Identification of reliable reference genes for quantitative gene expression studies in oral squamous cell carcinomas compared to adjacent normal tissues in the F344 rat model.

Oral squamous cell carcinomas (OSCCs) induced in F344 rats by 4-nitroquinoline-1-oxide (4-NQO) demonstrate considerable phenotypic similarity to human oral cancers and the model has been widely used for carcinogenesis and chemoprevention studies. Molecular characterization of this model needs reliable reference genes (RGs) to avoid false- positive and -negative results for proper interpretation of gene expression data between tumor and adjacent normal tissues. Microarray analysis of 11 pairs of OSCC and site-matched phenotypically normal oral tissues from 4-NQO-treated rats identified 10 stably expressed genes in OSCC compared to adjacent normal tissues (p>0.5, CV<15%) that could serve as potential RGs in this model. The commonly used 27 RGs in the rat were also analyzed based on microarray data and most of them were found unsuitable for RGs in this model. Traditional RGs such as ACTB and GAPDH were significantly altered in OSCC compared to adjacent normal tissues (p<0.01, n=11); however, the Hsp90ab1 was ranked as the best RG candidate and the combination of Hsp90ab1 and HPRT1 was identified by NormFinder to be a superior reference for gene normalization among the commonly used RGs. This result was also validated by RT-PCR based on the selected top RG candidate pool. These data suggest that there are no common RGs suitable for different models and RG(s) should be identified before gene expression analysis. We successfully identified Hsp90ab1 as a stable RG in 4-NQO-induced OSCC compared to adjacent normal tissues in F344 rats. The combination of two stably expressed genes may be a better option for gene normalization in tissue samples.