Cerebrum, liver, and muscle regulatory networks uncover maternal nutrition effects in developmental programming of beef cattle during early pregnancy

Abstract

The molecular basis underlying fetal programming in response to maternal nutrition remains unclear. Herein, we investigated the regulatory relationships between genes in fetal cerebrum, liver, and muscle tissues to shed light on the putative mechanisms that underlie the effects of early maternal nutrient restriction on bovine developmental programming. To this end, cerebrum, liver, and muscle gene expression were measured with RNA-Seq in 14 fetuses collected on day 50 of gestation from dams fed a diet initiated at breeding to either achieve 60% (RES, n = 7) or 100% (CON, n = 7) of energy requirements. To build a tissue-to-tissue gene network, we prioritized tissue-specific genes, transcription factors, and differentially expressed genes. Furthermore, we built condition-specific networks to identify differentially co-expressed or connected genes. Nutrient restriction led to differential tissue regulation between the treatments. Myogenic factors differentially regulated by ZBTB33 and ZNF131 may negatively affect myogenesis. Additionally, nutrient-sensing pathways, such as mTOR and PI3K/Akt, were affected by gene expression changes in response to nutrient restriction. By unveiling the network properties, we identified major regulators driving gene expression. However, further research is still needed to determine the impact of early maternal nutrition and strategic supplementation on pre- and post-natal performance.

Figure 1

Bioinformatics workflow of the multi-tissue RNA-Seq-based gene co-expression network. TS—tissue-specific genes; *DEGs—differentially expressed genes; TFs—transcription factors; CON—control; RES—restricted.

Figure 2

Tissue-to-tissue co-expression network of the bovine fetal organ transcriptome. (A) Heat map of 553 tissue-specific genes from the liver, muscle, and cerebrum tissues. (B) Co-expression network of 992 significantly co-expressed genes. Only nodes with a correlation greater than |0.9| are shown. Overlapped genes between analysis were colored based on the tissue with its maximum expression. Transcription factors are represented by a diamond shape. Nodes with few connections not linked to the main network are not showed. Heat map was constructed using pheatmap v.1.010. on R, whereas gene network was created on Cytoscape v.3.7.

Figure 3

Muscle regulatory network of differentially co-expressed genes from the bovine transcriptome. Nodes are genes with significant changes in the correlation between control and nutrient restricted fetus (q ≤ 0.05). The node size and color (from light to dark) are proportional to the number of connections for each gene. Nodes with few connections not linked to the main network are not shown. Transcription factors are represented by a diamond shape. Differentially expressed genes are labeled in blue. Edges are colored based on the differential correlation class (+/− , red; +/0, salmon; − /+ , green; − /0, green-yellow; 0/+ , magenta; 0/− , brown). Gene network was created on Cytoscape v.3.7.

Figure 4

Cerebrum regulatory network of differentially co-expressed genes from the bovine transcriptome. Nodes are genes with significant changes in the correlation between control and nutrient restricted fetus (q ≤ 0.05). The node size and color (from light to dark) are proportional to the number of connections for each gene. Nodes with few connections not linked to the main network are not showed. Only those gene pairs assigned as DEG, TS, or TF are shown. Transcription factors are represented by a diamond shape. Differentially expressed genes are labeled in blue. Edges are colored based on the differential correlation class (+/− , red; + /0, salmon; − /+ , green; − /− , deep sky blue; − /0, green-yellow; 0/+ , magenta; 0/− , brown). Gene network was created on Cytoscape v.3.7.

Figure 5

Liver regulatory network of differentially co-expressed genes from the bovine transcriptome. Nodes are genes with significant changes in the correlation between control and nutrient restricted fetus (q ≤ 0.05). The node size and color (from light to dark) are proportional to the number of connections for each gene. Nodes with few connections not linked to the main network are not showed. Transcription factors are represented by a diamond shape. Differentially expressed genes are labeled in blue. Edges are colored based on the differential correlation class (+/− , red; + /0, salmon; − /+ , green; − /0, green-yellow; 0/+ , magenta; 0/− , brown). Gene network was created on Cytoscape v.3.7.

Figure 6

Network topology of co-expressed genes between control and nutrient restricted fetuses from the liver, muscle, and cerebrum tissues in bovine. (A) Cumulative distribution functions of the gene connectivity. (B) Central reference union networks. Only those gene pairs assigned as DEG or TF are shown. Unique nodes are shown in green (control) or red (restricted). Shared nodes are shown in white; (C) Overlapping genes among the different analyses. TF—transcription factors; DE – differentially expressed gene; DC—differentially co-expressed; and DK—differentially connected. Cumulative distribution was created on R v.3.5.1; Gene network was created on Cytoscape v.3.7; and Venn diagram was created using Venny v.2.1. (https://rb.gy/jxxufy).