Exploring a DNA methylation field effect in renal cell carcinoma and its implications for biomarker research

A field effect, the phenomenon of molecular alterations, such as DNA methylation, in normally appearing tissue surrounding a malignant tumor, has been suggested to be associated with malignant transformation. Nevertheless, normally appearing tissue adjacent to the tumor is frequently used as control tissue in biomarker studies. Here, we aimed to evaluate DNA methylation alterations in tissue surrounding malignant renal cell carcinoma (RCC), and demonstrate the impact of these alterations on the evaluation of biomarker potential. We evaluated five potential DNA methylation markers ( ANGPTL6 , ANKRD34B , CAIX , NHLH2 , and ZIC1 ) across several sampled areas of eleven kidneys using qMSP. To evaluate the impact of the field effect on biomarker performance, the percentage of methylated reference (%PMR) was calculated and samples were analyzed with ROC curve analysis using RCC tissues as cases with either adjacent normal (AN) or normal kidney (NK) tissues as controls. DNA methylation was present in RCC tissues (ranging from 26 to 59%), and matched AN tissues (ranging from 0 to 54%) as compared to NK (2% in all genes). All AN samples (100%) with NHLH2 methylation also showed methylation in the corresponding RCC tissue, whereas this was true for only 42.8–75% of the other studied genes. No methylation was detected in the AN samples for ANKRD34B . The methylation rates varied widely between markers and patients. A gradual decline in %PMR all genes was found with increasing distance from the tumor (12.4%) towards the area furthest from the tumor (5.8%). Analyzing biomarker performance using AN as control tissues yielded lower sensitivities for most biomarkers (11.6–55.8%) compared to using NK as controls (25.6–59.3%). Here, we evaluated DNA methylation alterations in normally appearing tissue surrounding RCC, and support the possibility of an RCC-associated DNA methylation field effect. More importantly, this data emphasizes the importance of using appropriate control tissues to avoid underestimation of biomarker potential. Therefore we advocate to use normal tissue of non-cancerous patients as controls rather than AN in diagnostic biomarker studies, to avoid underestimation of biomarker performance.