Maternal high sugar and fat diet benefits offspring brain function via targeting on the gut–brain axis

Changes to the gut microbiota during the growth stage

To study the impact of the maternal diet during pregnancy on the growth, development, health, and disease of the offspring, we compared the microbiota in fimo at different growth phases, or from mothers fed on different diets during pregnancy, at least eight samples for each subgroup. The alpha diversity (Figure 1 and Supplementary Figures 1A–1F, 2A, 2C, 3) and the beta diversity (Supplementary Figures 2B, 2D, 3A–3G) could be distinguished between each of the standard diet, high sugar and high fat diet (HSHF), and dietary supplement groups. We first analyzed the constitutive changes between the standard diet groups: the number of operational taxonomic units (OTU) increased from low (35) to high (253) as the offspring grew up (Figure 1A and Supplementary Figures 1, 4A–4G), while the quantity variance between individuals decreased (Supplementary Figure 4A–4G). This indicated that the composition of the gut microbiota became more stable from birth to adulthood, the abundance of bacteria increased, and individual differences gradually decreased. Principal component analysis (PCA) showed that the 7- (N1) and 14- (N2) day samples could be successfully distinguished from the 21- (N3), 28- (N4), and 56- (N5) day samples; after 21 days of age, the differences became smaller (Supplementary Figure 3). The HSHF groups from the different growth phases (at 7 (M1), 14 (M2), 21 (M3), 28 (M4), and 56 (M5) days of age) showed the same variation tendency (Supplementary Figures 2A–2D, 4A–4G).

Figure 1. Influences of maternal high sugar and fat diet on the gut microbiota of offspring during the growth stage. (A) The OTUs number of high sugar and fat diet (HSHF) group were low than in the standard diet; the 16S rRNA analysis showed significant differences for the abundance-based covered estimator (ACE, B), Chao 1 (C) and Shannon (D) indexes (p < 0.05); (E) the heatmap of genus bacterial of the whole gut microbiota compositions of the offspring birth from the standard diet and HSHF-diet fed mother at different growth phase (7, 14, 21, 28, 56 days). Also see in Supplementary Figures 1–4. The symbol of N1A is the 7-day control male samples, N1B is the 7-day control female samples, and N2A for 14-day, N3A for 21-day, N4A for 28-day, N5A for 56-day male samples, N2B for 14-day, N3B for 21-day, N4B for 28-day, N5B for 56-day female samples; M1A is the 7-day HSHF male samples, M1B is the 7-day HSHF female samples, and M2A for 14-day, M3A for 21-day, M4A for 28-day, M5A for 56-day male samples, M2B for 14-day, M3B for 21-day, M4B for 28-day, M5B for 56-day female samples. Data are presented as the means ± SD of 8 independent experiments. ^*p <0.05 and ^**p <0.01 vs. the model group by one-way ANOVA, followed by the one-way analysis of variance (ANOVA) at p < 0.05.

The heatmap of bacterial genera in Figure 1E shows the whole gut microbiota compositions of the offspring from mothers fed with the standard diet (Supplementary Figure 4A) and the HSHF diet (Supplementary Figure 4B) were similar. The variation tendencies were similar in these two groups, with the relative abundance of the following genera highest at 7 days after birth, before gradually declining as the offspring grew up: Pasteurella; Delftia; Stenotrophomonas; Streptomyces; Lactobacillus; Corynebacterium; Streptococcus; Pseu-domonas; Aeromonas; Raoultella; Rubellimicrobium; Staphylococcus; Rothia; Acinetobacter; Buchnera; Rickettsia; and Citrobacter. Subsequently bacteria from the following genera grew to become the dominant populations: Atopostipes; Sporosarcina; Ruminococc-aceae UCG-013; Alistipes; Rikenella; Erysipelo-trichaceae; Porphyromonadaceae; Enterorhabdus; Parvibacter; Ruminiclostridium 6; Clostridiales vadin BB60 group; Faecalitalea; Odoribacter; Ruminococcus 1; Mollicutes RF9; [Eubacterium] ventriosum group; Tyzzerella 3; Candidatus Saccharimonas; Ruminoc-occaceae UCG-014; Butyricicoccus; Roseburia; [Eubacterium] xylanophilum group; [Eubacterium] nodatum group; Lachnospiraceae; Alloprevotella; Marvinbryantia; Blautia; Lachnospiraceae UCG-001; Lachnospiraceae UCG-006; Oscillibacter; Lachno-spiraceae FCS020 group; Family XIII UCG-001; Lachnospiraceae NK4A136 group; Coprococcus 1; Ruminiclostridium 9; Ruminococcaceae UCG-009; Peptococcus; Ruminococcaceae UCG-003; Akkerma-nsia; Mucispirillum; [Eubacterium] coprostanoligenes group; Paraprevotella; Prevotellaceae NK3B31 group; Gastranaerophilales; Butyricimonas; Prevotellaceae UCG-001; Bacteroidales S24-7 group; Lachnospi-raceae NK4B4 group; Flavobacteriaceae; Rikenell-aceae RC9 gut group; Ruminococcaceae UCG-010; Ruminococcaceae UCG-005; Peptococcaceae; Rumino-coccaceae NK4A214 group; Ruminiclostridium 5; Ruminiclostridium; Desulfovibrio; and Ruminococc-aceae.

Further details of the different growth phases given in Supplementary Figure 4A–4G show that the relative abundances and compositions of the different bacterial genera were changed, along with the variances of stages T1 and T2 (7 and 14 days of age) compared with the other stages (21, 28, and 56 days of age). The variances in the offspring born from the HSHF diet were enlarged, indicating that feeding the mother with the HSHF diet during pregnancy influenced the hereditary stability of gut microbiota. As such, individual outliers are more common in the HSHF-diet groups than in the standard-diet groups, especially at the T1 (Supplementary Figure 4C), T2 (Supplementary Figure 4D), and T3 (Supplementary Figure 4E) stages. The heatmap appears a bit chaotic and random at the T4 stage (Supplementary Figure 4F), indicating that the gut microbiota changed dramatically and then gradually stabilized at 1 month of age. This randomness might be induced by weaning or by other factors, but this requires further investigation. After this, the HSHF and standard diet groups were significantly different at the T5 stage, with the following genera increased in the male, standard-diet mice (Supplementary Figure 4G): [Eubacterium] ventriosum group; Tyzzerella 3; Veillonella; Family XIII UCG 001; Ruminococcaceae UCG 005; Rikenellaceae RC9 gut group; Ruminococcaceae UCG 010; Lachnospiraceae FCS020 group; Ruminococcus 1; [Eubacterium] coprostanoligenes group; Parvibacter; Mollicutes RF9; Desulfovibrio; Lachnospiraceae NK4A136 group; Odoribacter; Flavobacteriaceae; Lachnospiraceae NK4B4 group; Candidatus Saccharimonas; Roseburia; [Eubacterium] brachy group; Ruminococcaceae UCG 014; Mucispirillum; [Eubacterium] nodatum group; and Faecalitale. In contrast, the following genera were increased in the female, standard-diet mice: Proteus; Faecalibaculum; Bifidobacterium; Christensenellaceae R-7 group; Lactobacillus; Enterorhabdus; Bacteroidales S24-7 group; Escherichia/Shigella; Alloprevotella; and Clostridium sensu stricto 1. However, the intestinal microbial diversities of the HSHF-diet groups were reduced (Figure 1; p < 0.05).

Changes in autophagy and neurodevelopment during the growth stage

Before 56 days of age, the body weights of the HSHF-diet groups were higher than those of the standard-diet groups (Figure 2Aa). While there were no obvious differences in body weight in adult males (Figure 2Ab), there were significant differences in the body weights of adult females (Figure 2Ac). This indicates that body weight in females is more related to the maternal diet. As they grew older, only a few inflammatory factors, including interleukin 1β (IL-1β), epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), MIP-1α, and lipopolysaccharide (LPS) were changed in the serum (Figure 2B; p < 0.05). However, we found that a high serum alpha-fetoprotein (AFP) level was most common in overweight individuals from the HSHF-diet groups (Figure 2B). In addition, we found that some of these individuals had spontaneous tumors, including liver cancer, colorectal cancer, subcutaneous sarcoma, and splenomegaly (data not shown). This suggests that the maternal HSHF diet substantially increases the risk of carcinogenesis and increases body weight in females.

Figure 2. Influences of maternal high sugar and fat diet on autophagy and neurodevelopment of offspring during the growth stage. (A) Changes on the body weight; (B) Changes on inflammatory factors in serum; (C) the marked proteins expression of KGA/GAD, MAP2, GFAP, Vimentin, IBA1, GAD65, DYNC1H1, Myclin, Tubulin, CHAT, OLIG2 and WFS1 in brain tissues, (D) and the expression LC3A I/LC3A II/ LC3B I/LC3B II protein in brain tissues. The symbol of NA is the 16-month control male samples, NB is the 16-month control female samples; MA is the 16-month HSHF male samples, MB is the 16-month HSHF female samples. Data are presented as the means ± SD of more than 8 independent experiments, and more than 3 independent experiments in Western bolting. ^*p <0.05 and ^**p <0.01 vs. the model group by one-way ANOVA, followed by the one-way analysis of variance (ANOVA) at p < 0.05.

We monitored the development of neurons using the following marker proteins: KGA/GAD; MAP2; GFAP; vimentin; IBA1; GAD65; DYNC1H1; myclin; tubulin; CHAT; OLIG2; and WFS1. As shown in Figure 2C (Figures 2C1–C4), the HSHF diet markedly influences some types of neurons, alongside vimentin and GFAP in astrocytes, and IBA1 in activated microglia (Supplementary Figure 5A). The same trend was observed for CHAT in cholinergic neuron, DYNC1H1 in axons, and WFS1 in the hippocampal CA1 subfield (Figure 2C; p < 0.05). In addition, there were some opposing trends, such as for OLIG2 in oligodendrocytes, and tubulin in neurons (Figure 2C; p < 0.05). We also monitored the levels of autophagy, as shown in Figure 2D. We found that the ratio of the relative expression levels of LC3A I/LC3A II/ LC3B I/LC3B II protein in the HSHF-diet groups were higher than those in the standard diet groups before 28 days of age, after which they decreased (Figure 2D2). Taken together, these results show that the maternal diet influenced brain development in the offspring.

H&E staining showed the pathological structure of the small intestine (Supplementary Figure 5B), with the pathomorphological changes being different at different growth phases (7, 14, 21, 28, and 56 days of age). This indicates that the gut microbiota or diet can affect the maturation of the intestinal tissue, but the mechanisms underlying this effect need much more in-depth research.

Long-term potentiation in hippocampal slices

After ten months, subsets of mice were randomly selected to perform standard field potential recordings. Repetitive stimulation (0.33 Hz) of Schaffer collaterals evoked fEPSPs in the hippocampal CA1 region (Figure 3Aa; p < 0.05). The input–output relationship curves (Figure 3Ab; p > 0.05) and linear slopes (Figure 3Ac; p > 0.05) revealed that compared to control mice (standard-diet mice, control; HSHF-diet mice, model), the hippocampal CA1 neurons of model mice showed no differences. Synaptic short-term plasticity was measured using a PPF protocol in hippocampal slices. Statistical analyses of data from all slices demonstrated an inhibition of the ratio of P2/P1 at all tested P1 and P2 intervals in model mice compared to control mice (Figure 3Ad; p > 0.05), suggesting HSHF-diet no impaired hippocampal synaptic short-term plasticity. Finally, hippocampal CA1 synaptic LTP were compared between model and control mice. Plotting of the recording time to normalized fEPSP slopes (baseline set as 1) from pooled data showed impaired LTP induction (after theta-burst stimulation at 0–10 min) and maintenance (after theta-burst stimulation at 50–60 min) in all treated groups (Figure 3Ae- p > 0.05, unpaired t test). The mean LTP induction results are shown in Figures 3Ad and f (p < 0.05, unpaired t test). This suggests that consuming a HFHS diet before and during pregnancy could improve LTP in the offspring and suggests that the gut microbiota may play an important role.

Figure 3. Influences of maternal high sugar and fat diet on LTP, gut microbiota and metabolome of older mice. (A) LTP in hippocampal slices, 16 months later, subsets of mice were randomly selected to perform standard field potential recordings. (B) Changes on the gut microbiota; (C) Changes on the metabolome in fimo. The pathological structure of brain and small intestine (Supplementary Figure 5). Data are presented as the means ± SD of more than 6 independent experiments. Significant differences between two groups of LTP were evaluated by two-tailed unpaired Student’s t-tests or two-tailed Welch’s t-test. Significant differences between treatments were analyzed by one-way analysis of variance (ANOVA) at p < 0.05.

Changes in the microbiota and metabolomes of older mice

Analysis with 16S rRNA on some of the mice who were fed until 16 months of age showed that there were no obvious differences in the gut microbiota in the HSHF-diet and standard-diet groups. Only a little of bacteria were changed, as the abundances of Lachnospiraceae, Clostridia, Clostridiales, and Rikenellaceae were different, with the following genera more abundant in the HSHF diet groups than in the standard diet groups (Figure 3B; p < 0.05): Lachnospiraceae NK4A136 group; Lachnospiraceae UCG-006; Lachnospiraceae UCG-008; and Clostridiales vadinBB60 group.

Metabolic analysis of colonic contents from the HSHF-diet mice showed 860 different metabolites via LC/MS (Figure 3C). Pathway analysis of the different metabolites using MetaboAnalyst 4.0 (see metabolome in Supplementary Figure 6A) showed significant differences in the following pathways: Pyrimidine metabolism (p = 3.96E-05); Cysteine and methionine metabolism (p = 4.14E-05); Sphingolipid metabolism (p = 4.14E-05); Aminoacyl-tRNA biosynthesis (p = 5.01E-05); Arginine biosynthesis (p = 6.32E-05); Purine metabolism (p = 6.32E-05); Glycine, serine and threonine metabolism (p = 6.72E-05); and Glutathione metabolism (p = 9.17E-05).

Cholinergic and GABAergic neuron counts were increased by upregulation of LHX8 in the HSHF-diet offspring as they grew older

The mRNA sequence of whole brain tissues showed that 6 mRNAs were downregulated (TCONS_00085226, ENSMUSG00000096953, ENSMUSG00000099908, TCONS_00104776, and TCONS_00139120, Kpna2), and 37 mRNAs were upregulated, including Lhx8, predicted pseudogene 10184 (Gm10184), Cd4, adenosine A2a receptor (Adora2a), synapse differentiation-inducing 1-like (Syndig1l), and synaptotagmin VI (Syt6) (Figure 4A and Table 1; p < 0.05).

Figure 4. The cholinergic and GABAergic neurons were increased by up-regulated the LHX8 in the HSHF diet offspring when they getting older. (A) The mRNA sequence of whole brain tissues; (B, C) the cholinergic and GABAergic cells marked protein of ACEh, Amp, CHRNA1, CHRNB1 and GAD65; the synaptic functional markers of Dynamin 1 and PSD95, and the AD biomarkers of Tau, p-Tau, APOE, CD33 and TREM2 (D). The KEGG analysis results see in the Supplementary Figure 6. Data are presented as the means ± SD of more than 3 independent experiments in Western bolting. ^*p <0.05 and ^**p <0.01 vs. the model group by one-way ANOVA, followed by one-way analysis of variance (ANOVA) at p < 0.05.

L3/Lhx8 is involved in the determination of cholinergic or GABAergic cell fate [27]. The mRNA sequence results showed that expression of Lhx8 (Figure 4A and Table 1; p < 0.05) and the LHX8 protein (Figure 4B, p < 0.05) were upregulated. Cholinergic and GABAergic cells regulated by this protein showed upregulated levels of the marker proteins ACEh, Amp, CHRNA1, CHRNB1, and GAD65 (Figure 4B, 4C, p < 0.05).

Expression of the synaptic functional markers dynamin 1 and PSD95 differed (Figure 4B; p < 0.05), while the AD biomarkers Tau, p-Tau, APOE, CD33, and TREM2 showed no obvious differences in expression (Figure 4D; p > 0.05), except for Aβ₄₂ (Figure 4B, p = 0.0001). Furthermore, levels of IBA1, GFAP, and NF-κB did not differ (Figure 4D; p > 0.05), indicating that the maternal HSHF-diet had no obvious harmful effects on the offspring in terms of AD.

The mRNAs of Rgs9, Gpr88, Drd2, Sh3rf2, Tac1, Serpina 9, Rxrg, Drd1, Rasgrp2, Six3, Gpr6, Pdyn, Gng7, and Pde1b were all upregulated (Figure 4A and Table 1; p < 0.05); all of these have been reported to be more or less related to nervous system diseases. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis showed that significantly altered pathways, including: Dopaminergic synapse; cAMP signaling pathway; Gap junction; Rap1 signaling pathway; Retrograde endocannabinoid signaling; Glutamatergic synapse; Neuroactive ligand-receptor interaction; and Calcium signaling pathway (Supplementary Figure 6B, 6C).

Special maternal diet changes the gut microbiome trace

The use of specific dietary supplements (Hericium erinaceus and Ganoderma lucidum) during pregnancy changed the inheritance of microorganisms (Supplementary Figures 2A, 2B). In addition, exposure to different microorganisms in early life from the outdoor soil (T-) and from SAMP8 mice (S-) also changed the inherent abundances of different species (Supplementary Figures 2C, 2D).

The heatmap of bacterial genera in Supplementary Figure 7 shows that the whole gut microbiota compositions of the offspring born from mothers fed on a specific dietary supplement during pregnancy were similar, and their variation tendencies were also the same. However, the relative abundances of some bacterial genera were very different: those of Akkermansia, Blautia, Butyricicoccus, Lactobacillus, Lachnospiraceae, and Roseburia were changed at 28 days and/or 56 days in the Hericium erinaceus dietary supplement group (Figure 5A); and Helicobacter, Odoribacter, Blautia, Butyricicoccus, Lactobacillus, and Roseburia were changed in the Ganoderma lucidum dietary supplement group (Figure 5B). This indicates that specific dietary supplement with Hericium erinaceus or Ganoderma lucidum changes the gut microbiota composition induced by the HSHF diet.

Figure 5. Special maternal diet changes the gut microbiome trace. (A) The dietary supplement of Hericium erinaceus group (A); the dietary supplement of Ganoderma lucidum (B); the heatmap of genus bacterial influenced by the postpartum environmental exposure (C). The symbol of N1 is the 28-day control samples, N2 is the 56-day control samples; M1 is the 28-day HSHF samples, M2 is the 56-day HSHF samples; HE1 is for Hericium erinaceus treated 28-day samples, HE2 is for Hericium erinaceus treated 56-day samples, LZ1 is for Ganoderma lucidum treated 28-day samples, LZ2 is for Ganoderma lucidum treated 56-day samples. Data are presented as the means ± SD of more than 8 independent experiments. ^*p <0.05 and ^**p <0.01 vs. the model group by one-way ANOVA, followed by the one-way analysis of variance (ANOVA) at p < 0.05. Also see in the Supplementary Figures 7, 8

The heatmap of bacterial genera in Supplementary Figure 8 shows that the gut microbiota compositions are also influenced by postpartum environmental exposure, with the relative abundances of the following genera significantly changed after co-housing during pregnancy (Figure 5C): Parabacteroides; Clostridium sp. ND2; Gastranaerophilales; Enterococcus faecalis; Helicobacter; [Eubacterium] coprostanoligenes group; Flavobacteriaceae; Bacteroides; Parasutterella; Lactobacillus; Turicibacter; Faecalibaculum; [Eubac-terium] ventriosum group; Candidatus Saccharimonas; Acinetobacter; Odoribacter; Ruminiclostridium 5; Mollicutes RF9; Erysipelotrichaceae; and Erysipelato-clostridium (more details are shown in Supplementary Figure 8). Influences of gut microbiota on brain function were shown in our previous paper [38].

All of these results indicate that the diet during pregnancy is very important; however, the majority of the functions of the bacteria remain unknown.

Choline plays an important role in the cholinergic system

The cholinergic hypothesis plays an important role in AD therapy. To investigate the role of the gut microbiota on the conversion rate of choline, 50 mg/kg/d choline was fed to mice with different gut microbiota, including SAMP8 mice, KM mice, C57 mice, and HSHF-diet mice (at 5 weeks of age). As shown in Figure 6, body weight was slightly lowered after treatment with choline (Figure 6A), and levels of acetylcholine (Figure 6B; p < 0.05) and NADPH/NADP⁺ (Figure 6B; p < 0.05) were also changed. In addition, TMAO levels in serum (Figure 6Ca) and in brain tissue (Figure 6Cb) were changed, especially in mice pre-fed the HSHF diet (p < 0.05). Expression levels of the proteins MAOA, MAOB, and COMT in liver (Figure 6D) and in brain (Figure 6E; p < 0.05) tissues were changed; in particular, choline significantly increased the levels of these proteins in the liver of SAMP8 mice (Figure 6D; p < 0.05). Expression of the proteins AChE, AMP, CHRNA1, and CHRNB1 in the brain were also affected (Figure 6E). Together, these observations indicate that the gut microbiota influence the conversion rate of choline to TMAO, and that the TMAO level influences the cholinergic system. Furthermore, the effect of a decrease in body weight was confirmed in this study.

Figure 6. Gut microbiota plays an important role in the bioconversion of choline into TMAO and Ach. (A) The body weight changes after administration of choline; (B) Levels of acetylcholine and NADPH/NADP+ in the brain samples after administration of choline; (C) TMAO levels in serum and brain tissues after administration of choline mice; (D) Expression of the proteins MAOA, MAOB and COMT in liver; (E) Expression of the proteins MAOA, MAOB and COMT in brain tissues, and expression of the proteins AChE, AMP, CHRNA1 and CHRNB1 in the brain tissues. Data are presented as the means ± SD of more than 8 independent experiments, and more than 3 independent experiments in Western bolting. ^**p <0.05 and ^*p <0.01 vs. the control group by one-way ANOVA test.

Animal model establishment and treatment

Adult male KM mice (18–22 g, 6 weeks) obtained from the Center of Laboratory Animals of Guangdong Province (SCXK [Yue] 2008-0020, SYXK [Yue] 2008-0085) were pair-housed in plastic cages in a temperature-controlled (25 ± 2° C) colony room under a 12/12-h light/dark cycle. Food and water were available ad libitum. All experimental protocols were approved by the Center of Laboratory Animals of the Guangdong Institute of Microbiology. All efforts were made to minimize the number of animals used.

Mice were randomly allocated into two groups: control and model. The mice in the control group were fed with the standard diet, and the mice in the model group were fed a HSHF diet, and water was available freely. Mother mice were fertilized after feeding for 1 month, and continued to be fed on the HSHF diet until birth before returning back to the standard diet. All the offspring follow were fed with the standard diet. The gut microbiota, and intestinal and brain functions of the offspring were dynamically monitored at 7, 14, 28, and 56 days of age, until 16 months of age.

The components of high sugar and fat diet include 20% of sucrose, 15% of fat, 1.2% of cholesterol and 0.2 % of bile acid sodium, 10% of casein and 0.6% of calcium hydrogen phosphate, 0.4% of stone powder, 0.4% of premix, 52.2% of basic feed. Heat ratio: protein 17%, fat 17%, carbohydrate 46%.

Microbiome analysis

Fresh intestinal content samples were collected from the mouse before it fasted for 12 h, and stored at -80° C. Microbial DNA isolated from the samples, with total masses ranging from 1.2 to 20.0 ng, were stored at -20° C. Microbial 16S rRNA genes were amplified using the forward primer 5′-ACTCCTACGGGAGGCAGCA-3′ and the reverse primer 5′-GGACTACHV GGG TWTCTAAT-3′. Each amplified product was concentrated via solid-phase reversible immobilization and quantified by electrophoresis using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). After NanoDrop quantification of the DNA concentration, each sample was diluted to a concentration of 1 × 10⁹ molecules/μL in TE buffer and pooled. Twenty μL of the pooled mixture was sequenced on an Illumina MiSeq sequencing system (Illumina, San Diego, CA, USA) according to the manufacturer’s instructions. Raw pyrosequencing reads obtained from the sequencer were denoised using Titanium PyroNoise software. The resulting pyrosequencing reads were analyzed by common analysis methods, as described previously [38].

Metabolome analysis

After acquiring serum from the mouse, 80 μL serum was added to 240 μL cold methanol/acetonitrile (2:1, v/v). 10 μL internal tagging standard (L-2-chlorine-phenylalanine, 0.3 mg/mL, dissolved in methanol) was added, the samples were vortexed for 2 min, and then ultrasonic extraction was performed for 5 min. The samples were allowed to stand at -20° C for 20 min, and then centrifuged for 10 min (14,000 RPM, 4° C). 200 μL of the supernatant was loaded into a sample bottle with a lining tube for LC/MS analysis. A Waters UPLC I-class system equipped with a binary solvent delivery manager and sample manager, coupled to a Waters VION IMS Q-TOF Mass Spectrometer equipped with an electrospray interface (Waters Corporation, Milford, MA, USA) was used for LC/MS analysis. An Acquity BEH C18 column (100 mm × 2.1 mm; i.d., 1.7 μm; Waters, Milford, USA) was used for LC separation. Information on the peak picking, alignment, deconvolution, and further processing of raw LC-MS data can be found in previously published protocols [35, 40] (Sah et al. 2017; Sonnerburg et al., 2016).

Electrophysiological recordings

Standard field potential recordings were performed on the hippocampal cornu ammonis 1 (CA1) region using borosilicate glass micropipettes pulled to a tip diameter of about 1 μm and filled with 2 mol/L NaCl [49]. To record synaptic potentials, a recording electrode was placed at the CA1 apical dendrite region. Stimulus intensity was set based upon input–output relationships and was 50% of the maximal response. For testing paired pulse facilitation (PPF), two stimuli with 50% of the maximal intensity were given at 15-, 50-, 100-, and 400-ms intervals. For recording long-term potentiation (LTP), stable baseline synaptic potentials (50% of the maximal intensity) were recorded for 20 min, and then a theta-burst tetanic stimulation that contained 15 burst trains at 5 Hz was delivered (each train contained five pulses at 100 Hz). Thereafter, baseline intensity-evoked field excitatory postsynaptic potentials (fEPSPs) were recorded for 60 min with 0.33 Hz. A custom bipolar platinum wire electrode (0.08-mm diameter) was placed at the Schaffer collateral pathway, and stimulation was delivered using a Model 2100 A-M Systems Isolated Pulse Stimulator (Carlsborg, WA, USA). All evoked responses were recorded using an Axoclamp-2B amplifier, and data acquisition was controlled with pClamp 10.2 software (Molecular Devices, Sunnyvale, CA, USA).

Western blot analysis

In order to target the proteins, we used tandem mass tag (TMT) labeling for quantitative proteomic analysis. Global brain tissue was dissected from the offspring of high-fat-diet mice and proteins were extracted with radioimmunoprecipitation assay (RIPA) lysis buffer. Then, total protein samples (from the offspring of the model and control groups) were prepared. The proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membranes. After blocking with 5% nonfat dry milk in Tris-buffered saline (20 mM Tris-HCl, 500 mM NaCl, pH 7.4) with 0.2% Tween-20 (Aladdin, T104863), the membranes were probed with antibodies overnight at 4° C, followed by incubation with a horseradish peroxidase-conjugated goat anti-mouse (Servicebio; G2211-1-A) or goat anti-rabbit (Servicebio; G2210-2-A) IgG secondary antibody (1:2,000). Band intensity was quantified using ImageJ software (NIH).

The antibody of KGA/GAD (Proteintech, 12855-1-AP, 1:600), MAP2 (Proteintech, 17490-1-AP, 1:1000), GFAP (Proteintech, 20746-1-AP, 1:2000), vimentin (Proteintech, 10366-1-AP, 1:5000), IBA1 (Proteintech, 10904-1-AP, 1:1000), GAD65 (Proteintech, 20746-1-AP, 1:1000), DYNC1H1 (Proteintech, 12345-1-AP, 1:1000), myelin (Proteintech, 10458-1-AP, 1:3000), tubulin (Abcam, ab7291, 1:10000), CHAT (Proteintech, 207471-1-AP, 1:1000), OLIG2 (Proteintech, 13999-1-AP, 1:400), WFS1 (Proteintech, 11558-1-AP, 1:600), LC3A (CST, 4599S, 1:5000), LC3B (CST, 3868S, 1:5000), ACEh (Proteintech, 17975-1-AP, 1:1000), Amp (Proteintech, 13379-1-1P, 1:2000), CHRNA1 (Proteintech, 10613-1AP, 1:1000), CHRNB1 (Proteintech, 11553-1-AP, 1:1000), Tau (Affinity, AF6141, 1:1000), p-Tau (CST, 11837S, 1:1000), APOE (Absic, 5841, 1:1000), CD33 (Absic, AB32577, 1:1000), TREM2 (Abcam, ab86491, 1:500), LHX8 (Affinity, DF6778, 1:1000), PSD95 (Affinity, AF5283, 1:1000) and NF-κB (Abcam, 16502, 1:1000), MAOA (Proteintech, 10539-1-AP, 1:1000), MAOB (Proteintech, 12602-1-AP, 1:1000), and COMT (Proteintech, 14754-1-AP, 1:500) were measured.

RNA sequencing

We prepared RNA sequencing libraries using whole brain samples and performed 150-bp paired-end sequencing using an Illumina HiSeq platform (Illumina). RNA sequencing (RNA-seq) libraries were prepared from 2 μg of total RNA using the TruSeq Kit (Illumina), with the modifications. Instead of purifying poly-A RNA using poly-dT primer beads, we removed rRNA using the Ribo-Zero rRNA Removal Kit (Illumina). All other steps were performed according to the manufacturer’s protocols. RNA-seq libraries were analyzed for QC; the average size of inserts was approximately 200 to 300 bp. The sequencing library was then sequenced on the HiSeq platform.

Histopathology and immunostaining

The brain and intestine were removed and fixed in 4% paraformaldehyde at pH 7.4 for further observation of the pathology. Tissue samples were prepared as paraffin sections after drawing materials, fixation, washing, dehydration, transparency, dipping in wax, and embedding. Immunofluorescent staining of microglia and astrocytes was performed using antibodies to IBA-1, and paraffin-embedded 3-μm sections were stained with hematoxylin-eosin (H&E). Slides were observed using a light microscope and a fluorescence microscope.

Statistical analysis

All data are described as the mean ± standard deviations (SD) of at least three independent experiments. Significant differences between two groups of LTP were evaluated by two-tailed unpaired Student’s t-tests or two-tailed Welch’s t-test. Significant differences between treatments were analyzed by one-way analysis of variance (ANOVA) at p < 0.05 using the Statistical Package for the Social Sciences software (SPSS; Abacus Concepts, Berkeley, CA, USA) and Prism 5 software (GraphPad, San Diego, CA, USA).

Study limitations

There are several strengths of this study, but there are also limitations. For example, the study lacks a rescue test with single bacteria. As such, we have not identified one-on-one contacts between LHX8 and single bacteria, and therefore could not determine the signaling pathway or linking ligament responsible. Positive verification of this in future could use multi-omics studies of targeted strains in germfree mice to reveal the interaction of LHX8 with the microbiome–gut–brain axis, or use knockout LHX8 mice, to prove the high sugar and fat diet activating LHX8 to modulate cholinergic and GABAergic neurons via the gut-brain axis.

Consent for publication

No human data are reported in this manuscript.

Ethics approval

The animal protocols used in this work were approved by the Institutional Animal Care and Use committee of the Center of Laboratory Animals of the Guangdong Institute of Microbiology.