Effects of Dietary Vitamin B6 Restriction on Hepatic Gene Expression Profile of Non-Obese and Obese Mice

Four-week old C57BL/6J male mice were purchased from Raon Bio (Yongin, Korea) and housed in a temperature (22 ± 3 °C) and humidity (50 ± 10%) controlled room with a 12 h light–dark cycle. After 12 days of acclimation [5 days with a chow diet and 7 days with a control low-fat (12% kcal from fat) diet with 7 mg pyridoxine hydrochloride (PN)/kg diet], mice were randomly divided into the four groups (n = 8–9) and fed an LF7 (low-fat diet with 7 mg PN/kg diet), LF1 (low-fat diet with 1 mg PN/kg diet), HF7 [high-fat (45% kcal from fat) diet with 7 mg PN/kg diet], or HF1 (high-fat diet with 1 mg PN/kg diet) diet. The experimental diets were provided ad libitum for 16 weeks. The composition of the diets is shown in Supplementary Materials Table S1. The recommended amount of PN is a 7 mg/kg diet according to the AIN-93 formulation [21] and a 1 mg PN/kg diet has been reported as the minimum level to prevent a growth retardation [22]. At the end of the experiments, the animals were fasted for 12 h and sacrificed after intraperitoneal injection of 30 mg/kg Zoletil (Virbac, Carros, France) and 10 mg/kg Rompun (Bayer Korea, Seoul, Korea). Blood samples were collected by cardiac puncture. To obtain serum, blood was left at room temperature for 1 h and centrifuged at 3000 rpm, 4 °C for 20 min. The organs were excised, frozen in liquid nitrogen, and stored at −80 °C until analysis. All animal experiments were approved by Institute of Laboratory Animals Resources of Seoul National University (SNU-161111-1-2), and conducted in accordance with the guidelines of Institutional Animal Care and Use Committee of Seoul National University.

Serum glucose, triglyceride (TG), total cholesterol, and glutamate-pyruvate transaminase (GPT) concentrations were analyzed using commercial colorimetric assay kits (Asan Pharmaceutical Co., Seoul, Korea) according to the manufacturer’s protocol. After extraction of hepatic total lipids [23], cholesterol concentrations were determined using the same commercial kit. Hepatic TG and total bile acid concentrations were determined by thin-layer chromatography [24] and a total bile assay reagent kit (Gentaur, San Jose, CA, USA), respectively. Hepatic total homocysteine concentrations were measured using high-performance liquid chromatography, as previously described [25].

The dissected liver tissues were fixed with formalin, embedded in paraffin, sectioned into 4 μm-thick sections, and stained with hematoxylin and eosin (H&E) for histological analysis. The morphology was examined under an Olympus BX50 microscope using a DP-72 digital camera (Olympus, Tokyo, Japan) and captured using the Image-Pro Plus program (Media Cybernetics Inc., Rockville, MD, USA).

Total hepatic RNA was extracted using RNAiso Plus (Takara Bio Inc., Shiga, Japan) and RNA purity and integrity were confirmed by Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Transcriptome analysis was performed with Affymetrix Mouse Clariom S arrays (Thermo Fisher Scientific, Waltham, MA, USA) as described previously [26]. To identify differentially expressed genes (DEGs) among groups, statistical significance of the expression data was determined using one-way analysis of variance (ANOVA, p < 0.05). DEGs between two groups were selected based on the cut-off criteria (|fold change| ≥ 1.5 and p < 0.05). Functional analysis of DEGs was performed using the Gene Ontology (GO) database, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway, and WikiPathways. False discovery rate was corrected by p-value using the Benjamini–Hochberg algorithm. Heatmaps and volcano plots were constructed using the PermutMatrix software and R package, respectively.

Experimental results except for microarray data were analyzed using SPSS software (version 23; IBM SPSS Inc., Armonk, NY, USA). Data are expressed as the mean ± SEM and were analyzed by one-way ANOVA followed by Duncan’s multiple range test. Differences with p < 0.05 were considered statistically significant.

Consumption of a high-fat diet significantly increased body weight in both vitamin B6-sufficient and -deficient groups (Figure 1a). Relative weights of liver and epididymal fat were also significantly increased by high-fat feeding (Figure 1b,c). Both body weight and relative organ weights were not significantly changed by dietary vitamin B6 amount.

High-fat feeding significantly increased serum glucose concentrations, which were decreased by dietary vitamin B6 restriction (Figure 1d). Serum triglyceride concentrations were significantly reduced by high-fat feeding, which tended to increase in mice fed a vitamin B6-restricted diet (Figure 1e). No significant effect of dietary vitamin B6 restriction was observed in serum cholesterol, serum GPT, hepatic triglyceride, and hepatic cholesterol levels, which were significantly higher in HF-fed mice compared to LF-fed mice (Figure 1f–i, Supplementary Materials Figure S1). Hepatic bile acid levels were not significantly different between the groups (Figure 1j). Previous studies have reported that mice with hepatic steatosis exhibited increased free fatty acid uptake and TG synthesis [27]. Furthermore, stimulation of lipoprotein lipase activity has been shown to be associated with reduced serum triglyceride concentrations in subjects with hepatic steatosis [28].

Global gene expression was analyzed to identify signatures in hepatic genes related to experimental diets. Based on microarray analysis, expressions of 4000 genes were significantly altered and were largely dependent on fat content rather than vitamin B6 content in the diet (Figure 2a). Figure 2b,c show volcano plots of DEGs altered by vitamin B6 amount in mice fed a low-fat and high-fat diet, respectively. Among 108 DEGs in the LF1/LF7 comparison and 44 DEGs in HF1/HF7 comparison, Venn diagram shows only three overlapping DEGs, Asl, Nnmt, and Sds (Figure 2d). All of these genes encode PLP-binding proteins that are involved in amino acid metabolism, including the urea cycle, amino acid methylation, and serine metabolism, respectively.

Furthermore, we found the distinct effects of dietary vitamin B6 restriction in mice fed either LF or HF. According to the KEGG enrichment heatmap, “Jak-STAT signaling pathway” and “steroid biosynthesis”, were significantly associated with DEGs in the LF1/LF7 comparison (Figure 2e). On the other hand, KEGG pathways, such as “carbon metabolism”, “biosynthesis of amino acids”, “alanine, aspartate and glutamate metabolism”, “cysteine and methionine metabolism”, and “phenylalanine, tyrosine and tryptophan biosynthesis” were significantly associated with DEGs in the HF1/HF7 comparison.

Gene functional analysis revealed that 54 upregulated DEGs in the LF1/LF7 comparison were enriched in GO terms in the biological process (BP) category, including “sterol biosynthetic process”, “secondary alcohol biosynthetic process”, and “organic hydroxy compound metabolic process”. The top 10 enriched GO terms are described in Supplementary Materials Table S2. Expression levels of 11 DEGs enriched in the top 10 GO terms are visualized in the heatmap (Figure 3a). We also identified DEGs involved in cholesterol metabolism based on the Wikipathway and KEGG pathway (Figure 3b). Expression levels of Hmgcr, Idi1, Fdps, Lss, Cyp51, Msmo1, Nsdhl, and Cyp39a1 were significantly higher, whereas those of the Hsd3b7 gene were significantly lower in the LF1 group compared to the LF7 group.

We observed the enriched BP terms related to immune responses, including “regulation of viral process”, “response to cytokine”, “response to interferon-gamma”, and “innate immune response” among the significantly enriched GO terms of 54 downregulated DEGs (Supplementary Materials, Table S3). Expression levels of 20 DEGs enriched in the top 10 GO terms are visualized in the heatmap (Figure 3c).

Using functional enrichment analysis of 26 upregulated DEGs in the HF1/HF7 comparison, we identified GO terms in the BP category associated with amino acid and organic acid metabolism, including “cellular amino acid catabolic process”, “alpha-amino acid metabolic process”, and “carboxylic acid catabolic process”. Furthermore, “pyridoxal phosphate binding”, a GO term in molecular function category, was annotated. The top 10 enriched GO terms were described in Supplementary Materials, Table S4. Expression levels of seven genes enriched in the top 10 GO terms are visualized in the heatmap (Figure 4a).

Expression profile of significant genes (one-way ANOVA) enriched in KEGG “cysteine and methionine metabolism” pathway is shown in Figure 4b. Among these genes, mRNA levels of Cth, Got1, Lao1, Mat2a, Sds, and Tat, were significantly upregulated in the HF1 group compared to the HF7 group (Figure 4b). In addition, consumption of a vitamin B6-restricted diet increased the hepatic homocysteine concentration in both LF and HF-fed mice (Figure 4c). Positive associations were observed between hepatic homocysteine levels and mRNA levels of Cth (p = 0.012), Lao1 (p = 0.005), and Mat2a (p = 0.042) in high-fat fed mice.

There were no significantly enriched GO terms observed in the enrichment analysis of 18 downregulated DEGs.

We examined the expression of genes involved in vitamin B6 metabolism. Expression levels of Pdxk, Pdxp, and Pnpo, which convert B6 vitamers to each other, were not significantly different among the groups. On the other hand, expression levels of Aox1 and Aox3, which convert pyridoxal to 4-pyridoxate, a form excreted in urine, were significantly lower in the LF1 group compared to the other groups (Figure 5a). Significant differences were not observed in Aox1 and Aox3 mRNA levels of mice fed a high-fat diet in response to dietary restriction of vitamin B6.

In addition, we observed the overall down-regulation of PLP-dependent enzyme-encoding genes in response to high-fat consumption (Figure 5b). Among 11 DEGs enriched in the GO term “pyridoxal phosphate binding”, expression of Sptlc2 and Sptlc3 genes involved in lipid metabolism were upregulated and the remaining nine DEGs were downregulated in the high-fat groups compared to the low-fat groups, implicating the overall downregulation of pathways involving PLP as a cofactor in response to a high fat consumption.