Apolipoprotein E genotype-dependent nutrigenetic effects to prebiotic inulin for modulating systemic metabolism and neuroprotection in mice via gut-brain axis

Abstract

Objective:

The goal of the study was to identify the potential nutrigenetic effects to inulin, a prebiotic fiber, in mice with different human apolipoprotein E (APOE) genetic variants. Specifically, we determined the similar and different responses to inulin for the potential modulation of the systemic metabolism and neuroprotection via gut-brain axis in mice with human APOE ε3 and ε4 alleles.

Method:

We performed experiments with young mice expressing the human APOE3 (E3FAD mice) and APOE4 gene (E4FAD mice). We fed mice with either inulin or control diet for 16 weeks starting from 3 months of age. We determined gut microbiome diversity and composition using16s rRNA sequencing, systemic metabolism using in vivo MRI and metabolomics, as well as blood-brain barrier (BBB) tight junction expression using Western blot.

Results:

In both E3FAD and E4FAD mice, inulin altered the alpha and beta diversity of the gut microbiome, increased beneficial taxa of bacteria (Lactobacillus and Prevotella), and elevated cecal short chain fatty acid and hippocampal scyllo-inositol (an Aβ inhibitor). However, E3FAD mice had altered metabolism related to tryptophan and tyrosine, while E4FAD mice had changes in the tricarboxylic acid cycle, pentose phosphate pathway, and bile acids. Differences were also found in levels of brain metabolites related to oxidative stress reduction, and levels of Claudin-1 and Claudin-5 BBB tight junction expressions between the two groups.

Discussion:

We found that inulin had many similar beneficial effects in the gut and brain for both E3FAD and E4FAD mice, which may be protective for brain functions and reduce risk for neurodegeneration, such as Alzheimer’s disease. E3FAD and E4FAD mice also had distinct responses in several metabolic pathways, suggesting an APOE-dependent nutrigenetic effects in modulating systemic metabolism and neuroprotection. Our findings indicate the importance of precision nutrition as response to diet may be highly dependent on the host’s genotype.

Introduction

Nutrigenetics refers to the potential impacts of diet that may be driven by individual’s genetic background [1]. In contrast to the “one-size-fits-all” assumption, it has been suggested that the heterogeneous response of gene variants to nutrients and dietary components may need to be taken into consideration in the nutritional studies [1].

The goal of the study is to understand the nutrigenetic effects of the Apolipoprotein E gene (APOE) variants to prebiotic inulin. ApoE is a protein involving cholesterol and lipid transports [2]. Humans possess three genetic isoforms of APOE – APOE2, APOE3, and APOE4 – that confer differential risk for developing metabolic syndrome and neurodegenerative disorders [3]. Specifically, APOE4 carriers have higher risk than the other two allele carriers for developing disorders such as diabetes and Alzheimer’s disease [3]. They showed early signs of impairments in mitochondrial function, tricarboxylic acid cycle activity, glucose oxidative metabolism, and cognitive capability [4, 5]. Further, recent studies showed that APOE4 carriers are also associated with alternated and imbalanced gut microbiota (a.k.a., dysbiosis), which in turns exacerbates the systemic inflammation and metabolic dysfunction [6–8].

In a previous study, we showed that inulin, a prebiotic fiber, is effective in restoring systemic metabolism, enhancing gut microbiome balance, reducing neuroinflammation, and increasing levels of short chain fatty acids (SCFAs), tryptophan-derived metabolites, bile acids, glycolytic metabolites, and scyllo-inositol (inhibiting Aβ aggregation) in an APOE4 mouse model (E4FAD mice) [6]. In this study, our goal is to further determine if inulin has nutrigenetic effects for mice with different human APOE variants. In particular, we wanted to identify if mice with human APOE3 gene (E3FAD mice; neutral AD risk) would have different responses to inulin compared to the E4FAD mice.

As APOE4 carriers develop early deficits in brain and systemic metabolism, it would be important to identity interventions that may mitigate the deficits at early stage [4, 5]. Therefore, an additional goal of the study was to determine the effects of inulin as a preventive intervention. We fed young mice (3 months of age) with either inulin or control diet for 16 weeks. We determined the effects of inulin on gut microbiome diversity and composition, brain and systemic metabolism, as well as blood-brain barrier (BBB) tight junction expression in the study.

Methods

Animals and Study Design

We used a C57BL/6 mouse model which accumulates human Aβ₄₂ due to co-expression of 5 familial-AD (5xFAD) mutations in conjunction with human targeted replacement APOE (ε4 in the E4FAD line and ε3 in the E3FAD line). Each mouse was genotyped by Transnetyx Inc. (Cordova, TN, USA) to verify the genotype after weaning. The mice were separated into 4 groups: E3FAD-control, E3FAD-inulin, E4FAD-control, and E4FAD-inulin. We determined the sample size via power analysis with a comparison at a 0.05 level of significance and a 90% chance of detecting a true difference of each measured variable between groups; N= 15 per group (male= 7; female = 8) for gut microbiome experiments, a subset of mice (N= 8/group; male: female= 1:1) for the brain imaging and SCFAs experiments, and the other subset of mice (N= 7/group) for the BBB tight junction measurements. The prebiotic inulin diet contained 8% of fiber from inulin, and the control diet contained 8% of fiber from cellulose (Table 1). We started to feed mice at 3 months of age and fed them for 16 weeks (till they reached 7 months of age). As we previously reported that the mice had not developed severe AD pathology nor cognitive impairment at 7–8 months of age [6], the feeding period allowed us to determine the effects of inulin on systemic metabolism as a preventive intervention. Both the inulin diet and the control diet are modifications of TestDiet’s AIN-93G Semi-Purified Diet 57W5 (cellulose:9GLK; inulin:9GLL). Details of the composition of the control (cellulose) and inulin diets can be found in Table 1. Evidence shows that 8% inulin increases short chain fatty acids and improves enzyme activity in the mouse cecum compared to 4% inulin [9]. In addition, studies indicate that 8% inulin intake (40 g fiber per day) is the maximum amount for western humans to tolerate without side effects [10]. Each mouse was single housed (to avoid feces exchange) with ad libitum access to food and water. Food intake and body weight were recorded biweekly. The study was approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Kentucky (UK).

Table 1.

Diet composition

Diet	Prebiotic Diet	Vehicle Control Diet
Protein %	18.2	18.2
Carbohydrate %	67.8	60.2
Fat %	7.1	7.1
Fiber %	8.0 (Inulin)	8.0 (Cellulose)
Energy (kcal/g)	4.08	3.78

Fecal Sample Collection and Gut Microbiome Analysis

Fecal DNA Amplification

Fecal samples were collected from all the mice (N= 15/group) and frozen at −80°C until further use. A PowerSoil DNA Isolation Kit (MO BIO Laboratories, Inc.) was used for fecal DNA extraction, according to the manufacturer’s protocol [11]. First stage amplifications were performed followed by the second-stage PCR amplification with Access Array Barcode Library for Illumina Sequencers (Fluidigm, South San Francisco, CA; Item# 100–4876) at the University of Illinois at Chicago Sequencing Core. Sequencing was performed at the W.M. Keck Center for Comparative and Functional Genomics at the University of Illinois at Urbana-Champaign. The gene amplicon sequence data generated as part of this study have been submitted to the NCBI BioProject database (PRJNA540508).

Microbial Analysis

Forward and reverse reads were merged using PEAR [12]. Sequences were then trimmed based on quality scores using a modified Mott algorithm with PHRED quality threshold of p = 0.01, and sequences shorter than 300 bases after trimming were discarded. QIIME v1.8 was used to generate OTU tables and taxonomic summaries for all phyla, classes, orders, families, genera, and species present in the dataset [13]. Shannon and Bray-Curtis indices were calculated with default parameters in R using the vegan library [14]. The rarefied species data, taxonomic level 7, were used to calculate both indices. Significant difference among tested groups was determined using the Kruskal-Wallis one-way analysis of variance. The group significance tests were performed on the rarefied species data, taxonomic level 6 (genus), using the group_significance.py script within the QIIME v1.8 package.

In vivo Brain Metabolites Measurements

¹H-MRS was conducted on a 7T ClinScan MR scanner (Siemens, Germany) at the Magnetic Resonance Imaging & Spectroscopy Center of UK. MRS was utilized to measure metabolites in the hippocampus, as we previously reported [6]. N= 8 per group (male: female=1:1) were used for the experiment. The following metabolites were measured: alanine, total choline, glutamate-glutamine complex, myo-inositol, scyllo-inositol, lactate, NAA, phosphocreatine, total creatine, and taurine using the LCModel software to find the absolute concentration of the metabolites [15].

Short Chain Fatty Acid Analysis

After euthanizing the mice, we collected cecal and whole blood samples and sent them to Metabolon Inc. (Durham, NC) for SCFAs analysis. The same group of mice from the ¹H-MRS study (N= 8/group) were used for the experiment. The unit of cecal samples was μg/g and the unit of blood samples was ng/mL. Samples were processed by liquid chromatography/mass spectrometry (LC-MS/MS) for acetate, propionate, and butyrate. Samples were labelled with internal standards and homogenized in an organic solvent. They were then centrifuged followed by an aliquot of the supernatant used to derivatize to form SCFAs hydrazides. This reaction mixture was subsequently diluted, and an aliquot was injected onto a C18 column in Agilent 1290 UHPLC interfaced to an AB Sciex QTrap 5500 LCMS/MS system with electrospray ionization operating in negative mode using electrospray ionization. The raw data were analyzed by AB SCIEX software (Analyst 1.6.2).

Non-targeted Metabolomics Profiling

The whole blood and brain tissue (from cortex and hippocampus) were processed for metabolomics profiling at Metabolon Inc. as well. Metabolon’s standard solvent extraction method was used to prepare the samples, which were then equally split for analysis via LC/MS or gas chromatography/mass spectrometry (GC/MS) using their standard protocol [16]. Non-targeted UPLC-MS/MS and GC-MS analyses were performed at Metabolon, Inc.

Tight junction measurements

Brain Capillary Isolation

Brain capillaries were isolated from another subset of mice (N= 7/group) as described previously [17]. Isolated capillary pellets from each group were resuspended in CelLytic™ lysis buffer (Millipore-Sigma) containing cOmplete™ protease inhibitor cocktail (Roche, Mannheim, Germany). Suspensions were homogenized, centrifuged (30,000 g for 30 min; 95,000 g for 2h; 4 °C) in a fixed-angle rotor (TLA 100.2; Beckman Coulter, Indianapolis, IN, USA), resuspended and frozen at −20°C until further analysis.

Western Blotting

Sample protein concentrations were determined by Bradford assay. Membranes were blocked and incubated with primary antibody overnight (β-actin: 1 mg/ml, ab8226, Abcam, Cambridge, MA, USA; Claudin-1: 0.5 mg/ml, ab56417, Abcam, Cambridge, MA, USA; Claudin-5: 0.5 mg/mL, 35–2500, Thermo Fisher Scientific; Occludin: 0.5 mg/mL, 71–1500, Thermo Fisher Scientific). Membranes were washed and incubated with horseradish peroxidase-conjugated ImmunoPure™ secondary IgG (1:10,000; Pierce, Rockford, IL, USA) for 1 h at room temperature. Protein bands were visualized using Pierce SuperSignal™ West Pico and West Femto chemiluminescent substrates (Thermo Fisher Scientific) using a ChemiDoc™ XRS imager (Bio-Rad Laboratories, Hercules, CA, USA). Optical density was measured with ImageLab software (v.4.6.9; Bio-Rad Laboratories). Four replicate blots were obtained from pooled tissue of the seven mice in each group.

Statistical Analysis

All statistical analyses were completed using GraphPad Prism (GraphPad, San Diego, CA, USA). We performed a 2-way ANOVA to determinate the differences between groups. To determine which comparison the statistical differences were coming from we used post hoc Tukey’s multiple comparisons test. For Metabolon, log transformations were conducted followed by ANOVA for identification of biochemicals that were significantly different between groups. Between group differences were assessed with p-values and q-values (with the Benjamini Hochberg FDR) with a p-value or q-value (when applicable) less than 0.05 showing statistical significance.

Results

Inulin maintains bodyweight but alters the gut microbiome of the E3FAD and E4FAD mice.

We first determined if the inulin diet changed the food intake and bodyweight of the mice compared to the controls. We found that there was no difference in either food intake (Fig. 1A) or body weight (Fig. 1B) between the inulin- and the control-fed mice after 16 weeks of feeding. We further looked at the gut microbiome and found that there was a significant decrease in α-diversity in E3FAD-inulin mice (Gausian link function results, p < 0.001) (Fig. 1C) and in E4FAD-inulin mice compared to their controls (Gausian link function results, p = 0.019) (Fig. 1D). This shows species richness and evenness were decreased within each group following inulin treatment. Similarly, E3FAD- and E4FAD-inulin mice had significant differences in beta-diversity compared to the controls (E3FAD mice: ANOSIM R statistic = 0.877, p-value = 0.001) (Fig. 1E) (E4FAD mice: ANOSIM R statistic = 0.454, p-value = 0.001) (Fig. 1F). This shows the ratio of species was different between the two groups following inulin treatment.

Figure 1. Food intake and body weight and Gut microbiota diversity.

Inulin feeding did not change the (A) food intake and (B) body weight of the mice. Inulin induced significant differences of the α-diversity in the (C) E3FAD and (D) E4FAD mice, compared to their control littermates. Inulin also induces significant differences in β -diversity in the (E) E3FAD and (F) E4FAD mice, compared to their controls. N= 15/group. Data are presented as Mean ± SEM.

Significant changes were also observed in numerous bacterial taxa due to the beneficial effects of the prebiotic inulin. Table 2 shows the fold change between the group comparison; red indicates significant increases while green indicates significant decreases (p < 0.05). Both E3FAD-inulin and E4FAD-inulin mice had decreased Turicbacter, but increased Lactobacillus and Prevotella, compared to their respective control groups. Moreover, inulin increased Bifidobacterium in E3FAD mice compared to its control. These results indicate that the prebiotic inulin alters the gut microbiota in E3FAD and E4FAD mice in a different manner.

Table 2.

Gut microbiota taxonomy changed by inulin. Taxonomic differences at the genus level between E3FAD and E4FAD prebiotic and control fed mice. q-values were calculated for each taxon along with each group with a significant difference (q < 0.05). Fold change was calculated for each taxon of each group. Red and green shaded cells indicate p ≤ 0.05 (red specifies that the mean values are significantly higher for that comparison; green values significantly lower).

	E3FAD		E4FAD
Taxonomy	Fold Change	Q value	Fold Change	Q value
	Inulin Ctrl		Inulin Ctrl
Turicibacter	−2.11	2.96E-06	−0.47	0.03
Lactobacillus	+1.21	9.69E-04	+1.48	0.02
Prevotella	+1.48	1.11E-03	+1.78	2.17E-03
Bifidobacterium	+2.10	2.20E-06	0.11	0.95

Inulin increased SCFAs in E3FAD and E4FAD mice.

Both E3FAD-inulin and E4FAD-inulin mice showed a significant increase in SCFAs levels in the cecum, including acetate (Fig. 2A), propionate (Fig. 2B), and butyrate (Fig. 2C). In the blood samples, we found that acetate (Fig. 2D) was also significantly higher in the E4FAD-inulin group. Blood propionate (Fig. 2E) and butyrate (Fig. 2F) were trending higher in the Inulin-fed groups though did not reach statistical significance.

Figure 2. Short chain fatty acids (SCFAs).

Inulin significantly increased (A) cecal acetate, (B) cecal propionate, and (C) cecal butyrate of the E3FAD and E4FAD mice compared to their controls. For SCFAs in the blood, significantly elevation was found in (D) acetate from the E4FAD-inulin group; (E) blood propionate or (F) blood butyrate did not show changes in either of the E3FAD or E4FAD mice. N= 8/group. Data are Mean ± SEM; ns = not significant (p ≥ 0.05), *p < 0.05, **p < 0.01.

Inulin altered systemic metabolism differently in the E3FAD and E4FAD mice.

From the blood samples, we found that inulin induced different patterns of metabolic changes among the E3FAD and E4FAD mice. Inulin increased tryptophan-related metabolites while decreasing tyrosine-related metabolism in the E3FAD-inulin compared to the E3FAD-control mice (Table 3). The tryptophan metabolites included indoleacrylate, indolepropionate (IPA), serotonin, indoleacetylglycine, and N-acetyltryptophan. Tyrosine metabolites included phenol sulfate, phenol glucuronide and p-cresol glucuronide. In the E4FAD-inulin mice, only changes in indoleacrylate, IPA and p-cresol glucuronide were seen compared to the E4FAD-control mice.

Table 3.

Tryptophan and tyrosine metabolism, Pentose metabolism, TCA cycle, and bile acid changes induced by inulin. A table showcasing the metabolite changes in tryptophan and tyrosine metabolism, pentose metabolism, TCA cycle, and bile acid metabolism due to the prebiotic inulin and differences and differences between E3FAD and E4FAD mice in the blood. Welch’s two-sample t-test was performed. Heat map of statistically significant biochemicals profiled when comparing groups are labeled as follows: Red and green shaded cells indicate p ≤ 0.05 (red specifies that the mean values are significantly higher for that comparison; green values significantly lower).

		Fold Change
		Inulin Ctrl
Pathway	Biochemical Name	E3FAD	E4FAD
Tryptophan Metabolism	indoleacrylate	2.91	5.68
	indolepropionate	6.42	9.68
	serotonin	1.37	1.29
	indoleacetylglycine	1.55	0.65
	N-acetyltryptophan	1.3	0.81
Tyrosine Metabolism	phenol sulfate	0.23	0.87
	phenol glucuronide	0.21	2.11
	p-cresol glucuronide	0.08	0.14
Pentose Metabolism	ribose	1.67	1.41
	ribitol	0.99	2.01
	ribonate	1.46	2.08
	ribulose/xylulose	1	1.37
	arabitol/xylitol	0.87	2.67
	arabonate/xylonate	1.22	1.36
	sedoheptulose	0.82	1.68
TCA Cycle	isocitric lactone	1.28	1.73
	alpha-ketoglutarate	1.73	2.53
	succinate	1.03	2
	fumarate	1.23	1.6
	malate	1.24	1.44
Primary Bile Acid Metabolism	cholate	1.06	10.92
	beta-muricholate	1.82	16.81
Secondary Bile Acid Metabolism	deoxycholate	0.8	8.78
	ursodeoxycholate	1.37	82.41
	7-ketodeoxycholate	0.86	82.44
	ursocholate	1.76	22.69

In contrast, E4FAD-inulin mice had significant changes in metabolism involved in the pentose phosphate pathway, TCA cycle, and primary and secondary bile acids. Table 3 shows that E4FAD-inulin mice had significantly increased levels of ribose, ribitol, ribonate, ribulose, arabitol, arabonate, and sedoheptulose in pentose metabolism compared to E4FAD-control. In TCA Cycle metabolites, riboaconitate, isocitric lactone, succinate, fumarate, and malate increased in the E4FAD-inulin mice. Inulin also increased bile acid and their associated metabolites in E4FAD-inulin mice compared to E4FAD-control mice, including cholate and beta-muricholate (primary bile acids), and deoxycholate, ursodeoxycholate, 7-ketodeoxycholate and ursocholate (secondary bile acids). On the other hand, in the E3FAD-inulin mice, we found increases only in ribose (pentose metabolism) and alpha-ketoglutarate (TCA cycle) compared to the E3FAD-control mice.

Inulin changes brain metabolites in the E3FAD and E4FAD mice.

We further looked at the effect of the inulin on brain metabolites. We found that inulin reduced the level of 5-hydroxyindoleacetate in E4FAD-inulin mice, but not in the E3FAD mice (Fig. 3A). Inulin increased acetylcarnitine (C2) concentration in the inulin-fed groups, though only reached statistical significance in the E3FAD-inulin group (Fig. 3B). In contrast, myo-inositol concentration decreased in the inulin-fed mice, but only reached statistical significance in the E4FAD-inulin mice (Fig. 3C). Scyllo-inositol level increased in both E3FAD-inulin and E4FAD-inulin mice compared to their controls (Fig. 3D). The increase in scyllo-inositol from the brain tissue was confirmed by in vivo brain imaging using ¹H-MRS. Figure 3E shows the voxel that was used for obtaining the spectra and a representative spectrum is shown in Figure 3F. An arrow indicates the position of scyllo-inositol (3.34 ppm). We observed the level of scyllo-inositol to be significantly higher in the hippocampus in both E3FAD- and E4FAD-inulin mice relative to their controls (Fig. 3G).

Figure 3. Metabolite changes in brain tissue in response to inulin and Scyllo-inositol levels in the hippocampus, fecal culture, and plasma.

(A) 5-hydroxyindoleacetate was reduced in the E4FAD-inulin mice compared to the E4FAD-control. (B) Acetylcarnitine (C2) was increased in the E3FAD-inulin mice compared to E3FAD-control. (C) Myo-inositol was significantly lower in the E4FAD-inulin mice compared to the controls. (D) Scyllo-inositol was elevated in both E3FAD-inulin and E4FAD-inulin mice compared to their controls. (E) ¹H-MRS voxel on hippocampus. (F) A representative ¹H-MRS spectrum, showing N-acetyl-aspartate (NAA), glutamate (Glu) and glutamine (Gln), creatine (Cr), glycerophosphocholine (GPC) and phosphocholine (PCh), scyllo-inositol (Scyllo), taurine (Tau), and myo-inositol (mI) in parts per million (ppm). (G) Scyllo-inositol was dramatically increased in the hippocampus of E3FAD-inulin and E4FAD-inulin mice compared to E3FAD-control and E4FAD-control mice, respectively. N= 8/group. Data are Mean ± SEM, ns = not significant (p ≥ 0.05), *p < 0.05, **p <0.01, ****p < 0.0001.

Inulin has differential effects on blood-brain barrier tight junction protein levels in the E3FAD and E4FAD mice.

We have previously reported that dietary inulin reduces neuroinflammation in the E4FAD mouse model [6]. To examine the effect of inulin on the brain microvasculature, we compared tight-junction protein expression levels in brain capillaries isolated from E3FAD-control and E4FAD-control mice with E3FAD-inulin and E4FAD-inulin mice (Fig. 4A). Inulin did not significantly alter occludin protein levels in E3FAD mice or E4FAD mice but occludin protein levels were significantly lower in E4FAD-inulin mice compared to E3FAD-control mice (p = 0.0045) (Fig. 4B). Further, Claudin-1 protein expression levels were significantly higher in E4FAD-control mice compared to E3FAD-control mice (p = 0.0012) (Fig. 4C). Inulin did not alter Claudin-1 protein expression levels in E3FAD mice whereas Claudin-1 levels were significantly lower in E4FAD-inuiln mice compared to E4FAD-control mice (p = 0.0036). E3FAD-inulin mice also had a significantly higher protein expression for Claudin-5 compared to E3FAD-control mice (p = 0.0122) (Fig. 4D). Likewise, Claudin-5 protein expression levels were significantly higher in E3FAD-control mice compared to E4FAD-inulin mice (p = 0.0026). These values indicate that inulin affects tight-junction proteins differently in E3FAD and E4FAD mice.

Figure 4. Western blot analysis of tight-junction protein expression levels.

(A) Representative Western blot for occludin (~55–65 kDa), claudin-1 (~23 kDa), and claudin-5 (~23 kDa). β-actin (~42 kDa) served as protein loading control. Normalized protein expression levels for (B) Occludin, (C) Claudin-1, and (D) Claudin-5 are given as fold change in percent (%). Values are represented as Mean ± SEM for replicate blots (N = 4; pooled tissue from E3FAD-Control (N = 7 mice), E3FAD-Inulin (N = 7 mice), E4FAD-Control (N = 7 mice), E4FAD-Inulin (N = 7 mice); *p < 0.0332; ** p < 0.0021; ***p < 0.0002.

Discussion

In this study, we demonstrated that inulin had systemic benefits for both E3FAD and E4FAD mice. We further demonstrated that the E3FAD and E4FAD mice had very different responses to inulin, indicating a nutrigenetic effect depending on APOE genotype. Figure 5 summarizes the similarities and differences of E3FAD and E4FAD mice responding to inulin.

Figure 5.

Summary of the similarities and differences of responses to inulin between E3FAD and E4FAD mice.

The similar effects that inulin had on E3FAD and E4FAD mice included the alpha- and beta- diversity of the gut microbiome, increased the abundance of Lactobacillus and Prevotella taxa, and reduced abundance of Tuicibacter taxa (Fig. 5; middle column). In both groups, inulin also increased the cecal SCFAs and scyllo-inositol level in the hippocampus. Lactobacillus and Prevotella are known to produce SCFAs including acetate, butyrate, and propionate [18–20]. This is consistent with our findings that SCFAs were increased in the inulin-fed mice, specifically acetate in the E4FAD mice. Though being present mostly in the periphery system [21, 22], SCFAs have been shown to have a dramatic impact on brain vascular and metabolic function [23]. Specifically, butyrate was shown to be able to restore BBB tight junctions and consequently restored BBB function [24], which is consistent with our results of changes in Claudin-5 expression in the BBB of E3FAD mice. Further, scyllo-inositol was found to be increased in hippocampus in vivo and from the brain extracts. Scyllo-inositol has been shown to inhibit Aβ₄₂ aggregation in both animal and clinical trials [25, 26], suggesting that inulin might be able to reduce AD pathology when the mice get older.

Inulin induced different responses between the E3FAD (Fig. 5, left column) and E4FAD mice (Fig. 5, right column). Notably, E3FAD mice had altered metabolic pathways in the periphery related to tryptophan and tyrosine, while E4FAD mice had changes related to TCA, PPP and bile acids. Inulin increased Bifidobacterium in the E3FAD mice but showed no change in the E4FAD mice. There were also different changes in brain metabolites, and levels of Claudin-1 and Claudin-5 expressions between the two groups.

Tryptophan and tyrosine metabolism enhance the nervous system. Tryptophan is an essential amino acid and is the precursor of serotonin. Indole-3-propionic acid, a tryptophan-derived metabolite, can inhibit Aβ fibril formation in neurons and neuroblastoma cells [27]. Metabolites that play a role in tyrosine metabolism including phenol sulfate, phenol glucuronide, and p-cresol glucuronide were found to be lower in the inulin-fed mice, suggesting that inulin decreases inflammation in the E3FAD mice [28]. The higher abundance of Bifidobacterium found in the E3FAD mice compared to E4FAD mice suggests a reduced level of anxiety [29], which is consistent with our previous findings that E3FAD mice had lower anxiety compared to the E4FAD mice [30]. In contrast, E4FAD mice had enhanced TCA, PPP and bile acid metabolism compared to E3FAD mice. Both TCA and PPP pathways involve energy production through mitochondria [31]. Research shows that the PPP regulates the chronic neuroinflammation in the brain and may therapeutically improve neurodegeneration in the brain such as Parkinson’s disease [31]. Further, bile acids significantly impact pathways involved in host cholesterol, lipid and glucose metabolism, and inflammation, and have the potential to alter the host immunity [32] and circadian rhythms [33]. Collective evidence also shows that primary and secondary bile acid metabolism regulates the innate immune system by activation of bile acid activated receptors [34] and adaptive immune system by regulation of Th17 cells and regulatory T cells [35]. These findings are consistent with neuroimaging markers, showing that E4FAD mice had decreases in myo-inositol [36, 37], an inflammatory marker, in the hippocampus.

In brain metabolites, we found increased acetylcarnitine in E3FAD-inulin mice. Acetylcarnitine has been well studied to prevent oxidative stress and aging [38], suggesting inulin may reduce oxidative stress for the E3FAD mice. In contrast, we found reduced 5-hydroxyindoleacetate in the E4FAD-inulin mice. 5-hydroxyindoleacetate has been shown to be related to aggressive behaviors and toxification by affecting brain glutathione-S-transferase [39, 40]. Therefore, inulin might benefit the E3FAD mice by oxidative stress reduction, while it might benefit the E4FAD mice by increasing detoxification in the brain and lowering the risk for aggressive behavior development.

Previous studies found that the ApoE4 genetic background is associated with BBB dysfunction [41]. In the present study, we observed that inulin differentially affects tight-junction protein expression levels in isolated brain capillaries from E3FAD and E4FAD mice. Specifically, inulin decreased Claudin-1 expression in E4FAD mice, while increasing Claudin-5 expression in E3FAD mice. A previous study in a chronic stroke mouse model showed that increased Claudin-1 protein expression is associated with leaky brain capillaries and an endothelial proinflammatory phenotype [42]. Claudin-5 is the most enriched tight junction protein and its dysfunction has been implicated in many neurological disorders including AD, multiple sclerosis, depression, and schizophrenia [43]. It is possible to abrogate disease symptoms in many of these disorders by regulating levels of Claudin-5.

Though we studied the mice at pre-symptomatic stage, our study implied that inulin may reduce risk for developing AD for both APOE3 and APOE4 carriers. Emerging evidence further shows that AD is associated with brain metabolic impairment [44], gut microbiota dysbiosis [45], and bile acid profile alterations [46]. Further, tryptophan metabolism is also seen to be altered in patients with AD [47], which impedes the capacity to inhibit Aβ fibril formation in neurons and neuroblastoma cells [27]. The findings from the current studies suggest that AD risk might be able to be mitigated with early interventions of inulin. Future studies will be needed to verify the speculation with older animals.

In summary, we demonstrated that early interventions of inulin would be beneficial for enhancing brain and systemic metabolism through gut-brain axis, and the responses to the inulin diet are APOE genotype-dependent. Our findings indicate the importance of nutrigenetics and precision nutrition as the responses to a diet may be highly dependent on the genotype of the host.