RNA-seq reveals insights into molecular mechanisms of metabolic restoration via tryptophan supplementation in low birth weight piglet model

Abstract

Low birth weight (LBW) is associated with metabolic disorders in early life. While dietary l-tryptophan (Trp) can ameliorate postprandial plasma triglycerides (TG) disposal in LBW piglets, the genetic and biological basis underlying Trp-caused alterations in lipid metabolism is poorly understood. In this study, we collected 24 liver samples from 1-mo-old LBW and normal birth weight (NBW) piglets supplemented with different concentrations of dietary Trp (NBW with 0% Trp, N0; LBW with 0% Trp, L0; LBW with 0.4% Trp, L4; LBW with 0.8% Trp, L8; N = 6 in each group.) and conducted systematic, transcriptome-wide analysis using RNA sequencing (RNA-seq). We identified 39 differentially expressed genes (DEG) between N0 and L0, and genes within "increased dose effect" clusters based on dose-series expression profile analysis, enriched in fatty acid response of gene ontology (GO) biological process (BP). We then identified RNA-binding proteins including SRSF1, DAZAP1, PUM2, PCBP3, IGF2BP2, and IGF2BP3 significantly (P < 0.05) enriched in alternative splicing events (ASE) in comparison with L0 as control. There were significant positive and negative relationships between candidate genes from co-expression networks (including PID1, ANKRD44, RUSC1, and CYP2J34) and postprandial plasma TG concentration. Further, we determined whether these candidate hub genes were also significantly associated with metabolic and cardiovascular traits in humans via human phenome-wide association study (Phe-WAS), and analysis of mammalian orthologs suggests a functional conservation between human and pig. Our work demonstrates that transcriptomic changes during dietary Trp supplementation in LBW piglets. We detected candidate genes and related BP that may play roles on lipid metabolism restoration. These findings will help to better understand the amino acid support in LBW metabolic complications.

Keywords: RNA sequencing; lipid metabolism; low birth weight; piglets; tryptophan.

Figure 1.

Growth tendency of 4 groups of piglets (kg). All 7-d-old piglets had an adaptation from days 0 to 3, and experimental diets were supplied from day 4.

Figure 2.

Identification of DEG and corresponding BP enrichment. (A) From top left to bottom left: volcano plots of DEG in L0 vs. N0, L0 vs. L4 and L0 vs. L8 comparisons, respectively (FDR < 0.1). Top 5 downregulated (blue) and upregulated genes are labeled. (B) Enrichment of BP in GO analysis of 3 comparisons (P < 0.01).

Figure 3.

Gene characteristics of DEG in L0 vs. N0. (A) Heatmap of DEG clustering based on complete linkage clustering algorithm. (B) Three categories of expression profiles of DEG based on the response to Trp dose effects. (C) BP enrichment of 3 DEG clusters (P < 0.01). GO terms in red, blue, and green represent the results from “Counter effect”, “Sensitive dose”, and “Increased dose effect” clusters, respectively. (D) Co-expression network based on Pearson correlation. Notes with purple and yellow circles represent DEG upregulated and downregulated in N0, respectively. Red and blue edges show the positive and negative relation between DEG, respectively.

Figure 4.

Identification of RBP regulating the LBW-associated ASE. (A) The count of ASE in 5 splicing patterns when L0 was control among 3 comparisons (FDR < 0.1). (B) Biological process enrichment of ASE overlapped genes in L0 vs. N0 (P < 0.05). C. Boxplot of percent spliced inclusion (PSI) level of TDO2 among 4 groups, * means P < 0.05 between L0 and N0. D. Motif enrichment analysis of LBW-associated RBP. RBP with P < 0.05 were considered as enriched in Fisher’s exact test.

Figure 5.

PID1 may play a role in metabolic recovery by Trp supplementation. (A) The bar-plot shows the top 5 averaged gene conservation scores of 5 hub genes among seven mammalian species. (B) Hierarchical tree of 7 mammalian species based on the conservation scores of PID1 compared with Sus scrofa. (C) Scatter plot shows the correlation (r) between expression levels (TPM) of PID1 and levels of postprandial plasma TG (mg/dL) across 4 groups. (D) Phe-WAS results for PID1, where P-values are determined by the t-test between metabolic traits and the corresponding types of traits (P < 0.01).