Discrimination of distinct chicken M cell subsets based on CSF1R expression

In mammals, a subset of follicle-associated epithelial (FAE) cells, known as M cells, conduct the transcytosis of antigens across the epithelium into the underlying lymphoid tissues. We previously revealed that M cells in the FAE of the chicken lung, bursa of Fabricius (bursa), and caecum based on the expression of CSF1R. Here, we applied RNA-seq analysis on highly enriched CSF1R-expressing bursal M cells to investigate their transcriptome and identify novel chicken M cell-associated genes. Our data show that, like mammalian M cells, those in the FAE of the chicken bursa also express SOX8, MARCKSL1, TNFAIP2 and PRNP. Immunohistochemical analysis also confirmed the expression of SOX8 in CSF1R-expressing cells in the lung, bursa, and caecum. However, we found that many other mammalian M cell-associated genes such as SPIB and GP2 were not expressed by chicken M cells or represented in the chicken genome. Instead, we show bursal M cells express high levels of related genes such as SPI1. Whereas our data show that bursal M cells expressed CSF1R-highly, the M cells in the small intestine lacked CSF1R and both expressed SOX8. This study offers insights into the transcriptome of chicken M cells, revealing the expression of CSF1R in M cells is tissue-specific.

Exploring the value of genomic predictions to simultaneously improve production potential and resilience of farmed animals

Sustainable livestock production requires that animals have a high production potential but are also highly resilient to environmental challenges. The first step to simultaneously improve these traits through genetic selection is to accurately predict their genetic merit. In this paper, we used simulations of sheep populations to assess the effect of genomic data, different genetic evaluation models and phenotyping strategies on prediction accuracies and bias for production potential and resilience. In addition, we also assessed the effect of different selection strategies on the improvement of these traits. Results show that estimation of both traits greatly benefits from taking repeated measurements and from using genomic information. However, the prediction accuracy for production potential is compromised, and resilience estimates tends to be upwards biased, when families are clustered in groups even when genomic information is used. The prediction accuracy was also found to be lower for both traits, resilience and production potential, when the environment challenge levels are unknown. Nevertheless, we observe that genetic gain in both traits can be achieved even in the case of unknown environmental challenge, when families are distributed across a large range of environments. Simultaneous genetic improvement in both traits however greatly benefits from the use of genomic evaluation, reaction norm models and phenotyping in a wide range of environments. Using models without the reaction norm in scenarios where there is a trade-off between resilience and production potential, and phenotypes are collected from a narrow range of environments may result in a loss for one trait. The study demonstrates that genomic selection coupled with reaction-norm models offers great opportunities to simultaneously improve productivity and resilience of farmed animals even in the case of a trade-off.