# Assessment of references for the quantitative analysis of LINE-1 and Alu methylation in cellular DNA and circulating cell-free DNA of cancer patients

## TL;DR

This study shows that biased reference sequences can distort methylation measurements of LINE-1 and Alu elements in cancer DNA, affecting biomarker accuracy.

## Contribution

First demonstration of how biased MIP primers impact LINE-1 and Alu methylation profiles in cancer and cfDNA.

## Key findings

- Biased MIP primers cause significant shifts in methylation status of LINE-1 and Alu in cancer tissues and cfDNA.
- Unexpected shifts were observed in cfDNA even with unbiased MIP primers, depending on reference sequences.
- Impartial references are needed for accurate methylation quantitation of repetitive elements in cfDNA.

## Abstract

The LINE-1 and Alu retrotransposon elements, with more than 90% of their sequences being methylated, contribute to 30% of the human genome. Their hypomethylation profile, representing global methylation in cellular and cell-free DNA (cfDNA) from cancer, has been considered an attractive noninvasive biomarker of cancer. LINE-1 and Alu methylation profiling has preferentially been performed by real-time methylation-specific PCR (qMSP), pyrosequencing, and methylation-sensitive high-resolution melting (MS-HRM), which are bisulfite-based PCR approaches that require reference sequences amplified by the Methylation Independent PCR (MIP) primers to normalize the quantification data. A technical weakness of MIP primers is unequal amplification, termed PCR amplification bias, leading to an under- or overestimation of expected methylation levels, and thus, hindering the effectiveness of DNA methylation-based biomarkers. To date, the PCR amplification bias of MIP primers that may affect the methylation analysis of repeat sequences such as LINE-1 and Alu has not yet been described. Our study demonstrated for the first time the detrimental impact of biased MIP primers on LINE-1 and Alu methylation profiles, causing a significant shift from the hypomethylated status to hypermethylated in cancer tissues and in cfDNA from cancer patients. Unexpectedly, this shift was also observed in cfDNA, even when quantified by the unbiased MIP primers, depending on the reference sequences. Our results suggest that an impartial reference for the methylation quantitation of repetitive elements, most importantly in cfDNA, should be further established to ensure cross-platform consistencies in DNA methylation profiling through bisulfite-based PCR techniques.