Abstract
Accurate gene expression analysis via reverse transcription quantitative real-time PCR is vital for studying diabetes mellitus. Accurate and reliable normalization of RT-qPCR data, in accordance with the MIQE guidelines, demands a group of at least two stably expressed reference genes, especially in diabetic models where metabolic dysregulation often alters the profile of gene expression. To identify the best reference genes, this study systematically evaluates the stability of eleven candidate reference genes, including ACT, B2M, GAPDH, HPRT1, PPIA, RPL13A, RPLP0, TBP, UBC, YWHAZ, and 18SRNA, in liver, pancreas, and both types of tissues across four mouse groups of control, obesity, STZ-induced diabetic and treated STZ-diabetic groups. The identification of a robust group of reference genes was conducted by four widely used statistical algorithms of geNorm, NormFinder, BestKeeper, and comparative ΔCt methods. Each applies different statistical algorithms: geNorm uses pairwise variation, NormFinder considers inter- and intra-group variance, BestKeeper uses Ct variability through standard deviation and correlation analysis, and the ΔCt method evaluates expression consistency by evaluating the geometric mean of the stability rankings of each gene across all methods. Our analysis identified ACT and RPL13A as the most appropriate reference genes for liver, while UBC and RPL13A were most stable in pancreas. Overall, RPL13A and UBC demonstrated the highest expression stability across both liver and pancreatic tissues. Furthermore, our findings laid foundations for future studies in the analysis of gene expression in obese, diabetic, and diet-changed mice, as well as other liver- and pancreas-related diseases in mouse models.