Time course of the response to ACTH in pig: biological and transcriptomic study

Abstract

Background: HPA axis plays a major role in physiological homeostasis. It is also involved in stress and adaptive response to the environment. In farm animals in general and specifically in pigs, breeding strategies have highly favored production traits such as lean growth rate, feed efficiency and prolificacy at the cost of robustness. On the hypothesis that the HPA axis could contribute to the trade-off between robustness and production traits, we have designed this experiment to explore individual variation in the biological response to the main stress hormone, cortisol, in pigs. We used ACTH injections to trigger production of cortisol in 120 juvenile Large White (LW) pigs from 28 litters and the kinetics of the response was measured with biological variables and whole blood gene expression at 4 time points. A multilevel statistical analysis was used to take into account the longitudinal aspect of the data.

Results: Cortisol level reached its peak 1 h after ACTH injection. White blood cell composition was modified with a decrease of lymphocytes and monocytes and an increase of granulocytes (F D R<0.05). Basal level of cortisol was correlated with birth and weaning weights. Microarray analysis identified 65 unique genes of which expression responded to the injection of ACTH (adjusted P<0.05). These genes were classified into 4 clusters with distinctive kinetics in response to ACTH injection. The first cluster identified genes strongly correlated to cortisol and previously reported as being regulated by glucocorticoids. In particular, DDIT4, DUSP1, FKBP5, IL7R, NFKBIA, PER1, RGS2 and RHOB were shown to be connected to each other by the glucocorticoid receptor NR3C1. Most of the differentially expressed genes that encode transcription factors have not been described yet as being important in transcription networks involved in stress response. Their co-expression may mean co-regulation and they could thus provide new patterns of biomarkers of the individual sensitivity to cortisol.

Conclusions: We identified 65 genes as biological markers of HPA axis activation at the gene expression level. These genes might be candidates for a better understanding of the molecular mechanisms of the stress response.

Fig. 1

Mean evolution of the biological variables overtime. (*): time measurement at which the expression of the variable is significantly different (F D R<0.05) from the expression at t=0. Vertical bars correspond to + and – SEM at each time point

Fig. 2

PCA on the biological variables identified as responding to ACTH, the birth and weaning weights at t=0. Colors symbolize the sex: Black = Female; Red = Male a Projection of the individuals on dimensions 1 and 2; b Projection of the variables on dimensions 1 and 2; BW = birth weight; WW = weaning weight; Lympho = lymphocyte ratio; Mono = monocytes ratio; Granulo = granulocyte ratio; RC = red cell count; Gluc = glucose; FFA = free fatty acids

Fig. 3

Multilevel PCA on the biological variables responding to ACTH. Colors symbolize the time of measurement; Black: t=0; Red: t=+1; Green: t=+4; Blue: t=+24; a Projection of the individuals on dimensions 1–2; b Projection of the variables on dimensions 1–2; Lympho = lymphocyte ratio; Mono = monocyte ratio; Granulo = granulocyte ratio; RC = red cell counts; Gluc = glucose; FFA = free fatty acids

Fig. 4

Hierarchical ascending classification of the 65 unique DE genes. A Ward method was used with an Euclidean distance matrix based on the correlations between genes. Genes are shown in column. Observations are shown in line with one line being a combination pig × time. Colors on the row dendrogram identify the time of measurement. Black: t=0; Red: t=+1; Green: t=+4; Blue: t=+24. Numbers on the column dendrogram identify each cluster

Fig. 5

Average evolution of the genes in each of the cluster identified by the HAC on the 65 unique DE genes. Evolution of each gene is translated so that it is equal to 0 at t=0; Gray: Average evolution of each of the genes in the cluster. Red: Average evolution over all genes in the cluster (cluster 1: 18; cluster 2: 17; cluster 3: 8; cluster 4: 22)

Fig. 6

PLS regression predicting the biological variables responding to ACTH from the DEG expression a projection of the observations with one point being a combination: time × pig; Black: t=0; Red: t=+1; Green: t=+4; Blue: t=+24. b projection of the variables; Blue: biological variables; Red: gene expressions; 10 genes/components were kept using a sparse approach

Fig. 7

Gene network related to glucocorticoid response in whole blood transcriptome 1 h after ACTH injection This network corresponds to the genes up-regulated 1 h after ACTH injection (cluster 1, green nodes). It combines bibliographic (best enrichment score network =45) and regulatory relationships (genes co-regulated by the same regulator in blue with the highest enrichment score (p-value=2.00E−09), green lines) proposed by Ingenuity software. Cortisol has the highest plasma level at t=+1 and acts through the glucocorticoid receptor NR3C1