DNA Methylation of Two Long Non-Coding RNA Loci Tracks Metastatic Disease in Kidney Cancer

Researchers have identified two long non-coding RNA genes whose DNA methylation patterns shift markedly as clear cell renal cell carcinoma progresses toward metastatic disease. A study published in Clinical Epigenetics reports that the LINC01354 and LINC01510 loci become progressively hypermethylated in high-stage and high-grade tumors, and that methylation at these sites is most pronounced in established metastatic lesions. The work, led by Pouriya Faraj Tabrizi, Inga Mareike Schimansky and colleagues at Eberhard Karls University Tübingen and Hanover Medical School, adds two RNA regulatory genes to a growing list of epigenetically altered loci that track aggressive renal cancer, while stopping short of claiming any causal role in disease progression.

Clear cell renal cell carcinoma is the most common histological subtype of kidney cancer, accounting for roughly 75 percent of all renal cell carcinoma cases. More than a quarter of patients already carry metastatic disease at the time of diagnosis, and validated biomarkers for predicting metastatic progression remain limited. The study’s authors argue that this gap justifies a systematic search for epigenetic alterations that distinguish localized tumors from those that have spread. Their starting point was a discovery experiment in which genome-wide DNA methylation profiles were generated using Illumina EPIC arrays for 22 primary tumors without distant metastases and 22 primary tumors from patients who had developed distant metastases.

That profiling exercise, analyzed with the Bioconductor packages minfi and bumphunter, yielded ten candidate differentially methylated regions with family-wise error rates of 0.03 or lower. Two of these regions stood out: one within LINC01510 on chromosome 7, ranked first with a family-wise error rate below 0.001, and one within LINC01354 on chromosome 1, ranked fourth with a value of 0.025. At the CpG sites within both candidate regions, median methylation values were significantly higher in tumors from metastasis-positive patients, and the confidence intervals of the medians did not overlap between the two groups. Seven of the eight candidate CpG sites identified in the study were not covered by the older HM450K array used in The Cancer Genome Atlas kidney cancer project, meaning the improved EPIC arrays allowed the team to examine sites that previous datasets could not address.

To confirm that the array findings were not artifacts of a single measurement platform, the researchers designed three pyrosequencing assays interrogating the candidate CpG sites or sites immediately adjacent to them. The validation cohort was substantially larger than the discovery set: 145 primary renal cell carcinoma tissues, 124 paired tumor-adjacent normal tissues, and 194 metastatic tissue samples obtained from 101 patients with metastatic disease. Before turning to patient tissues, the team also profiled 18 cell models, including primary renal epithelial cells, six renal cell carcinoma lines, prostate, urothelial and cervical cancer cell lines. Renal cell carcinoma and cervical cancer cells showed generally high methylation levels of 50 to 98 percent at the candidate regions, while prostate and urothelial lines were more variable, and the aggressive DU-145 prostate line exceeded 90 percent at both loci.

The paired comparison of tumors with their adjacent normal tissues produced a more nuanced picture. Of the eight CpG sites analyzed, only two, both within LINC01510, showed statistically significant hypermethylation in tumors, and the effect sizes were small by Cohen’s d criteria. Strip plots of the paired data revealed considerable heterogeneity: for LINC01354, substantial subsets of tissue pairs showed either moderate hypomethylation or more pronounced hypermethylation in the tumor, and a similar although smaller hypermethylated subgroup appeared for LINC01510. This intertumoral variability, the authors note, limited the overall effect sizes and left several LINC01354 sites without statistical significance in the paired analysis.

Comparisons between independent tumor groups told a clearer story. When tumors were stratified by stage, all measured CpG sites in both candidate regions showed significantly higher methylation in high-stage tumors, defined as stage T3 or above, with odds ratios of 1.03 to 1.05 per one percent increase in methylation and a uniform p-value of 0.001 after Benjamini–Hochberg adjustment for multiple testing. High-grade tumors, grade G3 or above, showed the same pattern across all sites, with odds ratios of 1.03 to 1.05 and p-values of 0.003. The association with metastatic status was weaker: only one site, PS_307-CG1 within LINC01510, reached significance when primary tumors from patients with distant metastases were compared with those from patients without, at an odds ratio of 1.04 per percent methylation and p equal to 0.031. The authors attribute the limited power of this comparison to the imbalance between group sizes and the high variability of methylation values, which widened confidence intervals, and they caution that larger cohorts are needed to determine whether methylation at these loci is associated with metastatic status in primary tumors.

The most striking differences emerged when metastatic tissues were compared directly with localized primary tumors in a more balanced design. All candidate CpG sites in both loci were substantially more methylated in metastatic samples, with odds ratios of 1.05 to 1.09 per one percent increase in methylation and p-values below 0.001. Mean methylation values differed by roughly 25 to 30 percentage points at some sites, and when expressed as differences between group means, the odds ratios reached as high as 8.9 for the PS_306-CG2 site in LINC01510. The hypermethylation effects were most pronounced for the LINC01510 sites, which the authors describe as showing the strongest metastasis-associated methylation among all the RNA regulatory loci their group has analyzed, including the previously reported LINC00404. Age and sex contributed significantly to none of the models except for a possible age effect at one site.

Because the discovery and validation cohorts came from a single research program, the team turned to The Cancer Genome Atlas kidney renal clear cell carcinoma collection for independent support. One CpG site, cg21546522, was shared between the two datasets, and two adjacent sites were available only in the atlas data. In silico analysis of these three sites showed associations between hypermethylation and nearly all adverse clinical and outcome parameters examined, including distant metastatic status, high stage and high grade, though not lymph node metastasis. Overall survival was significantly associated with methylation at all three sites in univariate analyses, and at two of the three sites in multivariate Cox regression. For LINC01354, methylation at the three sites showed significant negative correlations with mRNA expression, with correlation coefficients between −0.41 and −0.51 in tumor tissues and −0.50 and −0.59 in adjacent normal tissues. The authors are careful to note that these modest correlation coefficients and the observational nature of the analysis do not establish methylation-mediated silencing, even though the pattern is consistent with epigenetic regulation. Results for LINC01510 were less consistent in the atlas data, partly because suitable CpG sites were unavailable and expression differences, though trending in the same direction, failed to reach statistical significance.

An exploratory database analysis added a further layer of context. Querying the RNAInter database revealed that LINC01354 and LINC01510 share a large fraction of their reported interactions with proteins and epigenetically modified histones, and that all reported significant interaction partners of LINC01510 were also among the targets of LINC01354 and LINC00404. Pathway enrichment pointed to processes including gastrulation, granulopoiesis, SUMOylation and RUNX1 transcriptional regulation, and a substantial portion of the putative target proteins showed significant expression alterations, often losses, in high-stage or metastatic tumors in the atlas data. The authors emphasize that these analyses are hypothesis-generating only and require experimental validation.

The study’s conclusion is deliberately measured. The data demonstrate tumor- and metastasis-associated hypermethylation at the two lncRNA loci and support an association with established metastatic disease, but the less consistent differences between primary tumors from metastasis-positive and metastasis-negative patients mean the findings do not provide predictive values for metastatic risk. The authors also flag the observational design as a major limitation: the statistical relationships cannot establish causal links between DNA methylation, lncRNA expression and disease progression, and functional studies are required to determine the biological impact of methylation at these loci. Within those limits, the work extends the catalog of epigenetically deregulated non-coding RNA genes in clear cell renal cell carcinoma and suggests that DNA methylation at LINC01354 and LINC01510 may merit evaluation as a potential biomarker of aggressive disease in future prognostic studies.

Subject of Research: Metastasis-associated DNA methylation of the long non-coding RNA genes LINC01354 and LINC01510 in clear cell renal cell carcinoma

Article Title: Metastasis-associated DNA methylation of the lncRNAs LINC01354 and LINC01510 in clear cell renal cell carcinoma

Article References: Faraj Tabrizi, P., Schimansky, I. M., Peters, I., Kanaan, M. R., Hennenlotter, J., Franke, A., Tezval, H., Kuczyk, M. A., & Serth, J. (2026). Metastasis-associated DNA methylation of the lncRNAs LINC01354 and LINC01510 in clear cell renal cell carcinoma. Clinical Epigenetics, 18(1), Article 185. https://doi.org/10.1186/s13148-026-02252-4

Image Credits: AI Generated

DOI: 10.1186/s13148-026-02252-4

Keywords: clear cell renal cell carcinoma, DNA methylation, LINC01354, LINC01510, long non-coding RNA, metastasis, epigenetics, biomarker, pyrosequencing, TCGA-KIRC, kidney cancer, tumor stage