Transcriptome Analysis in Spleen Reveals Differential Regulation of Response to Newcastle Disease Virus in Two Chicken Lines

Abstract

Enhancing genetic resistance of chickens to Newcastle Disease Virus (NDV) provides a promising way to improve poultry health, and to alleviate poverty and food insecurity in developing countries. In this study, two inbred chicken lines with different responses to NDV, Fayoumi and Leghorn, were challenged with LaSota NDV strain at 21 days of age. Through transcriptome analysis, gene expression in spleen at 2 and 6 days post-inoculation was compared between NDV-infected and control groups, as well as between chicken lines. At a false discovery rate <0.05, Fayoumi chickens, which are relatively more resistant to NDV, showed fewer differentially expressed genes (DEGs) than Leghorn chickens. Several interferon-stimulated genes were identified as important DEGs regulating immune response to NDV in chicken. Pathways predicted by IPA analysis, such as "EIF-signaling", "actin cytoskeleton organization nitric oxide production" and "coagulation system" may contribute to resistance to NDV in Fayoumi chickens. The identified DEGs and predicted pathways may contribute to differential responses to NDV between the two chicken lines and provide potential targets for breeding chickens that are more resistant to NDV.

Figure 1

Validation analysis by Fluidigm Biomark assay of log₂ fold change (Log₂FC) of selected genes that were significant for different contrasts in RNA-seq analysis. Contrasts between different line by treatment combinations are marked in different colors and each combination is labeled as Line_Treatment (FA: Fayoumi, LE: Leghorn, CT: Control, NDV: Newcastle disease virus). Pearson correlation coefficient is labeled as “r”. Log₂FC in Biomark assay equals −ΔΔCt for each comparison. Average cycle threshold (Ct) value for each group is the mean of samples in that group. Average expression of three housekeeping genes including GAPDH, ACTB and HPRT1 were used for normalization of Ct values.

Figure 2

Principal component analysis (PCA) plots generated with ggbiplot in R showing variation and clustering of samples in different groups. (A) PCA plot for samples at 2 dpi. (B) PCA plot for samples at 6 dpi. Horizontal and vertical axis show two principal components that respectively explain variation between different lines and that between different treatments. Sample clusters in different groups are circled with ellipses. Different groups are represented in different shapes and colors: Fayoumi (blue), Leghorn (red), NDV (triangle), Control (circle).

Figure 3

Number of significant differentially expressed genes (DEGs) at a false discovery rate <0.05 at 2 and 6 dpi. (A) Number of DEGs for NDV vs. Control comparisons within each line; (B) Number of DEGs for Fayoumi vs. Leghorn comparisons within each treatment. Upregulated and downregulated DEGs are represented in red and green color, respectively. DEGs within different range of Log₂FC are shown in the color of different intensities.

Figure 4

Differences of significant DEGs and biofunctions between Fayoumis and Leghorns are altered by NDV from 2 to 6 dpi. (A) Venn diagram showing unique DEGs for Fayoumi vs. Leghorn induced by NDV (red color) at different time points. (B) Heat map showing Diseases and Biofunctions predicted by IPA based on the DEGs. Prediction of lower or higher activity of a disease or biofunction in Fayoumis than Leghorns is calculated as negative or positive z-score and colored in blue or orange, respectively, in the heat map. The intensity of the color in the heat map based on |z-score| indicates degree of the predicted difference.

Figure 5

Comparison between two chicken lines on Diseases and Biofunctions predicted by IPA for NDV vs. Control at 2 dpi. Prediction of activation or inhibition of a disease or biofunction by NDV is calculated as negative or positive z-score and colored in blue or orange, respectively, in the heat map. The intensity of the color in the heat map based on |z-score| indicates robustness of the prediction.

Figure 6

WGCNA co-expression analysis revealed important modules with strong correlations with treatment, line, and viral load in tears. (A) Cluster dendrogram showing 19 modules of genes identified by WGCNA with similar expression patterns across different group. (B) Heatmap showing correlation of each module eigengene with treatment, line, and log₁₀ (viral copy number in lachrymal fluid). Red and green colors respectively represent positive and negative correlations. Intensity of the color is determined by the absolute value of the correlation coefficient, which is also labeled in each box with p-value in parenthesis. (C) Genes in lightcyan module showed significant enrichment (enrichment score >2) of several immune cell types (colored in red) in CTen analysis. Red line in the circle indicates enrichment score calculated as −log₁₀ (Benjamini-Hochberg adjusted P value). The color around the circle changes from blue to purple as the enrichment score decreases.