Validation of Reference Genes for Oral Cancer Detection Panels in a Prospective Blinded Cohort

Background Reference genes are needed as internal controls to determine relative expression for clinical application of gene expression panels. Candidate constitutively expressed genes must be validated as suitable reference genes in each body fluid and disease entity. Prior studies have predominantly validated oral squamous cell carcinoma associated messenger RNAs (mRNAs) based on quantitative polymerase chain reaction (qPCR) quantification cycle (Cq) values without adjustment for housekeeping genes. Methods One hundred sixty eight patients had saliva collected before clinically driven biopsy of oral lesions suspicious for cancer. Seven potential housekeeping mRNAs and six pre-specified oral cancer associated mRNAs were measured with qPCR by personnel blinded to tissue diagnosis. Housekeeping gene stability was determined with the NormFinder program in a training set of 12 randomly selected cancer and 24 control patients. Genes with stability indices <0.02 were then tested in the validation set consisting of the remaining cancer and control patients and were further validated by the geNorm program. Cancer gene delta Cqs were compared in case and control patients after subtracting the geometric mean of the reference gene raw Cqs. Results B2M and UBC had stability indices >0.02 in the training set and were not further tested. MT-ATP6, RPL30, RPL37A, RPLP0 and RPS17 all had stability indices <0.02 in the training set and in the verification set. The geNorm M values were all <= 1.10. All six pre-specified cancer genes (IL8, IL1, SAT, OAZ1, DUSP1 and S100P) were up-regulated in cancer versus control patients with from nearly twofold to over threefold higher levels (p<0.01 for all based on delta Cq values). Conclusions Five reference genes are validated for use in oral cancer salivary gene expression panels. Six pre-specified oral carcinoma associated genes are demonstrated to be highly significantly up-regulated in cancer patients based on delta Cq values. These cancer and reference genes are suitable for inclusion in gene expression panels for research and clinical applications.