Assessment of reference genes for reliable analysis of gene transcription by RT-qPCR in ovine leukocytes

Abstract

With the availability of genetic sequencing data, quantitative reverse transcription PCR (RT-qPCR) is increasingly being used for the quantification of gene transcription across species. Too often there is little regard to the selection of reference genes and the impact that a poor choice has on data interpretation. Indeed, RT-qPCR provides a snapshot of relative gene transcription at a given time-point, and hence is highly dependent on the stability of the transcription of the reference gene(s). Using ovine efferent lymph cells and peripheral blood mono-nuclear cells (PBMCs), the two most frequently used leukocytes in immunological studies, we have compared the stability of transcription of the most commonly used ovine reference genes: YWHAZ, RPL-13A, PGK1, B2M, GAPDH, HPRT, SDHA and ACTB. Using established algorithms for reference gene normalization “geNorm” and “Norm Finder”, PGK1, GAPDH and YWHAZ were deemed the most stably transcribed genes for efferent leukocytes and PGK1, YWHAZ and SDHA were optimal in PBMCs. These genes should therefore be considered for accurate and reproducible RT-qPCR data analysis of gene transcription in sheep.

Introduction

The advent of next-generation sequencing has led to a vast increase in available genomic and transcriptomic data, which combined with the paucity of available reagents for analysis of protein expression in most veterinary species, has led to an increased use of quantitative reverse transcription PCR (RT-qPCR) for gene transcription profiling (Bustin et al., 2005, Bustin et al., 2009). Despite well-documented shortcomings of this technique associated with major differences in mRNA quantity, quality and optimal assay design (Nolan et al., 2006), RT-qPCR is becoming the method of choice for RNA quantification. Measuring gene transcription levels in different tissues and cell types requires accurate normalization, most often achieved using an internal reference gene. Ideal reference genes should exhibit a relatively consistent transcription profile in different cell cycle stages and experimental conditions, as failing to do this will result in erroneous interpretation of the results (Vandesompele et al., 2002).

Large animals in general, and sheep in particular, are increasingly being used in immunological research (Davenport et al., 2014, Mahakapuge et al., 2015, Neeland et al., 2014, Scheerlinck et al., 2008). In sheep, PBMCs and lymph are the most commonly used sources of leukocytes, as they are readily and repetitively accessible and arguably represent the level of immune activity within the animal. Using pre-femoral efferent lymphatic cells and PBMCs from ten physiologically normal sheep we compared the transcriptional stability of reference genes with or without Con-A mediated T-cell activation, to determine the most appropriate reference genes in sheep leukocytes.

To identify the most consistently transcribed reference genes among the selected reference genes, we used statistical algorithms, geNorm version 3.5 (Center for Medical Genetics, Ghent University Hospital, Belgium) (Vandesompele et al., 2002) and Norm Finder (Version 20) (Andersen et al., 2004) according to developers’ recommendations. GeNorm software identifies the most consistently transcribed reference genes in a set of samples tested using average pairwise variation. The outcome is expressed as an “M-value”, which can be defined as the average pairwise variation of a particular reference gene compared to all other reference genes. The ideal reference gene for a particular tissue used in different experimental conditions should be transcribed equally in all the samples selected (i.e. minimum M-value). Additionally, the software also calculates pairwise variations, providing a geNorm V-value, which reflects levels of variation in average reference gene stability with the addition of n + 1^th reference gene (Vandesompele et al., 2002). Similarly, the Norm Finder algorithm also identifies optimal normalization genes amongst a given set of candidates based on overall transcription variation (Andersen et al., 2004). The software provides a stability value for each gene (as a direct measure of gene transcription variation) (Andersen et al., 2004). We used eight reference genes and 10 samples for identification of the most consistently transcribed reference genes.