Quantitative TaqMan real-time PCR assays for gene expression normalisation in feline tissues

Gene expression analysis is an important tool in contemporary research, with real-time PCR as the method of choice for quantifying transcription levels. Co-analysis of suitable reference genes is crucial for accurate expression normalisation. Reference gene expression may vary, e.g., among species or tissues; thus, candidate genes must be tested prior to use in expression studies. The domestic cat is an important study subject in both medical research and veterinary medicine. The aim of the present study was to develop TaqMan real-time PCR assays for eight potential reference genes and to test their applicability for feline samples, including blood, lymphoid, endocrine, and gastrointestinal tissues from healthy cats, and neoplastic tissues from FeLV-infected cats.

RNA extraction from tissues was optimised for minimal genomic DNA (gDNA) contamination without use of a DNase treatment. Real-time PCR assays were established and optimised for v-abl Abelson murine leukaemia viral oncogene homolog (ABL), beta-actin (ACTB), beta-2-microglobulin (B2M), beta-glucuronidase (GUSB), hydroxymethyl-bilane synthase (HMBS), hypoxanthine phosphoribosyltransferase (HPRT), ribosomal protein S7 (RPS7), and tryptophan 5-monooxygenase activation protein, zeta polypeptide (YWHAZ). The presence of pseudogenes was confirmed for four of the eight investigated genes (ACTB, HPRT, RPS7, and YWHAZ). The assays were tested together with previously developed TaqMan assays for feline glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and the universal 18S rRNA gene. Significant differences were found among the expression levels of the ten candidate reference genes, with a ~106-fold expression difference between the most abundant (18S rRNA) and the least abundant genes (ABL, GUSB, and HMBS). The expression stability determined by the geNorm and NormFinder programs differed significantly. Using the ANOVA-based NormFinder program, RPS7 was the most stable gene in the tissues studied, followed by ACTB and ABL; B2M, HPRT, and the 18S rRNA genes were the least stable ones.

The reference gene expression stability varied considerably among the feline tissues investigated. No tested gene was optimal for normalisation in all tissues. For the majority of the tissues, two to three reference genes were necessary for accurate normalisation. The present study yields essential information on the correct choice of feline reference genes depending on the tissues analysed.