Formation of 5hmC following exposure to Ionizing Radiation

Abstract

Eukaryotic DNA methylation (DNAm), which is the addition of a methyl group on cytosine bases (C), is an epigenetic mechanism involved in several biological processes, including replication, transcription, and carcinogenesis. Besides methylcytosine (5mC), hydroxymethylcytosine (5hmC) is another modified C moiety. 5hmC is formed by the enzymatic activity of TET (ten eleven translocation) enzymes oxidizing 5mC to 5hmC and further to formyland carboxyl-cytosine, the latter being finally replaced by C via base excision repair [1]. Conversion of 5mC to 5hmC could change the DNA binding affinity of several proteins and thus influence regulation of gene transcription [2]. Previous works reported that ionizing radiation (IR) and DNA double strand break (DSB) repair processes are able to induce changes of DNAm levels [3, 4, 5]. However, these studies were performed several population doublings after IR exposure and could not rule out replication-dependent DNAm changes. Whether the latter may take place also at short times after exposure to IR remains controversial. Increased DNA methylase levels at later times (≥ 24 hours) after IR exposure have been recently reported [6]. These findings suggest that increased DNAm levels would not occur until DNA repair is accomplished. Based on such observation, we hypothesize that IR may rather reduce DNAm levels at earlier times by triggering enzymatic activities regulating such process, e.g. TET enzymes. Therefore, we examined DNAm changes within one replication cycle time frame after IR (≤ 24 hours), focusing on DNA demethylation. Namely, we investigated 5hmC formation as an intermediate of DNA demethylation and whether this modification was involved in the DSB repair. We investigated human fibroblasts 30 minutes after irradiation with carbon ions and observed 5hmC formation along the ion trajectory. Both fluorescence microscopy images and intensity profiles indicate that 5hmC signal colocalizes with γH2AX (DSBs marker) (Fig. 1). Similar results were obtained upon irradiation with X-rays (Fig. 2A). Colocalization was statistically verified using intensity correlation analysis (ICA) [7]. Briefly, PDM (Product of the Differences from the Mean) values were calculated for each pixel in both channels. If the signals were positively correlated, this would result in a positive PDM value and, thus, in a positive correlation. Our results show that most PDM values are positive, displaying a moderate positive correlation between 5hmC and γH2AX at DSBs (RPearson = 0.425, Fig. 2B).