The biological signal the team hunted for is DNA methylation, a chemical modification in which methyl groups are attached to cytosine bases, most often at cytosine-guanine dinucleotides known as CpG sites. Methylation patterns act as a layer of gene regulation, silencing or activating genomic regions without altering the underlying DNA sequence. Because tumors shed DNA and because immune and other blood cells undergo measurable epigenetic shifts in response to cancer, methylation signatures detectable in blood have attracted growing interest as biomarkers for a range of malignancies. Yet while several methylation biomarkers associated with renal cell carcinoma have been found in tumor tissue, blood-derived methylation markers for the disease have remained scarce, leaving clinicians dependent on computed tomography scans and ultrasonography for early detection.

The Japanese team’s methodological choice proved decisive. Most epigenome-wide association studies, or EWAS, rely on microarrays such as the Illumina Infinium MethylationEPIC chip, which can interrogate only about 3 to 3.5 percent of the CpG sites in the human genome. That limitation means potentially valuable biomarkers may simply be invisible to conventional screens. Ohmomo and colleagues instead used targeted bisulfite sequencing, or TB-seq, a technique in which DNA fragments are treated with bisulfite to convert unmethylated cytosines into uracils, then captured and sequenced at millions of CpG positions. This approach proved essential for one region in particular: the segment of chromosome 5 that ultimately yielded the study’s key markers is densely homologous to other parts of the genome, making it impossible to amplify with standard PCR primers. Custom hybridization probes allowed the researchers to capture and sequence the region accurately where conventional methods would have failed.

The study was structured in two phases. In the discovery phase, the team analyzed whole blood DNA from 50 ccRCC patients treated at the National Cancer Center Hospital and 50 age- and sex-matched healthy controls drawn from the Tohoku Medical Megabank Community-Based Cohort. They performed an EWAS using a linear regression model adjusted for age, sex, and the estimated cellular composition of each blood sample, a critical correction because different leukocyte populations carry distinct methylation profiles. The model showed minimal statistical inflation, with a lambda value of 1.030, indicating that the results were not driven by confounding from cell-type mixtures. In the replication phase, the same analysis was repeated in an entirely independent set of 48 ccRCC patients and 48 controls to test whether the initial findings would hold.

The discovery screen identified eleven CpG sites with p-values below the suggestive threshold of 1.00 × 10⁻⁶, and six of those crossed the stringent Bonferroni-corrected significance threshold of 1.59 × 10⁻⁸. All six mapped to a single region on chromosome 5 spanning the PCBD2 gene and the adjacent pseudogene MTND4P12, within a stretch of DNA between positions 134,923,256 and 134,928,594. Notably, the six sites were all hypomethylated in the cancer patients, showing roughly 10 percent lower methylation levels than in the healthy controls. The surrounding genomic landscape is densely populated with DNase I hypersensitive sites, transcription factor binding sites, and binding sites for CCCTC-binding factor (CTCF), an architectural protein that organizes chromatin into loops. That structural context suggests the methylation changes could influence chromatin conformation and, in turn, gene regulation across the neighborhood.

The replication phase delivered strong confirmation. Four of the six CpGs again crossed the genome-wide significance threshold, set at 3.42 × 10⁻⁸ for the smaller number of sites tested, and all six showed the same directional pattern of approximately 10 percent hypomethylation in patients. The association survived additional statistical corrections for metabolic differences between groups, including hemoglobin A1c, cholesterol measures, diabetes, dyslipidemia, and chronic kidney disease, in both phases. A Jonckheere–Terpstra trend test revealed that methylation levels differed significantly between controls and patients at every cancer stage, although the levels did not track progressively with advancing stage, suggesting the marker reflects the presence of the disease rather than its extent.

To translate the finding into a practical test, the researchers evaluated how well the six CpGs could discriminate patients from controls using receiver operating characteristic analysis. Individually, each site performed well, but the sum of methylation levels across all six sites proved the strongest classifier. In the discovery cohort, the combined score achieved an area under the ROC curve of 0.922, with a 95 percent confidence interval of 0.871 to 0.973. In the independent replication cohort, the score remained robust at 0.871, with a confidence interval of 0.800 to 0.941. An AUC near 0.9 approaches the performance of clinically useful screening tests, and the fact that the signal persisted in a separate group of patients and controls lends credibility that many single-cohort biomarker discoveries lack. Only a small fraction of previously reported cancer methylation biomarkers, the authors note, have ever been validated, a problem they attribute in part to the false-positive burden of microarray-based discovery.

The team then probed what the methylation changes might actually do. Using the iMETHYL database, which integrates methylation, gene expression, and genetic variation data from blood, they performed cis-expression quantitative trait methylation analysis, examining 56 annotated genes within one million base pairs of the PCBD2/MTND4P12 region. Two emerged as significantly linked: TCF7, located 0.8 million base pairs upstream, and VDAC1, located 0.9 million base pairs upstream. Hypermethylation at one CpG was associated with increased TCF7 expression, while hypermethylation at another correlated with decreased VDAC1 expression, meaning the hypomethylation seen in ccRCC patients would correspond to reduced TCF7 and elevated VDAC1 activity. Both genes have plausible ties to cancer biology. VDAC1 encodes a voltage-dependent anion channel on the outer mitochondrial membrane that governs the transport of adenine nucleotides, calcium ions, and metabolites, and its overexpression has been documented in multiple cancer cell lines and linked to mitochondrial abnormalities in ccRCC. TCF7 encodes the transcription factor TCF1, a regulator of T cell and innate lymphoid cell development whose overexpression has been reported in ccRCC tumor tissue.

The authors are careful to frame the work as a promising beginning rather than a finished diagnostic. The study population was exclusively Japanese, so whether the six CpGs behave identically in other ethnic groups remains unknown. The analysis covered only clear cell renal cell carcinoma, leaving open the question of whether the markers are specific to ccRCC or shared across cancer types. And because the methylation changes were measured in blood cells rather than tumor tissue, the findings do not by themselves explain the etiology of the disease; the researchers did not directly measure TCF7 or VDAC1 expression in the patients’ blood cells, an experiment they flag as necessary. Why hypomethylation of PCBD2/MTND4P12 arises in the blood of ccRCC patients is a question for future work.

Even with those caveats, the study makes a methodological statement that resonates beyond kidney cancer. By demonstrating that sequencing-based methylation profiling can uncover disease-associated regions invisible to commercial microarrays, and by validating those regions in an independent cohort with strong discriminatory power, the Japanese team has shown a path toward blood-based epigenetic screening that could extend to other cancers and diseases. The six CpGs in PCBD2/MTND4P12 now stand as candidate biomarkers awaiting larger, multi-ethnic, prospective trials. If they survive that scrutiny, a simple blood test could join imaging as a first line of defense against a cancer that today is too often found by accident, or too late.

Subject of Research: Whole blood DNA methylation biomarkers for detecting clear cell renal cell carcinoma

Article Title: Potential DNA methylation biomarkers for the detection of clear cell renal cell carcinoma identified by a whole blood-based epigenome-wide association study

Article References: Ohmomo, H., Komaki, S., Sutoh, Y., Hachiya, T., Ono, K., Arai, E., Fujimoto, H., Yoshida, T., Kanai, Y., Asahi, K., Sasaki, M., & Shimizu, A. (2022). Potential DNA methylation biomarkers for the detection of clear cell renal cell carcinoma identified by a whole blood-based epigenome-wide association study. Epigenetics Communications, 2(1), Article 2. https://doi.org/10.1186/s43682-022-00009-7

Image Credits: AI Generated

DOI: 10.1186/s43682-022-00009-7

Keywords: clear cell renal cell carcinoma, DNA methylation, epigenome-wide association study, biomarkers, targeted bisulfite sequencing, whole blood, PCBD2, MTND4P12, kidney cancer, early detection, TCF7, VDAC1