diet, black represents unchanged, and brown represents unavailable measurements. Dotted lines indicate the presence of intermediate steps between two metabolites, while solid lines are a direct conversion.

Fig. 4. Bipartite graph between CSF and serum metabolites.

Each node represents a metabolite whose size reflects the magnitude of its fold change post-MMKD and whose color represents its biochemical group. Ellipses depict metabolites measured in serum, and diamonds represent metabolites measured in CSF. Each edge represents a significant (FDR < 5%) correlation of the magnitude of changes pre- to post-diet between two metabolites and is weighted to reflect the magnitude of that correlation.